FormalPara Key Points

This was the first-in-human study of BI 853520 in patients with nonhematologic malignancies.

A maximum tolerated dose of 200 mg once daily was identified; dose-limiting toxicities included grade 3 proteinuria and fatigue.

Preliminary pharmacokinetic findings support once-daily dosing, which was associated with a manageable safety profile and some evidence of antitumor activity.

1 Introduction

Cellular interactions with the extracellular matrix play a pivotal role in tumor initiation, progression, and metastasis [1]. Focal adhesion kinase [FAK; also known as protein tyrosine kinase 2 (PTK2)] is a widely expressed, nonreceptor, cytoplasmic tyrosine kinase [1,2,3] that is a key component of the focal adhesion complex, which plays an important role in integrating signals from integrins and growth factor receptors in both normal and cancer cells [1, 4], among other functions. A role for FAK in cancer development is suggested by its amplification, overexpression, and/or activation in a variety of human malignancies [1, 3, 5], including breast [6, 7], colorectal [7, 8], ovarian [9], prostate [10], and sarcoma [7, 11]. Research suggests that FAK signaling is involved in cellular proliferation, migration, angiogenesis, invasion, survival, and epithelial-mesenchymal transition, via a variety of different pathways [4]. In addition, FAK appears to play a role in tumor metabolism, and can promote glucose consumption, lipogenesis, and glutamine dependency to promote cancer cell proliferation, motility and survival [12]. FAK has been shown to be a central driver of the fibrotic and immunosuppressive tumor microenvironment that protects pancreatic ductal adenocarcinomas from immune surveillance [13]. Furthermore, through its effects on chemokine transcription in the cell nucleus, FAK can promote antitumor immune evasion [14].

Given its putative role in tumor development and progression, FAK has become a logical therapeutic target, and numerous attempts have been made to inhibit the oncogenic activity of FAK [4]. Preclinical studies have demonstrated antitumor activity with FAK blockade [15,16,17,18,19,20], and several orally bioavailable, adenosine triphosphate (ATP)-competitive, small-molecule inhibitors of FAK are undergoing evaluation in early-phase clinical trials in patients with cancer, including VS-6063, GSK2256098, and PF-00562271 [21,22,23,24].

BI 853520 is a novel, potent, highly selective, ATP-competitive inhibitor of FAK that has demonstrated activity in a variety of preclinical human tumor xenograft models [25]. In vitro, BI 853520 inhibited recombinant FAK with a half maximal (50%) inhibitory concentration (IC50) of 1 nM [26], which is largely comparable to the IC50 values reported for VS-6063 (0.6 nM) [23], PF-00562271 (1.5 nM) [18], and GSK2256098 (2–15 nM) [27]. BI 853520 has also demonstrated high selectivity; IC50 for the related kinase proline-rich tyrosine kinase 2, PYK2, was > 50,000 nmol/L [26]. The antitumor activity of BI 853520 was found to vary widely across a diverse panel of 16 murine subcutaneous adenocarcinoma xenograft models, from complete tumor inhibition to an absolute lack of sensitivity [26]. Biomarker analysis suggests that the in vivo efficacy of BI 853520 in these models is linked to a mesenchymal tumor phenotype characterized by low E-cadherin messenger RNA (mRNA) and protein levels, and by low expression of the microRNA hsa-miR-200c-3p, an epithelial-specific microRNA that promotes E-cadherin expression [26]. All xenograft models that were highly sensitive to BI 853520, including kidney, lung, ovary, pancreas and prostate adenocarcinomas, were found to lack E-cadherin expression or to express low levels of E-cadherin [26]. Conversely, of five BI 853520-resistant models, three were E-cadherin positive. In murine breast cancer models, BI 853520 most effectively prevented the establishment of metastases in tumors in which E-cadherin was either deficient or downregulated [28].

We report here the first-in-human phase I study of BI 853520, which comprised a dose-escalation phase followed by an expansion phase in patients with advanced nonhematologic cancers.

2 Methods

2.1 Patients

Patients aged ≥ 18 years with a confirmed diagnosis of advanced, measurable or evaluable, nonresectable and/or metastatic nonhematologic malignancy that was progressive within 6 months prior to study entry, as demonstrated by serial imaging, and refractory to standard therapy, or for which no effective standard treatment was available, were eligible for enrollment. Recruitment to the expansion phase was restricted to patients with metastatic pancreatic adenocarcinoma, metastatic esophageal carcinoma, metastatic soft tissue sarcoma, or metastatic platinum-resistant ovarian carcinoma. This selection was based on preclinical efficacy and prevalence data on E-cadherin loss in these tumor types. The complete inclusion/exclusion criteria are listed in the electronic supplementary material (ESM).

The trial was approved by the Institutional Review Board at each participating institution, and conducted in accordance with the Declaration of Helsinki, Good Clinical Practice guidelines, and applicable regulatory requirements (including International Conference on Harmonization guidelines). The Centrale Commissie Mensgebonden Onderzoek, Den Haag, The Netherlands approved the trial, and all patients provided written informed consent.

2.2 Study Design and Treatment

This open-label, phase I trial was conducted at five sites in The Netherlands and Canada between July 2011 and December 2015 (NCT01335269). The primary objective was to evaluate the safety and tolerability of BI 853520, determining the maximum tolerated dose (MTD), and identifying a recommended dose for further clinical development. Secondary objectives included pharmacokinetics (PK), pharmacodynamics (PD), assessment of predictive biomarkers, and efficacy.

Patients received oral BI 853520 in a continuous once daily (QD) dosing schedule in 28-day cycles. In the dose-escalation phase, sequential cohorts of three to six patients received escalating doses of BI 853520 (starting dose of 10 mg) in a standard 3 + 3 design. MTD was defined based on the occurrence of dose-limiting toxicities (DLTs) in the dose-escalation phase during the first cycle only. The MTD was defined as the dose of BI 853520 that was one dose level below the dose at which two or more of six patients experienced a DLT. Following determination of the MTD, additional patients were enrolled into four expansion cohorts with selected tumor types. Criteria for DLTs, dose reductions, and discontinuations are listed in the ESM.

2.3 Safety

The primary endpoint was the MTD of BI 853520. Safety was assessed by monitoring adverse events (AEs; National Cancer Institute Common Terminology Criteria for Adverse Events [CTCAE] version 4.03 [29]) and clinical laboratory parameters. Patients were included in the safety analysis if they had taken at least one dose of trial medication, and were evaluable for DLT if they had been observed for one treatment cycle and had undergone all safety assessments.

2.4 Pharmacokinetics

The PK profile of BI 853520 in plasma and urine was determined in all eligible patients (both the escalation and expansion phase) after a single oral dose and after repeat dosing (steady state). Full details of PK methods are listed in the ESM.

2.5 Pharmacodynamics and Analysis of E-Cadherin Loss as a Potential Predictive Biomarker

Levels of phosphorylated and total FAK in tumor biopsies were determined by enzyme-linked immunosorbent assay (ELISA), and E-cadherin expression was assessed by immunohistochemistry. Full details of the PD methods are listed in the ESM.

2.6 Efficacy

Patients were evaluable if they had at least one tumor assessment after baseline; the first was planned after 8 weeks of treatment, at the end of cycle 2. The following efficacy endpoints were evaluated: investigator-assessed objective response rate and disease control rate [complete response, partial response, or stable disease (SD)] according to Response Evaluation Criteria in Solid Tumors (RECIST) v1.1 [30], duration of disease control, and tumor shrinkage. Tumors were assessed by computed tomography/magnetic resonance imaging (CT/MRI) scans at screening (baseline), and at the end of every two cycles until treatment discontinuation.

2.7 Statistical Analyses

All analyses in this trial were exploratory and data were analyzed using descriptive statistics. The overall analysis of safety and efficacy was based on patients who had received at least one dose of BI 853520 (treated set).

3 Results

3.1 Dose-Escalation Phase

A total of 41 patients provided informed consent and 33 patients were treated with BI 853520 in the dose-escalation phase. Among the treated patients, 58% were female, median age was 60 years (range 33–82), and 9%/91% had an Eastern Cooperative Oncology Group performance status (ECOG PS) score of 0/1. Patients had various advanced solid tumors, including colorectal cancer (n = 10, 30%) and soft tissue or bone sarcoma (n = 4, 12%). Full baseline demographics for the dose-escalation phase are listed in Electronic Supplementary Table 1.

Patients were treated at the following dose levels: 10 mg QD (n = 3), 25 mg QD (n = 3), 50 mg QD (n = 3), 100 mg QD (n = 6), 200 mg QD (n = 11), and 300 mg QD (n = 7). Median duration of treatment among patients in the dose-escalation phase was 35 days (range 6–281). Twenty-five patients were evaluable for determination of the MTD.

Three patients experienced DLT during cycle 1: two of four evaluable patients in the 300-mg QD cohort (grade 3 proteinuria and grade 3 renal disorder; grade 3 fatigue), and one of nine evaluable patients in the 200-mg QD cohort (grade 3 proteinuria). The MTD of BI 853520 was therefore determined to be 200 mg QD.

Thirty-two (97%) patients experienced drug-related AEs, the most frequent of which were nausea (64%), fatigue (46%), diarrhea (36%), vomiting (36%), and decreased appetite (21%). Drug-related AEs were all CTCAE grade 1 or 2, with the exception of the DLTs mentioned above. Except for proteinuria (15% grade 1–2, 6% grade 3), no notable findings for laboratory assessments and vital signs were observed.

Response data in the escalation phase are described in Table 1.

Table 1 Confirmed best overall tumor response during the dose-escalation phase (treated set)
Full size table

3.2 Expansion Phase

3.2.1 Patients

Sixty-three patients were included in the expansion phase. Baseline characteristics by tumor type are summarized in Table 2. All eligible patients had a comprehensive history of previous anticancer therapies; 23 (37%) patients had received three or more prior systemic therapies.

Table 2 Patient demographics and baseline characteristics (expansion cohort; 200 mg QD)
Full size table

3.2.2 Treatment Exposure

All 63 patients received at least one dose of BI 853520 200 mg QD. At the time of data cut-off (29 October 2015), 61 patients had discontinued treatment due to progressive disease (n = 47; 75%), refusal to continue with the medication (n = 8; 13%), AEs (n = 5; 8%), or failure to comply with the study protocol (n = 1; 2%), and two patients remained on treatment. These two patients both discontinued within 6 weeks of database lock due to progressive disease. The maximum treatment duration for one of these patients, both of whom had pancreatic cancer, was 251 days. Among all 63 patients, the median number of cycles received was two (range 1–8), and the mean dose intensity was 83% (Table 3). Seven (11%) patients had received more than four cycles of treatment.

Table 3 BI 853520 treatment exposure by tumor cohort (expansion cohort; 200 mg QD)
Full size table

3.2.3 Safety

Drug-related AEs were reported in 61 (97%) patients, most commonly nausea, proteinuria, fatigue, diarrhea, and vomiting (Table 4). The majority of drug-related AEs were grade 1 or 2. Grade 3 drug-related AEs were reported in 22 (35%) patients, and included grade 3 proteinuria in 13 (21%) patients (Table 4). In all but one case, grade 3 proteinuria improved to grade 1 upon treatment interruption, and treatment could be reinitiated at a reduced dose of 100 mg. In the other case, proteinuria was reduced to grade 2. Two patients required a second dose reduction due to the recurrence of grade 3 proteinuria, and two patients underwent a kidney biopsy to further evaluate the proteinuria. Briefly, dysjunction of podocytes from the glomerular basement membrane and moderate-to-marked podocyte effacement were observed (details are provided in the ESM). No drug-related grade 4 or 5 events were observed.

Table 4 Most common drug-related adverse eventsa (expansion cohort; 200 mg QD)
Full size table

With regard to the laboratory assessments, no notable findings were observed, except for proteinuria (drug-related in 57% of patients). Serious drug-related AEs were reported in three (5%) patients [nausea (n = 2), vomiting (n = 3), and decreased appetite (n = 1)].

Nineteen (30%) patients underwent dose reduction due to AEs. Four (6%) patients had drug-related AEs leading to permanent discontinuation: grade 2 depressed mood (n = 1), grade 3 fatigue (n = 1), grade 2 vomiting (n = 1), and grade 2 abdominal pain combined with grade 1 diarrhea (n = 1).

3.2.4 Pharmacokinetics

The PK profile of BI 853520 was determined in 95 patients who received single and repeated oral doses of 10–300 mg. BI 853520 was rapidly absorbed, followed by at least bi-exponential disposition; maximum plasma concentrations (Cmax values) were observed at a median of 2 h (Fig. 1). Cmax and exposure (area under the curve values) increased with increasing doses. Dose-normalized geometric mean PK parameters suggested a trend toward an overproportional increase of exposure with dose; however, interpatient variability was moderate to high (Fig. 2). For unbound BI 853520, a low-to-moderate interpatient variability of exposure PK parameters was detected over all dose groups. Table 5 summarizes the PK parameters for BI 853520, and unbound BI 853520, after single- and multiple-dose administration (200 mg QD; see Electronic Supplementary Tables 2–5 for all dose groups).

Fig. 1
figure 1

Plasma concentration–time profiles for BI 853520 after single- and multiple-dose administration in the first cycle. Pharmacokinetic profiles after single- and multiple-dose administration were assessed during the first treatment cycle, i.e. after the first dose on day 1 and after repeated dosing on day 28

Full size image
Fig. 2
figure 2

Individual (circle) and geometric mean (cross) PK parameters for BI 853520 after single- and multiple-dose administration in the first cycle. PK profiles after single- and multiple-dose administration were assessed during the first treatment cycle, i.e. after the first dose on day 1 and after repeated dosing on day 28. AUC area under the plasma concentration–time curve, Cmax maximum plasma concentration, gMean geometric mean, PK pharmacokinetic

Full size image
Table 5 Pharmacokinetic parameters of BI 853520 after single- and multiple-dose administration (200 mg QD)a
Full size table

3.2.5 Pharmacodynamics

Paired tumor biopsies for analysis of FAK target engagement were available in 21 patients, however only eight of these patients had a complete set of data that allowed calculation of phosphorylated FAK (pFAK) and total FAK levels, both at baseline and during treatment. The lack of a complete data set for the remaining 13 patients was primarily due to FAK and pFAK levels being below the limits of quantification, partly due to limited amounts of tissue being available for both timepoints.

In five of the eight cases for whom complete sets of data were available [soft tissue sarcoma (n = 3), metastatic adenocarcinoma of the pancreas (n = 1), and esophageal cancer (n = 1)], the ratio of pFAK/total FAK was substantially decreased at the end of the first cycle compared with baseline (Fig. 3). There was a non-significant association between durable SD and decrease of the pFAK/total FAK ratio in one patient, while two patients with SD lasting two cycles also had a decrease of the pFAK/total FAK ratio. Given the limited number of patients with complete PD data and the absence of objective responses, it is not possible to draw a sound conclusion regarding correlation between decrease of the pFAK/total FAK ratio and efficacy.

Fig. 3
figure 3

Change from baseline to end of the first cycle in levels of phosphorylated FAK/total FAK in tumor tissue (expansion cohort; 200 mg QD). FAK focal adhesion kinase, QD once daily

Full size image

3.2.6 Efficacy

Of 63 patients treated in the expansion phase, 17 (27%) achieved a best response of SD [pancreatic adenocarcinoma (n = 6), platinum-resistant ovarian carcinoma (n = 5), soft tissue sarcoma (n = 4), and esophageal carcinoma (n = 2)] and 32 (51%) patients had progressive disease; 14 were not evaluable for response due to the absence of an on-treatment tumor assessment (Fig. 4). No objective tumor responses were observed; thus, the disease control rate was 27%. Median duration of disease control was 99 days (95% confidence interval 87–141 days), and four patients achieved disease control for ≥ 150 days [pancreatic adenocarcinoma (n = 2), platinum-resistant ovarian carcinoma (n = 1), and soft tissue sarcoma (n = 1)].

Fig. 4
figure 4

Confirmed best overall tumor response and treatment duration by tumor type (expansion cohort; 200 mg QD). AE adverse event, QD once daily

Full size image

Fifty of the 63 patients were evaluable for assessment of tumor shrinkage. Median best percentage change from baseline in the sum of longest diameters of target lesions was 8.97% (range − 25.7 to 168.6%).

3.2.7 Evaluation of E-Cadherin as a Potential Predictive Biomarker

Of the 57 patients for whom E-cadherin expression could be determined, 17 (30%) had low expression [soft tissue sarcoma (n = 13), esophageal carcinoma (n = 2), pancreatic adenocarcinoma (n = 1), and ovarian cancer (n = 1)]. Approximately one-quarter of patients in both the low- and high-expression E-cadherin groups achieved SD, as confirmed best overall tumor response (H-score ≤ 100: 4/17 patients, 23.5%; and H-score > 100: 11/40 patients, 27.5%). Median duration of disease control was also similar in both groups (low expression 103 days; high expression 95 days).

4 Discussion

This phase I trial demonstrated that continuous monotherapy with BI 853520 is feasible in heavily pretreated patients with advanced or metastatic nonhematologic malignancies; the MTD is determined at 200 mg QD.

The tolerability profile for BI 853520 is acceptable and is similar to that observed with other FAK inhibitors [e.g. VS-6063 (defactinib) and GSK2256098] [21, 22, 24]. The most common drug-related AEs included gastrointestinal events (nausea, diarrhea, vomiting), proteinuria, and fatigue, and were predominantly mild to moderate in severity.

No grade 4 or 5 drug-related AEs were reported; drug-related grade 3 AEs occurred in 35% of patients. Thirteen (21%) patients experienced grade 3 proteinuria, which was unexpected as preclinical toxicology studies did not identify any kidney abnormalities [31]. After the first occurrence of grade 3 proteinuria in a patient treated at the 300-mg dose, frequent monitoring of urine protein levels was implemented to facilitate early detection. Importantly, proteinuria was generally improved to grade 1 upon treatment interruption and subsequent dose reduction, suggesting a dose-dependent effect. To date, no predisposing or risk factors for the development of BI 853520-associated proteinuria have been identified, and no clear correlation was identified with prior nephrotoxic therapy. In kidney biopsies from two patients with grade 3 proteinuria, dysjunction of podocytes from the glomerular basement membrane and podocyte effacement was observed. FAK has been localized to the cytoplasm and nuclei of glomerular podocytes, and studies suggest that proteinuria may be related to activation of FAK in the glomerulus [32, 33]. However, the underlying mechanism of proteinuria following FAK inhibition remains unclear and requires further investigation. Proteinuria has also been reported as an AE following treatment with GSK2256098 [24] and PF-00562271 [21].

PK analysis demonstrated that BI 853520 was rapidly absorbed and exhibited at least biphasic disposition kinetics. A trend was observed toward an overproportional increase in mean exposure within the entire dose range investigated; however, interpatient variability of PK parameters was moderate to high. While this variability cannot yet be fully explained, it was within the expected range for an orally administered anticancer drug in the first dose-finding study. The PK data demonstrate oral bioavailability of BI 853520 and support a QD dosing schedule.

In five of eight cases, including three patients with soft tissue sarcoma, the ratio of active pFAK/total FAK was substantially reduced at the end of the first cycle, thus providing clinical evidence supporting target engagement with BI 853520 in patients treated with the 200 mg QD dose.

Disease control was observed in 27% of patients with a median duration of 99 days, and four patients achieved disease control for ≥ 150 days. No objective tumor responses were observed. In order to identify patients most likely to derive benefit from FAK inhibitors, we undertook an exploratory analysis of E-cadherin expression as a potential predictive biomarker [34]. There were no clear differences in disease control rate with respect to E-cadherin levels. However, it should be noted that the majority of patients with E-cadherin-negative tumors had soft tissue sarcomas, which are of mesenchymal origin and would not be expected to express E-cadherin at all. Based on this study, the potential predictive value of E-cadherin could not be determined, and further studies are warranted to identify other potential biomarkers to guide patient selection for FAK inhibitor treatment.

Given the cytostatic activity of BI 853520, the combination with other compounds such as immunotherapy [13, 19] should be considered for further development. In vitro and clinical studies suggest that combining FAK inhibitors with other biologic or chemotherapeutic agents such as MEK inhibitors, vascular endothelial growth factor (VEGF) inhibitors, immune checkpoint inhibitors, gemcitabine, paclitaxel, or cisplatin may overcome resistance, reduce metastasis and/or improve antitumor efficacy [13, 35,36,37,38,39]. A separate phase I study of BI 853520 has been conducted in 21 Japanese and Taiwanese patients with advanced or metastatic solid tumors [40]. In this study, a similar safety and PK profile was observed, with an equivalent MTD of 200 mg QD. The most common drug-related AEs were proteinuria and gastrointestinal events. One patient with gastric cancer who received BI 853520 100 mg QD achieved a confirmed partial response, with disease control seen in 29% of patients.

5 Conclusion

These data demonstrate that BI 853520 has a manageable and acceptable safety profile, favorable PK, and preliminary antitumor activity, including, for some patients, SD lasting ≥ 150 days, and PD modulation at the MTD of 200 mg QD in patients with selected advanced nonhematologic malignancies. As single-agent activity is modest, future development should focus on BI 853520 in combination with other agents.