the scope, progress, results and costs of developing our product candidates or any other future product candidates, and conducting preclinical studies and clinical trials, including our RMC-4630 Phase 1/2 clinical program;

the scope, progress, results and costs related to the research and development of our pipeline;

the timing of and costs involved in obtaining and maintaining regulatory approval for any of current or future product candidates, and any related restrictions, limitations and/or warnings in the label of an approved product candidate;

our expectations regarding the potential market size and size of the potential patient populations for RMC-4630, our other product candidates and any future product candidates, if approved for commercial use;

our ability to maintain existing and establish new collaborations, licensing or other arrangements and the financial terms of any such agreements, including our collaboration with Sanofi;

our commercialization, marketing and manufacturing capabilities and expectations;

the rate and degree of market acceptance of our product candidates, as well as the pricing and reimbursement of our product candidates, if approved;

the implementation of our business model and strategic plans for our business, product candidates and technology, including additional indications for which we may pursue;

the scope of protection we are able to establish and maintain for intellectual property rights covering our product candidates, including the projected term of patent protection;

estimates of our expenses, future revenue, capital requirements, our needs for additional financing and our ability to obtain additional capital;

developments and projections relating to our competitors and our industry, including competing therapies and procedures;

regulatory and legal developments in the United States and foreign countries;

the performance of our third-party suppliers and manufacturers;

our ability to attract and retain key scientific or management personnel;

our expectations regarding the period during which we will qualify as an emerging growth company under the Jumpstart Our Business Startups Act of 2012;

our expectations regarding our ability toto obtain, maintain, enforce and defend our intellectual property protection for our product candidates; and

other risks and uncertainties, including those listed under the caption “Risk Factors.”

The COVID-19 pandemic, or other epidemic and pandemic diseases or governmental or other actions taken in response to them, could significantly disrupt our business.

We are a clinical-stage precision oncology company with a limited operating history and no products approved for commercial sale. We have incurred significant losses since inception. We expect to incur losses for at least the next several years and may never achieve or maintain profitability, which, together with our limited operating history, makes it difficult to assess our future viability.

We have never generated revenue from product sales and may never be profitable.

We will require substantial additional financing to achieve our goals, which may not be available on acceptable terms, or at all. A failure to obtain this necessary capital when needed could force us to delay, limit, reduce or terminate our product development or commercialization efforts.

We are early in our development efforts. Our business is dependent on the successful development of our current and future product candidates. If we are unable to advance our current or future product candidates through clinical trials, obtain marketing approval and ultimately commercialize any product candidates we develop, or experience significant delays in doing so, our business will be materially harmed.

Preclinical development is uncertain. Our preclinical programs may experience delays or may never advance to clinical trials, which would adversely affect our ability to obtain regulatory approvals or commercialize our product candidates on a timely basis or at all, which would have an adverse effect on our business.

Historically, direct inhibition of any RAS protein has been challenging due to a lack of tractable, or “druggable,” binding pockets and we are not aware of any programs in clinical development that have successfully targeted any RAS(ON) protein. Given this approach is unproven, it may not be successful.

The results of preclinical studies and early-stage clinical trials may not be predictive of future results.

If we encounter difficulties enrolling patients in our clinical trials, our clinical development activities could be delayed or otherwise be adversely affected.

We are dependent on our collaboration with Sanofi for the development of RMC-4630 and may depend on Sanofi for the development and commercialization of any other future SHP2 inhibitor product candidates. Under certain circumstances, Sanofi may unilaterally terminate the collaboration for convenience, which would materially and adversely affect our business.

We are currently developing and may, in the future, develop RMC-4630 and other product candidates in combination with other therapies, which exposes us to additional risks.

We face significant competition, and if our competitors develop and market products that are more effective, safer or less expensive than the product candidates we develop, our commercial opportunities will be negatively impacted.

If we and our collaborators are unable to obtain and maintain sufficient patent and other intellectual property protection for our product candidates and technology, our competitors could develop and commercialize products and technology similar or identical to ours, and we may not be able to compete effectively in our market or successfully commercialize any product candidates we may develop.

Establish tri-complex inhibition of RAS(ON) as a first-in-class therapeutic modality. There are dozens of RAS variants that previously have been difficult or

3

impossible to drug. We believe our current and future product candidates, when used in specialized dosing paradigms and rational in-pathway combinations, will have the potential to promote profound and sustainable clinical benefit, combat adaptive resistance mechanisms and, in some cases, supplant the current standardimplicated as molecular drivers of care for patients with tumors driven by these pathways.

Our most advanced product candidate, RMC-4630, is a potent and selective inhibitor of SHP2, based on preclinical evidence described in this Annual Report on Form 10-K. SHP2 is a protein that plays a central role in modulating cell survival and growth by transmitting signals from upstream receptor tyrosine kinases, or RTKs, to RAS. In collaboration with Sanofi, we are evaluating RMC-4630 in a multi-cohort Phase 1/2 clinical program, which includes our ongoing Phase 1 study of RMC-4630 as monotherapy in patients with advanced cancers, including those with tumors harboring genetically defined mutations in the RAS signaling pathway and our ongoing Phase 1b/2 study of RMC-4630 in combination with the MEK inhibitor cobimetinib. Based on our own data, and supported by observations by others, we are evaluating intermittent dosing schedules in our clinical program to allow us to maximize dose intensity in order to achieve the greatest depth of response. We also plan to explore the potential clinical benefit of RMC-4630 in combination with other in-pathway agents such as RTK (initially epidermal growth factor receptor, or EGFR), KRAS^G12C inhibitors, and ERK inhibitors as well as in combination with PD-1 inhibitors. In November 2019, we entered into an agreement with Amgen to evaluate the combination of RMC-4630 and Amgen’s KRAS^G12C(OFF) inhibitor, AMG 510, in a Phase 1b trial that will be conducted by Amgen. Site activation, including appropriate institutional review board approval has begun (www.clinicaltrials.gov). Patient dosing has not started and may be delayed due to the COVID-19 pandemic. In March 2020, the Pancreatic Cancer Collective (a strategic partnership between Lustgarten Foundation and Stand Up To Cancer) announced that it had awarded funding to the Netherlands Cancer Institute, or the NKI, for its study using our SHP2 inhibitor, RMC-4630 and an investigational ERK inhibitor in patients with pancreatic cancer. We would provide RMC-4630 to support this study.  Although we are at an early stage of clinical testing and product candidate development, we believe RMC-4630 is well-positioned to become the backbone of targeted therapy combinations for the treatment of various RAS-dependent tumors. Under our collaboration with Sanofi on our SHP2 program, we have a 50-50 profit share and a co-promote right in the United States and are eligible to receive royalties on net sales outside of the United States. Sanofi is responsible for reimbursing substantially all of our research costs and all of our development costs for the SHP2 program.

We are also developing a portfolio of what we believesmall molecules designed to be the first potent, selective and cell-active inhibitors of mutant RAS(ON) proteins. Historically, direct inhibition of any RAS protein has been challenging due to a lack of tractable, or “druggable,” binding pockets. Recently, selective inhibitors of inactive, GDP-boundtarget multiple oncogenic forms of RAS, or RAS(OFF), have demonstrated encouraging preliminary anti-tumor effects and thus provide clinical validation for targeting mutant RAS in cancer. Our small molecule inhibitors of mutant RAS(ON) that are derived from our proprietary tri-complex technology platform, which enables us to target proteins lacking intrinsic drug binding sites by inducing new druggable pockets. Initially, we will prioritize four mutant RAS(ON) targets—KRAS^G12C, KRAS^G13C, KRAS^G12D and NRAS^G12C—and expect to nominate our first development candidate in 2020. We plan to evaluate our RAS(ON) inhibitors alone and in combination with other drugs and investigational new drugs, particularly in-pathway agents.platform. We believe that targeted inhibition of various oncogenic RAS(ON) mutantsvariants represents a highly differentiated approach for treating the large population of patients with diverse RAS mutations, including non-small cell lung cancer, or NSCLC, colorectal, pancreatic and other cancers.

Establish our proprietary SHP2 inhibitor, RMC-4630, as the backbone of targeted therapy combinations for the treatment of RAS-dependent tumors. RMC-4630 is an allosteric inhibitor of SHP2, a convergent node within the oncogenic RAS-signaling pathway. We are evaluating RMC-4630 as both a monotherapy and as a backbone of combination therapies designed to maximize clinical benefit by combatting drug resistance pathways that depend on SHP2. In the future, we may leverage our complementary pipeline of RAS(ON) and RAS Companion Inhibitors by evaluating RMC-4630 in combination with certain RAS(ON) inhibitors, which could include RMC-6291 or RMC-6236.

Demonstrate clinical utility of our RAS Companion Inhibitors targeting other key nodes in the RASmTORC1/4EBP1 (RMC-5552) and mTOR signaling pathways. Our program targeting SOS1 a protein that plays a key role in converting RAS(OFF) to RAS(ON) in cells, is currently in lead generation stage. In addition, our preclinical development candidate,(RMC-5845) through precision oncology. RMC-5552 is designed as a potent, selective inhibitor of mTORC1/4EBP1 that we expect to selectivelyevaluate in a clinical trial beginning as a monotherapy in the first of half 2021. In initial clinical trials, we plan to test RMC-5552 both as a monotherapy in patients with cancers harboring genotypes linked to hyperactivated mTORC1 signaling, and deeply inhibit mTORC1, thereby preventing phosphorylationas a combination partner for RAS inhibitors in patients with cancers harboring one or more RAS mutations and inactivation of 4EBP1, a downstream proteinco-occurring mutations in the mTOR signaling pathwaypathway. RMC-5845 is designed as a potent, selective inhibitor of SOS1 that normally suppresses expression of certain oncogenes such as C-MYC. We advanced RMC-5552 into IND-enabling development in June 2019.

Our management team has significant experience in oncology and in progressing products from early stage research to clinical trials, and ultimately to regulatory approval and commercialization. Dr. Steve Kelsey, our President of Research and Development, was previously President of Onkaido Therapeutics, a Moderna venture focused on oncology mRNA therapeutics, and has held senior positions at Medivation, Geron and Genentech, where he played a significant roleis in the developmentIND-enabling stage of Perjeta, Kadcyla and Erivedge. Our President and Chief Executive Officer, Dr. Mark Goldsmith, served as Chief Executive Officer of Constellation Pharmaceuticals, where he led the creation of its oncology pipeline and drove the development of a strategic alliance with Genentech. He also has led four other companies spanning early discovery through development, including Global Blood Therapeutics, where he led the discovery and early development of voxelotor.

Our company was founded and continues to be supported by three world-class scientific advisors: Dr. Kevan Shokat (Professor and Chair of the Department of Cellular and Molecular Pharmacology at University of California, San Francisco, Professor of Chemistry at the University of California, Berkeley and an investigator at the Howard Hughes Medical Institute), Dr. Martin Burke (Professor of Chemistry at the University of Illinois at Urbana-Champaign) and Dr. Michael Fischbach (Associate Professor in the Department of Bioengineering at Stanford University and a Stanford ChEM-H Institute Scholar). Dr. Shokat is widely recognized for his seminal contributions to the field of kinase biology, using chemistry, protein engineering and genetic tools to pioneer novel therapeutic approaches to target key signaling pathways in cancer. He led the discovery of the first KRAS^G12C(OFF) inhibitor.

4

Our strategy

Our goal is to develop novel targeted therapies to outsmart cancer for the benefit of patients.preclinical development. We plan to pursue the following strategies:

5

Our opportunity: multiple large unmet needs in RAS-addicted cancers with driver mutations in notorious growth and survival pathways

Background

The RAS and mTOR signaling cascades are among the most frequently exploited by human cancers. Cancer cells often carry mutations in proteins in these pathways that subvert normal cell growth and survival by causing excessive or aberrant signal transduction. These proteins can be directly or indirectly involved in signal transduction. For example, many tumors of different types exhibit excessive activation of the RAS signaling cascade as a result of mutations in RTKs, RAS, NF1 and/or RAF.

We have built a portfolio of compounds that inhibit select signaling nodes within the RAS and mTOR pathways. To date, our discovery and development efforts have focused on SHP2, RAS, SOS1 and mTORC1 (these targets are shaded dark green or blue in the figure above).

SHP2

SHP2 is a protein tyrosine phosphatase that plays a critical role in the transduction of intracellular signals downstream of a wide variety of RTK growth factor receptors to promote cell survival and growth. SHP2 acts as a central signaling node that regulates growth signals within normal cells and, in certain circumstances, cancer cells. Some mutant forms of RAS, such as KRAS^G12C and KRAS^G12A, exert their oncogenic effects by amplifying or exaggerating normal RTK-mediated growth signals transmitted via SHP2, and as a result they can be suppressed by inhibiting SHP2. There are other cancer-causing mutations that result in, or are dependent upon, activation of wild-type RAS and are likewise dependent on SHP2, including amplification of wild-type RAS or mutations in the gene encoding the GTPase-activating protein (GAP) neurofibromin 1 (NF1) which reduce activity of NF1 (so called NF1 loss-of-function or NF1^LOF), and class 3 mutations in the downstream effector BRAF (BRAF^Class3).

6

RAS

RAS proteins drive normal cell proliferation, differentiation and survival in response to growth factors acting through RTKs, and they can also be direct drivers of cancer. Normally RAS proteins cycle between an inactive form (RAS(OFF)), which is bound to GDP and unable to transmit signals, and an active conformation (RAS(ON)), which is induced upon binding GTP in response to growth factor receptor stimulation, causing it to become competent to interact physically with downstream effector proteins such as RAF. The magnitude of cell signals transmitted by the RAS activation cycle is proportional to the intracellular level of RAS(ON). In a healthy, normal cell RAS(ON) represents a small fraction of the total RAS pool within a cell. Signals arising from RTKs upstream of the RAS cycle act through SHP2 to promote the substitution of GTP for GDP in association with RAS, thereby increasing RAS(ON) levels. In cancers with abnormally elevated RTK activity, increased signaling via the RAS activation cycle is a major driver of tumor cell growth. Likewise, oncogenic mutations of RAS itself result in a significant slowing of the enzymatic conversion of RAS-bound GTP to GDP and thus drive cancer by raising RAS(ON) significantly above normal levels. In some cells harboring a KRAS^G12C mutation, 80% or more of cellular KRAS^G12C is in the GTP-bound state (RAS(ON)), representing a >15-fold increase compared to wild-type KRAS. As a general principle, RAS-dependent cancer cells exploit a high level of RAS(ON) for continued survival and growth.

SOS1

SOS1 is a member of a family of proteins that activate RAS. SOS1 directly activates RAS proteins by promoting the release of tightly bound GDP and facilitating the binding of GTP, which is present at a much higher intracellular concentrations than GDP, to generate RAS(ON). SOS1 itself is activated by RAS through the binding of RAS(ON) to an allosteric site on the SOS1 protein. As a result, there is a positive feedback loop between SOS1 and RAS that increases RAS signaling. The activation of RAS by SOS1 is ‘processive’; that is, once a single molecule of SOS1 is activated it can sequentially activate multiple RAS molecules until it eventually becomes inactive.

4EBP1/mTORC1

mTORC1 and mTORC2 are large protein complexes that share mTOR kinase but contain distinct additional components and cellular functions. mTORC1 is a critical regulator of metabolism, growth and proliferation within cells, including cancer cells. Two of the main substrates of mTORC1 are eukaryotic initiation factor 4E-binding protein 1 (4EBP1) and ribosomal S6 kinase (S6K). Under resting conditions non-phosphorylated 4EBP1 functions as a suppressor of the translation of proteins that are required for cell growth, proliferation and survival. Phosphorylation of 4EBP1 by activated mTORC1 inhibits this suppressive regulatory function and thereby upregulates translation of these proteins. One of the most important proteins regulated by 4EBP1 is the oncogenic protein C-MYC, which for many years has been thought to be central to cancer cell growth and survival. The abnormal activation of mTORC1, and subsequent inactivation of the tumor suppressor 4EBP1, is a mechanism that is frequently harnessed by cancer cells to gain a growth and proliferation advantage over normal cells. A number of upstream proteins in the mTOR signaling pathway that regulate mTORC1 activity are frequently mutated, or deleted, in cancer cells, resulting in increased activation of mTORC1 and upregulation of translation; these targets of mutation include PTEN, PI3 kinase (PI3K) and hamartin (TSC1) and tuberin (TSC2).

RAS mutant epidemiology in the United States

MutationsVariants in RAS proteins account for approximately 30% of all human cancers in the United States, many of which are fatal. Diverse oncogenic RAS mutationsvariants in three different RAS isoforms (KRAS, NRAS and HRAS) drive distinct human cancers. KRAS mutations are commonly found in NSCLC and account for approximately 85% of RAS-mutant cancers. NRAS mutations are commonly found in melanoma and account for approximately 11% of RAS-mutant cancers. HRAS mutations are commonly found in bladder cancer and account for 4% of RAS-mutant cancers. The tableFigure 1 below summarizes the frequency of RAS cancer mutationsestimated new RAS-addicted cancers diagnosed each year in the United States. There continues toUS organized into convenience groups based on tumor genetics. Based on these data, we believe there are an estimated 230,000 new cancers diagnosed per year that could potentially be a high unmet medical need for patients bearing tumors with these mutations. We believe our programs, including our SHP2,addressed by RAS and SOS1 inhibitors may be useful in addressing this unmet medical need.

7

Projected number of new cases in 2019 in the United States (frequency % shown in parentheses below)US alone.

Figure 1.

Limitations of approved drugs treating RAS-dependent cancers and our opportunity

A major goal in contemporary oncology treatment is to replace relatively unselective chemotherapy regimens—which in many cases remain the standard of care today but provide only partial benefit with many side effects—with more effective and better tolerated targeted therapeutic options. Targeted therapies directed against RAS-dependent cancers, which include drugs that inhibit RTKs, RAF and MEK, have been approved for use in lung cancer, melanoma and colorectal cancer. These targeted therapies have shown the capacity to drive deeper and more durable responses than conventional chemotherapy regimens while minimizing unwanted side effects and damage to normal tissues.

Two treatment gaps remain in RAS-dependent cancers. First, several oncogenic proteins are not addressed by current targeted therapies. Second, cancers driven by oncogenic proteins that are addressed by current targeted therapies often progress in the face of drug therapy due to adaptive resistance mechanisms.

Specific examples of frontier cancer drivers are RAS, NF1, and selected RAF mutants (BRAF^Class3). Historically, direct inhibition of any RAS protein has been challenging due to a lack of tractable, or “druggable,” binding pockets. Recently reported initial clinical results from two RAS(OFF) inhibitors targeting mutant KRAS^G12C suggest significant clinical benefit and provide strong pharmacologic validation of this oncoprotein as a cancer driver. These results, along with other preclinical data, provide a compelling basis for our commitment to targeting oncogenic mutant forms of RAS(ON). We are using our innovation engine to develop a portfolio of mutant-selective RAS(ON) inhibitors and, initially, we will prioritize four mutant RAS(ON) targets—KRAS^G12C, KRAS^G13C, KRAS^G12D and NRAS^G12C.

A common source of treatment failure with existing targeted therapies is that cancer cells exhibiting “oncogene addiction” exploit cell signaling circuitry to bypass the drug’s effect and sustain growth and survival. This phenomenon is particularly well recognized in RAS-dependent cancers, and may be especially active in certain tumor histotypes, making them less sensitive to a drug from the outset and/or more likely to progress over time. Certain RAS-dependent cancers have been treated with two RAS pathway targeted agents to achieve combinatorial benefit by attenuating adaptive resistance mechanisms. For example, melanomas driven by BRAF^Class1 mutations can be treated with a combination of a BRAF inhibitor and a MEK inhibitor. SHP2 is believed to be a central node that can be targeted to disrupt bypass signaling pathways that may involve activation of multiple RTKs. Therefore, although we are at an early stage of clinical testing and product candidate development, we believe RMC-4630, a potent and selective inhibitor of SHP2, is well positioned to become the backbone of targeted therapy combinations for the treatment of various RAS-dependent tumors, and plan to explore this paradigm in our ongoing RMC-4630 clinical program.

We are using our innovation engine to develop novel targeted therapies and combination regimens to address these treatment gaps.

8

Our innovation engine

We have built an innovation engine that enables us to discover and develop novel targeted therapies for elusive high-value frontier cancer targets with particular focus on a cohesive set of disease targets within notorious growth and survival pathways. This engine consists ofis centered around our proprietary tri-complex platform and is bolstered by three complementary drivers:pillars:

Our chemical biology and cancer pharmacology know-howTri-complex platform

We test our inhibitors across a diverse set of human cancer cell and patient-derived in vitro and/or in vivo models of cancer. This is complemented by targeted implementation of bioinformatics and functional genomics. The biological insights we generate helpOur proprietary tri-complex technology enables us to unraveldiscover small molecule inhibitors of targets lacking intrinsic drug binding sites by inducing new druggable pockets. This occurs through small molecule-driven formation of a high affinity ternary complex (tri-complex) between the complex molecular circuitry in human cancers. We also explore mechanismstarget protein, the small molecule, and a widely expressed cytosolic protein called a chaperone (e.g., FKPB12 or cyclophilin A). This platform technology is the foundation of adaptive resistance that RAS-addicted cancer cells use to circumvent inhibitionour RAS(ON) Inhibitor programs. In this context, the inhibitory effect of tri-complex formation on the RAS(ON) target is mediated by steric occlusion of the pathway,site where RAS(ON) binds its downstream effector molecules, such as RAF, which are required for propagating the oncogenic signal. Thus, tri-complex formation with RAS(ON) targets disrupts RAS effector binding and develop innovative mechanism-based dosing paradigms and rational in-pathway combinations with our proprietary compounds and/or other agents. We evaluate such dosing and combination approaches in our preclinical in vitro and in vivo models to define their pharmacologic opportunities and limitations, and to prioritize therapeutic strategies for translation to the clinic.

terminates oncogenic signaling. Our structure-based drug discovery capabilities

We enlist various technologies and capabilities that give us chemical access to challenging sites thatRAS(ON) tri-complex inhibitors, which are generally not accessible using conventional small molecule drug discovery approaches. For each target, we consider the specific structural, physico-chemical, functional and dynamic properties of the target and deploy the approach(es) that appears most likely to yield viable development candidates. In some instances the compounds we discover and developinspired by natural products, are small molecules (e.g., less than 500 mw) with properties that generally satisfy conventional pharmaceutical “Rule of 5” criteria, while in other cases, they are larger (e.g., 500-1000 mw) “Beyond Rule of 5” compounds. In either case, we use various structure-based design tools to discover the initial chemical matter, drive optimization using iterative medicinal chemistry, and generate structure-activity and structure-property relationships to identify development candidates. In order to prosecute effective medicinal chemistry campaigns within complex chemical space, we use our deep experience and make the necessary investments to design and develop scalable modular chemical synthesis, purification and analytical methods for selected scaffolds to routinely and efficiently analogue our “Beyond Rule of 5” chemical series.

Although we are at an early stage of clinical testing and product candidate development, we believe our differentiated chemical approaches to discovering inhibitors for challenging frontier cancer targets

6

Pipeline

Our pipeline is exemplified by our current portfolio. Each of our current programs takes advantage of allosteric regulation to inhibit the target of interest by exploiting one or more of three distinct mechanisms. We use the term allosteric inhibitors to describe those that “act at a distance,” meaning that the inhibitory effect occurs at a protein site or domain distinct from the compound’s binding site.summarized below:

RAS(ON) Inhibitors

Overview

Our RAS(ON) Inhibitors are based on our proprietary tri-complex technology platform, which enables a highly differentiated approach to inhibiting the active, GTP-bound form of RAS, or RAS(ON). We are developing a portfolio of compounds that we believe are the first and only RAS(ON) inhibitors to use this mechanism of action. Our RAS(ON) compounds, RMC-6291 and RMC-6236, are in the IND-enabling stage of preclinical development as KRAS^G12C(ON)/NRAS^G12C (ON) and RAS^MULTI(ON) inhibitors, respectively. These inhibitors have exhibited anti-tumor activity in vivo in preclinical models. We believe that direct inhibitors of RAS(ON) will suppress cell growth and survival and be less susceptible to adaptive resistance mechanisms recognized for RAS(OFF) inhibitors. We plan to evaluate our RAS(ON) Inhibitors alone and in combination with other drugs and investigational new drugs, particularly in-pathway agents. We also plan to nominate a third RAS(ON) Inhibitor as a development candidate in the second half of 2021.

We believe tailored RAS(ON) inhibitors will be necessary in order to serve the diverse landscape of RAS-driven cancers. For some RAS variants like KRAS^G13C and KRAS^G12D, we are pursuing additional RAS mutant-selective inhibitors in parallel with the IND-enabling development of RMC-6291 and RMC-6236. For variants like KRAS^G12V and KRAS^G12A, we believe discovery of a mutant-selective inhibitor is unlikely, and we plan to focus on optimizing therapeutic regimens for cancers bearing these mutations using our RAS^MULTI(ON) inhibitor.

RMC-6291

RMC-6291, our first mutant-selective RAS(ON) inhibitor development candidate, is in the IND-enabling stage of preclinical development and we currently plan to submit an IND for this drug candidate in the first half of 2022. RMC-6291 is designed as a first-in-class, potent, oral and selective tri-complex inhibitor of KRAS^G12C/NRAS^G12C(ON). It exhibits subnanomolar potency for suppressing RAS pathway signaling and growth of KRAS^G12C-bearing cancer cells and is engineered to be highly selective for KRAS^G12C/NRAS^G12C over wild type RAS and other cellular targets. We believe that, in conjunction with its drug-like properties including oral bioavailability across multiple species, the results from our preclinical studies to date evaluating RMC-6291 support advancement to clinical development.

We believe that RMC-6291 is differentiated by its mechanism of inhibiting the KRAS^G12C(ON) form. KRAS^G12C(OFF) inhibitors have been shown to disrupt RAS-RAF signaling by binding to and sequestering the RAS(OFF) form; their activity is dictated by

7

availability of the KRAS^G12C(OFF) target, which is inherently limited by the turnover rate of RAS(ON) to RAS(OFF) forms. In contrast, consistent with its direct effect on KRAS^G12C(ON) proteins, disruption of RAS-RAF interaction was observed to be nearly instantaneous in cells exposed to RMC-6291 (Figure 2).

Figure 2. RMC-6291 drove immediate termination of RAS signaling in tumor cell line.

Cellular characterization of the disruption of interaction between KRAS^G12C and the RAS binding domain, or RBD, of CRAF by KRAS^G12C(OFF) inhibitors AMG 510 and MRTX849, and the KRAS^G12C(ON) inhibitor RMC-6291. An engineered NanoBRET system was used in U2OS cells with a nanoluciferase-KRAS^G12C fusion and HaloTag-CRAF RBD fusion labeled with Halo618. 100% interaction is defined by the BRET signal before addition of inhibitors. Results of a single experiment are shown, and are representative of results from three independent experiments.

Oral administration of RMC-6291 produced deep and durable suppression of RAS pathway activity in KRAS^G12C tumor models and drove profound tumor regressions in vivo at well-tolerated doses. Across multiple tumor xenograft models, we observed that RMC-6291 at equivalent, and sometimes lower, doses outperformed selected KRAS^G12C(OFF) inhibitors by driving deeper and more durable anti-tumor effects (Figures 3-4).

Figure 3. RMC-6291 caused deep regressions in preclinical xenograft model of tumors harboring KRAS^G12C.

8

Anti-tumor activity of RMC-6291 in non-small cell lung cancer cell line-derived (NCI-H358, KRAS^G12C) xenograft model in mice. Data represent (left) individual end of treatment responses, with tumor volume expressed as a percentage of initial tumor volume at time of study start (truncated at 300%) and (right) time to post-treatment tumor re-growth expressed as the percent of tumors in each group with tumor volumes less than 200 mm³ over time, as measured by number of days post-implant. Number of mice per group = 10. In (left) each animal is represented as a separate bar. Numbers indicate number of complete regressions (CR, defined as ≥ 80% reduction in tumor volume from initial volume) in each group. po = per os (oral administration) and qd = daily.

Figure 4. RMC-6291 caused regressions in patient-derived xenograft model of tumors harboring KRAS^G12C.

Anti-tumor activity of RMC-6291 in non-small cell lung cancer patient-derived (LUN092, KRAS^G12C) xenograft model in mice. Data represent (upper left) mean tumor volume over time, (upper right) waterfall plot of individual end of study responses, with tumor volume expressed as a percentage of initial tumor volume at time of study start (truncated at 300%) and (bottom) percent body weight change over time. Number of mice per group = 10. In (upper left) and (bottom), data represent mean and errors bars represent standard error of the mean. In (upper right) each animal is represented as a separate bar. Numbers indicate number of regressions (R, defined as ≥ 10% reduction in tumor volume from initial volume) in each group. po = per os (oral administration) and qd = daily.

RMC-6236

RMC-6236, our first RAS^MULTI(ON) inhibitor development candidate, is also in the IND-enabling stage of preclinical development and we currently plan to submit an IND for this drug candidate in the first half of 2022. RMC-6236 is designed as a first-in-class, potent, oral, RAS-selective tri-complex inhibitor of multiple RAS(ON) variants including KRAS^G12V(ON) and KRAS^G12D(ON). In preclinical studies, it exhibited low nanomolar inhibition of both RAS pathway signaling and growth in RAS-dependent cells and a selectivity window of more than 1,000 times over RAS-independent cells (Figure 5). In addition, RMC-6236 has a drug-like profile, including oral bioavailability across multiple species, based on preclinical results. We believe the results from our preclinical studies evaluating the RMC-6236 support advancement to clinical development.

Figure 5. RMC-6236 inhibited proliferation in vitro in RAS-dependent tumor cell lines.

9

Cellular characterization of the effect of RMC-6291 on the proliferation of a variety of RAS-dependent and RAS-independent cell lines in preclinical studies. Cell proliferation was monitored in 2D cell cultures using CellTiter-Glo (CTG). Data shown represent the mean of at least two independent preclinical studies, each performed in duplicate (error bars show the standard deviation).

We believe the preclinical results for RMC-6236 show that this product candidate has the potential to treat patients whose cancers harbor a range of RAS variants. First, RMC-6236 may have the potential to be used to treat cancers driven by RAS variants for which we believe a mutant-selective inhibitor is unlikely to be discovered based on the properties of the mutated amino acid side chains, such as the very common KRAS^G12V (Figure 6). In this context, oral administration of RMC-6236 was observed to produce deep and durable suppression of RAS pathway activity in a KRAS^G12V tumor model and to have driven profound tumor regressions in vivo at well-tolerated doses in NSCLC colorectal and pancreatic KRAS^G12V tumor models in mice (Figures 7-9). RMC-6236 may also be useful for targeting RAS variants for which we believe mutant-selective inhibitors are likely to be discovered in the future, such as KRAS^G12D. In a pancreatic cancer tumor xenograft model with the KRAS^G12D variant, oral administration of RMC-6236 drove profound tumor regressions (Figure 10). Across all tumor models presented in Figures 8-10, RMC-6236 was well-tolerated at the dose regimens shown. In our preclinical evaluation, RMC-6236 exhibited significant anti-tumor activity at well-tolerated doses in murine tumor models.

Figure 6. Numerous unmet needs in RAS-addicted cancers may be served by a RAS^MULTI inhibitor.

Figure 7. RMC-6236 drove sustained, dose-dependent suppression of the RAS pathway in vivo.

Oral administration of a single dose of RMC-6236 at 10 mg/kg (blue) or 25 mg/kg (green) produced a dose-dependent inhibition of tumor DUSP6 mRNA relative to control tumors in the cell line-derived (NCI-H4411, KRAS^G12V) xenograft model. Number of mice per group = 3. Data mean and errors bars represent standard error of the mean. po = per os (oral administration).

10

Figure 8. RMC-6236 caused regressions in preclinical xenograft model of tumors harboring KRAS^G12V .

Anti-tumor activity of RMC-6236 in non-small cell lung cancer cell line-derived xenograft NCI-H441 KRAS^G12V xenograft model in mice. Data represent (upper left) mean tumor volume over time, (upper right) waterfall plot of individual end of study responses, with tumor volume expressed as a percentage of initial tumor volume at time of study start (truncated at 300%), and (bottom) percent body weight change over time. Number of mice per group = 10. In (upper left) and (bottom), data represent mean and errors bars represent standard error of the mean. In (upper right) each animal is represented as a separate bar. Numbers indicate number of regressions (R, defined as ≥ 10% reduction in tumor volume from initial volume) or number of complete regressions (CR, defined as ≥ 80% reduction in tumor volume from initial volume) in each group in each group. po = per os (oral administration) and qd = daily.

Figure 9. RMC-6236 caused regressions in preclinical xenograft model of tumors harboring KRAS^G12V

Anti-tumor activity of RMC-6236 in (left) pancreatic ductal adenocarcinoma (PDAC) cell line-derived (Capan-2, KRAS^G12V) and (right) colorectal cancer cell line-derived (SW404, KRAS^G12V) xenograft models in mice. Data represent waterfall plots of individual end of study responses, with tumor volume expressed as a percentage of initial tumor volume at time of study start (truncated at 300%) Number of mice per group = 8. Each animal is represented as a separate bar. Numbers indicate number of regressions (R, defined as ≥ 10% reduction in tumor volume from initial volume) or number of complete regressions (CR, defined as ≥ 80% reduction in tumor volume from initial volume) in each group. po = per os (oral administration), qd = daily and q2d = every other day.

11

Figure 10. RMC-6236 caused regressions in preclinical xenograft model of tumors harboring KRAS^G12D.

Anti-tumor activity of RMC-6236 in pancreatic ductal adenocarcinoma (PDAC) cell line-derived xenograft (CDX) HPAC KRAS^G12D xenograft model in mice. Data represent (upper left) mean tumor volume over time, (upper right) waterfall plot of individual end of study responses, with tumor volume expressed as a percentage of initial tumor volume at time of study start (truncated at 300%), and (bottom) percent body weight change over time. Number of mice per group = 9 or 10, as indicated. In (upper left) and (bottom), data represent mean and errors bars represent standard error of the mean. In (upper right) each animal is represented as a separate bar. Numbers indicate number of complete regressions (CR, defined as ≥ 80% reduction in tumor volume from initial volume) in each group. po = per os (oral administration), qd = daily and q2d = every other day.

KRAS^G12V and KRAS^G12D are the two most common RAS variants implicated in driving human cancers and together are believed to cause almost 100,000 new cases of cancer each year in the US alone. Beyond direct targeting of numerous RAS(ON) mutants, we believe RMC-6236 has the potential, if approved, to be used to treat cancers driven by wild type RAS, such as those in which amplification of wild type RAS drives cancer growth. Along similar lines, we believe RMC-6236 has the potential to be deployed in combination with other RAS pathway inhibitors to combat various resistance mechanisms that depend on wild type RAS activation.

We have identified potent, cell-active RAS(ON) Inhibitors for variants thought to be responsible for the vast majority of RAS-addicted cancers. Two of these programs, one designed to target KRAS^G13C(ON) selectively and one designed to target KRAS^G12D(ON) selectively, are currently in the lead optimization stage of preclinical development.

RAS Companion Inhibitors

Our RAS Companion Inhibitor RMC-4630 is designed as a potent and selective inhibitor of SHP2. The clinical development program for RMC-4630 currently consists of four active clinical trials, one of which includes two arms studying different combinations:

We also anticipate that RMC-4630 will be studied in combination with:

12

Our precision medicine approach

We interrogate the biology of different cancers and their associated mutational drivers to help inform patient selection, therapeutic treatment regimens, and appropriate outcome measures. To identify patient subsets that may benefit most from our treatment strategies, we use genomics, transcriptomics and proteomics data from human tumor samples and/or broad panels of human cancer cell lines. We also pursue development of drug combinations where combinatorial benefits are predicted and confirmed in preclinical models. Innovative, mechanism-based dosing paradigms are explored for each combination and refined using pharmacokinetic and pharmacodynamic modeling techniques and sophisticated continuous reassessment dosing methodology. We also identify and monitor pharmacodynamic biomarkers and surrogates of clinical activity to help measure target inhibition.

Our pipeline

We are using our innovation engine to develop a deep pipeline of novel targeted therapies to inhibit elusive, high-value frontier targets within the notorious RAS and mTOR signaling pathways. Our pipeline includes one product candidate that is in clinical development and all of our other programs are in the preclinical stage. Under our collaboration with Sanofi on our SHP2 program,RMC-4630, we have a 50-50 profit share and a co-promote right in the United States and are eligible to receive royalties on net sales outside of the United States. Sanofi is responsible for reimbursing substantially all of our research costs and all of our development costs for the SHP2 program. For all other programs, we retain worldwide commercial rights.

10

Our SHP2 inhibitor, RMC-4630

Overview

Our most advanced product candidate, RMC-4630, is a potent and selective inhibitor of SHP2, based on preclinical evidence described in this Annual Report on Form 10-K. SHP2 is a protein that plays a central role in modulating cell survival and growth by transmitting signals from upstream RTKs to RAS. In collaboration with Sanofi, we are evaluating RMC-4630 in a multi-cohort Phase 1/2 clinical program, which includes our ongoing Phase 1 study of RMC-4630 as monotherapy in patients with advanced cancers, including those with tumors harboring genetically defined mutations in the RAS signaling pathway, and our ongoing Phase 1b/2 study of RMC-4630 in combination with the MEK inhibitor cobimetinib. Based on our own data, and supported by observations by others, we are evaluating intermittent dosing schedules in our clinical program to allow us to maximize dose intensity in order to achieve the greatest depth of response. We also plan to explore the potential clinical benefit of RMC-4630 in combination with other in-pathway agents targeting RTK (initially EGFR), KRAS^G12C, and ERK, as well as in combination with PD-1 inhibitors. Although we are at an early stage of clinical testing and product candidate development, we believe RMC-4630 is well-positioned to become the backbone of targeted therapy combinations for the treatment of various RAS-dependent tumors.

Preclinical profile of RMC-4630

RMC-4630 potently inhibits SHP2 phosphatase activity in a biochemical assay that monitored dephosphorylation of a probe substrate (IC₅₀ 1.29 nM, range 0.9 nM to 2.2 nM, number of experiments = 13) and SHP2 function in cellular assays, as measured by inhibition of ERK1/2 phosphorylation at Thr202/Tyr204, a read out for RAS pathway activation (IC₅₀ values of 14 nM in PC9^EGFRex19del cells, range 5.3 nM to 63 nM, number of experiments = 11; and 20 nM in NCI-H358 KRAS^G12C cells, range 14.5 nM to 27.5 nM, number of experiments = 4). In standard in vitro assays of target selectivity, no significant interaction with kinases or other phosphatases was observed. RMC-4630 (up to a test concentration of 10 µM) exhibited no inhibition of full length SHP1 phosphatase (number of experiments = 3), or the catalytic domains of SHP1 and thirteen other phosphatase enzymes (single experiment conducted in duplicate). RMC-4630 exhibited over 3,000-fold (range 7054 to >19,000, single experiment conducted in duplicate) selectivity for SHP2 (as measured by biochemical potency) over a panel of over 450 kinases (as measured by displacement of probe binding).  

Preclinical anti-tumor activity

Consistent with the role of SHP2 as a regulator of the RAS cycle, we observed that RMC-4630 suppresses tumor growth in a dose-dependent manner in human cell-line or patient-derived preclinical xenograft models of tumors harboring KRAS^G12C, NF1^LOF,or BRAF^Class3 mutations or wild-type KRAS amplifications (Figure 1). Moreover, RMC-4630 at 30 mg/kg daily administered orally induced regression in some tumor models.

Figure 1: RMC-4630 suppresses tumor growth in preclinical xenograft models of tumors harboring KRAS^G12C, NF1^LOF, or BRAF^Class3 mutations or wild-type KRAS amplifications.

Daily oral administration (po qd) of RMC-4630 at 10 mg/kg (red) or 30 mg/kg (green) produces a dose-dependent inhibition of tumor growth in multiple solid human tumor cell line-derived or patient-derived xenograft (CDX or PDX, respectively) models bearing RAS pathway activating mutations of interest. All human xenograft models implanted in immune-deficient mice: (a) non-small cell lung cancer (NSCLC) PDX LUN#092 KRAS^G12C, (b) NSCLC BRAF^Class 3 PDX LUN#023 (BRAF^D594N), (c) gastric cancer PDX STO#332 wild-type KRAS amplification (KRAS Amp, copy number, CN = 4) and (d) NSCLC CDX NCI-H1838 NF1^LOF (NF1^I184fs). Control animals are shown in black. Data represent waterfall plots of individual end of study tumor responses, with tumor volume expressed as a percentage of initial tumor volume at time of study start (truncated at 400%). Each animal is represented as a separate bar (number of mice per group = 9 to 10). Numbers indicate number of regressions (defined as > 10% reduction in tumor volume

11

from starting volume) in each group. Dotted line references 50% reduction in tumor volume. The duration of treatment with RMC-4630 or vehicle control was within a range of 30 days to 50 days across the four models.

Optimizing dosing and scheduling

Using an intermittent dosing schedule, which permits deep but discontinuous inhibition of the SHP2 target, significantly higher doses of RMC-4630 were tolerated than could be delivered with daily dosing. These higher doses of RMC-4630 led to increased tumor growth inhibition and resulted in more frequent and deeper tumor regressions (Figure 2).

Figure 2: Intermittent dose regimens of RMC-4630 produce deeper and more frequent tumor regressions than daily dosing at maximal tolerated dose in a preclinical xenograft model of NSCLC tumors harboring KRAS^G12C mutations.

KRAS^G12C

Anti-tumor activity of daily (qd) and intermittent, every other day, (q2d) oral (po) dose regimens for RMC-4630 in NSCLC NCI-H358 KRAS^G12C cell line-derived xenograft model in mice. Graphs show tumor volume data for individual animals, expressed as a percentage of initial tumor volume at time of study start, for (a) 30 mg/kg qd, (b) 60 mg/kg q2d, (c) 120 mg/kg q2d and (d) 200 mg/kg q2d dose regimens (number of animals per group = 9 to 10). Changes in tumor volume of greater than 10% (grey zone) are considered significant. Dotted line references 50% reduction in tumor volume. All dose regimens were well-tolerated.

Rationale for combining with other targeted agents

Certain cancer treatments that inhibit components of the RAS signaling pathway are often unable to achieve the desired clinical effect as single agents due to the rapid development of adaptive resistance. These resistance mechanisms often involve hyperactivation of various RTKs that drive oncogenic signals via SHP2. Given that SHP2 is required for RAS signaling pathway activation by many RTKs, it might represent a viable target to limit potential resistance to other single-agent treatments. Inhibition of SHP2 in cell culture experiments abrogated RTK signaling and, in preclinical studies, RMC-4630 demonstrated combinatorial activity when given with other RAS signaling pathway inhibitors, such as MEK, KRAS^G12C or EGFR.

MEK inhibitors, such as cobimetinib, are approved for the treatment of certain types of melanoma but only in combination therapy. As single agents they have shown limited clinical effect, particularly in lung cancers carrying RAS mutations, which is believed to be due in part to adaptive resistance mechanisms.

In several preclinical tumor xenograft models either RMC-4630 or cobimetinib, dosed as single agents at doses lower than the maximally tolerated dose for each agent, inhibited tumor growth but induced few tumor regressions. However, the number and depth of tumor regressions was markedly increased upon treatment with a combination of these low-doses of RMC-4630 and cobimetinib (Figure 3).

12

Figure 3: Combination benefit for RMC-4630 and cobimetinib in preclinical xenograft models of tumors harboring KRAS^G12C mutations or wild-type KRAS amplifications.

KRAS^G12C

Wild-Type KRAS Amp

Anti-tumor activity of RMC-4630 (10 mg/kg, red) and cobimetinib (2.5 mg/kg, blue) dosed daily by oral administration (po, qd) as single agents or in combination (purple) in (a and b) NSCLC CDX NCI-H358 KRAS^G12C and (c and d) gastric cancer PDX STO#332 wild-type KRAS amplification (KRAS Amp, CN = 4) xenograft models in mice. Data represent (a and c) mean tumor volume over time or (b and d) waterfall plots of individual end of study responses with tumor volume expressed as a percentage of initial tumor volume at time of study start (truncated at 300% in d). In (a) and (c) data represent mean and errors bars represent standard error of the mean. Each animal represented as a separate bar in (b and d). Number of animals per group = 10. Respective doses (in parentheses) of RMC-4630 (10 mg/kg) and cobimetinib (2.5 mg/kg) are lower than the corresponding maximally-tolerated dose for each agent. Numbers indicate number of regressions (defined as > 10% reduction in tumor volume from starting volume) in each group.

There are three important implications of these observations: first, the anti-tumor effects of RMC-4630 may be significantly greater in human cancers that are already predicted to be sensitive to SHP2 inhibition if RMC-4630 is combined with a MEK inhibitor. Second, the effects of the combination may be observed at doses or exposures of RMC-4630 or a MEK inhibitor that are significantly below the maximum tolerated dose for each agent. Third, there may be a higher probability of invoking tumor cell death with the combination, and thus seeing tumor regressions.

In addition, two of our academic collaborators have demonstrated that tumors with KRAS^G12D or KRAS^G13D mutations may be responsive to a SHP2 inhibitor combined with a MEK inhibitor. Based on their published results, the RMC-4630 and MEK inhibitor combination may be active in some tumors with mutations that may not be sensitive to SHP2 inhibition alone.

The combination of RMC-4630 and cobimetinib has been relatively well tolerated in preclinical studies. We have also sought to maximize potent anti-tumor activity of RMC-4630, and reduce potential side effects, by deploying an intermittent dosing schedule.

An academic collaborator has presented preclinical data on the combination of a SHP2 inhibitor and an ERK inhibitor, providing rationale for the potential clinical evaluation of the combination.

13

Recently reported initial clinical results from two KRAS^G12C(OFF) inhibitors suggest significant clinical benefit and provide strong pharmacologic validation of this oncoprotein as a cancer driver. Preclinical studies have demonstrated that KRAS^G12C(OFF) inhibitors also cause a rapid increase in signaling through RTKs that are typically SHP2-dependent. Thus, the magnitude and durability of effect of an inhibitor of KRAS^G12C(OFF)may be significantly increased when combined with a SHP2 inhibitor that disrupts signaling from the activated RTKs. Recent data have demonstrated that a combination of our proprietary SHP2 inhibitor with KRAS^G12C(OFF) inhibitors can drive significant tumor regression in two distinct KRAS^G12C driven tumor models that exhibit only partial anti-tumor responses to either compound alone (Figure 4).

Figure 4: Combination benefit for SHP2 inhibitor and KRAS^G12C(OFF) inhibitor in preclinical xenograft models of tumors harboring KRAS^G12C mutations.

Anti-tumor activity of a representative SHP2 allosteric inhibitor (RMC-4550) and a KRAS^G12C(OFF) inhibitor (MRTX849 or MRTX1257) dosed daily by oral administration (po, qd) as single agents or in combination in (a) esophageal carcinoma KYSE-410 and (b) NSCLC NCI-H358 KRAS^G12C cell line-derived xenograft models in mice. Data represent (a) mean tumor volume over time or (b) waterfall plot of individual end of study responses, with tumor volume expressed as a percentage of initial tumor volume at time of study start (truncated at 300%). Data in panel (a) taken from an August 2019 Mirati corporate presentation; show evidence for tumor regressions, in addition to reduced rate of tumor regrowth after ‘dosing stopped’, for the SHP2 plus KRAS^G12C(OFF) inhibitor combination group relative to either single agent group. For our data in panel (b) each animal is represented as a separate bar (number of animals per group = 10). Numbers indicate number of regressions (> 10% reduction in tumor volume from starting volume) in each group. RMC-4550 is a potent and selective SHP2 allosteric inhibitor tool compound (see Nichols et al., 2018). MRTX849 is Mirati’s KRAS^G12C(OFF) clinical candidate and MRTX1257 is a potent and selective KRAS^G12C(OFF) inhibitor tool compound.

In approximately 25% of NSCLC in North and South America, EGFR is mutated and drives tumor growth. EGFR inhibitors are used to treat these types of lung cancer, but emergence of resistance is a clinical problem. With recently approved EGFR inhibitors such as osimertinib (marketed as Tagrisso by AstraZeneca), emergent resistance is frequently due to mutation or amplification of signaling proteins other than EGFR. Similar to the adaptive resistance pathways activated by MEK inhibitors, several of these escape drivers have been shown to signal through SHP2.

RMC-4630 enhanced the anti-tumor activity of osimertinib in preclinical models of osimertinib-sensitive and osimertinib-resistant EGFR-mutant tumors (Figures 5 and 6). RMC-4630 accelerated and increased the magnitude of tumor regression in an osimertinib-sensitive tumor and delayed and/or reduced tumor regrowth upon cessation of treatment in this model. RMC-4630 also inhibited tumor growth in a patient-derived tumor xenograft that had become resistant to osimertinib via amplification of the oncogene c-MET, an RTK that has been shown to drive some forms of cancer and that signals through SHP2. This suggests that, under circumstances where escape from osimertinib occurs via a SHP2-dependent mechanism, RMC-4630 may have clinical activity.

14

Figure 5: Combination benefit for RMC-4630 and the EGFR inhibitor, osimertinib, in an EGFR^L858R/T790M osimertinib-sensitive NSCLC xenograft model.

EGFR^L858R/T790M

Anti-tumor activity of osimertinib (5 mg/kg) dosed daily by oral administration (po, qd) as a single agent (a and b) or in combination with RMC-4630 (30 mg/kg po, qd) (c and d) in NSCLC NCI-H1975 (EGFR^L858R/T790M) cell line-derived xenograft model in mice. Graphs show tumor volume data for individual animals, expressed as a percentage of initial tumor volume at time of study start (number of animals per group =12). Horizontal dotted lines reference the starting tumor volume (0%) and a 50% reduction in tumor volume. Vertical dotted line marks time at which dosing was stopped. Panels (a and c) show the same data as in (b and d) up to the time point of dosing cessation but on an expanded time scale.

Figure 6: RMC-4630 suppresses tumor growth in an osimertinib-resistant NSCLCpatient-derived xenograft model (EGFR^L858R/T790/MET^amplified ).

Daily oral administration of RMC-4630 (30 mg/kg po, qd) inhibits tumor growth in an osimertinib-resistant NSCLC patient-derived xenograft model (EGFR^L858R/T790/MET^amplified) wherein the EGFR^T790M allele was no longer detected and the patient tumor exhibited genomic amplification of the MET receptor tyrosine kinase. The human tumor xenograft model was implanted in immune deficient mice. Data represent mean and errors bars represent standard error of the mean. Number of animals per group = 10. Osimertinib 25 mg/kg, 5 days on/2 days off had no significant impact on tumor growth as anticipated.

Rationale for combining with immune checkpoint inhibitors

Immune checkpoint inhibitors, such as inhibitors of PD-1, have been useful against a variety of tumor types, including melanomas, breast and lung cancers, certain types of colon cancer and bladder cancers. It has been proposed in the scientific literature that SHP2 interacts with the PD-1 receptor and mediates at least part of its immune suppressive signals. We have observed that SHP2 inhibition phenocopies some of the effects of PD-1 blockade in certain in vitro and in vivo models.

15

We have also seen that SHP2 inhibition inhibits the viability of pro-tumorigenic (M2) macrophages in vitro. In the tumor microenvironment in vivo, SHP2 inhibition reduced the number of M2 macrophages while also promoting increases in the anti-tumor M1 macrophages population. Therefore, RMC-4630 may increase the ability of the innate and adaptive arms of the immune system to control or even eradicate cancer cells. In models of cancer in immunocompetent mice, SHP2 inhibition activated the murine immune system to slow tumor growth, even in tumors that are not intrinsically sensitive to direct cellular effects of SHP2 inhibition. In preclinical models, the combination of a SHP2 inhibitor with an immune checkpoint inhibitor, such as a PD-1 inhibitor, occasionally induced an immune response that is sufficient for mice to ‘reject’ their tumors completely and elicit immunological memory.

Significant anti-tumor effects of SHP2 inhibition, both alone and in combination with PD-1 inhibition, were also observed in tumors intrinsically sensitive to SHP2 inhibition in vitro (Figure 7). A SHP2 inhibitor such as RMC-4630 may, therefore, elicit anti-tumor effects via two separate biologic mechanisms: targeted inhibition of RAS-dependent tumor growth, and liberation of anti-tumor immune responses by transformation of the tumor microenvironment.

Figure 7: Anti-tumor effects of SHP2 inhibition alone and in combination with PD–1 checkpoint blockade in the EMT6 syngeneic model.

EMT6 Mouse Tumor Model

RMC-4550 (30 mg/kg) was administered daily by oral administration for the duration of the study starting at day 6 post-implant; anti-PD-1 (10 mg/kg) was administered every three days by intra-peritoneal administration, for a total of 7 doses starting at day 6 post-implant, or a combination of both was administered to EMT6 tumor bearing immunocompetent mice. Control animals received the isotype control for the anti-PD-1 antibody. Data represent (a) tumor growth of individual mice for each experimental group and (b) Kaplan–Meier curves showing percentage of animals with tumor burden < 2000 mm³ in each treatment group for the duration of the study. RMC-4550 is a potent and selective SHP2 allosteric inhibitor tool compound. Number of animals per group = 10.

Development strategy

In summary, preclinical research suggests that RMC-4630 has the potential to cause significant anti-tumor effects:

16

Although we are at an early stage of clinical testing and product candidate development, we believe RMC-4630 is well-positioned to become the backbone of targeted therapy combinations for the treatment of various RAS-dependent tumors (Figure 8).

Figure 8

Phase 1/2 clinical program

In collaboration with Sanofi, we are evaluating RMC-4630 in a multi-cohort Phase 1/2 clinical program. This RMC-4630 Phase 1/2 program currently consists of two active clinical trials: RMC-4630-01, a Phase 1 study of RMC-4630 as a single agent, and RMC-4630-02, a Phase 1b/2 study of RMC-4630 in combination with the MEK inhibitor cobimetinib.RMC-4630.

RMC-4630-01 study of single agent RMC-4630 in patients with advanced solid tumors

RMC-4630-01 is aan ongoing Phase 1 study in patients with advanced cancers, including those with tumors harboring genetically defined mutations in the RAS signaling pathway, that is evaluating the safety, pharmacokinetics and pharmacodynamic effects of RMC-4630 as a single agent under two differentat the recommended phase 2 dose administration schedules: daily and twice weekly dosing. A preliminary evaluationschedule (RP2DS) of anti-tumor activity is also being made in patients who have tumors harboring mutations in200 mg administered on Day 1 and Day 2 (D1D2) weekly. Data from the RAS pathway that are predicted to be sensitive to SHP2 inhibition, including KRAS^G12C, KRAS^G12A, NF1^LOF, and BRAF^Class3 and others (e.g., KRAS^amp).

The RMC-4630-01 study was designed to evaluate two different schedules: a daily dosing schedule andsuggest that an intermittent dosing schedule (Day 1 and Day 4 of every week). The intermittent schedule was intended to achieve intermittentstrategy for RMC-4630 both exceeds target coverage which, inplasma exposure (apoptotic threshold) based on preclinical models was associated with similar or superiorof RAS pathway-driven cancers that project potential clinical activity and better tolerability. The RMC-4630-01 trial is currently being conducted at 12 clinical study sitesfacilitates low levels of target plasma exposure (cytostatic threshold) later in the United States. dosing interval to potentially allow normal tissue to recover. We are evaluating this dose and schedule in an expansion cohort of the RMC-4630-01 study, consisting of patients with gynecologic tumors harboring NF1^LOF mutations, in addition to a small safety/tolerability cohort representing a broader set of histotypes and RAS pathway genotypes. We currently expect to report a dose escalation safety data set for this trial in the first half of 2021.

As of the data cut-off on November 6, 2019,May 4, 2020, we reported the following preliminary data from RMC-4630-01:

In RMC-4630-01, 6687 patients had been enrolled and 63 had received study drug and were evaluable for safety: 14 with the38 patients received an intermittent schedule and 49 with the daily schedule (Tables 1 5 and 8). Dose escalation has been completed for2) and the daily dosing schedule. Dose escalation continues using intermittent schedules. Preliminary data suggest that an intermittent schedule will be the preferred schedule for the further development of RMC-4630. Therefore, safety, tolerability and pharmacokinetic data for patients treated with intermittent schedules are reported separately from patients treated withremainder on the daily schedule.

17

Interim safety and tolerability – intermittent dosing scheduleschedules

Fourteen31 patients have been dosed with the intermittent weekly D1D4 schedule and 7 patients have been dosed with the weekly D1D2 schedule. Across both intermittent schedules 38 patients have been evaluated for safety after a median RMC-4630 exposure of 1.61.8 months (range 0.3-5.00.2-10.6 months). Demographic and baseline characteristics information is shown in Table 1.

13

Table 1: Demographics and baseline characteristics—intermittent schedule in RMC-4630-01 study.

Data as of November 6, 2019.May 4, 2020.

The emerging safety profile of RMC-4630 when dosed on an intermittent schedule is consistent with the mechanistic effects of the product candidate on SHP2 and hence the RAS signaling cascade, including edema, gastrointestinal toxicity, reduced red cell production (low hemoglobin concentration and worsening of pre-existing anemia), reduced platelet production (thrombocytopenia), hypertension and fatigue. This safety profile was largely predictable from preclinical studies and clinical studies of other well-known inhibitors of this pathway.

Treatment-related and emergent adverse events, or related AEs, occurring in greater than or equal to 10% of patients are listed in Table 2. No relatedOne grade 4 or grade 5 AEs have been reported for this schedule. One treatment-related serious adverse event, or SAE,AE of thrombocytopenia has been reported in a patient with pancreatic cancer receiving 200240 mg twice weekly who was hospitalized withon D1D4. No related grade 3 abdominal distension (see Table 3); the SAE was unresolved at the time the patient withdrew from the study to transfer to hospice care.5 AEs have been reported.

Table 2:2: Related AEs occurring in ≥ 10% of dosed patients by grade—intermittent schedule in RMC-4630-01 study.

Data as of November 6, 2019.May 4, 2020. Abbreviations: SMQ, Standardized MedDRA Query; CMQ, Customized MedDRA Query.

Table 3: Related SAEs—intermittent schedule in RMC-4630-01 study.

Data as of November 6, 2019.

18

Pharmacokinetics—intermittent dosing schedule

The pharmacokinetic profile of RMC-4630 after dosing on the intermittent schedule is shown in Table 4 and Figure 9. No accumulation from day 1 to day 15 was observed. Plasma exposure at both dose levels was within the range anticipated to be biologically active from preclinical models. After a single dose of 140 mg, the plasma concentration of RMC-4630 remained above the in vivo EC₅₀ for pERK for 72 hours. We estimate the half-life of RMC-4630 to be approximately 25 hours.

Table 4: Pharmacokinetics – intermittent schedule in RMC-4630-01 study.

Data as of October 8, 2019.

Figure 9: RMC-4630 pharmacokinetics—intermittent schedule in RMC-4630-01 study.

Pharmacokinetic profile of RMC-4630 dosed at either 140 mg or 200 mg on D1 and D4 of each week. Steady state is considered to be day 15 of cycle 1. EC₅₀/fu and EC₇₅/fu are the total estimated plasma concentrations in humans that correspond to 50% and 75% inhibition of pERK in KRAS^G12C tumor models. Steady state in this instance is defined as at least five half lives after starting dosing at which time the pharmacokinetic profile of RMC-4630 on any given day would be considered to be representative of subsequent dosing days.

19

Interim safety and tolerability—daily dosing schedule

Forty-nineFour patients have been treated in RMC-4630-01 with the daily schedule. Median RMC-4630 exposure is 1.8 months (range 0.0-13.0 months). Demographic and baseline characteristics information is shown in Table 5.

Table 5: Demographics and baseline characteristics—daily schedule in RMC-4630-01 study.

Data as of November 6, 2019.

The ECOG (Eastern Cooperative Oncology Group) performance status is a scale often used to assess how a patient’s disease is progressing.

Although both dosing regimens have been reasonably well tolerated, daily dosing has been associated with more frequent and severe AEs than the intermittent schedule. As with the intermittent schedule, the safety profile from the daily dosing schedule has been consistent with the mechanistic effects of the product candidate on SHP2 and the RAS signaling pathways. We have not determined a maximum tolerated dose for daily dosing, although dose escalation will not continue beyond the 80 mg daily level already evaluated. If further development with this schedule was pursued, we expect the recommended Phase 2 dose for the daily schedule would be 60 mg daily.

Treatment-related AEs occurring in greater than or equal to 10% of patients who received the daily schedule are shown in Table 6. No toxicities consistent with ‘off-target’ effects have been reported. Increases in liver enzymes such as alanine transaminase and aspartate transaminase have been observed at all grades. These have been attributed, wholly or in part, to RMC-4630 in 10% or 16% of patients treated with the daily schedule, respectively. In two patients, the increase in alanine transaminase or aspartate transaminase was either grade 3 or grade 4.

Table 6: Related AEs occurring in ≥ 10% of dosed patients by grade—daily schedule in RMC-4630-01 study.

Data as of November 6, 2019.

20

Eight patients treated with the daily schedule have experienced toxicities involving the lungs or respiratory system that were attributed in part to RMC-4630 by the treating investigator. These were generally moderate or mild. Two additional cases of grade 4 respiratory failure are discussed in more detail below in the description of SAEs. No data has been reported suggesting that systemic activation of the immune system is associated with toxicity in subjects treated with RMC-4630. There have been no reports of pneumonitis. Related(11%) had serious adverse events, involving other important organs such as the heart, brain and kidneys have been uncommon and generally mild to moderate in severity.

There have been threeor SAEs, thought to be possibly or probably related to the study drug as assessed by the trial sponsor (Table 7)across all intermittent dosing cohorts. Three SAEs were reported among three patients receiving 200 mg D1D4 (Grade 3 abdominal distention, Grade 3 anemia, Grade 2 deep vein thrombosis, one patient each). Four SAEs were reported in one patient receiving 240 mg D1D4 (pleural effusion, pulmonary embolism, nausea, and vomiting, all Grade 3).

Three additional SAEs across three patients (cerebrovascular accident, multifocal pneumonia, and pleural effusion) were reported in which the investigator was unable to rule out an association with the study drug, but where the evidence for causality by RMC-4630 was absent or considered unlikely by the study sponsor. One patient with extensive metastasespresented symptoms five days after starting study drug where a total of tumor in the lungs developed grade 4 respiratory failuretwo doses were administered. This patient had major risk factors for developing a cerebrovascular accident including underlying bladder cancer, pancreatic adenocarcinoma, recent major thromboembolic event of pulmonary embolism less than two months prior to start of study treatment, elderly age, history of coronary artery calcification, smoking and was hospitalized and treated with oxygen. The SAE was ongoing when the patient was withdrawn from the study.hypertension. A second patient with fever and radiologic evidencehad pneumonia which is not an expected event for RMC-4630. After taking two doses of infectious pneumonia developed grade 4 respiratory failure andstudy drug, patient was treated with oxygen, systemic antibiotics and corticosteroids. The SAE was ongoing when the patient diedhospitalized due to progressionsuspicion of underlying cancer.post-obstructive pneumonia which was later updated to multifocal pneumonia. No pathogens were identified from the infection work-up. Given the history of lung cancer, the sponsor considers this event unlikely related to RMC-4630. A third patient developedhad NSCLC and a single reading of grade 3 prolongation of QTc. This patient had been receiving 60 mg daily of RMC-4630 but had not received any dose for three days at the time of the reading. The patient had a previous history of prolonged QTc, underlying systemic lupus,multiple pleural effusions (including prior to starting study treatment) with multiple thoracentesis performed. Given patient history and was taking ondansetron. QTc was prolonged (grade 1) at baseline. Five hoursthe rapid accumulation of pleural fluid after the prolonged QTc reading,drug was held, the patient had two follow-up ECGssponsor believes that showed normal QTc interval.this

Table 7: Related SAEs—daily schedule in RMC-4630-01 study.

Data as of November 6, 2019.

Table 8: Early data suggest intermittent schedule may be better tolerated than daily schedule in RMC-4630-01 study.

Data as of November 6, 2019.

21

Pharmacokinetics—daily dosing scheduleevent is most likely due to underlying disease progression and is unlikely related to RMC-4630. No grade 4 or 5 related SAEs have been reported in either intermittent cohort.

With daily dosing, plasma concentrations of RMC-4630 reached a steady state by day 22 (Table 9 and Figure 10). Plasma concentrations of RMC-4630 in the blood at all daily dose levels were consistently higher than the in vivo EC₅₀ for pERK in tumor models. Exposure increased approximately proportionally with increasing dose. The total exposure to RMC-4630 over a 24-hour period at 60 mg daily was 14.6 uM.hr. This is more than twice the exposure thatRMC-4630-01 study has been required to see anti-tumor effects, particularly tumor stasis, in animal models (6.44 uM.hr).

Table 9: Pharmacokinetics—daily schedule in RMC-4630-01 study.

Data as of October 8, 2019. Number of patients evaluated for parameters Cmax and AUC are shown in N(Cmax/AUC) column.

Figure 10: Pharmacokinetics—daily schedule in RMC-4630-01 study.

Data as of October 8, 2019.

Pharmacokinetic profile of RMC-4630 dosed at either 20 mg, 40 mg, 60 mg or 80 mg daily. Steady state is considered to be day 22 of cycle 1. EC₅₀/fu and EC₇₅/fu are the total estimated plasma concentrations in humans that correspond to 50% and 75% inhibition of pERK in KRAS^G12C tumor models.

Pharmacodynamic effects of RMC-4630—daily and intermittent dosing schedules

Activation of the protein ERK, which is an important protein in the RAS signaling pathway and a substrate for MEK, is a good surrogate for the inhibition of pathway activity by a SHP2 inhibitor. The pharmacodynamic effects of RMC-4630 on activation of ERK were studied in the blood cells of patients being treated with RMC-4630. Despite considerable assay variability and inter-patient variability, which is common for these types of dynamic assays in patients, there was a trend in favor of inhibition of activated ERK in peripheral blood cells at all dose levels tested. These effects are consistent with engagement and inhibition of the SHP2 target and downstream RAS signaling by RMC-4630.

22

Phosphorylation of ERK has been assessed in tumors before and during RMC-4630 administration (Figure 11). In three cases, there was a reduction in phosphorylation of cytoplasmic and nuclear ERK in the tumor while RMC-4630 was at steady state. One patient’s tumor showed no reduction in tumor pERK, but this tumor showed very little phosphorylation in the pre-treatment sample and the patient had not received any RMC-4630 for eight days prior to the second tumor biopsy.

Figure 11: pERK inhibition in tumors on RMC-4630 in RMC-4630-01 study.

Tumor pERK Staining

Quantitation of phospho-ERK (pERK) in tumor tissue taken from patients treated with daily RMC-4630 at either 20 mg, 40 mg or 60 mg daily. Panel A represents the H score for pERK before and after dosing in four patients. H score is the product of percentage of tumor cells staining positive for pERK and the intensity of staining per cell. Both nuclear and cytoplasmic pERK are shown. Panel B shows the immunohistochemistry sections from which the H score is estimated. pERK stains brown. Panel C provides information for each patient on whom paired biopsies were obtained.

Allelic burden of circulating KRAS^G12C tumor DNA, or ctDNA, has been assessed prior to the study and at least once during the study in seven patients with tumors harboring KRAS^G12C. KRAS^G12C ctDNA was detected in four of seven patients prior to the study. In two patients, with NSCLC and either partial response or stable disease as best response, there was a reduction in circulating KRAS^G12C ctDNA. In one patient with colon cancer who had a progressive disease the allelic frequency of KRAS^G12C ctDNA increased.

Interim evidence of clinical activity—daily and intermittent schedules

There isprovided preliminary evidence that RMC-4630 has single agent anti-tumor activity in KRAS mutant NSCLC. One patient, with KRAS^{G12C MUTANT}NSCLC treated at 60 mg daily, had a confirmed partial response, with a 45% reductionand in tumor volume as measured by CT imaging and is ongoing on study therapy at 16 weeks. A second, NSCLC patient with KRASNF1^G12DLOF and a mutation in the SHP2 gene (SHP2^V428M), treated with 140 mg on the intermittent schedule, had an unconfirmed partial response (unconfirmed by RECIST 1.1, recorded as best responsetumors. As of stable disease). Disease control rate, or DCR, the sum of best response of partial response and stable disease cases forMay 4, 2020, seven patients with KRAS^G12C NSCLC as of the cut-off date was 6/8 (75%).

23

As of the cut-off date, five patients with KRAS^G12C NSCLC had follow-up CT scans of target lesions andlesions. Among them, one had either partial response, orfive had stable disease, and one had disease progression (Figure 12)11); three patients had not reported follow-up measurements of target lesions, of which one has been recorded as best response of stable disease and two of progressive disease. Disease control rate, or DCR, the sum of best response of partial response and stable disease cases for patients with KRAS^G12C NSCLC as of the extract date was 7/10 (70%).

For all patients with KRAS mutant^MUTANT NSCLC, DCR was 12/18 (67%17/29 (59%) (Figure 13)12). One patient with KRAS^G12V NSCLC had beenwas on treatment for over 1416.3 months with stable disease (and approximately 15% reduction in tumor volume) as of the cut-off date. Duration of treatment, time to first and best response, duration of response and time to progression in NSCLC patients with any KRAS mutation are shown in Figure 14.

In histotypes other than NSCLC, the best response for tumors harboring a KRAS mutation, as of the cut-off date, was stable disease.

Figure 12:11. Best change in tumor burden from baseline in KRAS^G12C NSCLC in RMC-4630-01 study.non-small cell lung cancer.

Data as of November 6, 2019.May 4, 2020.

Waterfall plot of best tumor response for five patients with KRAS^G12CNSCLC non-small cell lung cancer (NSCLC) who had baseline target lesions assessed and at least one radiologic follow-up assessment of target lesion size. Percentage (Y axis) represents the percentage change from baseline in the Sum of Longest Diameters of target lesions using RECIST 1.1. Colors represent different dose levels.

Data are presented for the efficacy evaluable population (N=8)10) defined as participants with baseline and at least one post-baseline scan or who died or had clinical progression prior to first post-baseline scan. Three patients are not represented in this figure: 2 PD (1 death due to clinical progression prior to first scan and 1 did not have measurement for one of the target lesions but had new lesion) and 1 SD (pending(had partial missing tumor measurements in the database at the time of data entry in EDC)extract). PD = progressive disease; SD = stable disease; PR = partial response; PD, SD, and PR each per RECIST 1.1.

2415

Figure 1312:. Best change in tumor burden from baseline for NSCLC with anyin KRAS mutation in RMC-4630-01 study.^MUTANT non-small cell lung cancer.

Data as of November 6, 2019.May 4, 2020.

Waterfall plot of best tumor response for fourteen patients with KRAS^MUTANTmutant NSCLC, non-small cell lung cancer (NSCLC), including KRAS^G12C, who had baseline target lesions assessed and at least one radiologic follow-up assessment of target lesion size. Percentage (Y axis) represents the percentage change from baseline in the Sum of Longest Diameters of target lesions using RECIST 1.1. Colors represent different KRAS mutations. Data are presented for the efficacy evaluable population (N=18)29) defined as participants with baseline and at least one post-baseline scan or who died or had clinical progression prior to first post-baseline scan. FourFive patients are not represented in this figure: 2 patients1 patient had death due to clinical progression prior to first scan, 1 patient did not have measurements for one of the target lesions but progressed developing new lesion, and 1 patient3 patients had missing tumor measurements in the database at the time of data extract. PD = progressive disease; SD = stable disease; PR = partial response; PD, SD, and PR each per RECIST 1.1.

Figure 1413: Duration of treatment, time to and duration of response. Best change in NSCLC with any KRAS mutationtumor burden from baseline in RMC-4630-01 studyNF1^LOF cancers.

25

Data as of November 6, 2019.May 4, 2020.

SwimmerWaterfall plot of duration of treatment, time to first and best tumor response duration of response and time to progression for eighteen patients with KRAS mutant NSCLC, including KRAS^G12C. Percentage (Y axis) represents each individual patient. Colors represent different KRAS mutations. Data are presented for the efficacy evaluable population (N=18)6) defined as participants with baseline and at least one post-baseline scan or who died or had clinical progression prior to first post-baseline scan. One patient (NSCLC) with death due to clinical PD prior to first scan is not represented in this figure. NF1^LOF is loss, or

16

significant reduction, in neurofibromin protein function is presumed from nature of mutation. NSCLC = non-small cell lung cancer. PD = progressive disease; SD = stable disease; PR = partial response; PD, SD, and PR each per RECIST 1.1.

RMC-4630-02 study of RMC-4630 in combination with cobimetinib in patients with advanced solid tumors

RMC-4630-02 is a Phase 1b/2 study of RMC-4630 that includes an arm studying RMC-4630 in combination with the MEK inhibitor cobimetinib in patients with advanced cancers that harbor mutations in the RAS signaling pathway. The study is designed to evaluate the safety, tolerability, pharmacokinetic, and pharmacokineticspharmacodynamic profiles of RMC-4630 and cobimetinib under two different dose administration schedules.schedules for each agent. We currently expect to provide preliminary safety and clinical activity data for this combination in expansion cohorts in KRAS^MUTANT colorectal cancer in 2022.

We reported interim data from this ongoing study in a plenary session at the EORTC-NCI-AACR 32nd Symposium on Molecular Targets and Cancer Therapeutics (ENA 2020). These interim data as of September 21, 2020 provide preliminary evidence of anti-tumor activity in patients with colorectal cancer driven by KRAS mutations (Figure 14).

Interim results also suggest that a dual intermittent dosing strategy for RMC-4630 and cobimetinib exceeds target plasma exposures for each compound based on preclinical models of RAS pathway-driven cancers that project potential clinical activity (Figure 15). The objectiveadverse event profile of this study isthe combination, which was consistent with expected on-pathway effects of both drugs, was tolerable under the dual intermittent dosing schedule.

 Figure 14. Best change in tumor burden from baseline in KRAS^MUTANT colorectal cancer.

Data as of 9/21/2020; Bendell et al, ENA 2020; PD = progressive disease; SD = stable disease; PR = partial response; PD, SD, and PR each per RECIST 1.1.

17

Figure 15. Plasma concentrations of RMC-4630 and cobimetinib in combination at recommended Phase 2 dose and schedule.

Mean pharmacokinetic profiles of RMC-4630 (left; 140 mg) and cobimetinib (right; 40 mg) in humans at the recommended phase 2 dose and schedule. The EC₅₀ lines represent the approximate plasma concentrations required to determine ainhibit RAS pathway activity in tumors in vivo by 50% in preclinical models.

Data as of 9/21/20. Bendell et al., ENA 2020.

We have selected the recommended dose and schedule and further test clinical activity of the combination. Initially, the study assesses twice weekly RMC-4630 (D1,D4) with daily cobimetinib (21 days on, 7 off). In the second schedule, which has not begun dosing, both RMC-4630for this combination (RMC-4630 140 mg and cobimetinib 40 mg administered on a D1D2 weekly schedule), and are dosed intermittently. A preliminary evaluation of anti-tumor activity is also being made.

At the latest data cut-off on November 14, 2019, 8 patients had been enrolledevaluating this combination at this dose and all had received study medication at the first dose level and were evaluable for safety. Dose escalation to the next highest dose level has occurred and enrollment is ongoing.

Interim safety and tolerability

Eight patients have been evaluated for safety with a median RMC-4630 exposure of 1.4 months (range 0.1 – 2.5 months). Demographic and baseline characteristics information is shownschedule in Table 10.

Table 10: Demographics and baseline characteristics for RMC-4630-02 study.

Data as of November 14, 2019.

The ECOG (Eastern Cooperative Oncology Group) performance status is a scale often used to assess how a patient’s disease is progressing.

The emerging safety profile is consistent with the mechanistic effects of both SHP2 inhibition and MEK inhibition, including edema, diarrhea and other gastrointestinal toxicity, anemia and rash. This safety profile was largely predictable from single agent clinical studies of both agents.

26

Table 11: Related AEs attributed to RMC-4630 in RMC-4630-02 study.

Data as of November 14, 2019.

Table 12: Related AEs attributed to cobimetinib in RMC-4630-02 study.

Data as of November 14, 2019.

Pharmacokinetics

The pharmacokinetic profiles of RMC-4630 and cobimetinib are shown in Table 13 and Figure 14. Plasma levels of RMC-4630 were consistent with those obtained in the RMC-4630-01 study and were continuously greater than our predicted EC₅₀ for pERK inhibition in preclinical tumor models.

Table 13: Summary of pharmacokineticsexpansion cohorts in the RMC-4630-02 study.

Data asstudy in patients with colorectal cancer harboring KRAS^G12V or KRAS^G12D mutations and others drawing from a broader set of November 1, 2019.

Number of patients evaluated for parameters Cmaxhistotypes and AUC are shown in N(Cmax/AUC) column.

27

Figure 15:RMC-4630 pharmacokinetics in RMC-4630-02 study.

Data as of November 1, 2019.

Clinical activity

Only two patients have been evaluated for efficacy in this study. No efficacy data or ctDNA data were available in the electronic database as of the cut-off date.

Ongoing and planned clinical studies with RMC-4630RAS pathway genotypes.

The RMC-4630-01 Phase 1 studyEGFR inhibitor arm of RMC-4630-02 is continuing to enroll patients to determine the maximum tolerated dose and recommended Phase 2 dose for the intermittent dosing schedule. Alternative intermittent schedules may be explored as part of the RMC-4630-01 study and these data may be translated to ongoing combination studies or future studies.

The RMC-4630-02 study, evaluating intermittent dosing of RMC-4630 in combination with cobimetinib, has been activated and dose escalation is ongoing.

28

In collaboration with others, during 2020, we intend to start dosing patients in Phase 1b studies evaluating the combination of RMC-4630 with the KRAS^G12C(OFF) inhibitor AMG 510, with the EGFR inhibitor osimertinib (Tagrisso^®) in patients with EGFR positive NSCLC. The arm is designed to evaluate the safety, tolerability, and pharmacokinetics of RMC-4630 and osimertinib with different doses of RMC-4630, administered on a PD-1 inhibitorD1D2 weekly schedule, with the approved dose of osimertinib. Overlapping on-pathway toxicities with the combination of RMC-4630 and osimertinib were anticipated and have been observed. As with MEK inhibitors, many EGFR inhibitors are less selective for a mutated cancer driver than are targeted inhibitors for other drivers, such as pembrolizumab (Figure 16).

Figure 16: Phase 1/2 plannedAMG 510 (sotorasib) for KRAS^G12C, and thus the combination of EGFR antagonists with SHP2 inhibitors presents a potential clinical development programchallenge. We continue to enroll patients in the osimertinib combination study for RMC-4630.

Our RAS(ON) portfolio

Overview

evaluation of safety and tolerability. We are also developing a portfoliocurrently expect to release initial tolerability and pharmacokinetic data for the RMC-4630 and osimertinib combination in the second half of what2021. In the future, we may consider evaluating RMC-4630 in combination with other receptor tyrosine kinase (RTK) inhibitors, which we believe to be the first potent, selective and cell-active inhibitorsmay have less potential for overlapping toxicities.

Additional clinical studies with RMC-4630

RMC-4630 is also being evaluated by Amgen in combination with its investigational KRAS^G12C(OFF) agent AMG 510 (sotorasib) as an arm of mutant RAS(ON) proteins. These inhibitors have also exhibited anti-tumor activity in vivo in preclinical models.Amgen’s CodeBreaK 101 study. We believe that direct inhibitorsthe mechanism of RAS(ON)action and preclinical data that we have observed provide a viable rationale for this study. We currently expect that a recommended Phase 2 dose and schedule for further testing of this combination will suppress cell growthbe selected in the first half of 2021 and survivalthat preliminary clinical activity data will be provided in the second half of 2021.