Introduction

Non-small cell lung cancer (NSCLC) accounts for approximately 85% of lung cancer diagnoses [1] and the majority of lung cancer deaths worldwide [1, 2]. Recommended systemic treatment options for patients with advanced or metastatic NSCLC have historically included cytotoxic chemotherapy, with targeted agents directed against the vascular endothelial growth factor (VEGF) and the epidermal growth factor receptor (EGFR) representing more recent therapeutic additions for use in specific treatment settings or populations. Maintenance monotherapy with a chemotherapeutic or targeted agent has also emerged as an important consideration following the use of first-line chemotherapy [3].

With respect to cytotoxic chemotherapy, various platinum-based doublets have demonstrated comparable efficacy in terms of palliative benefits and improvements in survival when used in the treatment of patients with advanced NSCLC [3, 4]. Although no single platinum-based doublet has been considered superior in the overall advanced unselected NSCLC population, histology has emerged as an important consideration for individualizing regimen selection. The benefit of histological screening is particularly evident in the improved activity seen with the antifolate pemetrexed (Alimta®, Eli Lilly and Company) when used in patients with adenocarcinoma and large-cell carcinoma versus squamous cell carcinoma [5].

Over the past decade, targeted agents have emerged as novel therapeutic options for NSCLC, both as a part of combination therapies and as single agents. Bevacizumab (Avastin®, Genentech), a monoclonal antibody that targets VEGF, is the only United States (US) Food and Drug Administration (FDA)-approved adjunct to first-line platinum-based chemotherapy for NSCLC. Bevacizumab is specifically indicated for patients with nonsquamous tumors [6] and is recommended for patients with a good performance status and no recent hemoptysis [3]. The EGFR tyrosine kinase inhibitors (TKIs) erlotinib (Tarceva®, Genentech) and gefitinib (Iressa™, AstraZeneca) also have FDA-approved indications for use as monotherapy in patients with chemotherapy-pretreated advanced NSCLC [7, 8]. However, due to inconsistent benefit observed with gefitinib, its indication in the United States is limited to patients who are or have previously benefited from gefitinib therapy [7]. This article provides an overview of the rationale for selection of VEGF, platelet-derived growth factor (PDGF), and fibroblast growth factor (FGF) pathway signaling as a therapeutic target. Past and ongoing trials of bevacizumab and other developmental antiangiogenic agents for the treatment of advanced NSCLC are summarized.

Rationale for antiangiogenic VEGF-targeted therapy

Angiogenesis, the development of new blood vessels, is a critical component of tumor development and progression, as this process provides access to the nutrients and oxygen needed for tumor growth [9, 10]. This is a key therapeutic target because in order to grow larger than a few millimeters in diameter, solid tumors must develop a vascular supply [11]. The process of angiogenesis relies on a multitude of signaling cascades and angiogenic factors [12–15]. Of the many proangiogenic signaling molecules, VEGF (or VEGF-A) is the most widely studied to date [14, 15]. VEGF-A belongs to a larger family of 6 secreted glycoproteins that also includes VEGF-B, -C, and -D, and placenta growth factor (PlGF)-1 and −2 [16, 17]. Although the spatiotemporal expression of these isoforms varies, all are involved in angiogenesis and/or lymphangiogenesis in normal human development and in pathological states [16].

Binding of VEGF family ligands to 3 cognate receptor tyrosine kinases (VEGF receptor-1 [VEGFR-1], VEGFR-2, and VEGFR-3) results in the activation of numerous signaling factors that ultimately promote cellular migration, survival, and proliferation (Fig. 1) [12, 18]. VEGF-A has been shown to activate both VEGFR-1 and VEGFR-2, two highly similar receptors that have distinct roles in angiogenesis. VEGFR-1 is required for proper formation of the vasculature during early embryonic development and can also bind VEGF-B and PlGF [16, 17]. This receptor is expressed in the vasculature and in other cell types and can serve as both a positive and negative regulator of angiogenesis [17]. VEGFR-2 is primarily expressed on the vascular endothelium, where it plays a substantial role in VEGF–A-mediated angiogenesis, although VEGF-C and VEGF-D can also serve as ligands. VEGFR-2 signaling mediates microvascular permeability and proliferation, migration, and survival of endothelial cells [16, 17]. VEGFR-3 preferentially binds VEGF-C and VEGF-D. VEGFR-3 is expressed in the embryonic vasculature, but throughout development and in the adult, its expression is limited to lymphatic endothelial cells. VEGFR-3 is thought to be involved in remodeling of primary vascular networks during embryogenesis and in lymphangiogenesis in the adult [16, 17]. Coordination of VEGFR activity with other signaling pathways, such as the PDGF and FGF pathways, may be required to support tumor angiogenesis. These pathways appear to function in a synergistic manner to promote the formation of new vessels that provide nutrients and oxygen to tumors, in addition to promoting dissemination of tumor cells [19].

Fig. 1
figure 1

The VEGF pathway and sites of action for inhibitors. Binding of VEGF family ligands to their cognate receptors, VEGFR tyrosine kinases (VEGFR-1, VEGFR-2, and VEGFR-3), results in the activation of numerous signaling factors that ultimately promote cellular migration, survival, and proliferation. The sites of action for the antiangiogenic inhibitors discussed herein are also illustrated. The VEGF pathway and sites of action for inhibitors. Used with permission of Society for Experimental Biology and Medicine, from Antiangiogenic therapy in lung cancer: focus on vascular endothelial growth factor pathway by Korpanty G, Smyth E, Sullivan LA, Brekken RA, Carney DN. Exp Biol Med (Maywood) 235:3–9 (2010); permission conveyed through Copyright Clearance Center, Inc

Full size image

The VEGF signaling axis is now a validated therapeutic target in NSCLC. Early preclinical models have illustrated the ability of VEGF inhibition to slow tumor growth and angiogenesis [20]. Observations supporting high levels of VEGF and VEGFR expression or upregulation in human malignancies have further supported research efforts to develop anti-VEGF/VEGFR therapies [21–23]. Inhibitors targeting the VEGF pathway, as well as other pathways involved in angiogenesis could potentially lead to substantial improvements in outcomes for patients with NSCLC.

Bevacizumab

Bevacizumab is currently the best characterized and only approved VEGF inhibitor in clinical use for NSCLC. Approval of bevacizumab for the treatment of NSCLC was based on the results of 2 phase III trials combining bevacizumab with first-line chemotherapy. In a phase II trial the combination of bevacizumab with cytotoxic chemotherapy increased response rate (RR), time to progression (TTP), and overall survival (OS) compared with chemotherapy alone in patients with advanced nonsquamous NSCLC [24].

In the phase III E4599 registration trial (N = 878) in patients with advanced NSCLC receiving front-line therapy, a significant OS benefit was shown for bevacizumab (15 mg/kg) plus paclitaxel/carboplatin followed by maintenance bevacizumab until disease progression versus standard chemotherapy alone [25]. Significant improvements in OS (12.3 vs 10.3 months; hazard ratio [HR], 0.79; 95% confidence interval [CI], 0.67–0.92; P = 0.003) and progression-free survival (PFS, 6.2 vs 4.5 months; HR, 0.66; 95% CI, 0.57–0.77; P < 0.001) were observed for the bevacizumab arm, as well as an objective RR that was more than twice that of chemotherapy alone (35% vs 15%; P < 0.001) [25]. However, bevacizumab treatment was associated with significant increases in grade ≥3 hematologic toxicities, including neutropenia (25.5% vs 16.8%), febrile neutropenia (5.2% vs 2.0%), and thrombocytopenia (1.6% vs 0.2%), as well as grade ≥3 hypertension (7.0% vs 0.7%), proteinuria (3.1% vs 0%), bleeding (4.4% vs 0.7%), hyponatremia (3.5% vs 1.1%), rash (2.3% vs 0.5%), and headache (3.0% vs 0.5%) over chemotherapy alone (P < 0.05). In addition, 5 of 15 treatment-related deaths in the bevacizumab arm were attributed to pulmonary hemorrhage [25].

In the subsequent AVAiL trial (N = 1,043), bevacizumab was administered at 1 of 2 doses (7.5 mg/kg or 15 mg/kg) and added to a first-line doublet of gemcitabine/cisplatin in chemotherapy-naive patients with advanced NSCLC [26]. Following first-line therapy, patients in the bevacizumab arms continued to receive bevacizumab monotherapy and patients in the placebo arm continued to receive placebo until disease progression, with no crossover to bevacizumab allowed. In addition to patients with squamous tumors, patients with tumors invading or adjacent to major blood vessels were excluded from the trial. The AVAiL trial found that, independent of dose, bevacizumab was associated with significant improvements in the objective RR (20.1%; 34.1%, P < 0.0001; and 30.4%, P = 0.0023 for placebo and bevacizumab 7.5 mg/kg and 15 mg/kg, respectively) and PFS (HR, 0.75; 95% CI, 0.62–0.91; P = 0.003 for 7.5 mg/kg and HR, 0.82; 95% CI, 0.68–0.98; P = 0.03 for 15 mg/kg) [26]. However, in contrast to E4599, no OS benefit was evident in either of the bevacizumab groups [27]. Addition of bevacizumab to chemotherapy in the AVAiL trial modestly increased the incidences of grade ≥3 neutropenia (placebo, 32%; 7.5 mg/kg, 40%; 15 mg/kg, 36%), vomiting (4%; 7%; 9%), hypertension (2%; 6%; 9%), proteinuria (0%; <1%; 1%), and bleeding (2%; 4%; 4%) over chemotherapy alone [26]. The incidences of fatal pulmonary hemorrhage were approximately 1% in each bevacizumab arm and 0.3% in the placebo arm [26].

The phase IV SAiL trial (MO19390, N = 2,212) assessed the efficacy and safety of the addition of bevacizumab to standard chemotherapy regimens, primarily carboplatin doublets (49%) and cisplatin doublets (38%) [28]. Similar efficacy was generally observed across chemotherapy regimens, with the exception of a lower median OS rate for patients who received non-platinum doublets (8.1 months; 95% CI, 5.7–13.0) or monotherapy (9.4 months; 95% CI, 5.3–14.7), who had slightly lower median OS than patients receiving other carboplatin doublet (14.3 months; 95% CI, 13.2–15.6) or cisplatin doublet (14.7 months; 95% CI, 13.7–16.0) regimens. No new safety signals were identified, and grade ≥3 toxicities included thromboembolism (8%), hypertension (6%), bleeding (4%), proteinuria (3%), pulmonary hemorrhage, pulmonary embolism, epistaxis, neutropenia, febrile neutropenia, and deep vein thrombosis (1% each). Thromboembolism (1%) and bleeding (1%) represented the most common cause of treatment-associated deaths [28]. Another observational phase IV clinical trial (ARIES) evaluated the safety and efficacy of bevacizumab plus chemotherapy as first-line therapy in patients with locally advanced nonsquamous NSCLC [29]. Median PFS and OS were 6.7 months and 13.6 months overall (N = 1,970); 6.8 months and 12.6 months among patients ≥70 years of age (n = 650), 5.8 months and 8.1 months among patients with performance status ≥2 (n = 182). Incidence of severe pulmonary hemorrhage (observed in 0.8% overall, 3% of those ≥70 years of age, and 1% of those with performance status ≥2, respectively), grade ≥3 bleeding excluding pulmonary hemorrhage (3%; 3%; 4%), arterial thromboembolism (2%; 3%; 3%), and grade ≥3 central nervous system (CNS) bleeding (0.1%; 0%; 0.5%) were reported. Among patients with CNS metastases (n = 150), median PFS and OS were 6.0 months and 11.7 months, respectively, and no grade ≥3 CNS bleeding was reported in this subgroup.

In addition to these findings, results are also available from a prospective biomarker study of the E4599 trial [30]. Patients with high plasma levels of VEGF were shown to have a higher probability of response to chemotherapy plus bevacizumab compared with chemotherapy alone (33% vs 7%; P = 0.01). However, baseline VEGF levels were not predictive of the OS benefit observed with bevacizumab. In addition, intracellular adhesion molecule-1 (ICAM-1) levels were predictive of response to chemotherapy alone or with bevacizumab and were also prognostic for survival [30].

A number of additional phase II and III clinical trials of bevacizumab have been completed or are ongoing. Emerging results from these trials may support the use of bevacizumab in previously excluded NSCLC patient populations. For example, although bevacizumab is not recommended for patients with treated CNS metastasis due to the risk of CNS bleeding [6], the phase II PASSPORT trial reported no grade ≥2 CNS hemorrhage events among 106 evaluable patients with treated brain metastases [31].

Combination of bevacizumab with other targeted therapies may also have clinical benefit. In the placebo-controlled BeTa phase III trial of bevacizumab plus erlotinib as second-line therapy, no benefit in median OS was observed with the combination versus erlotinib alone (9.3 vs 9.2 months; HR, 0.97; 95% CI, 0.80–1.18; P = 0.76) [32]. The combination numerically improved the objective RR (13% vs 6%) and median PFS (3.4 vs 1.7 months; HR, 0.62; 95% CI, 0.52–0.75), but statistical significance could not be determined because the study prespecified that the primary endpoint be significant before testing secondary endpoints..

Bevacizumab is used as maintenance therapy based on the E4599 and AVAiL trials, and there are combinations with bevacizumab also being evaluated in this setting. The placebo-controlled ATLAS trial (NCT00257608) is currently evaluating bevacizumab plus erlotinib as maintenance therapy following bevacizumab/chemotherapy for advanced NSCLC without progression. Preliminary data from this trial (n = 768) support a significant PFS benefit with bevacizumab and erlotinib versus bevacizumab and placebo (median, 4.8 vs 3.7 months; HR, 0.722; 95% CI, 0.592–0.881; P = 0.0012) [33]. The phase III POINTBREAK study [34], which will evaluate bevacizumab in combination with paclitaxel/carboplatin as induction treatment, followed by bevacizumab/pemetrexed versus bevacizumab as maintenance therapy in patients with advanced NSCLC, is currently recruiting patients. A phase III trial is currently evaluating adjuvant chemotherapy with and without bevacizumab for surgically resected stage IB-IIA NSCLC [35]. Interim safety data showed no unexpected toxicities; however, the addition of bevacizumab to chemotherapy resulted in significantly increased grade 3/4 hypertension (19.6% vs 2.0% with chemotherapy alone; P < 0.001), proteinuria (3.2% vs 0.7%; P = 0.03), abdominal pain (4.6% vs 0.3%; P = 0.001), and overall grade 3/4 toxicity (84.0% vs 68.0%; P < 0.001) [35]. Together, these studies indicate that bevacizumab may represent an effective therapeutic agent for NSCLC in multiple settings and support the rationale for further study of antiangiogenic therapy in this tumor type.

Investigational antiangiogenic agents for NSCLC

In addition to the VEGF pathway, other signaling pathways, such as those mediated by PDGF and FGF, have also been implicated in angiogenesis [36, 37]. These pathways may also play a role in resistance to VEGF-directed treatment [38–41], suggesting that multitargeted TKIs may potentially offer therapeutic benefit in the treatment of advanced NSCLC. Completed and ongoing trials of investigational antiangiogenic agents under evaluation for their therapeutic potential in NSCLC, including various multitargeted TKIs and developmental antibody-based agents, are discussed herein.

Multitargeted antiangiogenic TKIs

Multitargeted TKIs have emerged as potential therapeutics for NSCLC due to the ability of these agents to target multiple pathways involved in angiogenesis. However, efficacy and safety outcomes for these agents vary substantially, and identification of patients most likely to benefit from these agents remains under investigation.

Sorafenib (Nexavar®, Bayer), inhibits VEGFR-2 and −3 and the PDGF receptor (PDGFR)-β and possesses additional antiproliferative potential through inhibition of (v-raf 1 murine leukemia viral oncogene homolog 1) Raf, stem cell factor receptor (c-kit), and fms-like tyrosine kinase 3 (flt-3) [42]. Despite early promise, phase III trials of sorafenib as an adjunct to first-line chemotherapy for advanced NSCLC have failed to demonstrate prolongation of OS relative to chemotherapy alone. The phase III ESCAPE trial (N = 926) of paclitaxel/carboplatin alone or with sorafenib was terminated at interim analysis due to low likelihood of reaching the primary endpoint and revealed an adverse safety signal among patients with squamous tumors [43]. Per protocol amendment, patients with squamous tumors were subsequently excluded from the phase III NExUS trial of gemcitabine/cisplatin alone or with sorafenib (NCT00449033), but efficacy results were consistent with those of ESCAPE, with no OS benefit in the sorafenib arm [44].

Clinical development of sorafenib for advanced NSCLC is currently focused on patients with chemotherapy-refractory disease, a setting in which activity was suggested in prior clinical studies. In a placebo-controlled phase II discontinuation trial in a population of heavily pretreated patients with NSCLC, sorafenib monotherapy was associated with significant prolongation of PFS (3.6 vs 1.9 months with placebo; P = 0.01) and a significantly higher rate of stable disease (29% vs 5% with placebo; P = 0.002). Grade 3 to 4 toxicities in all patients receiving sorafenib prior to randomization to further sorafenib or placebo included rash/hand-foot syndrome (15%) and fatigue (11%) [45]. In addition, a phase III trial (NCT00863746) of sorafenib monotherapy as third-line or fourth-line therapy for advanced NSCLC is currently underway. Phase II randomized studies have also evaluated the potential benefit of adding sorafenib to pemetrexed [46] or erlotinib [47] but have not met prespecified efficacy measures.

Clearly, the development of a predictive biomarker to identify patients likely to benefit or not benefit from antiangiogenic therapy is the next step. Results of a biomarker analysis from a phase II trial of sorafenib plus erlotinib as first-line treatment for advanced NSCLC suggest a potential role for CD133-positive (CD133+) circulating hematopoietic progenitor cells (HPCs) for predicting response and TTP [48]. This study found that baseline levels of CD133+ HPCs were significantly lower among responders versus nonresponders (P = 0.01), and values lower than the median were correlated with prolonged TTP (P = 0.037). However, baseline CD133+ HPCs were not associated with baseline VEGF levels [48].

Biomarkers predictive of sorafenib response were also addressed in the MD Anderson Cancer Center BATTLE (Biomarker-integrated Approaches of Targeted Therapy for Lung Cancer Elimination) program, which was designed to evaluate the role of predictive biomarkers in guiding the selection of targeted therapy for advanced NSCLC [49, 50]. This adapted randomized study screened 11 biomarkers relevant to the EGFR, Ras/Raf, VEGF, and cyclin D1/retinoid X receptor (RXR) pathways in core biopsy specimens. Biomarker results from patients initially enrolled into the umbrella study determined whether patients would enroll in a clinical trial of sorafenib, vandetanib (Zactima™, AstraZeneca), erlotinib, or erlotinib/bexarotene [49]. The disease control rate at 8 weeks (primary endpoint) was 46% across all 4 regimens, ranging from 33% with vandetanib to 58% with sorafenib [49]. These biomarker results will be further evaluated in the upcoming BATTLE-2 trial (NCT01248247). Sorafenib has also been shown to be associated with a high disease control rate in patients with Kirsten ras sarcoma viral oncogene homolog (KRAS) mutation-positive tumors (61% disease control rate at 8 weeks), but significantly lower disease control rates in patients with EGFR mutation-positive tumors (23% vs 64% for EGFR mutation-negative tumors; P = 0.012) or high EGFR polysomy tumors (27% vs 62% for low-polysomy; P = 0.048) [50]. These results suggest that biomarker-based selection of patients most likely to benefit may lead to improved efficacy of sorafenib.

Sunitinib (Sutent®, Pfizer) is a multikinase inhibitor that targets VEGFR-1, -2, and −3 and PDGFR-α/β, as well as c-kit, flt-3, and the rearranged during transfection (RET) receptor. Data from phase II studies support activity of sunitinib monotherapy in patients with chemotherapy-pretreated advanced NSCLC. Socinski and colleagues evaluated 4-weeks-on/2-weeks-off sunitinib dosing in 63 patients pretreated with 1 or 2 prior chemotherapy regimens (which may have included an EGFR TKI). Sunitinib monotherapy was associated with an objective RR of 11.1% (7/63), a rate of stable disease for ≥8 weeks of 28.6% (18/63), and median PFS and OS of 12.0 weeks and 23.4 weeks, respectively. In this study, which excluded patients at high bleeding risk, the most common toxicities were fatigue/asthenia (29%), pain/myalgia (17%), dyspnea (11%), and nausea/vomiting (10%) [51]. Novello and colleagues evaluated continuous dosing of sunitinib in 47 patients pretreated with 1 or 2 prior chemotherapy regimens, with an objective RR of 2.1% (1/47), a rate of stable disease for ≥8 weeks of 23.4% (11/47), and median PFS and OS of 11.9 weeks and 37.1 weeks, respectively. The most common grade 3 to 4 toxicities in this study were fatigue/asthenia (17%), hypertension (9%), and dyspnea (6%) [52]. Sunitinib monotherapy has also been studied in the first-line setting in 63 elderly patients with advanced nonsquamous NSCLC, reporting an overall RR of 5%, disease control rate of 58%, and median PFS of 3.0 months. The most common treatment-related grade 3/4 toxicity was fatigue (18%) and quality of life worsened during the study [53].

A phase III trial (N = 960) has investigated sunitinib as second-line therapy in combination with erlotinib [54]. Addition of sunitinib to erlotinib resulted in a significant increase in PFS (15.5 vs 8.7 weeks for placebo; HR, 0.807; 95% CI, 0.695–0.937; P = 0.0023), but not OS (9.0 vs 8.5 months; HR, 0.922; 95% CI, 0.797–1.067; P = 0.1388). Treatment-related grade ≥3 toxicities included diarrhea (10.4% vs 1.7% for placebo), rash (8.2% vs 2.5%), decreased appetite (3.4% vs 1.3%), fatigue (3.8% vs 1.9%), and nausea (1.3% vs 0.6%). Interim results from a subset analysis of East Asian patients (n = 103) showed significantly longer OS (not reached vs 9.4 months; P = 0.0042). significantly higher RR (38.5% vs 13.7%; P = 0.0083), and numerically longer PFS (31.2 vs 15.2 weeks; HR, 0.723; 95% CI, 0.451–1.161; P = 0.0889) with sunitinib plus erlotinib versus erlotinib alone, with similar toxicity to that reported in the primary analysis [55]. Additional phase III trials are ongoing in patients with advanced NSCLC to determine the role of sunitinib as maintenance therapy (NCT00693992 [CALGB 30607] and NCT01210053).

BIBF 1120 (Boehringer Ingelheim) targets VEGFR-1, -2, and −3, PDGFR-α/β, FGFR-1, -2, and −3, as well as members of the v-src sarcoma viral oncogene homolog (Src) family and flt-3 [56, 57]. BIBF 1120 has been evaluated in a phase II study involving 73 patients with advanced NSCLC, including patients with an Eastern Cooperative Oncology Group Performance Status (ECOG PS) of 0 to 2 [58]. Median OS was 21.9 weeks overall and 37.7 weeks for the 56 patients with an ECOG PS 0–1, with corresponding median PFS of 6.9 weeks and 11.6 weeks, respectively. Grade 3 to 4 toxicities included gastrointestinal complaints (eg, diarrhea [8.2%], nausea [6.8%], abdominal pain [2.7%], vomiting [2.7%]), anorexia (1.4%), fatigue (1.4%), alanine aminotransferase elevation (9.6%), aspartate aminotransferase elevation (1.4%), and gamma glutamyl transferase elevation (4.1%) [58]. Phase III trials have been initiated to evaluate BIBF 1120 as a component of second-line chemotherapy with docetaxel (LUME-Lung 1; NCT00805194) or pemetrexed (LUME-Lung 2; NCT00806819).

Cediranib (Recentin™, AstraZeneca) is a multikinase targeted agent that inhibits VEGFR-1, -2, and −3, c-kit, and PDGFR-β [59]. A phase II/III trial (BR24; N = 296) has evaluated cediranib with paclitaxel/carboplatin as first-line therapy in patients with advanced NSCLC [60]. Interim results from this trial demonstrated a significantly higher response rate with cediranib versus placebo (38% vs 16%; P < 0.0001). During the trial, the original cediranib dose of 45 mg was reduced to 30 mg because of toxicities, but the 30-mg dose was still associated with excessive adverse events, including grade ≥3 hypertension (19% versus 2% with placebo-chemotherapy), grade ≥3 diarrhea (15% versus 2%, respectively). Based on these toxicities, the study was halted to review imbalances in assigned causes of death. In light of the results of BR24, the ongoing phase III BR29 trial was designed to evaluate cediranib 20 mg plus paclitaxel/carboplatin (NCT00795340) but was recently reported to have closed [61].

Cediranib is also under evaluation in combination with pemetrexed for treatment of NSCLC [62]. Results from this study revealed a RR of 29% (95% CI, 17%–45%) and a disease control rate of 74% (95% CI, 58%–85%) for 38 evaluable patients. Grade 3/4 toxicities for this regimen included neutropenia (14%), febrile neutropenia (5%), fatigue (22%), diarrhea (14%), and infection (8%).

Motesanib (Amgen) inhibits VEGFR-1, -2, and −3, PDGFR, c-kit, and RET [63]. A phase II trial (N = 186) has evaluated motesanib (125 mg once daily in arm A or 75 mg twice daily in arm B) plus paclitaxel/carboplatin versus bevacizumab plus paclitaxel/carboplatin (arm C) as first-line therapy for advanced nonsquamous NSCLC [64]. The objective RR was 30%, 23%, and 37%, median PFS was 7.7 months, 5.8 months, and 8.3 months, and median OS was 14.0 months, 12.8 months, and 14.0 months for arms A, B, and C, respectively. Incidence of grade ≥3 adverse events was higher in arms A (71%) and B (79%) than in arm C (63%). The most common grade 3/4 toxicities occurring in ≥20% of patients and in more than 1 treatment arm included diarrhea (19%, 11%, and 3% in arms A, B, and C, respectively), dehydration (17%, 11%, and 3%), and fatigue (17%, 5%, and 8%). Clinical development of motesanib in advanced NSCLC has followed a pattern similar to that for bevacizumab and sorafenib with respect to histology, with temporary suspension of the phase III MONET1 trial of paclitaxel/carboplatin alone or with motesanib in light of increased hemoptysis and mortality among the squamous histology subset [65]. The trial reopened and completed accrual (N = 1,090), though enrollment was restricted to patients with nonsquamous tumors. Preliminary results reported significantly longer PFS (5.6 vs 5.4 months; HR, 0.785; 95% CI, 0.684–0.901; P = 0.0006) and higher RR (40% vs 26%; P < 0.0001) with chemotherapy plus motesanib versus chemotherapy plus placebo; however, the study did not meet its primary endpoint of improving OS (13.0 vs 11.0 months; HR, 0.897; 95% CI, 0.776–1.035; P = 0.137) [66, 67].

Axitinib (Pfizer) inhibits VEGFR-1, -2, and −3, PDGFR-β, and c-kit [68]. A phase II trial of axitinib monotherapy has been conducted in 32 patients with advanced NSCLC, including chemotherapy-naive or chemotherapy-pretreated patients [69]. In this trial, 3 patients (9%) had a partial response and 10 (31%) had stable disease for ≥16 weeks. Median PFS was 4.9 months overall and 9.2 months in the first-line axitinib subset, with median OS of 14.8 months for all patients (14.8 months for previously untreated patients; 15.5 months for previously treated patients). Grade 3 fatigue and hypertension were reported in 22% and 9% of patients, respectively. Current early phase trials are investigating the combination of axitinib with chemotherapy for advanced NSCLC, including a phase I/II trial in nonsquamous tumors (NCT00768755) and a phase II trial in squamous tumors (NCT00735904).

Pazopanib (GlaxoSmithKline) is a multikinase inhibitor that targets VEGFR-1, -2, and −3, PDGFR-α/β, and c-kit [70, 71]. A phase II trial of pazopanib monotherapy as preoperative therapy for stage I/II NSCLC revealed that 3 of 35 patients (9%) had a partial response, and the most common grade ≥3 toxicity was alanine transaminase elevation (6%) [72]. A biomarker analysis of this trial showed that 11 cytokine and angiogenic factors were significantly impacted (reduced or increased) by pazopanib [73]. Serum VEGFR-2 levels showed the most pronounced reduction, and changes in VEGFR-2 and interleukin-4 correlated with tumor shrinkage. A phase II/III trial is currently evaluating pazopanib as adjuvant monotherapy for surgically resected stage I NSCLC (NCT00775307) and a second phase II/III trial is evaluating pazopanib monotherapy in patients with advanced NSCLC as second-line therapy (NCT01208064). Additional phase II trials of pazopanib as monotherapy (NCT01049776) and as a component of combination regimens (with chemotherapy [NCT00866528] or an EGFR inhibitor [NCT01027598]) for advanced NSCLC are also underway.

Antiangiogenic antibody-based agents

In addition to multitargeted TKIs, novel antibody-based therapies have also been studied as antiangiogenic agents with mixed results. Aflibercept or AV0005 (VEGF Trap, Regeneron) is a peptide-antibody fusion containing portions of human VEGFR-1 and −2 [74]. A phase II study (N = 98) that evaluated aflibercept in patients with platinum- and erlotinib-resistant adenocarcinoma of the lung demonstrated an overall RR of 2% and a median PFS and OS of 2.7 months and 6.2 months, respectively [75]. The most common grade ≥3 toxicities included proteinuria (10%), hypertension (23%), and dyspnea (21%). Two treatment-related fatal adverse events, both grade 5 hemoptysis, occurred on study. In a recently-completed phase III trial (NCT00532155), aflibercept was evaluated in combination with docetaxel as second-line therapy for advanced NSCLC. Initial results suggested that this trial did not meet its primary endpoint of improving OS versus docetaxel (HR, 1.01; CI, 0.868–1.174). However, both PFS (HR, 0.82; CI, 0.716–0.937) and ORR (23.3% for aflibercept vs 8.9% for placebo) were improved in patients treated with aflibercept versus placebo. The most common grade 3/4 toxicities in patients treated with aflibercept were fatigue, stomatitis, disease progression, hypertension, febrile neutropenia, dyspnea, neutropenia, and asthenia [76]. In a phase III clinical trial (VITAL), aflibercept/docetaxel and placebo/docetaxel were evaluated as second-line therapy (1 prior platinum-based regimen) in advanced, nonsquamous NSCLC [77]. Aflibercept was associated with improvement in PFS (HR, 0.82; 95% CI, 0.72–0.94; P = 0.0035) and RR (23.3% vs 8.9%; P < 0.0001); however, the primary endpoint of improved OS was not met (HR, 1.01; 95.1% CI, 0.87–1.17; P = 0.898). Stomatitis, weight decrease, hypertension, epistaxis, and dysphonia were reported as occurring at greater than 10% higher incidence with aflibercept compared with placebo.

Ramucirumab or IMC-1121B (ImClone Systems, Inc.) is an anti-VEGFR-2 monoclonal antibody [78]. Preliminary results are available from an open-label phase II study of ramucirumab plus first-line paclitaxel/carboplatin in patients with advanced NSCLC [79]. For 15 evaluable patients, the preliminary objective RR was 67%, including 1 complete response, and median PFS was 5.7 months. Two potentially ramucirumab-related serious adverse events were reported, grade 4 febrile neutropenia and grade 2 pneumothorax. A phase II trial (NCT01160744) of platinum-based chemotherapy (pemetrexed plus carboplatin/cisplatin or gemcitabine plus carboplatin/cisplatin) with or without ramucirumab in patients with previously untreated advanced NSCLC is currently recruiting patients, and a phase III trial (NCT01168973) of docetaxel plus ramucirumab or placebo as second-line therapy for stage IV NSCLC is planned.

Conclusions

The VEGF axis continues to be a key therapeutic target for NSCLC-directed treatments, with substantial interest in simultaneous targeting of other signaling pathways in an effort to improve clinical outcomes. The anti-VEGF monoclonal antibody bevacizumab is currently the only approved and recommended antiangiogenic agent for use in patients with advanced NSCLC, although a number of other antiangiogenic strategies including antibodies and multitargeted TKIs are under clinical evaluation in various settings in NSCLC. Selection of patients most likely to benefit from antiangiogenic therapies may lead to improved outcomes, and toward this end biomarker evaluations are becoming a more prevalent component of clinical trials of antiangiogenic agents. Advances are being made in this regard (as illustrated by the BATTLE program) with the ultimate goal of identifying predictive factors to truly individualize therapy for patients.