This phase I, open-label, multicenter, multiple-dose study was approved by the institutional review boards at all nine centers in the US. Patients were enrolled between May 26, 2015 and February 7, 2017 (study completion October 10, 2017). The study was conducted in compliance with the ethical principles originating in or derived from the Declaration of Helsinki and in compliance with the International Council for Harmonization Good Clinical Practice Guidelines. All patients provided written informed consent. The study is registered on ClinicalTrials.gov (NCT02444793).

The primary objective of the study was to estimate the maximum tolerated dose (MTD) of utomilumab in combination with mogamulizumab in patients with advanced solid tumors. Secondary objectives included assessment of the safety profile, PK, immunogenicity, and antitumor activity of the combination. Exploratory objectives included the pharmacodynamic effect on immune parameters in blood.

Refractory patients had a previously documented best overall response (BOR) of non-complete response (CR)/partial response (PR)/stable disease (SD) on PD-1/PD-L1 treatment (includes progressive disease and clinical deterioration); relapsed patients had documented BOR of CR/PR/SD but later progressed on PD-1/PD-L1 treatment (includes progressive disease and clinical deterioration).

Patients were not eligible if they had a history of autoimmune disease; systemic anticancer therapy within 28 days prior to registration; radiation therapy within 14 days prior to registration; therapeutic or experimental mAbs within 28 days prior to registration; active and clinically significant bacterial, fungal, or viral infection; live vaccine within 30 days prior to registration; or systemic corticosteroid therapy or any other form of immunosuppressive therapy within 14 days prior to registration.

The starting dose for intravenous utomilumab was 1.2 mg/kg every 4 weeks, with escalation to 2.4 mg/kg and 5 mg/kg in the subsequent cohorts following the time-to-event continual reassessment method (TITE-CRM). A flat dosing of 100 mg utomilumab was also assessed. Intravenous mogamulizumab 1 mg/kg was administered weekly for 4 consecutive weeks and biweekly thereafter, following utomilumab dosing. Treatment with study drugs was to continue until the first occurrence of one of the following: completion of 24 months of treatment, disease progression, patient refusal to continue, unacceptable toxicity, or study termination by the sponsor.

In all, 24 patients received mogamulizumab 1 mg/kg in combination with utomilumab dosed as follows: 1.2 mg/kg (n = 11), 2.4 mg/kg (n = 4), 5 mg/kg (n = 3), and 100-mg flat dose (n = 6). Most patients were male (79.2%) and white (79.2%). The mean (range) age was 63.9 (53–75) years. There were 11 patients with SCCHN, 10 with NSCLC (n = 7 squamous and 3 adenocarcinoma), 2 patients with colorectal cancer, and 1 patient with ovarian cancer, assessed by Response Evaluation Criteria in Solid Tumors (RECIST 1.1). The majority (91.7%) of patients had received at least 2 lines of anticancer drug therapy (Table 1 ). Median (range) duration of treatment was 16 (4.0–41.3) weeks. All of the patients with squamous NSCLC (n = 7), 1 with lung adenocarcinoma, and 7 with SCCHN were relapsed or refractory to anti–PD-1/PD-L1 checkpoint inhibitor therapy. Nine (37.5%) and 15 (62.5%) patients had baseline Eastern Cooperative Oncology Group performance status 0 and 1, respectively. The dose-expansion phase of the study was not initiated due to marginal efficacy.Tab1

Primary diagnosis and prior anti-cancer treatment

No DLTs were observed at any utomilumab dose (1.2 mg/kg, 2.4 mg/kg, 5 mg/kg, 100 mg flat dose) in combination with mogamulizumab 1 mg/kg. Although no DLTs were observed up to 5 mg/kg, the estimated recommended Phase II dose was at least 2.4 mg/kg per the TITE-CRM method; as the 5 mg/kg cohort only enrolled 3 patients, this dose was not fully explored in this respect. The most common (in ≥25% of patients), all-causality AEs were fatigue (45.8%), rash (29.2%), and diarrhea (25.0%), all of grade 1 or grade 2 severity. Eight (33.3%) patients experienced all-causality grade 3–4 AEs. Ten (41.7%) patients experienced serious AEs (SAEs), all determined to be unrelated to utomilumab or mogamulizumab; AE causality was initially assessed by the site Principal Investigator and all SAEs were adjudicated during regular conferences involving all sites and sponsor. The majority of the treatment-related AEs were grade 1 or 2, and none were grade 4 or 5. Two (8.3%) patients in the utomilumab 100 mg/mogamulizumab 1 mg/kg treatment group experienced three grade 3 AEs determined to be related to treatment: pneumonitis (utomilumab-related), hypophosphatemia (mogamulizumab-related), and anemia (both treatments). Three (12.5%) patients experienced grade 5 AEs, determined to be unrelated to either treatment. Of these patients, 2 occurred within 30 days after the last dose of study treatment and were due to malignant neoplasm progression/disease progression. The third patient died due to sepsis during the follow-up period (within 60 days after the last dose of study treatment).

Five patients had sufficient data to calculate PK parameters at Cycle 5. Utomilumab systemic exposure based on area under the serum concentration–time curve to last measureable dose and maximum serum concentration values appeared to increase with increasing dose. Due to the low number of patients, the relationship between serum PK parameters and dose could not be fully determined (Table 2 ).Tab2

Descriptive summary of serum utomilumab pharmacokinetic parameter values for Cycle 5

Following co-administration with utomilumab, mogamulizumab PK was similar across groups, with accumulation observed following multiple-dose administration (Additional file 1 and Additional file 2).

Thirteen of 24 (54.2%) patients that received utomilumab tested positive for treatment-induced ADA and 11 of 24 (45.8%) were positive for NAb. One (4.2%) patient who received mogamulizumab was confirmed positive for treatment-induced ADA; no one in this group tested positive for NAb. The median onset for treatment-induced ADA against utomilumab was 56.01 days (interquartile range [Q1, Q3]: 55.88, 56.13) and the median onset for NAb was 56.13 days (interquartile range [Q1, Q3]: 55.96, 56.97). The median duration of ADA and NAb was 0 and 0 days, respectively (interquartile range [Q1, Q3]: 0, 55.95 for ADA and 0, 62.84 for NAb). There was no substantial impact of ADA/NAb on PK and safety when utomilumab and mogamulizumab were administered in combination.

Patients treated with combination utomilumab/mogamulizumab showed a transient reduction in circulating T cells at 6 h (p < 0.0001) (Fig. 1a), possibly due to cytokine-induced adhesion of T cells to endothelial cells [34]. There was not a significant (p < 0.05) relationship between combination dose, T-cell reduction, and cytokine levels observed in this study (data on file; Pfizer). Levels of circulating CD8⁺ T cells largely returned to baseline by 7 days (168 h) after start of dosing; however, circulating CD4⁺ T cells did not fully recover by 336 h. Statistically significant (p < 0.001) decreases in Tregs (CD3 + CD4 + CD25 + CD127^low/−) were observed at 6, 168, and 336 h (Fig. 1b), as predicted based on data from a phase I study of single-agent mogamulizumab in patients with CCR4⁻ lung and esophageal cancers [25]. Reductions in effector memory (CD45RA⁻CCR7⁻) and central memory (CD45RA⁻CCR7⁺) CD4 T cells were also observed at the same time points (p < 0.01). Naïve (CD45RA⁺CCR7⁺) CD4⁺ T cells were less strongly affected. Within the CD8⁺ T-cell compartment, the central memory subpopulation was reduced, relative to baseline, at 6, 168, and 336 h (p < 0.005) to a greater degree than either the naïve or effector memory compartments (Fig. 1c). These results are largely concordant with earlier findings [25] and consistent with observations that CCR4 is expressed by central memory T cells [35, 36].Fig1Fig. 1

Fold-changes relative to baseline. Fold-changes are shown by lymphocyte populations in peripheral blood following treatment with utomilumab and mogamulizumab. Results were aggregated across all utomilumab doses, as statistically significant differences between utomilumab doses were not observed. a Major T-cell subpopulations relative to all white blood cells; b Treg and other major CD4⁺ T- cell populations; and (c) major CD8⁺ T-cell populations. White blood cells were defined by forward and side light scatter. T cells were defined by co-expression of CD3, CD4, and CD8. Naïve, central memory, and effector memory T-cell subpopulations were defined as CD45RA⁺CCR7⁺, CD45RA⁻CCR7⁺, and CD45RA⁻CCR7⁻, respectively [32]. Tregs were defined as CD3⁺CD4⁺CD25⁺CD127^low/ −[33]. Treg, regulatory T cell

Baseline tumor biopsies from patients with NSCLC (n = 1) and SCCHN (n = 4) were analyzed for the presence of PD-L1, CD8, FoxP3, and 4-1BB/CD137 (Table 3 ). All biopsies were negative for PD-L1 expression on tumor cells, except for one SCCHN specimen that had 10% PD-L1⁺ tumor cells. The biopsies had low levels of infiltrating CD8⁺ (range, 1–17%) and FoxP3⁺ (range, 1–9%) cells, with CD8/FoxP3 ratios ranging from 1 to 9. Also, 4-1BB/CD137 was observed on small percentages (9, 10, and 14%) of cells in the IM. The limited number of available specimens precludes generalizations about the cohort.Tab3

Immunohistochemistry analysis of whole tumor and IM of pretreatment tumor biopsies

The ORR was 4.2% (95% CI: 0.1–21.1%). Best percent change from baseline in sum of longest diameters (SLD) for target lesions is shown in the waterfall plot (Fig. 3a). The spider plot (Fig. 3b) shows percentage change from baseline in sum of SLD for target lesions over time. One patient in the utomilumab 100 mg/mogamulizumab 1 mg/kg group with PD-1 refractory squamous NSCLC achieved PR, which occurred at the first tumor assessment with a duration of response of approximately 2 months.Fig3Fig. 3

Antitumor efficacy. A BOR of partial response was observed in 1 patient with PD-1–refractory squamous NSCLC. a Waterfall plot of best percent change from baseline in the SLD for target lesions, full analysis set, with BOR indicated for each patient. b Spider plot of percent change from baseline in SLD for target lesions over time, full analysis set. c Baseline and post-treatment scan images. The baseline scan was taken ~ 4 weeks prior to treatment initiation. The post-treatment scan was taken at ~ 7 weeks after treatment initiation. Arrows point to tumor location in right lower lobe of the lung. CRC, colorectal cancer; NSCLC, non–small-cell lung cancer; PD-1, programmed cell death 1; SCCHN, squamous cell cancer of the head and neck; SLD, sum of the longest diameter