|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 23, No 6 (February 20), 2005: pp. 1295-1311 © 2005 American Society of Clinical Oncology. DOI: 10.1200/JCO.2005.10.022
Combination of Antiangiogenic Therapy With Other Anticancer Therapies: Results, Challenges, and Open QuestionsFrom the Division of Medical Oncology, S. Filippo Neri Hospital, Rome, Italy; and Genzyme Corporation, Framingham, MA Address reprint requests to Giampietro Gasparini, MD, Division of Medical Oncology, Azienda Complesso Ospedaliero S. Filippo Neri, Via C. Martinotti, 20 00135 Rome, Italy; e-mail: gasparini.oncology{at}tiscalinet.it
Angiogenesis is necessary for tumor growth. Drug discovery efforts have identified several potential therapeutic targets on endothelial cells and selective inhibitors capable of slowing tumor growth or producing tumor regression by blocking angiogenesis in in vivo tumor models. Certain antiangiogenic therapeutics increase the activity of cytotoxic anticancer treatments in preclinical models. More than 75 antiangiogenic compounds have entered clinical trials. Most of the early clinical testing was conducted in patients with advanced disease resistant to standard therapies. Several phase III trials have been undertaken to compare the efficacy of standard chemotherapy versus the same in combination with an experimental angiogenesis inhibitor. Preliminary results of the clinical studies suggest that single-agent antiangiogenic therapy is poorly active in advanced tumors. Although some of the results of combination trials are controversial, recent positive outcomes with an antivascular endothelial growth factor antibody combined with chemotherapy as front-line therapy of metastatic colorectal cancer have renewed enthusiasm for this therapeutic strategy. This article presents an overview of experimental and clinical studies of combined therapy with antiangiogenic agents and highlights the challenges related to the appropriate strategies for selection of the patients, study design, and choice of proper end points for preclinical and clinical studies using these agents.
A tumor requires angiogenesis to grow beyond 1 to 2 mm3 in size and to develop metastasis.1 Angiogenesis may occur as a result of genetic changes or be triggered by local alterations such as hypoxia, glucose deprivation, and oxidative and mechanical stresses.1 Several genetic alterations regulate angiogenesis: RAS, MYC, RAF, HER-2/neu, c-JUN, and SRC upregulate vascular endothelial growth factor (VEGF) or downregulate thrombospondin-1 (TSP-1), a naturally occurring angiogenesis inhibitor (AI).2 Studies of angiogenesis documented for the first time that stromal components also contribute to tumor development and progression1,2 (Fig 1).
Angiogenesis can occur by sprouting, nonsprouting, or intussusception.1,2 The abnormality of tumor vasculature and the value of working with endothelial cells (ECs) isolated from solid tumors have been recognized.3 There are at least four potential mechanisms leading to tumor angiogenesis: secretion by tumor and/or stromal cells of proangiogenic factors, co-option of pre-existing vasculature, vasculogenesis from circulating AC133+/CD34+ endothelial precursor cells (EPCs), and vascular mimicry, that is the formation of vascular channels by tumor cells.1–4 Angiogenesis, the process of formation of new vessels arising from sprouts of existing vessels is distinct from vasculogenesis, vessels arising from EPCs (angioblasts).5 Asahara et al6 isolated angioblasts from human peripheral blood undergoing postnatal vasculogenesis and pathologic neovascularization. Studies in allogeneic bone marrow transplant recipients confirmed that circulating EPCs in peripheral blood originate from bone marrow.7 Recent studies have formally tied circulating EPCs to tumor angiogenesis.8 In mice bearing human breast carcinoma xenografts, both circulating and intratumor proliferating EPCs have been detected.9 NOD/SCID mice transplanted with human bone marrow and bearing human Namalwa or Granta 519 Burkitt's lymphoma xenografts had a seven-fold increase in circulating EPCs compared with non–tumor-bearing mice.10
Physiologic angiogenesis is tightly regulated by pro- and antiendothelial growth factors and occurs by a series of complex and interrelated steps.1,2 Proangiogenic growth factors, such as VEGF, fibroblast growth factors, and platelet-derived growth factor, are released into the microenvironment by malignant, inflammatory, and other stromal cells in response to various stimuli. The released growth factors activate local ECs and EPCs from bone marrow that enter circulation to generate new blood vessels.1,2 Activated ECs, as well as local stromal cells and EPCs, secrete several enzymes, including metalloproteinases (MMPs) that break down extracellular matrix and allow ECs to invade surrounding tissue, proliferate in response to growth factors, and migrate toward the malignant stimulus.1,11 Plasmin generated by the urokinase plasminogen activator, regulated by the urokinase plasminogen activator receptor and plasminogen activator inhibitor-1, is a key mediator of these processes.12 The migration of both ECs and EPCs is regulated by adhesion molecules. The receptors of integrins
Tumor endothelium is phenotypically different from normal vessels, and it is characterized by increased fenestration and leakiness, abnormal architecture with arteriovenous shunts, multiple loops, and fan and spiral motifs.1,2,15 Tumor ECs divide up to 50 times more frequently and express higher levels of specific cell surface molecules, such as integrin  vß3, E-selectin, endoglin, endosialin, and VEGFRs, than normal ECs.2,16,17 Transcriptional profiles of tumor ECs show significant differences in gene expression (angiomics) compared with ECs isolated from the corresponding normal tissue.3 Tumor ECs express surface receptors and secrete factors that sustain their own growth (by autocrine pathways) as well as the growth of tumor parenchyma (by paracrine pathways). A mutual stimulation occurs between the stroma and tumor parenchyma that sustains malignant growth, progression, and metastasis.2,18 Therefore, tumor cell/vascular system should be considered a functional unit regarding tumor growth. The tissue oxygen diffusion limit is 100 to 200 µm, corresponding to three to five cell layers around a blood vessel.19 AIs are biologic response modifiers acting by cytostatic mechanisms targeting activated ECs and EPCs that inhibit tumor growth indirectly by blocking the formation of new vasculature.3,19 Antiangiogenic therapy may represent a new promising anticancer therapeutic strategy. First, like most normal tissues, normal ECs are quiescent under physiologic conditions, whereas tumor ECs and EPCs are actively proliferating and with an angiogenic phenotype.2,18 Consequently, AIs tend to have moderate toxicity, because angiogenesis is infrequent in adults, except during inflammation, ovulation, pregnancy, wound healing, and ischemia.2 ECs are quiescent in hypoxic and necrotic regions, whereas in the areas of progressive malignant disease, they are active, proliferating, and more sensitive to therapy.1 Second, ECs are genetically stable, although under the influence of the malignant environment, tumor ECs express an abnormal phenotype.1,2,20 Third, tumor endothelium expresses high levels of specific molecular targets and antigens that may be targeted by selective inhibitors.2 Fourth, ECs are readily accessible to selective AIs given by systemic administration.2 Finally, like other anticancer therapies, some AIs may have synergistic therapeutic effects in combination with conventional cytotoxic therapies.21
Antiangiogenic therapy targets genetically stable cells, and direct evidence of acquired resistance has not yet been clearly demonstrated in preclinical studies and likewise in the treatment of certain nonmalignant human tumors. However, there is some evidence of gradual loss of activity of AIs, especially when they are administered in monotherapy.22,23 Postulated mechanisms of acquired resistance to AIs include the following: multiplicity of tumor- and stromal cell–secreted growth factors,2,24 antiapoptotic/prosurvival functions of tumor ECs,25,26 epigenetic changes in tumor ECs,27 vascular channel assembly by tumor cells,28 and genetic alterations of tumor cells, such as p53 inactivation or mutation and changes in hypoxia-inducible factor-1 alpha (HIF-1 ) or survivin pathways.2,22,27,29 Wild-type p53 protein inhibits angiogenesis through upregulation of TSP-1, whereas inactivation of wild-type p53 reduces ECs and tumor cells susceptibility to apoptosis.29 p53–/– tumor cells had a survival advantage under hypoxic conditions compared with the p53+/+ ones.22 Alteration of the HIF-1 pathway as a result of upstream oncogenic changes (eg, activated ras, src, and HER-2) increases cellular response to hypoxia and enhances the survival of tumor cells under stress conditions.2,30,31 The role of survivin in resistance of ECs mediated by VEGF was studied by Tran et al,27 who found that VEGF-upregulated survivin ensured the integrity of microtubule dynamics with drug-protective effects. Strategies to improve the antitumor efficacy by blocking angiogenesis include combination of antiangiogenic therapy with the following: cytotoxic therapies, targeted therapies directed toward malignant cells, inhibitors of oncogene-mediated signal transduction directed toward the malignant cells, and cytotoxic agents active also in hypoxic conditions. Vascular targeting is an alternative strategy to obtain vessel obstruction and rapid necrosis of tumor mass.2
AIs can be classified as direct, indirect, or mixed inhibitors. Direct AIs target the ECs involved in the malignant disease by inhibiting their ability to proliferate, migrate, or form new blood vessels. The action of direct AIs may be independent of the type of cancer cell with low probability of acquired resistance. Indirect AIs interfere with production of angiogenic factors by malignant cells, stromal cells, and inflammatory cells or with extracellular processes. Resistance to indirect AIs may be more likely than resistance to direct AIs because they target genetically unstable tumor cells.32 Mixed AIs, such as multitargeting kinase inhibitors, epidermal growth factor receptor (EGFR) inhibitors or neutralizing agents, protein kinase C inhibitors, and others, as well as cytotoxic anticancer agents, target both tumor ECs and malignant cells.
Folkman32 proposed a classification system based on the efficacy of AIs in preclinical tumor models: first-generation AIs, such as interferons, TNP-470, thalidomide, and matrix metalloproteinases inhibitors (MMPIs), only slow tumor growth; second-generation AIs, such as anti-VEGF and anti-integrin However, the results of experimental studies on endostatin are controversial in part because of the instability of the molecule and the production from different laboratories of lots with diverse activity.33 A recent Italian study suggests that endostatin sequence exhibits peptides with both angiosuppressive and angiostimulatory effects. Two fragments with high angioinhibitory activity have been sequenced.34
Human endostatin (h-Endostatin) is the NC1 domain of the alpha-1 chain of type XVIII collagen, and it is cleaved by proteases such as elastase and cathepsins. It circulates in the blood at concentrations of 20 to 35 ng/mol. h-Endostatin inhibits EC migration, whereas mouse endostatin blocks migration and causes Gap1 EC cycle arrest. The reason for this difference is presently unknown. h-Endostatin binds to several cell surface proteins, including heparan sulfate proteoglycans, glypicans, VEGFR2, and integrins. Although many different intracellular pathways have been identified as possible mediators of h-Endostatin action, a functional receptor has not been identified.33 Recently, Sudhakar et al33 demonstrated that h-Endostatin binds AIs can also be classified by the mechanism of action: inhibitors of angiogenic factors secretion, inhibitors of EC intracellular signaling transduction, inhibitors of EC proliferation, inhibitors of MMPs, agents cytotoxic toward ECs, and inhibitors of mobilization of EPCs from bone marrow.36
Compelling experimental data suggest that selective inhibitors block tumor angiogenesis with regression of many tumor types.18 The first generation of AIs was identified using specific strategies and assays that used normal, mature ECs. Specific cell-based assays of mature ECs, such as human umbilical vein endothelial cells or human microvascular endothelial cells, have been used for evaluation of proliferation, migration, invasion, and ability to form tube-like capillary structures on gelatinized disks or matrigel.37 Through transcriptional analysis, the recognized immature phenotype of tumor ECs suggests the need for improved cell-based assays in the field.1,3 As initial step, the activity of potential AIs was tested using assays focused on normal neoangiogenesis in a variety of models, including inhibition of the chicken chorioallantoic membrane neovascularization,38 inhibition of vascularization of a matrigel plug containing angiogenic factors implanted in a mouse that can be quantified histologically,39 and inhibition of vascular growth toward an angiogenic stimulus implanted in the rat or mouse cornea.40 It has not been established whether these assays involving developing vasculature in the embryo or the neovascularization induced by an angiogenic stimulus are really comparable with in vivo tumor angiogenesis. A second generation of matrigel plug assay uses human EPCs suspended in the matrigel before the implant onto the animal.2 It is possible to observe the effects of blocking angiogenesis on tumor transplanted onto animals.41 The activity of AIs on the primary tumor as well as on metastasis is usually evaluated by comparing the diameter and number of the lesions of the treated animals versus those receiving placebo.42 A major drawback of these models is that many experiments have not been performed in orthotopic models and that the tested vasculature is murine. Therefore, the EC molecular targets are the mouse homologs of the desired human protein targets. Although the mouse protein and the human protein targets often have high homology, an incomplete cross-reactivity of antibodies may occur in certain circumstances. Several approaches of incorporating human EC target molecules in mouse include development of transgenic animals, transplantation of immunodeficient mice with human bone marrow, transfusion of immunodeficient mice with human EPCs, and transplantation of human foreskin onto immunodeficient mice.
Certain studies conducted in preclinical tumor models have documented advantages of combining AIs with cytotoxic chemotherapeutic agents or radiation therapy. These combined regimens produced additive or synergistic antitumor activity.43–45 Potentiation of the therapeutic effects with combined regimens may be related to increased access into the tumor mass of cytotoxic drugs or to enhanced oxygen pressure, as a result of the enhanced permeability induced by AIs.46,47 The greater-than-additive therapeutic effects may result from indirect effects on tumor ECs in addition to direct effects on tumor cells. The multitargeted kinase inhibitor SU11248 blocks the activity of receptor tyrosine kinases located on both ECs and malignant cells.48 Combinations of indirect AIs with radiation therapy can block tumor growth by inhibiting VEGF secretion stimulated by hypoxia.49 In animals bearing Lewis lung carcinoma, the seminal studies by Teicher et al21,46,47 demonstrated that TNP-470,50 minocycline, suramin, and genistein, alone or in two-agent combinations with cytotoxic agents and radiation therapy, enhanced the regression of primary subcutaneous tumors and reduced the number and size of lung metastases.51–54 There were no clear benefits of using combinations of different AIs, and AIs were more effective in combination with cytotoxic therapies when used as two-agent combinations rather than as single agents.21 In an orthotopic animal model of transitional cell carcinoma, docetaxel, administered before TNP-470, significantly increased the complete response rate of nonestablished and established tumors compared with either compound alone.54 The combined treatment inhibited angiogenesis by upregulation of basic fibroblast growth factor (bFGF) and MMP-9 and enhanced apoptosis, without altering the expression of VEGF, interleukin 8, MMP-2, and E-cadherin. Combinations of TNP-470 with various cytotoxic chemotherapeutic agents, such as paclitaxel and carboplatin in non–small-cell lung cancer (NSCLC) and breast cancer models, paclitaxel in NSCLC, cisplatin in liver metastasis of human pancreatic cancer, and fluorouracil in liver metastasis of colorectal cancer, produced additive or synergistic antitumor activity.55–57 Inoue et al58,59 found that combined regimens with the monoclonal antibody C225, which blocks EGFR function, or the rat monoclonal antibody DC101, which blocks VEGFR2 function, with paclitaxel had enhanced antitumor activity through inhibition of both angiogenesis and induction of apoptosis. Using male nude mice implanted with PC3-MM2 cells in the tibia, Kim et al60 found that the combination of oral administration of PKI 166, a selective EGFR tyrosine kinase inhibitor, and low-dose paclitaxel reduced the incidence and size of bone metastasis from prostate cancer and inhibited EGFR phosphorylation on tumor cells and ECs with enhanced apoptosis. Administration of angiostatin or endostatin along with cytotoxics produced marked antitumor effects in an ovarian carcinoma model and the RIPTag transgenic mouse pancreatic adenocarcinoma model.47,61 Some chemotherapeutic agents, such as camptothecin analogs, vinca alkaloids, and taxanes, may have antiangiogenic effects at lower concentrations than those frequently used to kill malignant cells. This has given rise to the development of regimens that apply frequent dosing of cytotoxic agents with the possibility of a predominant effect in the vascular compartment (metronomic chemotherapy), although it is possible that effects of the low doses may also occur in the tumor as well as vascular compartment.62 However, despite interesting antiangiogenic results in cell-based culture assays, only a few cytotoxic agents retain antiangiogenic activity in vivo.62 The most promising antiangiogenic chemotherapeutic agents identified so far are cyclophosphamide, vinblastine, paclitaxel, and docetaxel.62 Studies using subcutaneously implanted tumors in mice documented the antiangiogenic activity of cytotoxic chemotherapeutic agents when administered continuously at low doses.62,63 Broder et al63 administering a combined regimen with cyclophosphamide in the drinking water by a low-dose metronomic schedule and TNP-470, eradicated Lewis lung carcinoma, a cyclophosphamide exquisitely sensitive tumor model, in the majority of treated mice. Other potential mechanisms by which AIs may enhance activity of chemotherapy are listed in Table 1. Similar results have been reported by Clement et al,64 who combined continuous low doses of vinblastine with a rat VEGFR2-neutralizing monoclonal antibody. In vitro synergistic antiangiogenic activity was reported for docetaxel and a recombinant humanized monoclonal antibody directed toward VEGF or 2-methoxyestradiol. Docetaxel inhibited EC migration and proliferation with a concentration that inhibits 50% of 10 ppm, which is similar to its cytotoxic concentration that inhibits 50% against cancer cells in culture.65
SU5416 was under development as a selective inhibitor of VEGFR2 (Flk-1, KDR) kinase activity; SU6668 and SU11248 are under development as broad-spectrum receptor kinase inhibitors being able to block VEGFR2, bFGF factor receptors, and platelet-derived growth factor receptor kinase activities.66,67 Early in vivo work with SU5416 suffered from the use of dimethyl sulfoxide as a vehicle for the compound administered by intraperitoneal injection in mice.68 Gang et al69 found that SU5416 increases the sensitivity of murine B16 melanoma and murine GL261 glioma to radiation therapy. SU5416 and SU6668 have been tested as single agents or in combination with fractionated radiation therapy in C3H mice bearing SCC VII squamous carcinomas.70 Like STI571, SU5416, SU6668, and SU11248 inhibit c-kit (KIT), the stem-cell factor receptor tyrosine kinase, and FLT3.66,67,71 C-kit is a key factor for development of normal hematopoietic cells and has a functional role in acute myeloid leukemia. The potential therapeutic utility of SU11248, alone or in combination with docetaxel, fluorouracil, or doxorubicin, was evaluated in different models of breast carcinoma: MMTV-v-Ha-ras transgenic mouse, DMBA carcinogen-induced rat mammary cancer, MX-1 human breast carcinoma subcutaneously implanted xenograft, and MDA-MB-435 human breast carcinoma subline, and in the 435/HAL-Luc line, selected from bone metastases. The combined regimens resulted in longer survival times than either single chemotherapeutic agent or SU11248 alone.66,67 Other small molecule tyrosine kinase inhibitors showing promising activity in early clinical trial include PTK787/ZK222584 and ZD6474.72–74 Daily oral treatment with PTK787/ZK222584 resulted in a significant decrease in primary murine renal cell carcinoma grown in the surrenal capsule of Balb/c mice. The occurrence of lung metastases was reduced 98% and 78% on days 14 and 21, respectively, and development of lymph node metastases was delayed.73 Protein kinase C isoforms are involved in the signaling transduction pathways that regulate cell cycle, apoptosis, angiogenesis, differentiation, invasiveness, senescence, and drug efflux.75,76 Nude mice bearing human SW2 small-cell lung carcinoma subcutaneous xenografts treated with LY317615, a potent and selective inhibitor of protein kinase Cß, show a dose-dependent decrease in tumor microvessel density.77 Plasma VEGF levels in LY317615-treated SW2-bearing animals were significantly lower as compared with the control group.78 VEGF levels in the control Caki-1 renal cell carcinoma–bearing nude mice treated with LY317615 remained suppressed throughout all the treatment period (d21–39) and until day 53, when the experiment was terminated.78 Treatment of SW2 small-cell lung carcinoma bearing nude mice with paclitaxel followed by LY317615 resulted in more than 60 days of tumor growth delay and a 2.5-fold increased duration of tumor response. The antitumor activity of LY317615 alone and in combination with cytotoxic agents has been explored in several human tumor xenografts in nude mice.77,79 In most tumor models, LY317615, as a single agent, induced tumor growth delay. The combined schedules suggested higher activity and LY317615 has currently completed phase I clinical studies.80 A major strategy to inhibit VEGF signaling pathway consists of VEGF neutralizing monoclonal antibodies.81–83 Bevacizumab, a recombinant humanized anti-VEGF monoclonal antibody, is showing promise in clinical trial.84 Indeed, preclinical studies have shown that the antitumor activity of some cytotoxic agents is potentiated by cyclo-oxygenase 2 (COX-2) inhibitors.85
More than 75 AIs entered clinical evaluation in cancer patients. At least 12 agents entered or completed phase III trials.2 Based on the positive results of the phase III clinical trial AVF 2107, bevacizumab has been the first antiangiogenic compound approved by the US Food and Drug Administration with fluoropyrimidine-combined regimens in metastatic colorectal cancer on February 26, 2004.86 Other AIs evaluated in clinical trials alone and in combination regimens include kinase inhibitors, MMPIs, natural inhibitors, COX-2 inhibitors, and thalidomide analogs.
Bevacizumab (Avastin; Genentech Inc, South San Francisco, CA) was tested in phase I studies in combination with chemotherapy and showed a good safety profile.82,83 In phase II studies (two randomized studies and one nonrandomized study), bevacizumab combined with chemotherapy was evaluated in advanced colon cancer,84,86 stage IIIB/IV NSCLC,87 advanced breast cancer,88 and metastatic renal cancer.89 The combined regimens resulted in increased response rate and prolonged time to progression (TTP) compared with chemotherapy alone. In 99 patients with stage IIIB or IV NSCLC,87 carboplatin/paclitaxel chemotherapy was compared with the same regimen plus bevacizumab at 7.5 mg/kg or 15 mg/kg. In the high-dose group, the response rate was increased by approximately 10%, and TTP was prolonged for 3 months compared with carboplatin/paclitaxel alone. However, six patients developed severe hemoptysis (four episodes were fatal), four of whom had centrally located squamous cell cancer. Bevacizumab is now under testing in phase III studies in NSCLC in patients with nonsquamous cell histology. In 104 previously untreated patients with metastatic colorectal cancer, the combination of bevacizumab (5 mg/kg every 2 weeks) with fluorouracil/leucovorin, resulted in higher response rate, longer TTP, and increased median survival compared with fluorouracil/leucovorin alone.84 Thrombosis was the most significant adverse event and was fatal in one patient. Hypertension, proteinuria, epistaxis, headache, rash, and chills were other reversible side effects. Two phase III trials are ongoing in colorectal cancer. The first compared fluorouracil/leucovorin/irinotecan (n = 412 patients) with fluorouracil/leucovorin/irinotecan/bevacizumab (n = 403 patients) or fluorouracil/leucovorin/bevacizumab as front-line therapy, followed by irinotecan as second-line therapy. The addition of bevacizumab to bolus fluorouracil/leucovorin/irinotecan, even in patients not selected for the expression of the target, resulted in increased survival (20.3 months v 15.6 months; P < .00003), progression-free survival (10.6 months v 6.2 months), response rate (44.9% v 34.7%), and duration of response (10.4 months v 7.1 months) as compared with chemotherapy alone.86 Grade 3 hypertension occurred in 11% of patients receiving bevacizumab, compared with 2.3% of the patients receiving only chemotherapy. Gastrointestinal tract perforation occurred in six patients receiving bevacizumab with one death, two discontinued therapies, and three interrupted therapies. The second phase III colorectal cancer trial compares fluorouracil/leucovorin (Roswell Park regimen) with fluorouracil/leucovorin/bevacizumab in patients who are not candidates for first-line irinotecan. An Eastern Cooperative Oncology Group–sponsored trial (E3200) is testing bevacizumab versus bevacizumab plus fluorouracil/leucovorin/oxaliplatin in patients with progressive disease after previous chemotherapy with fluorouracil/leucovorin/irinotecan. In a randomized, double-blind, phase II trial, bevacizumab, at doses of 3 and 10 mg/kg given every 2 weeks, was compared with placebo in patients with metastatic renal cell carcinoma who experienced disease progression after interleukin-2. The interim analysis showed a significant prolongation of TTP in the high-dose antibody group as compared with the placebo group (P < .001) and a small difference, of borderline significance, between the TTP in the low-dose antibody group and that in the placebo group (P = .053). There were no significant differences in overall survival (P > .20 for all comparisons), probably related to the permitted cross-over from placebo to antibody treatment. Minimal toxic effects were reported, with transient hypertension and asymptomatic proteinuria.89 Several phase III trials of combination of bevacizumab with chemotherapy are ongoing in other tumor types (http://www.nci.nih.gov/clinicaltrials). No clinical data are available regarding DC101, soluble VEGFR1 (Flt-1), and dominant-negative VEGFR2 (Flk-1/KDR). Several small molecules that selectively block phosphorylation of VEGF receptors entered clinical evaluation. The first tested in humans was SU5416 (semoxinal).90–93 In untreated patients with metastatic colorectal cancer, SU5416 was administered in combination with fluorouracil/leucovorin (Roswell Park or Mayo Clinic regimens), at two different dose levels, 85 and 145 mg/m2 twice weekly. The toxicity observed was that expected for the fluorouracil/leucovorin regimen (mucositis), with only few patients reporting mild headache. Six patients achieved a major objective response, and another nine patients had durable stable disease.90 In a randomized, international, multicenter prospective phase III trial in untreated metastatic colorectal cancer patients, SU5416 was administered with fluorouracil/leucovorin (Roswell Park regimen) compared with fluorouracil/leucovorin alone. The final analysis after the enrollment of 737 patients indicated no improvement of clinical outcome in the SU5416 arm.2 The combination of SU5416 at the dose of 145 mg/m2 biweekly with cisplatin and gemcitabine in patients with advanced solid tumors was associated with a surprisingly high toxicity, resulting in a severe rate of thromboembolic events.91 SU5416 may also be useful for the treatment of patients with von Hippel-Lindau syndrome,94 but, taking into account the negative results of phase II/III clinical trials, the compound is no longer in clinical development. SU11248, a multitargeted receptor tyrosine kinase inhibitor, blocks the kinase activity of VEGFR2, platelet-derived growth factor receptor, KIT, and Flt3.95–100 In four phase I clinical studies, decreased levels of circulated VEGF were observed in treated patients.95,96,98,99 Responses were observed in two phase I studies, especially in renal cell carcinoma, neuroendocrine tumor, and thyroid cancer, and nearly 50% of patients had stable disease.95,96 In a phase I study of gastrointestinal stromal tumors in patients resistant to imatinib (Gleevec; Novartis Pharmaceuticals Corp, East Hanover, NJ), five of the 32 treated cases had measurable responses, and approximately 60% of the patients had stable disease for more than 4 months.97,98 Clinical phase I studies of PTK787/ZK222584 alone or in combination with cytotoxic agents have been conducted in patients with metastatic colorectal cancer, renal cell carcinoma, brain tumors, and acute myelogenous leukemia or myelodysplastic syndrome.101–106 The once per day oral dose of PTK787/ZK222584 tolerated in most combination regimens was 1,200 to 1,250 mg/d. International multicenter phase III trials are comparing the bolus fluorouracil plus leucovorin, oxaliplatin, and infused fluorouracil (FOLFOX-4) regimens versus the same regimen plus PTK787/ZK222584 in previously untreated or treated patients with advanced colorectal cancer. The phase I study of ZD6474 concluded that 100 to 300 mg/d would be the appropriate dose range for phase II testing.107 Marimastat, a synthetic inhibitor of MMP-1, -2, -3, -7, and -9, was the first orally bioavailable MMPI tested in the clinic.108–116 Marimastat was evaluated in phase I and II studies at doses from 2 to 100 mg bid, either alone or in combination with chemotherapy.108–112,116 The dose-limiting toxicity of marimastat is musculoskeletal disorders at doses of 10 mg or greater twice daily. In a phase III trial, 414 patients with advanced pancreatic cancer were randomly assigned to receive marimastat (5 mg v 10 mg v 25 mg twice a day) or gemcitabine. Forty-four percent of the patients treated with marimastat experienced musculoskeletal toxicity versus 12% of those who received gemcitabine. Survival was not significantly different in the two arms, whereas a better progression-free survival was reported in the gemcitabine group.109 A large randomized study of marimastat as maintenance therapy of small-cell lung cancer was conducted.111 Among the 532 eligible patients (266 patients receiving marimastat and 266 patients receiving placebo), the median TTP for patients receiving marimastat was 4.3 months compared with 4.4 months for placebo (P = .81). Median survival for marimastat and placebo was 9.3 months and 9.7 months, respectively (P = .90). Toxicity was generally limited to musculoskeletal symptoms (18% grade 3/4 for marimastat). Patients receiving marimastat had significantly poorer quality of life at 3 and 6 months. Despite promising preclinical data, results of clinical phase III trials have been disappointing because of the narrow therapeutic index of MMPIs. There are currently five MMPIs in clinical development: marimastat in radically resected pancreatic cancer, BMS-275291 in advanced NSCLC, prinomastat in diverse tumor types and earlier stages of disease, Metastat in Kaposi's sarcoma, and Neovastat in unresectable renal cell carcinoma.114
Phase I study results have been reported for angiostatin and endostatin.117–121 Endostatin was administered by intravenous infusion or by subcutaneous administration and generally was well tolerated. Although the maximum-tolerated dose (MTD) was not reached for these proteins, biologic assessment showed trends toward decreased VEGF/bFGF urinary levels, reduction of tumor blood flow by dynamic magnetic resonance imaging (MRI), reduction of circulating EPCs, and increase in EC apoptosis.117–121 In a phase II study of subcutaneously administered endostatin in patients with advanced neuroendocrine tumors, there was minimal toxicity, and more than 60% of patients had stable disease.122 These results should be interpreted with caution, because the study did not use a randomized phase II model and the authors did not clearly define the clinical value of stable disease. Recently, independent groups reported either lack of efficacy of endostatin gene therapy in different tumors or failed to replicate the original findings of tumor growth inhibition with the recombinant protein.123
Thalidomide is a compound with antiangiogenic, immunomodulatory, and antitumor effects. The more promising clinical results with thalidomide have been observed in plasma cell malignancies, particularly multiple myeloma.124–128 The clinical data with thalidomide in multiple myeloma have been confirmed with a dose-dependent therapeutic effect.125–127 The mechanisms of thalidomide activity in multiple myeloma have not been well defined. Thalidomide in combination with chemotherapy did not improve clinical outcome of patients with acute myeloid leukemia or high-risk myelodysplastic syndrome.128 Several phase II studies have found a moderate activity of thalidomide in renal cell carcinoma.129,130 Dose levels of 800 to 1,200 mg/d were achieved; however, toxicities including somnolence, constipation, fatigue, neuropathy, and thromboembolism occurred. The addition of thalidomide to gemcitabine and fluorouracil did not improve the objective response rate observed with gemcitabine and fluorouracil, but added significant toxicity.129,130 In 17 assessable patients with recurrent glioblastoma, thalidomide, at 400 mg/d, was well tolerated, with constipation, somnolence, and peripheral neuropathy being the most common side effects.131 One minimal response and eight cases of stable disease were observed, with an overall clinical benefit of 52.9%. Median TTP and overall survival for responders were 25 and 36 weeks, respectively.131 Phase II clinical trials of thalidomide alone or in combination regimens have been reported in prostate cancer, head and neck cancer, malignant melanoma, and brain tumors.124,132–136 Most studies observed little activity with thalidomide alone, but suggest that further investigation in combination regimens may be justified with close follow-up of toxicities, especially peripheral neuropathy. Numerous clinical trials are ongoing to test the efficacy of nonsteroidal anti-inflammatory COX-2 inhibitors in combination regimens for therapy of advanced solid tumors.85 These compounds exhibit anti-inflammatory, analgesic, and antipyretic activities, as well as block of angiogenesis in animal models. Phase II clinical studies have combined celecoxib with a taxane, either docetaxel or paclitaxel, for treatment of NSCLC.85 Each study found the combination to be well tolerated, with response rates trending toward improved activity with celecoxib, without additional toxicity. In breast cancer, celecoxib in combination with exemestane has been reported.85 The combination was well tolerated, with a trend toward more efficacy for the combination. Celecoxib combinations have been studied for therapy of esophageal cancer with irinotecan/cisplatin/concurrent radiation therapy, pancreatic cancer with gemcitabine, renal cell carcinoma with low-dose cyclophosphamide, and malignant glioma with temozolomide.85 Celecoxib was well tolerated in all of the combination regimens. In a phase II study, Altorki et al137 evaluated the combination of celecoxib with paclitaxel/carboplatin regimen as preoperative chemotherapy in early-stage NSCLC. In comparison with historically reported data, the addition of celecoxib enhanced response rate and normalized the prostaglandin E2 tissue levels. This is the first published study suggesting a possible additive therapeutic effect by combining chemotherapy and anti–COX-2 agents in human solid tumors. Promising are the preliminary results of a phase I/II study ongoing at the San Filippo Neri Hospital in Rome, Italy, testing the combination of rofecoxib (50 mg/d) with weekly irinotecan and infusional fluorouracil. The dose-finding study on 15 cases demonstrated a good tolerability up to the irinotecan dose of 125 mg/m2/wk. The phase II study enrolled up to now 37 cases, and among the 30 assessable patients, the objective response rate was 36.7%, with a clinical benefit of 76.7%. Median TTP and overall survival were 4+ and 9+ months, respectively. The combination seems to be feasible and safe, with a reduced rate of mucositis and diarrhea.85,138 In an ongoing phase II study, the activity and tolerability of weekly paclitaxel and celecoxib was tested in 58 pretreated patients with NSCLC. The preliminary analysis on the 48 assessable patients showed a response rate of 27% and a stable disease rate of 48%, with a TTP and an overall survival of 4+ and 6+ months, respectively. This schedule was well tolerated, with a low incidence of grade 3/4 neutropenia and peripheral neuropathy.85 The rationale supporting the antitumor activity of selective inhibitors of COX-2 and an overview of preliminary data of the phase I/II clinical studies of combined therapy of anti–COX-2 agents with chemotherapy in advanced solid tumors have been recently reviewed.85 The main clinical trials with AIs combination are listed in Table 2.
On the basis of the results of experimental studies,63,64 Hanahan et al139 proposed the term of metronomic chemotherapy for schedules of cytotoxic agents given regularly at subcytotoxic doses and having the activated endothelium as principal target (ie, the antiangiogenesis chemotherapy paradigm). However, cytotoxic effects of so-called metronomic regimens in the tumor parenchyma could still contribute to the observed efficacy of such treatments and may require evaluation of the relatively most important compartment with each regimen and, potentially, tumor type. There are currently no published clinical studies that compare a true metronomic schedule of chemotherapy with conventional schedules. Several phase I and II studies were carried out involving oral, low, continuous doses of cytotoxic agents, with interesting results.140 There are some theoretical advantages to be explored with regard to this new schedule of chemotherapy (Table 3), but there are also potential problems and challenges in terms of appropriate experimental study design and clinical testing.62 cDNA microarrays and proteomic studies will better clarify the genetic basis of responsiveness of tumors to metronomic, antiangiogenic schedules.141 Finally, the identification of more specific surrogate markers is warranted, allowing the selection of patients for such treatments and the monitoring of the biologic effects as early or intermediate end points of treatment efficacy.142
Noninvasive assessment of tumor vascularization represents a novel approach for monitoring the activity of AIs. MRI, computed tomography, and positron emission tomography (PET) provide dynamic images of microvascularization and have also a potential utility for assessing the efficacy of AIs.143–146 Dynamic contrast-enhanced MRI uses gadolinium chelates as a marker of extravasation. Pharmacokinetic analysis of time-intensity curves quantifies the degree of tumor vascularization and its modifications after therapy.147 PET is another approach used to assess blood flow in human tumors.148 Two radiolabeled molecules are of particular interest: a radioactive form of water labeled with 15O, used to calculate blood flow within tumors, and radiolabeled carbon monoxide (11CO) that irreversibly binds RBCs and distributes in accordance with vascular volume.
Also, membrane proteins were selectively expressed by tumor ECs, such as integrins ( The utility of MRI and PET are under clinical evaluation in phase I/II studies. Preliminary results with AIs monitored with dynamic MRI or PET imaging demonstrated changes in vascular permeability, volume fraction, or metabolism after therapy.151 However, these changes do not always predict clinical efficacy of AIs. In patients with advanced colorectal cancer, Morgan et al have demonstrated a significant reduction in dynamic contrast-enhanced MRI parameters within a few hours after administration of PTK787/ZK222584152; moreover, there was a significant relationship between reduction of contrast enhancement and tumor regression. A phase I trial was performed with combretastatin A4 phosphate in 34 patients with different solid tumors measuring tumor blood flow parameters by either PET or dynamic MRI.153,154 Significant dose-dependent reduction in tumor blood flow perfusion was seen with PET or MRI a few hours after therapy. These two studies suggest that functional imaging obtained by PET or dynamic MRI could help to assess whether the drug achieves the target, as demonstrated by the reduction of tumor perfusion; select an adequate dosage for phase II studies, in relation to the identification of the doses able to reduce tumor perfusion; identify the better schedule of administration for phase II studies; and, finally, distinguish responsive versus unresponsive patients to AIs.
The clinical testing of first-generation AIs has assumed universal applicability rather than disease/target selectivity. The hypothesis was that angiogenesis is a necessary and ubiquitous step of progression of all tumors. The clinical study design for most of the randomized clinical trials of first-generation AIs consisted of the comparison of a standard chemotherapeutic regimen along with the test compound administered on a dose schedule established as a monotherapy versus the standard chemotherapeutic regimen alone. Clinical development of SU5416 (semaxanib) was an example of this strategy. The Roswell Park fluorouracil-based chemotherapeutic regimen was compared with the same regimen plus SU5416. The study was stopped when an interim data analysis found increased toxicity without any additional clinical benefit in the |
TOP
ABSTRACT
INTRODUCTION