Riley_washington_0250E_26512-html.html

Protein-level regulation of oncogenic RIT1 in non-small cell lung cancer

Amanda K. Riley

A dissertation

submitted in partial fulfillment of the

requirements for the degree of

Doctor of Philosophy

University of Washington

2024

Reading Committee:

Alice H. Berger, Chair

Susan Biggins

Lucas Sullivan

Program Authorized to Offer Degree:

Molecular and Cellular Biology

University of Washington

Abstract

Protein-level regulation of oncogenic RIT1 in non-small cell lung cancer

Amanda K. Riley

Chair of the Supervisory Committee:

Associate Professor Alice H. Berger

Human Biology Division, Fred Hutchinson Cancer Center

Lung cancer is the leading cause of cancer-related deaths worldwide. Non-small cell lung

cancer is the most diagnosed type of lung cancer, and lung adenocarcinoma is the most prevalent

subtype. Approximately 50% of lung adenocarcinoma tumors harbor druggable mutations in

genes such as

EGFR

and

ALK

, and targeted therapies are highly effective at reducing tumor

burden. Indeed, targeted therapies have revolutionized cancer treatment and are becoming

standard of care over cytotoxic chemotherapy; however, many mutations are not clinically

actionable. Up to 15% of lung adenocarcinoma tumors are driven by mutation or amplification of

the RAS-family gene

RIT1

, and

RIT1

mutations do not co-occur with other canonical driver

mutations. There is a growing understanding that the protein abundance of RIT1 is essential for

its function. Therefore, inhibiting positive regulators of RIT1 abundance could be a tractable

means of reducing tumor burden and abrogating the growth of tumors driven by

RIT1

mutations

and amplifications.

Development of a RIT1-specific inhibitor is unlikely to succeed due to the structure of

RIT1 as a GTPase. In 2013, groundbreaking work on KRAS resulted in the development of the

first mutant-specific inhibitors, which represents a major advance in this field and for patient

care. Such an approach for RIT1, however, would be quite difficult due to the resources required

and our lack of knowledge pertaining to RIT1 biology and oncogenic mechanisms. Because of

this, innovative approaches are needed to understand RIT1 genetic dependencies and uncover

druggable targets.

The Berger Lab developed a CRISPR screening approach to discover genes required for

RIT1

-driven cellular transformation. From this work, we found that

RIT1

-mutant cells are

uniquely dependent on genes associated with the Spindle Assembly Checkpoint (SAC),

including Aurora kinases A and B.

RIT1

-mutant cells are more sensitive than

KRAS

-mutant cells

to alisertib (an Aurora kinase A inhibitor) and barasertib (an Aurora kinase B inhibitor).

Expression of mutant

RIT1

weakens the SAC, prompting cells to prematurely exit mitosis and

accumulate mitotic abnormalities. In addition to the SAC vulnerability, we identified the

deubiquitinase

USP9X

as a top essential gene in

RIT1

-mutant cells. This was particularly

intriguing given that previous work has suggested that the protein abundance of RIT1 is

important for its function. Indeed, although RIT1 shows high sequence homology to KRAS,

RIT1 does not appear to be regulated in a similar manner (i.e. at the level of GAP resistance).

Instead, RIT1 appears to be regulated at the level of protein abundance.

Here, I explore the hypothesis that RIT1 is a substrate of USP9X and found that USP9X

binds to and deubiquitinates RIT1. I find that USP9X depletion decreases RIT1 protein

abundance and stability, and loss of USP9X abrogates

RIT1

-driven cell growth and proliferation.

vii

ACKNOWLEDGEMENTS

I want to thank current and former members of the Berger Lab for helping me design this

project, conceive of and carry out experiments, and interpret the analyses presented in this

dissertation. I first met Dr. Alice Berger during the interview weekend for the MCB program,

and I was inspired by her enthusiasm and excitement for lung cancer genomics. Throughout my

PhD, Alice’s support and guidance has helped me become a better scientist. When I entered

graduate school, I didn’t have many lab skills beyond tissue culture, and being in Alice’s lab

provided me with opportunities to learn an incredible array of techniques and skills. I also want

to thank a former postdoc in the lab—Dr. Athea Vichas—who mentored me during my rotation

and provided many useful reagents and cell lines for the USP9X experiments.

Thank you to my former mentors—Dr. Marten Edwards, Dr. Amy Hark, and Dr.

Elizabeth McCain—from Muhlenberg College, who guided me throughout my undergraduate

training. The foundational knowledge that I obtained in biology helped me build upon this

knowledge in graduate school. Thank you to Dr. Cyril Benes for being a wonderful PI while I

was a research technician at the Massachusetts General Hospital in Boston. In his lab, I

discovered my interest in cancer research, and I gained so much valuable insight about how

science is done in this field.

Thank you to the MCB program, particularly Maia Low, for their unwavering support

and for responding so promptly to all of my emails over the years. Also thank you to all of the

admins and IT support at Fred Hutch, especially Luna Yu for fixing nearly all of my computer

troubles throughout graduate school.

Finally, thank you to all of my incredible friends and partners. I could not have

completed this PhD without your love and support.

Chapter 1.

INTRODUCTION

A version of this chapter will be submitted in 2024 as a review article to the journal

Oncogene

1.1

URRENT STATE OF LUNG CANCER RESEARCH

Lung cancer is the most common cause of cancer-related deaths worldwide (1).

Approximately 75% of lung cancer cases are caused by mutations in the Receptor Tyrosine

Kinase (RTK)/RAS signaling pathway (2). Mutations within this pathway—in genes such as

EGFR

(epidermal growth factor receptor) and

KRAS

—are mutually exclusive, meaning that

tumors are often driven by one key oncogene (2). In the past decade, targeted therapies have

revolutionized cancer treatment by offering a means to kill cancer cells that harbor specific

oncogenic mutations. Targeted therapies are more effective and less toxic than non-selective,

cytotoxic chemotherapy. Despite this progress, however, there are still oncogenes that cannot be

targeted with currently available inhibitors.

1.2

RIT1

IN CANCER

Over the past 40 years, Ras guanosine triphosphate hydrolases (GTPases) have been

characterized as potent oncogenes through their perturbation of the mitogen activated protein

kinase (MAPK) signaling pathway (3–6). In addition to classical RAS proteins such as KRAS

and NRAS, many other RAS-related proteins have been identified, including RIT1 (Ras-like in

all tissues) (7–9).

RIT1

was first discovered in a study of gene expression from mouse retinas (8).

As its name implies,

RIT1

is expressed across all tissues in embryos and adults, although it is

more highly expressed in adult animals (8). In 2014,

RIT1

was identified as an oncogene in lung

adenocarcinoma, and to date

RIT1

mutations and amplifications have been linked to several other

cancer types, including myeloid malignancies and endometrial cancer (9–11). In the context of

lung cancer,

RIT1

is considered to be a “rare” oncogene since it occurs in <5% of cases (12).

Given that close to 240,000 people will be diagnosed with lung cancer in 2023, even rare

subtypes affect thousands of patients (13). Because of this and given the prevalence of RIT1 in

other cancer types, further understanding RIT1 biology and oncogenic mechanisms is crucial for

improving treatment options and patient outcomes.

1.3

OMPARISON OF

RIT1

TO OTHER

RAS

PROTEINS

Figure 1.1

. Venn diagram illustrating RIT1 functions that are similar to RAS and those that are

unique to RIT1.

Figure created with Biorender.com

RIT1 shows high sequence homology to other RAS proteins such as KRAS, NRAS, and

HRAS (14). Expression of mutant

RIT1

(9) and mutant

RAS

(15–17) transforms cells, primarily

via activation of PI3K/MAPK signaling to promote cell proliferation (9,18–21). Furthermore,

RIT1

and

KRAS

mutants both confer resistance to EGFR tyrosine kinase inhibitors such as

erlotinib and osimertinib (18,22). This suggests that RIT1 functions within the MAPK signaling

pathway, but the exact role of RIT1 within this pathway appears to be cell-type and context-

dependent.

RIT1

-mutant NCI-H2110 cells, knockdown of

RIT1

reduces phosphorylation of AKT,

MEK, and ERK, thereby suggesting that RIT1 plays an important role in the PI3K and MEK

pathways in these cells (9). Similarly, in HEK293T cells, expression of

RIT1

mutants increases

activation of ERK1/2 (23). Expressing RIT1

M90I

in immortalized lung epithelial cells increases

MEK phosphorylation in some, but not all, cell lines (9). Some groups have found that

expression of mutant

RIT1

in NIH3T3 cells does not induce phosphorylation of AKT, MEK, or

ERK (18,24). Interestingly, co-expression of RIT1 and RIN (a neuron-specific form of RIT1)

induces colony formation in NIH3T3 cells, but the exact mechanism of this is not fully

understood (8,25). Given this variability, there is a clear need to further study RIT1 and

understand the context- and cell-line dependent factors that govern its function.

Similarly to RAS proteins, RIT1 binds GDP and GTP to cycle between inactive and

active forms, respectively (7). In cells not stimulated with mitogenic signals, most cellular RIT1

is bound to GTP (26). This is in contrast to KRAS, which is predominantly bound to GDP in

cells not stimulated with growth factors (3). High abundance of GDP-bound KRAS is mediated

through the action of GAPs (GTPase-activating proteins) (3), but GAPs and GEFs (guanine

nucleotide exchange factors) that regulate RIT1 have yet to be identified. RIT1 mutants have

similar or increased GTP loading compared to wild-type RIT1, but there is not a clear connection

between the GTP-loaded state of RIT1 and its oncogenic function (26). This is in contrast to

KRAS mutations, which tend to occur in GTP-binding regions of the protein, thereby blocking

GAP activity and increasing the GTP-to-GDP ratio (3). RIT1 mutations, however, cluster in the

switch II domain of the protein, which is important for protein-protein interactions (9).

Over 90% of oncogenic RAS mutations occur in “hot-spot” regions, defined at Gly

Gly

, and Gln

(27). These mutants confer GAP resistance, meaning that GTP hydrolysis is

prevented and RAS remains in the active GTP-bound state (28,29). The RAS Q61 site is

homologous to Q79 in RIT1 (7). Experimentally engineered RIT1 Q79L transforms NIH3T3

cells (30) and activates MAPK signaling (31,32). However, Q79L is not observed in patient

tumors because the Q79 codon occurs at a splice site, so the mutation would disrupt the splice

site (9). G12 and G13 of RAS are homologous to G30 and G31 in RIT1, but to date no G30 or

G31

RIT1

mutations have been found in cancer. Ectopic expression of G31R activates MAPK

signaling in HEK293T cells, as does the G30V mutation, but to a lesser extent (23,33).

Of the currently identified

RIT1

mutations, the M90I variant (RIT1

M90I

) is the most

recurrent RIT1 variant in cancer (9). Instead of GAP resistance, mutant RIT1 appears to alter

RIT1 regulation via dysregulation of RIT1 protein abundance (26,34,35). Wild-type RIT1 is

subject to polyubiquitination and proteasomal degradation by the cullin 3 RING E3 ubiquitin

ligase (CUL3) and the adaptor protein LZTR1 (leucine zipper-like transcription regulator 1) (26).

Mutant forms of RIT1 evade this regulation, thereby increasing the protein abundance of RIT1

(26). Increased RIT1 protein abundance has been linked to Noonan syndrome (26) and blood

cancers (34). The importance of protein abundance, as opposed to GAP resistance, appears to be

a key difference in RIT1 and RAS and likely partially explains why GAPs and GEFs that

regulate RIT1 have not been identified. The link between RIT1 protein abundance and oncogenic

phenotypes is further discussed in

Section 1.7.1

The G domain of RIT1 and RAS share 51% sequence homology, suggesting that RIT1

and RAS could interact with similar effector proteins (36). RAS family members show varying

degrees of binding affinity to RASSF (RAS association domain family) proteins, which are

involved in many different signaling pathways, including cytoskeleton dynamics and cell cycle

progression (37,38). RASSF1 and RASSF5 show high affinity interactions with RAS, but not

with RIT1 (39). Instead, RIT1 preferentially binds to RASSF7 and RASSF9 (39).

Of the currently identified RAS effector proteins, RAF kinases—and their effects on

downstream signaling—have been extensively studied (40,41). At the plasma membrane, it is

well-known that KRAS interacts with RAF kinases to mediate downstream signaling (42). RAF

proteins (ARAF, BRAF, and CRAF/RAF1) contain a high-affinity RAS-binding domain (RBD)

that is essential for interaction with RAS and activation of downstream signaling (40,41). Unlike

other RAS proteins, RIT1 lacks a carboxy-terminal CAAX prenylation motif that is used by RAS

for plasma membrane tethering (8). Despite this, the C-terminal hypervariable region (HVR) of

RIT1 mediates plasma membrane localization through charge complementarity (8,43–46).

Deletion of the C terminus–but not the N terminus–disrupts this plasma membrane interaction

(44). Previous work suggested that RIT1 was unable to bind RAF proteins (7,39), but others have

found that RIT1-GTP can weakly bind BRAF (26,31,47). Comprehensive characterization of

RIT1 and RAF kinases revealed that RIT1 preferentially binds to CRAF (RAF1) over ARAF or

BRAF (44). The C terminus of RIT1 is essential its interaction with CRAF (44).

Mutant forms of RIT1 accumulate at the plasma membrane, which promotes recruitment

of RAF kinases and stimulates MAPK signaling (26,44). Pharmacological inhibition of RAF

abrogates this RIT1-driven MAPK signaling (44). Even the constitutively active RIT1

Q79L

mutant–structurally analogous to RAS

Q61L

–is unable to activate MAPK signaling in the absence

of RAF binding (44). This suggests that RIT1-RAF binding is essential for downstream

mitogenic signaling. Furthermore, RIT1 is unable to activate MAPK in the absence of RAS (44).

These data suggest a model whereby RIT1-RAF binding drives RAS-RAF interactions and

activates RAS signaling (44). Together, these findings suggest that RIT1 and RAS signaling is

highly intertwined.

The Berger Lab has employed methods to systematically compare and contrast the

biology of

KRAS

-driven versus

RIT1

-driven lung adenocarcinoma in cell line models (48). Using

human lung epithelial cells expressing

KRAS

and

RIT1

variants, we profiled the transcriptome,

proteome, and phosphoproteome (48). As expected,

KRAS

- and

RIT1

-mutant cells exhibited

increased phosphorylation of RAS/MAPK signaling effectors, including MEK and ERK (48).

Overall, the global proteome in

KRAS

- and

RIT1

-mutant cells were largely similar (48).

However,

KRAS

-mutant cells showed decreased phosphorylation of splicing factor proteins,

whereas these proteins were largely unchanged in

RIT1

-mutant cells (49). Interestingly, over-

expression of wild-type

RIT1

phenocopied mutant

RIT1

, whereas over-expression of wild-type

KRAS

was not phenotypically similar to mutant

KRAS

(48). This suggests that wild-type

RIT1

amplification and overexpression could result in a similar phenotype to mutant

RIT1

expression.

In addition to the transcriptome and proteome analyses described above, the Berger Lab

has employed genomics methods to compare genetic dependencies of

RIT1

- and

KRAS

-mutant

cells. Genome-wide CRISPR screening in

KRAS

- and

RIT1

-mutant isogenic lung

adenocarcinoma cells revealed that

RIT1

-mutant cells–but not

KRAS

-mutant cells–are dependent

on genes involved in the Spindle Assembly Checkpoint (SAC), such as Aurora kinase A

(AURKA) and Aurora kinase B (AURKB) (18). This dependency renders

RIT1

-mutant cells

uniquely sensitive to Aurora kinase inhibitors (18). Many of the SAC genes were not significant

dependencies in

KRAS

-mutant cells, and

KRAS

-mutant cells were not as sensitive to AURKA

and AURKB inhibition (18). I was involved in analysis of the CRISPR screen data and

performed experiments to explore RIT1’s perturbation of mitosis and dependency on SAC genes.

These efforts and my contributions are further discussed in

Section 2.2

In summary, RIT1 and RAS are GTPase proteins with some overlapping functions in

terms of effector molecules and activation of downstream signaling pathways (

Figure 1.1

However, the mutational landscape of RIT1 and RAS are unique and likely reflect key

differences in functional outcomes. RAS is regulated at the level of GAP resistance, whereas the

protein abundance of RIT1 appears to be important for its function. As we gain more knowledge

about RIT1 regulation, we are uncovering new genetic vulnerabilities and nominating druggable

genes for the development of targeted therapies.

1.4

PPROACHES TO STUDYING

RIT1

Studying the role of RIT1 in cancer has become an important question, not only to

understand the biology underlying

RIT1

-driven oncogenesis but also to identify better treatment

options. Answering these questions, however, poses a major challenge due to the lack of cell line

and mouse models. Out of all commercially available lung adenocarcinoma cell lines, only

one—NCI-H2110—harbors an endogenous RIT1

M90I

mutation (9). Because of this, innovative

approaches are needed to better understand RIT1 biology and oncogenic mechanisms.

The Berger Lab developed a unique genome-wide screening approach to identify genes

required for RIT1 function (18,50). This screen took advantage of the observation that

expression of RIT1

M90I

confers resistance to EGFR inhibitors in PC9 lung adenocarcinoma cells

(18,22). PC9 cells harbor an

EGFR

mutation that renders them dependent on EGFR for survival

and thus sensitive to EGFR tyrosine kinase inhibitors such as erlotinib and osimertinib (51,52).

In this system, the drug resistance phenotype is directed related to RIT1 function, thus

facilitating the discovery of essential genes and cooperating factors in

RIT1

-mutant cells. This

method is robust, but there are some notable drawbacks.

RIT1

mutations are almost always

mutually exclusive with other mutations in the RTK/RAS pathway, including

EGFR

(2).

Because of this, the physiological relevance of the

EGFR

-mutant PC9 system might be

questioned. However, it is important to note that this PC9 system is similar to the widely used

Ba/F3 model system (53). Ba/F3 cells are a mouse pro-B cell line dependent on interleukin 3

(IL-3) for survival (53). Expression of a driver oncogenic mutation—such as

EGFR

—renders

Ba/F3 cells IL-3 independent (53). This system has been used since the late 1980s to understand

the biology of driver genes and test sensitivity to targeted therapies (54). The Berger Lab’s PC9-

based system is analogous to the Ba/F3 model since expression of oncogenes renders PC9 cells

resistance to EGFR inhibitors. PC9 cells, however, are a human lung cell line and thus more

physiologically relevant than the mouse-based Ba/F3 system. As such, the PC9 system is a useful

and valuable system for understanding RIT1 genetic dependencies.

Another challenge to studying RIT1 has been the lack of mouse models. Substantial

progress was made in 2019, when a RIT1

M90I

-mutant mouse model was generated via germline

knock-in of the mutant allele (26). Heterozygous RIT1

M90I

-mutant mice showed phenotypes

resembling Noonan syndrome, including craniofacial abnormalities and increased heart weight-

to-body weight ratios (26). In mouse embryonic fibroblast (MEF) cell lines established from

these mice, RIT1

M90I

-mutant cells stimulated with fetal bovine serum showed increased

phosphorylation of MEK1/2 and ERK1/2, as well as increased mRNA expression of MAPK-

related transcription factors

Dusp6

and

Spry2

(26). The establishment of this mouse model

provided key insight into RIT1 disease mechanisms in the context of Noonan syndrome.

In 2022, a leukemia mouse model was established to investigate the mechanisms of

LZTR1

and

RIT1

mutations in hematological malignancies (11,34,55). In embryos with complete

loss of LZTR1, death occurred between embryonic day 17.5 (E17.5) and birth (34). Cells derived

from E14.5

LZTR1

knockout (KO) liver cells showed increased expression of RIT1 and RAS

proteins, including KRAS (34). Upon transplantation of hematopoietic

LZTR1

KO cells into

animals expressing wild-type

LZTR1

, the KO cells eventually out-competed WT cells, and

animals developed hematologic malignancies, including myeloproliferative neoplasms (MPN)

(34). Expression of RIT1

M90I

in hematopoietic cells phenocopies what is observed upon NRAS-

and KRAS-expression, notably enhanced granulocyte-macrophage colony-stimulating factor

(GM-CSF) colony formation (26,56,57). Both loss of LZTR1 and expression of RIT1

M90I

increased the proliferation of hematopoietic stem/progenitor cells (HSPC) (34). In myeloid cells

derived from this mouse model,

LZTR1

deletion and RIT1

M90I

expression resulted in similar gene

expression profiles and reduced expression of many tumor suppressor genes such as

EZH2

(34).

After six weeks, induction of RIT1

M90I

caused MPN and other myeloid neoplasms (34). This

mouse model represents the first

in vivo

model to study

RIT1

-induced transformation.

The Noonan syndrome and leukemia mouse models described above represent major

advances in our understanding of RIT1 disease mechanisms

in vivo

. More models such as these

are needed to understand the mechanism(s) of RIT1 in other cancer types, such as non-small cell

lung cancer. There is a need to study both the effects of

RIT1

amplifications and mutations in

murine models. With a lung cancer model of

RIT1

-driven cancer, we would be able to test

genetic dependencies found

in vitro

and assess the efficacy of proposed targeted therapies.

1.5

RIT1

MUTATIONS IN CANCER

When

RIT1

mutations were first discovered, they were primarily studied in the context of

the developmental disease Noonan syndrome (NS). In NS patients, germline

RIT1

mutations

generally do not increase an individual’s risk of developing cancer (58). In rare cases, juvenile

myelomonocytic leukemia has been reported in patients with

RIT1

-associated NS (58). Beyond

NS, most of our knowledge pertaining to

RIT1

mutations in adult cancers has come from tumor

sequencing studies from the past decade.

Whole-exome sequencing of lung adenocarcinoma (LUAD) tumors in The Cancer

Genome Atlas Research Network (TCGA) study revealed that

RIT1

was mutated in 2.2% of

tumors (2). These alterations did not co-occur with other common LUAD driver genes such as

KRAS

, suggesting that

RIT1

itself drives oncogenesis (2). A follow-up study found that

expression of mutant

RIT1

transforms cells, and injection of

RIT1

-mutant cells in mice induces

tumor formation (9). This study marked the first in-depth analysis of

RIT1

as an oncogene. Since

then,

RIT1

mutations have been studied in several different cancer types.

RIT1

mutations have been found in numerous types of myeloid malignancies (11,34,59).

This was first discovered in a next-generation sequencing study of patients with

myelodysplastic/myeloproliferative neoplasms (MDS/MPN), including chronic myelomonocytic

leukemia (CMML) (11).

RIT1

mutations (predominantly at the F82 locus, but also at M90 and

E81) were found in 7% of cases (11). In a retrospective sequencing study of patients with

myeloid neoplasms,

RIT1

mutations were detected in approximately 15% of MDS/MPN cases

(59). Treatment options for these disorders are currently very limited and primarily focused on

symptom management (60).

In rare cases,

RIT1

mutations have been found in the context of resistance to tyrosine

kinase inhibitors (61). In an analysis of

ALK

-positive tumors in patients that had progressed on

ALK inhibitor therapy, an acquired RIT1 K139N mutation was detected in one patient (61). This

mutation has not yet been characterized, and it is unclear how it may affect RIT1 activity.

In parallel with the identification of

RIT1

mutations, many studies have also observed

RIT1

amplifications (

Figure 1.2

). In a study of endometrial cancer,

RIT1

mutations (including

missense mutations and in-frame mutations) were found in ~1% of cases, while

RIT1

amplifications were documented in 10% of cases (10). In the TCGA cohort,

RIT1

amplifications

were found in 14% of NSCLC tumors (2). In the leukemia study described above,

RIT1

amplifications were found in 4% of cases (11). In the cohort of ALK inhibitor-resistant tumors,

one sample showed increased copy number variation of

RIT1

(61). Cases of

RIT1

amplifications

are not confined to these specific studies; instead, there is a growing understanding that

RIT1

copy number changes are likely oncogenic.

Figure 1.2

. Landscape of

RIT1

alterations in cancer.

Approximate prevalence of

RIT1

mutations and amplifications across several different cancer

types. Percentages are based on genome sequencing studies (2,10,11,62–64).

1.6

RIT1

AMPLIFICATIONS IN CANCER

RIT1

is overexpressed in several cancer types, including adrenocortical carcinoma,

hepatocellular carcinoma, glioblastoma, and myeloid malignancies (11,62,63,65,66). Tumors

with elevated

RIT1

expression show increased activation of RAS/MAPK pathway genes (62,67).

In studies of hepatocellular carcinoma, copy number amplification of

RIT1

was found in 13% of

patients, which is a higher frequency than any other RAS family member in this cancer type (62).

Studies in patients with oral squamous cell carcinoma revealed that upregulation of

RIT1

was

found in 15% of patients who smoked tobacco products (64). In cases of

RIT1

amplification,

both

RIT1

mRNA and protein abundance are upregulated (10,62).

In most cases, high expression of

RIT1

is associated with poor prognosis (10,64–66,68).

Overexpression of

RIT1

in glioblastoma cell lines caused increased proliferation, invasion, and

colony formation compared to control groups (67). Furthermore,

RIT1

expression is associated

with poor survival in lower grade gliomas (68). Subcutaneous injection of

RIT1

-depleted

glioblastoma cells abrogated tumor formation in nude mice, suggesting that RIT1 is important

for driving tumor growth in these cells (67). Conversely, in kidney renal clear cell carcinoma,

RIT1

expression is correlated with better overall survival (68). In esophageal squamous cell

carcinoma,

RIT1

expression is associated with lower levels of MAPK signaling and better

prognosis (69). To date, these two studies are the only instances of

RIT1

acting as a tumor

suppressor.

In the field of RIT1 biology, we are beginning to appreciate that

RIT1

amplifications,

similarly to

RIT1

mutations, are oncogenic. Global proteome and phosphoproteome analyses

revealed that overexpression of RIT1

and expression of RIT1

M90I

phenocopied each other in

terms of down- and up-regulated proteins (48). Conversely, over-expression of wild-type

KRAS

did not phenocopy mutant

KRAS

(48). As such, this suggests that the abundance of RIT1 protein

is related to unique mechanisms of RIT1 oncogenesis. The relevance and data supporting this

hypothesis are further discussed in

Section 1.7.1

1.7

RIT1

ONCOGENIC MECHANISMS

The studies discussed above support an oncogenic role for RIT1, but the exact

mechanisms that lead to transformation and tumor growth are not clear. Despite this, recurrent

themes have been emerging. Note that the mechanisms discussed in this section could occur

simultaneously or only in certain cellular contexts, and future work is needed to understand

context-specific states.

1.7.1

Role of protein abundance

In 2019, evidence emerged that the protein abundance of RIT1 appeared to be crucial for

its function (26). This study found that wild-type RIT1 binds to leucine zipper-like transcription

regulator 1 (LZTR1), an adaptor for the cullin 3 RING E3 ligase (70). This binding promotes

polyubiquitination and degradation of wild-type RIT1 through the ubiquitin-proteasome system

(26). Mutations in

LZTR1

and in

RIT1

, including RIT1

M90I

, disrupt interaction with LZTR1,

thereby preventing RIT1 degradation (26). In addition to targeting RIT1, LZTR1 has been shown

to bind other RAS proteins, including KRAS (71–73). However, LZTR1 preferentially binds

RIT1 over other RAS proteins (35). In primary skin fibroblasts derived from children with

homozygous

LZTR1

mutations and diagnosed Noonan syndrome, RIT1 protein abundance was

markedly increased compared to fibroblasts with heterozygous

LZTR1

mutations (26). This was

the first study to link the protein abundance of RIT1 to a disease state.

Beyond Noonan syndrome, the protein abundance of RIT1 has also been implicated in

driving tumor growth. In human myeloid leukemia cells, expression of mutant

RIT1

(including

the M90I variant) resulted in increased RIT1 protein abundance—via evasion of LZTR1-

mediated degradation—and increased phosphorylation of MEK1/2 and ERK1/2 (34). Pathogenic

levels of RIT1 are thus important for maintaining and perhaps initiating disease states. Given the

prevalence of

RIT1

mutations in lung adenocarcinoma, it is also important to understand how

RIT1 oncoproteins are regulated in lung cancer cells.

In analyzing data from the PC9 lung adenocarcinoma cell-based CRISPR screen, I found

that the deubiquitinase

USP9X

emerged as a top genetic dependency in

RIT1

-mutant cells (18).

This was an intriguing finding and suggested that RIT1 could be a substrate of USP9X. To

follow up on this hypothesis, a recent preprint from the Berger Lab investigated the role of

USP9X in

RIT1

-mutant cancers (74). This study is a key component of this dissertation and is

further discussed in

Chapter 3

. In brief, loss of

USP9X

abrogated proliferation and anchorage-

independent growth of

RIT1

-mutant cells and resensitized these cells to EGFR inhibition (74).

USP9X

knockout reduced the abundance of both wild-type and mutant RIT1, suggesting that

USP9X targets multiple forms of RIT1 (74). This work suggests that USP9X and LZTR1 are

opposing the action of one another on the regulation of wild-type RIT1, while USP9X and a

currently unknown E3 ligase are targeting mutant RIT1 (74). Based on these findings and the

importance of protein abundance for RIT1 oncogenesis, USP9X could be a promising

therapeutic target in

RIT1

-driven diseases. Future work is needed to assess the preclinical utility

of USP9X inhibitors in NSCLC and other cancers characterized by

RIT1

mutations and

amplifications.

1.7.2

Perturbation of mitosis

Although genome instability is considered to be a hallmark of cancer (75), it is not

completely clear how genomic changes such as aneuploidy contribute to oncogenesis.

Chromosome and whole genome duplications acquired during aberrant cell division can hinder

cell growth due to accumulation of deleterious mutations; however, these mutations can also

promote cell growth and oncogenic mechanisms (76–79). In the context of

RIT1

-mutant cancers,

several groups have found that RIT1 perturbs mitosis (18,80,81), but the exact contribution of

this to oncogenesis has yet to be elucidated.

The Berger Lab’s CRISPR screen revealed that RIT1

M90I

-mutant cells are uniquely

dependent on cell cycle genes, including Aurora kinase A (

AURKA

) and Aurora kinase B

(

AURKB

) (18). Genetic knockout of

AURKA

AURKB

RIT1

-mutant–but not

KRAS

-mutant–

lung adenocarcinoma cells hindered proliferation and anchorage-independent growth (18).

Furthermore,

RIT1

-mutant cells were significantly more sensitive to alisertib (an AURKA

inhibitor) and barasertib (an AURKB inhibitor) compared to

KRAS

-mutant cells (18). Expression

of RIT1

M90I

in HeLa cells resulted in faster progression through mitosis and increased

chromosomal abnormalities, suggesting that cells were bypassing the Spindle Assembly

Checkpoint (SAC) (18). Based on these results, it appears as though RIT1

M90I

weakens the SAC,

allowing cells to aberrantly pass through this checkpoint. Because of this, it has been

hypothesized that RIT1

M90I

-mutant cells are dependent on AURKA and AURKB to maintain

sufficient mitotic fidelity, thereby resulting in heightened vulnerability to Aurora kinase

inhibitors (18). Specific data pertaining to these results are further discussed in

Section 2.2

In parallel with the Berger Lab’s findings, Dr. Frank McCormick’s group also

demonstrated that RIT1 weakens the SAC (80). Comprehensive biochemical analyses revealed

that RIT1 physically interacts with MAD2 and p31

comet

, components of the Spindle Assembly

Checkpoint (80). During mitosis, MAD2 amplifies the signal at unattached kinetochores to

promote formation of the Mitotic Checkpoint Complex (MCC) (82). The MCC inhibits the

anaphase-promoting complex/cyclosome (APC/C) to prevent cells from progressing through

mitosis before all chromosomes are properly aligned (83). P31

comet

promotes the removal of

MAD2 from kinetochores and dissociation of the MCC (84). It is thought that this RIT1-MAD2

binding sequesters MAD2 away from its role in the MCC and accelerates progression through

mitosis (80). The physical interaction of RIT1 with MAD2 and p31

comet

is not affected by the

GTP-loaded state or mutational status of RIT1 (80). Importantly, these interactions are unique to

RIT1 and are not observed with other RAS proteins (80).

As cells enter mitosis, RIT1 diffuses into the cytoplasm and is phosphorylated by CDK1,

which prevents interaction with p31

comet

and MAD2 (80). However, when RIT1 is expressed at

pathogenic levels, CDK1 is not able to phosphorylate all RIT1 molecules present, and complexes

of RIT1/MAD2/p31

comet

are formed (80). This mechanism provides additional, important context

for further understanding how the protein abundance of RIT1 contributes to its pathogenic

function. This was first explored in the context of lung adenocarcinoma, but RIT1’s perturbation

of mitosis has also been observed in other cancer types.

In hepatocellular carcinoma (HCC) cells, RIT1 localizes to chromosomes during mitosis

(81). Interestingly, in these cells, evidence for physical interaction of RIT1 and MAD2 has not

been detected (81). Instead, RIT1 was found to interact with other proteins important for

chromosome separation, including SMC3 (structural maintenance of chromosome 3) (81). SMC3

must be acetylated in order to ensure proper progression through mitosis (85). During mitosis,

RIT1 interacts with PDS5 (precocious dissociation of sisters 5), which maintains SMC5

acetylation (81,86). RIT1’s interaction with SMC3 appears to be vital to proper progression

through mitosis, for

RIT1

knockdown arrests HCC cells in G2/M and causes mitotic

abnormalities (81). Based on these studies, RIT1’s regulation of mitosis may be context and cell-

type dependent. Future work is needed to understand how RIT1 affects the cell cycle in various

disease states.

Although it is not exactly clear how RIT1’s perturbation of mitosis is linked to

oncogenesis, it has been hypothesized that this could be related to other work pointing towards

RIT1’s ability to evade the cellular stress response, as discussed in

Section 1.7.3

1.7.3

Oxidative stress response

Increased oxidative stress, due to excessive reactive oxygen species (ROS), can cause

cellular damage that promotes or maintains disease states (87). In response to ROS, signaling

pathways can lead to cell survival or cell death (88). MEFs derived from

RIT1

knockout mice

showed increased apoptosis in response to ROS accumulation (89). Notably, treatment with a

MEK/ERK inhibitor did not affect the survival of

RIT1

-mutant cells exposed to hydrogen

peroxide, but these cells were potently sensitive to a p38 inhibitor (89). This suggests that the

p38 cascade is important in

RIT1

-driven oxidative stress response and survival.

P38 is known to activate pro-survival signaling through numerous pathways, including

via AKT (90–92).

RIT1

-driven AKT activation can occur via RIT1’s interaction with mTORC2

(mammalian target of rapamycin) (93). Knockdown of mTORC2 blocks AKT phosphorylation in

RIT1

-mutant cells, rendering cells more vulnerable to ROS-induced cell death (93). Although

other RAS proteins are known to signal through AKT, ROS-related RAS signaling appears to be

PI3K-dependent (94,95). In contrast, RIT1-mediated AKT signaling in the stress response

appears to uniquely occur through p38 (93,96). This work suggests that p38 inhibitors may be a

viable treatment option for patients with

RIT1

-driven cancers, but future work is needed to

understand the risks and benefits of such a treatment regimen (96).

The involvement of RIT1 in the p38 signaling cascade is not unique to human cells. In

mouse hippocampal neurons, mutant

RIT1

expression improves cell survival upon exposure to

hydrogen peroxide, which is consistent with what has been observed in human cell lines as

discussed above (97). This group also found that p38 inhibitors, but not MEK/ERK inhibitors,

promoted ROS-induced cell death in

RIT1

-mutant cells (97). Together, these results suggest that

RIT1 regulates the oxidative stress response in numerous cellular contexts.

The connection between RIT1-mediated oxidative stress response and oncogenesis is not

completely clear, but it may be related to the mitotic phenotypes described in

Section 1.7.2

RIT1’s ability to promote progression through the SAC (18,80) could be important for promoting

cell survival, even in the presence of oxidative stress. Interestingly, Aurora kinase A can become

hyperphosphorylated and inactivated by ROS, which negatively affects spindle assembly and can

cause mitotic delay (98). This suggests a model in which perhaps ROS accumulation inactivates

a portion of Aurora kinase A molecules within the cell, so

RIT1

-mutant cells are more dependent

on the active Aurora kinase A still present. This hypothesis, and the role of Aurora kinase B in

this model, requires future experimentation.

1.7.4

Cellular invasion & epithelial-to-mesenchymal transition

Inducing vasculature formation is a well-known hallmark of cancer, and there is evidence

that RIT1 can affect this process (75). BALB/C mice injected with

RIT1

-overexpressing liver

cells showed higher density of capillaries compared to control conditions (65). Additional

experiments in this HCC mouse model revealed that RIT1 was able to regulate tumor vasculature

and promote metastasis through activation of VEGFA (vascular endothelial growth factor A)

(62). Evidence for

RIT1

-driven regulation of vasculature has also been found in endometrial

cancer, where high

RIT1

expression is associated with increased vascular invasion (10).

Analysis of the global proteome of

RIT1

-mutant cells revealed that epithelial-to-

mesenchymal transition (EMT) pathway genes were the most significantly up-regulated

pathways compared to parental cells (48). Both RIT1

- and RIT1

M90I

-overexpressing lung

epithelial cells showed heightened migration capacity in scratch assays (48). This is consistent

with work showing that RIT1-induced MEK/ERK signaling promoted migration and outgrowth

of human neuronal cells (99). In liver cancer cells,

RIT1

overexpression promoted cell migration

and invasion (100). In A549 lung cancer cells, knockout of

LZTR1

promoted cell invasion, while

RIT1

knockdown suppressed invasion (101). Together, these studies suggest that increased RIT1

protein abundance is important for promoting and modulating EMT phenotypes.

Figure 1.3

. Summary of currently known and proposed RIT1 oncogenic mechanisms.

Figure created with Biorender.com.

1.7.5

RIT1/YAP synergy

In addition to understanding genetic dependencies, the Berger Lab’s CRISPR screening

approach identified unique synergies in

RIT1

-driven lung adenocarcinoma cells. Most strikingly,

activation of YAP1 (yes-associated protein 1) or loss of components of the Hippo signaling

pathway synergized with RIT1

M90I

to promote proliferation and xenograft tumor formation (18).

Hippo signaling is a well-studied tumor suppressive pathway, but it had not previously been

linked to RIT1 biology (102). Analysis of TCGA data revealed that

RIT1

alterations co-occurred

with loss of at least one Hippo pathway gene in 75% of cases (18). Additionally,

immunohistochemistry analysis of

YAP1

expression in patient tumors revealed that

RIT1

-mutant

tumors had significantly more nuclear-localized (i.e. activated) YAP1 (18). The mechanisms

underlying this RIT1/YAP synergy is an area of ongoing research, but this holds significant

promise for the potential use of YAP inhibitors for the treatment of

RIT1

-driven diseases.

1.8

TTEMPTS TO IDENTIFY DRUGGABLE TARGETS

To date, no RIT1 inhibitors have been developed. This is due in part to the structure of

RIT1, which–as a GTPase–is difficult to drug due to the lack of ATP binding pockets (103).

Breakthrough developments have been made to target KRAS

G12C

via covalent reactions with the

cysteine residue of G12C (104,105). Innovative approaches to directly targeting RIT1 might be

possible, but a more promising and informative approach is to identify RIT1 genetic

dependencies that could also serve as viable drug targets.

In cellular models of hepatocellular carcinoma,

RIT1

-expressing cells are more sensitive

than

RIT1

knockout cells to the multi-kinase inhibitor sorafenib (62). This drug targets BRAF,

VEGFR and other kinases in the RAS/RTK pathway (106). Sorafenib treatment suppressed ERK

activation in

RIT1

-mutant cells; however, AKT levels were not affected (62). Given that RIT1

induces AKT activation in multiple different cellular contexts, the combination of sorafenib with

the AKT inhibitor MK226 was tested in HCC mouse models (62). In this model, the combination

treatment resulted in almost complete abrogation of tumor burden (62).

These findings in HCC are consistent with other attempts to block

RIT1

-driven

oncogenesis via targeting RAS/MAPK pathway members. In PC6 pheochromocytoma cells

expressing mutant

RIT1

, phosphorylation of MEK, ERK, and AKT were higher relative to

control cells (9). Treating these cells with the MEK inhibitor PD98059 or the PI3K/mTOR

inhibitor LY294002 reduced ERK and AKT phosphorylation, respectively (9). These findings

are consistent with results in

RIT1

-mutant NCI-H2110 cells: knockdown of

RIT1

reduced MEK,

ERK, and AKT phosphorylation (9). In mice injected with NCI-H2110 cells, treatment with the

PI3K/mTOR inhibitor GDC-0941 significantly reduced tumor growth (9). This work suggests

that inhibition of PI3K and MEK could be viable treatment options for

RIT1

-driven lung cancer.

Another approach to treating

RIT1

-driven diseases could focus on decreasing the protein

abundance of RIT1. In NSCLC,

USP9X

knockout reduces the abundance of wild-type and

mutant RIT1 (74), suggesting that USP9X inhibitors could be effective in treating diseases

characterized by

RIT1

amplifications and mutations. This is further discussed in

Chapter 3

1.9

HERE DO WE GO FROM HERE

In the past five years, our knowledge of RIT1 biology and oncogenic mechanisms has

greatly expanded. While initially

RIT1

alterations were studied in the context of the

developmental disease Noonan syndrome, we now understand that

RIT1

mutations and

amplifications underlie many different cancer types. We have begun to uncover the molecular

mechanisms that contribute to RIT1 oncogenesis, and this remains an area of important and

active research.

The Berger Lab’s genome-wide CRISPR screen in PC9 lung adenocarcinoma cells

provided key insight pertaining to RIT1 genetic dependencies. This screen revealed that

RIT1

mutant cells are uniquely dependent on components of the Spindle Assembly Checkpoint

(SAC). Other groups have independently demonstrated that RIT1 weakens the SAC, but it is

currently not clear if this is a side-effect or a direct mechanism of RIT1 oncogenesis.

Furthermore, although Aurora kinase inhibitors are available, none are currently approved for

cancer treatment. As such, RIT1’s mitotic perturbation is an intriguing mechanism of RIT1

biology but may not represent the most promising avenue for targeted therapies.

There is a growing understanding that disrupting the protein abundance of RIT1 may be a

means of abrogating tumor growth. Several groups have published work to support the

hypothesis that the protein abundance of RIT1 is important for its function. One of the current

hypotheses is that mutant RIT1 accumulates at the plasma membrane and recruits RAF kinases,

which then activates RAS proteins in the vicinity and activates downstream MAPK signaling.

This is an intriguing model, and future work is required to understand if this mechanism is

occurring across different cell types and cancer contexts.

The model of

RIT1

-driven oncogenesis is gradually being filled in, and pieces are coming

together. Despite this progress, there is a need to understand how these oncogenic mechanisms

intersect or are unique across cancer types. Evidence suggests that RIT1 weakens the SAC,

synergizes with YAP, and promotes EMT. Are these mechanisms happening concurrently, and

are some or all of them facilitated by increased RIT1 protein abundance? Does mutant RIT1 and

over-expression of wild-type RIT1 activate these pathways? My thesis work has provided key

insight into understanding the protein-level regulation of both wild-type and mutant RIT1, as

well as implications for future treatment options.

In Chapter 2, I further discuss the Berger Lab’s CRISPR screening system in PC9 lung

adenocarcinoma cells and present data I generated to explore the mechanism of RIT1’s mitotic

perturbation. I also present the finding of USP9X as a genetic dependency in

RIT1

-mutant cells.

In Chapter 3, I share my work demonstrating that wild-type and mutant RIT1 are substrates of

the deubiquitinase USP9X. Overall, this work identifies USP9X as a novel regulator of RIT1

protein abundance and reveals USP9X as a potential druggable target in diseases characterized

RIT1

mutations and amplifications. This work can inspire future studies on USP9X inhibitors

to address the major unmet clinical need for targeted therapies in

RIT1

-driven cancers.

Chapter 2.

IDENTIFICATION OF RIT1 GENETIC

DEPENDENCIES

A version of this chapter has been published based on content in the following peer-reviewed
manuscripts:

Riley, A. K. & Berger, A. H. (2021). Genome-wide CRISPR screening reveals novel therapeutic

targets in RIT1-driven lung cancer.

Mol Cell Oncol,

8(6).

http://doi.org/10.1080/23723556.2021.2000318

I wrote the manuscript cited above and generated all associated figures.

Vichas, A., Riley, A. K., Nkinsi, N. T., Kamlapurkar, S., Parrish, P. C. R., Lo, A., Duke, F.,

Chen J., Fung, I., Watson, J., Rees, M., Gabel, A. M., Thomas, J. D., Bradley, R. K., Lee,
J. K., Hatch, E. M., Baine, M. K., Rekhtman, N., Ladanyi, M., Piccioni, F., & Berger, A.
H. (2021). Integrative oncogene-dependency mapping identifies RIT1 vulnerabilities and
synergies in lung cancer.

Nat Comm

12(4789).

https://doi.org/10.1038/s41467-021-

24841-y

For the manuscript cited above, I performed mitotic timing assays, contributed methods, and
helped with revision experiments. I assembled the final revised manuscript for re-submission.
Portions of this manuscript are presented in

Section 2.2

2.1

PC9

SYSTEM FOR UNCOVERING

RIT1

ESSENTIAL GENES

Figure 2.1

. Isogenic CRISPR screening identifies genetic vulnerabilities in

RIT1

-driven lung

cancer.

Only one

RIT1(Ras-like in all tissues)

-mutant lung cancer cell line is commercially available,

thereby making

RIT1

mutations difficult to study. Cell line numbers obtained from the

Dependency Map (DepMap) portal.

Left, in PC9 lung adenocarcinoma cells, a mutation in

EGFR

(epidermal growth factor

receptor) renders cells sensitive to the EGFR inhibitor erlotinib. Right, expression of
RIT1

M90I

in PC9 cells confers erlotinib resistance and restores cell survival.

This erlotinib resistance phenotype was used to perform CRISPR/Cas9 screens in isogenic
PC9 cells expressing lung cancer driver oncogenes, including RIT1

M90I

From these CRISPR screens, we found that RIT1

M90I

weakens the Spindle Assembly

Checkpoint (SAC), rendering cells vulnerable to inhibitors of SAC components such as the
Aurora kinases. We also found that RIT1

M90I

synergizes with YAP1 (yes-associated protein

1) to induce transcription of YAP1 targets, suggesting that YAP1 inhibitors should be
investigated for the treatment of diseases caused by mutation of

RIT1

. Figure created with

Biorender.com.

As introduced in

Section 1.4

, the Berger Lab implemented a unique CRISPR-screening

approach to assess genetic dependencies and synergies in

RIT1

-mutant lung cancer. We

performed genome-wide CRISPR/Cas9 knockout screens in PC9 lung adenocarcinoma cells

stably expressing a lung cancer oncogene (RIT1

M90I

, KRAS

G12V

, or EGFR

T790M/L858R

) (18). These

screens took advantage of the observation that each introduced oncogene confers resistance to

EGFR inhibition (

Figure 2.1B-C

) (9,18,22). We used this drug resistance phenotype to probe the

requirements for oncogene-driven survival. We computationally analyzed the screen data to

identify genetic dependencies (genes that, when knocked out, confer a growth disadvantage) and

cooperating factors (genes that, when knocked out, confer a growth advantage). We found that

RIT1

-mutant cells were more dependent than

KRAS

-mutant cells on several genes implicated in

the Spindle Assembly Checkpoint (SAC) (18). To further investigate this, we conducted

experiments to investigate the effects of RIT1

M90I

on mitotic progression and SAC integrity.

2.2

PINDLE

SSEMBLY

HECKPOINT

(SAC)

VULNERABILITY

Altering the SAC in normal cells accelerates mitotic timing (107), and in conditions of

mitotic stress can either promote mitotic cell death or result in mitotic slippage (the exit from

mitosis before proper chromosome alignment is complete) (108). Given that many of the

RIT1

dependencies identified in the CRISPR screen were components of the SAC, we hypothesized

that RIT1

M90I

might weaken the SAC, enhancing the vulnerability of the cells to further loss of

SAC activity. To test this hypothesis, we adapted a model system commonly used for mitotic

timing experiments (109): HeLa cells expressing a nuclear H2B-GFP fusion protein (110)

(

Figure 2.2a

). We used live cell fluorescence microscopy to time the duration of mitosis from

nuclear envelope breakdown (NEBD) to anaphase onset (107) (

Figure 2.2b

). In parental H2B-

GFP cells the median duration of mitosis was 70.5 min (95% CI = 63 – 82 min), while in

RIT1

M90I

-mutant cells the median duration of mitosis was reduced to 48 min (95% CI = 45 – 51

min) (

Figure 2.2c,d

). Overall mitotic index was unaffected, suggesting that mitotic entry is not

regulated by RIT1

M90I

(

Figure 2.2e

). The difference in mitotic timing between RIT1

M90I

-mutant

cells and parental cells was eliminated by treatment with reversine, an inhibitor of the MPS1

kinase involved in establishing the SAC kinetochore signal (111), demonstrating that RIT1

M90I

perturbs mitotic timing at the level of the SAC (

Figure 2.2c,d

). A weakened SAC has been

associated with mitotic errors including misaligned chromosomes, chromosome bridges,

micronuclei formation, and aneuploidy (112–115). Consistent with RIT1

M90I

suppression of the

SAC,

RIT1

-mutant cells showed significantly higher prevalence of chromosomal abnormalities

compared to parental cells (

Figure 2.2f,g

The vulnerability of

RIT1

-mutant cells to Aurora kinase A inhibition and RIT1’s ability

to weaken the SAC led us to hypothesize that Aurora kinase A inhibition may be able to

overcome the mitotic phenotype induced by RIT1

M90I

. To test this hypothesis, we performed

mitotic timing analysis in HeLa H2B-GFP cells treated with the Aurora kinase A inhibitor

alisertib. Whereas RIT1

M90I

alone accelerated mitosis compared to parental cells, mitotic timing

in RIT1

M90I

cells was increased compared to parental cells in the setting of alisertib (

Figure

2.2h

). RIT1

M90I

-mutant cells accumulated more mitotic errors than parental cells in alisertib

(

Figure 2.1f

). Taken together, these data indicate that oncogenic RIT1 weakens the SAC,

creating a vulnerability to Aurora kinase inhibitors. Because Aurora kinases are required for full

activation of the SAC (116–118), we propose a working model by which combined RIT1

M90I

and

Aurora kinase A inhibition leads to cellular toxicity and cell death (

Figure 2.2i

Figure 2.2

. RIT1

M90I

weakens the spindle assembly checkpoint.

Western blot of RIT1 expression in parental and RIT1

M90I

-expressing HeLa H2B-GFP cells.

Vinculin was used as a loading control.

Duration of mitosis was measured as time from nuclear envelope breakdown (NEBD) to the
onset of anaphase. Each frame represents movie stills from time-lapse live-cell imaging of
parental HeLa H2B-GFP cells undergoing mitosis, scale bar = 5µm.

Time-lapse fluorescence microscopy of time in mitosis in asynchronous parental and
RIT1

M90I

-expressing HeLa H2B-GFP cells in normal media conditions (control) or treated

with 0.5 µM reversine (Rev) two hours before imaging. n = 50 cells per condition. Mitotic
timing was measured from time of NEBD to anaphase onset.

Alternative representation of data from (c). Data shown is the mean and individual data
points of each condition. ****p = 9.06e

-9

, n.s., p = 0.1754 by unpaired two-tailed t-test.

Mitotic index calculated as the percentage of mitotic cells in a frame at a chosen time point.
Box plots show the median (center line), first and third quartiles (box edges), and the min and
max range (whiskers). n = 2 biological replicates, 6 time points per condition, n.s., p =
0.3587 by unpaired two-tailed t-test.

Comparison of mitotic error rates in HeLa H2B-GFP cells expressing RIT1

M90I

compared to

parental HeLa H2B-GFP cells in vehicle (DMSO)-treated and alisertib-treated (1 μM)
conditions. n = 3 biological replicates, error bars indicate s.d., ***p

= 0.0004, **p = 0.0029

by unpaired two-tailed t-test.

Representative images of mitotic errors from HeLa cells expressing RIT1

M90I

and quantified

in (f), scale bar = 5µm.

Time-lapse fluorescence microscopy of time in mitosis in asynchronous parental and
RIT1

M90I

-expressing HeLa H2B-GFP cells in normal media conditions (control) or treated

with 1 µM alisertib two hours before imaging. n = 50 cells per condition.

Proposed model explaining enhanced efficacy of Aurora kinase inhibitors in RIT1

M90I

-mutant

cells.

2.3

ATERIALS AND METHODS

2.3.1

Genome-wide CRISPR knockout screen

The human CRISPR Brunello lentiviral pool was obtained from the Broad Institute

Genetic Perturbation Platform and is also available from Addgene (73179-LV). The library

contains 76,441 sgRNAs targeting 19,114 protein-coding genes and 1,000 non-targeting control

sgRNAs. For genome-wide CRISPR screening, 320 million PC9-Cas9-Luciferase, PC9-Cas9-

RIT1

M90I

, PC9-Cas9-KRAS

G12V

, or PC9-Cas9-EGFR

T790M/L858R

cells were infected with the

Brunello Library lentivirus (119) at a low MOI (<0.3). At 24 h after infection, the medium was

replaced with fresh media containing 1 μg/mL puromycin (Sigma). After selection on day 7, cells

were split into 2 replicates containing 40 million cells each and treated with either DMSO

(Sigma-Aldrich) or 40 nM erlotinib (Selleckchem). Cells were then passaged every 3 days and

maintained at 500-fold coverage. For early time point analysis (day 7) an initial pool of 60

million cells was harvested for genomic DNA extraction from each of the cell lines. After ~12

doublings, a final pool of 60 million cells was harvested in ice-cold PBS and stored at -80°.

Genomic DNA was extracted using the QIAamp DNA Blood Maxi Kit (QIAGEN) and

the sgRNAs from each sample were PCR amplified by dividing gDNA into multiple 100 μl

reactions containing a maximum of 10 μg gDNA (as recommended by Broad Institute standard

protocols). Per 96-well plate, a master mix consisted of 150 μl ExTaq polymerase (Takara Bio),

1,000 μl of 10x ExTaq buffer (Takara Bio), 800 μl of dNTP (Takara Bio), 50 μl of P5 primer

(stock at 100 μM concentration), and 2,075 μl water. Each well consisted of 50 μl gDNA plus

water, 40 μl PCR master mix, and 10 μl of P7 primer (stock at 5 μM concentration). PCR cycling

conditions: an initial 5 min at 95 °C; followed by 30 s at 95 °C, 30 s at 53 °C, 20 s at 72 °C, for

28 cycles; and a final 10 min extension at 72 °C. PCR samples were purified with Agencourt

AMPure XP SPRI beads (Beckman Coulter). Samples were sequenced on a HiSeq 2500

(Illumina). Raw FASTQ files were demultiplexed and sgRNA counts were calculated using

PoolQ v2.2.0.

2.3.2

Mitotic timing and chromosomal aberration analysis

RIT1

M90I

-expressing HeLa H2B-GFP cells were generated by transduction with a

pLX303-RIT1

M90I

lentivirus and selection with puromycin. Protein expression was confirmed by

Western blotting. One day before imaging, cells were seeded at a density of 30,000 cells per well

of an 8-well Ibidi glass-bottomed plate. For drug-treated populations, 0.5 μM Reversine

(Selleckchem) was added two hours before imaging. Live-cell imaging was performed using a

20×/0.70 Plan Apo Leica objective on an automated Leica DMi8 microscope outfitted with an

Andor CSU spinning disk unit equipped with Borealis illumination, an ASI automated stage with

Piezo Z-axis top plate, and an Okolab controlled environment chamber (humidified at 37°C with

5% CO2). Long term automated imaging was driven by MetaMorph software (v7.10.0.119).

Images were captured with an Andor iXon Ultra 888 EMCCD camera. Images were captured

every minute for 18 hours. Time in mitosis was measured as time from nuclear envelope

breakdown to the onset of anaphase. Imaging experiment was repeated four times with distinct

biological replicates and 50 cells were analyzed per cell line per condition.

Mitotic index was calculated from one time point selected from live-imaging experiments

in HeLa H2B-GFP parental and HeLa H2B-GFP-RIT1

M90I

cells described above. At each time-

point, three independent, representative images were collected and the mitotic index was

calculated based on the number of cells undergoing mitosis (in any phase between prometaphase

to telophase) divided by the total number of cells. The same time point was analyzed in two

independent experiments for a total of six total mitotic index analyses per cell line.

For the mitotic abnormality analysis, HeLa H2B-GFP cells were plated in a 4-well

Nunc

Lab-Tek

chamber slide (ThermoFisher Scientific) at a density of 80,000 cells per well.

The next day, cells were treated with either DMSO vehicle or 1 μM alisertib for 6 hours prior to

being washed with 1X PBS and fixed with 4% paraformaldehyde (from 16% paraformaldehyde

[wt/vol]) in 1X PBS for 30 minutes at RT. Coverslips were mounted with Vectashield antifade

mounting medium with 1.5 μg/mL DAPI (Vector Laboratories). Slides were analyzed using a

40x/1.25 Oil PL Apo Leica objective on a Leica DMi8 outfitted with a TCS SPE scan head with

spectral detection. Cells were visualized using the LAS X software platform (v3.5.7.23225).

Representative images were captured using a 40×/1.30 Plan Apo Leica objective on a Leica

DMi8 outfitted with a TCS SPE scan head with spectral detection. Images were acquired using

the LAS X software platform (v3.5.5.19976). Images were corrected for brightness and contrast

using FIJI (v2.1.0/1.53c). Images are single sections. For each biological replicate, 60 cells were

analyzed per cell line. The prevalence of chromosome bridges, lagging and chromosome

misalignment, micronuclei, aneuploidy (i.e. notable, unequal separation of chromosomes),

polyploidy (i.e. evidence of unsuccessful cytokinesis) and normal separation of chromosomes

were recorded.

2.1

CKNOWLEDGMENTS

The mitotic timing analysis was included as part of a larger published manuscript (18). I

completed the experiments presented in

Figure 2.2

. I would like to thank all the co-authors on

this paper, particularly Dr. Athea Vichas who completed many of the other experiments for this

manuscript. All the authors and I thank Amy Goodale (Broad Institute) for technical assistance in

the design and synthesis of the targeted validation screening library and Dr. Daphne Avgousti

(Fred Hutchinson Cancer Center) for providing HeLa H2B-GFP cells. We thank the Broad

Institute Genetic Perturbation Platform for technical assistance and advice and the Fred Hutch

Genomics Shared Resource and Comparative Medicine Shared Resource (supported by NIH/NCI

Cancer Center Support Grant P30 CA015704) and the Northwest Genomics Center. We thank

the Fred Hutch Cellular Imaging Shared Resource (supported by the Fred Hutch/University of

Washington Cancer Consortium P30 CA015704) for assistance with microscopy and image

analysis.

2.1.1

Funding

This research was funded in part through NCI R00CA197762 and R37CA252050 to

A.H.B, donations from the Smith family to A.H.B., NIH/NCI Cancer Center Support Grant P30

CA015704 New Investigator support to A.H.B., a Lung Cancer Research Foundation Research

Grant to A.V., and the Hutch United postdoctoral fellowship to A.V. A.R. was supported in part

by PHS NRSA T32GM007270 from NIGMS. P.C.R.P. was supported in part by NSF DGE-

1762114 and E.M.H was supported by NIGMS grant R35GM124766.

2.1.2

Author contributions

A.H.B. conceived of the project. A.H.B., R.K.B., E.M.H., and M.L. supervised the

project. A.V., F.P., E.M.H., M.L., and A.H.B. designed experiments. A.V., N.T.N., A.K.R.,

P.C.R.P., A.L., M.K.B., A.G., and A.H.B. analyzed the data. A.V., N.N., A.K.R., P.C.R.P., S.K.,

I.F., J.D.T., J.K.L., F.D., J.C., J.W., M.R., N.R., M.K.B., A.G., J.D.T and A.H.B. conducted

experiments. A.V. and A.H.B. wrote the manuscript. All the authors critically reviewed the

manuscript and approved the final version.

2.1.3

Competing interests

F.P. is a current employee of Merck Research Laboratories and A.V. is a current

employee of Bristol Myers Squibb. All other authors declare no competing interests.

2.2

ONCLUSION

The Berger Lab’s genome-wide CRISPR/Cas9 screen revealed novel mechanisms of

RIT1

-driven cell survival. Beyond classic RTK-RAS signaling components, we uncovered a

surprising vulnerability of

RIT1

-mutant cells to perturbation of mitotic regulators, particularly

components of the spindle assembly checkpoint. We found that

RIT1

-mutant cells showed

heightened sensitivity to loss of mitotic regulators such as

AURKA, USP9X,

MAD2L1BP,

and

PLK1

, whether by genetic inactivation or small molecule inhibition, despite no differences in cell

proliferation or mitotic index. We showed that RIT1

M90I

weakens the spindle assembly

checkpoint, leaving cells vulnerable to Aurora kinase inhibitors. When our manuscript was

published, another group published work showing discovery of RIT1 as a MAD2-binding protein

that inhibits the mitotic checkpoint complex to accelerate mitotic timing (80). Using a similar

mitotic timing assay, they showed that oncogenic RIT1

M90I

accelerates mitosis in U2-OS and

HeLa cells (80). One difference was that the assay was performed in nocodazole; otherwise the

result is nearly identical to the data we show in HeLa cells, suggesting that regulation of mitotic

timing and induction of mitotic errors are a general function of oncogenic RIT1. They also

showed that wild-type RIT1 participates in the spindle assembly checkpoint and that knockout of

endogenous

RIT1

extends mitotic timing in multiple cell types. Therefore, RIT1 normally

participates in the spindle assembly checkpoint and pathogenic levels of RIT1

M90I

alter this

normal regulation. The results of our study further imply that this mitotic phenotype confers a

targetable vulnerability to

RIT1

-mutant cells. Future genotype-directed clinical trials could

determine if patients with

RIT1

-mutant tumors would uniquely benefit from treatment with

Aurora kinase inhibitors or other modulators of mitosis. Interestingly, Aurora A activation has

been found to drive resistance to the third-generation EGFR inhibitor osimertinib (120). It is

possible that RIT1

M90I

is harnessing this same mechanism to drive erlotinib resistance and

cellular transformation.

2.3

XPLORING THE HYPOTHESIS OF

USP9X-

MEDIATED

RIT1

REGULATION

Upon publication of the CRISPR screen and SAC vulnerability data, the question of

USP9X’s role in

RIT1

-mutant cells remained. When we first identified USP9X in the CRISPR

screen, we classified it as a mitotic regulator given that USP9X is known to strengthen the SAC

by stabilizing CDC20 (118). CDC20 is a component of the MCC (mitotic checkpoint complex),

which is formed when chromosomes are not properly aligned during metaphase (121,122). The

MCC inhibits the APC/C (anaphase-promoting complex/cyclosome) until proper chromosome

alignment is achieved (121,122). In human osteosarcoma U2-OS cells, loss of

USP9X

accelerates progression through mitosis and increases the abundance of chromosomal

abnormalities (118). In this way, USP9X protects cells against excessive chromosomal

instability. Because of this, it is possible that loss of

USP9X

RIT1

-mutant cells has a similar

effect as loss of Aurora kinase A and Aurora kinase B, i.e. accumulation of chromosomal

abnormalities (

Figure 2.1f

). This may contribute to cell death given that RIT1 is already

weakening the SAC and promoting the accumulation of chromosome errors (

Figure 2.2f,g,i

This hypothesis remains to be explored. In the scope of this thesis work, given that new evidence

was emerging to link the protein abundance of RIT1 to its function, the decision was made to

explore the hypothesis that RIT1 is a direct substrate of USP9X. This is not in contradiction to

the findings that USP9X and RIT1 function at the SAC, but it is not clear if these are overlapping

or independent functions. This is further discussed in

Chapter 4.

Chapter 3.

RIT1 IS A SUBSTRATE OF THE

DEUBIQUITINASE USP9X

A version of this chapter is under peer review at iScience. A pre-print is available on bioRxiv:

Riley, A. K., Grant, M., Snell A., Vichas A., Moorthi S., Urisman, A., Castel P., Wan., L., &

Berger, A. H. The deubiquitinase USP9X regulates RIT1 protein abundance and
oncogenic phenotypes. (2023).

https://doi.org/10.1101/2023.11.30.569313

3.1

VERVIEW OF THE UBIQUITIN

PROTEASOME SYSTEM

Figure 3.1

. Schematic of protein degradation mediated by E3 ligases and deubiquitinases in the

ubiquitin-proteasome system.

Image created with Biorender.com.

The ubiquitin-proteasome system is a well-characterized cellular process involved in

regulating the intracellular abundance of protein substrates (

Figure 3.1

) (123). This system is

composed of many important cellular players, including E3 ubiquitin ligases and deubiquitinases

(DUBs) (123). E3 ligases catalyze the addition of ubiquitin molecules to proteins, primarily at

lysine residues (124). This reaction can lead to mono-ubiquitination or poly-ubiquitination

through a diverse range of ubiquitin linkages (124). Ubiquitin linkages dictate the fate of the

protein substrate, with branched K48 and K11 chains associated with proteasomal degradation

(124).

E3 ligases are protein complexes that can be classified into four main groups: HECT

(homologous to the E6AP carboxyl terminus), RING (really interesting new gene), U-box, and

RBR (RING-IBR-RING) (125). These groups are characterized based on different substrate

types and structures (125). The RING E3 ligases are the largest group, and cullin-RING ligases

(CRLs) are a unique subtype of RING E3 ligases that incorporate a Cullin subunit (126). CRL3

incorporates Cul3, which is known to associate with substrate-specific adaptors that contain BTB

(Bric-a-brac, Tramtrack, and Broad complex) domains (126). LZTR1—the negative regulator of

RIT1

M90I

discussed in

Section 1.7.1

—is a BTB-containing protein and a substrate adaptor for

CRL3 (CRL

LZTR1

) (26).

Unlike E3 ligases, DUBs do not function within a larger protein complex. Hundreds of

deubiquitinase enzymes exist in human cells and there are five main DUB families: USPs

(ubiquitin-specific proteases), OTUs (ovarian tumor proteases), UCHs (ubiquitin C-terminal

hydrolases), Josephin, and MINDYs (motif interacting with ubiquitin-containing) (127). These

families are distinguished from one another based on varying degrees of specificity for cleaving

ubiquitin linkages (127). Most DUB enzymes contain a catalytic triad composed of Cystine-

Histidine-Aspartic acid/Asparagine (128). USP9X is a member of the USP family and is highly

conserved across evolution (129).

3.2

ECHANISMS AND BIOLOGY OF

USP9X

USP9X

–first discovered in fruit flies (

Drosophila melanogaster

) as the

fat facets

gene–is

essential for embryogenesis in flies and mice (

Mus musculus

)

(130,131).

USP9X

can replace

fat

facets

during fly development (130) and shares 44% identity and 88% similarity to the fruit fly

gene (132). Although USP9X’s catalytic domain is consistent with DUBs in yeast,

fat facets

the earliest

USP9X

ortholog characterized (133). From flies to mammals,

USP9X

is highly

constrained across evolution (129). Of note, orthologs of

RIT1

and

LZTR1

have been

documented in fruit flies but not in less complex lab models such as

Caenorhabditis elegans

(35). This is in contrast to

KRAS

, which is highly conserved in yeast (35). Given this, it is

possible that a regulatory network involving

USP9X

RIT1

, and

LZTR1

arose in flies and has

been conserved in other more complex animals, including mammals. Rigorous evolutionary

analysis is required to support this hypothesis.

Although it was initially characterized in the context of development (134,135), USP9X

has been implicated in apoptosis (136–138), protein trafficking (139–142), and polarity (143–

145). USP9X has been found across a wide range of cellular compartments, including the

cytoplasm (142), nucleus (146,147), and mitochondria (138). This diversity highlights that

USP9X function is largely dictated by cell type. As a deubiquitinase, USP9X positively

maintains the abundance of proteins by removing polyubiquitin chains and preventing

proteasomal degradation (138,148,149). USP9X is known to remove polyubiquitin (138,150–

152) and monoubiquitin (139,153) chains. Despite these diverse functions, structural analysis of

USP9X suggests that the catalytic domain preferentially binds to and cleaves polyubiquitin

chains with K48- and K11-linkages (150).

In the context of cancer,

USP9X

can be upregulated or downregulated, depending on the

target substrates (133). In lung adenocarcinoma,

USP9X

has been found amplified (2,154),

deleted (2,154), and mutated (154–158). The exact consequences of these alterations have yet to

be fully elucidated. In NSCLC,

USP9X

has been characterized as an oncogene (159–161), and

high

USP9X

expression is associated with poorer overall survival (162).

Outside of cancer,

USP9X

mutations underlie X-linked developmental disability (XID)

(163,164). In the neurons of individuals afflicted with XID,

USP9X

knockout causes cytoskeletal

disruptions that hinder cell growth and migration (163). RIT1 is known to regulate neuronal

growth and survival (36) and can also affect actin dynamics in fibroblast-like cells via regulation

of p21-activated kinase (PAK1) (23). It is possible that USP9X-mediated regulation of RIT1 is

important in the context of XID, but this requires further experimentation. These findings, in

combination with the observation that USP9X can cleave a diverse range of ubiquitin linkages

(150), suggest that cellular context is important for determining key USP9X substrates and

potential relevance in disease states.

3.3

USP9X

IS AN ESSENTIAL GENE IN

RIT1

MUTANT CELLS

In prior work, we identified genes required for RIT1

M90I

to promote resistance to the

EGFR tyrosine kinase inhibitor (TKI) erlotinib in

EGFR

-mutant PC9 lung adenocarcinoma cells

(18) (

Figure 3.2A

). When comparing the CRISPR scores in erlotinib-treated vs. DMSO-treated

screens,

RIT1

emerged as a top essential gene, as expected (

Figure 3.2B, Figure 3.10A,

and Supplementary Table 1

). Another top essential gene was the deubiquitinase

USP9X

(

Figure 3.2B, Figure 3.10A, and Supplementary Table 1

To validate the result of the screen

that USP9X is necessary for RIT1-induced resistance to EGFR inhibition, we generated pooled

populations of RIT1

M90I

-mutant PC9-Cas9 cells harboring a guide RNA targeting

USP9X

RIT1

(

Figure 3.2C

). Knockout of

USP9X

RIT1

resensitized cells to erlotinib and osimertinib

(

Figure 3.2D-E and Figure 3.10B-C

). As an orthogonal approach to CRISPR knockout, we

utilized siRNAs to knockdown

USP9X

. Knockdown of

USP9X

resensitized RIT1

M90I

-mutant

cells to erlotinib (

Figure 3.2F

) and osimertinib (

Figure 3.10D

). Together, these experiments

show that USP9X is required for

RIT1

-driven drug resistance in RIT1

M90I

-mutant PC9 cells.

Gene rank plot of previously published CRISPR/Cas9 whole-genome screen performed
in RIT1

M90I

-mutant PC9-Cas9 cells. ΔCRISPR Score is the difference between CRISPR

scores in the erlotinib screen vs. CRISPR Scores in the DMSO screen.

Western blot of PC9-Cas9 cells treated with siCtrl or si

USP9X

for 48 hours. RIT1 bands

are shown from a short exposure (SE) and long exposure (LE) to visualize both wild-type
and mutant RIT1 abundance. Vinculin serves as a loading control.

Dose-response curves of PC9-Cas9 Parental cells and RIT1

M90I

-mutant PC9-Cas9 cells

with indicated gene knockouts (sg

RIT1

and sg

USP9X

) treated with erlotinib for 72 hours.

Knockouts confirmed from Western blot in (C). CellTiterGlo was used to quantify viable
cell fraction determined by normalization to DMSO control. Data shown are the mean +
s.d. of two technical replicates. Data are representative results from n = 2 independent
experiments.

Area-under-the-curve (AUC) analysis of dose response curves shown in (D). p-values
calculated by unpaired two-tailed t-tests.

Dose-response curves of RIT1

M90I

-mutant PC9 cells treated with siCtrl or si

USP9X

for 48

hours, prior to treatment with erlotinib for 72 hours. Knockdowns validated by Western
blot in (C). CellTiterGlo was used to quantify viable cell fraction determined by
normalization to DMSO control. Data shown are the mean + s.d. of two technical
replicates. Data are representative results from n = 3 independent experiments.

Given that the protein abundance of RIT1 is known to be important for its function (26),

we were intrigued to see the deubiquitinase

USP9X

as a top essential gene in

RIT1

-mutant cells

(

Figure 3.2B, Figure 3.10A, and Supplementary Table 1

). We hypothesized that USP9X may

be positively regulating RIT1 levels, and

USP9X

knockout would reduce RIT1 protein

abundance. Indeed,

USP9X

knockout reduced RIT1

M90I

protein abundance, and complete

ablation of

USP9X

with an additional siRNA resulted in further reduction of RIT1

M90I

(

Figure

3.2C

). This suggests that USP9X positively regulates RIT1 protein abundance. In addition to

EGFR TKI resistance, expression of RIT1

M90I

is known to promote anchorage-independent

growth (9). Given this, we investigated how USP9X regulates proliferation phenotypes in

RIT1

mutant PC9 cells.

Under normal media conditions, genetic depletion of

RIT1

and

USP9X

did not affect the

proliferation of

RIT1

-mutant cells (

Figure 3.11A

). In the context of erlotinib, PC9-Cas9-

RIT1

M90I

cells depend on RIT1 for growth; therefore, genes required under erlotinib treatment

are RIT1 dependency genes (

Figure 3.2A

). Because of this, we expect that the effect of

USP9X

depletion will be most pronounced when cells are treated with an EGFR inhibitor. Under

erlotinib treatment, RIT1

M90I

+ sg

USP9X

cells and RIT1

M90I

+ sg

RIT1

cells proliferated slower

than RIT1

M90I

cells (

Figure 3.3A

). In addition to 2D growth on tissue culture plates, we explored

3D growth via soft agar colony formation assays in DMSO and erlotinib (

Figure 3.3B

). After

normalizing to the DMSO condition, we found that RIT1

M90I

+ sg

USP9X

and RIT1

M90I

+ sg

RIT1

cells formed significantly fewer colonies compared to RIT1

M90I

cells (

Figure 3.3C

Figure 3.3

. USP9X regulates

RIT1

-driven proliferation and anchorage-independent growth.

Proliferation of PC9-Cas9 Parental cells and RIT1

M90I

-mutant PC9-Cas9 cells with

indicated gene knockouts (sg

RIT1

and sg

USP9X

) treated with 40 nM erlotinib. Data

shown are the mean + s.d. of three technical replicates per cell line. Data are
representative results from n = 2 independent experiments. p-value calculated by multiple
unpaired two-tailed t-tests.

Representative images of soft agar colony formation assay in and RIT1

M90I

-mutant PC9-

Cas9 cells with indicated gene knockouts (sg

RIT1

and sg

USP9X

) treated with DMSO or

500 nM erlotinib. Images captured at 4X after 10 days of growth. Scale bar is 100 μM.

Normalized counts of colonies per well formed by indicated cell lines treated with
500nM erlotinib. All data were normalized to DMSO for each cell line, and then
normalized to RIT1

M90I

. Counts taken after 10 days of growth. Data shown are the mean

+ s.d. of three technical replicates per cell line. p-values calculated by unpaired two-tailed
t-tests.

Normalized counts of colonies per well formed by indicated PC9-Cas9 cell lines
expressing RIT1

M90I

, EGFR

T790M/L858R

, or KRAS

G12V

with or without sg

USP9X

. All data

were normalized to DMSO for each cell line, and then normalized to parental (no
sg

USP9X

) cell lines. Counts taken after 10 days of growth. Data shown are the mean +

s.d. of three technical replicates per cell line. p-values calculated by unpaired two-tailed t-
tests.

To confirm that this USP9X dependency is specific to

RIT1

-mutant cells, we performed

the soft agar experiment in erlotinib-resistant

KRAS

- and

EGFR

-mutant PC9-Cas9 cells with or

without sg

USP9X

(

Figure 3.11B

). As expected,

USP9X

knockout did not affect the number of

colonies formed by KRAS

G12V

and EGFR

T790M/L858R

-mutant cells (

Figure 3.3D

). We also

assessed the size of colonies formed in erlotinib and found no difference in

USP9X

knockout

KRAS

- or

EGFR

-mutant PC9 cells, while

USP9X

knockout decreased the colony size of

RIT1

mutant cells (

Figure 3.3E

). To further confirm that this USP9X dependency is specific to

RIT1

mutant cells and not influenced by potential other known USP9X substrates, we confirmed that

USP9X

knockout did not affect the abundance of the pro-survival protein MCL1 in PC9 cells

(

Figure 3.11C

) (138). These data demonstrate that USP9X is important for maintaining

RIT1

driven proliferation and anchorage-independent growth.

3.4

USP9X

REGULATES

RIT1

ABUNDANCE AND STABILITY IN MULTIPLE CELL

LINES

We hypothesized that

USP9X

knockout may reduce the abundance of RIT1, thus

explaining USP9X’s ability to counteract RIT1 function observed above. We already observed

that

USP9X

knockout reduced RIT1 abundance in RIT1

M90I

-mutant PC9 cells (

Figure 3.2C

), and

we expanded upon these findings. In parental PC9 cells (which express endogenous, wild-type

RIT1), si

USP9X

significantly reduced wild-type RIT1 protein abundance (

Figure 3.4A-B

). In

PC9 cells expressing ectopic RIT1

M90I

, CRISPR-mediated depletion of

USP9X

also significantly

reduced the abundance of RIT1

M90I

(

Figure 3.4C-D

). Importantly, we found no difference in

RIT1

mRNA expression in si

USP9X

-treated parental and RIT1

M90I

-mutant PC9 cells (

Figure

3.12A-B

), suggesting that differences in RIT1 protein abundance are due to post-translational

modifications.

In addition to protein abundance, we explored the stability of RIT1 in the context of

USP9X

depletion. Cells were treated with cycloheximide (CHX) to inhibit protein translation,

and the level of RIT1 was monitored over time by western blot. In

USP9X

knockout PC9 cells,

RIT1 level decreased faster over the course of 12 hours compared to parental cells (

Figure

3.4E

). The half-life of RIT1 in parental cells was approximately 7.6 hours compared to 2.4 hours

USP9X

knockout cells (

Figure 3.4F

). By the 6 hour time-point, we consistently found that

RIT1 protein abundance was significantly lower in

USP9X

knockout cells compared to parental

(

Figure 3.4G

). Together, these data show that USP9X is important for maintaining RIT1 protein

abundance.

Figure 3.4

. USP9X controls RIT1 abundance and stability in PC9 lung adenocarcinoma cells.

Western blot of PC9-Cas9 Parental cells treated with indicated siRNAs for 48 hours.
Vinculin serves as a loading control.

Quantification of Western blot band intensity of RIT1 bands in (A). Data shown are the
mean + s.d. of three independent experiments with 2-3 biological replicates per condition.
p-value calculated by paired two-tailed t-test.

Western blot of RIT1

M90I

-mutant PC9-Cas9 cells (Control) and a CRISPR-engineered

clonal

USP9X

KO (sg

USP9X

) cell line. Vinculin serves as a loading control.

Quantification of Western blot band intensity of RIT1 bands in (C). Data shown are the
mean + s.d. of two independent experiments with 2 biological replicates per condition. p-
value calculated by paired two-tailed t-test.

PC9-Cas9 parental and sg

USP9X

cells treated with 100 μg/mL cycloheximide (CHX) for

the indicated time periods before harvesting for Western blot. CDC20 serves as a positive
control for USP9X activity. Tubulin serves as a loading control. Data are representative
of n = 3 independent experiments.

Half-life analysis of RIT1 protein abundance over time based on RIT1 band intensity in
(E).

Comparison of RIT1 protein abundance based on Western blot band intensity of
aggregated cycloheximide-chase experiments in cells treated with CHX for 6 hours. Data
shown are the mean + s.d. of three independent experiments with 3 technical replicates
per cell line. p-value calculated by paired two-tailed t-test.

To extend these findings to another cell context, we utilized NCI-H2110 cells, which is

currently the only commercially available lung adenocarcinoma cell line that harbors an

endogenous RIT1

M90I

mutation (9). Similar to our results in PC9 cells, treatment with si

USP9X

significantly reduced the abundance of RIT1

M90I

(

Figure 3.12C-D

). As orthogonal confirmation,

we generated NCI-H2110 cell lines incorporating an inducible (iCas9) system where expression

of Cas9 is under control of doxycycline (dox) (165). Following dox treatment and Cas9

expression,

USP9X

knockout significantly decreased RIT1 abundance (

Figure 3.12E-F

). To rule

out the possibility that changes to USP9X protein abundance are due to changes in gene

expression, we performed RT-qPCR. No difference in

RIT1

mRNA expression was observed in

USP9X

cells (

Figure 3.12G

), indicating that USP9X’s regulation of RIT1 occurs post-

transcriptionally and also confirming this relationship occurs in other cell contexts.

To further validate USP9X’s regulation of RIT1 in other cell types, we analyzed the

effects of USP9X depletion in AALE lung epithelial cells. AALE cells are an immortalized, non-

transformed human lung epithelial cell line (166). We expressed FLAG-tagged RIT1

and

RIT1

M90I

in these cells and treated with si

USP9X

(

Figure 3.5A-B

). As expected,

USP9X

knockdown reduced the abundance of both wild-type and mutant RIT1. Of note, we also saw

reduction of endogenous wild-type RIT1 in parental AALE cells (

Figure 3.5A-B

). Although the

difference in FLAG-RIT1

abundance was not statistically significant in these experiments, the

marked reduction of endogenous RIT1 in parental AALE cells suggests that USP9X is also

targeting wild-type RIT1 in this system (

Figure 3.5A-B

). It is possible that the FLAG tag or

other factors prevent USP9X’s interaction with FLAG-RIT1

, but this requires further

experimentation. Together, these data indicate that USP9X regulates RIT1 across numerous lung

cancer cell types.

3.5

AIRED

GUIDE APPROACH TO KNOCKOUT

USP9X

When trying to achieve knockout of

USP9X

, we employed multiple different methods,

including CRISPR-mediated knockout and siRNA knockdowns. Overall, we found that most

single-guide CRISPR knockouts were incomplete (

Figure 3.6A

). The siRNA method proved

effective, but as a transient approach it is not always applicable for longer time-course

experiments. To improve knockouts, in H2110iCas9 cells we implemented a paired-guide

approach incorporating two unique guides per gene target (

Figure 3.6B-D

). We predicted that

transducing cells with two guide RNAs targeting the same gene would give a better overall

knockout. We compared single guide (sgRNA) and paired-guide (pgRNA) approaches (

Figure

3.6E-F

). Overall, the paired guides performed slightly better than single guide approaches. This

pgRNA method for improving genetic knockouts could be useful for other difficult targets.

Genetic knockdowns and knockouts remain a useful tool for understanding the role(s) of certain

genes in cellular mechanisms, and improving our knockout technology will help facilitate these

studies. Most genes are easily targetable with efficient guide RNAs, but other genes—such as

USP9X

—could require the implementation of multi-guide approaches to establish stable

knockout cell lines.

Figure 3.6

. Paired-guide approach for genetic knockout.

Western blot of H2110iCas9 cells harboring single-gene (sg) guide RNAs targeting
indicated genes. Cells were treated with doxycycline/dox (to induce Cas9 expression) for
48 hours or 5 days. Actin serves as a loading control.

Western blot of H2110iCas9 cells harboring paired-gene (pg) guide RNAs targeting
indicated genes. Cells were transfected with either 300 μL or 600 μL of lentivirus for
pgRNA delivery. Plasmid # indicates the paired-guide construct used to generate the cell
line. Cells were treated with doxycycline/dox for 48 hours. Actin serves as a loading
control.

line. Cells were treated with doxycycline/dox for 5 days. Actin serves as a loading
control.

Quantification of Western blot band intensity for single-guide and paired-guide
approaches for RIT1 knockout from (A-C).

Quantification of Western blot band intensity for single-guide and paired-guide
approaches for USP9X knockout from (A) and (D). Construct 1.1 was chosen due to its
higher knockout efficiency.

3.6

RIT1

UBIQUITINATION IS MEDIATED BY

USP9X’

S CATALYTIC ACTIVITY

To identify potential DUBs of RIT1 in an unbiased manner, we fused RIT1 to a ubiquitin

molecule through a flexible peptide linker to obtain a constitutively ubiquitinated form of RIT1

that we termed RIT1

~Ub

(

Figure 3.7A

). This construct acts as a molecular trap, by stabilizing

interaction with DUBs, which are unable to cleave the peptide bond (167). Next, we undertook

affinity-purification mass spectrometry in HEK293T cells using our RIT1

~Ub

mutant. In cells

expressing RIT1

~Ub

, we found enrichment of RIT1 and LZTR1 peptides compared to control

cells transfected with empty vector (

Figure 3.7B and Figure 3.13A-C

). We also found a similar

enrichment of USP9X peptides (

Figure 3.7B, Figure 3.13A and Figure 3.13D

), indicating that

USP9X is physically interacting with ubiquitinated RIT1.

Figure 3.7

. USP9X binds to and deubiquitinates RIT1.

Schematic of affinity-purification/mass-spectrometry (AP/MS) experiment performed
in HEK293T cells transfected with RIT1

~Ub

vector. This experiment was designed to

identify proteins that bind to ubiquitinated RIT1. Figure created with Biorender.com.

Mean abundance (log

-transformed) of peptides across biological replicates (n = 7 for

Empty Vector/EV and n = 4 for RIT1

~Ub

) of affinity purification/mass spectrometry

experiment. p-values calculated by paired two-tailed t-tests.

Co-immunoprecipitation in HEK293T cells transfected with GFP control or RIT1

~Ub

Vinculin serves as a loading control. Data shown is representative of n = 2
independent experiments.

Western blot of whole cell lysates (WCL) and anti-Flag immunoprecipitates (IP)
derived from HEK293T cells transfected with Flag-RIT1

or Flag-RIT1

M90I

together

with the HA-USP9X construct. 36 hours post-transfection, cells were pretreated with
10 μM MG132 for 10 hours before harvesting. Data shown are representative of n = 4
replicates for RIT1

and n = 1 replicates for RIT1

M90I

Co-immunoprecipitation experiment in HEK293T cells transfected with indicated
GST-tagged RIT1 variants or a GFP transfection control. Vinculin serves as a loading
control. Data shown are representative of n = 2 independent experiments.

Western Blot of WCL and subsequent His-tag pull-down in 6 M guanine-HCl
containing buffer derived from HEK293T cells transfected with the indicated
plasmids. Cells were pretreated with 10 μM MG132 for 16 hours to block the
proteasome pathway before harvesting. Data shown are representative of n = 3
independent experiments.

Ubiquitination experiment as described in (F) in HEK293T cells transfected with
RIT1

and RIT1

M90I

, as well as wild-type or catalytically dead (CD) USP9X. Data

shown are representative of n = 3 independent experiments.

To further validate the physical interaction of USP9X and RIT1, we performed co-

immunoprecipitation experiments in HEK293T cells expressing FLAG-tagged RIT1 (wild-type

and RIT1

M90I

) and HA-tagged USP9X and saw evidence for interaction of RIT1

and RIT1

M90I

with USP9X (

Figure 3.7D

). To rule out potential confounding factors related to tagged USP9X,

we performed this experiment with endogenous USP9X in HEK293T cells and found that

endogenous USP9X also interacts with RIT1

and RIT1

M90I

(

Figure 3.7E

To investigate if USP9X is regulating RIT1 ubiquitination, we transiently transfected

HA-tagged RIT1 and His-tagged ubiquitin in HEK293T cells. We titrated increasing amounts of

Flag-USP9X and found that ubiquitinated RIT1 decreased in a dose-dependent manner in

relation to the amount of USP9X expressed (

Figure 3.7F

). To expand upon these findings, we

performed ubiquitin-pulldown experiments with a catalytically dead (CD) form of USP9X that

harbors an active site mutation (C1566A) which ablates its deubiquitinase activity (118).

Expression of wild-type USP9X reduced ubiquitination of RIT1

and RIT1

M90I

, but expression

of USP9X

did not affect RIT1 ubiquitination (

Figure 3.7G

). These experiments confirm that

the deubiquitinase activity of USP9X is responsible for modulating ubiquitination of RIT1.

3.1

USP9X

COULD BE A PROMISING THERAPEUTIC TARGET FOR

RIT1

DRIVEN

DISEASES

Our findings suggest that

USP9X

genetic depletion reduces RIT1 protein abundance and

abrogates

RIT1

-driven oncogenic phenotypes. As such, pharmacological inhibition of USP9X is

predicted to have similar effects, and USP9X could be a promising drug target in diseases

characterized by

RIT1

mutations and amplifications. Of note, these implications are not limited

to lung adenocarcinoma. Analysis of the Cancer Dependency Map and associated proteomics

datasets (168) revealed a positive correlation between USP9X protein abundance and RIT1

protein abundance (

Figure 3.8A

), suggesting that this regulation may extend to multiple cancer

types driven by

RIT1

alterations. Importantly, this correlation was also seen with CDC20–a

known USP9X substrate (118) (

Figure 3.8B

), whereas no correlation was observed when

comparing USP9X protein abundance to mRNA expression of

RIT1

(

Figure 3.8C

) or

CDC20

(

Figure 3.8D

) (169). Additionally, analysis of

RIT1

copy number data in lung cancer cell lines

revealed that cell lines with

RIT1

amplifications (i.e. copy number greater than 2) were more

dependent on USP9X (

Figure 3.14A

). A similar analysis of

RIT1

expression data showed that

lung cancer cell lines with increased expression of RIT1 had lower

USP9X

CRISPR scores

(

Figure 3.14B

). Although we did not find any statistical differences across these groups, trends

in these data are consistent with the copy number analysis (

Figure 3.14A

).Thus,

pharmacological inhibition of USP9X could be an intriguing strategy to promote RIT1

degradation, which would be detrimental to the growth and proliferation of tumors driven by

RIT1

mutations and amplifications (

Figure 3.8E

Together, we propose a model whereby USP9X positively regulates wild-type and mutant

RIT1 (

Figure 3.8E

). We predict that this regulation counteracts the effects of LZTR1 on wild-

type RIT1 (

Figure 3.8E

). LZTR1 promotes K48 polyubiquitination of RIT1

at K187 and

K135 (26). Given this, it is possible that USP9X deubiquitinates RIT1

at these lysine sites.

Future work is needed to identify the precise lysines targeted by USP9X, and whether this varies

between wild-type and mutant forms of RIT1. The E3 ligase responsible for ubiquitinating

mutant RIT1 has yet to be identified (

Figure 3.8E

). In summary, our work builds upon the

understanding that the protein abundance of RIT1 is key to its function, and we have identified

USP9X as a positive regulator of wild-type and mutant RIT1.

Correlation of DepMap gene expression data (Expression Public 23Q2) (169) for RIT1
and USP9X proteomics data (Q93008) (168). Pearson r and p-values calculated in Prism.

Correlation of DepMap gene expression data (Expression Public 23Q2) (169) for CDC20
and USP9X proteomics data (Q93008) (168). Pearson r and p-values calculated in Prism.

Proposed model (left) of RIT1 protein regulation. RIT1

is ubiquitinated by LZTR1,

while RIT1

M90I

is ubiquitinated by a currently unknown E3 ligase. USP9X counteracts

the ubiquitination of both wild-type and mutant RIT1. Increased RIT1 abundance and
stability are important for RIT1 function and disease progression. The exact biological
consequences of RIT1

amplification have yet to be elucidated. Genetic knockout

(right) of

USP9X

prevents RIT1 deubiquitination, thereby promoting RIT1 degradation

and abrogating oncogenic phenotypes. Figure created with Biorender.com.

3.2

ISCUSSION

Our group previously performed CRISPR screens to explore

RIT1

genetic dependencies

(18) and identified the deubiquitinase

USP9X

as a genetic regulator of RIT1 function (18). This

finding was particularly interesting given that there is a growing understanding that the protein

abundance of RIT1 is important for its function (26,34,44). Mutant forms of RIT1 evade

regulation by the CRL3

LZTR1

complex, thereby increasing RIT1 protein abundance (26). As such,

it is highly probable that the activity of RIT1 is also regulated by DUBs.

Given the function of USP9X as a deubiquitinase, we predicted that USP9X would

physically interact with and modify the ubiquitination status of RIT1. Our unbiased AP/MS

approach revealed that USP9X interacts with mono-ubiquitinated RIT1 (

Figure 3.7A-B, Figure

3.13A, and Figure 3.13D

). As expected, this assay also detected LZTR1–a known RIT1

interactor–thereby increasing the robustness of our findings (

Figure 3.7B, Figure 3.13A, and

Figure 3.13C

). We further validated the physical interaction of USP9X and RIT1 in HEK293T

cells (

Figure 3.7D-E

) and confirmed that the catalytic activity of USP9X regulates the

ubiquitination status of RIT1

and RIT1

M90I

(

Figure 3.7F-G

). Our findings suggest that

USP9X deubiquitinates RIT1

M90I

, but we have yet to identify the E3 ubiquitin ligase(s) that is

promoting ubiquitination. In our CRISPR screen, we identified 22 E3 ligases that were positively

selected (

Supplementary Table 1

), meaning that individual genetic knockout of these ligases

conferred a growth advantage in

RIT1

-mutant PC9 cells. Therefore, these genes are candidate E3

ligases that may target RIT1

M90I

independently or potentially cooperate with CRL

LZTR1

ubiquitinate RIT1. Systematic characterization and investigation of these 22 E3 ligases is

required to further elucidate the protein-level regulation of RIT1

M90I

Even though mutant RIT1 is the predominant form of RIT1 expressed in PC9 cells,

LZTR1 was observed as a cooperating factor in our CRISPR screen (18). In other words,

knockout of

LZTR1

conferred a growth advantage in RIT1

M90I

-mutant PC9 cells (

Figure 3.2B,

Figure 3.10A, and Supplementary Table 1

). This result was somewhat unexpected given that

the M90I mutation in RIT1 prevents its interaction with LZTR1 (26). However, it is possible that

LZTR1 is acting on the endogenous wild-type RIT1 that is also expressed in our engineered

RIT1

-mutant PC9 cells. Our work predicts that both USP9X and LZTR1 are acting on wild-type

RIT1, while only USP9X is targeting mutant RIT1 (

Figure 3.8E

). The findings presented here

support currently known mechanisms of RIT1 regulation (26,34,35,101), while also identifying

USP9X as a novel regulator of both wild-type and mutant RIT1.

To confirm the dependency of

RIT1

-mutant cells on USP9X, we turned to our CRISPR

screening system where we initially identified USP9X. This screen was performed with

RIT1

M90I

-mutant PC9 cells, which depend on RIT1

M90I

to confer resistance to EGFR inhibitors

such as erlotinib or osimertinib (

Figure 3.2A

). In this setting, genes required under erlotinib

treatment are RIT1 dependency genes. Given this context, we replicated the screen conditions

when exploring USP9X-mediated regulation of RIT1. We validated this dependency in erlotinib-

and osimertinib-treatment experiments (

Figure 3.2D-F and Figure 3.10B-D

). Notably, complete

knockdown of

USP9X

with siRNA resulted in a dramatic resensitization to erlotinib (

Figure

3.2C and Figure 3.2F

) and osimertinib (

Figure 3.2C

and Figure 3.10D

), while incomplete

knockout of

USP9X

via CRISPR editing only partially reverted this resistance phenotype

(

Figure 3.2C-E and Figure 3.10B-C

). These differences in sensitivity appear to be directly

related to the knockout efficiency of various techniques and further support the conclusion that

USP9X is important for regulating

RIT1

-driven drug resistance.

As expected,

USP9X

knockout impaired the proliferation of

RIT1

-mutant cells in

erlotinib (

Figure 3.3A

). In soft agar colony formation assays,

RIT1

-mutant

USP9X

knockout

cells formed fewer and smaller colonies (

Figure 3.3C-E

). These phenotypic effects appear to be

directly related to USP9X’s regulation of RIT1 protein abundance. Importantly, USP9X

knockout did not affect colony formation in erlotinib-resistant

EGFR

- or

KRAS

- mutant cells

(

Figure 3.3C-E

). This coincides with our previous CRISPR screen results, where we found that

USP9X is uniquely essential to

RIT1

-mutant cells (18). Given our knowledge of RIT1 biology

and the importance of protein abundance, these data suggest that RIT1 itself is the substrate

driving USP9X dependency.

Overall, we consistently found that USP9X depletion corresponded with decreased RIT1

protein abundance (

Figure 3.2C

Figure 3.4A-D

and

Figure 3.12C-F

USP9X

knockdown did

not affect

RIT1

gene expression (

Figure 3.12A-B, S3G

), suggesting that this regulation is

occurring at the protein level. Of note, we recognize the limitations of the PC9 cell system in

studying RIT1. PC9 cells harbor a mutation in

EGFR

, but in patient tumors

RIT1

mutations are

almost always mutually exclusive with other mutations in the RTK/RAS pathway (2). Although

the PC9+erlotinib/osimertinib system is an

in vitro

-based cell line model, it offers valuable insight

into RIT1 regulation, genetic dependencies, and oncogenic mechanisms. Furthermore, our work

in NCI-H2110 cells (

Figure 3.12

), AALE cells (

Figure 3.5

) and DepMap analyses (

Figure 3.8A-

D and Figure 3.14A-B

) underscores that RIT1 is a substrate of USP9X in other human lung cancer

cell models and may be relevant across a wide range of cancer types. Indeed, USP9X could be a

promising therapeutic target for diseases characterized by

RIT1

amplifications and mutations.

RIT1 regulation by a DUB could open up opportunities to inhibit DUB function and

decrease RIT1 protein levels. Attempts have been made to develop small molecule inhibitors

against USP9X. The compound WP1130 has been shown to inhibit USP9X as well as other

DUBs including USP5, USP14, and UCH37 (170). In cells expressing high abundance of

oncoproteins targeted by USP9X, WP1130 treatment abrogates growth and proliferation

(171,172). However, the pre-clinical practicality of WP1130 is limited due to low solubility and

poor bioavailability in animal models (173,174). The compound G9 was developed after

WP1130 and is more soluble and less toxic than WP1130 (174). Experiments with G9 have

shown promising results for breast cancer, leukemia, and melanoma cells harboring specific

mutations (148,175–177). Notably, G9 has also been shown to target USP24 (178) and USP5

(179), so it is difficult to directly link the effects of this drug to USP9X inhibition. In 2021, a

more specific USP9X inhibitor FT709 was developed with a nanomolar range IC50 (180).

Unlike WP1130 or G9, FT709 does not target USP24 or USP5 (180). It will be intriguing to test

if FT709 destabilizes RIT1 in NSCLC cells and whether FT709 sensitizing EGFRi-resistant

NSCLC cells to EGFR or MAPK inhibition.

In summary, we identified USP9X as a positive regulator of RIT1 function. USP9X

deubiquitinates wild-type and mutant RIT1 (

Figure 3.7F-G

), thereby increasing RIT1 abundance

and stability. Given that protein abundance of RIT1 is important for its function (26), USP9X is a

key factor in mediating

RIT1

-driven oncogenic phenotypes. Our work supports previously

known mechanisms of RIT1 regulation by LZTR1 (26,34,35), and we suggest that USP9X and

LZTR1 oppose the action of one another in controlling the ubiquitination status of wild-type

RIT1 (

Figure 3.8E

). We found that USP9X also targets RIT1

M90I

(

Figure 3.2C, Figure 3.4C-D,

Figure 3.7D-E, Figure 3.7G, Figure 3.12C-F, and Figure 3.5A-B

) and future work is needed

to identify other players within the ubiquitin-proteasome system that may be regulating mutant

forms of RIT1 (

Figure 3.8E and Supplementary Table 1

). Additionally, more experimentation

is required to understand the biological consequences of RIT1

amplification in disease states.

Overall, this work builds upon our knowledge of RIT1 biology and the mechanisms underlying

how

RIT1

mutations and amplifications cause disease and improves our understanding of the role

of USP9X in lung cancer. These insights can be leveraged in the future to develop robust novel

therapies for diseases characterized by

RIT1

alterations.

3.3

ONCLUSION

Overall, this work supports previously known mechanisms of RIT1 protein regulation.

The discovery of a DUB that targets mutant RIT1 may seem in contradiction with the finding

that mutant RIT1 evades regulation by the cullin RING E3 ligase and the adaptor protein

LZTR1. However, we are confident that our findings build upon current literature. To conclude

this chapter, the interplay of LZTR1, USP9X, and RIT1 will be further addressed.

Our model predicts that LZTR1 may be acting on endogenous, wild-type RIT1 or other

RAS proteins in this system. To explore this, we generated a clonal

LZTR1

knockout (KO) cell

line (

Figure 3.9A

). We transduced PC9-Cas9 parental cells and the

LZTR1

KO cell line with

FLAG-tagged RIT1

or RIT1

M90I

(

Figure 3.9A

). As expected,

LZTR1

KO increased the

abundance of wild-type RIT1 (

Figure 3.9A

). Upon

USP9X

knockdown, the abundance of

RIT1

and RIT1

M90I

decreased (

Figure 3.9A

). Dual knockout of

LZTR1

and

USP9X

rescued

the abundance of RIT1

but not RIT1

M90I

(

Figure 3.9A

). Together, these results support

previously published findings on LZTR1’s regulation of wild-type RIT1 (26) and suggest that,

unlike LZTR1, USP9X is able to act on both wild-type and mutant RIT1. In order to further

confirm that USP9X–but not LZTR1–are modulating

RIT1

-driven EGFR inhibitor resistance, we

treated FLAG- RIT1

M90I

-expressing

LZTR1

KO cells with osimertinib and found that these cells

have a similar osimertinib sensitivity to FLAG-RIT1

M90I

-expressing cells, based on Area-Under-

the-Curve (AUC) analysis (

Figure 3.9B

USP9X

knockdown resulted in a similar reversion of

osimertinib sensitivity regardless of whether

LZTR1

was expressed (

Figure 3.9B

Figure 3.9

. Evidence for dual LZTR1/USP9X regulation of wild-type RIT1.

Western Blot of PC9-Cas9 Parental cells and a clonal

LZTR1

KO cell line (sg

LZTR1

) expressing

FLAG-tagged RIT1 variants and treated with indicated siRNAs for 48 hours. Actin serves as a
loading control.

Area-Under-the-Curve (AUC) analysis of osimertinib-treated FLAG-RIT1

M90I

-expressing PC9-

Cas9 cells. p-values calculated by unpaired two-tailed t-tests. ns = not significant by unpaired
two-tailed t-test. Data shown include n = 2 technical replicates from n = 2 biological replicates.

USP9X

: - + - + - + - + - + - +

+RIT1

USP9X

LZTR1

FLAG

RIT1

Actin

+RIT1

M90I

+RIT1

M90I

PC9-Cas9 Parental

PC9-Cas9 + sg

LZTR1

250kD

100kD

25kD

37kD

LZTR1

: - + - - + +

USP9X

: - - - + - +

AUC

*p=0.0105

*p=0.0173

+RIT1

M90I

In summary, this work presents USP9X as a novel regulator of RIT1 protein abundance

and suggests a model in which

RIT1

amplifications could be maintained by the opposing actions

of LZTR1 and USP9X.

3.4

ATERIALS AND METHODS

3.4.1

Cell lines

PC9 cells were a gift from Dr. Matthew Meyerson (Broad Institute). PC9-Cas9 cells were

generated as previously described (18). NIH3T3, NCI-H2110, and HEK293T cells were obtained

from ATCC (CRL-1658, CRL-5924, and CRL-3216, respectively). PC9 and NCI-H2110 cells

were cultured in RPMI-1640 (Gibco) supplemented with 10% Fetal Bovine Serum (FBS).

NIH3T3 and HEK293T cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM,

Genesee Scientific) supplemented with 10% FBS (Peak Serum, PS-FB2). All cells were

maintained at 37°C in 5% CO

and confirmed mycoplasma-free.

3.4.2

Cell line generation

RIT1

M90I

, RIT1

M90I

+ sg

USP9X

, and RIT1

M90I

+ sg

RIT1

cells were generated as

previously described (18). PC9-Cas9 Parental + sg

USP9X

cells were generated by co-

transfecting with lipofectamine CRISPR max (Life Technologies) and a synthetic guide RNA:

TCATACTATACTCATCGACA

Single cells were plated in a 96-well cell culture plate (Falcon), and clones were

expanded and validated by Western blot analysis and Sanger sequencing. H2110iCas9 cells were

generated as previously described (18).

PC9-Cas9 KRAS

G12V

- and EGFR

T790M/L858R

-mutant cells expressing sgNTC and

USP9X

were generated by transduction with pXPR003 and sgNTC or sg

USP9X

guide RNAs.

Lentivirus was generated as previously described (18).

3.4.3

Transformation and plasmid preparation

To propagate plasmids, One Shot TOP10 Chemically Competent

E. coli

(ThermoFisher

Scientific) were transformed with 1 μg of plasmid. Bacteria were propagated and plasmid was

isolated using the Plasmid Plus Midi kit (Qiagen) as per the manufacturer’s protocol.

3.4.4

siRNA treatment

Lyophilized siRNAs were ordered from Dharmacon and resuspended to make 100 μM

stock solutions. For siCtrl conditions, ON-TARGETplus Non-targeting siRNA #1 was used. The

sequence for si

USP9X

is as follows:

Sense: 5' ACACGAUGCUUUAGAAUUUUU 3'

Antisense: 5' PAAAUUCUAAAGCAUCGUGUUU 3'

Transfections were performed using Lipofectamine RNAiMAX (Life Technologies)

following the manufacturer’s protocol.

3.4.5

Dose response curves

For drug treatment experiments in PC9 cells, cells were plated in white-bottom 384-well

plates (Falcon) at a density of 400 cells per well in 40 μL of media. For siRNA experiments,

1x10

cells were plated in a 10 cm cell culture dish (ThermoFisher Scientific). 24 hours later,

cells were transfected with 120 pmol of siCtrl or si

USP9X

(Dharmacon) following the

Lipofectamine RNAiMAX transfection procedure (Life Technologies). 48 hours later, cells were

plated in 384-well plates as described above. 24 hours after cell plating, a serial dilution of

erlotinib or osimertinib was performed using a D300e dispenser (Tecan). 72 hours post-

treatment, 10 μL of CellTiterGlo reagent (Promega) was added to each well and luminescence

was quantified on an Envision MultiLabel Plate Reader (PerkinElmer). The viable cell fraction

was calculated by comparing the viability of drug-treated cells to the average viability of cells

treated with DMSO only (Sigma-Aldrich), normalized by fluid volume. Curve fitting was

performed using GraphPad Prism (v10.1.0). Inhibitors were obtained from SelleckChem:

Erlotinib-OSI-774 (S1023) and Osimertinib-AZD92921 (S7297). Area-under-the curve (AUC)

analyses were performed in Prism 10 (v10.0.3).

3.4.6

Cell lysis and immunoblotting

Whole-cell extracts for immunoblotting were prepared by washing cells with cold PBS

(Corning) supplemented with phosphatase inhibitors (ThermoFisher) on ice and then scraping

cells in RTK lysis buffer [20 mM Tris (pH 8.0), 2 mM EDTA (pH 8.0), 137 mM NaCl, 1%

IGEPAL CA-630, 10% Glycerol, and ddH

0] supplemented with phosphatase inhibitors

(ThermoFisher Scientific) and protease inhibitors (ThermoFisher Scientific, EDTA-free).

Lysates were incubated on ice for 20 min. Following centrifugation (13,000 rpm for 20 min),

lysates were quantified using the Pierce BCA Protein Assay Kit (ThermoFisher Scientific) in a

96-well plate and read on an Accuris Smartreader 96 (MR9600). Lysates were separated by

SDS-PAGE and transferred to PVDF membranes using the Trans-blot Turbo Transfer System

(BioRad). Membranes were blocked in Intercept PBS blocking buffer (LiCOR) for 1 h at room

temperature followed by overnight incubation at 4°C with primary antibodies diluted in blocking

buffer. IRDye (LiCOR) secondary antibodies were used for detection and were imaged on the

LiCOR Odyssey DLx. Images were acquired using the Licor Acquisition Software (v1.1.0.61)

from LiCOR Biosciences. Loading control and experimental proteins were probed on the same

membrane unless indicated otherwise. For clarity, loading control is shown below experimental

conditions in all panels regardless of the relative molecular weights of the experimental

protein(s). Quantification and normalization of Western blot band intensity was performed

following protocols from Invitrogen (iBright Imaging Systems).

For immunoprecipitation experiments in Figures 4D and 4F-G: cells were lysed in EBC

buffer (50 mM Tris pH 7.5, 120 mM NaCl, 0.5% NP-40/IGEPAL CA-630) supplemented with

protease inhibitors (Thermo Scientific) and phosphatase inhibitors (Thermo Scientific). To

prepare the Whole Cell Lysates (WCL), 3 × SDS sample buffer was directly added to the cell

lysates and sonicated before being resolved on SDS-PAGE and subsequently immunoblotted

with primary antibodies. The protein concentrations of the lysates were measured using the Bio-

Rad protein assay reagent on a Bio-Rad Model 680 Microplate Reader. For immunoprecipitation,

1 mg lysates were incubated with the appropriate agarose-conjugated primary antibody for 3-4 h

at 4°C or with unconjugated antibody (1-2 mg) overnight at 4°C followed by 1 h incubation with

Protein G Sepharose beads (GE Healthcare). Immuno-complexes were washed four times with

NETN buffer (20 mM Tris, pH 8.0, 100 mM NaCl, 1 mM EDTA and 0.5% NP-40) before being

resolved by SDS-PAGE and immunoblotted with indicated antibodies.

Primary antibodies used for immunoblotting: β-Actin 1:1000 (Cell Signaling

Technology, 4970), USP9X 1:500 (Proteintech, 55054-1-AP), RIT1 1:1000 (Abcam, Ab53720),

Vinculin 1:1500 (Sigma-Aldrich, V9264), Cyclophilin A 1:1000 (Bio-Rad, VMA00535), Cas9

1:1000 (Cell Signaling Technology, 14697), CDC20 1:2000 (Santa Cruz, 13162), Tubulin

1:2000 (Sigma-Aldrich, T5168), Flag 1:2000 (Sigma-Aldrich, F1804), HA 1:2000 (Biolegend,

901503). MCL1 1:1000 (Cell Signaling Technology, 94296), EGFR 1:1000 (Cell Signaling

Technology, 2232), KRAS 1:500 (Sigma-Aldrich, WH0003845M1).

3.4.7

Proliferation assay

PC9-Cas9 cells (Parental, RIT1

M90I

, RIT1

M90I

+ sg

USP9X

, and RIT1

M90I

+ sg

RIT1

) were

seeded in triplicate in 6-well tissue culture-treated dishes (CytoOne) at a density of 1x10

cells

per well. Cells were counted and passaged every 2-4 days and replated at a density of 1x10

cells

per well. Cumulative population doublings were calculated in excel, and statistical analyses were

performed in Prism (v10.1.0).

3.4.8

Soft agar assays

For soft agar colony formation assays, 4x10

PC9-Cas9 cells (Parental,

RIT1

M90I

, RIT1

M90I

+ sg

USP9X

, and RIT1

M90I

+ sg

RIT1

) were suspended in 1.3 mL media and

2.7 mL 0.5% select agar (Sigma-Aldrich) in RPMI+10% FBS. 1 mL of this cell suspension was

plated into 3 wells on a bottom layer of 0.5% select agar in RPMI+10% FBS in 6-well non-tissue

culture treated dishes (Thermo Scientific). For soft agar inhibitor experiments, erlotinib was

suspended in the top agar solution at a final concentration of 500 nM. DMSO control conditions

were prepared to normalize by DMSO volume. After 10 days of growth, brightfield images were

acquired on an ImageExpress (Molecular Devices) microscope using a 4x/0.2 NA objective.

Fields of view were tiled in a 9x9 grid to cover the entire well with no overlap. A z-stack with

1125 μm range and 25 μm step size were acquired for each field of view and saved as a 2D

minimum projection. All images were analyzed in ImageJ (1.53t) using a custom macro

(available upon request). Images were excluded if they were obstructed by the 6-well plate

and/or if the agar in view contained bubbles or other abnormalities.

3.4.9

Cycloheximide-chase

PC9-Cas9 Parental and sg

USP9X

cells treated with 100 μg/mL cycloheximide (CHX) for

the indicated time periods before harvesting protein for Western blot. Cycloheximide (Tocris

Bioscience, Cat. No. 0970, CAS No. 66-81-9), was dissolved in DMSO as 100 mg/ml stock

solution freshly before each use.

3.4.10

RT-qPCR

For RT-qPCR experiments in siRNA-treated PC9 cells (parental and RIT1

M90I

-mutant),

150,000 cells were plated in each well of a 6-well plate (CytoOne). The next day, siRNA

treatment was performed as described above. 48 hours post-transfection, each well was washed

with 1 mL cold PBS, and 700uL of Trizol reagent (Life Technologies) was added to each well.

After 1 minute, the suspension was collected, and RNA was isolated using the Direct-Zol RNA

Miniprep Plus (Zymo Research). Reverse transcription was performed with 1 μg RNA and the

SuperScript IV First-Strand Synthesis System (Invitrogen). cDNA was amplified using TaqMan

gene expression assays (ThermoFisher Scientific): RIT1 (assay Hs00608424_m1) and 18S (assay

Hs99999901_s1). For RIT1 reactions, 30 ng of cDNA was used. For 18S reactions, 5 ng of

cDNA was used. Reactions were run on the BioRad CFX384 Real-Time system and analyzed via

the standard curve method.

For the H2110iCas9 RT-qPCR experiments, total RNA was extracted from two

biological replicates of parental H2110iCas9 cells and H2110iCas9 + sg

USP9X

cells treated with

siCtrl or si

USP9X

. Reverse transcription was performed with 1 μg RNA and the SuperScript IV

First-Strand Synthesis System (Invitrogen). 20 ng of cDNA was used for each RT-PCR reaction.

cDNA was amplified using TaqMan gene expression assays (ThermoFisher Scientific): RIT1

(assay Hs00608424_m1) and 18S (assay Hs99999901_s1). Reactions were run on the BioRad

CFX384 Real-Time system. Expression was normalized to 18S within each sample in the same

experiment, and relative expression was quantified using the 2

-ΔΔCt

method.

3.4.11

Co-immunoprecipitation

For RIT1

~Ub

pulldown, 2 million HEK293T cells were plated in 10 cm cell culture dishes

(ThermoFisher Scientific). The next day, cells were transfected with 10 μg of indicated plasmids

and jetPRIME reagent (Polyplus), following the manufacturer’s protocol. 24 hours post-

transfection, HA pulldown was performed using EZview Red Anti-HA Affinity Gel (Millipore

Sigma). In brief, cells were washed with cold PBS supplemented with phosphatase inhibitors

(ThermoFisher Scientific). Cells were scraped in 1 mL NP40 lysis buffer (150 mM NaCl, 1%

NP40, 10% glycerol, 10mM Tris (pH 8.0), and ddH

0) supplemented with protease and

phosphatase inhibitors (ThermoFisher Scientific). 50 μL of lysate was reserved for the Whole

Cell Lysate. 30 μL of pre-washed EZview beads were added per condition, and samples were

incubated for 2 hours at 4°C with shaking. Samples were washed 3X and then prepared for SDS-

PAGE and Western blot as described above. For RIT1

~Ub

pull downs used for affinity

purification mass spectrometry, a similar protocol was used with few substitutions. First, the

number of cells was scaled up to 15 million in two 15 cm plates and magnetic anti-HA beads

(ThermoFisher Scientific) were used instead. After washing beads with lysis buffer, two

additional washes were performed using PBS to remove residual detergent present in the beads.

For each experimental condition, four biological replicates were used.

For IP:RIT1 experiments with endogenous USP9X, 3 million HEK393T cells were plated

in 10 cm cell culture dishes (ThermoFisher Scientific). The next day, cells were transfected with

2.5 μg of GST-tagged RIT1 plasmids or GFP control plasmid using Lipofectamine 3000 Reagent

(ThermoFisher Scientific) following the manufacturer’s protocol. 24 hours later, cells were

washed with cold PBS supplemented with phosphatase inhibitors (ThermoFisher Scientific).

Cells were scraped in 700 μL lysis buffer (50mM Tris (pH 7.5), 1% IGEPAL-CA-630)

supplemented with protease and phosphatase inhibitors (ThermoFisher Scientific). Protein

lysates were quantified using the Pierce BCA Protein Assay Kit (ThermoFisher Scientific), and 1

mg of protein was used for each IP condition. 50 μg of protein was set aside for the Whole Cell

Lysate. For IP conditions, each lysate was pre-cleared with 20 μL of pre-washed Protein A

agarose beads (Cell Signaling 9863S). Next, 1 μg of RIT1 antibody (Abcam Ab53720) or 1 μg of

control rabbit IgG antibody (R&D Systems AB-105-C) was added, and samples were incubated

for 2 hours at 4°C with shaking. 20 μL of beads were added to each tube, and samples were

incubated overnight at 4°C with shaking. The next day, samples were washed 3X and prepared

for SDS-PAGE and Western blot as described above.

3.4.12

In vitro ubiquitination assay

HEK293T cells were transfected with RIT1, USP9X, and His-ubiquitin constructs. 36

hours after transfection, 10 μM MG132 was added to block proteasome degradation, and cells

were harvested in denatured buffer (6 M guanidine-HCl, 0.1 M Na

HPO

/NaH

, 10 mM

imidazole), followed by Ni-NTA (Ni-nitrilotriacetic acid) purification and immunoblot analysis.

MG-132 (Selleck Chemicals, Cat. No. S2619, CAS No. 1211877-36-9), was dissolved in DMSO

as a 10 mM stock solution and stored in -20°C.

3.4.13

Affinity purification/mass spectrometry

On-bead trypsin digests were performed as previously described (26), and digested

tryptic peptides were analyzed by LC-MS/MS on Orbitrap Fusion Lumos Tribrid Mass

Spectrometer (Thermo Fisher Scientific) using the same configuration and settings as previously

reported (181). Acquired MS data were analyzed using a workflow previously described

(80,181). Briefly, spectra were searched in Protein Prospector (version 6.2.4 (182)) against

human proteome (SwissProt database downloaded on 01/18/2021) and decoy database of

corresponding randomly shuffled peptides. Search engine parameters were as follows: “ESI-Q-

high-res” for the instrument, trypsin as the protease, up to 2 missed cleavages allowed,

Carbamidomethyl-C as constant modification, default variable modifications, up to 3

modifications per peptide allowed, 15 ppm precursor mass tolerance, and 25 ppm tolerance for

fragment ions. The false discovery rate was set to <1% for peptides, and at least 3 unique

peptides per protein were required. Protein Prospector search results formatted as BiblioSpec

spectral library were imported into Skyline (v21) to quantify peptide and protein abundances

using MS1 extracted chromatograms (183). Statistical analysis of observed protein abundances

was performed using MSstats package integrated in Skyline (184).

Abundance per peptide represents the log

-abundance of the peak intensity (AUC) from

mass-spectrometry for each peptide. Mean per-peptide abundance is the average enrichment of

abundance for each peptide across replicates. Mean of mean-peptide abundance was calculated

as the average of mean-peptide abundance for all peptides for the protein across repeats. Mean of

mean-peptide abundance was used to generate the heatmap (

Figure 3.13A

). Heatmap shows a

subset of those proteins with at least a 5-fold enrichment over empty vector (EV).

3.4.14

CRISPR data analysis

CRISPR scores were calculated as previously described (18). In brief, all data were

scaled so that the median of non-essential genes (based on previously published lists in DepMap

(185)) is 0 and the median of essential genes is -1. CRISPR scores were defined as this scaled,

normalized log-fold-change data. All data from this CRISPR screen are available within the main

figures and supplemental information from the associated published manuscript (18).

3.4.15

DepMap analyses

For correlation analyses, data were explored in the Cancer Dependency Map (DepMap)

online portal (

https://depmap.org/portal/

). Proteomics data were captured from (168) and

available within DepMap. RIT1, CDC20, and USP9X expression were evaluated in the

Expression Public 23Q2 datasets and previously published (169). Correlation and statistical tests

were performed in Prism (v10.1.0). For the Copy Number analyses, data was obtained from the

Copy Number Absolute dataset, as well as the Expression Public dataset in DepMap. USP9X

CRISPR scores were obtained from DepMap Public 23Q4+Score, Chronos.

3.4.16

Quantification and statistical analysis

Data are expressed as mean + s.d. unless otherwise noted. Exact numbers of biological

and technical replicates for each experiment are reported in the Figure Legends. p-values less

than 0.05 were considered statistically significant based on the appropriate statistical test for the

experiment in question. For all data, *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. Data were

analyzed using Prism Software 10.0 (GraphPad).

3.5

CKNOWLEDGEMENTS

We would like to thank Dr. Lena Schroeder of the Fred Hutchinson Cancer Center

Cellular Imaging Shared Resource for assistance with microscopy and image analysis. The wild-

type and catalytically dead USP9X constructs used in ubiquitination experiments were a kind gift

from Dr. Lindsey Allan at the University of Dundee and were previously published (118).

3.5.1

Funding

This research was funded in part through NCI F31CA271637 to A.K.R., NCI

R37CA252050 and a Pardee Foundation award to A.H.B, R01CA255398 and ACS RSG-19-226-

01-TBE to L.W., and ACS TLC-21-009-01-TLC to A.H.B. & L.W. This research was supported

by the Cellular Imaging Shared Resource RRID:SCR_022609 of the Fred Hutch/University of

Washington/Seattle Children’s Cancer Consortium (P30 CA015704).

3.5.2

Author contributions

A.K.R., L.W., and A.H.B. conceived of and designed the study. A.K.R., M.G., and A.S.

performed the cellular and biochemistry experiments. A.V. generated

RIT1

-mutant PC9-Cas9

knockout cell lines. S.M. analyzed the AP/MS data and generated associated plots. L.W. and

A.S. provided panels 3E and 4D, and L.W. and M.G. provided panels 4F-G. P.C. contributed

reagents. A.U. and P.C. performed AP/MS experiments and provided the associated methods.

A.K.R. wrote the manuscript, which was reviewed by all co-authors.

3.5.3

Competing interests

The authors declare no competing interests.

3.6

ATA AND MATERIALS AVAILABILITY

Data and materials used for the USP9X/RIT1 study are outlined in

Table 3.1

. Key Resources Table.

REAGENT or RESOURCE

SOURCE

IDENTIFIER

Antibodies

β-Actin

Cell Signaling
Technology

4970

USP9X

Proteintech

55054-1-AP

RIT1

Abcam

Ab53720

Vinculin

Sigma-Aldrich

V9264

Cyclophilin A

Bio-Rad

VMA00535

Cas9

Cell Signaling
Technology

14697

CDC20

Santa Cruz

13162

Tubulin

Sigma-Aldrich

T5168

Flag

Sigma-Aldrich

F1804

MCL1

Cell Signaling
Technology

94296

EGFR

Cell Signaling
Technology

2232

KRAS

Sigma-Aldrich

WH0003845M1

Biolegend

901503

Rabbit IGG antibody

R&D Systems

AB-105-C

IRDye secondary antibodies

LiCOR

922-322/680

Bacterial and virus strains

One Shot TOP10 Chemically Competent

E. coli

ThermoFisher
Scientific

C404010

Chemicals, peptides, and recombinant proteins

Erlotinib-OSI-774

SelleckChem

S1023

Osimertinib-AZD92921

SelleckChem

S7297

MG132

Selleck Chemicals

Cat. No. S2619,
CAS No.
1211877-36-9

Cycloheximide

Tocris Bioscience

Cat. No. 0970,
CAS No. 66-
81-9

Complete protease inhibitor cocktail tablets

ThermoFisher
Scientific

A32955

Phosphatase inhibitor tablets

ThermoFisher
Scientific

A32957

Trypsin

Corning

MT 25-053-CI

jetPRIME Reagent

Polyplus

101000027

Lipofectamine 3000 Reagent

ThermoFisher
Scientific

L3000008

RNAiMAX

ThermoFisher
Scientific

13778075

ANTI-FLAG M2 Affinity Gel

Millipore Sigma

A2220

EZview Red Anti-HA Affinity Gel

Millipore Sigma

E6779

Protein A agarose beads

Cell Signaling

9863S

Protein G Sepharose beads

GE Healthcare

17-0618-01

DMSO

Sigma-Aldrich

D2650

Critical commercial assays

Pierce BCA Protein Assay Kit

ThermoFisher
Scientific

23225

Bio-Rad Protein Assay Reagent

Biorad

5000001

CellTiter-Glo Luminescent Cell Viability Assay

Promega

G7572

Lipofectamine CRISPRMAX

Invitrogen

CMAX00008

SuperScript IV First-Strand Synthesis System

Invitrogen

18091050

Plasmid Plus Midi kit

Qiagen

12941

Trans-blot Turbo Transfer System

Biorad

1704274

Taqman Gene Expression Assay: RIT1

ThermoFisher
Scientific

Hs00608424

Taqman Gene Expression Assay: 18S

ThermoFisher
Scientific

Hs99999901_s1

Deposited data

Proteomic data

ProteomeXchange
Consortium
PRIDE(186)

PXD047228

CRISPR screen

Published(18)

Experimental models: Cell lines

PC9

Dr. Matthew
Meyerson (Broad
Institute)

NIH3T3

ATCC

CRL-1658

NCI-H2110

ATCC

CRL-5924

HEK293T

ATCC

CRL-3216

Oligonucleotides

USP9X

: TCATACTATACTCATCGACA

Synthego

USP9X

Sense: 5'
A.C.A.C.G.A.U.G.C.U.U.U.A.G.A.A.U.U.U.U.U 3'
Antisense: 5' 5'-
P.A.A.A.U.U.C.U.A.A.A.G.C.A.U.C.G.U.G.U.U.U
3'

Dharmacon

CTM-511558

siCtrl: ON-TARGETplus Non-targeting siRNA #1

Dharmacon

D-001810-01-
05

Software and algorithms

Prism

Graphpad

v10.1.0

ImageJ

NIH

1.53t

ImageStudio

Licor

v5.2.5

Licor Acquisition Software

Licor

v1.1.0.61

Other

DMEM

Genesee Scientific

25-500

RPMI 1640

Gibco

11875119

Fetal Bovine Serum

Peak Serum

PS-FB2

96-well cell culture plate

Falcon

353075

6-well cell culture plate

CytoOne

CC7682-7506

6-well non-treated plate

ThermoFisher
Scientific

150239

10cm cell culture dish

ThermoFisher
Scientific

12556002

White-bottom 384-well cell culture plate

Falcon

08-772-116

PBS

Corning

21-040-CV

Tris pH 7.5

Invitrogen

15-567-027

Tris pH 8.0

Lonza

51238

EDTA 0.5 M

Hoefer

GR123-100

NaCl 5 M

Growcells

MRGF-1207

IGEPAL CA-630

Sigma-Aldrich

18896

NP-40

GBiosciences

072N-A

Glycerol

ThermoFisher
Scientific

3563501000M

Intercept PBS Blocking Buffer

LiCOR

927-70003

Select Agar

Sigma-Aldrich

A5054

3.6.1

Lead contact

Further information and requests for resources and reagents should be directed to and will

be fulfilled by the lead contact, Dr. Alice Berger (

ahberger@fredhutch.org

3.6.2

Materials availability

This study did not generate new unique reagents. Reagents used in this study are

commercially available or available upon request to the lead author.

3.6.3

Data and code availability

The data supporting the findings of this study are available within the article and its

supplementary materials.

The mass spectrometry proteomics data have been deposited to the ProteomeXchange

Consortium via the PRIDE(186) partner repository with the dataset identifier

PXD047228.

Analysis of CRISPR screen data is based on a previously published manuscript(18).

This study did not generate code. The custom macro for image analysis is available upon

request.

Any additional information required to reanalyze the data reported in this paper is

available from the lead contact (Dr. Alice Berger,

ahberger@fredhutch.org

) upon request.

3.7

UPPLEMENTAL DATA

3.7.1

Supplemental tables

All Supplemental Tables referenced in this chapter can be found at the following link:

https://doi.org/10.1101/2023.11.30.569313

3.7.2

Supplemental figures

Figure 3.10

. Supplementary figure associated with

Figure 3.2

Box plots showing average CRISPR score of indicated sgRNAs in PC9-Cas9 Parental
(+Luciferase/Control) cells or RIT1

M90I

-mutant PC9-Cas9 cells. Box plots show the

median (center line) and the min and max range of replicates. For Control conditions, n =
3 biological replicates. For RIT1

M90I

-mutant cells, n = 2 biological replicates. p-values

calculated by unpaired two-tailed t-tests.

Dose-response curves of PC9-Cas9 Parental cells and RIT1

M90I

-mutant PC9-Cas9 cells

with indicated gene knockouts (sg

RIT1

and sg

USP9X

) treated with osimertinib for 72

hours. CellTiterGlo was used to quantify viable cell fraction determined by normalization
to DMSO control. Data shown are the mean + s.d. of two technical replicates. Data are
representative results from n = 2 independent experiments.

Area-under-the-curve (AUC) analysis of dose response curves shown in (D). p-values
calculated by unpaired two-tailed t-tests.

Dose-response curve of RIT1

M90I

-mutant PC9 cells treated with siCtrl or si

USP9X

for 48

hours, prior to treatment with osimertinib for 72 hours. CellTiterGlo was used to quantify

Figure 3.11

. Supplementary figure associated with Figure 3.3.

Proliferation of PC9-Cas9 Parental cells and RIT1

M90I

-mutant PC9-Cas9 cells with

indicated gene knockouts (sg

RIT1

and sg

USP9X

) treated with DMSO (vehicle). Data

shown are the mean + s.d. of three technical replicates per cell line. Data are
representative results from n = 2 independent experiments. p-value calculated by multiple
unpaired two-tailed t-tests.

Western blot of

EGFR

- and

KRAS

-mutant PC9-Cas9 cells with or without sg

USP9X

. “-“

lanes represent cells harboring sgNTC. Vinculin serves as a loading control.

Western blot of parental and RIT1

M90I

-mutant PC9 cells with sg

USP9X

or sg

RIT1

Vinculin serves as a loading control.

Figure 3.12

. Supplementary figure associated with Figure 3.4.

Relative expression of

RIT1

as determined by qPCR and standard curve-based

quantification. Parental PC9 cells were treated with siCtrl or si

USP9X

for 48 hours before

RNA collection. Relative normalized expression calculated by BioRad software. Each dot
represents a biological replicate (individual cDNA preparations). Each qPCR sample was
run in triplicate for housekeeping (18S) and RIT1 probes.

Relative expression of

RIT1

as determined by qPCR and standard curve-based

quantification. RIT1

M90I

-mutant PC9 cells were treated with siCtrl or si

USP9X

for 48

hours before RNA collection. Relative normalized expression calculated by BioRad
software. Each dot represents a biological replicate (individual cDNA preparations). Each
qPCR sample was run in triplicate for housekeeping (18S) and RIT1 probes.

Western blot of NCI-H2110 cells treated with indicated siRNAs for 72 hours. Vinculin
serves as a loading control.

Quantification of Western blot bands based on (A) and additional replicates. Data shown
are the mean + s.d. of three independent experiments with 2-3 technical replicates per
condition. p-value calculated by paired two-tailed t-test.

Western blot of NCI-H2110iCas9 cells treated with 1 μg/mL Dox for 7 days to induce
Cas9 expression. Cyclophilin A serves as a loading control.

of the peak intensity from the MS. The mean of all peptides was combined across
biological replicates. For EV samples, n = 7 biological replicates. For RIT1

~Ub

samples, n

= 4 biological replicates.

Abundance (log

-transformed) of individual RIT1 peptides in Empty Vector (EV) control

and RIT1

~Ub

conditions from AP/MS.

Abundance (log

-transformed) of individual LZTR1 peptides in EV and RIT1

~Ub

conditions from AP/MS.

Abundance (log

-transformed) of individual USP9X peptides in EV and RIT1

~Ub

conditions from AP/MS. For B-D, some peptides were detected in all replicates while
other peptides were only detected in some replicates.

Figure 3.14

. Supplementary figure associated with Figure 3.8.

Data from the Cancer Dependency Map (DepMap) comparing the CRISPR score of
USP9X (DepMap Public 23Q4+ Score, Chronos) in lung cancer cells with normal copy
number of RIT1 (CN = 1) versus cells with high copy number of RIT1 (CN > 1). Copy
Number data were based on the Copy Number (Absolute) data set. p-value calculated by
unpaired two-tailed t-test.

DepMap data comparing the CRISPR score of USP9X (DepMap Public 23Q4+ Score,
Chronos) to RIT1 expression in the Expression Public 23Q4 dataset. For visualization
purposes, cell lines were binned based on indicated ranges.

Chapter 4.

DISCUSSION

4.1

ONCLUSIONS

In summary, my dissertation research has expanded our knowledge of RIT1 biology and

protein regulation. This project began with analysis and publication of the Berger Lab’s CRISPR

screen data in PC9 lung adenocarcinoma cells. From that work, I explored the role of RIT1 in the

Spindle Assembly Checkpoint (SAC) and found that mutant RIT1 weakens the SAC, rendering

cells vulnerable to loss of SAC genes. This added to other research in the field showing that

RIT1 comprises mitotic fidelity and promotes the accumulation of chromosomal abnormalities. I

focused the remainder of my dissertation research on identifying novel therapeutic targets in

RIT1

-driven non-small cell lung cancer.

Through my dissertation work, I have identified the deubiquitinase USP9X as a positive

regulator of both wild-type and mutant RIT1. This builds upon previous studies showing that the

cullin 3 RING E3 ligase and the adaptor protein LZTR1 bind and ubiquitinate wild-type—but

not mutant—RIT1. To date, USP9X is the first RIT1 deubiquitinase that has been characterized.

Although my work focused on non-small cell lung cancer, analysis of cancer cell lines within the

Cancer Dependency Map revealed a positive correlation between RIT1 protein abundance and

USP9X. This suggests that my findings could be applicable across numerous cancer types driven

RIT1

mutations and amplifications.

Of note, I acknowledge that my findings related to RIT1’s mitotic perturbation and RIT1

as a USP9X substrate are not necessarily mutually exclusive. During mitosis, USP9X could be

stabilizing RIT1, promoting the formation of RIT1/MAD2/p31

comet

complexes. This would

accelerate progression through mitosis and weaken the SAC, as is observed in

RIT1

-mutant cells.

Although this is possible, it appears to be in contradiction to the observation that USP9X

strengthens the SAC by preventing the degradation of CDC20 (118). There are a few possible

explanations to resolve these apparent contradictions. First, USP9X-mediated regulation of RIT1

might only occur outside of mitosis. During the cell cycle, modifications such as phosphorylation

could dictate USP9X’s substrate specificity and may direct it to CDC20 over RIT1. Indeed,

phosphorylation has been shown to alter USP9X’s activity (149). Another possibility is that, in

lung cancer cells, USP9X preferentially binds RIT1 over other substrates. Although our work

suggests that CDC20 is also a USP9X substrate in PC9 cells (

Figure 3.4E

), this has not been

explored in mitotic cells. To resolve these possibilities, it will be important to conduct

experiments in mitotic PC9 cells to ascertain if USP9X is binding to and deubiquitinating RIT1

during mitosis.

Altogether, my dissertation work adds to our understanding of RIT1 biology and the

importance of RIT1 protein regulation in disease states. My research also opens opportunities to

explore USP9X-mediated RIT1 regulation in other cancers and investigate the potential

preclinical utility of USP9X inhibitors.

4.2

ROADER IMPACT

Ten years ago,

RIT1

was identified as an oncogene in lung cancer. Since then, somatic

RIT1

alterations have been found in a diverse range of cancer types, including hepatocellular

carcinoma, endometrial cancer, and myeloid malignancies. Understanding RIT1 in cancer has

been particularly challenging given that 1) we know very little about RIT1 function and 2) few

model systems exist to allow us to study RIT1 in the lab.

In the face of these challenges, the Berger Lab developed a novel screening approach in

PC9 lung adenocarcinoma cells to probe essential genes underlying RIT1 function. This

screening system is based on the observation that expression of RIT1

M90I

confers resistance to

EGFR inhibition. This PC9 screening system also has important implications for other

oncogenes that confer resistance to EGFR tyrosine kinase inhibitors. We have already published

CRISPR screen results in erlotinib-resistant

KRAS

- and

EGFR

-mutant PC9 cells, and our

screening system can also be used to understand oncogene-driven drug resistance in other

cellular contexts. For example, over-activation of the PI3K/AKT/mTOR pathway is known to

confer EGFR inhibitor resistance (187), and our PC9 system could be used to understand genetic

dependencies underlying these resistance mechanisms. This information could nominate genes as

novel therapeutic targets to mitigate drug resistance.

Of note, it is important to acknowledge the limitations of targeted therapies in cancer

treatment. In general, targeted therapies are an improvement over cytotoxic chemotherapy

because targeted therapies are designed to only kill transformed cells while ensuring that normal,

healthy cells remain viable. Many side-effects of chemotherapy—namely hair loss and

gastrointestinal discomfort—are primarily caused by the effects of chemotherapy on fast-

dividing, non-cancer cells. Targeted therapies are designed to reduce these effects and more

efficiently kill tumor cells. Despite this, small molecule inhibitors can be difficult to design. In

the context of deubiquitinases (DUBs), this remains a complex challenge given the number of

DUBs in human cells (over 100) and the fact that many of these DUBs show high sequence

homology. Because of this, a small molecule DUB inhibitor is likely to inhibit at least more than

one DUB family member. This has been an issue in the development of USP9X inhibitors, and

even the newest inhibitors bind other targets. Despite this, an inhibitor that only binds one target

is unlikely to exist. Instead, it is important to determine that any off-target effects do not

negatively affect the viability of non-transformed cells. Nevertheless, the discovery of USP9X as

a novel RIT1 regulator represents one of the most promising therapeutic targets to date for

cancers driven by

RIT1

mutations and amplifications.

Moving forward, research into oncogenic

RIT1

amplifications is crucial. Tumor

sequencing studies have found

RIT1

amplifications across many cancer types, and studies have

found that

RIT1

over-expression phenocopies mutant

RIT1

expression. These findings are in

concordance with the hypothesis that RIT1 protein abundance is important for its function. My

dissertation work presents USP9X as an essential gene for maintaining the abundance—and

potential oncogenicity—of wild-type RIT1.

Over the course of my dissertation research, I have applied my skills in biochemistry and

molecular biology to undercover novel mechanisms of oncogenic RIT1 protein regulation. I

acknowledge that there are many steps ahead to translate this work to clinical outcomes. Despite

this, I recognize the importance of my contributions, both for expanding our understanding of

RIT1 biology and for nominating new therapeutic targets in

RIT1

-driven diseases. I hope that my

work can inspire more research focused on RIT1 oncogenic mechanisms and that this knowledge

can be used to develop new therapies and improve patient outcomes.

BIBLIOGRAPHY

1. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer

Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36
Cancers in 185 Countries. CA Cancer J Clin. 2021 May;71(3):209–49.

2. Cancer Genome Atlas Research Network. Comprehensive molecular profiling of lung

adenocarcinoma. Nature. 2014 Jul 31;511(7511):543–50.

3. Trahey M, McCormick F. A cytoplasmic protein stimulates normal N-ras p21 GTPase, but

does not affect oncogenic mutants. Science. 1987 Oct 23;238(4826):542–5.

4. Cox AD, Der CJ. Ras history: The saga continues. Small GTPases. 2010 Jul;1(1):2–27.

5. Simanshu DK, Nissley DV, McCormick F. RAS Proteins and Their Regulators in Human

Disease. Cell. 2017 Jun 29;170(1):17–33.

6. Campbell SL, Khosravi-Far R, Rossman KL, Clark GJ, Der CJ. Increasing complexity of

Ras signaling. Oncogene. 1998 Sep 17;17(11 Reviews):1395–413.

7. Shao H, Kadono-Okuda K, Finlin BS, Andres DA. Biochemical characterization of the Ras-

related GTPases Rit and Rin. Arch Biochem Biophys. 1999 Nov 15;371(2):207–19.

8. Lee CH, Della NG, Chew CE, Zack DJ. Rin, a neuron-specific and calmodulin-binding

small G-protein, and Rit define a novel subfamily of ras proteins. J Neurosci. 1996 Nov
1;16(21):6784–94.

9. Berger AH, Imielinski M, Duke F, Wala J, Kaplan N, Shi G-X, et al. Oncogenic RIT1

mutations in lung adenocarcinoma. Oncogene. 2014 Aug 28;33(35):4418–23.

10. Xu F, Sun S, Yan S, Guo H, Dai M, Teng Y. Elevated expression of RIT1 correlates with

poor prognosis in endometrial cancer. Int J Clin Exp Pathol. 2015 Sep 1;8(9):10315–24.

11. Gómez-Seguí I, Makishima H, Jerez A, Yoshida K, Przychodzen B, Miyano S, et al. Novel

recurrent mutations in the RAS-like GTP-binding gene RIT1 in myeloid malignancies.
Leukemia. 2013 Sep;27(9):1943–6.

12. Harada G, Yang S-R, Cocco E, Drilon A. Rare molecular subtypes of lung cancer. Nat Rev

Clin Oncol. 2023 Apr;20(4):229–49.

13. Siegel RL, Miller KD, Wagle NS, Jemal A. Cancer statistics, 2023. CA Cancer J Clin. 2023

Jan;73(1):17–48.

14. Reuther GW, Der CJ. The Ras branch of small GTPases: Ras family members don’t fall far

from the tree. Curr Opin Cell Biol. 2000 Apr;12(2):157–65.

15. Santos E, Tronick SR, Aaronson SA, Pulciani S, Barbacid M. T24 human bladder

carcinoma oncogene is an activated form of the normal human homologue of BALB- and
Harvey-MSV transforming genes. Nature. 1982 Jul 22;298(5872):343–7.

16. Parada LF, Tabin CJ, Shih C, Weinberg RA. Human EJ bladder carcinoma oncogene is

homologue of Harvey sarcoma virus ras gene. Nature. 1982 Jun 10;297(5866):474–8.

17. Der CJ, Krontiris TG, Cooper GM. Transforming genes of human bladder and lung

carcinoma cell lines are homologous to the ras genes of Harvey and Kirsten sarcoma
viruses. Proc Natl Acad Sci U S A. 1982 Jun;79(11):3637–40.

18. Vichas A, Riley AK, Nkinsi NT, Kamlapurkar S, Parrish PCR, Lo A, et al. Integrative

oncogene-dependency mapping identifies RIT1 vulnerabilities and synergies in lung cancer.
Nat Commun. 2021 Aug 9;12(1):4789.

19. Molina JR, Adjei AA. The Ras/Raf/MAPK pathway. J Thorac Oncol. 2006 Jan;1(1):7–9.

20. Mitin N, Rossman KL, Der CJ. Signaling interplay in Ras superfamily function. Curr Biol.

2005 Jul 26;15(14):R563-74.

21. Zenonos K, Kyprianou K. RAS signaling pathways, mutations and their role in colorectal

cancer. World J Gastrointest Oncol. 2013 May 15;5(5):97–101.

22. Berger AH, Brooks AN, Wu X, Shrestha Y, Chouinard C, Piccioni F, et al. High-throughput

Phenotyping of Lung Cancer Somatic Mutations. Cancer Cell. 2016 Aug 8;30(2):214–28.

23. Meyer Zum Büschenfelde U, Brandenstein LI, von Elsner L, Flato K, Holling T, Zenker M,

et al. RIT1 controls actin dynamics via complex formation with RAC1/CDC42 and PAK1.
PLoS Genet. 2018 May;14(5):e1007370.

24. Sakabe K, Teramoto H, Zohar M, Behbahani B, Miyazaki H, Chikumi H, et al. Potent

transforming activity of the small GTP-binding protein Rit in NIH 3T3 cells: evidence for a
role of a p38gamma-dependent signaling pathway. FEBS Lett. 2002 Jan 30;511(1–3):15–
20.

25. Hoshino M, Yoshimori T, Nakamura S. Small GTPase proteins Rin and Rit Bind to PAR6

GTP-dependently and regulate cell transformation. J Biol Chem. 2005 Jun
17;280(24):22868–74.

26. Castel P, Cheng A, Cuevas-Navarro A, Everman DB, Papageorge AG, Simanshu DK, et al.

RIT1 oncoproteins escape LZTR1-mediated proteolysis. Science. 2019 Mar
15;363(6432):1226–30.

27. Burge RA, Hobbs GA. Chapter Two - Not all RAS mutations are equal: A detailed review

of the functional diversity of RAS hot spot mutations. In: O’Bryan JP, Piazza GA, editors.
Advances in Cancer Research. Academic Press; 2022. p. 29–61.

28. Eccleston JF, Moore KJ, Morgan L, Skinner RH, Lowe PN. Kinetics of interaction between

normal and proline 12 Ras and the GTPase-activating proteins, p120-GAP and
neurofibromin. The significance of the intrinsic GTPase rate in determining the
transforming ability of ras. J Biol Chem. 1993 Dec 25;268(36):27012–9.

29. Scheffzek K, Ahmadian MR, Wiesmüller L, Kabsch W, Stege P, Schmitz F, et al. Structural

analysis of the GAP-related domain from neurofibromin and its implications. EMBO J.
1998 Aug 3;17(15):4313–27.

30. Rusyn EV, Reynolds ER, Shao H, Grana TM, Chan TO, Andres DA, et al. Rit, a non-lipid-

modified Ras-related protein, transforms NIH3T3 cells without activating the ERK, JNK,
p38 MAPK or PI3K/Akt pathways. Oncogene. 2000 Sep 28;19(41):4685–94.

31. Shi G-X, Andres DA. Rit contributes to nerve growth factor-induced neuronal

differentiation via activation of B-Raf-extracellular signal-regulated kinase and p38
mitogen-activated protein kinase cascades. Mol Cell Biol. 2005 Jan;25(2):830–46.

32. Shi G-X, Rehmann H, Andres DA. A novel cyclic AMP-dependent Epac-Rit signaling

pathway contributes to PACAP38-mediated neuronal differentiation. Mol Cell Biol. 2006
Dec;26(23):9136–47.

33. Aoki Y, Niihori T, Banjo T, Okamoto N, Mizuno S, Kurosawa K, et al. Gain-of-function

mutations in RIT1 cause Noonan syndrome, a RAS/MAPK pathway syndrome. Am J Hum
Genet. 2013 Jul 11;93(1):173–80.

34. Chen S, Vedula RS, Cuevas-Navarro A, Lu B, Hogg SJ, Wang E, et al. Impaired Proteolysis

of Noncanonical RAS Proteins Drives Clonal Hematopoietic Transformation. Cancer
Discov. 2022 Oct 5;12(10):2434–53.

35. Cuevas-Navarro A, Rodriguez-Muñoz L, Grego-Bessa J, Cheng A, Rauen KA, Urisman A,

et al. Cross-species analysis of LZTR1 loss-of-function mutants demonstrates dependency
to RIT1 orthologs. Elife [Internet]. 2022 Apr 25;11. Available from:
http://dx.doi.org/10.7554/eLife.76495

36. Shi G-X, Cai W, Andres DA. Rit subfamily small GTPases: regulators in neuronal

differentiation and survival. Cell Signal. 2013 Oct;25(10):2060–8.

37. Chan JJ, Flatters D, Rodrigues-Lima F, Yan J, Thalassinos K, Katan M. Comparative

analysis of interactions of RASSF1-10. Adv Biol Regul. 2013 May;53(2):190–201.

38. Donninger H, Schmidt ML, Mezzanotte J, Barnoud T, Clark GJ. Ras signaling through

RASSF proteins. Semin Cell Dev Biol. 2016 Oct;58:86–95.

39. Rezaei Adariani S, Kazemein Jasemi NS, Bazgir F, Wittich C, Amin E, Seidel CAM, et al.

A comprehensive analysis of RAS-effector interactions reveals interaction hotspots and new
binding partners. J Biol Chem. 2021 Apr 28;296:100626.

40. Leevers SJ, Paterson HF, Marshall CJ. Requirement for Ras in Raf activation is overcome

by targeting Raf to the plasma membrane. Nature. 1994 Jun 2;369(6479):411–4.

41. Stokoe D, Macdonald SG, Cadwallader K, Symons M, Hancock JF. Activation of Raf as a

result of recruitment to the plasma membrane. Science. 1994 Jun 3;264(5164):1463–7.

42. Terrell EM, Morrison DK. Ras-Mediated Activation of the Raf Family Kinases. Cold

Spring Harb Perspect Med [Internet]. 2019 Jan 2;9(1). Available from:
http://dx.doi.org/10.1101/cshperspect.a033746

43. Heo WD, Inoue T, Park WS, Kim ML, Park BO, Wandless TJ, et al. PI(3,4,5)P

and

PI(4,5)P

Lipids Target Proteins with Polybasic Clusters to the Plasma Membrane. Science.

2006;314(5804):1458–61.

44. Cuevas-Navarro A, Wagner M, Van R, Swain M, Mo S, Columbus J, et al. RAS-dependent

RAF-MAPK hyperactivation by pathogenic RIT1 is a therapeutic target in Noonan
syndrome-associated cardiac hypertrophy. Sci Adv. 2023 Jul 14;9(28):eadf4766.

45. Migliori AD, Patel LA, Neale C. The RIT1 C-terminus associates with lipid bilayers via

charge complementarity. Comput Biol Chem. 2021 Apr;91:107437.

46. Li L, He L, Wu B, Yu C, Zhao H, Zhou Y, et al. Structural Determinants for Light-

Dependent Membrane Binding of a Photoswitchable Polybasic Domain. ACS Synth Biol.
2021 Mar 19;10(3):542–51.

47. Fang Z, Marshall CB, Yin JC, Mazhab-Jafari MT, Gasmi-Seabrook GMC, Smith MJ, et al.

Biochemical Classification of Disease-associated Mutants of RAS-like Protein Expressed in
Many Tissues (RIT1). J Biol Chem. 2016 Jul 22;291(30):15641–52.

48. Lo A, Holmes K, Kamlapurkar S, Mundt F, Moorthi S, Fung I, et al. Multiomics

characterization of oncogenic signaling mediated by wild-type and mutant RIT1. Sci Signal.
2021 Nov 30;14(711):eabc4520.

49. Lo A, McSharry M, Berger AH. Oncogenic KRAS alters splicing factor phosphorylation

and alternative splicing in lung cancer. BMC Cancer. 2022 Dec 16;22(1):1315.

50. Riley AK, Berger AH. Genome-wide CRISPR screening reveals novel therapeutic targets in

RIT1-driven lung cancer. Mol Cell Oncol. 2021 Nov 16;8(6):2000318.

51. Paez JG, Jänne PA, Lee JC, Tracy S, Greulich H, Gabriel S, et al. EGFR mutations in lung

cancer: correlation with clinical response to gefitinib therapy. Science. 2004 Jun
4;304(5676):1497–500.

52. Lynch TJ, Bell DW, Sordella R, Gurubhagavatula S, Okimoto RA, Brannigan BW, et al.

Activating mutations in the epidermal growth factor receptor underlying responsiveness of
non-small-cell lung cancer to gefitinib. N Engl J Med. 2004 May 20;350(21):2129–39.

53. Koga T, Suda K, Mitsudomi T. Utility of the Ba/F3 cell system for exploring on-target

mechanisms of resistance to targeted therapies for lung cancer. Cancer Sci. 2022
Mar;113(3):815–27.

54. Daley GQ, Baltimore D. Transformation of an interleukin 3-dependent hematopoietic cell

line by the chronic myelogenous leukemia-specific P210bcr/abl protein. Proc Natl Acad Sci
U S A. 1988 Dec;85(23):9312–6.

55. Pich O, Reyes-Salazar I, Gonzalez-Perez A, Lopez-Bigas N. Discovering the drivers of

clonal hematopoiesis. Nat Commun. 2022 Jul 23;13(1):4267.

56. Li Q, Haigis KM, McDaniel A, Harding-Theobald E, Kogan SC, Akagi K, et al.

Hematopoiesis and leukemogenesis in mice expressing oncogenic NrasG12D from the
endogenous locus. Blood. 2011 Feb 10;117(6):2022–32.

57. Van Meter MEM, Díaz-Flores E, Archard JA, Passegué E, Irish JM, Kotecha N, et al. K-

RasG12D expression induces hyperproliferation and aberrant signaling in primary
hematopoietic stem/progenitor cells. Blood. 2007 May 1;109(9):3945–52.

58. Cavé H, Caye A, Ghedira N, Capri Y, Pouvreau N, Fillot N, et al. Mutations in RIT1 cause

Noonan syndrome with possible juvenile myelomonocytic leukemia but are not involved in
acute lymphoblastic leukemia. Eur J Hum Genet. 2016 Aug;24(8):1124–31.

59. Zhao F, Bosler DS, Cook JR. Designing Myeloid Gene Panels. Arch Pathol Lab Med. 2022

Aug 1;146(8):1004–11.

60. Clara JA, Sallman DA, Padron E. Clinical management of myelodysplastic

syndrome/myeloproliferative neoplasm overlap syndromes. Cancer Biol Med. 2016
Sep;13(3):360–72.

61. McCoach CE, Le AT, Gowan K, Jones K, Schubert L, Doak A, et al. Resistance

Mechanisms to Targeted Therapies in ROS1+ and ALK+ Non-small Cell Lung Cancer. Clin
Cancer Res. 2018 Jul 15;24(14):3334–47.

62. Sun L, Xi S, Zhou Z, Zhang F, Hu P, Cui Y, et al. Elevated expression of RIT1

hyperactivates RAS/MAPK signal and sensitizes hepatocellular carcinoma to combined
treatment with sorafenib and AKT inhibitor. Oncogene. 2022 Jan;41(5):732–44.

63. Li J-T, Liu W, Kuang Z-H, Chen H-K, Li D-J, Feng Q-S, et al. [Amplification of RIT1 in

hepatocellular carcinoma and its clinical significance]. Ai Zheng. 2003 Jul;22(7):695–9.

64. Mohanta S, Sekhar Khora S, Suresh A. Cancer Stem Cell based molecular predictors of

tumor recurrence in Oral squamous cell carcinoma. Arch Oral Biol. 2019 Mar;99:92–106.

65. Song Z, Liu T, Chen J, Ge C, Zhao F, Zhu M, et al. HIF-1α-induced RIT1 promotes liver

cancer growth and metastasis and its deficiency increases sensitivity to sorafenib. Cancer
Lett. 2019 Sep 28;460:96–107.

66. Khalil A, Nemer G. The potential oncogenic role of the RAS-like GTP-binding gene RIT1

in glioblastoma. Cancer Biomark. 2020;29(4):509–19.

67. Che M, Lan Q. RIT1 Promotes Glioma Proliferation and Invasion via the AKT/ERK/NF-ĸB

Signaling Pathway. J Mol Neurosci. 2022 Aug;72(8):1547–56.

68. Li N, Zhan X. Machine Learning Identifies Pan-Cancer Landscape of Nrf2 Oxidative Stress

Response Pathway-Related Genes. Oxid Med Cell Longev. 2022 Feb 17;2022:8450087.

69. Feng Y-F, Lei Y-Y, Lu J-B, Xi S-Y, Zhang Y, Huang Q-T, et al. RIT1 suppresses

esophageal squamous cell carcinoma growth and metastasis and predicts good prognosis.
Cell Death Dis. 2018 Oct 22;9(11):1085.

70. Paladino A, D’Angelo F, Noviello TMR, Iavarone A, Ceccarelli M. Structural Model for

Recruitment of RIT1 to the LZTR1 E3 Ligase: Evidences from an Integrated Computational
Approach. J Chem Inf Model. 2021 Apr 26;61(4):1875–88.

71. Bigenzahn JW, Collu GM, Kartnig F, Pieraks M, Vladimer GI, Heinz LX, et al. LZTR1 is a

regulator of RAS ubiquitination and signaling. Science. 2018 Dec 7;362(6419):1171–7.

72. Steklov M, Pandolfi S, Baietti MF, Batiuk A, Carai P, Najm P, et al. Mutations in LZTR1

drive human disease by dysregulating RAS ubiquitination. Science. 2018 Dec
7;362(6419):1177–82.

73. Abe T, Umeki I, Kanno S-I, Inoue S-I, Niihori T, Aoki Y. LZTR1 facilitates

polyubiquitination and degradation of RAS-GTPases. Cell Death Differ. 2020
Mar;27(3):1023–35.

74. Riley AK, Grant M, Snell A, Vichas A, Moorthi S, Urisman A, et al. The deubiquitinase

USP9X regulates RIT1 protein abundance and oncogenic phenotypes [Internet]. bioRxiv.
2023 [cited 2024 Jan 5]. p. 2023.11.30.569313. Available from:
https://www.biorxiv.org/content/10.1101/2023.11.30.569313v1

75. Hanahan D. Hallmarks of Cancer: New Dimensions. Cancer Discov. 2022 Jan;12(1):31–46.

76. Gordon DJ, Resio B, Pellman D. Causes and consequences of aneuploidy in cancer. Nat

Rev Genet. 2012 Jan 24;13(3):189–203.

77. Geurts van Kessel A. The cancer genome: from structure to function. Cell Oncol . 2014

Jun;37(3):155–65.

78. Holland AJ, Cleveland DW. Boveri revisited: chromosomal instability, aneuploidy and

tumorigenesis. Nat Rev Mol Cell Biol. 2009 Jul;10(7):478–87.

79. Simonetti G, Bruno S, Padella A, Tenti E, Martinelli G. Aneuploidy: Cancer strength or

vulnerability? Int J Cancer. 2019 Jan 1;144(1):8–25.

80. Cuevas-Navarro A, Van R, Cheng A, Urisman A, Castel P, McCormick F. The RAS

GTPase RIT1 compromises mitotic fidelity through spindle assembly checkpoint
suppression. Curr Biol. 2021 Sep 13;31(17):3915-3924.e9.

81. Su Y, Lin H, Yu J, Mao L, Jin W, Liu T, et al. RIT1 regulates mitosis and promotes

proliferation by interacting with SMC3 and PDS5 in hepatocellular carcinoma. J Exp Clin
Cancer Res. 2023 Nov 29;42(1):326.

82. Musacchio A. The Molecular Biology of Spindle Assembly Checkpoint Signaling

Dynamics. Curr Biol. 2015 Oct 19;25(20):R1002-18.

83. Primorac I, Musacchio A. Panta rhei: the APC/C at steady state. J Cell Biol. 2013 Apr

15;201(2):177–89.

84. Hagan RS, Manak MS, Buch HK, Meier MG, Meraldi P, Shah JV, et al. p31(comet) acts to

ensure timely spindle checkpoint silencing subsequent to kinetochore attachment. Mol Biol
Cell. 2011 Nov;22(22):4236–46.

85. Zhang J, Shi X, Li Y, Kim B-J, Jia J, Huang Z, et al. Acetylation of Smc3 by Eco1 is

required for S phase sister chromatid cohesion in both human and yeast. Mol Cell. 2008 Jul
11;31(1):143–51.

86. Chan K-L, Gligoris T, Upcher W, Kato Y, Shirahige K, Nasmyth K, et al. Pds5 promotes

and protects cohesin acetylation. Proc Natl Acad Sci U S A. 2013 Aug 6;110(32):13020–5.

87. Wagner EF, Nebreda AR. Signal integration by JNK and p38 MAPK pathways in cancer

development. Nat Rev Cancer. 2009 Aug;9(8):537–49.

88. D’Autréaux B, Toledano MB. ROS as signalling molecules: mechanisms that generate

specificity in ROS homeostasis. Nat Rev Mol Cell Biol. 2007 Oct;8(10):813–24.

89. Cai W, Rudolph JL, Harrison SMW, Jin L, Frantz AL, Harrison DA, et al. An

evolutionarily conserved Rit GTPase-p38 MAPK signaling pathway mediates oxidative
stress resistance. Mol Biol Cell. 2011 Sep;22(17):3231–41.

90. Rane MJ, Pan Y, Singh S, Powell DW, Wu R, Cummins T, et al. Heat shock protein 27

controls apoptosis by regulating Akt activation. J Biol Chem. 2003 Jul 25;278(30):27828–
35.

91. Zheng C, Lin Z, Zhao ZJ, Yang Y, Niu H, Shen X. MAPK-activated protein kinase-2

(MK2)-mediated formation and phosphorylation-regulated dissociation of the signal
complex consisting of p38, MK2, Akt, and Hsp27. J Biol Chem. 2006 Dec
1;281(48):37215–26.

92. Wu R, Kausar H, Johnson P, Montoya-Durango DE, Merchant M, Rane MJ. Hsp27

regulates Akt activation and polymorphonuclear leukocyte apoptosis by scaffolding MK2 to
Akt signal complex. J Biol Chem. 2007 Jul 27;282(30):21598–608.

93. Cai W, Andres DA. mTORC2 is required for rit-mediated oxidative stress resistance. PLoS

One. 2014 Dec 22;9(12):e115602.

94. Rodriguez-Viciana P, Sabatier C, McCormick F. Signaling specificity by Ras family

GTPases is determined by the full spectrum of effectors they regulate. Mol Cell Biol. 2004
Jun;24(11):4943–54.

95. Yuan TL, Cantley LC. PI3K pathway alterations in cancer: variations on a theme.

Oncogene. 2008 Sep 18;27(41):5497–510.

96. Shi G-X, Jin L, Andres DA. A rit GTPase-p38 mitogen-activated protein kinase survival

pathway confers resistance to cellular stress. Mol Cell Biol. 2011 May;31(10):1938–48.

97. Cai W, Rudolph JL, Sengoku T, Andres DA. Rit GTPase regulates a p38 MAPK-dependent

neuronal survival pathway. Neurosci Lett. 2012 Dec 7;531(2):125–30.

98. Wang G-F, Dong Q, Bai Y, Yuan J, Xu Q, Cao C, et al. Oxidative stress induces mitotic

arrest by inhibiting Aurora A-involved mitotic spindle formation. Free Radic Biol Med.
2017 Feb;103:177–87.

99. Hynds DL, Spencer ML, Andres DA, Snow DM. Rit promotes MEK-independent neurite

branching in human neuroblastoma cells. J Cell Sci. 2003 May 15;116(Pt 10):1925–35.

100. Tian Y, Shao L, Wang Q, Ru G, Zhai C, Zhou L. NECAB3 promotes the migration and

invasion of liver cancer cells through HIF-1α/RIT1 signaling pathway. Open Med. 2023
May 16;18(1):20230700.

101. Abe T, Kanno S-I, Niihori T, Terao M, Takada S, Aoki Y. LZTR1 deficiency exerts high

metastatic potential by enhancing sensitivity to EMT induction and controlling KLHL12-
mediated collagen secretion. Cell Death Dis. 2023 Aug 25;14(8):556.

102. Meng Z, Moroishi T, Guan K-L. Mechanisms of Hippo pathway regulation. Genes Dev.

2016 Jan 1;30(1):1–17.

103. Gysin S, Salt M, Young A, McCormick F. Therapeutic strategies for targeting ras

proteins. Genes Cancer. 2011 Mar;2(3):359–72.

104. Moore AR, Rosenberg SC, McCormick F, Malek S. RAS-targeted therapies: is the

undruggable drugged? Nat Rev Drug Discov. 2020 Aug;19(8):533–52.

105. Ostrem JM, Peters U, Sos ML, Wells JA, Shokat KM. K-Ras(G12C) inhibitors

allosterically control GTP affinity and effector interactions. Nature. 2013 Nov
28;503(7477):548–51.

106. Llovet JM, Ricci S, Mazzaferro V, Hilgard P, Gane E, Blanc J-F, et al. Sorafenib in

Advanced Hepatocellular Carcinoma. N Engl J Med. 2008 Jul 24;359(4):378–90.

107. Meraldi P, Draviam VM, Sorger PK. Timing and checkpoints in the regulation of mitotic

progression. Dev Cell. 2004 Jul;7(1):45–60.

108. Topham CH, Taylor SS. Mitosis and apoptosis: how is the balance set? Curr Opin Cell

Biol. 2013;25(6):780–5.

109. Sivakumar S, Daum JR, Gorbsky GJ. Live-cell fluorescence imaging for phenotypic

analysis of mitosis. Methods Mol Biol. 2014;1170:549–62.

110. Kanda T, Sullivan KF, Wahl GM. Histone–GFP fusion protein enables sensitive analysis

of chromosome dynamics in living mammalian cells. Curr Biol. 1998 Mar 26;8(7):377–85.

111. Santaguida S, Tighe A, D’Alise AM, Taylor SS, Musacchio A. Dissecting the role of

MPS1 in chromosome biorientation and the spindle checkpoint through the small molecule
inhibitor reversine. J Cell Biol. 2010 Jul 12;190(1):73–87.

112. Dobles M, Liberal V, Scott ML, Benezra R, Sorger PK. Chromosome missegregation and

apoptosis in mice lacking the mitotic checkpoint protein Mad2. Cell. 2000 Jun
9;101(6):635–45.

113. Iwanaga Y, Chi Y-H, Miyazato A, Sheleg S, Haller K, Peloponese J-M Jr, et al.

Heterozygous deletion of mitotic arrest-deficient protein 1 (MAD1) increases the incidence
of tumors in mice. Cancer Res. 2007 Jan 1;67(1):160–6.

114. Li M, Fang X, Wei Z, York JP, Zhang P. Loss of spindle assembly checkpoint-mediated

inhibition of Cdc20 promotes tumorigenesis in mice. J Cell Biol. 2009 Jun 15;185(6):983–
94.

115. Zeng X, Xu WK, Lok TM, Ma HT, Poon RYC. Imbalance of the spindle-assembly

checkpoint promotes spindle poison-mediated cytotoxicity with distinct kinetics. Cell Death
Dis. 2019 Apr 5;10(4):314.

116. Yu F, Jiang Y, Lu L, Cao M, Qiao Y, Liu X, et al. Aurora-A promotes the establishment

of spindle assembly checkpoint by priming the Haspin-Aurora-B feedback loop in late G2
phase. Cell Discov. 2017 Jan 10;3:16049.

117. Courtheoux T, Diallo A, Damodaran AP, Reboutier D, Watrin E, Prigent C. Aurora A

kinase activity is required to maintain an active spindle assembly checkpoint during
prometaphase. J Cell Sci [Internet]. 2018 Apr 12;131(7). Available from:
http://dx.doi.org/10.1242/jcs.191353

118. Skowyra A, Allan LA, Saurin AT, Clarke PR. USP9X Limits Mitotic Checkpoint

Complex Turnover to Strengthen the Spindle Assembly Checkpoint and Guard against
Chromosomal Instability. Cell Rep. 2018 Apr 17;23(3):852–65.

119. Doench JG, Fusi N, Sullender M, Hegde M, Vaimberg EW, Donovan KF, et al.

Optimized sgRNA design to maximize activity and minimize off-target effects of CRISPR-
Cas9. Nat Biotechnol. 2016 Feb;34(2):184–91.

100

120. Shah KN, Bhatt R, Rotow J, Rohrberg J, Olivas V, Wang VE, et al. Aurora kinase A

drives the evolution of resistance to third-generation EGFR inhibitors in lung cancer. Nat
Med. 2019 Jan;25(1):111–8.

121. Alfieri C, Chang L, Zhang Z, Yang J, Maslen S, Skehel M, et al. Molecular basis of

APC/C regulation by the spindle assembly checkpoint. Nature. 2016 Aug
25;536(7617):431–6.

122. Yamaguchi M, VanderLinden R, Weissmann F, Qiao R, Dube P, Brown NG, et al. Cryo-

EM of Mitotic Checkpoint Complex-Bound APC/C Reveals Reciprocal and Conformational
Regulation of Ubiquitin Ligation. Mol Cell. 2016 Aug 18;63(4):593–607.

123. Ciechanover A, Schwartz AL. The ubiquitin-proteasome pathway: the complexity and

myriad functions of proteins death. Proc Natl Acad Sci U S A. 1998 Mar 17;95(6):2727–30.

124. French ME, Koehler CF, Hunter T. Emerging functions of branched ubiquitin chains.

Cell Discov. 2021 Jan 26;7(1):6.

125. Yang Q, Zhao J, Chen D, Wang Y. E3 ubiquitin ligases: styles, structures and functions.

Mol Biomed. 2021 Jul 30;2(1):23.

126. Canning P, Cooper CDO, Krojer T, Murray JW, Pike ACW, Chaikuad A, et al. Structural

Basis for Cul3 Protein Assembly with the BTB-Kelch Family of E3 Ubiquitin Ligases*. J
Biol Chem. 2013 Mar 15;288(11):7803–14.

127. Mevissen TET, Komander D. Mechanisms of Deubiquitinase Specificity and Regulation.

Annu Rev Biochem. 2017 Jun 20;86:159–92.

128. Hanpude P, Bhattacharya S, Dey AK, Maiti TK. Deubiquitinating enzymes in cellular

signaling and disease regulation. IUBMB Life. 2015 Jul;67(7):544–55.

129. Cooper GM, Stone EA, Asimenos G, NISC Comparative Sequencing Program, Green

ED, Batzoglou S, et al. Distribution and intensity of constraint in mammalian genomic
sequence. Genome Res. 2005 Jul;15(7):901–13.

130. Noma T, Kanai Y, Kanai-Azuma M, Ishii M, Fujisawa M, Kurohmaru M, et al. Stage-

and sex-dependent expressions of Usp9x, an X-linked mouse ortholog of Drosophila Fat
facets, during gonadal development and oogenesis in mice. Gene Expr Patterns. 2002
Nov;2(1–2):87–91.

131. Naik E, Webster JD, DeVoss J, Liu J, Suriben R, Dixit VM. Regulation of proximal T

cell receptor signaling and tolerance induction by deubiquitinase Usp9X. J Exp Med. 2014
Sep 22;211(10):1947–55.

132. Jones MH, Furlong RA, Burkin H, Chalmers IJ, Brown GM, Khwaja O, et al. The

Drosophila developmental gene fat facets has a human homologue in Xp11.4 which escapes
X-inactivation and has related sequences on Yq11.2. Hum Mol Genet. 1996
Nov;5(11):1695–701.

101

133. Murtaza M, Jolly LA, Gecz J, Wood SA. La FAM fatale: USP9X in development and

disease. Cell Mol Life Sci. 2015 Jun;72(11):2075–89.

134. Fischer-Vize JA, Rubin GM, Lehmann R. The fat facets gene is required for Drosophila

eye and embryo development. Development. 1992 Dec;116(4):985–1000.

135. Wood SA, Pascoe WS, Ru K, Yamada T, Hirchenhain J, Kemler R, et al. Cloning and

expression analysis of a novel mouse gene with sequence similarity to the Drosophila fat
facets gene. Mech Dev. 1997 Apr;63(1):29–38.

136. Ryoo HD, Bergmann A, Gonen H, Ciechanover A, Steller H. Regulation of Drosophila

IAP1 degradation and apoptosis by reaper and ubcD1. Nat Cell Biol. 2002 Jun;4(6):432–8.

137. Wing JP, Schreader BA, Yokokura T, Wang Y, Andrews PS, Huseinovic N, et al.

Drosophila Morgue is an F box/ubiquitin conjugase domain protein important for grim-
reaper mediated apoptosis. Nat Cell Biol. 2002 Jun;4(6):451–6.

138. Schwickart M, Huang X, Lill JR, Liu J, Ferrando R, French DM, et al. Deubiquitinase

USP9X stabilizes MCL1 and promotes tumour cell survival. Nature. 2010 Jan
7;463(7277):103–7.

139. Savio MG, Wollscheid N, Cavallaro E, Algisi V, Di Fiore PP, Sigismund S, et al. USP9X

Controls EGFR Fate by Deubiquitinating the Endocytic Adaptor Eps15. Curr Biol. 2016 Jan
25;26(2):173–83.

140. Chen H, Polo S, Di Fiore PP, De Camilli PV. Rapid Ca

-dependent decrease of protein

ubiquitination at synapses. Proc Natl Acad Sci U S A. 2003 Dec 9;100(25):14908–13.

141. Marx C, Held JM, Gibson BW, Benz CC. ErbB2 trafficking and degradation associated

with K48 and K63 polyubiquitination. Cancer Res. 2010 May 1;70(9):3709–17.

142. Murray RZ, Jolly LA, Wood SA. The FAM deubiquitylating enzyme localizes to

multiple points of protein trafficking in epithelia, where it associates with E-cadherin and
beta-catenin. Mol Biol Cell. 2004 Apr;15(4):1591–9.

143. Strutt H, Searle E, Thomas-Macarthur V, Brookfield R, Strutt D. A Cul-3-BTB

ubiquitylation pathway regulates junctional levels and asymmetry of core planar polarity
proteins. Development. 2013 Apr;140(8):1693–702.

144. Théard D, Labarrade F, Partisani M, Milanini J, Sakagami H, Fon EA, et al. USP9x-

mediated deubiquitination of EFA6 regulates de novo tight junction assembly. EMBO J.
2010 May 5;29(9):1499–509.

145. Luton F, Klein S, Chauvin J-P, Le Bivic A, Bourgoin S, Franco M, et al. EFA6, exchange

factor for ARF6, regulates the actin cytoskeleton and associated tight junction in response to
E-cadherin engagement. Mol Biol Cell. 2004 Mar;15(3):1134–45.

102

146. Trinkle-Mulcahy L, Boulon S, Lam YW, Urcia R, Boisvert F-M, Vandermoere F, et al.

Identifying specific protein interaction partners using quantitative mass spectrometry and
bead proteomes. J Cell Biol. 2008 Oct 20;183(2):223–39.

147. Urbé S, Liu H, Hayes SD, Heride C, Rigden DJ, Clague MJ. Systematic survey of

deubiquitinase localization identifies USP21 as a regulator of centrosome- and microtubule-
associated functions. Mol Biol Cell. 2012 Mar;23(6):1095–103.

148. Potu H, Peterson LF, Kandarpa M, Pal A, Sun H, Durham A, et al. Usp9x regulates Ets-1

ubiquitination and stability to control NRAS expression and tumorigenicity in melanoma.
Nat Commun. 2017 Feb 15;8:14449.

149. Dietachmayr M, Rathakrishnan A, Karpiuk O, von Zweydorf F, Engleitner T, Fernández-

Sáiz V, et al. Antagonistic activities of CDC14B and CDK1 on USP9X regulate WT1-
dependent mitotic transcription and survival. Nat Commun. 2020 Mar 9;11(1):1268.

150. Paudel P, Zhang Q, Leung C, Greenberg HC, Guo Y, Chern Y-H, et al. Crystal structure

and activity-based labeling reveal the mechanisms for linkage-specific substrate recognition
by deubiquitinase USP9X. Proc Natl Acad Sci U S A. 2019 Apr 9;116(15):7288–97.

151. Wrobel L, Siddiqi FH, Hill SM, Son SM, Karabiyik C, Kim H, et al. mTORC2 Assembly

Is Regulated by USP9X-Mediated Deubiquitination of RICTOR. Cell Rep. 2020 Dec
29;33(13):108564.

152. Wang P, Wang J, Yao S, Cui M, Cheng Y, Liu W, et al. Deubiquitinase USP9X stabilizes

RNA m6A demethylase ALKBH5 and promotes acute myeloid leukemia cell survival. J
Biol Chem. 2023 Aug;299(8):105055.

153. Rott R, Szargel R, Haskin J, Bandopadhyay R, Lees AJ, Shani V, et al. α-Synuclein fate

is determined by USP9X-regulated monoubiquitination. Proc Natl Acad Sci U S A. 2011
Nov 15;108(46):18666–71.

154. Imielinski M, Berger AH, Hammerman PS, Hernandez B, Pugh TJ, Hodis E, et al.

Mapping the hallmarks of lung adenocarcinoma with massively parallel sequencing. Cell.
2012 Sep 14;150(6):1107–20.

155. Gillette MA, Satpathy S, Cao S, Dhanasekaran SM, Vasaikar SV, Krug K, et al.

Proteogenomic Characterization Reveals Therapeutic Vulnerabilities in Lung
Adenocarcinoma. Cell. 2020 Jul 9;182(1):200-225.e35.

156. Lengel HB, Mastrogiacomo B, Connolly JG, Tan KS, Liu Y, Fick CN, et al. Genomic

mapping of metastatic organotropism in lung adenocarcinoma. Cancer Cell. 2023 May
8;41(5):970-985.e3.

157. Chen J, Yang H, Teo ASM, Amer LB, Sherbaf FG, Tan CQ, et al. Genomic landscape of

lung adenocarcinoma in East Asians. Nat Genet. 2020 Feb;52(2):177–86.

103

158. Rizvi NA, Hellmann MD, Snyder A, Kvistborg P, Makarov V, Havel JJ, et al. Cancer

immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small
cell lung cancer. Science. 2015 Apr 3;348(6230):124–8.

159. Chen X, Yu C, Gao J, Zhu H, Cui B, Zhang T, et al. A novel USP9X substrate TTK

contributes to tumorigenesis in non-small-cell lung cancer. Theranostics. 2018 Mar
22;8(9):2348–60.

160. Wang T, Jing B, Sun B, Liao Y, Song H, Xu D, et al. Stabilization of PTGES by

deubiquitinase USP9X promotes metastatic features of lung cancer via PGE2 signaling. Am
J Cancer Res. 2019 Jun 1;9(6):1145–60.

161. Jie X, Fong WP, Zhou R, Zhao Y, Zhao Y, Meng R, et al. USP9X-mediated KDM4C

deubiquitination promotes lung cancer radioresistance by epigenetically inducing TGF-β2
transcription. Cell Death Differ. 2021 Jul;28(7):2095–111.

162. Wang Y, Liu Y, Yang B, Cao H, Yang C-X, Ouyang W, et al. Elevated expression of

USP9X correlates with poor prognosis in human non-small cell lung cancer. J Thorac Dis.
2015 Apr;7(4):672–9.

163. Homan CC, Kumar R, Nguyen LS, Haan E, Raymond FL, Abidi F, et al. Mutations in

USP9X are associated with X-linked intellectual disability and disrupt neuronal cell
migration and growth. Am J Hum Genet. 2014 Mar 6;94(3):470–8.

164. Reijnders MRF, Zachariadis V, Latour B, Jolly L, Mancini GM, Pfundt R, et al. De Novo

Loss-of-Function Mutations in USP9X Cause a Female-Specific Recognizable Syndrome
with Developmental Delay and Congenital Malformations. Am J Hum Genet. 2016 Feb
4;98(2):373–81.

165. Cao J, Wu L, Zhang S-M, Lu M, Cheung WKC, Cai W, et al. An easy and efficient

inducible CRISPR/Cas9 platform with improved specificity for multiple gene targeting.
Nucleic Acids Res. 2016 Nov 2;44(19):e149.

166. Lundberg AS, Randell SH, Stewart SA, Elenbaas B, Hartwell KA, Brooks MW, et al.

Immortalization and transformation of primary human airway epithelial cells by gene
transfer. Oncogene. 2002 Jul 4;21(29):4577–86.

167. Asimaki E, Petriukov K, Renz C, Meister C, Ulrich HD. Fast friends - Ubiquitin-like

modifiers as engineered fusion partners. Semin Cell Dev Biol. 2022 Dec;132:132–45.

168. Nusinow DP, Szpyt J, Ghandi M, Rose CM, McDonald ER 3rd, Kalocsay M, et al.

Quantitative Proteomics of the Cancer Cell Line Encyclopedia. Cell. 2020 Jan
23;180(2):387-402.e16.

169. Ghandi M, Huang FW, Jané-Valbuena J, Kryukov GV, Lo CC, McDonald ER 3rd, et al.

Next-generation characterization of the Cancer Cell Line Encyclopedia. Nature. 2019
May;569(7757):503–8.

104

170. Kapuria V, Peterson LF, Fang D, Bornmann WG, Talpaz M, Donato NJ. Deubiquitinase

inhibition by small-molecule WP1130 triggers aggresome formation and tumor cell
apoptosis. Cancer Res. 2010 Nov 15;70(22):9265–76.

171. Peddaboina C, Jupiter D, Fletcher S, Yap JL, Rai A, Tobin RP, et al. The downregulation

of Mcl-1 via USP9X inhibition sensitizes solid tumors to Bcl-xl inhibition. BMC Cancer.
2012 Nov 21;12:541.

172. Wang S, Kollipara RK, Srivastava N, Li R, Ravindranathan P, Hernandez E, et al.

Ablation of the oncogenic transcription factor ERG by deubiquitinase inhibition in prostate
cancer. Proc Natl Acad Sci U S A. 2014 Mar 18;111(11):4251–6.

173. Perry JW, Ahmed M, Chang K-O, Donato NJ, Showalter HD, Wobus CE. Antiviral

activity of a small molecule deubiquitinase inhibitor occurs via induction of the unfolded
protein response. PLoS Pathog. 2012 Jul 5;8(7):e1002783.

174. Charbonneau M-E, Gonzalez-Hernandez MJ, Showalter HD, Donato NJ, Wobus CE,

O’Riordan MXD. Small molecule deubiquitinase inhibitors promote macrophage anti-
infective capacity. PLoS One. 2014 Aug 5;9(8):e104096.

175. Jaiswal A, Murakami K, Elia A, Shibahara Y, Done SJ, Wood SA, et al. Therapeutic

inhibition of USP9x-mediated Notch signaling in triple-negative breast cancer. Proc Natl
Acad Sci U S A [Internet]. 2021 Sep 21;118(38). Available from:
http://dx.doi.org/10.1073/pnas.2101592118

176. Pal A, Dziubinski M, Di Magliano MP, Simeone DM, Owens S, Thomas D, et al. Usp9x

Promotes Survival in Human Pancreatic Cancer and Its Inhibition Suppresses Pancreatic
Ductal Adenocarcinoma In Vivo Tumor Growth. Neoplasia. 2018 Feb;20(2):152–64.

177. Akiyama H, Umezawa Y, Watanabe D, Okada K, Ishida S, Nogami A, et al. Inhibition of

USP9X Downregulates JAK2-V617F and Induces Apoptosis Synergistically with BH3
Mimetics Preferentially in Ruxolitinib-Persistent JAK2-V617F-Positive Leukemic Cells.
Cancers [Internet]. 2020 Feb 10;12(2). Available from:
http://dx.doi.org/10.3390/cancers12020406

178. Peterson LF, Sun H, Liu Y, Potu H, Kandarpa M, Ermann M, et al. Targeting

deubiquitinase activity with a novel small-molecule inhibitor as therapy for B-cell
malignancies. Blood. 2015 Jun 4;125(23):3588–97.

179. Potu H, Peterson LF, Pal A, Verhaegen M, Cao J, Talpaz M, et al. Usp5 links suppression

of p53 and FAS levels in melanoma to the BRAF pathway. Oncotarget. 2014 Jul
30;5(14):5559–69.

180. Clancy A, Heride C, Pinto-Fernández A, Elcocks H, Kallinos A, Kayser-Bricker KJ, et al.

The deubiquitylase USP9X controls ribosomal stalling. J Cell Biol [Internet]. 2021 Mar
1;220(3). Available from: http://dx.doi.org/10.1083/jcb.202004211

105

181. Lopez J, Bonsor DA, Sale MJ, Urisman A, Mehalko JL, Cabanski-Dunning M, et al. The

ribosomal S6 kinase 2 (RSK2)-SPRED2 complex regulates the phosphorylation of RSK
substrates and MAPK signaling. J Biol Chem. 2023 Jun;299(6):104789.

182. Chalkley RJ, Baker PR. Use of a glycosylation site database to improve glycopeptide

identification from complex mixtures. Anal Bioanal Chem. 2017 Jan;409(2):571–7.

183. Pino LK, Searle BC, Bollinger JG, Nunn B, MacLean B, MacCoss MJ. The Skyline

ecosystem: Informatics for quantitative mass spectrometry proteomics. Mass Spectrom Rev.
2020 May;39(3):229–44.

184. Choi M, Chang C-Y, Clough T, Broudy D, Killeen T, MacLean B, et al. MSstats: an R

package for statistical analysis of quantitative mass spectrometry-based proteomic
experiments. Bioinformatics. 2014 Sep 1;30(17):2524–6.

185. Dempster JM, Rossen J, Kazachkova M, Pan J, Kugener G, Root DE, et al. Extracting

Biological Insights from the Project Achilles Genome-Scale CRISPR Screens in Cancer
Cell Lines [Internet]. bioRxiv. 2019 [cited 2023 Oct 25]. p. 720243. Available from:
https://www.biorxiv.org/content/10.1101/720243v1

186. Perez-Riverol Y, Bai J, Bandla C, García-Seisdedos D, Hewapathirana S,

Kamatchinathan S, et al. The PRIDE database resources in 2022: a hub for mass
spectrometry-based proteomics evidences. Nucleic Acids Res. 2022 Jan 7;50(D1):D543–52.

187. Fumarola C, Bonelli MA, Petronini PG, Alfieri RR. Targeting PI3K/AKT/mTOR

pathway in non small cell lung cancer. Biochem Pharmacol. 2014 Aug 1;90(3):197–207.

106

VITA

Amanda Riley grew up in Cary, North Carolina. She received her Bachelor of Science degree

in Biology,

summa cum laude

, and graduated as valedictorian of her class at Muhlenberg College

in 2016. As an undergraduate, she gained research experience during summer internships. In the

summer of 2014, she worked at Novartis Vaccines (now Seqirus) in the Manufacturing Science

and Technology Division. There, she worked in an analytical chemistry lab and gained experience

using LC-MS to test influenza vaccine samples for hemagglutinin protein levels. In the summer of

2015, she worked at Duke University in Dr. David Hinton’s lab in the Nicolas School of the

Environment. In the Hinton Lab, she studied the effects of pollution adaptation on liver

development in killifish. From this work, she published her first peer-reviewed, first-author journal

article. After graduating from college in 2016, Amanda worked as a research technician in Dr.

Cyril Benes’ lab at the Massachusetts General Hospital Cancer Center in Boston, Massachusetts.

In the Benes lab, she generated patient-derived cell lines to study resistance mechanisms to

tyrosine kinase inhibitors. From this work, she is recognized as a co-author on five peer-reviewed

publications. She enrolled in the Molecular and Cellular Biology PhD program at the University

of Washington in 2018 and joined Dr. Alice Berger’s lab at the Fred Hutch Cancer Center in 2019.

During her dissertation work, she applied her skills in biochemistry and cell biology to understand

genetic vulnerabilities in lung cancer. Throughout her time in graduate school, she received many

accolades for her work, including two poster presentation prizes (including first place and the

coveted Galloway Cup!). She also received several competitive training grants, including an F31

grant from the NIH. She was recognized as a Promising Young Scientist in an article from the

Brotman Baty Institute, which highlighted her achievements and academic successes. In her free

time, she enjoys yoga, running, swimming, live music, and outdoor picnics.