Phase I Study of U3-1287, a Fully Human Anti-HER3 Monoclonal Antibody, in Patients with Advanced Solid Tumors

The EGF receptor (EGFR)/HER family of receptor tyrosine kinases, consisting of HER1 (EGFR), HER2, HER3, and HER4, plays an important role in oncogenesis. By homodimerizing and heterodimerizing, HER family members activate signaling pathways that are involved in cellular processes, including morphogenesis, proliferation, angiogenesis, and survival (1–4). A number of tumor types require EGFR or HER2 signaling for growth and survival (5–7), prompting the development of various therapeutic agents that target the EGFR and HER2 receptors. Current anti-EGFR and anti-HER2 agents, however, have proven to be effective only in a subset of patients, and initial response is frequently followed by relapse and the development of therapeutic resistance. Recent studies suggest that HER3 upregulation or reactivation is important to the development of resistance to EGFR- and HER2-targeted treatments (8–11).

Elevated HER3 expression has been detected in many tumor types and is associated with negative clinical prognosis, including shorter relapse-free survival time in breast cancer (12) and shorter overall survival times in lung, ovarian, and colon cancers (13–16). HER3 is also overexpressed and occasionally amplified in non–small cell lung cancer (NSCLC), and autocrine loops with its ligands have been documented (13, 17–20). Although lacking intrinsic kinase activity itself (21, 22), HER3 activates signaling pathways through heterodimerization with other HER family members. HER2:HER3 heterodimers show particularly potent mitogenic signaling compared with other HER family heterodimers (23). Targeting HER3, therefore, may negatively affect tumors and improve the efficacy of current anti-HER agents.

U3-1287 (NCT00730470) is a first-in-class, fully human monoclonal antibody directed against HER3. In vitro, U3-1287 binds to the extracellular domain of HER3 and promotes receptor internalization, inhibiting both basal and ligand-induced receptor signaling (24–26). Tumor cells treated with U3-1287 show reduced cellular migration, proliferation, and anchorage-independent growth. In addition, U3-1287 and anti-EGFR antibody treatment of EGFR- and HER3-expressing xenografts significantly reduced tumor growth compared with anti-EGFR antibodies alone or with an isotype immunoglobulin G (IgG) control in both EGFR wild-type models and resistant models with the T790M EGFR mutation (27).

A phase I, open-label, first-in-human study of single-agent U3-1287 was initiated to assess its safety, tolerability, and pharmacokinetic profile in patients with advanced solid tumors. The study was conducted in 2 parts. Part 1 was a dose-escalation, dose-finding study designed to identify the maximum-tolerated dose (MTD) by treating successive cohorts with escalating doses and evaluating the incidence of dose-limiting toxicities (DLT). Part 2 was a dose-expansion phase conducted in patients with advanced malignancies along with an exploratory NSCLC cohort that further evaluated the safety, tolerability, and efficacy of the possible dose levels identified from part 1 pharmacokinetic in the intended phase II population. This report describes results from both parts of this study.

For parts 1 and 2, eligible patients had a confirmed diagnosis of an advanced solid tumor that was refractory to standard treatment or for which no standard therapy existed and that was of a type known to express HER3. For part 2, the same eligibility criteria were used, but an exploratory cohort was included in which 15 or more of the 30 patients enrolled had to have NSCLC. Patients were not preselected on the basis of RAS or EGFR status. All patients were 18 years of age or more, had an Eastern Cooperative Oncology Group (ECOG) performance status of 0 to 2, and a life expectancy of more than 3 months. Additional inclusion criteria included controlled hypertension (diastolic ≤ 100 mmHg; systolic ≤ 140 mmHg), left ventricular ejection fraction 50% or more, cumulative anthracycline exposure 360 mg/m² or less, and adequate neutrophil count (≥1,500/μL), hemoglobin count (≥9 g/dL), platelet count (≥100,000/μL), calculated creatinine clearance level (modified Cockcroft–Gault equation; ≥60 mL/minute), bilirubin level (≤2 × upper limit of normal; ULN), prothrombin time or partial prothrombin time (≤1.5 × ULN), alkaline phosphatase level (≤2 × ULN with no liver or bone involvement or <5 × ULN with liver or bone involvement), and alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels (≤2.5 × ULN with no liver involvement or <5 × ULN with liver involvement). Patients were ineligible if they had previously received an anti-HER3–targeted antibody or if they had received small-molecule tyrosine kinase inhibitors within 2 weeks; major surgery, antibody therapy, retinoid therapy, hormonal therapy, or therapeutic or palliative radiotherapy within 4 weeks; or nitrosoureas or mitomycin C chemotherapies within 6 weeks of the first dose of U3-1287. Patients with hematopoietic malignancies, untreated or symptomatic brain metastases, or known HIV or active hepatitis B or C virus infection were ineligible. The study protocol was approved by participating Institutions' Research Ethics Boards, and each patient signed a consent form.

This was a multicenter study conducted in 2 parts. In part 1, dose escalation followed a modified 3 + 3 study design: 3 or 4 patients were enrolled initially in sequential cohorts of 0.3, 1, 3, 6, 9, 14, and 20 mg/kg. U3-1287 was administered as an intravenous infusion over 60 minutes. After the first intravenous administration of U3-1287 on day 1, a 21-day observation period occurred for each patient. If no DLTs were observed during the first 21 days, U3-1287 administration began on day 22 and was administered every 14 days. In part 2, the initial dose for dose expansion was determined on the basis of preliminary safety and pharmacokinetic through cohort 5 (9 mg/kg) from the dose-escalation phase of the study. The dose of 9 mg/kg was the first dose that produced a trough drug level 10-fold more than the trough level that corresponded to maximal preclinical efficacy. Escalating doses administered in part 2 of the study were 9, 14, and 20 mg/kg i.v. on days 1 and 15 of a 28-day cycle.

DLT was defined as drug-related grade 3 or higher hematologic or nonhematologic toxicity (except alopecia); for fatigue, only grade 3 fatigue that persisted for more than 7 days or grade 4 fatigue of any length was included. For patients with bone or liver metastases and baseline levels 2.5 × ULN or more (for alkaline phosphatase >2 × ULN), DLTs did not include elevations in AST, ALT, or alkaline phosphatase unless the following criteria were met: AST or ALT more than 8 × ULN (if baseline was 2.5–5 × ULN in patients with liver metastases); alkaline phosphatase > 8 × ULN (if baseline level was 2–5 × ULN in patients with bone or liver metastases). DLT assessment followed the guidelines provided in the Common Terminology Criteria for Adverse Events (CTCAE), v.3.0. MTD was defined as the highest dose with an observed incidence of DLT in less than 33% of the patients in the cohort.

In part 1 of the study, disease response assessment using modified Response Evaluation Criteria in Solid Tumors (RECIST; ref. 28) was conducted at screening, days 35 and 64, and every 8 weeks thereafter. In part 2 of the study, disease response assessment occurred at screening, days 42 and 70, and every 8 weeks thereafter. 2-deoxy-2-(¹⁸F)fluoro-D-glucose positron emission tomography/computed tomography scans were conducted after eligibility was determined and within 14 days before the first dose of study drug (baseline) and on day 42 in part 1 (cohorts 3 through 7) and part 2 (all cohorts), once dose levels that might be pharmacodynamically active had been achieved. Metabolic response was assessed as described by Young and colleagues (29).

All adverse events were recorded from the time of informed consent until 30 days after the last dose of U3-1287 and were graded according to CTCAE v.3.0. Scheduled 12-lead electrocardiograms were conducted during screening at days 1, 15, 29, and 57, and every 4 weeks thereafter. QTc interval was evaluated centrally and corrected for heart rate by the Fridericia formula (QTcF).

During part 1, pharmacokinetic evaluation was conducted on all patients on day 1 of treatment [blood samples taken at end of infusion (EOI), and at 3, 6, 24, and 72 hours after EOI], and on days 7, 14, and 21. Samples were also obtained immediately before the third dose (day 36), at EOI, at 3 and 6 hours after EOI, and on days 37, 39, and 42. Preinfusion and EOI pharmacokinetic samples were also taken on days 22, 50, and 64, and every 8 weeks thereafter. The pharmacokinetic parameters [area under the plasma concentration–time curve (AUC), maximum-observed serum concentration (C_max), and half-life (t_1/2)] were derived from serum concentration–time data following doses 1 and 3 during part 1 and then calculated by noncompartmental analysis using WinNonlinEnterprise v5.1.1 (Pharsight Corp.).

Blood samples for determining anti-U3-1287 antibodies in serum were collected in part 1 before dosing on days 1, 21, and 64, and every 8 weeks thereafter, as well as during an end-of-study visit. In part 2, blood samples were collected before dosing on days 1, 29, and 57, and then every 8 weeks and at the end-of-study visit.

HER3 gene amplification status and protein expression were assessed by FISH and immunohistochemistry (IHC), respectively, in archived, formalin-fixed, paraffin-embedded (FFPE) tumor tissues. For FISH, tissue sections were evaluated through incubation with the dual color probe set ZytoLight SPEC HER3/CEN 12 (Zytovision) and scored both by absolute copy number and amplification ratio. For IHC, tissue sections were incubated with mouse IgG antihuman HER3 (sc81455, Santa Cruz Biotechnology) and scored as a percentage of positive cells and on a semiquantitative scale using H-scores for total HER3 in the cytoplasm and membrane. Samples were analyzed by a single pathologist who was blinded to clinical outcomes.

Twenty-six patients were enrolled in part 1, and 31 patients were enrolled in part 2. Demographics and clinical characteristics are summarized in Table 1. The most common primary tumor sites were colorectal (50.9%) and NSCLC (35.1%); 17 of the 20 patients with NSCLC were specifically enrolled into the dose-expansion cohorts.

All patients had received chemotherapy, with a median of 5 previous regimens. Forty-one patients (71.9%) received prior EGFR-targeted therapy, of which 3 patients (7.3%) had partial response (PR) as the best response to prior EGFR-targeted treatment, 14 patients (34.1%) had the best response of stable disease, and 16 patients (39.0%) had progressive disease. The best response on prior EGFR-targeted therapy was unknown for 8 patients (19.5%).

All patients received at least 1 dose of U3-1287. Fifty-two patients (91.2%) withdrew from study because of disease progression, 2 patients (3.5%) because of full withdrawal of consent (unrelated to any study-related event), 2 patients (3.5%) because of nonfatal adverse events [grade 2 prolonged QTc interval (20 mg/kg) and grade 3 pulmonary embolism (20 mg/kg)], and 1 patient because of a fatal adverse event [grade 5 respiratory insufficiency (14 mg/kg)]; all were considered unrelated to U3-1287.

Most patients [55 (96.5%)] had at least 1 adverse event. Approximately half of the patients (47.3%) had adverse events with a highest grade of 1 or 2. The most common adverse events overall (≥20% of patients) included fatigue (42.1%), diarrhea (24.6%), nausea (22.8%), and dyspnea (21.1%; Table 2). On the basis of this limited sample set, it did not seem that any adverse event was dose dependent. U3-1287–related adverse events were reported in 26 patients (45.6%) and included fatigue (21.1%), diarrhea (12.3%), nausea (10.5%), decreased appetite (7.0%), and dysgeusia (5.3%). Nearly all drug-related events were grade 1 or 2; only 1 patient (3 mg/kg) had a drug-related adverse event that was grade ≥3 (a nonserious grade 3 event of hyponatremia on day 20 and a nonserious grade 3 event of hypophosphatemia on day 23). Both events were possibly related to study medication. There were no grade 4 or 5 U3-1287–related adverse events.

Overall, 6 patients experienced 7 grade ≥4 events; none were related to U3-1287. Nineteen patients (33.3%) had a serious adverse event; none were deemed related to U3-1287. Eight patients discontinued study treatment due to 10 adverse events, none of which were judged to be related to treatment with the study drug.

One patient had asymptomatic grade 2 QTcF prolongation and 3 female patients had asymptomatic QTcF between 450 and 470 ms (normal for females). Analysis of QTcF prolongation as a function of U3-1287 serum concentration showed no concentration dependence (see Supplementary Fig. S1).

No DLTs were reported for any of the cohorts during the DLT observation window. Thus, the MTD was not reached and is therefore not lower than the maximum dose administered in this study (20 mg/kg).

Pharmacokinetic parameters for dosing cohorts in part 1 are summarized in Table 3. The C_max and AUC of U3-1287 increased more than dose proportionally across the tested range (0.3–20 mg/kg). Pharmacokinetic approached linearity at the higher doses studied (6–20 mg/kg), indicating that saturation of clearance pathways occurs at doses more than 6 mg/kg. Mean serum t_1/2 was approximately 8 days at doses more than 6 mg/kg. Data from pharmacokinetic analysis of part 2 indicate that in patients treated with 9, 14, or 20 mg/kg every 2 weeks, steady state was achieved after approximately 3 doses. At dose levels 9 mg/kg or more, the mean minimum-observed serum concentration (C_min) beyond the second dose was at least 10-fold more than the C_min associated with maximal efficacy in mouse xenograft models (IC₉₀ ≈ 3 μg/mL; Fig. 1; ref. 30). Owing to this, as well as the simulation of the every 2 week and every 3 week dosing data, the pharmacokinetic profile of U3-1287 supports intravenous administration at or above 9 mg/kg every 2 weeks or every 3 weeks.

No patient developed human antihuman antibodies (HAHA) after dosing with U3-1287. One of the 26 patients tested during part 1 and 1 of the 31 patients tested during part 2 had positive HAHA results before and after exposure to U3-1287, as detected by immunoassay. No neutralizing antibodies were detected at any visit for either patient; therefore, the HAHAs detected were present before dosing and were non-neutralizing.

In the 51 patients evaluable for tumor response, stable disease at the earliest observation time point was the best single-evaluation response in 50.9% of patients. In part 1, stable disease was recorded in 1 or more evaluations in 53.8% of patients and maintained for at least 16 weeks in 7.7% of patients. In part 2, stable disease was recorded in at least 1 evaluation in 48.4% of patients and maintained for at least 16 weeks in 16.1% of patients (Fig. 2). Twenty-two patients (38.6%) had progressive disease. No patient had complete response or PR per modified RECIST. One patient with NSCLC (20 mg/kg) had tumor shrinkage (26.3% decrease) that approached but did not reach the 30% reduction threshold for PR (Fig. 3).

The overall metabolic response rate was 3.5%: the previously mentioned patient with NSCLC and 1 patient in part 1 (14 mg/kg) had partial metabolic responses. Both patients were nonsmokers who had received prior EGFR-targeted therapy. The first patient previously had the best response of stable disease while receiving erlotinib-based therapy and, in this study, had stable disease for 42 weeks. The second patient previously had progressive disease while receiving cetuximab-based therapy and, in this study, was removed for progressive disease after the first tumor assessment posttherapy; however, this patient had stable disease per central review that lasted 32 days. No clear dose response could be discerned given the limited number of responders. Overall in this study, 24.6% of the patients had stable disease, 40.4% had progressive disease, and 31.6% were not evaluable for metabolic tumor response.

Expression of HER3 was investigated at the gene copy number level by FISH and at the protein expression level by IHC in FFPE tissue sections. Patient subgroups defined using varying cut-off values for gene copy number or amplification ratios, or by protein expression levels, were evaluated for clinical outcome (progressive disease or stable disease) to identify possible correlations. Results show a trend toward higher proportions of patients with stable disease than patients with progressive disease belonging to the subgroups defined by the highest cut-off values for gene copy number or protein expression levels, but this exploratory result requires independent confirmation.

Phospho-HER3 was examined as a potential pharmacodynamic and tissue biomarker; however, research using tissue microarrays indicated that the phospho-HER3 epitope is not stable in archived tissue, and fresh tissue could not be examined because patients were not prospectively biopsied. Attempts to assay the HER3 ligand heregulin by either IHC or quantitative PCR are still ongoing.

This first-in-human study of U3-1287, a first-in-class HER3-targeted fully human monoclonal antibody, supports dosing regimens at a range of 9 to 20 mg/kg U3-1287 every 2 weeks or every 3 weeks. U3-1287 exhibited a favorable safety profile and preliminary evidence of antitumor activity.

The most frequently reported adverse events included fatigue, diarrhea, nausea, and dyspnea. This small patient cohort did not provide clear evidence of a dose-dependent relationship for toxicity. Most incidences of U3-1287–related events were mild to moderate (grades 1 or 2). There were no grade 4 or 5 adverse events related to U3-1287, and no treatment discontinuations were attributed to a U3-1287–related adverse event.

The development of neutralizing antibodies against therapeutic monoclonal antibodies in a patient can lead to the loss of efficacy or to more serious clinical effects. As U3-1287 is a fully human antibody, immunogenicity was not expected to be a concern, and consistent with this, no neutralizing anti-U3-1287 antibodies were detected in this patient population following exposure to U3-1287. Pharmacokinetic results and simulations from every 2 week dosing suggest that doses 9 mg/kg every 3 weeks and above achieve steady state trough levels 10-fold in excess of those corresponding to maximal efficacy in mice. Preclinical experiments showed a correlation between trough concentrations and pharmacodynamics, as assessed by reductions in phospho-HER3 levels (see Supplementary Fig. S2). Variation of treatment schedule around a constant dose suggested that trough levels are more important to pharmacodynamics than peak levels in the murine system. However, a number of factors affecting the ability of an antibody to penetrate solid tumors (e.g., variable intervessel distances and high interstitial pressure; ref. 31) may be less prominent in the smaller tumors of mice. Therefore, the results from murine studies cannot be directly translated into recommended dose levels for humans, and thus, 9 mg/kg every 3 weeks may be viewed as the minimal dose level needed to achieve a significantly higher concentration in humans than is required in murine systems. U3-1287 has showed an acceptable safety profile in doses up to 20 mg/kg every 2 weeks; therefore, it is recommended that doses of 9 to 20 mg/kg every 2 weeks or every 3 weeks be used for collecting further safety, efficacy, and biodistribution data. Ongoing clinical studies are using U3-1287 every 3 week dosing regimens for enhanced patient convenience. Although pharmacodynamic activity could not be effectively assessed in this study to confirm dose selection, 2 independent studies are in progress to refine dose selection: a radiolabeled biodistribution study and a dose-ranging phase I/II study in combination with erlotinib, another HER family inhibitor.

HER3 has been shown to be a major activator of phosphoinositide 3-kinase/Akt signaling in EGFR-addicted and HER2-amplified cancers (11, 32, 33). Previous studies have correlated elevated HER3 expression with poor prognosis in a number of tumor types (12–16). There was a trend in this patient population toward the association of clinical benefit with high levels of HER3 amplification or expression. However, due to the large number of exploratory comparisons, independent confirmation is required. This association was not sufficient to allow patient stratification and needs further examination in additional studies, including in more homogeneous populations and in the context of combination therapy. Total HER3 expression might also be considered for incorporation into a multicomponent classifier for patient selection in the future. Assays to look at pathway activation and presence of ligand, rather than simple receptor expression, are under development but remain a challenge in archived tissues.

Early evidence of clinical activity for U3-1287 in a selected population was suggested by several patients with prolonged stable disease and/or metabolic responses. It has been suggested that activated HER3 plays a role in the development of resistance to HER2 and EGFR inhibitors and that the use of a HER3 inhibitor, such as U3-1287, may prevent or overcome such resistance if used in combination with these agents. The safety and pharmacokinetic profile of U3-1287 monotherapy in this population favor future studies of this agent in combination regimens, particularly with anti-EGFR– and anti-HER2–based therapies used in colorectal, head and neck, and breast cancers (18, 34, 35). A phase I/II trial is ongoing to examine U3-1287 (9 or 18 mg/kg every 3 weeks) in combination with erlotinib in patients with NSCLC.

P.M. LoRusso has a commercial research grant from U3 Pharma. P.A. Janne is a consultant/advisory board member of Amgen, Boehringer Ingelheim, Roche, Pfizer, Astra Zeneca, Clovis Oncology, and Lab Corp. V.L. Keedy has other commercial research support from Amgen and Daiichi Sankyo Pharmaceutical. L. Yee is a consultant/advisory board member of Dendreon, Spectrum Pharmaceutical, and Sanofi Aventis. C. Copigneaux is employed (other than primary affiliation; e.g., consulting) as a director (clinical development) at Daiichi Sankyo. A. Ang has ownership interest (including patents) in Amgen Inc. R.A. Beckman is employed (other than primary affiliation; e.g., consulting) as Chief Scientific Officer of Onco-Mind, LLC. J. Berlin is a consultant/advisory board member of Amgen. No potential conflicts of interest were disclosed by the other authors.

Conception and design: P.M. LoRusso, M.R. de Oliveira, T. Hettmann, R.A. Beckman, D.M. Beaupre

Development of methodology: V.L. Keedy, T. Hettmann, A.-B. Halim, D.M. Beaupre

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P.M. LoRusso, P.A. Janne, M.R. de Oliveira, N.A. Rizvi, L. Malburg, V.L. Keedy, L. Yee, J. Berlin

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P.M. LoRusso, P.A. Janne, V.L. Keedy, L. Yee, T. Hettmann, C.-Y. Wu, A. Ang, A.-B. Halim, R.A. Beckman, J. Berlin

Writing, review, and/or revision of the manuscript: P.M. LoRusso, P.A. Janne, M.R. de Oliveira, N.A. Rizvi, L. Malburg, V.L. Keedy, L. Yee, C. Copigneaux, T. Hettmann, A. Ang, R.A. Beckman, D.M. Beaupre, J. Berlin

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P.M. LoRusso, M.R. de Oliveira, V.L. Keedy

Study supervision: V.L. Keedy, L. Yee, C. Copigneaux, T. Hettmann, R.A. Beckman, D.M. Beaupre

The authors thank Xiaoping Jin and Erik Rasmussen for statistical support, Emily Liu for clinical operational support, Alin Chen for pharmacokinetic calculations, Daniel Freeman and Jeanne Mendell-Harary for advice on pharmacokinetic and pharmacodynamic issues, and Cheryl Chun (BlueMomentum) for medical writing assistance.

This study was sponsored by Daiichi Sankyo Pharmaceutical Development and Amgen, Inc. Funding for medical writing support was provided by Daiichi Sankyo.

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.