Phase I Evaluation of Telatinib, a Vascular Endothelial Growth Factor Receptor Tyrosine Kinase Inhibitor, in Combination with Irinotecan and Capecitabine in Patients with Advanced Solid Tumors

The importance of angiogenesis in solid tumor growth is well established. Inhibition of the vascular endothelial growth factor (VEGF)/VEGF receptor (VEGFR) pathway is now an approved and generally accepted treatment of renal cell, hepatocellular, breast, colorectal, and lung cancer (1–6). Interestingly, bevacizumab, a monoclonal antibody against VEGF-A, is predominantly administered in combination with chemotherapy, whereas the small-molecule inhibitors of the VEGFRs, sunitinib and sorafenib, are used as single agent (2, 4–6).

The mechanism underlying the synergism between the combination of bevacizumab and chemotherapy is not completely understood, but preclinical and early clinical research point to possible explanations (7, 8). First, improving or normalization of the leaky and ineffective vasculature by the addition of a VEGF-inhibiting agent is an emerging concept to enhance the efficacy of concomitantly administrated cytotoxic therapies (9, 10). Second, addition of antiangiogenic agents within the drug-free periods between chemotherapy cycles might inhibit the tumor cell division and tumor regrowth in the chemotherapy-free periods (11). Finally, it has recently been shown in preclinical and clinical studies that certain anticancer treatments can induce an immediate mobilization of endothelial progenitor cells (EPC) from the bone marrow to the tumor within hours after start of the treatment. Interestingly, in mice, this phenomenon prevented necrosis induced by therapy and could be inhibited by an antibody against the VEGFR-2, restoring or enhancing the antitumor effect induced by therapy (12–14).

Although disagreement still exists on the contribution of these cells to the actual growth of the tumor (15–20), it is univocally shown that these cells have a crucial function in metastasis formation, the transition of micrometastasis to macrometastasis, and prevention of necrosis after therapy (21, 22). Addition of a VEGF (endothelial cell)–inhibiting agent to conventional chemotherapy regimens might therefore act synergistically.

The major breakthrough for combinatorial treatment regimens was constituted by the clinically meaningful improvement in survival observed in metastatic colon cancer patients treated with irinotecan, capecitabine, leucovorin, and bevacizumab (1).

In contrast to the established added value of bevacizumab to chemotherapy in the first-line treatment of metastasized colon cancer, small-molecule tyrosine kinase inhibitors (TKI) targeting the VEGFR have not been shown to enhance the efficacy of conventional chemotherapy yet (23). We therefore embarked on a clinical study to investigate the combination of the VEGFR-TKI telatinib with a combination of capecitabine and irinotecan in patients with advanced solid tumors.

Telatinib (BAY 57-9352) is an orally available, highly potent, small-molecule inhibitor targeting the tyrosine kinase domain of the VEGFR (KDR/VEGFR-2/VEGFR-3), platelet-derived growth factor receptor β, and c-Kit. It has low affinity for the Raf kinase pathway, epidermal growth factor receptor family, the fibroblast growth factor receptor (FGFR) family, or the Tie-2 receptor. The antitumor activity of telatinib has been shown in a range of preclinical models and the safety of telatinib monotherapy has already been shown in a phase I trial (24). We studied the feasibility and evaluated safety of telatinib in combination with capecitabine and irinotecan in a phase I study. Secondary objectives included the determination of the pharmacokinetic (PK) profile of telatinib in combination with capecitabine and irinotecan, investigation of the effect of telatinib on markers of biological activity (VEGF/sVEGFR-2 and circulating endothelial cell measurements), and preliminary evaluation of efficacy.

In two centers in the Netherlands, adult patients with histologic or cytologic evidence of advanced solid tumors refractory to or failing standard treatment or patients with advanced colorectal cancer eligible for second-line chemotherapy treatment were recruited. Patients were required to have progressive disease within 6 mo before study entry based on radiological assessment; at least one measurable lesion; WHO status of ≤1; a life expectancy of at least 12 wk; and an adequate bone marrow, renal, and liver function. The most important exclusion criteria were a history of central nervous system tumors or metastases, a history of cardiac disease, congestive heart failure New York Heart Association class of >2, active coronary artery disease, cardiac arrhythmias requiring antiarrhythmic therapy, poorly controlled hypertension, uncontrolled infections, patients with serious nonhealing wounds, patients with baseline coagulation disorders, gastrointestinal disorders resulting in malabsorbtion, pregnant or breast feeding women, and patients with toxicity suggestive of dihydropyrimidine dehydrogenase deficiency or UGT1A1 polymorphisms (*6, *7, *27, *28, and *29).

The study was approved by both institutional ethics committees and all patients provided written informed consent. The trial was conducted in accordance with the Declaration of Helsinki.

In this phase I, two-center, open-label, dose escalation study, patients were included in successive cohorts of three patients with increasing dose of telatinib or irinotecan. Capecitabine was administered at a fixed dose of 1,000 mg/m² twice daily every first 14 d of each cycle in all four cohorts (Table 1). Telatinib treatment was started on day 5 of cycle one and was given twice daily continuously. Patients in the first dose escalation cohort were treated with 300 mg telatinib twice daily, 125 mg/m² irinotecan infusion once every 21 d, and 1,000 mg/m² capecitabine twice daily every first 14 d of each cycle, both starting at day 1 of cycle one. Predefined maximum doses (telatinib and irinotecan) and fixed dose (capecitabine) based on previously performed phase I studies of telatinib alone and of the combination of irinotecan and capecitabine were 900 mg twice daily, 180 mg/m², and 1,000 mg/m², respectively (24–26). In all four cohorts, patients received telatinib until tumor progression or when uncontrollable toxicity was encountered. The chemotherapy regimens were administered up to a maximum of six cycles. From that moment on, patients were treated with monotherapy telatinib until disease progression, unacceptable toxicity, or withdrawal of consent. Individual dose modifications as a consequence of toxicity were done according to predefined guidelines.

Dose limiting toxicity (DLT) was defined as any combination regimen or telatinib related nonhematological adverse event of at least Common Terminology Criteria for Adverse Events (CTCAE) version 3.0 grade 3 occurring during the first and/or second cycle (one treatment cycle is equivalent to 3 wk) of treatment with the exception of alopecia, nausea/diarrhea well controlled by intervening treatment, and liver function disturbances no longer persisting than 3 wk. Hypertension grade 3 refractory to antihypertensive treatment according to the predefined hypertension management protocol or grade 4 was considered to be a DLT. Hematologic adverse events considered as DLT were as follows: neutropenia defined as <0.5 × 109/L neutrophils for >7 d, neutropenia with fever of ≥38.5°C, absolute neutrophil count of <0.5 × 10⁹/L, and platelets of <25 × 10⁹/L or thrombocytopenic bleeding CTCAE grade 3. In case of a DLT, the cohort was expanded to six patients. If DLT was observed in more than one of the six patients within a dose level a, that dose was considered above the maximum tolerated dose, and dose escalation was stopped. Safety review meetings were held for each dose level before entering the next dose level.

At every biweekly visit during the course of the study, a physical examination, assessment of adverse events, clinical chemistry, hematology, and urinalysis were done.

Cardiac function was monitored before each treatment cycle by an electrocardiogram. Tumor assessment was done before the start of the study and every 6 wk thereafter or at the discretion of the investigator. Response was assessed using the Response Evaluation Criteria in Solid Tumors guidelines (27).

Blood samples were collected to determine the plasma concentrations of irinotecan and SN 38 in the dose escalating cohorts on day 1 of cycle 1 and on day 1 of cycle 2 before dosing and at 1, 1.5 h (just before the end of the infusion), 1.75, 2, 2.5, 4, 6, 8, 12, 24, 48, and 72 h thereafter; of capecitabine and 5-fluorouracil (5-FU) on day 1 of cycle 1 and on day 1 of cycle 2 before dosing and at 0.5, 1, 2, 4, 6, 8, and 12 h thereafter; and of telatinib and its metabolite M2 (BAY 60-8246) on day 21 of cycle 1 and on day 1 of cycle 2 before dosing and at 0.5, 1, 2, 4, 6, 8, and 12 h thereafter.

The plasma concentrations of telatinib, BAY 60-8246, capecitabine, and 5-FU were determined using specific high-performance liquid chromatography tandem mass spectrometry assays with a lower limit of quantification of 0.002 mg/L (telatinib and BAY 60-8246), 25 ng/mL (capecitabine), or 5.0 ng/mL (5-FU). For the determination of plasma concentrations of irinotecan and SN 38, a specific high-performance liquid chromatography assay with fluorescence detection was applied with an lower limit of quantification of 2.0 ng/mL for both compounds.

The primary PK characteristics of area under the curve (AUC) and C_max (irinotecan), AUC(0-24) and C_max (SN 38), AUC(0-12) and C_max (capecitabine and 5-FU), or AUC(0-12) and C_max (telatinib and metabolite M2), respectively, were analyzed assuming log-normally distributed data.

The logarithms of these PK characteristics were analyzed using ANOVA. Based on these analyses point estimates (LS-Means) and exploratory 90% confidence intervals for the ratios of parameters after administration of all drugs simultaneously versus administration of chemotherapy and telatinib alone were calculated by retransformation of the logarithmic data.

Blood samples for the measurement of circulating endothelial (progenitor) cells were collected on cycle 1 day 1 (before dosing) and on day 14. Mononuclear cells were isolated by means of a 8-mL CPT tube (Becton Dickinson). Additional plasma samples were stored for the determination of soluble VEGFR-2 and VEGF before dosing and 8 h after dosing cycle 1 on day 1, 3, 4, and 21; cycle 2 on day 1 and day 14; and subsequent cycles on day 1.

Endothelial (progenitor) cells were quantified by four-color flow cytometry using CD45, CD31, CD146, and CD133 as markers as previously reported (28). Plasma VEGF and sVEGFR-2 levels were measured using commercially available sandwich ELISA kits (R&D Systems) following the manufacturer's instructions. Statistical comparisons between baseline and each of subsequent time points were done using the Student's t test. All tests were two sided. P values lower than 0.05 were considered as statistically significant (GraphPad prism 4.0/SPSS version 15).

A total of 23 patients were enrolled in the study in four different dose escalating cohorts (Table 1). All patients were valid for safety analysis and 17 patients were valid for PK analysis. The median age of the patients was 57 years (range, 20-71). Additional patient characteristics are provided in Table 2.

Dose level I enrolled three patients (Table 1). The combination at this dose level was well tolerated. Dose level II enrolled seven patients in total. Due to a sudden death of the first patient in this cohort that occurred after just a few days of treatment, the study was interrupted for 4 months in expectation of the autopsy results, PK analysis and UGT1A1 polymorphism analysis from the deceased patient. Based on detailed analysis of this patient, it was decided that the death was unrelated and that it was considered safe to proceed with the study. Although the event was eventually not assessed as a DLT, for safety reasons, it was decided to expand the cohort to six patients. Because another patient experienced an acute anticholinergic syndrome due to irinotecan infusion, the patient was replaced. In total, five patients in this cohort tolerated treatment well and it was decided to increase the dose of telatinib to 600 mg twice daily according to the protocol.

Dose level III enrolled six patients. Three patients withdrew their consent before the observation period of two cycles (due to non-DLT toxicity, difficulty to swallow, and personal reasons) and had to be replaced. Once more, the combination at this dose level was well tolerated and because of the absence of DLTs, the dose of telatinib was increased to the recommended phase II dose of 900 mg twice daily (24).

Dose level IV at start enrolled three patients. After 3 months of continuous telatinib administration, all three patients showed diverse cardiotoxicity such as electrocardiogram changes, a myocardial infarction, and a significant systolic dysfunction. It was decided to add three additional patients with intensive cardiac monitoring. One of these patients withdrew consent after the first day of treatment because of personal reasons and had to be replaced. No further signs of cardiotoxicity were observed at this dose level. The study was, as outlined in the protocol, completed at this dose level because the recommended doses for telatinib and irinotecan from phase I studies was attained.

All 23 patients enrolled in the study received at least one dose of study medication and therefore were assessable for safety analysis. Treatment emergent adverse events observed in ≥25% of the patients were vomiting (83%), nausea (78%), fatigue (70%), diarrhea (61%), alopecia (52%), hand-foot syndrome (48%), constipation (48%), and voice changes (35%). Grade 3 and 4 toxicities are presented in Table 3. Serious adverse events reported related to study treatment were cardiac ischemia/infarction, aspecific cardiac complaints with normal cardiac ultrasound, left ventricular systolic dysfunction, sudden death, and diarrhea. Following the per protocol definitions, no DLTs were encountered.

Two deaths during treatment were reported. In dose level II, the first patient suddenly died after 2 days of combination treatment. Although not likely related to the study drug, a relation could not be ruled out and results from the autopsy could not provide a cause of death. Because of the fact that in the past, the patient was treated for a heart rhythm disorder and before his death this patient suffered from an atrial fibrillation, a cardiac cause of death seemed to be likely. PK analysis showed no significant abnormalities and there was no UGTA1 polymorphism present. The second patient died of disease progression after 107 days of treatment in dose level IV.

In dose level IV, one patient experienced a silent myocardial infarction 9 weeks after the start of the study, confirmed by ultrasound registration. After discontinuation of the study drug, the electrocardiogram changed back to normal. In the same dose level, two cases of low left ventricular ejection fraction (LVEF; CTCAE grade 2 and 3) were observed, respectively, 16 and 19 weeks after the start of study treatment. In both patients, the left ventricular dysfunction was preceded by symptoms of dyspnoea d' effort, and on ultrasound, the ejection fraction of the left ventricle was 45% and 25%, respectively. Cardiac follow-up of these two patients after the discontinuation of the study drug showed improvement of the left ventricle function to 63% and 53%, respectively, within 6 to 12 weeks. Remarkably, all these cardiac events started with minimal, clinically not significant electrocardiogram disturbances and without the presence of symptoms, and were reversible after discontinuation of the study drug. In addition, none of these patients had a history of heart problems or cardiac risk factors. Intensive cardiac monitoring in the extra three patients at this dose level showed no further cardiac toxicity.

The median numbers of days on treatment for the four different cohorts for telatinib were 174, 60, 65, and 96, respectively (Table 4). In dose level I, no dose modifications occurred. Due to hand-foot syndrome and neutropenia in dose level II, two dose reductions of capecitabine or irinotecan occurred in two patients. In dose level III, in two patients, two dose reductions in capecitabine and irinotecan, respectively, occurred because of hand-foot syndrome and liver function abnormalities (increased aspartate aminotransferase). No dose reductions occurred in the forth cohort. Primary reason for permanent discontinuation was disease progression (13 of 23) followed by adverse events (4 of 23) and consent withdrawn (4 of 23; Table 4).

Eighteen patients were assessable for antitumor activity of which 17 patients had tumor measurements by Response Evaluation Criteria in Solid Tumors. Five patients discontinued the study before the first radiological assessment due to a sudden death, consent withdrawn (n = 2), and adverse event (n = 2). Five of 23 patients showed a partial response with a median duration of 2.2 months (95% confidence interval, 1.3-12.7) and 9 of 23 patients showed stable disease with a median duration of 4.3 months (95% confidence interval, 2.7-9.9), cumulating in a clinical benefit rate of 61% (14 of 23). The group of the patients with a confirmed partial response consisted of three patients with colorectal cancer, two patients with an adenocarcinoma of an unknown primary and one patient with a chordoma. Tumor shrinkage was present in 11 of 17 patients (65%). Although small patient numbers are prohibiting any definite conclusions, the highest shrinkage rate was observed in the 900 mg telatinib dose level.

Seventeen (telatinib, capecitabine) and 16 (irinotecan) of the 23 patients enrolled were evaluable for PK analysis. Geometric mean plasma concentration time courses of telatinib from cycle 1, day 21 (telatinib alone) and from cycle 2, day 1 (telatinib upon coadministration with irinotecan and capecitabine) were analyzed. Table 5 shows the primary PK parameters of all six compounds analyzed.

Following oral administration, telatinib was rapidly absorbed with a median t_max of 2 to 4 hours. The systemic exposure data correspond well with previously published data from the phase I trial with telatinib (24). The simultaneous administration of 300, 600, or 900 mg twice daily telatinib, 1,000 mg/m² twice daily capecitabine, and of 125 or 180 mg/m² irinotecan revealed no clear effect of the comedication on the PK of telatinib and of its metabolite BAY 60-8246. Either no or not relevant changes in PK of irinotecan and of SN 38, and small to moderate changes in PK of capecitabine and of 5-FU were observed, although individual effects might be disguised by the mean values due to the high interidividual and intraindividual variability of the PK especially of capecitabine and 5-FU. The nonappearance of any significant interaction is consistent with the independent mechanism of metabolism and transport for all these agents.

Plasma biomarker analysis consisting of (circulating) endothelial (progenitor) cells by flow cytometry analysis showed that the addition of telatinib to chemotherapy stabilizes progenitor cell/EPC levels in patients with progressive disease. Furthermore, this stabilization seemed to be dose dependent. (Fig. 1A and B) Measurements of sVEGFR-2 levels revealed a clear reduction starting at cycle 1 day 21 through the entire course of treatment (total group: P < 0.0009). Plasma VEGF levels had a tendency to increase during treatment, with a generally higher variability regarding their absolute levels and relative changes, compared with sVEGFR-2. (Fig. 1C and D).

The addition of bevacizumab to chemotherapy regimens has proven its clinical benefit in the treatment of colorectal, breast, and lung cancer. In contrast to bevacizumab, small-molecule TKIs targeting the VEGFR have not yet shown to enhance the efficacy of conventional chemotherapy in clinical trials (23). Nevertheless, it might be favorable to combine chemotherapy with VEGFR-2–inhibiting agents that are available in oral formulations and which have an apparently milder toxicity profile, expressed in a lower incidence of acute disorders such as gastrointestinal perforations and coagulation disorders. Furthermore, the majority of bevacizumab-treated patient will become resistant to treatment during treatment. The VEGFR-targeting TKIs have in general a unique but diverging target specificity profile. From that point of view, one could speculate that TKIs, targeting multiple tyrosine kinases of other potentially to be upregulated proangiogenic factors (i.e., PLGF, PDGF, etc.) during VEGF-inhibiting treatment, might block compensatory resistance pathways (23, 29–32). In this study, we combined the VEGFR-2 TKI (with affinity also for platelet-derived growth factor receptor β and c-Kit) telatinib with a chemotherapy regimen consisting of irinotecan and capecitabine to maximize the therapeutic effect compared with treatment with the chemotherapeutic regimen alone. In the phase I telatinib monotherapy trials, maximum tolerated dose was set at 900 mg twice daily in a continuous regimen (24). From these phase I studies, telatinib toxicity was considered as mild and combining this agent with chemotherapy treatment was expected to be safe.

The results from the present study indeed confirm that the combination of telatinib and a chemotherapy regimen consisting of irinotecan and capecitabine is tolerated and sufficiently safe provided that cardiac monitoring is included during the course of treatment. The most frequent toxicities of this combination treatment reported were vomiting, nausea, fatigue, diarrhea, alopecia, hand-foot syndrome, and constipation indicative for the fact that the toxicity profile of the study drug combination consists mainly of the known toxicities caused by irinotecan and capecitabine. The addition of telatinib to the combination did not seem to increase the frequency or the severity of this well-known toxicity caused by the chemotherapy (33–35). In particular, the presumed increase of diarrhea caused by both telatinib as well as the combination irinotecan/capecitabine possibly impeding adequate resorption of the TKI was not observed. Hypertension did occur at a frequency one would expect for a VEGF-inhibitor of this class and grade 3 hypertension was observed at lower frequencies than in the monotherapy phase I trials with telatinib (17% versus 23%; refs. 24, 36–40). Strikingly, in contrast to combinatorial regimens consisting of chemotherapy and other VEGFR TKIs, no significant myelosuppression was observed (n = 2, 9%; refs. 41–43). This might be explained by differences in TKI affinity or the composition of the chemotherapy regimens.

Single-agent studies with telatinib, sunitinib, and sorafenib showed, respectively, in 1.9%, 42%, and 31% of the patients any grade bone marrow suppression (24, 44, 45). This may indicate that telatinib may be more suitable to combine with chemotherapy than other VEGFR-TKI.

Cardiac toxicity was reported in three cases, consisting of a silent myocardial infarction and two cases of decreased LVEF. The LVEF decreases normalized again after the discontinuation of the study drugs. Due to the small numbers in this study and the heavily pretreated patient population, a final assessment about the actual cardiotoxic potential for the telatinib/irinotecan/capecitabine combination is not possible. However, cardiotoxicity is a frequently reported phenomenon for this class of anticancer agents, although varying incidences have been reported for the clinically approved VEGFR-TKI (38, 39, 46, 47). Further insight and revelation of the exact underlying mechanisms is of great importance. Successive phase II studies with this combination should include cardiac monitoring on a regularly basis to address this research question.

No DLTs were reported in this study; thus, the maximum tolerated dose was defined as for the combination of telatinib (900 mg twice daily), 180 mg/m² irinotecan, and 1,000 mg/m² capecitabine at the applied schedule. Consequently, the recommended phase II dose for the combination of telatinib with capecitabine and irinotecan is 900 mg telatinib twice daily continuously, 180 mg/m² irinotecan thrice weekly, and 1,000 mg/m² capecitabine twice daily on day 1 to 14.

The Colorectal Oral Novel Therapy for the Inhibition of Angiogenesis and Retarding of Metastases (CONFIRM) 1 and 2 trials, in which vatalanib, VEGFR-2 TKI was combined with FOLFOX-4 regimen as first-line and second-line treatment for metastasized colorectal cancer, respectively, showed no enhanced activity for the combination (48, 49). In our study, a clinical benefit rate of 61% was observed in a typical heterogeneous, heavily pretreated phase I population. In six patients with colorectal cancer, three partial responses occurred (50%). In comparison with clinical trials combining capecitabine or 5-FU and irinotecan as second-line therapy in metastasized colorectal cancer patients, in which a clinical benefit rate of 34% and objective response rates of 4% were reported, we might conclude that the combination has antitumor activity (26, 50).

The PK profiles of telatinib as well as of irinotecan, capecitabine, and their metabolites were not meaningfully altered by coadministration. Incidental changes observed were of low magnitude and within the usual range of interpatient variability. Pharmacodynamic analysis showed a decrease in sVEGFR-2 and a more variable pattern but with a trend toward upregulation of VEGF during the course of treatment both as reported before in literature. Analysis of EPC levels showed stabilized levels during the course, possibly suggesting that addition of telatinib might blunt chemotherapy induced EPC release. The absence of a proper control (telatinib only) prohibits a definitive conclusion on this part and the findings should be considered as exploratory. In the last dose level, inhibition of EPCs was most effective, possibly reflected by the highest observed tumor shrinkage at this level.

In conclusion, this study reveals that the combination of telatinib and irinotecan plus capecitabine was sufficiently tolerated at relevant single-agent doses of all three agents, and antitumor activity was found in severely pretreated patients. These results support the further development of this regimen as treatment of metastasized colon cancer under the condition that regular cardiac monitoring is incorporated in following studies.

No potential conflicts of interest were disclosed.

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.