Edward S. Kim, MD, and Emer Hanrahan, MD
Department of Medicine
The University of Texas M.D. Anderson Cancer Center, Houston, Texas
Combining Targeted Therapies
The molecular pathways involved in the proliferation of cancer cells and tumor angiogenesis are highly complex. Inhibiting a single step in these pathways may therefore not be the optimal approach to block disease progression. After prolonged inhibition of one aspect of a pathway, tumor cells can develop an escape mechanism. However, it is anticipated that combining agents that block multiple different aspects of the same pathway may improve therapeutic benefit. This concept is supported by substantive preclinical evidence of synergistic antitumor activity when targeted agents that block multiple signaling pathways are combined.48-52
The epidermal growth factor receptor (EGFR) shares common downstream signaling pathways with the vascular endothelial growth factor (VEGFR), and evidence suggests both that VEGF is downregulated by EGFR inhibition and that blockade of VEGF may inhibit EGFR autocrine signalling.53,54 Therefore, it has been hypothesized that dual blockade of these molecular targets would have additive or synergistic antitumor effects, a theory supported by several preclinical studies.50-52,55 The EGFR tyrosine kinase inhibitors (TKIs) gefitinib and erlotinib initially were thought to offer the greatest hope for targeted therapies in non-small cell lung cancer (NSCLC). Although erlotinib was shown in a randomized trial to improve survival relative to supportive care in refractory stage IIIB/IV NSCLC, gefitinib, in a similar trial, did not.56,57 The failure of these agents to improve survival in combination with first-line chemotherapy has been most disappointing.2,58-60 Similarly, the anti-EGFR monoclonal antibody cetuximab, when added to cisplatin in the first-line treatment of patients with recurrent or metastatic squamous cell carcinoma of the head and neck, improved response rate, but not overall survival.61
Data from clinical trials combining agents that target both VEGF (either bevacizumab or VEGFR tyrosine kinase inhibitors) and EGFR signaling (cetuximab, gefitinib, and erlotinib) have already shown promise (Tables 6 and 7).62-65 Bevacizumab has been combined with chemotherapy or erlotinib in patients with second-line NSCLC.66 Results are encouraging for combinations of chemotherapy with bevacizumab, as well as the biologic combination of bevacizumab with erlotinib. A randomized phase III trial is assessing the activity of erlotinib with bevacizumab in salvage NSCLC based on the design illustrated in Figure 9.
Bevacizumab and sorafenib in combination are also under study. This antiangiogenesis strategy targets the ligand (VEGF), receptors VEGFR2 and VEGFR3, and one of the downstream signaling pathways (Raf). Preliminary results of an ongoing phase I study in patients with refractory, metastatic, or unresectable solid tumors were recently reported.67 Thirty-four of 37 evaluable patients (92%) had clinical benefit. Four patients, all with ovarian cancer, had a partial response.
Dual Inhibitors of EGFR and VEGFR
ZD6474
One of the most promising VEGFR TKIs in NSCLC is ZD6474, an orally bioavailable potent inhibitor of VEGFR2. This agent also has activity against EGFR, fibroblast growth factor receptor 1 (FGFR-1), and RET.9,68,69 Inhibition of EGFR signaling has been shown to inhibit the secretion of VEGF, bFGF, and TGF-.53 The antiangiogenic potential of ZD6474 may be due to a reduction in the level of proangiogenic agents through its anti-EGFR activity, as well as direct inhibition of the VEGFR-2 TKI.
Preclinical studies have shown that the use of ZD6474 results in inhibition of lung tumor growth and metastasis in human xenografts models.9,68 Combining ZD6474 with paclitaxel or radiation therapy has been shown to produce greater growth inhibition than either agent alone.68-70 ZD6474 was evaluated in two phase I studies of patients with refractory solid tumors, one involving a Western population (n=49), the other, a Japanese population (n=18).71,72 Therapy was well tolerated at doses of 300 mg/day. The main reported side effects were facial flushing, facial rash, fatigue, diarrhea, and asymptomatic QTc interval prolongation. More than 40% of Western patients had stable disease ≥ 8 weeks, and tumor regression wasobserved in four of nine Japanese patients (44%) with NSCLC. Several phase II studies of ZD6474 alone or in combination with chemotherapy in patients with previously treated NSCLC have subsequently been conducted. In a phase II randomized trial, 168 patients with locally advanced or metastatic, platinum-refractory NSCLC were allocated to receive either daily oral ZD6474 (300 mg) or gefitinib (250 mg) until disease progression or limiting toxicity (part A).10 On progression, patients had the option of crossing over to the alternative therapy (part B). The adverse event profile of ZD6474 was similar to that seen in previous trials (diarrhea grade 3/4, 8.4%; rash grade 3/4, 4.8%; asymptomatic QTc prolongation, 20.5%). There was a statistically significant improvement in median time to progression with ZD6474, compared with gefitinib (11.0 vs. 8.1 weeks, respectively; p= 0.025). In part B, disease control >8 weeks was achieved in 16 of 37 patients who switched from gefitinib to ZD6474 and in 7 of 29 patients who switched from ZD6474 to gefitinib. No significant difference was noted in overall survival between patients initially randomized to either ZD6474 or gefitinib, which is not surprising given the crossover design.
In a phase II trial assessing ZD6474 combined with chemotherapy, 127 patients (37 with squamous histology) with advanced, platinum-refractory NSCLC were randomized to receive once daily oral ZD6474 (100 or 300 mg) or placebo in combination with intravenous docetexal (75 mg/m2) every 21 days.43 The toxicity profile for ZD6474 was similar to that seen in earlier studies. Four patients with squamous histology (two patients in the ZD6474 arm and two patients in the placebo arm) experienced nonfatal hemoptysis. This study met its primary endpoint of prolonged median progression-free survival in the ZD6474-containing arm (19 weeks with 100 mg, p = 0.074; 17 weeks with 300 mg, p = 0.461; 12 weeks with docetaxel alone), and a phase III trial using ZD6474 (100 mg) is now being planned. A three-arm phase II trial combining ZD6474 with chemotherapy is ongoing: carboplatin plus paclitaxel combined with ZD6474 (200-300 mg), chemotherapy alone, or ZD6474 alone.
AZD2171
AZD2171 is an orally bioavailable highly potent TKI of VEGFR2, with lesser activity against platelet-derived growth factor receptor ß (PDGFR-ß) and c-Kit. This agent has been investigated in several phase I trials and is well tolerated at doses ≤ 45mg/day.5,44,73 One phase I trial specific to NSCLC evaluated AZD2171 in combination with carboplatin and paclitaxel in patients with advanced NSCLC (any histology).74 Twenty patients were enrolled, and toxicities were manageable. Nine patients were treated at an AZD2171 dose of 30 mg, with one dose-limiting toxicity (grade 3 elevation in ALT) and hypertension ≥ grade 2 in six patients. Eleven patients received a dose of 45 mg, with one dose-limiting toxicity reported (grade 3 febrile neutropenia with grade 3 mucositis). Other common toxicities included fatigue, anorexia, mucositis, and diarrhea; no hemoptysis was reported. Myelosuppression was not greater than expected with chemotherapy alone. Of 15 evaluable patients, six patients experienced partial response and eight patients had stable disease. A phase II/III trial of carboplatin and paclitaxel +/- AZD2171 for first-line treatment in stage IIIB (wet) or IV NSCLC is now being planned.
Conclusions
Studies of anti-VEGF agents in combination with other targeted therapies, most notably EGFR inhibitors, are advancing rapidly, as are studies of newer agents that target both VEGF and other kinases (i.e., PDGFR, c-Kit) or EGFR tyrosine kinases. At present, these approaches are mainly in phase I/II trials, although phase III studies evaluating the combination of erlotinib and bevacizumab in advanced stage NSCLC are currently being conducted following very promising results in earlier trials.
Discussion
Dr. Lynch: How do we know which agents to combine? Are there models that are helpful?
Dr. Socinski: I'm hopeful that we will be able to identify some preclinical or xenograft models that will reliably tell us the importance of the individual targets, as well as what combinations may be best.
Dr. Lynch: So, what do you see as the challenge in the next few years?
Dr. Socinski: Because we have so many potential agents that could be used in combination, the challenge is fairly mind boggling. We probably need two to four good agents that work, which will increase competition in the marketplace and give clinicians the flexibility of using an active class of drugs and tailoring therapy to the individual patient.