A Phase I Study with Neratinib (HKI-272), an Irreversible Pan ErbB Receptor Tyrosine Kinase Inhibitor, in Patients with Solid Tumors Free

Novel therapeutic agents that target growth factor pathways dysregulated in tumors may offer alternative treatment options for patients who are chemo-intolerant or refractory to standard therapies. In particular, agents targeting members of the human epidermal growth factor receptor (EGFR) family (ErbB-1/EGFR and ErbB-2) have shown encouraging therapeutic efficacy. The first to be approved by the U.S. Food and Drug Administration in 1998 was trastuzumab, a humanized monoclonal antibody that binds to the extracellular domain of ErbB-2, for the treatment of ErbB-2–positive breast cancer. Later, small molecules that specifically inhibit EGFR tyrosine kinase were approved, including erlotinib and gefinitib, for advanced or metastatic non–small cell lung cancer (NSCLC; refs. 1, 2).

However, these targeted agents have limitations. Trastuzumab shows only 15% to 26% response as monotherapy and 38% response when used in combination with paclitaxel in metastatic breast cancer (3, 4)_. Trastuzumab treatment is also associated with heart failure in some patients, particularly those pretreated with anthracyclines (4, 5)_. Gefinitib and erlotinib have objective response rates of only 12% and 18%, respectively, in patients with advanced NSCLC; the benefit from these drugs occurs largely in the subgroup of patients whose cancers harbor activating EGFR kinase domain mutations (6, 7).

Preclinical data suggest that different tyrosine kinase inhibitors have different potencies against various EGFR and ErbB-2–activating mutations, and clinical data with reversible and irreversible tyrosine kinase inhibitors on patients with refractory solid tumors have been conflicting. Preclinical data with neratinib [HKI-272; a potent (8), orally administered, small-molecule, irreversible pan ErbB inhibitor that inhibits ErbB-1, -2, -4] suggest that this irreversible tyrosine kinase inhibitor can potentially overcome the acquired resistance of EGFR T790M mutation. This mutation developed in the tumors of lung cancer patients that have the EGFR kinase domain sensitizing mutation after initial response to either gefitinib or erlotinib treatment and subsequent progression (9–12).

Neratinib targets a conserved cysteine residue (Cys 797) within the catalytic cleft of the ErbB receptors and irreversibly inhibits its kinase activity (13). Neratinib is expected to inhibit tumors that express high (+3) and intermediate (+2) levels of ErbB-2. Tumors that express low levels of ErbB-2 are resistant to trastuzumab treatment.

The primary purpose of this study was to assess the tolerability and safety and to define the maximum tolerated dose of neratinib administered orally in patients with advanced-stage tumor types that express ErbB-2 or ErbB-1/EGFR. In addition, the pharmacokinetic profile and the antitumor activity of orally administered neratinib in patients with advanced stage ErbB-2 or ErbB-1/EGFR tumors were assessed.

Trial design. This was a phase I, open-label study wherein patients with ErbB-2– or ErbB-1/EGFR–positive tumors initially received single oral doses of neratinib, followed by 1 wk of observation to assess single-dose pharmacokinetic profiles and adverse events. Patients continued on the same dose level if neratinib was well tolerated with no evidence of progressive disease.

Patients. Eligible patients had to have previous histologic/cytologic diagnosis of metastatic or advanced-stage ErbB-2– or ErbB-1/EGFR–positive cancer that had failed standard effective therapy. Patients also had to have tumors for which +1, +2, or +3 levels of ErbB-2 or EGFR were documented by immunohistochemistry (Ventana). ErbB-2–positive tumors had staining intensities of ≥2 on a scale of 0 to 3 and percent positive tumor cells ≥10%; EGFR-positive tumors had staining intensities ≥1 with percent positive tumor cells ≥10% or any percentage of cells staining ≥2+. EGFR mutation analyses were done on the six NSCLC patients with stable disease using heteroduplex analyses combined with Surveyor endocnuclease digestion (14).

Patients had to have measurable disease (Response Evaluation Criteria In Solid Tumors), an Eastern Cooperative Oncology Group performance status of 0 to 2, and adequate bone marrow and organ functions.

The study protocol was approved by the institutional review boards of the participating institutions, and all patients gave written informed consent. This study was conducted according to the Declaration of Helsinki and its amendments.

Dose escalation. Dose escalation followed a modified Fibonacci scheme with three to six patients in each cohort. The planned dose-escalation schedule was 40, 80, 120, 180, 240, 320, 400, and 500 mg of neratinib (cohort 1-8) daily with food.

The decision for dose escalation was made after three to six patients in a cohort were evaluated for adverse events through day 14 of continuous daily administration. In previous studies with other tyrosine kinase inhibitors, gastrointestinal toxicity was reported soon after dosing began, and therefore, 14 d was selected as the safety assessment period for acute toxicities for dose escalation. Adverse events were graded based on the National Cancer Institute Common Terminology Criteria for Adverse Events version 3.0.

A dose-limiting toxicity was defined as any neratinib-related grade 2 diarrhea lasting >2 d on medical therapy or associated with fever or dehydration (maximum tolerated dose); nonhematologic grade 3 adverse event excluding grade 3 nausea, vomiting, or diarrhea (unless patients were receiving appropriate medical therapy); or any grade 4 adverse event. If none of three to six patients had a dose-limiting toxicity by day 14 of continuous daily administration, then dose escalation proceeded to the next level. If one of three to six patients had a dose-limiting toxicity, then the cohort was expanded to six required patients. If two or more of three to six patients in a cohort had a dose-limiting toxicity, dose escalation was stopped and the previous dose level was the maximum tolerated dose.

Dose reduction was recommended if a patient had a neratinib-related grade 3 or 4 adverse event, except in the case of nausea, vomiting, or rash wherein dose reduction occurred only if the patient was receiving appropriate medical therapy, and if a patient had grade 2 or 3 diarrhea lasting >2 d on medical therapy or associated with fever or dehydration. Doses were reduced by dropping back one dose level. No more than two dose reductions were allowed for any patient.

Up to 40 additional patients, including 10 patients with NSCLC who had progressed following ≥8 wks of treatment with either erlotinib or gefitinib and 10 patients with breast cancer, were treated at the maximum tolerated dose to better define the tolerability and preliminary activity of neratinib.

Evaluation of patients. All enrolled patients had radiologic tumor assessments done at screening and after every two cycles (i.e., cycles 2, 4, etc), according to the modified Response Evaluation Criteria In Solid Tumors guideline (15).

The primary efficacy parameter assessed in the trial was the best overall response, recorded from the start of treatment until disease progression/recurrence. Progression-free survival was also measured, and safety evaluations were based on the occurrence of adverse events. The efficacy evaluable population included all patients who received two or more consecutive weeks of treatment, and had one or more tumor assessment ∼8 wks after beginning continuous daily dosing or patients who received one or more dose of neratinib but discontinued the study because of early progressive disease, symptomatic deterioration, or death.

Pharmacokinetic analyses. Timed blood samples for pharmacokinetic analyses of neratinib were collected after the administration of the first single dose on day 1 and after 14 days of continuous daily administration (study day 21 ± 2 d). Samples were obtained at predose and at 1, 2, 3, 4, 5, 6, 8, and 24 h postdose on days 1 and 21, and a sample was also taken 48 h postdose on day 1 of the single-dose period.

Plasma concentrations were measured using a validated liquid chromatography/tandem mass spectrometry method. The bioanalytic method used 0.250 mL of plasma and was linear over the range of 3 to 250 ng/mL (lower limit of quantitation (LLQ) was 3 ng/mL). The mean interday variabilities (coefficient of variation) of neratinib quality control samples were ≤7.8%, and intraday variabilities were ≤8.8%. Mean interday accuracy was within the range of 98.4% to 103.2%, and mean intraday accuracy was within the range of 100.0% to 106.6%. Mean accuracy at the LLQ of 3 ng/mL (104.0%) indicated that variability was acceptable (8.8%). No interferences were observed in blank plasma or plasma spiked with internal standard. The plasma samples were stored at −70°C until analysis.

Pharmacokinetic analyses were done using WinNonLin Enterprise application version 4.1 (Pharsight Corporation). Standard noncompartmental method (16) was used to calculate the peak plasma concentration (C_max), time to C_max (t_max), the total area under the concentration-time curve (AUC = AUC_0-∞), the area under the concentration-time curve to the 24 h postdose (AUC_{0-24 hr}) or at steady state (AUC_ss), and the elimination half-life (t_1/2).

Patients. Seventy-three patients were enrolled from November 2003 until December 2005. The last patient completed the study in January 2007. One patient in the expanded maximum tolerated dose cohort was not treated because she withdrew consent. Of the 72 patients who were treated with neratinib, 72% were female with a median age of 57 years (Table 1). The predominant primary cancer diagnoses were breast cancer (40%) and NSCLC (21%). All had previous chemotherapy. Most of the patients were heavily pretreated, with 34% patients having four or more previous cytotoxic therapy regimens in the metastatic setting.

Dose escalation of neratinib. Diarrhea was the primary dose-limiting toxicity in this study. Patients at the 40- to 120-mg doses had no dose-limiting toxicities; one patient at the 180-mg dose had grade 3 diarrhea. Four patients at the 400-mg dose had grade 3 diarrhea, and per protocol, the maximum tolerated dose was determined to be 320 mg. The 320-mg cohort was expanded to include an additional 39 patients to confirm the safety and tolerability of the maximum tolerated dose.

Safety. All 72 (100%) patients experienced adverse events. Those, of any grade, that occurred during the study in ≥10% patients are summarized in Table 2. The most common adverse events were diarrhea (88%), nausea (64%), fatigue (63%), vomiting (50%), and anorexia (40%). Neratinib-related adverse events were similar.

Grade 3 or higher neratinib-related adverse events occurred in 39% of all patients. The most common grade 3 or higher related adverse events were diarrhea (32%), fatigue (4%), and vomiting (4%; Table 3).

The median onset of diarrhea was 8.5 days (range, 1.0-22.0 days). One patient, in the 320-mg cohort, on day 338 of the study, experienced grade 3 pneumonitis considered related to neratinib. This patient was hospitalized and withdrew from the study but eventually recovered.

All patients discontinued treatment. Discontinuations were due mainly to adverse events 13 (18%), disease progression 40 (56%), and symptomatic deterioration 7 (10%). The most common adverse events leading to treatment discontinuation were diarrhea in 10 patients (14%) and fatigue in 2 (3%). Nine (9) of 10 patients discontinued due to diarrhea during the first cycle and the tenth did so during cycle 4. A total of 15 patients died during the study; 14 of the deaths (19%) were within 30 days of the last dose of neratinib, and all of deaths were attributed to progressive disease. A total of 22 (31%) patients had dose reductions; 18 patients had one reduction, and four patients had two reductions. Diarrhea was the cause of the dose reduction in 19 of these patients. Fourteen (36%) patients had dose reductions at the maximum tolerated dose of 320 mg, and, therefore, 240 mg neratinib was designated the therapeutic dose.

Pharmacokinetics. Neratinib pharmacokinetic parameters are summarized in Table 4. Following treatment with single doses ranging from 40 to 400 mg, absorption of neratinib was relatively slow, with a median t_max of 3 to 6.5 hours. On study day 1, following single oral doses of 40 to 400 mg, neratinib exposure increased in a dose-dependent manner. For the same respective doses on study day 21 at steady state, the mean C_max ranged from 5.8 to 119 ng/mL, and the mean AUC_ss ranged from 76.0 to 1,704 ng·h/mL. Interpatient variability (coefficient of variation) estimates for neratinib C_max and AUC were small to moderate. In general, the steady state C_max and AUC of neratinib increased with increasing dose but in a nonlinear fashion because of a plateau between 320- to 400-mg doses. The mean accumulation ratio, R (AUC_ss/AUC_{0-24 hr}), was about one- to two-fold when comparing the multiple dose exposures to the single dose exposures (40-400 mg). The R value was 1.14 following daily administration of 240 mg neratinib, indicating no major accumulation of neratinib after repeated daily administration in patients with cancer at the therapeutic dose of 240 mg. The mean elimination t_1/2 on day 1 following oral administration of therapeutic dose of 240 mg neratinib with food was 14 h, and the half-life supports once-a-day dosing regimen.

Antitumor activity. A total of 60 patients were considered evaluable for efficacy, 25 patients with breast cancer and 14 patients with NSCLC. Among 25 evaluable patients with breast cancer, partial response was observed in 8 (32%) and one patient showed stable disease ≥24 weeks (Table 5). All responders had previous treatment with trastuzumab, anthracyclines, and taxanes, and tumors with a baseline ErbB-2 immunohistochemical staining intensity of 2+ or 3+ (seven of eight responders had a 3+ score). Among 14 evaluable patients with NSCLC, 6 (43%) showed stable disease ≥24 weeks. Among the six patients with stable disease ≥24 weeks, five patients had tumors with EGFR staining intensity of 3 and all had failed at least one regimen of an EGFR inhibitor (gefitinib or erlotinib). The six patients with stable disease had the following EGFR kinase domain mutations: del L747-750del, insP; L858R; E746-P753delinsVS; L747-S752del, P753S; del L747-751C; and del L747-S752, T790M.

Maximum percentage decrease from baseline in tumor size (sum of measured lesions) was evaluated separately for all patients. Thirteen patients with breast cancer and five patients with NSCLC showed reductions from baseline in tumor size (Fig. 1).

The median duration of response was 4.8 months [95% confidence interval (95% CI): 1.9, 9.5] in the eight breast cancer patients with partial response. The median duration of stable disease was 5.8 months (95% CI: 3.7, 8.1) in patients with breast cancer (n = 14) and 9.0 months (95% CI: 7.4, 9.2) in patients with lung cancer (n = 8). The median progression-free survival was 3.6 months (95% CI: 1.7, 5.6) in patients with breast cancer and 3.5 months (95% CI: 1.2, 9.0) in patients with lung cancer.

In this phase I study, neratinib was orally administered to patients with solid tumors. No dose-limiting toxicities were observed in doses ranging from 40 to 120 mg; but one patient in the 180-mg dose group and four patients in the 400-mg dose group experienced grade 3 diarrhea. Hence, the maximum tolerated dose was determined to be 320 mg. This cohort was expanded to include an additional 39 patients for evaluation of safety and tolerability. Neratinib was found to have an acceptable safety profile. The most common related adverse events were gastrointestinal in nature, mainly diarrhea. However, diarrhea was managed by use of antidiarrheal agents and dose reductions. Fourteen percent of patients had to discontinue treatment because of diarrhea.

The frequency and severity of rash seem to be significantly less than with erlotinib (17) or with gefitinib (18). One patient had grade 3 pneumonitis and discontinued the study but eventually recovered. Pharmacokinetic analyses showed that the steady state C_max and AUC of neratinib increased in a dose-dependent manner from 40 to 320 mg, but there was no further increase in exposure when the dose was increased from 320 to 400 mg. This could be due to low solubility of neratinib, and the saturable low solubility of the drug might be the main contributor to the absorption plateau at the high doses. Eight patients with metastatic breast cancer had a partial response. Six patients were in the 320-mg dose group with AUCs ranging from 532 to 2,752 ng·h/mL; one patient was in the 180-mg dose group with an AUC of 946 ng·h/mL; and one patient was in the 120-mg dose group with an AUC of 713 ng·h/mL. In the nonclinical efficacy model in nude mice, the exposure at the minimum efficacious dose was 431 ng·h/mL. Partial responses in all eight patients occurred at or above minimum efficacious dose exposure, and the mean steady-state exposure at the therapeutic dose of 240 mg was ∼2.2-fold higher than the minimum efficacious dose exposure in nude mice. Diarrhea seems, to some extent, related to neratinib dose and exposure. However, because of the small number of patients in this study, no clear relationships between the dose or exposure and the severity of major adverse events (i.e., diarrhea, nausea, or rash) were observed.

The efficacy results were very promising in breast cancer patients. Partial response was observed in 8 (32%) patients with breast cancer. Stable disease ≥24 weeks was observed in one breast cancer patient and six NSCLC patients.

Preliminary results from a phase II trial comprised of breast cancer patients with tumors positive for ErbB-2, who either had or had not undergone previous trastuzumab therapy, support that neratinib is efficacious. Among trastuzumab-experienced patients, 59% had received one previous trastuzumab-containing regimen, 34% received 2 to 3 regimens, and 6% received 4 regimens. In this phase II study, patients with previous trastuzumab therapy had objective response rates of 33% and patients with no previous trastuzumab had objective response rates of 49%. Corresponding 16-week progression-free survival rates of 53% (95% CI: 40, 65) and 72% (95% CI: 56, 83) were observed in the respective treatment groups (19). By comparison, objective responses of 15% to 26% were observed when patients with metastatic breast cancer were treated with monotherapeutic trastuzumab. Another dual ErbB-2/EGFR inhibitor, lapatinib, showed only 4.3% response as monotherapy in metastatic breast cancer (20).

The lack of partial responses for NSCLC patients was disappointing, but 43% showed stable disease ≥24 weeks. This lack of association of EGFR expression and clinical outcome is not unexpected. Another irreversible ErbB pan inhibitor CI-1033 showed marginal efficacy in patients with advanced NSCLC with a response rate of 2% to 4% and failed to meets its primary statistical end point (21). In addition, others have shown that protein expression is a poor predictor for response, time to progression, or overall survival (22, 23). Rather, studies with gefitinib and erlotinib have indicated that clinical response is correlated to specific somatic mutations in the tyrosine kinase domain of the EGFR gene, and indeed, disease progression can occur with the development of a secondary erlotinib/gefitinib acquired resistance mutation, T790M. It is interesting to note that all six NSCLC patients with stable disease had clinical responses to gefitinib/erlotinib before progression and enrollment into the neratinib trial, and their lung cancers had documented activating EGFR kinase domain mutations. Thus, neratinib may be beneficial to this subset of lung cancer patients.

The tolerable toxicity profile, pharmacokinetic properties, and encouraging antitumor activity in patients with advanced solid tumors warrant further evaluation of neratinib for the treatment of patients with solid tumors.

J. Vermette, R. Abbas, S. Quinn, C. Powell, and C. Zacharchuk are employed by Wyeth. T. Lynch is a member of the speakers' bureau for Wyeth. P. Munster has a patent on EGFR mutation testing. P. Fracasso and T. Lynch are consultants for Wyeth. T. Lynch is a consultant for Genentech, Bristol-Myers Squibb, Boehringer-Ingelheim, and Exelixis.

We thank the patients, their families, and the clinical personnel who participated in this study; and Tricia Gooljarsingh and Susan Leinbach for the assistance with article preparation.