An In-Depth Analysis of RLY-5836: A Case Study in Precision Oncology Strategy and Pipeline Prioritization

Executive Summary

This report provides a comprehensive analysis of RLY-5836, an investigational, allosteric, pan-mutant, and isoform-selective inhibitor of phosphoinositide 3-kinase alpha (PI3Kα) developed by Relay Therapeutics. The trajectory of RLY-5836, from its rational design on the company's proprietary Dynamo™ platform to the strategic deprioritization of its clinical development, offers a salient case study in modern precision oncology and disciplined portfolio management. The narrative of RLY-5836 is not one of failure, but rather a successful outcome of a deliberate "two shots on goal" strategy, where the emergence of a superior sibling compound, RLY-2608, rendered its continued development strategically untenable.

The PI3K/AKT/mTOR pathway, driven by activating mutations in the PIK3CA gene, is a validated and high-value target in oncology, particularly in hormone receptor-positive (HR+), HER2-negative breast cancer. However, the clinical utility of first-generation, orthosteric PI3Kα inhibitors, such as the FDA-approved alpelisib, has been severely constrained by a narrow therapeutic window. Their lack of selectivity for mutant versus wild-type (WT) PI3Kα leads to significant on-target, off-tumor toxicities, most notably severe hyperglycemia, which limits dosing and compromises efficacy.

Relay Therapeutics leveraged its Dynamo™ platform, which integrates computational modeling of protein dynamics with advanced experimental techniques, to design inhibitors that bind to a novel allosteric pocket, aiming to achieve mutant selectivity and overcome the limitations of prior agents. This effort produced two chemically distinct clinical candidates: RLY-2608 and RLY-5836. While RLY-5836 entered a first-in-human trial (NCT05759949) in April 2023, its development was overshadowed by the rapid and impressive progress of RLY-2608.

Clinical data from the ReDiscover trial (NCT05216432) revealed that RLY-2608 possesses a transformative clinical profile. It demonstrated not only promising antitumor activity, with a median progression-free survival (mPFS) of 11.0 months in second-line HR+/HER2- breast cancer patients, but also a remarkably favorable safety profile. The incidence of severe, dose-limiting hyperglycemia was exceptionally low (2.5-3%), decoupling the desired anti-tumor effect from the problematic metabolic toxicity that plagued its predecessors. This allowed for high dose intensity and low discontinuation rates, establishing a new and much higher benchmark for a "winnable" drug profile in this class.

In February 2024, Relay Therapeutics announced the deprioritization of RLY-5836. This decision was not precipitated by a negative safety event or lack of efficacy within the RLY-5836 program itself—for which no clinical data were ever publicly released—but was a direct consequence of RLY-2608's success. The superior clinical profile of RLY-2608, combined with its potential to overcome acquired resistance mutations, made it the unequivocal lead asset. The strategic pivot to halt RLY-5836 development reflects a disciplined allocation of capital and resources toward the candidate with the highest probability of success and best-in-class potential. This move validates Relay's discovery engine and strengthens its competitive position as it advances RLY-2608 toward pivotal studies, with the potential to redefine the standard of care for patients with PIK3CA-mutant cancers.

The PI3Kα Pathway: A High-Value but Challenging Target in Oncology

The phosphoinositide 3-kinase (PI3K) pathway is a fundamental intracellular signaling network that serves as a central regulator of a multitude of critical cellular functions. Its proper functioning is essential for normal cell physiology, governing processes such as cell growth, proliferation, survival, differentiation, and metabolism. The pathway is typically activated by upstream signals from receptor tyrosine kinases (RTKs) or G protein-coupled receptors (GPCRs). Upon activation, PI3K enzymes phosphorylate phosphatidylinositol-4,5-bisphosphate (PIP2​) to generate the second messenger phosphatidylinositol-3,4,5-trisphosphate (PIP3​). This event recruits and activates downstream effectors, most notably the serine/threonine kinase AKT (also known as protein kinase B), which in turn phosphorylates a wide array of substrates, including the mammalian target of rapamycin (mTOR). The coordinated action of the PI3K/AKT/mTOR cascade ultimately promotes anabolic processes and cell cycle progression while inhibiting apoptosis, making its tight regulation paramount for cellular homeostasis.

PIK3CA Mutations as a Key Oncogenic Driver

Dysregulation of the PI3K/AKT/mTOR pathway is one of the most common events in human cancer, leading to uncontrolled cell proliferation and survival, and contributing to resistance to both chemotherapy and radiotherapy. One of the primary mechanisms of this dysregulation is the acquisition of gain-of-function mutations in the PIK3CA gene, which encodes the p110α catalytic subunit of the Class I PI3K enzyme.

PIK3CA is one of the most frequently mutated oncogenes across all solid tumors. Its clinical significance is particularly pronounced in hormone receptor-positive (HR+), human epidermal growth factor receptor 2-negative (HER2-) breast cancer, the most common subtype of the disease, where activating PIK3CA mutations are present in approximately 35-40% of cases. These mutations are considered truncal driver events, occurring early in tumorigenesis and playing a key role in tumor progression. The mutations are not randomly distributed but are concentrated in specific "hotspots" within two key functional domains of the p110α protein: the helical domain (e.g., E542K and E545K mutations in exon 9) and the kinase domain (e.g., H1047R mutation in exon 20). These hotspot mutations constitutively activate the kinase, leading to persistent downstream signaling independent of upstream growth factor stimulation, thereby providing a sustained signal for cancer cells to grow and divide. The high prevalence and clear oncogenic role of these mutations have established PI3Kα as a highly validated and commercially attractive therapeutic target in oncology.

The Therapeutic Window Problem: On-Target, Off-Tumor Toxicity

Despite the clear rationale for targeting PI3Kα, the development of effective and well-tolerated inhibitors has been a significant challenge for the pharmaceutical industry. This difficulty is rooted in the fundamental problem of the therapeutic window, which is defined by the dose range that provides therapeutic benefit without causing unacceptable toxicity. For PI3Kα inhibitors, this window has historically been very narrow, a direct consequence of the biology of the target itself.

The first PI3Kα-selective inhibitor to gain FDA approval was alpelisib (brand name Piqray), which validated PI3Kα as a druggable target in solid tumors. It is approved for use in combination with fulvestrant for patients with HR+/HER2-, PIK3CA-mutated advanced breast cancer. However, the clinical impact of alpelisib and other similar orthosteric inhibitors—drugs that bind to the enzyme's active (ATP-binding) site—is severely limited by their lack of meaningful selectivity for the mutant form of the PI3Kα enzyme over its wild-type (WT) counterpart.

This lack of mutant selectivity is the root cause of the class's most significant clinical liability. The WT PI3Kα enzyme is not only present in tumor cells but is also ubiquitously expressed in healthy tissues throughout the body, where it plays a crucial role in normal metabolic processes, particularly insulin signaling and glucose homeostasis. When a non-selective inhibitor like alpelisib is administered, it blocks the activity of WT PI3Kα in tissues such as the liver, skeletal muscle, and adipose tissue. This inhibition disrupts the normal insulin signaling pathway, leading to decreased glucose uptake and utilization, which manifests clinically as hyperglycemia (high blood sugar).

This is not a minor or easily managed side effect. The hyperglycemia induced by alpelisib is frequent and often severe. In the pivotal SOLAR-1 clinical trial, 65% of patients experienced hyperglycemia of any grade, with 37% experiencing severe (Grade 3 or higher) events. This on-target, off-tumor toxicity necessitates intensive patient management, including frequent blood glucose monitoring, strict dietary modifications, and often the initiation or intensification of anti-hyperglycemic medications like metformin. More importantly, this toxicity is dose-limiting. The need to manage hyperglycemia frequently forces clinicians to reduce the dose of the PI3Kα inhibitor or to discontinue treatment altogether. In the SOLAR-1 trial, for instance, 21% of patients permanently discontinued alpelisib due to adverse events. This creates a detrimental clinical paradox: the toxicity prevents the administration of a sufficiently high and sustained dose needed to achieve optimal inhibition of the mutant PI3Kα enzyme within the tumor. Consequently, the drug's anti-cancer efficacy is compromised. This fundamental challenge—the narrow therapeutic index dictated by WT-mediated toxicity—defined the major unmet medical need in the field and set the stage for the development of a new generation of truly mutant-selective inhibitors.

Relay Therapeutics and the Dynamo™ Platform: A New Paradigm for Drug Design

Relay Therapeutics was founded with the mission to transform the drug discovery process by focusing on the dynamic nature of proteins, aiming to develop precision medicines that can solve therapeutic problems previously considered intractable or that were inadequately addressed by traditional methods. The company's strategy centers on creating highly selective small molecules that can overcome the common pitfalls of drug development, such as off-target toxicity and acquired resistance, which have long hindered progress in fields like oncology.

The Dynamo™ Platform

At the heart of Relay's research and development strategy is its unique and proprietary Dynamo™ platform. This platform represents a paradigm shift from traditional structure-based drug design, which largely relies on static, snapshot images of proteins, to a "Motion-Based Drug Design®" approach. The Dynamo™ platform integrates a sophisticated suite of leading-edge computational and experimental technologies to understand and exploit the full range of a protein's movements and conformational changes.

The key components of the platform include:

• Advanced Computational Methods: Relay utilizes cutting-edge supercomputing to run long time-scale molecular dynamics simulations. These simulations generate virtual "movies" of full-length proteins, revealing how they move, flex, and change shape over biologically relevant timescales. This is coupled with artificial intelligence and machine learning (AI/ML) algorithms to analyze these complex datasets and identify critical patterns.
• Advanced Experimental Techniques: The computational insights are continuously informed and validated by a range of high-resolution experimental methods. These include cryogenic electron microscopy (cryo-EM) and ambient temperature X-ray crystallography to visualize protein structures in different states, as well as a proprietary, machine learning-powered DNA-encoded library (DEL) platform for identifying chemical starting points.

The central innovation of the Dynamo™ platform is its ability to identify and target transient or "cryptic" allosteric binding pockets. Allosteric sites are locations on a protein distinct from the primary active (or orthosteric) site. These pockets are often not visible in static crystal structures but are revealed only when the protein is in motion. By targeting these unique, often conformation-specific sites, Relay aims to design drugs with much higher selectivity and novel mechanisms of action compared to conventional orthosteric inhibitors. This integrated, iterative process between computation and experimentation is designed to reduce drug discovery cycle times and increase the probability of clinical success.

Application to PI3Kα and the "Two Shots on Goal" Strategy

Relay Therapeutics identified the PI3Kα pathway as an ideal target for its Dynamo™ platform, given the clear unmet need for a mutant-selective inhibitor that could overcome the toxicity limitations of existing drugs. The company applied its full suite of capabilities to this challenge. Scientists at Relay successfully synthesized and solved the full-length cryo-EM structure of the PI3Kα enzyme and then deployed long time-scale molecular dynamics simulations to elucidate the subtle but critical differences in motion between the WT and mutant forms of the protein. These simulations revealed a novel, previously undiscovered allosteric pocket that could be leveraged to design a drug that would preferentially bind to and inhibit the mutant enzyme while sparing the WT version.

This discovery process led to the generation of not one, but two distinct clinical candidates: RLY-2608 and RLY-5836. Both were designed as orally available, allosteric, pan-mutant, and isoform-selective PI3Kα inhibitors. Critically, the two molecules were described as being "chemically distinct," meaning they were based on different molecular scaffolds but designed to target the same pathway through a similar allosteric mechanism.

Advancing two separate candidates against the same target is a well-known but capital-intensive drug development strategy often referred to as "two shots on goal." The rationale behind this approach is to mitigate the inherently high risk and attrition rates of clinical development. Different chemical structures, even when designed against the same target, can exhibit profoundly different pharmacological properties in humans, including variations in potency, selectivity, pharmacokinetics (absorption, distribution, metabolism, and excretion), and unforeseen safety issues. By advancing two distinct molecules in parallel, Relay Therapeutics significantly increased the statistical probability that at least one of them would emerge from early clinical testing with a desirable, validated, and ultimately "winnable" clinical profile.

However, this strategy carries an intrinsic and unavoidable consequence: it creates an internal competition, or "bake-off," between the two assets. From a strategic and financial perspective, it was never probable that the company would advance both candidates through late-stage development and commercialization for the same indication. The logical endpoint of a "two shots on goal" strategy is the selection of a single winner. Therefore, the moment one candidate began to generate clinical data that was clearly superior to the other, the strategic imperative and financial justification for continuing the parallel development of the "backup" candidate would be severely weakened. The deprioritization of RLY-5836 should be viewed through this lens—not as an isolated failure, but as the planned and rational conclusion of a competitive development strategy that successfully identified a lead asset.

Profile of an Investigational Agent: RLY-5836

RLY-5836 was the second of two investigational agents that Relay Therapeutics advanced into the clinic as part of its broad effort to develop a best-in-class, mutant-selective PI3Kα inhibitor. As a product of the Dynamo™ platform, it was designed with the same core principles as its sibling compound, RLY-2608, but possessed a distinct chemical structure.

Table 1: RLY-5836 Drug Profile
Name	RLY-5836
Synonyms	RLY 5836, RLY5836, pan-mutant-selective PI3K-alpha inhibitor RLY-5836, mutant-selective PI3Ka inhibitor RLY-5836
NCIT ID	C199577
Developer	Relay Therapeutics, Inc.
Originator	While developed by Relay, some early discovery work on pyridopyrimidinones as PI3Kα inhibitors was conducted at Mirati Therapeutics.
Drug Class	Allosteric, pan-mutant, and isoform-selective PI3Kα inhibitor
Formulation	Oral
Development Status	Deprioritized (February 2024)

Detailed Mechanism of Action

RLY-5836 is an orally bioavailable small molecule designed to function as a highly selective inhibitor of the alpha isoform of the Class I PI3K enzyme. Its mechanism is distinct from first-generation inhibitors like alpelisib. Instead of competing with ATP at the enzyme's active (orthosteric) site, RLY-5836 was engineered to bind to an allosteric site—a secondary, regulatory pocket on the protein.

The key therapeutic hypothesis underpinning RLY-5836 was its dual selectivity:

1. Isoform Selectivity: It was designed to inhibit PI3Kα while sparing the other Class I PI3K isoforms (β, δ, γ), thereby avoiding off-isoform toxicities.
2. Pan-Mutant Selectivity: Crucially, it was designed to selectively bind to and inhibit the various mutated forms of PI3Kα (including the common helical and kinase domain mutants) while having minimal activity against the wild-type (WT) enzyme.

By specifically targeting the aberrant, cancer-driving mutant protein, RLY-5836 was intended to prevent the downstream activation of the PI3K/AKT/mTOR signaling pathway in tumor cells, leading to the inhibition of cell growth and the induction of apoptosis. The intended clinical benefit of this highly selective mechanism was to achieve potent anti-tumor efficacy while simultaneously avoiding the dose-limiting toxicities, such as hyperglycemia, rash, and diarrhea, that are associated with the inhibition of WT PI3Kα in healthy tissues. This would theoretically create a much wider therapeutic window, allowing for more effective and durable treatment.

Preclinical Differentiation and a Key Omission

In its communications, Relay Therapeutics consistently described RLY-5836 as being "molecularly distinct with differentiated pharmaceutical properties" compared to its lead compound, RLY-2608. This highlights that the two drug discovery programs yielded different chemical matter, a key goal of the "two shots on goal" strategy. Early discovery work on a class of compounds known as pyridopyrimidinones, which selectively inhibit the H1047R PI3Kα mutant, was reported by researchers at Mirati Therapeutics and may represent the chemical class from which RLY-5836 was derived or related.

However, a critical element in the analysis of RLY-5836 is the striking asymmetry in the public disclosure of preclinical data for Relay's two candidates. The scientific literature and company presentations are replete with detailed preclinical evidence showcasing the profile of RLY-2608. These data demonstrate its potent and selective inhibition of mutant PI3Kα, its superior biochemical selectivity over WT and other isoforms, and, most importantly, its minimal impact on insulin and glucose homeostasis in preclinical models compared to orthosteric inhibitors like alpelisib and inavolisib. This body of evidence established a strong preclinical rationale for RLY-2608's differentiated profile.

In stark contrast, a similar portfolio of public-facing preclinical data for RLY-5836 is absent. In the highly competitive and transparent context of biopharmaceutical development, where positive data is a key asset for building investor confidence and scientific validation, this absence is conspicuous. It strongly suggests that in the company's internal, head-to-head preclinical assessments, RLY-2608 demonstrated a more favorable profile. This could have been due to superior potency, selectivity, pharmacokinetic properties, or a cleaner safety profile. This likely established RLY-2608 as the de facto lead candidate well before clinical development began for either molecule, positioning the RLY-5836 clinical program as a valuable but secondary backup, contingent on the performance of the lead asset.

Clinical Evaluation: The First-in-Human Study of RLY-5836 (NCT05759949)

The clinical development program for RLY-5836 was initiated to evaluate its safety, tolerability, and preliminary activity in humans. The program was centered on a single, multi-part, first-in-human (FIH) clinical trial, registered under the identifier NCT05759949.

Trial Design and Logistics

The study was designed as a Phase 1, open-label, non-randomized, multicenter trial intended to enroll patients in the United States. Key academic institutions, including the Dana-Farber Cancer Institute, served as clinical sites, with Dr. Andreas Varkaris of Massachusetts General Hospital listed as a Principal Investigator. Relay Therapeutics officially announced the initiation of the trial and the enrollment of the first patient in April 2023. The trial's design consisted of two main parts: a dose-escalation phase (Part 1) to determine the appropriate dose, followed by a dose-expansion phase (Part 2) to further evaluate the selected dose in specific patient populations. A Bayesian Optimal Interval (BOI) design was specified for the dose escalation portion, a sophisticated statistical method used to efficiently identify the optimal dose level.

Objectives and Endpoints

As is typical for a FIH study, the primary objectives were centered on safety and establishing a dose for future studies.

• Primary Endpoints:
• To determine the Maximum Tolerated Dose (MTD) and Recommended Phase 2 Dose (RP2D) of RLY-5836, both as a single agent and in combination with other therapies. The MTD was defined as the dose level with a dose-limiting toxicity (DLT) rate closest to a target of 30% during the first cycle of treatment.
• To evaluate the overall safety and tolerability profile of RLY-5836.
• Secondary Endpoints:
• To characterize the pharmacokinetics (PK) of RLY-5836 (how the body absorbs, distributes, metabolizes, and excretes the drug).
• To assess pharmacodynamics (PD) by measuring changes in biomarkers of the PI3K pathway, such as circulating markers of glucose metabolism.
• To obtain a preliminary signal of anti-tumor activity, measured by Objective Response Rate (ORR) and Duration of Response (DOR) according to the Response Evaluation Criteria in Solid Tumors (RECIST v1.1).

The trial was designed with multiple arms to explore RLY-5836 both as a monotherapy and in combination with established cancer therapies, reflecting a comprehensive development strategy.

Table 2: Design of the NCT05759949 Clinical Trial
Arm Number	Intervention(s)	Patient Population	Key Inclusion/Exclusion Criteria
Arm 1	RLY-5836 Monotherapy	Patients with unresectable or metastatic solid tumors	Documented oncogenic PIK3CA mutation; disease refractory to, intolerant of, or patient declined standard therapy; measurable disease per RECIST 1.1
Arm 2	RLY-5836 + Fulvestrant	Patients with HR+, HER2- locally advanced or metastatic breast cancer	Documented oncogenic PIK3CA mutation; prior treatment with ≥1 CDK4/6 inhibitor and ≥1 anti-estrogen therapy; evaluable disease
Arm 3	RLY-5836 + Fulvestrant + Palbociclib	Patients with HR+, HER2- locally advanced or metastatic breast cancer	Documented oncogenic PIK3CA mutation; prior treatment with ≥1 CDK4/6 inhibitor and ≥1 anti-estrogen therapy; evaluable disease
Arm 4	RLY-5836 + Fulvestrant + Ribociclib	Patients with HR+, HER2- locally advanced or metastatic breast cancer	Documented oncogenic PIK3CA mutation; prior treatment with ≥1 CDK4/6 inhibitor and ≥1 anti-estrogen therapy; evaluable disease
Arm 5	RLY-5836 + Fulvestrant + Abemaciclib	Patients with HR+, HER2- locally advanced or metastatic breast cancer	Documented oncogenic PIK3CA mutation; prior treatment with ≥1 CDK4/6 inhibitor and ≥1 anti-estrogen therapy; evaluable disease
All Arms			Age ≥18 years; ECOG Performance Status 0-1. Prior treatment with a PI3Kα inhibitor was an exclusion criterion for the dose expansion parts of all arms.

Public Data and the Information Vacuum

Despite the detailed and ambitious design of the clinical trial, the public dissemination of data from the RLY-5836 program has been exceptionally limited. The only formal scientific disclosure was the abstract for a poster presentation at the San Antonio Breast Cancer Symposium (SABCS) in December 2023.

This abstract is notable primarily for what it does not contain: any clinical results. It meticulously outlines the background, rationale, and methodology of the NCT05759949 study, as detailed above, but provides no information on the number of patients treated, safety observations, pharmacokinetic data, or any signals of anti-tumor activity. The program was officially deprioritized by Relay Therapeutics just two months later, in its Q4 2023 earnings report in February 2024.

The complete absence of any publicly shared clinical data for RLY-5836 is highly telling. It indicates that the decision to halt the program was made at a very early stage of development. Furthermore, it strongly reinforces the conclusion that the pivot was not driven by an adverse safety signal or a clear lack of efficacy observed within the RLY-5836 trial itself. Had there been a significant negative event, it is likely that some form of disclosure would have been necessary. Instead, the information vacuum surrounding RLY-5836's clinical performance points to an external, strategic driver for its discontinuation: the concurrent and overwhelming success of the RLY-2608 program.

The Decisive Factor: Comparative Analysis of the PI3Kα Inhibitor Landscape

The decision to deprioritize RLY-5836 cannot be understood in isolation. It was a direct result of a rigorous, data-driven assessment of the competitive landscape, a landscape dominated by its own sibling compound, RLY-2608. As clinical data from the RLY-2608 program emerged, it became clear that it possessed a profile that was not just promising, but potentially transformative for the class, setting a new and significantly higher bar for any competitor, including RLY-5836.

Table 3: Comparative Profile of Key PI3Kα Inhibitors in HR+/HER2- Breast Cancer (Post-CDK4/6i Setting)
Drug	Mechanism	Selectivity	mPFS (months)	Key Grade ≥3 Toxicities	Development Status
Alpelisib (Piqray)	Orthosteric	Non-selective	~5.5	Hyperglycemia (37%), Rash (10%), Diarrhea (7%)	Approved
Inavolisib (Itovebi)	Orthosteric	α-selective	17.2 (1L setting)	Hyperglycemia (12%), Stomatitis (6%), Diarrhea (not specified)	Approved
RLY-2608	Allosteric	Pan-mutant & Isoform-selective	11.0 (2L setting)	Hyperglycemia (2.5-3%), Diarrhea (1.7%), Rash (rare), Stomatitis (rare)	Phase 3 planned
RLY-5836	Allosteric	Pan-mutant & Isoform-selective	N/A	N/A	Deprioritized

Analysis of Competitors

• Alpelisib (Piqray): As the first-generation incumbent, alpelisib established the clinical proof-of-concept for PI3Kα inhibition. However, its efficacy in the heavily pre-treated post-CDK4/6 inhibitor setting is modest, with studies showing a median Progression-Free Survival (mPFS) of approximately 5.5 months. This limited benefit is further eroded by its severe toxicity profile. The high rates of Grade 3+ hyperglycemia (37%), rash (10%), and diarrhea (7%) make it a challenging drug for patients and clinicians to manage, leading to frequent dose interruptions and a high discontinuation rate that ultimately undermines its effectiveness.
• Inavolisib (Itovebi): Developed by Roche/Genentech, inavolisib represents the next evolution of orthosteric inhibitors. It gained FDA approval based on the impressive results of the INAVO120 trial, which evaluated it as a first-line therapy in combination with palbociclib and fulvestrant. In that setting, it demonstrated a robust mPFS of 17.2 months, a significant improvement over the control arm. While inavolisib raised the efficacy bar, it did not solve the fundamental toxicity problem of the class. The INAVO120 trial still reported clinically significant rates of Grade 3+ adverse events, including hyperglycemia (12%), stomatitis (6%), and diarrhea. This demonstrated that even with improved α-selectivity, the inhibition of WT PI3Kα via an orthosteric mechanism remains intrinsically linked to metabolic and other on-target toxicities.

Analysis of the Lead Candidate: RLY-2608 (ReDiscover Trial)

The clinical data from the ReDiscover trial (NCT05216432) for RLY-2608, particularly in combination with fulvestrant for patients with HR+/HER2- breast cancer who had progressed on prior CDK4/6 inhibitor therapy, proved to be the decisive factor.

• Efficacy: RLY-2608 demonstrated highly encouraging and clinically meaningful anti-tumor activity. In the target population (post-CDK4/6i, without PTEN/AKT co-alterations), the combination yielded an mPFS of 11.0 months for second-line patients. This result is highly competitive, especially considering the heavily pre-treated nature of the patient population. The activity was particularly pronounced in patients with kinase domain mutations, where the mPFS reached an impressive 18.4 months. The Objective Response Rate (ORR) was also strong, ranging from 39% overall to 67% in the kinase-mutant subgroup, with 81% of patients showing some degree of tumor reduction.
• Safety—The Key Differentiator: The most transformative aspect of the RLY-2608 data was its safety and tolerability profile. By successfully targeting an allosteric site and achieving high selectivity for mutant PI3Kα, RLY-2608 effectively decoupled the desired anti-tumor activity from the problematic WT-mediated toxicities.
• Hyperglycemia: The rate of Grade 3 hyperglycemia was exceptionally low, reported at just 2.5% to 3% across studies. This stands in stark contrast to the 37% seen with alpelisib and 12% with inavolisib. The majority of hyperglycemia events were low-grade (Grade 1) and required no medical intervention.
• Other Key Toxicities: Other class-specific adverse events like severe rash and stomatitis were reported as absent or rare. Grade 3 diarrhea was minimal at 1.7%.
• Dose Intensity and Discontinuation: This favorable safety profile translated directly into excellent treatment tolerability. Patients were able to stay on the drug at the intended dose for longer periods. The median dose intensity was reported to be a very high 92-98%, and discontinuations due to treatment-related adverse events were extremely rare.

The clinical profile demonstrated by RLY-2608 fundamentally redefined what a "winnable" profile for a PI3Kα inhibitor could be. The biopharmaceutical industry is in a constant state of re-evaluating the benchmarks for success. Alpelisib established a low bar for tolerability, where significant toxicity was accepted for modest efficacy. Inavolisib raised the efficacy bar but confirmed that considerable toxicity was still the price of admission for potent orthosteric inhibition. The data from RLY-2608, however, presented a new possibility: a drug that could deliver strong, competitive efficacy without the severe toxicity trade-off. It offered the potential for a significantly wider therapeutic index.

Once this new, clearly superior profile was established by RLY-2608, the development path for RLY-5836 became strategically untenable. To justify its continued development and the immense investment required, RLY-5836 would have needed to demonstrate a clinical profile that was not just good, but demonstrably better than RLY-2608's. Given that both were products of the same discovery platform, this was an exceedingly unlikely outcome. The success of RLY-2608 effectively raised the bar for its sibling compound to an insurmountable height, making its deprioritization the only logical and disciplined decision.

The Strategic Pivot: Unpacking the Deprioritization of RLY-5836

The culmination of the internal competition between Relay's two PI3Kα candidates was the formal announcement to halt the development of RLY-5836. This decision, while seemingly abrupt from an external perspective, was the logical outcome of a well-defined and data-driven portfolio management process.

The Announcement and its Timing

Relay Therapeutics disclosed its decision to "deprioritize further clinical development" of RLY-5836 in its fourth-quarter and full-year 2023 financial results press release, issued on February 22, 2024. This announcement came less than a year after the initiation of the FIH trial in April 2023 and only two months after the presentation of the methods-only poster at SABCS in December 2023. The rapid succession of these events underscores the swiftness of the company's decision-making process once the superiority of the RLY-2608 clinical data became clear.

Table 4: Timeline of Relay Therapeutics' PI3Kα Program (2022-2024)
Date / Quarter	Key Event	Source
Dec 2021	RLY-2608 FIH monotherapy arm initiated.
Apr 2022	RLY-2608 FIH combination arm with fulvestrant initiated.
Jun 2022	Relay discloses RLY-5836 as a new, chemically distinct PI3Kα inhibitor program.
Feb 2023	Relay announces RLY-5836 is expected to enter the clinic in Q2 2023.
Apr 2023	Relay presents initial clinical data for RLY-2608 at AACR, showing a favorable safety profile and selective target engagement.
Apr 2023	Relay initiates the first-in-human clinical trial for RLY-5836 (NCT05759949).
Jul 2023	Relay initiates the first dose expansion cohort for RLY-2608 + fulvestrant at 600mg BID.
Dec 2023	A poster describing the methodology of the RLY-5836 trial is presented at SABCS. No clinical data are shared.
Feb 2024	Relay announces the deprioritization of further clinical development for RLY-5836 in its Q4 2023 earnings report.

Synthesizing the Rationale for Deprioritization

The decision to halt the RLY-5836 program was not based on a single data point but rather on a cascade of strategic imperatives that became undeniable once the clinical profile of RLY-2608 was established.

1. Unequivocal Clinical Superiority: The primary driver was the quality of the data emerging from the RLY-2608 program. As detailed previously, RLY-2608 demonstrated a best-in-class potential by combining robust efficacy with a remarkably clean safety profile, particularly the near-absence of severe hyperglycemia. This profile represented a clear and significant differentiation from all existing and known investigational PI3Kα inhibitors. With RLY-2608 established as the clear lead candidate, continuing to invest in a second, likely inferior, asset made little scientific or commercial sense.
2. Financial Prudence and Capital Allocation: Clinical development is the most capital-intensive phase of drug development. Running parallel Phase 1 trials, especially complex multi-arm studies like those for RLY-2608 and RLY-5836, incurs substantial costs. Relay's financial statements consistently show a high cash burn rate, driven primarily by research and development expenses related to its clinical trials. By halting the RLY-5836 program, the company could immediately conserve capital and reallocate those financial resources to accelerate the development of its most promising asset, RLY-2608. This is a hallmark of disciplined financial management in the biotechnology sector.
3. Operational Focus and Resource Concentration: Beyond capital, clinical development demands immense operational resources, including the time and attention of key personnel, clinical trial site management, patient recruitment efforts, and chemistry, manufacturing, and controls (CMC) activities. By consolidating its PI3Kα efforts onto a single program, Relay could focus its entire operational apparatus on executing the RLY-2608 development plan with maximum speed and efficiency. This focus was critical for enabling the next strategic steps, such as initiating complex triplet combination studies (e.g., with CDK4/6 inhibitors) and planning for a large, resource-intensive pivotal Phase 3 trial.
4. Signaling Confidence to the Market: In the competitive biopharmaceutical landscape, strategic clarity is a valuable asset. By making a swift and decisive move to deprioritize RLY-5836, Relay Therapeutics sent a powerful and unambiguous message to investors, analysts, and the broader scientific community. The decision signaled immense confidence in the clinical profile of RLY-2608 and the company's conviction that it held a potential best-in-class medicine. This act of strategic discipline likely bolstered investor confidence in the company's ability to make tough, rational decisions to maximize the value of its pipeline.

The Future of PI3Kα Inhibition: Overcoming Acquired Resistance

While achieving a favorable initial safety and efficacy profile is a critical first step, the long-term success of any targeted cancer therapy is ultimately determined by its durability and its ability to contend with the inevitable emergence of drug resistance. The next frontier for PI3Kα inhibitors lies in understanding and overcoming these resistance mechanisms. Here too, the allosteric approach embodied by RLY-2608 offers a significant potential advantage over its orthosteric predecessors.

The Next Frontier: Durability and Resistance

Cancers are dynamic and evolve under the selective pressure of therapy. One of the key ways they develop resistance is through the acquisition of new mutations in the drug's target. Seminal research from investigators at the Mass General Cancer Center and their collaborators has elucidated the landscape of acquired resistance to orthosteric PI3Kα inhibitors like alpelisib and inavolisib.

Through analysis of serial liquid biopsies and tumor samples from patients whose breast cancer had progressed on these drugs, the researchers identified that a major mechanism of resistance was the emergence of new, secondary mutations within the PIK3CA gene itself. These on-target mutations were found to occur at key amino acid residues within the catalytic pocket where the orthosteric drugs bind, specifically at residues such as Trp780 and Gln859. These mutations effectively "change the lock," altering the shape of the binding site so that the orthosteric inhibitor "key" no longer fits properly. This reduces the drug's binding affinity and allows the PI3K pathway to be reactivated, driving tumor progression despite continued treatment. Other resistance mechanisms identified include the loss of the tumor suppressor PTEN and the acquisition of activating mutations in the downstream effector AKT1.

The Allosteric Advantage

The novel allosteric mechanism of RLY-2608 is uniquely suited to circumvent this primary on-target resistance mechanism. Because RLY-2608 binds to a completely different site on the PI3Kα protein—a cryptic allosteric pocket far from the orthosteric ATP-binding site—mutations that arise in the catalytic pocket to block orthosteric inhibitors do not affect RLY-2608's ability to bind and exert its inhibitory effect.

This was demonstrated experimentally in the same study that identified the resistance mutations. The researchers showed that while mutations at W780 and Q859 conferred resistance to alpelisib and inavolisib, they had no effect on the potency of RLY-2608. This provides RLY-2608 with a second, potentially profound, clinical advantage over its orthosteric competitors. Beyond its superior initial tolerability, RLY-2608 may also offer a more durable benefit by preempting and overcoming a major class of acquired resistance mutations. This combination of better safety and potentially longer duration of response is the quintessential profile of a next-generation therapeutic.

The Path Forward: Rational Combinations

The understanding of resistance also illuminates the path forward for the field: rational combination therapies designed to block multiple escape pathways simultaneously. While RLY-2608 can overcome on-target resistance, tumors can still escape through off-target mechanisms, such as activating AKT1 mutations. This highlights the need to combine PI3Kα inhibition with agents that block other critical pathways.

Relay's clinical development strategy for RLY-2608 reflects this forward-looking approach. The ReDiscover trial is not only evaluating RLY-2608 with the endocrine therapy fulvestrant but is also actively exploring triplet combinations with CDK4/6 inhibitors like ribociclib (Kisqali) and a selective CDK4 inhibitor, atirmociclib, from Pfizer. By simultaneously targeting the PI3K pathway, the estrogen receptor pathway (with fulvestrant), and the cell cycle pathway (with CDK4/6 inhibitors), Relay aims to create a more comprehensive blockade of tumor growth and survival signals, potentially delaying resistance and further improving outcomes for patients with HR+/HER2- breast cancer.

Conclusion: Lessons from RLY-5836

The story of RLY-5836 is a compelling illustration of disciplined, data-driven strategy in modern biopharmaceutical development. Its journey from a promising, rationally designed molecule to a deprioritized asset should not be interpreted as a failure of the compound or the science behind it. Instead, it represents a successful scientific and strategic exercise—a testament to the power of a discovery platform capable of generating multiple high-quality candidates and the corporate discipline required to select the single best one to advance.

In today's highly competitive oncology landscape, the goal is no longer simply to develop a drug that "works." The objective is to create a therapeutic with a decisively superior clinical profile that can displace existing standards of care and provide a clear, meaningful benefit to patients. The pivotal moment in the RLY-5836 narrative was when clinical data from its sibling compound, RLY-2608, established a new and dramatically higher benchmark for what a PI3Kα inhibitor could achieve. RLY-2608 demonstrated the potential to deliver robust efficacy while largely eliminating the severe, dose-limiting toxicities that had plagued the drug class for a decade. It proved that potent anti-tumor activity and good tolerability were not mutually exclusive.

Once RLY-2608 achieved this high bar, the continuation of the RLY-5836 program became strategically indefensible. The decision to deprioritize RLY-5836 was a moment of strategic clarity for Relay Therapeutics, reflecting a prudent allocation of capital and a sharp focus of operational resources on the asset with the greatest potential to transform patient care. This decisive action validates the strength of the Dynamo™ platform and solidifies Relay's position with a highly promising lead candidate, RLY-2608, which is now poised to advance into pivotal trials and potentially redefine the standard of care for a large population of patients with PIK3CA-mutant cancers.