|Home | About | Journals | Submit | Contact Us | Français|
Alternatives to cytotoxic agents are desirable for patients with HIV-associated Kaposi's sarcoma (KS). Vascular endothelial growth factor-A (VEGF-A) contributes to KS pathogenesis. We evaluated the humanized anti–VEGF-A monoclonal antibody, bevacizumab, in patients with HIV-KS.
Patients with HIV-KS who either experienced progression while receiving highly active antiretroviral therapy (HAART) for at least 1 month or did not regress despite HAART for at least 4 months were administered bevacizumab 15 mg/kg intravenously on days 1 and 8 and then every 3 weeks. The primary objective was assessment of antitumor activity using modified AIDS Clinical Trial Group (ACTG) criteria for HIV-KS. HIV-uninfected patients were also eligible and observed separately.
Seventeen HIV-infected patients were enrolled. Fourteen patients had been receiving effective HAART for at least 6 months (median, 1 year). Thirteen patients had advanced disease (ACTG T1), 13 patients had received prior chemotherapy for KS, and seven patients had CD4 count less than 200 cells/μL. Median number of cycles was 10 (range, 1 to 37 cycles); median follow-up was 8.3 months (range, 3 to 36 months). Of 16 assessable patients, best tumor responses observed were complete response (CR) in three patients (19%), partial response (PR) in two patients (12%), stable disease in nine patients (56%), and progressive disease in two patients (12%). Overall response rate (CR + PR) was 31% (95% CI, 11% to 58.7%). Four of five responders had received prior chemotherapy for KS. Over 202 cycles, grade 3 to 4 adverse events at least possibly attributed to therapy included hypertension (n = 7), neutropenia (n = 5), cellulitis (n = 3), and headache (n = 2).
Bevacizumab is tolerated in patients with HIV-KS and has activity in a subset of patients.
Kaposi's sarcoma (KS) is a multifocal angioproliferative malignancy characterized by endothelial-derived spindle cells, vascular slits with enhanced permeability, and local inflammatory infiltrate. KS-associated herpes virus (KSHV), also called human herpesvirus-8, is a necessary but insufficient cause of KS.1–3 The majority of cells in KS lesions are KSHV infected. HIV is a cofactor that increases KS risk.4 Despite decline in KS incidence associated with highly active antiretroviral therapy (HAART) availability in the developed world, HIV-infected individuals remain at a markedly elevated risk of KS. In the United States, KS remains the second most common cancer among people with HIV.5 KS is one of the most common cancers in sub-Saharan Africa6 and is a major public health problem as a result of epidemic HIV.7 KS incidence increases with age, and the effect of an aging US HIV-positive population on KS incidence remains to be seen.
HAART is essential to HIV-associated KS (HIV-KS) therapy.8–10 Its effectiveness is largely a result of control of HIV and resulting improved KSHV-specific cellular immunity.11 In controlled KS trials, HAART alone induced responses in approximately 20% of patients,9,10,12,13 depending partly on immune reconstitution potential and extent of KS. Addition of systemic cytotoxic chemotherapy is indicated for advanced or symptomatic KS. Liposomal anthracyclines, with an overall response rate (ORR) of 55% to 76% in the HAART era,13–17 are considered first-line agents. However, KS is not curable, and 1-year progression-free survival (PFS) with liposomal doxorubicin is approximately 70%.17 Long-term administration of continuous or intermittent chemotherapy is often required. Given substantially improved survival of HIV-positive patients on HAART, long-term toxicities of anti-KS therapies must be considered. Indeed, cumulative therapy-associated toxicity, rather than therapy-refractory disease, frequently limits long-term KS management. High cumulative doses of anthracyclines are associated with irreversible cardiac toxicity.18 Although drugs such as interferon alfa, vincristine, vinblastine, etoposide, and paclitaxel are active in KS, they have lower activity than liposomal anthracyclines and/or greater toxicity. Improved therapies are urgently needed.19–21
KS, characterized by angiogenic proliferation of endothelial-derived cells, is a rational and potentially optimal tumor in which to consider antiangiogenic approaches. Vascular endothelial growth factor-A (VEGF-A) is an important paracrine and autocrine growth factor in KS.22,23 KSHV has developed redundant mechanisms for upregulation of VEGF-A. Viral gene products, including viral G protein–coupled receptor, viral interleukin (IL) -6, latency-associated nuclear antigen (LANA), and K1, all directly or indirectly upregulate VEGF-A production.24–29 VEGF-A seems to be responsible for leaky blood vessels, a common pathologic feature, as well as some clinical features, including tumor-associated edema and effusions.30–33 Given the role of VEGF-A in KS pathogenesis, we performed a phase II study of the humanized, monoclonal, anti–VEGF-A antibody, bevacizumab, in patients with HIV-KS.
Patients were adults with pathologically confirmed KS and at least five evaluable cutaneous lesions. HIV-positive patients must have been on HAART for at least 1 month with evidence of progressive disease (PD) or for at least 4 months without disease regression. Additional requirements included the following: Eastern Cooperative Oncology Group performance status of ≤ 2, life expectancy of at least 6 months, systolic blood pressure less than 160 mmHg, diastolic blood pressure less than 95 mmHg, urine protein less than 1+ on dipstick or less than 500 mg on 24-hour collection, absolute neutrophil count greater than 750 cells/μL, hemoglobin greater than 9 g/dL, and platelets greater than 75,000/μL. There were no CD4 count exclusion criteria. Patients with symptomatic visceral KS, concurrent malignancies not in remission for at least 1 year, or history of thromboembolic disease were excluded.
In this single-center phase II study, patients received bevacizumab 15 mg/kg loading dose, then bevacizumab 15 mg/kg every 3 weeks starting 1 week after the loading dose. Bevacizumab was temporarily held for systolic blood pressure greater than 160 mmHg, diastolic blood pressure greater than 95 mmHg, proteinuria greater than 2+ on dipstick or 2 g in a 24-hour collection, or platelet count less than 50,000/μL. If proteinuria did not resolve to less than 2+ or 2 g/24 hours within 4 weeks, bevacizumab was discontinued. Antihypertensive therapy was initiated for systolic blood pressure greater than 160 mmHg or diastolic blood pressure greater than 95 mmHg persisting for more than 1 week or for systolic blood pressure greater than 210 mmHg or diastolic blood pressure greater than 120 mmHg at any time. HIV-positive patients with CD4 count of less than 200 cells/μL received Pneumocystis jiroveci prophylaxis. Mycobacterium avium prophylaxis was considered if CD4 count was less than 75 cells/μL. HIV-infected patients continued HAART, with adjustments made as needed according to US guidelines.34 Bevacizumab was continued unless patients had PD requiring cytotoxic therapy, unacceptable toxicity, lack of adherence to protocol (including HAART), or patient-requested discontinuation (ie, for elective surgery). In HIV-KS, transient progression can be seen before improvement35; therefore, in patients with limited KS that during interval assessment was classified as PD by modified AIDS Clinical Trial Group (ACTG) criteria36,37 (Appendix, online only), bevacizumab could be continued for additional cycles at investigator discretion as long as the patient did not require cytotoxic chemotherapy. Filgrastim was used as clinically indicated. Bevacizumab was provided by the Cancer Therapy Evaluation Program of the National Cancer Institute (NCI) through a Cooperative Research and Development Agreement with Genentech (South San Francisco, CA). This protocol was approved by the NCI Institutional Review Board and is registered at ClinicalTrials.gov (NCT00055237). All patients provided written informed consent.
KS response was evaluated every cycle and categorized as complete response (CR), partial response (PR), stable disease (SD), or PD using modified ACTG criteria,36,37 as previously described.37 Response evaluations included lesion counts, measurement of the sum product of the diameters (SPD) of five marker lesions, and assessment of nodularity. PR required at least 50% decrease in number of lesions and/or sum product of the diameters of marker lesions and/or nodularity of lesions and no new lesions. CR required clinical resolution of all lesions and tumor-associated phenomenon, with biopsy confirmation. (See Appendix for detailed response criteria.) Both CR and PR had to be sustained for 4 weeks. Best response was evaluated for each patient. Patients who did not achieve SD for at least 3 weeks were considered to have PD as best response.
Safety was monitored each cycle and 3 to 6 weeks after completing therapy. Evaluations included CBCs with differential, serum chemistries, and urinalysis. Toxicities were graded using NCI Common Terminology Criteria for Adverse Events version 2.0. In HIV-positive patients, CD4 cell counts and HIV viral load were evaluated every 12 weeks.
Correlative assays were performed on stored biospecimens collected at baseline and time of best response. Serum VEGF-A was measured using Quantikine ELISA Kit (R&D Systems, Minneapolis, MN). Serum cytokines (IL-1β, IL-6, IL-8, IL-10, IL-12 p70, interferon gamma, tumor necrosis factor α) were evaluated using MSD 96-Well Multiarray Proinflammatory 7-plex Assay (Meso-Scale Discovery, Gaithersburg, MD) and Sector Imager (Meso-Scale Discovery). KSHV viral load was measured using previous described methodology.38
The primary objective was to determine the ORR (CR + PR) in patients with HIV-KS on HAART treated with bevacizumab. HIV-negative patients were also eligible; however, given likely differences between HIV-negative KS and HIV-KS, prespecified primary analysis was limited to HIV-KS. Entry criteria requiring SD or PD on HAART were designed to exclude patients most likely to respond to HAART alone. Sample size was determined using two-stage Simon optimal design39 (α = .10; β = .10; undesirably low ORR, 15%; targeted ORR, 45%). If two or more of the first eight patients had a PR or better, accrual continued to 17 HIV-positive patients. If five or more of 17 patients responded, bevacizumab would be considered sufficiently active to consider future studies. Statistical significance of differences in serum VEGF and cytokines at time of best response versus baseline was determined by Wilcoxon signed rank test. Significance of the difference in changes between patients with or without a clinical response was determined by an exact Wilcoxon rank sum test. PFS for evaluable patients with HIV was determined using Kaplan-Meier methods, censoring patients without progression at the off-study date. P values are two-tailed and presented without adjustment for multiple comparisons, because they are results of exploratory tests.
Between February 2003 and December 2008, 17 patients (16 men and one woman) with HIV-KS were enrolled (Table 1). Eight patients were black, six were white, and three were Hispanic. Median age was 44 years (range, 23 to 65 years). Thirteen patients (76%) had advanced KS (ACTG T1).40 Median CD4 count was 294 cells/μL (range, 7 to 654 cells/μL), and five patients (29%) had CD4 count less than 150 cells/μL. Patients had substantial prior treatment; 13 patients (76%) had received previous cytotoxic chemotherapy, including liposomal doxorubicin (n = 12) and paclitaxel (n = 5). Additionally, 11 patients had received immunotherapy (interferon alfa, IL-12, or thalidomide), and five patients had received radiation therapy (Table 1). One patient had bilateral KS-associated pleural effusions requiring indwelling catheter drainage (Fig 1) and bilateral lower extremity edema requiring daily diuretics. Additionally, two HIV-uninfected black men, age 49 and 65, were enrolled and analyzed separately.
Assessable patients with HIV-KS received a median of 11 cycles (range, four to 37 cycles) of bevacizumab and were observed on-study for a median of 9 months (range, 3 to 36 months). All patients with HIV-KS received HAART. The two HIV-uninfected patients received four and five cycles of bevacizumab. Two hundred two cycles are evaluable for safety and tolerability.
Sixteen of 17 patients with HIV-KS were assessable for tumor response; one patient did not return for tumor evaluation after initial doses of therapy. Best responses were CR in three patients (19%), PR in two patients (12%), SD in nine patients (56%), and PD in two patients (12%; Table 2). Best ORR was 31% (95% CI, 11% to 58.7%). One patient achieved CR after initial transient progression during the first month. Another patient had rapid durable resolution of bilateral effusions and lower extremity edema and achieved a CR (Fig 1). In patients with PR or CR, median time to best response was 5 months (range, 2.5 to 9.5 months). In two HIV-uninfected patients, best responses were SD (n = 1) and PD (n = 1). Overall, eight of 11 patients with baseline tumor-associated edema had evidence of improvement. Six patients had a greater than 2 cm decrease in circumference of affected limb at time of best response, and five patients had subjective improvement, including opiate discontinuation (n = 2), diuretic discontinuation (n = 1), and increased mobility and activity (n = 4).
All five responders were among a group of 12 patients with sustained HIV suppression on study (Table 2). In these 12 patients, ORR was 42%. Interestingly, 11 patients had evidence of ongoing immune reconstitution, with median change in CD4 count of +144 cells/μL (range, −51 to +352 cells/μL). Seven (42%) of these 12 patients had been on cytotoxic chemotherapy in the 12 months before starting bevacizumab, and immune reconstitution may have been facilitated by switching from additional cytotoxic therapy to bevacizumab. Responses did not seem to be merely a result of recent initiation of HAART, because all responders had been on a stable HAART regimen for 6 months or longer (median, 11 months) before starting bevacizumab. In contrast, four patients had increasing HIV viral load on study attributed to HAART nonadherence (n = 2), acquired HIV resistance (n = 1), or not taking HAART because of intercurrent illness (n = 1). In these four patients, median change in CD4 count was −19 cells/μL (range, −232 to +37 cells/μL; Table 2). In evaluable patients with HIV-KS, median PFS was 8.3 months (Fig 2). Overall survival on study was 100%.
Bevacizumab was generally well tolerated. Common adverse effects were proteinuria, headache, epistaxis, and hypertension (Table 3). Five patients started antihypertensive therapy. One heavily pretreated patient with lasting PR developed asymptomatic proteinuria after cycle 30 and discontinued bevacizumab after 37 cycles when urine protein reached 1,024 mg/24 h. Proteinuria resolved over 7.5 months of follow-up. A second patient with CR who discontinued therapy after cycle 26 was found to have an enlarged cardiac silhouette on chest x-ray 3 months after completing bevacizumab. On echocardiography, ejection fraction (EF) was 36%; no baseline cardiac studies were available for comparison. The patient was asymptomatic and continued on losartan and carvedilol. Follow-up echocardiography 6 years later showed an EF of 50% with no dilation. Three patients developed soft tissue infections associated with underlying KS requiring intravenous antibiotics; one patient had recurrent infections, one had infection in the setting of neutropenia, and one discontinued therapy as a result of severity of infection.
We assessed serum levels of VEGF-A and several cytokines possibly relevant to KS pathogenesis at entry and time of best response (Table 4). There was a decrease in serum IL-8 (median decrease, 11.2 ng/mL; P = .0023) but no significant changes in serum VEGF-A, IL-1β, IL-6, IL-8, IL-10, IL-12p70, interferon gamma, or tumor necrosis factor α. There were no significant differences in changes in these factors between those with or without a clinical response.
In this study, patients with HIV-KS on a stable HAART regimen received bevacizumab 15 mg/kg every 3 weeks after an initial loading dose. Best ORR was 31% (95% CI, 11% to 58.7%), meeting predefined criteria for consideration in future combination studies. Median PFS was 8.3 months. Compared with bevacizumab monotherapy studies in metastatic renal cell cancer or recurrent ovarian cancer, in which ORRs of 10% to 21% were seen,44–46 responses observed in HIV-KS are quite good. Moreover, these responses were seen in patients with poor prognosis KS,41 and four of five responding patients had received prior cytotoxic chemotherapy (Table 2). In contrast to most cytotoxic agents active in KS, bevacizumab does not seem to impair immune reconstitution, an important feature for therapeutic interventions for HIV-KS.
Patients who responded had controlled HIV and increases in CD4 counts while on bevacizumab (Table 2). Baseline CD4 lymphocytopenia may have been in part a result of prior chemotherapy in most patients, and cessation of hematotoxic chemotherapy seems to permit the increases in CD4 counts observed in responding patients. A limitation to any phase II study in HIV-KS is that possible immune reconstitution must be taken into consideration, and HAART alone can induce responses in KS (approximately 20% in controlled trials9,13). Meta-analysis suggests that most patients with HIV-KS who respond to HAART alone have T0 KS (limited disease); only five documented cases were identified in which patients with T1 KS (widespread disease) responded to HAART alone.10 In the current study, all responders had T1 KS, and most had prior cytotoxic chemotherapy, making it unlikely that responses were a result of HAART alone. Moreover, most KS responses to HAART occur soon after the initiation of HAART, although paradoxical worsening of HIV-KS after starting HAART is also described.47,48 To limit these potential biases, we required patients to have either KS not regressing while on HAART for 4 months or progressing on HAART for 1 month. Given the potential role of VEGF dysregulation in the pathophysiology of KS-associated pleural effusions49 and edema, rapid resolution of pleural effusions in one patient and common subjective improvement in tumor-associated edema were noteworthy observations. Nonetheless, definitive assessment of anti-KS efficacy of bevacizumab beyond that of HAART alone requires a randomized controlled trial.
With the exception of decrease from baseline in IL-8, assessment of serum VEGF-A and cytokines did not show substantial changes or association with responses. Bevacizumab binds to VEGF-A, and measurement of bound VEGF-A may affect assay results.50 Although difference between responders and nonresponders was not statistically significant, the decrease in IL-8 is interesting, because KSHV encodes a latently expressed gene, K13, that transcriptionally upregulates IL-851 and may have a role in mediating angiogenesis in KS.52 Additional studies will be needed to evaluate IL-8 as a biomarker and to sort out its possible biologic role in KS response to bevacizumab.
Bevacizumab was generally well tolerated over a relatively long time. Adverse events (Table 3) were comparable with those seen in other studies.53 Two noteworthy toxicities at least possibly attributable to long-term bevacizumab were proteinuria (> 1 g/d) in one patient and a decreased cardiac EF in another; in both cases, toxicities improved with bevacizumab discontinuation. In addition, five patients required initiation of antihypertensive agents. Three patients developed soft tissue infections; KS patients are susceptible to soft tissue infections, and it was unclear whether bevacizumab had a role in their pathogenesis. Overall, the toxicity profile observed in this HIV-positive population receiving HAART supports bevacizumab use in future studies in HIV-associated cancers, as well as its use in HIV-positive patients with cancers for which bevacizumab is US Food and Drug Administration approved. This is particularly important given the increasing burden of non–AIDS-defining malignancies such as lung cancer and colon cancer in HIV-infected individuals in the United States.5
KS response rates are affected by extent of disease, degree of immunosuppression, and control of HIV, and response rates can be difficult to compare among clinical trials. The observed ORR here is less than that reported with liposomal anthracyclines13–16 but is comparable to that seen using agents that inhibit angiogenesis through different mechanisms, such as TNP-47054 or the matrix metalloproteinase inhibitor COL-3.55,56 Given the important role of VEGF-A in KS pathogenesis, one must ask why bevacizumab was not more active. One likely reason is that redundant angiogenic and proliferative stimuli activate spindle cell proliferation. In addition to VEGF-A receptors 1 and 2, KS spindle cells express VEGF-A receptor 3 and the receptor for platelet-derived growth factor (PDGF) and proliferate in response to ligands for these receptors (VEGF-C and PDGF).57–60 Furthermore, a number of KSHV genes, such as latency-associated nuclear antigen (LANA), v-FLIP, v-cyclin, and kaposin-A, can inhibit apoptosis or directly contribute to KS spindle cell proliferation.61 Thus, optimal targeted therapy for KS may require targeting two or more of these pathways simultaneously.54–56
Although only a subset of patients responded in this trial, results should be considered in light of the fact that most patients had features making them unlikely to respond to any therapy. Overall, this study suggests that bevacizumab has utility in KS. In particular, bevacizumab may be of value in combination with other drugs or after initial reduction of the tumor burden with cytotoxic chemotherapy, or in patients who are approaching the maximal safe cumulative dose of anthracyclines. A second study of bevacizumab combined with liposomal doxorubicin, followed by bevacizumab maintenance (ClinicalTrials.gov identifier: NCT00923936), is under way, and a randomized trial of bevacizumab is worth considering. With increasing insight into KSHV biology and range of clinical presentations of KS38 and other KSHV-associated malignancies, rational therapeutic approaches such as bevacizumab offer hope for both cytotoxic-sparing treatment options and personalized approaches to difficult-to-manage specific tumor-associated symptoms like chronic edema and effusions.
We thank the patients who volunteered to participate; the medical, nursing, and support staffs of the HIV and AIDS Malignancy Branch, the Medical Oncology Branch, and the National Institutes of Health Clinical Center; and Adam Rupert, Randy Stevens, and others at Science Applications International Corporation, National Cancer Intitute-Frederick, Frederick, MD.
Modified AIDS Clinical Trial Group response assessments consisted of lesion counts, assessment of lesion nodularity, measurement of sum of the product of the diameters of five marker lesions, and evaluation for oral Kaposi's sarcoma (KS) and secondary effects of KS. For patients with less than 50 cutaneous lesions at baseline, the total number of lesions was counted at each response evaluation and assessed as raised or flat. In patients with more than 50 lesions at baseline, between one and three representative body areas were selected, and the total number of lesions within each of those areas was counted and assessed for nodularity. Total-body photographs, as well as detailed photographs of sentinel and marker lesions and any other notable features of the patient's KS, were performed at baseline, every fourth cycle, and at the time a patient went off study. Chest x-rays and other clinically indicated radiologic studies were obtained at baseline and whenever clinically indicated to assess changes. Complete response was defined as clinical resolution of all lesions and tumor-associated phenomena and required biopsy confirmation in patients with residual macular pigmentation.37 Partial response was defined as no evidence of progressive disease and at least one of the following: greater than 50% decrease in number and/or size of lesions; complete flattening of at least 50% of previously raised lesions; or 50% decrease in sum of the product of the diameters of marker lesions. Patients with residual tumor-associated edema or effusions otherwise meeting complete response criteria were also considered to have a partial response. Partial and complete responses had to be sustained for at least 4 weeks. Patients with a greater than 25% increase in any lesion characteristics defined earlier, new cutaneous or visceral disease, or new or increasing tumor-associated edema or effusions persisting for at least 1 week and interfering with normal activities were considered to have progressive disease. Patients not meeting any of these categories were considered to have stable disease.
Supported in part by the Intramural Research Program, National Cancer Institute (NCI), National Institutes of Health (NIH). Additional funding provided by the NCI, NIH, under Contract No. HHSN261200800001E.
Presented in part at the 17th Annual Conference on Retroviruses and Opportunistic Infections, February 16-19, 2010, San Francisco, CA.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Clinical trial information can be found for the following: NCT00055237.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: None Consultant or Advisory Role: None Stock Ownership: None Honoraria: None Research Funding: Robert Yarchoan, bevacizumab provided to National Cancer Institute by Genentech under a Cooperative Research and Development Agreement Expert Testimony: None Other Remuneration: Robert Yarchoan, patents on certain anti-HIV drugs, including didanosine, assigned to Department of Health and Human Services/National Institutes of Health under the Federal Technology Transfer Act
Conception and design: Richard F. Little, Robert Yarchoan
Collection and assembly of data: Thomas S. Uldrick, Kathleen M. Wyvill, Pallavi Kumar, Deirdre O'Mahony, Wendy Bernstein, Karen Aleman, Mark N. Polizzotto, Stefania Pittaluga, Vickie Marshall, Denise Whitby, Richard F. Little, Robert Yarchoan
Data analysis and interpretation: Thomas S. Uldrick, Kathleen M. Wyvill, Deirdre O'Mahony, Mark N. Polizzotto, Seth M. Steinberg, Robert Yarchoan
Manuscript writing: All authors
Final approval of manuscript: All authors