Initial Phase I Safety Study of Gedatolisib plus Cofetuzumab Pelidotin for Patients with Metastatic Triple-Negative Breast Cancer

Triple-negative breast cancer (TNBC) is a heterogeneous disease that comprises a minority (15%–20%) of breast cancer cases, but has a disproptionally higher mortality (1). TNBC is characterized by the absence of estrogen-receptor (ER), progesterone-receptor, and HER2 overexpression (2–4). Patients with TNBC have a higher probability of relapse in the first 3 years after surgery, higher rates of visceral metastases, and shortened overall survival (OS) upon onset of metastatic disease when compared with hormone receptor–positive and HER2-positive breast cancers (5). Despite the recent FDA approval of immune checkpoint inhibition using pembrolizumab for early-stage (6) and late-stage disease (7), cytotoxic chemotherapy remains the mainstay treatment for metastatic TNBC. Previous attempts to introduce targeted therapies in TNBC by inhibiting the EGFR and c-KIT receptors were unsuccessful (8–11). PARP inhibitors are effective, but only in the subset of patients with deleterious BRCA1/2 mutations (12, 13). Sacituzumab govitecan, an antibody–drug conjugate (ADC) targeting TROP2 was recently approved by the FDA for metastatic TNBC, based on an overall response rate (ORR) of 33.3% in a single-arm phase II study unselected for TROP2 expression (14). These results were recently confirmed in the phase III ASCENT study demonstrating an ORR of 35% with sacituzumab compared with 5% with single-agent chemotherapy in patients treated with two prior lines of therapy (15). Clearly, there remains a critical need to identify additional novel targeted therapies for TNBC.

The majority of TNBCs (∼70%) harbor a genomic aberration in the PI3K pathway (16), most commonly an activating mutation or somatic copy-number change in one of the canonical components (17). Somewhat surprisingly, single-agent inhibition of PI3K has been ineffective as a treatment for TNBC (18). We have previously reported that PI3K inhibition in TNBC results in an immediate compensatory upregulation of the Wnt pathway (19). Activation of the Wnt pathway is well known for its role in cancer metastases and confers resistance to PI3K inhibition in TNBC, pancreatic, and colorectal cell lines (20–22). Simultaneous dual targeting of both pathways is synergistic against TNBC models in vitro and in vivo (19).

Building on our preclinical observations, we initiated a phase I clinical trial of gedatolisib (23, 24), a pan-class I isoform PI3K/mTOR inhibitor, and cofetuzumab pelidotin (25, 26), an ADC against the cell-surface PTK7 protein with an auristatin payload. PTK7 is a pseudoreceptor tyrosine kinase involved in noncanonical Wnt signaling that maintains developmental planar cell polarity in normal cells (27). PTK7 is overexpressed in a variety of cancers including TNBC (25), and its expression is induced by PI3K inhibition therefore providing rationale for exploring the combination of a PI3K inhibitor with a PTK7-targeting ADC in TNBC. The payload for cofetuzumab pelidotin, auristatin, is in itself synergistic with gedatolisib, providing another potential mechanism of synergy for the combination (28).

Participants (pt) had metastatic TNBC or ER-low (ER and PgR <5%, HER2-negative) breast cancer, and had received at least one prior chemotherapy for metastatic disease. Pts were excluded if previously treated with a PI3K or an mTOR inhibitor or if they had previous exposure to cofetuzumab pelidotin. Pts were required to have an Eastern Cooperative Oncology Group performance status ≤1 with adequate hematologic, hepatic, and renal function. Pts with treated and stable central nervous system involvement were allowed. Given the potential for hypergylcemia with gedatolisib, pts with uncontrolled diabetes were excluded.

This single-center, phase I dose-escalation trial utilized a traditional 3+3 schema with a small expansion cohort at the recommended phase II dose to better characterize safety (Supplementary Fig. S1). The primary objective was safety as assessed by NCI CTC v4.0 criteria. Dose-limiting toxicity (DLT) was defined as grade 4 neutropenia lasting ≥ 7 days; febrile neutropenia; grade 3 thrombocytopenia associated with bleeding or grade 4 thrombocytopenia lasting ≥ 4 days; any nonhematologic grade 3 ≥ toxicity despite the use of adequate/maximal medical interventions and/or prophylaxis; alanine aminotransferase or aspartate aminotransferase > 3× upper limit of normal (ULN) and total bilirubin >2× ULN with serum alkaline phosphatase normal; any toxicity that resulted in a >14-day delay in treatment; or any death not clearly due to underlying disease or extraneous causes. Secondary endpoints included ORR, clinical benefit at 18 weeks [CB₁₈; prospectively defined as all pts with a complete response, partial response, or stable disease (SD) for at least 18 weeks], and progression-free survival (PFS). Exploratory analyses probed for association between genomic/pathologic features and clinical efficacy to identify putative biomarkers for consideration in subsequent trials. The Kaplan–Meier method was used to analyze for PFS and OS [with median times and 95% confidence intervals (CI) calculated] using SAS Version 9.4. The study was approved by the Institutional Review Board at Indiana University (Indianapolis, IN); all pts provided written informed consent prior to study entry. This study was conducted in accordance with U.S. Common Rule.

The recommended phase II dose (RP2D) and toxicity profile had been established for each agent as monotherapy. As overlapping toxicity and pharmacokinetic interactions were not expected, we planned a limited dose escalation, starting with 50% of the monotherapy RP2D for each agent in cohort 1. The full RP2D of each agent as monotherapy was delivered in cohort 3 (Table 1). Gedatolisib was infused first followed by cofetuzumab pelidotin. Pts were treated prophylactically with a steroid mouthwash prior to administration of gedatolisib to minimize mucositis. Pts were evaluated clinically each week during the first two cycles and on day 1 of each subsequent cycle. Serum chemistry was obtained at the start of each cycle; complete blood counts were obtained prior to each gedatolisib infusion. Dose modifications for hematologic and nonhematologic toxicity were prespecified. Response was evaluated according to RECIST 1.1 every two cycles (6 weeks) through week 18, then every three cycles thereafter. Tumor biopsies for correlative analyses were obtained at screening and cycle 1 day 15 in pts with accessible lesions.

PTK7 IHC was performed by Flagship Biosciences using a proprietary antibody for staining under a protocol developed by Pfizer (29). Staining for phospho-AKT (pAKT, Ser473), was performed using antibody #4060 procured from Cell Signaling Technology as described previously (30). PTK7 and pAKT IHC results were assessed by a certified pathologist and quantified by H-score.

DNA from tumor and matched blood normal was extracted using the Qiagen AllPrep DNA/RNA FFPE kit and QIAamp DNA Blood Mini kit, respectively. DNA library preparation utilized the SureSelect XTHS Target Enrichment System for Illumina Paired-End Multiplexed Sequencing Version A1, July 2017 (Agilent Technologies). Libraries were then hybridized, captured, and amplified with the Agilent Human All Exon V7 probe set (48 Mb, hg38). Libraries were sequenced on a NovaSeq 6000 sequencer using 150 bp paired-end chemistry (Illumina, Inc.). A Phred quality score (Q score) was used to measure the quality of sequencing. More than 90% of the sequencing reads reached Q30 (99.9% base call accuracy).

FASTQ files were aligned to the human reference genome (b37/hg19) using the Burrows-Wheeler Aligner (v. 0.7.17, http://bio-bwa.sourceforge.net/; ref. 31). PCR duplicates were removed and coverage metrics were calculated using Picard tools (v.2.21.2, http://picard.sourceforge.net/). The Genome Analysis Toolkit (GATK, v. 4.1.4.1, http://www.broadinstitute.org/gatk/) was used for SNP and INDEL discovery according to GATK best practices (32). Copy-number variants were analyzed using CODEX2 (v.1.3.0; ref. 33).

Raw sequencing data for this study were generated by The Center for Medical Genomics at Indiana University School of Medicine (Indianapolis, IN). Derived data supporting the findings of this study are available in the Supplementary Data.

We enrolled 18 pts: 10 in three dose-escalation cohorts and 8 in expansion at cohort 3. Baseline characteristics are shown in Table 2. Pts were heavily pretreated with the majority having prior exposure to an anthracycline, taxane, platinum, and capecitabine.

Dose escalation was completed with no DLTs; an additional pt was included in cohort 1 to accommodate an immediate clinical need for treatment. A small expansion cohort proceeded at the full RP2D for monotherapy for both agents. The most common adverse events of any grade were nausea (n = 16, 89%), anorexia (n = 13, 72%), constipation (n = 12, 67%), fatigue (n = 12), and mucositis (n = 12). Grade 3 or greater toxicity (nausea, n = 1; fatigue, n = 2) and myelosuppression (grade ≥3 neutropenia, n = 2) were uncommon (Table 3).

A total of 16 of 18 pts enrolled were evaluable for response; 2 pts discontinued treatment prior to the first disease assessment that was not due to toxicity or progression. Three pts achieved a confirmed partial response (ORR, 16.7%; 3/18 including the two nonevaluable pts). An additional 3 pts had SD; of these 2 had SD for > 18 weeks. The CB₁₈ was 27.8% (5/18; Fig. 1). Of note, 1 pt with SD with Inflammatory TNBC (Pt #: 0613-25) had their only target lesion biopsied and therefore was not eligible for measurement per RECIST; however, they had a prolonged exceptional response of approximately 2 years. The pt has been off all anti-cancer treatment for > 15 months. Overall, after a median follow-up of 7.9 months, median PFS was 2.0 months (95% CI for PFS: 1.2–6.2; Fig. 2A); median OS was 9.4 months (95% CI for OS: 4.3–15.8; Fig. 2B).

Using whole-exome sequencing, we set out to observe if somatic DNA variants were associated with clinical response (Table 4). Of the 18 pts on trial, 15 had successful whole-exome sequencing of the baseline biopsy sample. We did not detect abberations in the PI3K pathway or PTK7 in 6 of 15 pts. In pts who had a partial response (n = 3), we observed a PTEN I101N mutation in 1 pt and a high gain of AKT2 in another. In pts who had SD (n = 3), one harbored a high gain of AKT2, another had a PIK3CA E545A gain-of-function mutation and PTEN single-copy deletion, and a third had high gains in AKT2, PIK3CG, and PTK7. In those pts with progressive disease (n = 10), we observed an AKT2 S268 L gain-of-function mutation in 1 pt, a PIK3CA H1047R gain-of-function mutation in 1 pt, a PTK7 variant (of unknown significance), and high gains of AKT2 in 3 pts and PIK3CB in 2 pts. Overall, 5 of 5 pts with either a partial response (PR) or SD and matched exome sequencing harbored an aberration in the PI3K pathway or PTK7, compared with 4 of 9 pts with progressive disease.

Of note, in our pt with exceptional response to therapy (Pt #: 0613-25), we observed that the pt had a concurrent PIK3CA E545A gain-of-function mutation and a PTEN deletion. These data demonstrate two mutations that are consistent with potential activation of the PI3K pathway and may help explain the exceptional response seen in this pt.

Tissues from research biopsies obtained prior to therapy and at C1D15 were stained for the expression of PTK7 as well as pAKT, a pharmacodynamic marker of PI3K inhibition (Table 4). There was no correlation between IHC H-scores for either marker and timepoint with clinical benefit. However, due to the small sample size, this observation is preliminary.

In this study, we have shown that the novel combination of gedatolisib and cofetuzumab pelidotin can be delivered in combination at full doses with manageable toxicity and has clinical activity in pts with heavily pretreated metastatic TNBC. While the PI3K pathway is genomically and transcriptomically aberrant in the majority of TNBCs, single-agent treatment with inhibitors of this pathway has resulted in only modest clinical activity, contrary to the theory of oncogene addiction (34, 35). Transcriptome reprogramming has emerged as a common response mechanism in tumor cells when exposed to small-molecule perturbations (36). Four independent groups have reported that upregulation of the Wnt pathway is a common compensatory response to PI3K inhibition driving resistance (19–22). Inhibition of both pathways resulted in synergistic antitumor efficacy in preclinical models (19). We theorized that we could leverage this compensatory mechanism to inform clinical combination therapy by using a PI3K inhibitor (gedatolisib) to potentially drive upregulation of PTK7, such that synergy would be achieved with cotreatment with a PTK7-ADC (cofetuzumab pelidotin). Furthermore, previously reported data has also shown that the payload for cofetuzumab pelidotin, auristatin, is in itself synergistic with gedatolisib, providing another potential mechanism for synergistic action of this combination (28).

In this phase I trial, we report the first clinical experience of combining a PI3K pathway inhibitor with an agent that targets a component of the Wnt pathway. The trial successfully completed its dose escalation to RP2D of both agents with no DLTs. Safety data from the dose escalation and our expansion cohort found the combination to be generally well tolerated. Furthermore, while our sample size was limited, we observed clinical activity of the combination with a CB₁₈ of 27.8%. Given the paucity of targeted agents for the majority of patients with TNBC, this combination provides a promising avenue for increasing the clinical armamentarium. Our study was not powered to definitely detect biomarkers of response; however, patients with genomic aberrations in the PI3K pathway or PTK7 may preferentially benefit. Taken together, these data set the stage for larger phase II studies of the combination in metastatic TNBC.

M. Radovich reports other support from LifeOmic, Tyme Technologies, and Caris Life Sciences; personal fees and non-financial support from Eli Lilly; and grants from Boston Biomedical outside the submitted work. J.P. Solzak reports other support from Caris Life Sciences, Inc. outside the submitted work. C.J. Wang reports other support from Caris Life Sciences outside the submitted work. B.A. Hancock reports other support from LifeOmic outside the submitted work. S. Badve reports grants from Agilent and Lilly, as well as personal fees from BMS outside the submitted work. S.M. Bray reports other support from LifeOmic, Inc. outside the submitted work. T.J. Ballinger reports personal fees from Medscape and Novartis outside the submitted work. B.P. Schneider reports non-financial support from Genentech, Pfizer, Foundation Medicine, and Epic Sciences, as well as personal fees from Eli Lilly outside the submitted work. K.D. Miller reports non-financial support from Pfizer, as well as grants from Breast Cancer Research Foundation during the conduct of the study. K.D. Miller also reports personal fees from Merck, Roche/Genentech, and AstraZeneca, as well as grants from Astex, Pfizer, CytoMx, AbbVie, Amgen, British Biotech, Merck, and Seattle Genetics outside the submitted work. No disclosures were reported by the other authors.

M. Radovich: Conceptualization, resources, formal analysis, supervision, funding acquisition, investigation, methodology, writing–original draft, project administration, writing–review and editing. J.P. Solzak: Conceptualization, formal analysis, investigation, methodology, writing–original draft, writing–review and editing. C.J. Wang: Investigation, writing–review and editing. B.A. Hancock: Investigation, writing–review and editing. S. Badve: Investigation. S.K. Althouse: Data curation, formal analysis, visualization, writing–review and editing. S.M. Bray: Data curation, software, visualization, writing–review and editing. A.M.V. Storniolo: Investigation, writing–review and editing. T.J. Ballinger: Investigation, writing–review and editing. B.P. Schneider: Investigation, writing–review and editing. K.D. Miller: Conceptualization, resources, formal analysis, supervision, funding acquisition, investigation, methodology, writing–original draft, project administration, writing–review and editing.

This work was supported by the NIH/NCI R21CA229951 (M. Radovich); Breast Cancer Research Foundation (K.D. Miller); 100 Voices of Hope; The Vera Bradley Foundation for Breast Cancer Research; The Catherine Peachey Fund; and the Indiana University Precision Health Initiative. Drug support was provided by Pfizer.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.