First Published Online November 5, 2008 The Oncologist, Vol. 13, No. 11, 1193-1200, November 2008; doi:10.1634/theoncologist.2008-0188 © 2008 AlphaMed Press
Angiogenesis Inhibition in Non-GIST Soft Tissue SarcomasaErasmus University Medical Center, Rotterdam, The Netherlands; bRadboud University Nijmegen Medical Center, Nijmegen, The Netherlands; cCentre Léon Bérard, Lyon, France Key Words. Soft tissue sarcoma • Angiogenesis • Vascular endothelial growth factor (VEGF) Correspondence: Stefan Sleijfer, M.D., Ph.D., Department of Medical Oncology, Erasmus University Medical Center, Daniel den Hoed Cancer Center, Groene Hilledijk 301, 3075 EA Rotterdam, The Netherlands. Telephone: 31-10-7041733; Fax: 31-10-7041003; e-mail: s.sleijfer{at}erasmusmc.nl Received August 22, 2008; accepted for publication October 1, 2008; first published online in THE ONCOLOGIST Express on November 5, 2008.
Disclosure: Employment/leadership position: None; Intellectual property rights/inventor/patent holder: None; Consultant/advisory role: Jean-Yves Blay, GlaxoSmithKline; Honoraria: None; Research funding/contracted research: Stefan Sleijfer, noncommercial interest; Ownership interest: None; Expert testimony: None; Other: None.
This article is available for continuing medical education credit at CME.TheOncologist.com
Because angiogenesis is of crucial importance in the pathogenesis of cancer, blocking the function of proangiogenic factors has been shown to improve the outcomes of patients with several cancer types. Given the poor survival durations of patients with advanced soft-tissue sarcomas (STSs), which has remained stable at a median of 12 months over the last 20 year, there is an unmet need for novel agents active against these tumors. Like in other tumors, accumulating evidence points at an important role for angiogenic factors in STSs, rendering these factors attractive treatment targets. This review discusses the currently available evidence supporting a role for angiogenic factors in the pathogenesis of STSs and the first preliminary study results obtained with angiogenesis inhibitors.
The group of soft tissue sarcomas (STSs) encompasses a wide spectrum of >50 different tumor entities displaying great differences between each other in terms of pathogenesis, genetic makeup, clinical behavior, and sensitivity to systemic agents [1]. Given the latter, STSs are generally divided into three groups. Small blue round cell sarcomas, such as extraosseous Ewing's sarcoma/peripheral primitive neuroectodermal tumor and embryonal rhabdomyosarcoma, are highly sensitive to chemotherapy. As a consequence, chemotherapy is crucial in the management of these patients, not only for metastatic but also for localized disease. The advent of imatinib and the impressive antitumor effects achieved with this drug in patients with gastrointestinal stromal tumors (GISTs) have led to a second group that is treated differently from the other sarcomas. The third group, known as non-GIST STSs, is the largest and comprises all other entities. Despite the heterogeneity within this third group, and the consequences this may have for systemic therapies, almost all patients are treated in a similar way. For many years, doxorubicin and ifosfamide were considered the only drugs with consistent antitumor activity [2, 3]. And although the armamentarium of active regimens has recently been expanded for certain histological STS entities with trabectedin, approved only in Europe so far, and the combination of gemcitabine and docetaxel, which is not formally approved by health authorities for this indication, patients with advanced disease still face a dismal prognosis, underscoring the need for novel active compounds. Like all human tumor types, angiogenesis is an absolute prerequisite for the growth and dissemination of non-GIST STSs. Angiogenesis is a complex process, regulated by numerous pro- and antiangiogenic factors. Of all the proangiogenic factors, such as platelet-derived growth factor (PDGF), transforming growth factor, tumor necrosis factor, and interleukin-8, in particular, vascular endothelial growth factor (VEGF) and its receptor (VEGFR) have been elucidated as key players. Following successes with drugs interfering in the VEGF–VEGFR pathway in other tumor types, such as renal cell cancer, colorectal cancer, and breast cancer, trials exploring these drugs have been initiated in non-GIST STS as well, and recently the first preliminary results were presented. This review addresses the currently known role of angiogenesis in non-GIST STS and focuses on the studies with angiogenesis inhibitors performed so far.
The formation of the vascular network in tumors is regulated by numerous pro- and antiangiogenic factors, produced by tumor cells and by adjacent cells from the surrounding environment including endothelial cells, fibroblasts, and immune cells (Table 1). If the effects of the proangiogenic factors outweigh those of the antiangiogenic factors, endothelial cells are activated and start to proliferate. In addition, endothelial progenitor cells are mobilized from the bone marrow and home to the tumor site to form new endothelium. Furthermore, matrix metalloproteinases (MMPs) are produced to digest the extracellular matrix to allow the newly formed vasculature to expand.
Initial studies exploring the role of angiogenesis in STSs focused on microvessel density as the result of the balance between pro- and antiangiogenic factors. However, in most of these studies, no association of this parameter with outcome was revealed [4]. Driven by the increasingly recognized role of the VEGF–VEGFR pathway in angiogenesis, more recent studies focused on the determination of VEGF and its receptor as surrogates for angiogenic capacity. The VEGF family consists of VEGF-A, often referred to as VEGF, VEGF-B, VEGF-C, VEGF-D, and placental growth factor, of which VEGF is the most abundant. In addition, alternative splicing of VEGF results in the occurrence of four VEGF isoforms with different biological properties. There are three different VEGF receptors, VEGFR-1, VEGFR-2, and VEGFR-3. It is generally assumed that, in particular, VEGFR-2 and, albeit to a lesser extent, VEGFR-1 are involved in angiogenesis. In two studies using immunohistochemistry, there was an association between VEGF expression and histological tumor grade, but no association with disease-free or overall survival in patients with localized STS [5, 6]. In another study, enzyme-linked immunosorbent assay was used to assess VEGF expression, which enables a more quantitative determination. VEGF expression above the median level was an independent poor prognostic factor for metastasis-free and overall survival times in patients with localized disease [7]. High VEGF levels in STS patients have been found not only in tumor tissue, but also in the blood. In several studies, STS patients appeared to have elevated VEGF blood levels, compared with healthy controls [8–12]. Similar to the tissue levels, blood VEGF levels were related to histological grade [8, 10], and in several studies, though not in all, to a worse outcome as well in patients with localized disease [9, 10]. Not surprisingly, given the heterogeneity of non-GIST STSs, VEGF expression has been suggested to differ considerably among the diverse entities. The strongest VEGF expression was seen in tumors with epithelioid features, such as epithelioid sarcomas and alveolar soft part sarcomas, the latter being characterized by marked vascularity [13]. High VEGF expression levels have also been described in patients with malignant fibrous histiocytoma, dermatofibrosarcoma protuberans, and leiomyosarcoma [14]. In contrast to VEGF, data on VEGFR expression are rather scarce. In angiosarcomas, 65% had expression of VEGFR-2, and low expression was a poor prognostic factor for overall survival [15]. In addition to the VEGF–VEGFR pathway, several other factors involved in angiogenesis have been examined in STSs, though not so extensively. Using gene-expression profiling by microarray analysis, a variety of angiogenic genes were found to be upregulated in STS tumor tissue compared with normal tissue. These include the PDGF receptor (PDGFR), MMP-2, and Notch-1 and Notch-4 [12]. Several of these factors have been found to impact the outcome of non-GIST STS patients. For example, elevated expression of PDGF-B in tumors was associated with increased cell growth and histological grade [16]. Furthermore, levels of circulating basic fibroblast growth factor (bFGF) and angiopoietin-2 were higher in advanced STS patients than in healthy controls [8, 9, 11, 12]. In general, there was a relationship between these levels and grade [8], but an association with outcome was not consistently seen in all studies [11]. Also with respect to these angiogenic factors, there are differences among the diverse subtypes. The highest bFGF levels have been reported in fibrosarcoma and leiomyosarcoma patients [12]. Furthermore, alveolar soft part sarcoma, a highly vascular tumor, has a specific profile of upregulated angiogenic genes, probably induced by the specific fusion protein (ASPSCR1–TFE3) that is pathognomonic for this entity [17]. Factors involved in extracellular matrix breakdown, which is also essential for angiogenesis, are also related to outcome in STS patients. Patients with increased tumor tissue levels of MMP-2, MMP-9, and urokinase plasminogen activator had a worse outcome than patients with lower levels [18, 19]. In localized STS, high expression of CD100, a glycoprotein with a wide array of physiological functions, including promoting angiogenesis, was an independent prognostic factor for short disease-free survival and overall survival durations [20]. Collectively, many studies point to an important role for angiogenic factors in many STS entities (Table 2). In particular, for VEGF the data are accumulating, but for proangiogenic factors other than VEGF the evidence is less robust. Different assays, studies including small numbers of patients, and, above all, heterogeneity of the assessed tumor types are likely to account for this.
Of all the proangiogenic factors that may play a role in non-GIST STS, the VEGF–VEGFR pathway seems to be the most important one. The mechanisms underlying the increased activity of the VEGF–VEGFR pathway in STS remain to be unraveled and, again, are likely to differ by tumor type. Although not belonging to non-GIST STSs, in Ewing's sarcomas the specific fusion protein EWS–ETS activates the promoter of the VEGF gene [21]. Accordingly, similar mechanisms may induce VEGF expression in other STS entities, in particular in those types that are characterized by specific translocations, resulting in novel fusion proteins acting as transcription factors.
Another mechanism that may induce increased expression of VEGF is through hypoxia-inducible factor (HIF)-1
In addition to inducing angiogenesis, there is mounting evidence that VEGF expression can confer resistance against chemotherapy, which contributes to the generally worse outcome for patients with localized tumors exhibiting increased VEGF expression. Human STS cell lines stably transfected to produce VEGF did not exhibit increased growth and invasion capacity in vitro. However, in xenografts, these VEGF-producing cell lines formed highly vascular tumors with accelerated growth, displaying a higher potency to metastasize than the parental cell lines. Additionally, tumors from the VEGF-transfected cell lines exhibited less sensitivity to doxorubicin. An antibody directed towards VEGFR-2 increased the antitumor effects of doxorubicin, which underlines that increased activity of the VEGF–VEGFR pathway can confer chemoresistance [24]. In preclinical models from other tumor types, several mechanisms have been revealed by which increased VEGF–VEGFR activity yields chemoresistance. VEGF has been described to increase Bcl-2 levels, a well-known antiapoptotic factor [25]. Furthermore, activation of VEGFR yields resistance against several cytotoxic agents through induction of survivin, another antiapoptotic factor [26]. Another VEGF-driven mechanism that may partially account for lower sensitivity to systemic agents is through an effect on the interstitial fluid pressure. As a result of the fragile tumor vasculature that is more permeable than normal vasculature, almost all solid tumors are characterized by greater fluid pressure of the tumor interstitium than of adjacent tissues. This higher interstitial fluid pressure hinders the uptake of drugs from the peripheral circulation into tumors and consequently may contribute to resistance [27]. Inhibition of VEGFR has been shown to decrease the interstitial fluid pressure of tumors, thereby facilitating drug uptake [27]. Also, inhibition of PDGFR, for example, by imatinib, results in a reduced interstitial tumor pressure [27]. So, several mechanisms may underlie the generally worse outcome of patients with tumors with high VEGF expression (Table 3), but it is unlikely that the whole spectrum of mechanisms has been elucidated yet.
After recognizing the importance of angiogenesis in solid tumors, studies with drugs thought to affect angiogenesis were initiated. Although its precise mechanism of action is unclear, thalidomide is considered to exert its antitumor effects partially through inhibiting angiogenesis [28]. Thalidomide was one of the first drugs explored as an antiangiogenic drug in many tumor types, including sarcomas. In a small series of 17 patients with uterine sarcomas, including 10 carcinosarcomas, a tumor type that should nowadays be regarded as carcinoma instead of sarcoma, thalidomide did not induce antitumor activity in view of the median progression-free survival time of <2 months [29]. In case reports, however, thalidomide has been described to be able to produce dramatic antitumor effects [30]. The understanding of the involvement of the VEGF–VEGFR pathway in non-GIST STS, together with the antitumor activity of compounds targeting this pathway in other tumor types, prompted the exploration of these drugs in the clinic. Roughly, there are two drug classes by which the VEGF–VEGFR pathway can be inhibited: monoclonal antibodies and tyrosine kinase inhibitors (TKIs). So far, the most widely explored monoclonal antibody targeting the VEGF–VEGFR pathway in oncology is bevacizumab, an antibody directed toward VEGF. However, studies with bevacizumab as a single agent, or with other monoclonal antibodies against VEGF or VEGFR, are not available in STSs. With respect to TKIs, three drugs have been assessed for their effects against non-GIST STSs. Sorafenib is a TKI that targets the tyrosine kinase activity of several factors, including PDGFR and VEGFR-1, VEGFR-2, and VEGFR-3. In a relatively small study [31], published only in abstract form so far, sorafenib was assessed at a dose of 400 mg twice daily. Response rate was the primary endpoint, and three different STS tumor entities were examined: STSs of vascular origin (angiosarcomas and solitary fibrous tumors), leiomyosarcomas, and liposarcomas. In the 37 patients evaluable for response, only two responses were encountered: one in the nine patients with vascular STS tumors and one in the 19 accrued patients with leiomyosarcomas. The median progression-free survival times were 4.7, 1.7, and 1.8 months for patients with vascular sarcomas, liposarcomas, and leiomyosarcomas, respectively [31]. In a larger study of sorafenib, also available only in abstract form, 147 patients who had received one or fewer lines of prior treatment were entered [32]. Sorafenib was administered at 400 mg twice daily and six different strata were assessed: leiomyosarcomas, malignant fibrous histiocytomas, malignant peripheral nerve-sheath tumors, vascular sarcomas, synovial sarcomas, and a group with all remaining eligible STS tumor types. Again, response rate according to the Response Evaluation Criteria in Solid Tumors (RECIST) was the primary endpoint, and a Simon two-stage design was applied for each stratum separately. Of the 37 angiosarcoma patients, five experienced a response. In the leiomyosarcoma cohort, two of the 37 patients had a response, while in the other strata no responses according to the RECIST were seen. The median times to progression for the leiomyosarcoma, angiosarcoma, and other strata were 5.2, 5.5, and 2.8 months, respectively [32]. Toxicities with sorafenib were similar to those encountered in other tumor types [33, 34]. The outcomes of both these two studies look disappointing at first glance. However, the response rate served as the primary endpoint in both studies. It is increasingly being recognized that the antitumor activity of VEGFR TKIs is not adequately reflected in objective responses, but is better described in terms of progression-free survival. In advanced GIST patients who failed imatinib, sunitinib as a second-line therapy produced a response rate of only 7%, but induced a fourfold longer progression-free survival time than with placebo [35]. Likewise, sorafenib in patients with advanced renal cell carcinoma yielded a 10% response rate while the progression-free survival time was double what was seen with placebo [33]. Furthermore, sorafenib induced a 2% response rate in patients with advanced hepatocellular carcinoma but produced a longer median time to radiological progression and overall survival, by 2.7 and 2.8 months, respectively, compared with placebo [34]. As a result, many consider the progression-free rate (PFR) at a certain time point a more relevant endpoint when screening the antitumor activity of compounds targeting angiogenesis. Based on a large dataset from the European Organization for Research and Treatment of Cancer Soft Tissue and Bone Sarcoma Group (EORTC-STBSG), Van Glabbeke et al. [36] provided, for that purpose, PFRs at 3 and 6 months associated with active and inactive therapies for second-line treatment in advanced STS patients. According to this analysis, agents that induce a 3-month PFR >40% warrant further investigation. As in both studies on sorafenib, >50% of the angiosarcoma patients were progression free at 3 months, which applies also to leiomyosarcoma patients in one study [31, 32]; sorafenib seems to exhibit interesting antitumor activity in some STS entities. Data on overall survival are awaited with interest. Sunitinib is another TKI that inhibits the signal transduction of several factors including c-Kit, VEGFR, and PDGFR. Sunitinib was studied in two different phase II studies, both available only in abstract form. In a study on patients with advanced disease who received up to three different lines of prior chemotherapeutic regimens, continuous sunitinib at a daily dose of 37.5 mg resulted in one partial response in a patient with a desmoplastic small round-cell tumor, and seven of 39 patients remained progression free at 16 weeks [37]. Another study explored sunitinib at 50 mg daily from day 1 to 28 every 6 weeks in four different cohorts: liposarcomas, malignant fibrous histiocytomas, fibrosarcomas, and leiomyosarcomas. All patients received a maximum of three prior lines of therapy. In one patient, a partial response was seen and 29 of the 36 patients were progression free at 3 months after the initiation of sunitinib [38], meeting the criteria for an active drug according to the definitions based on the EORTC-STBSG analysis [36].
Another TKI that has been explored in STS is pazopanib. Pazopanib targets many factors, including VEGFR-1, VEGFR-2, and VEGFR-3, PDGFR- Recently, mTOR inhibitors have also been assessed in STS [23]. However, because it is uncertain whether or not these agents exert their antitumor effects through inhibiting angiogenesis, these agents are not discussed in this review. All together, VEGFR TKIs such as pazopanib, sunitinib, and sorafenib induce interesting antitumor activity in some STS tumor types. Of course, the true value of these drugs can be established only in randomized studies. For pazopanib, a worldwide placebo-controlled randomized phase III study in non-GIST STSs other than liposarcomas will start accrual shortly.
Besides as a single agent, inhibitors of the VEGF–VEGFR pathway are attractive agents to combine with conventional cytotoxic drugs as has successfully been done with bevacizumab in tumor types such as non-small cell lung cancer, colorectal cancer, and breast cancer. In addition to a potential synergistic interaction, VEGF-targeting agents and conventional cytotoxic agents both have antitumor activity as single agents and exhibit nonoverlapping toxicity profiles, justifying the exploration of such combinations in STS. In a single-arm phase II study, bevacizumab was combined with doxorubicin in 17 patients with metastatic STS [40]. Despite the use of dexrazoxane for cumulative doxorubicin doses >300 mg/m2, greater than expected cardiotoxicity was seen. Of the 17 patients, six experienced a grade >2 decline in ejection fraction. With respect to antitumor activity, only two of the 17 patients achieved a response, the primary endpoint. However, as mentioned above, the response rate may not be the most suitable parameter to assess antitumor activity of regimens containing VEGF-targeting agents. The finding that 11 patients had stable disease for >12 weeks, while, additionally, in some of these patients clear cystic alterations in tumor lesions occurred, suggests that this combination has antitumor activity [40]. Of note, the lack of a randomized study design renders it very difficult to draw firm conclusions on the precise antitumor activity of this drug combination. Furthermore, given the heterogeneity among the different histological tumor types in terms of chemosensitivity, inclusion of particular STS entities can greatly influence the outcomes. In this study, 65% of the patients had a leiomyosarcoma, a rather chemoresistant subtype, but this tumor type constitutes approximately 25% of all STSs. Nevertheless, despite the unacceptably high incidence of cardiotoxicity from the combination of bevacizumab and doxorubicin seen in this study [40], combinations of compounds targeting the VEGF–VEGFR pathway with conventional cytotoxic drugs seem worthwhile to explore in STS. Data on combined approaches with other angiogenesis inhibitors and conventional agents in STS have not been reported yet. Instead of doxorubicin, ifosfamide may be an attractive combination treatment partner to explore in STS given its lack of cardiotoxicity [3].
As holds true for all human solid tumor types, angiogenesis plays a role in the pathogenesis of STS. As an almost universal mediator of angiogenesis, in STS there also are strong indications that VEGF is involved. Not surprisingly, given the heterogeneity among the diverse STS entities, the importance of VEGF seems to differ among the different subtypes. It is likely that proangiogenic factors other than VEGF contribute as well, but strong evidence supporting this is lacking for the time being. VEGF expression, both in tumor tissue and in the blood, has been shown to be associated with tumor grade and, though not consistently, also with outcome in localized disease. The reasons why VEGF expression is adversely related to outcome in STS are not exactly known. Several mechanisms are probably involved, including a higher potency of these tumors to metastasize and resistance to chemotherapy.
Currently, there are preliminary findings strongly suggesting that VEGFR TKIs such as sorafenib and pazopanib exhibit antitumor activity in STS. Again, sensitivity seems to differ among the STS entities. Based on the outcome of a phase II study [39], a global placebo-controlled phase III study with pazopanib will be initiated by the end of 2008 in STS entities, with the exception of liposarcomas. In addition to establishing the true value of VEGFR TKIs in STS, there is a great need to identify prognostic and predictive factors for outcome using such agents. Besides histopathological entities, factors potentially worthwhile exploring include gene-expression profiling of tumor tissue, expression of components of the VEGF–VEGFR pathway, such as VEGF and VEGFR themselves, and also HIF1 In addition to agents impacting the VEGF–VEGFR pathway, novel agents targeting other components involved in angiogenesis are being developed. These include inhibitors of PDGFR, bFGF, MMPs, integrins, and many others. Given the assumed role of some of these targets, these drugs look worthwhile to explore in non-GIST STS. Not only as single agents, but also combined with other agents, compounds targeting the VEGF–VEGFR system deserve further study. The first study of such a combination consisting of bevacizumab and doxorubicin was characterized by a higher incidence of cardiomyopathy [40]. By using ifosfamide, which is considered a reasonable alternative to doxorubicin, in such a combination, the problem of cardiotoxicity may be circumvented. Furthermore, monitoring of the antitumor activity of angiogenesis inhibitors, especially when given as a single agent, should be carefully considered because the common and widely applied RECIST and World Health Organization response criteria are likely to underreport the efficacy of these novel agents. Angiogenesis inhibition has proven to be a valuable approach to treat numerous tumor types and it is likely that its use can be expanded to particular STS entities as well. Their true value, however, can be established only in appropriately designed and conducted trials. Given the rarity of STS in general and the specific subtypes in particular, this will be a major task, which can be accomplished only through the increasingly ongoing international collaboration.
Conception/design: Stefan Sleijfer; Winette T.A. van der Graaf; Jean-Yves Blay Financial support: Stefan Sleijfer; Winette T.A. van der Graaf; Jean-Yves Blay Administrative support: Stefan Sleijfer; Winette T.A. van der Graaf; Jean- Yves Blay Provision of study materials: Stefan Sleijfer; Winette T.A. van der Graaf; Jean-Yves Blay Collection/assembly of data: Stefan Sleijfer; Winette T.A. van der Graaf; Jean-Yves Blay Data analysis: Stefan Sleijfer; Winette T.A. van der Graaf; Jean-Yves Blay Manuscript writing: Stefan Sleijfer; Winette T.A. van der Graaf; Jean-Yves Blay Final approval of manuscript: Stefan Sleijfer; Winette T.A. van der Graaf; Jean-Yves Blay
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||