Phase I Study of Single-Agent Anti–Programmed Death-1 (MDX-1106) in Refractory Solid Tumors: Safety, Clinical Activity, Pharmacodynamics, and Immunologic Correlates

MDX-1106

MDX-1106 (BMS-936558/ONO-4538) is a genetically engineered, fully human immunoglobulin G4 (IgG4) mAb specific for human PD-1 (Appendix Fig A1A, online only). Mice transgenic for human Ig loci were immunized with Chinese hamster ovary cell PD-1 transfectants and a PD-1/human IgG1 Fc fusion protein. MDX-1106 contains an engineered hinge region mutation (S228P) designed to prevent exchange of IgG4 molecules; the IgG4 isotype minimizes cellular and complement-mediated cytolytic functions. MDX-1106 binds PD-1 with high affinity (K_D = 2.6 nmol/L by Scatchard analysis to polyclonally activated human T cells), blocks its interactions with both B7-H1 and B7-DC (Appendix Fig A1B), and enhances tumor antigen-specific T cell proliferation and secretion of cytokines in vitro.¹⁵

Patients

Eligible patients had treatment-refractory metastatic melanoma, castrate-resistant prostate cancer, renal cell carcinoma (RCC), non–small-cell lung cancer (NSCLC), or colorectal cancer (CRC), and had no cancer therapy for at least 4 weeks before enrollment. Patients were ≥ 18 years old with a life expectancy of ≥ 12 weeks, an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1, adequate organ function, and no ongoing systemic infections or history of autoimmune disease. Concurrent antineoplastic therapies, systemic steroids, and prior treatment with anti–CTLA-4 were not permitted. Patients with treated stable brain metastases were eligible.

Study Design and Procedures

This multi-institutional, first in-human, open-label, phase I, dose-escalation study was approved by local institutional review boards. All participating patients signed informed consent. The primary objectives were to characterize the safety and tolerability of a single dose of MDX-1106 in patients with selected malignancies and to determine the maximum-tolerated dose (MTD) and pharmacokinetics. Secondary objectives included assessing antitumor activity, pharmacodynamics, and immunologic end points. Sequential cohorts of six patients received a 60-minute intravenous infusion of MDX-1106 at 0.3, 1, 3, or 10 mg/kg and were evaluated for toxicities on a weekly basis for 8 weeks. Dose-limiting toxicity (DLT) was defined as a treatment-related grade ≥ 3 AE or laboratory abnormality occurring ≤ 28 days postdose. The MTD was the highest dose at which no more than one of six patients experienced a DLT. Fifteen additional patients were planned to be enrolled at the MTD or the highest planned dose (10 mg/kg) to confirm safety.

Patients were restaged radiographically (Response Evaluation Criteria in Solid Tumors [RECIST] 1.0) at 8 and 12 weeks. Patients with progressive disease were taken off study. Those with stable disease or evidence of lesional tumor regression, no AE grade ≥ 3, and no evidence of human antihuman Ab at a 1:10 serum dilution received additional doses of MDX-1106 at weeks 12 and 16 and were then observed for 3 months and restaged. Those with continued clinical benefit could receive two more doses, spaced by 4 weeks. Each re-treatment phase was 16 weeks. Patients with objective partial responses (PRs) or CRs were observed, with optional re-treatment on progression. Responses are reported as of January 2010.

Pharmacokinetics

Serum samples were collected serially before and up to 85 days after the first dose of MDX-1106. MDX-1106 serum concentrations were determined with a quantitative enzyme-linked immunosorbent assay (ELISA) capable of detecting ≥ 1.2 μg/mL, using 96-well plates coated with chimeric PD-1/human IgG1 Fc protein (R&D Systems, Minneapolis, MN).

Tumor Biopsies

Sections of formalin-fixed, paraffin-embedded tumor specimens archived before protocol entry or core-needle or excisional biopsies obtained immediately pre- and post-therapy were subjected to hematoxylin and eosin staining and immunohistochemistry (IHC) to detect lymphoid infiltrates (anti-CD3, anti-CD4, and anti-CD8) and B7-H1 expression (murine antihuB7-H1, clone 5H1; previously described¹³). Pigmented melanoma samples were bleached before staining and visualized with a red chromogen. Tumors were considered B7-H1–positive if ≥ 5% of tumor cells showed membranous staining with 5H1.

Immunologic Assessments

Delayed-type hypersensitivity reactions to Candida albicans and tetanus toxoid were assessed along with viral antigen recall reactions (details are included in the Appendix, online only). Peripheral blood lymphocyte (PBL) phenotypes were also assessed. Serially collected blood was analyzed for the presence and activation status of various lymphocyte subsets, as detailed in the Appendix.

PD-1 Receptor Occupancy (pharmacodynamics)

MDX-1106 binding to PD-1 molecules on circulating CD3⁺ PBLs was investigated with flow cytometric analysis of serially collected blood samples (see PBL phenotyping schedule in the Appendix). Peripheral blood mononuclear cells were preincubated (30 minutes at 4°C) with a saturating concentration (20 μg/mL) of either unlabeled huIgG4 (isotype control) or MDX-1106, washed extensively, and then costained with anti-CD3 fluorescein isothiocyanate and murine antihuIgG4 biotin (Invitrogen, Carlsbad, CA) plus streptavidin-phycoerythrin. PD-1 occupancy by infused MDX-1106 was estimated as the ratio of the percent of CD3⁺ cells stained with antihuIgG4 after in vitro saturation with isotype control Ab (indicating in vivo binding) to that observed after MDX-1106 saturation (indicating total available binding sites).

Patients and Treatments

Thirty-nine patients with advanced metastatic NSCLC, melanoma, castrate-resistant prostate cancer, RCC, or CRC received MDX-1106 in four escalating dose cohorts of 0.3 to 10 mg/kg and an expansion cohort at 10 mg/kg, from October 2006 through June 2009 (Tables 1 and 2). Their median age was 62 years. All had progressive treatment-refractory disease, and they had undergone a median of four prior therapies.

Treatment-Related Toxicities

MDX-1106 was well-tolerated: no DLTs were observed after one dose, and an MTD was not defined in this study. Grade ≥ 2 adverse clinical and laboratory events are summarized in Appendix Table A1 (online only). Most frequent were decreased CD4⁺ lymphocyte counts (14 patients, 35.9%), lymphopenia (10 patients, 25.6%), fatigue and musculoskeletal events (six patients each, 15.4%). No patient developed human antihuman Ab, even after multiple doses.

Immune-related AEs (irAEs) were of special interest because of the presumed mechanism of action of anti–PD-1 and prior experience with anti–CTLA-4.⁵ No grade ≥ 3 irAE occurred in the 28-day period following the first dose of anti–PD-1. One patient with metastatic ocular melanoma developed grade 3 inflammatory colitis following five doses (1 mg/kg) administered over 8 months (Appendix Fig A2, online only), which responded to steroids and infliximab. One patient (10 mg/kg) experienced grade 2 hypothyroidism requiring hormone replacement. Two patients (at 3 and 10 mg/kg) developed grade 2 polyarticular arthropathies requiring oral steroids and were not further treated; in retrospect, both had potentially contributory predisposing factors, one with a history of Lyme arthritis and polymyalgia rheumatica, the other with a preexisting antinuclear antibody titer > 1:1000.

Antitumor Activity

One patient with CRC (3 mg/kg) achieved a CR, and two patients with RCC (10 mg/kg) and melanoma (10 mg/kg) experienced PRs to therapy. A 67-year-old male with CRC metastatic to intra-abdominal lymph nodes received five doses of MDX-1106 and experienced a CR persisting 21+ months. A 72-year-old male with multiorgan metastatic RCC had a mixed response after one dose of MDX-1106, with progression in a pancreatic metastasis but regression in other sites; this evolved to an overall PR after two additional doses, lasting 16+ months without further therapy (Fig 1A). A 51-year-old female with melanoma metastatic to multiple lymph nodes and liver sites initially experienced a mixed response, with regression at all sites except an enlarging subpectoral lymph node, and achieved a PR after receiving 11 doses of MDX-1106 over 24 months (Fig 1B). Two additional patients had significant lesional or mixed tumor regressions (defined as regression in individual lesions with concomitant progression at other sites), including one with NSCLC (1 mg/kg) and another with melanoma (10 mg/kg). In total, 12 patients with stable disease or lesional tumor regressions at the first disease assessment received multiple doses of MDX-1106 (Table 2).

Tumor Biopsies

B7-H1 expression on tumors may affect the ability to respond to PD-1 blockade. To explore this, tumor biopsies from nine patients undergoing treatment with MDX-1106 were analyzed for B7-H1 expression with IHC (Appendix Table A2, online only). In two cases, both pre- and post-treatment samples were available; in the others, only pretreatment (six cases) or post-treatment (one case) samples were accessible. Tumor cell staining for B7-H1 expression fell into three patterns: negative, intracytoplasmic, or membranous (cell surface). Among nine patients studied, four exhibited membranous B7-H1 staining (Fig 2): three of these patients experienced tumor regressions following MDX-1106 therapy; the fourth was treated in the 0.3-mg/kg cohort where no responses were observed. Conversely, among five patients whose tumors failed to express B7-H1 at the cell surface, there was no evidence of clinical response. B7-H1 staining patterns were consistent in five patients from whom multiple biopsies were available (AppendixTable A2). In this small sample size, the correlation between membranous B7-H1 expression on tumor cells and the likelihood of tumor regression following PD-1 blockade suggested potential significance (two-sided P = .0476; Fisher's exact test).

Melanoma patient 3019, who experienced a PR to anti–PD-1 therapy, underwent pre- and post-treatment biopsies of an axillary lymph node metastasis for characterization of intratumoral lymphoid infiltrates by IHC. Whereas the pretreatment biopsy contained only sparse lymphoid cells, subsequent tumor regression was accompanied by a moderate infiltration of CD8⁺, but not CD4⁺, T cells (Fig 1B).

Reactivity to Recall Antigens and PBL Phenotypes

No significant effects of MDX-1106 therapy on delayed-type hypersensitivity responses against C albicans or tetanus toxoid or antiviral recall responses by interferon gamma ELISpot analysis were observed in 15 and six patients examined, respectively. The effects of a single 10-mg/kg dose of MDX-1106 on PBL numbers, subset profiles, and activation status were analyzed in 17 patients (Fig 3). Twenty-four hours postdose, total lymphocyte as well as CD3, CD4, and CD8 numbers declined and then rebounded from days 2 through 29 and declined again from days 29 through 85. These trends were not observed for CD19 (B lymphocyte) or CD56 (natural killer) cells (not shown), suggesting a selective effect on T cells. Percentages of CD3, CD4, and CD8 cells declined steadily over the 85-day observation period, with a reciprocal increase in CD19 and CD56 cell percentages (Appendix Fig A3A, online only). No significant changes were observed in expression of the T cell activation markers CD25, CD45RO, or HLA-DR by CD4⁺ or CD8⁺ PBLs (Appendix Fig A3B).

PD-1 Receptor Occupancy (pharmacodynamics)

PD-1 has predominant cell surface expression, unlike CTLA-4, which is displayed only transiently at the T cell surface. Thus, it was possible to develop flow cytometric methods to evaluate the pharmacodynamics of infused MDX-1106, estimating PD-1 occupancy on circulating T cells over time. Standard pharmacokinetic measurements of MDX-1106 serum concentrations yielded an approximate serum half-life (t_1/2) of 12 days (0.3-, 1-, or 3-mg/kg dose) to 20 days (10 mg/kg), with maximum concentration (C_max) and AUC directly related to dose. However, pharmacokinetics and pharmacodynamics were unexpectedly discordant in 15 patients studied (Fig 4). PD-1 occupancy appeared to be dose-independent, with a mean peak occupancy of 85% (range, 70% to 97%) and a mean plateau occupancy of 72% (range, 59% to 81%) observed at 4 to 24 hours and ≥ 57 days, respectively, after one infusion (Fig 4A). These data are consistent with the high affinity of MDX-1106 for PD-1—in vitro, 0.04 μg/mL MDX-1106 is sufficient to occupy > 70% PD-1 molecules on T cells (not shown)—suggesting that even when serum levels are undetectable (< 1.2 μg/mL), sufficient concentrations persist to maintain plateau PD-1 occupancy. Occupancy eventually decayed after 85 days (Fig 4B, top and middle panels). In patients receiving repeated infusions of MDX-1106 at 10 mg/kg, troughs and peaks of PD-1 occupancy around each dose were observed (Fig 4B, middle and bottom panels), although 100% occupancy was not achieved. In vitro experiments indicate that PD-1 occupancy analyses of cryopreserved PBLs may underestimate occupancy on fresh PBLs (not shown). It is unknown whether these findings in circulating lymphocytes reflect PD-1 occupancy on lymphocytes in the tumor, secondary lymphoid organs, and/or tissues.