Dose-Ranging and Cohort-Expansion Study of Monalizumab (IPH2201) in Patients with Advanced Gynecologic Malignancies: A Trial of the Canadian Cancer Trials Group (CCTG): IND221

Cancer immune evasion is largely regulated by immune checkpoint proteins. Cancers exploit immune checkpoint pathways as important mechanisms of immune resistance. Some cancers upregulate the expression of inhibitory ligands, resulting in a state of immune tolerance/immune exhaustion, and several immune suppressive pathways have been identified. These include inhibitory receptors, for example, CTLA-4 and PD-1, which, when engaged by their ligands (B7 and PD-L1, respectively) expressed on antigen-presenting cells or tumor cells, block T-cell activation. Epithelial ovarian, endometrial, and cervical cancers can upregulate immune checkpoints within the tumor microenvironment, abrogating an effective host immune response. As many of the immune checkpoints are initiated by ligand–receptor interactions, they can be blocked by antibodies or modulated by recombinant forms of ligands or receptors (1).

HLA-E is a nonclassical MHC class I molecule. It serves as a ligand to CD94/NKG2A, a major immune checkpoint receptor expressed on subsets of natural killer (NK) cells and some activated CD8⁺ T lymphocytes (2). The binding of HLA-E to CD94/NKG2A transduces inhibitory signals that suppress cytokine secretion and cytotoxicity. The normal immunologic role of this interaction is to prevent excessive immune-mediated tissue damage, for example in the setting of infection (3). HLA-E and CD94/NKG2A are upregulated in a variety of tumor types. In some cancers, this is associated with adverse outcomes, for example, in epithelial ovarian cancer (EOC) HLA-E expression abrogated the favorable effects of CD8⁺ cytotoxic T cells (CTL; ref. 4). A similar effect was also reported in non–small cell lung carcinoma (5), and hepatocellular carcinomas (6). In colorectal cancer, loss of HLA-E has been associated with improved survival, potentially as a result of susceptibility to NK-cell elimination (7, 8).

Monalizumab (IPH2201) is an mAb with high affinity for the CD94/NKG2A receptor; therefore, it has the potential to function as a therapeutic immune checkpoint inhibitor. Monalizumab binding to CD94/NKG2A can restore the function of NK cells by blocking the transduction of inhibitory signals. A phase I study of monalizumab in patients with rheumatoid arthritis had demonstrated minimal toxicity with intravenous and subcutaneous dosing of up to 10 mg/kg i.v. with evidence of pharmacodynamic effect (i.e., CD94/NKG2A receptor saturation) even at low doses (e.g., 0.005 mg/kg i.v.; ref. 9).

As monalizumab was of interest to test in patients with advanced/metastatic/recurrent gynecologic cancers including high-grade serous ovary/fallopian tube or peritoneal carcinoma (OVA-S and OVA-R; platinum-sensitive and platinum-resistant, respectively), and endometrial (END) and cervical (CX) cancers, we undertook a multicenter phase I dose-ranging trial of single-agent monalizumab, testing 3 prespecified dose levels. This 2-part trial included a randomized dose-ranging study (part 1), followed by a cohort expansion (part 2). Translational studies looking for biomarkers of response were undertaken. Herein we report the results of this trial.

Study accrual was from September 2015 to November 2016. Trial participants had histologically and/or cytologically confirmed advanced/metastatic/recurrent malignancy, measurable disease (RECIST 1.1 criteria; ref. 10), an ECOG PS of 0, 1, or 2, adequate hematologic (absolute neutrophil ≥ 1.5 × 10⁹/L, platelets ≥ 100 × 10⁹/L) and biochemical parameters [bilirubin ≤ 1.5 × ULN (upper limit of normal)], AST and ALT ≤2.5 × ULN (≤ 5 × ULN with liver metastases), serum creatinine (<1.25 × ULN or creatinine clearance ≥ 50 mL/min), and were ≥18 years of age. Patients with OVA-S/OVA-R, CX (squamous histology only), or END (epithelial cancers, exclusive of carcinosarcoma) were eligible; platinum-resistance was defined as progression <6 months from the last line of platinum-based therapy. All patients had archival tumor tissue and had at least one prior platinum-based regimen of chemotherapy for advanced, metastatic, or recurrent disease and had recovered from reversible treatment toxicities. Patients with more than 3 prior cytotoxic regimens were excluded; however, other therapies such as immunotherapy, hormone therapy, angiogenesis inhibitors, PARP inhibitors, or signal transduction inhibitors were also permitted. Exclusion criteria included: any serious illness or medical conditions that might be aggravated by treatment or limit active immune-mediated diseases or known HIV infection or hepatitis B or C, use of systemic corticosteroid therapy >5 mg prednisone/day. Cytokines or growth factors were not permitted. Patients could not have residual severe adverse effects from prior immunologically based treatment using monoclonal antibodies, or be receiving any other concurrent anticancer or investigational agents. The protocol was approved by the institutional review boards of participating institutions, and written informed consent was obtained from all patients prior to performing study-related procedures. The study was conducted in accordance with the Declaration of Helsinki.

Part 1 was a randomized (1:1:1), open-label, dose-ranging study of 3 dose levels of monalizumab: 1 mg/kg, 4 mg/kg, and 10 mg/kg (i.v.) once every 2 weeks. Eighteen patients were to be randomized. Dose-limiting toxicity (DLT) was not expected (9) and not predefined. As DLT was not anticipated based on data from other studies, the RP2D was to be determined after review of toxicity, pharmacokinetic, and pharmacodynamics. Part 2 of the trial was a 2-stage cohort expansion design with monalizumab given at the RP2D. Response to therapy was assessed using RECIST v 1.1 (10) criteria every 8 weeks until progression of disease (PD); in addition, as an exploratory measure, investigators were advised to continue treatment, in the absence of unacceptable toxicity, until unequivocal PD, and to confirm PD with a second CT scan (11). All patients who had at least one cycle of therapy and had their disease reevaluated were considered evaluable for response (those who exhibited objective disease progression prior to the end of cycle 1 were also to be considered evaluable). Adverse events were graded using CTCAE V4.0 (12).

Monalizumab was given by intravenous infusion over 1 hour on day 1 of each 2-week cycle. Dose reductions were not permitted but doses were held for adverse events considered related to the protocol therapy (≥grade 3 nonhematologic/organ toxicity, febrile neutropenia, grade 4 neutropenia lasting ≥7 days). Monalizumab was discontinued for any grade 4 nonhematologic/organ-related adverse events and if related adverse effects did not recover to ≤grade 2 within 2 weeks. The planned dose intensity (in units per week) was calculated based on the planned dose, planned number of doses per cycle, and planned cycle length in weeks.

The primary study objective of part 1 was to evaluate the RP2D of single-agent monalizumab, in patients with gynecologic malignancies. Secondary objectives were to characterize the pharmacokinetics, pharmacodynamics, safety, toxicity, efficacy, and immunogenicity of monalizumab. Part 2 evaluated, in an exploratory fashion, activity in OVCA-R, OVCA-S, END, and CX carcinomas.

Based on a stable disease (SD) hypothesis of H0 <5% and Ha >30%, the drug was to be rejected at the end of part 2, stage 1 of accrual, if 2 or fewer SDs were seen in any one cohort. Otherwise, per protocol, stage 2 expansion was permitted if at least 3 SDs were observed in any of the 4 cohorts in part 2. The significance level (i.e., the probability of rejecting H0 when it is true) of this procedure is α = 0.01 and the power (i.e., the probability of rejecting H0, i.e., deciding the regimen is active, when H1 is true) is 0.81.

Biomarker expression levels between patients with SD and PD, as well as between archival and baseline biopsy samples, were compared using 2 sample t tests.

Archival (primary or metastatic), formalin-fixed and paraffin embedded, tumor tissue blocks were required for study participation. Part 2 of the study required that at least 4 patients in each disease cohort undergo paired pre- and on-treatment biopsies (up to 7 days prior to day 1 of cycle 3) plus optional biopsy at the time of disease progression. Whole blood, plasma, and serum were collected on all patients at baseline, predose on day 1 of cycles 4 and 8, and at the time of confirmed PD. Samples were used for the evaluation of potential biomarkers of response and for evidence of a pharmacodynamics effect.

All available tumor samples were analyzed for IHC expression of the following exploratory biomarkers: HLA-E expression was assessed separately on tumor cells, lymphocytes, and endothelium; PD-L1 expression was analyzed on tumor cells and CD8⁺, NKp46⁺, and CD94⁺ cells were quantified separately in the tumor nest and the stroma.

Tumor sections were evaluated individually by a pathologist for intensity of signal and proportion of cells staining positive for the targeted protein. Intensity scores were assigned as follows: 0 = negative, 1 = faint to weak, 2 = intermediate, 3 = strong. Proportion scores were assigned as follows: 0 = no positive cells detected, 1 = between 1% and 33% of positive cells, 2 = between 34% and 66% of positive cells, and 3 = ≥67% of positive cells. An overall IHC score representing the product of the intensity and proportion scores was derived (thus, a numerical value of 0, 1, 2, 3, 4, 6, or 9). Results were grouped based upon treatment response and mean scores with standard deviations are reported.

Flow cytometry on whole blood, using the CELLSEARCH system was used to enumerate circulating tumor cells (CTC) at baseline and over the course of therapy. CTCs collection was mandatory pretreatment, on cycle 4, day 1, and at disease progression.

Samples for pharmacokinetics were drawn: day 1 cycles 1 and 4 (predose, end of infusion, 2 and 24 hours post-infusion), day 1 cycles 2, 3, 4, and 8 pre- and post-infusion in part 1 and pre- and post-infusion day 1 cycles 1, 4, and 8 for part 2. Concentration at the end of administration (C_eoi), concentration at predose before administration (C_trough), and accumulation ratio (C_eoi between cycle 1 and cycle 4) were evaluated. For pharmacokinetics, samples were incubated on coated monalizumab anti-idiotypic mAb, and bound monalizumab was detected using peroxidase-conjugated anti-human IgG4 Ab by colorimetric reaction (TMB); samples were assayed by batch with one calibration curve (0.5–100.0 ng/mL with minimal residual dilution of 10) and at least 2 sets of quality controls (QC). All other samples were drawn at baseline, cycles 4 and 8 day 1 (pre-dose) and at progression. For antidrug antibodies (ADA), samples were incubated on a mix of biotin labeled monalizumab and SULFO-TAG labeled monalizumab. Anti-monalizumab bound antibodies were detected using biotin fixation on 96-well streptavidin plate and electrochemical stimulation of SULFO-TAG; samples were assayed by batch with at least 2 sets of QCs, in duplicates. Each sample tested positive in ADA assay was confirmed in a confirmatory assay, in the presence of excess monalizumab. The cut point of each plate and the signal inhibition (for a confirmatory assay) of each sample was calculated.

Saturation of CD94/NKG2A receptors expressed on peripheral NK and CD8⁺ T cells were assessed by flow cytometry on whole blood, collected into EDTA tubes. Detection of monalizumab bound to CD94/NKG2A was performed by staining with PE-conjugated mouse antihuman IgG₄ (clone HP6025; Southern Biotechnology); detection of the total of CD94/NKG2A expressed on cell membrane was performed from a separate tube by staining ex vivoctg with saturating amount of monalizumab followed by PE-conjugated HP6025. The intensity of the PE fluorescence was normalized to molecules of equivalent soluble fluorochrome (MESF) using Quantum MESF beads (Bang Laboratories). Percentage of NKG2A saturation is the ratio of MESF of the bound compare to the total at each time point.

Table 1 summarizes all baseline patient characteristics for parts 1 and 2 of the study.

Eighteen patients with high-grade serous carcinoma of the ovary (14 patients), fallopian tube (2 patients), or peritoneum (2 patients) were randomized to part 1, 6 to each of 3 dose cohorts: 1, 4, 10 mg/kg. Median age was 60 years old (range, 36–74). All patients had had prior cytotoxic chemotherapy, with 78% (14/18) having >1 prior chemotherapy regimens. Eleven patients were platinum-resistant and 7 were platinum-sensitive. All 18 patients are off study: 15 for progressive disease, 1 for symptomatic progression, 1 for patient request, and 1 for investigator decision. Seventeen patients were evaluable for response; one patient experienced symptomatic progression and death before assessment of response could be completed.

Forty patients were enrolled and treated (1 additional patient was cancelled prior to treatment) 10 OVA-R, 9 OVA-S, 12 END (additional 2 enrolled to meet paired biopsy requirement due to omission of paired biopsies), 9 CX. The median age was 61 (range 33–82). All patients had received prior systemic therapy with 70% (28/40) having >1 prior chemotherapy regimens. All 40 patients are off study: 1 due to disease related death, 37 for disease progression, and 2 for symptomatic progression. Thirty-eight patients were evaluable for response: 1 patient had symptomatic progression and death before assessment of response could be completed, and 1 patient was ineligible (endometrial carcinosarcoma).

The treatment delivery data for parts 1 and 2 are summarized in Table 2.

For the 1/4/10 mg/kg cohorts, patients received 33/35/41 total cycles. The median number of cycles for all doses was 6. Dose delays were mostly for administrative/scheduling reasons, although 1 patient in the 4 mg/kg dose cohort had a delay due to fatigue, and 1 patient in the 10 mg/kg dose cohort had a delay due to dehydration and nausea. At least 90% planned dose intensity was achieved in 78% of patients.

For the OVA-R, OVA-S, END, and CX cohorts, patients received 86/57/89/58 total cycles. The median number of cycles for each cohort was: 7/5/4.5/4 (OVA-R/OVA-S/END/CX, respectively). Dose delays were mostly for administrative/scheduling reasons, although other reasons for delays included patient request (OVA-R 2, CX 1), anemia (OVCA-R 1), hyponatremia (OVA-S 1), intercurrent illness or management of disease related complications (END 2, CX 2), and investigator decision (OVA-R 1, END 1, CX 1). At least 90% of the planned dose intensity was delivered to 87.5% of patients.

All patients were evaluable for toxicity. The most common adverse events (of any causality) were fatigue, constipation and abdominal pain, and nausea, each in >60% of patients (Supplementary Table S1). The most common treatment-related adverse events, of any grade, were headache, fatigue, and vomiting, each in >20% of patients. Of the events considered related to monalizumab, 1 patient on the 10 mg/kg dose had grade 3 nausea, vomiting, and dehydration, and 1 patient on each of the 1 and 4 mg/kg doses had grade 3 fatigue. There was no obvious dose relationship for adverse events, although grade 2/3 fatigue, abdominal pain, and back pain seemed more common at the 4 mg/kg dose compared with the 1 and 10 mg/kg doses. Table 3 summarizes ≥grade 3 treatment-related nonhematologic adverse events.

Hematologic and biochemical changes were mild, with 1 patient on the 1 mg/kg dose cohort experiencing grade 3 hypoalbuminemia and 1 patient on the 10 mg/kg dose cohort experiencing grade 3 hypokalemia (with normal baseline values; Table 4).

Ten serious adverse events (SAE) were reported in 8 patients. Only one of these was considered possibly related to protocol treatment (grade 3 lung infection/hypotension/nausea/vomiting, at the 10 mg/kg dose cohort).

Based on these results, the 10 mg/kg dose was selected as the dose for part 2 of the study.

All 40 patients were evaluable for toxicity. The most common nonhematologic adverse events (any causality) were fatigue, constipation, abdominal pain, nausea, anorexia, headache, and back pain, each in >40% of patients. The most common adverse events considered related to protocol treatment were abdominal pain, headache, nausea, and fatigue, each in >15% of patients. One patient on the OVA-R cohort had grade 3 anorexia, nausea, and dyspnea, and 1 patient on the CX cohort had grade 3 proctitis, considered related to monalizumab (Table 4).

Forty patients were evaluable for hematologic and biochemical adverse events (1 patient on the OVA-R was cancelled prior to treatment). In patients with normal baseline values, on the OVA-S, 1 patient had grade 3 anemia and one on the OVA-S cohort had grade 4 hypokalemia. On the END cohort, 1 patient had disease related grade 4 bilirubin, grade 3/4 increases in AST/ALT, and grade 3 hyperglycemia in the context of preexisting type II diabetes mellitus. On the CX cohort, 1 patient had grade 3 anemia and grade 3 hypercalcemia (Table 4).

Thirty-seven SAEs were reported in 21 patients. Only one of these was considered related to protocol treatment (grade 2 infusion related reaction, END cohort).

Among 17 evaluable patients no responses were observed. Seven patients had SD (median duration 3.4 months, range: 1.4–5.5 months) and 10 had PD as a best response (Fig. 1A; Table 5). There were no significant differences in clinical factors between those experiencing PD and SD, and SD was observed at all dose levels. There was no evidence of pseudoprogression with late response.

Among 38 evaluable patients no responses were observed. Seven patients had SD (median duration 3.4 months, range: 2.6–14.8 months), and 31 patients had PD (Table 5; Fig. 1B; Supplementary Fig. S3). There was no evidence of pseudoprogression with late response. The number of SDs observed per cohort did not meet protocol-defined criteria for cohort expansion.

In part 1, pharmacokinetics were dose proportional for both C_eoi and C_trough (Supplementary Fig. S1; Supplementary Table S2). Low-to-moderate intersubject variability was observed. There was a 1.4- to 2-fold accumulation of monalizumab (in terms of C_eoi) after the administration of 4 intravenous doses, which was consistent with administration frequency and the t_½ seen with monalizumab in previous studies.

Complete and continuous saturation of CD94/NKG2A receptors expressed on peripheral blood cells was achieved following the first infusion of monalizumab in all patients and at all dose levels (detected from the first time point assessed 2-hour post-infusion in part 1 and 1-hour post-infusion in part 2).

None of the biomarkers assessed from baseline archival samples of all 56 patients (parts 1 and 2 together) were predictive of treatment response (Supplementary Table S3). However, there was a trend to statistical significance between tumor lymphocytic HLA-E intensity scores and SD (P = 0.095 and 0.089, respectively) in END patients (N = 11, archival samples part 2) with lower scores for patients with SD (Supplementary Table S4).

Pretreatment baseline biopsy samples were available for 15 patients (N = 2 part 1, and N = 13 part 2). Overall biomarker expression scores were low, including PD-L1, NKp46, and CD94, with stromal CD8 and lymphocytic HLA-E being the only markers to score above a total score of 3 (Supplementary Table S5A). None of the biomarkers assessed on these samples were predictive of response, although a trend to significance was found between the lymphocytic HLA-E proportion, total scores, and SD (P = 0.07 for both) in the 13 part 2 patients (Supplementary Table S5B).

Fourteen (N = 2 part 1, N = 12 part 2) paired pre- and posttreatment biopsy samples were available for analysis. A trend to significance was found between the reduction in lymphocyte HLA-E total score and SD among all patients (P = 0.059) and part 2 patients (P = 0.051; Supplementary Table S6A and S6B).

Exploratory analyses of the concordance of biomarker expression between archival and pretreatment biopsy samples were performed (N = 26). There was a trend for archival samples to be associated with higher levels of CD8 on infiltrating lymphocytes and tumor PD-L1 expression (P = 0.077 and 0.092, respectively). Statistically significant differences in expression of tumor, endothelial, and lymphocytes HLA-E were also observed, with mean expression levels being higher in archival samples as compared with pretreatment biopsy samples (P = 0.006, 0.016, 0.012, respectively).

Forty-four patients had blood collected for CTC analysis at baseline, cycle 4 day 1 (pre) and at progression. Baseline CTCs in 44 patients were 18 = 0, 8 = 1–2, 4 = 3–4, 14 ≥5, while the minimum CTCs on treatment were 23 = 0, 0 = 1–2, 12 = 3–4, 9 ≥5. There is a strong correlation between baseline and on treatment CTCs (P = 0.02). SD was observed on respectively 5 patients with 0 CTC at baseline, 1 with1–2, 0 with 3–4 and 5 with ≥5 and on respectively 4 patients with 0 CTC on treatment, 4 with1–2, and 3 with ≥5. No significant association was found between either baseline CTC or on-treatment CTC with SD (P = 0.49 and 0.47, respectively).

This is the first reported trial of single-agent monalizumab (IPH2201), a first-in-class mAb targeting CD94/NKG2A, in solid human malignancies. In part 1 of the study (dose-ranging), 10 mg/kg i.v. every 2 weeks was selected as the RP2D. In part 2 (cohort expansion), 40 patients with pretreated gynecologic cancers were treated. Monalizumab was well tolerated with few and mostly low-grade adverse events. None of the observed toxicities were felt to be immunologically mediated, in contrast to other immune checkpoint inhibitors (e.g., PD-L1/PD-1 and CTLA-4 inhibitors), that have a variety of well-defined, and potentially severe, immune-related adverse events caused by immune activation (13). However, in part 2, no cohort met the protocol-defined criteria for proceeding to stage 2, although some short interval disease stability was documented.

Modulation of the immune-microenvironment in gynecologic cancers is a rational therapeutic strategy. PD-1/PD-L1 immune checkpoint inhibition has been effective in several cancer types (14), including gynecologic cancers, harboring mismatch repair deficiencies (MMRd; e.g., MMRd endometrial carcinomas; ref. 15) and metastatic cervical cancers (16), both of which have FDA approvals for immune-checkpoint therapy. PD-1 inhibition in endometrial cancers with MMRd has led to response rates of >50%, with long-term disease control and occasional complete responders (15, 17, 18). PD-L1 expressing endometrial cancers, without MMRd, may also respond to PD-1 inhibition; although response rates are lower (∼13%; ref. 19). PD-L1 expressing cervical cancers benefit from PD-1 inhibition, with durable disease stabilization despite an overall response rate of <15% (15). In ovarian cancer, modest activity with immune checkpoint inhibition is reported (reviewed in ref. 20) with many ongoing studies.

HLA-E is expressed in most tissues and binds to its receptor, the CD94/NKG2 heterodimer complex. The CD94/NKG2A receptor complex is inhibitory, whereas the CD94/NKG2C complex is activating and both complexes are expressed on NK cells and a subset of CD8⁺ CTLs (2, 21). NK cells are part of the innate immune system, providing immediate reactivity against viral infections and also against tumors (21). However, NK cells also interact with T cells of the adaptive immune system and are involved in self-tolerance and can develop immunologic memory (23–27). In cancer tissues, this interaction may play a role in immune escape by tumor cells. HLA-E has been shown to be expressed in a variety of cancer types, and cytokines present in tumor microenvironments, such as IL15 and TGFβ, can induce expression of the inhibitory receptor, CD94/NKG2A (28–30).

The effect of HLA-E expression on solid cancer immune response and the impact on prognosis is not yet well understood. In breast and colorectal cancers, improved survival has been reported when HLA-E is lost (8, 31). In endometrial cancer, HLA-E expression was shown to be prognostic only when tumor NK cells were also present (32) with the level of HLA-E expression being important, as “normal” HLA-E expression led to a worse DFS, whereas “upregulated” HLA-E expression resulted in better DFS. It was proposed that the normal expression of HLA-E leads to the preferential binding of the inhibitory CD94/NKG2A complex, whereas upregulation of HLA-E expression results in simultaneous binding to the activating NK receptor, CD94/NKG2C, leading to overall NK-cell activation (32). It has been demonstrated that the inhibitor CD94/NKG2A receptor has a 6-fold greater affinity to HLA-E as compared with its activating counterpart, CD94/NKG2C, supporting the potential biologic importance of the level of HLA-E expression in tumors (33).

In ovarian cancer, high HLA-E expression was shown to abrogate the prognostic effect of CD8⁺ CTLs (4), and effect felt to be mediated by the inhibitory CD94/NKG2A complex expressed on intratumoral CTLs. The same group demonstrated that HLA-E expression in cervical cancers has been associated with a decreased risk of death on univariate analysis; although, this association was lost on multivariate analysis (4).However, upon histotype-specific analysis in cervical cancer, HLA-E expression was associated with improved overall survival for adenocarcinomas (34).

Although monalizumab failed to elicit clinical responses in women with recurrent gynecologic cancers, the biomarker studies on pretreatment samples revealed low overall expression of the therapeutic target receptor (CD94 IHC, as a marker of the CD94/NKG2 heterodimer) and presence of few NK cells (NKp46 IHC). We did not determine whether CD94 was associating with the inhibitory (CD94/NKG2A) or activating (CD94/NKG2C) receptor complex. Nonetheless, pretreatment stromal CD94, stromal Nkp46, and lymphocyte HLA-E expression were observed to be higher for patients with SD. Posttreatment, a drop in HLA-E expressing lymphocytes trended with SD. Based on the results of this study, although limited by small sample size and the disease heterogeneity, presence of stromal NK cells and target receptor expression (CD94) may improve patient selection for monalizumab therapy, but requires further study and validation. CD8 tumor and stromal lymphocyte expression were not associated with response.

An exploratory analysis of the correlation of biomarker expression levels in archival and pretreatment tumors demonstrated no relationship. Significantly, higher expression of tumor, endothelial, and lymphocyte HLA-E was seen in the archival samples. Expression of CD8 on tumor lymphocytes and tumor PD-L1 trended toward higher levels in archival samples. This observation supports the use of contemporaneous tumor biopsies for the study of biomarkers of immunotherapy response.

As a single agent, monalizumab had very little clinical activity in this patient population, although some long-term disease stabilization was observed. It is possible, however, that monalizumab could synergize with other immunomodulatory agents. For example, combined blockade of PD-L1 and NKG2A checkpoints in vivo enhanced antitumor CD8⁺ T-cell response in a murine model of cancer (35). Of interest, none of 12 pretreatment samples obtained for this trial had tumor PD-L1 expression, thus it is possible this therapeutic combination would need to be studied in an earlier treatment setting. Trials of monalizumab as cancer therapy are ongoing and include new combinations with PD-L1 inhibitors (NCT02671435), targeted therapies such as epidermal growth factor receptor inhibitors (NCT02643550), tyrosine kinase inhibitors (NCT02557516), and even as single-agent therapy in other, nongynecologic, solid tumors (NCT03833440).

In conclusion, single-agent monalizumab (IPH2201), a first-in-class mAb targeting the CD94/NKG2A and hence potentially abrogating the inhibitory effect induced by HLA-E binding, did not elicit clinical effect against recurrent, pretreated gynecologic cancers. The minimal toxicity observed with monalizumab supports further studies in combination with other agents. A rational approach to the use of monalizumab will require a deeper understanding of the immune regulatory environment of cells expressing CD94/NKG2A, hence biologic correlative studies remain a major research priority of this field.

A.V. Tinker reports receiving speakers bureau honoraria from and reports receiving commercial research grants from AstraZeneca. H.W. Hirte reports receiving speakers bureau honoraria from AstraZeneca, and is a consultant/advisory board member for AstraZeneca and Merck. M. Butler is an employee of GlaxoSmithKline; reports receiving speakers bureau honoraria from Novartis, Bristol-Myers Squibb, and Merck; is a consultant/advisory board member for Merck, Bristol-Myers Squibb, Novartis, Sanofi, EMD Serono, and GlaxoSmithKline; and reports receiving commercial research grants from Takara Bio, Inc., and Merck. H.A. Azim Jr has ownership interests (including patents) at Innate Pharma and is a consultant/advisory board member for Roche. L. Seymour reports receiving commercial research support from Innate in the form of funding to CCTG for trial support. No potential conflicts of interest were disclosed by the other authors.

Conception and design: A.V. Tinker, H. Ritter, D. Tu, L. Seymour

Development of methodology: A.V. Tinker, L. Seymour

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.V. Tinker, H.W. Hirte, D. Provencher, M. Butler, P. Paralejas, N. Grenier, S.-A. Hahn, J. Ramsahai, L. Seymour

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.V. Tinker, H.W. Hirte, D. Tu, L. Seymour

Writing, review, and/or revision of the manuscript: A.V. Tinker, H.W. Hirte, D. Provencher, M. Butler, D. Tu, H.A. Azim Jr, L. Seymour

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): H.W. Hirte, H.A. Azim Jr, L. Seymour

Study supervision: H. Hirte, L. Seymour

L. Seymour received funds from Innate Pharma on behalf of the Canadian Cancer Trials Group in support of this trial.

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.