点击显示 收起
1HIV and AIDS Malignancy Branch, 2Laboratory of Pathology, and 3Experimental Transplantation and Immunology Branch, Center for Cancer Research, National Cancer Institute, Bethesda, and 4Viral Epidemiology Laboratory, National Cancer Institute, Frederick, Maryland; 5Department of Microbiology, University of Minnesota, Minneapolis
Received 1 July 2002; revised 6 September 2002; electronically published 13 December 2002.
A clinical trial was conducted to test the activity of cidofovir (CDV), a drug with in vitro activity against Kaposi sarcoma (KS)associated herpesvirus (KSHV), in KS. Five patients with human immunodeficiency virusassociated KS (4 receiving antiretroviral therapy) and 2 patients with classical KS were administered CDV (5 mg/kg/dose) weekly for 2 weeks and then every other week. All 7 patients had progression of their KS at a median of 8.1 weeks (range, 527 weeks). Skin biopsy specimens of KS lesions showed no change in expression of latent or early lytic genes, but, in the 1 assessable patient, there was decreased expression of a late lytic gene. There was no decrease in the virus load of KSHV in peripheral blood mononuclear cells. This study does not provide proof of principle for the treatment of KS with CDV. However, it remains possible that antiherpesvirus therapy can be developed for herpes-induced tumors.
Presented in part: 5th International AIDS Malignancy Conference, Bethesda, Maryland, April 2001 (abstract 52).
The protocol for the study was approved by the institutional review board of the National Cancer Institute. Written informed consent was obtained from all study participants, in accordance with the human experimentation guidelines of the US Department of Human Services.
Financial support: National Cancer Institute intramural support and National Cancer Institute (contract N01-CO-12400). Cidofovir was provided by Gilead Sciences under a Material Transfer Agreement with the National Cancer Institute.
Present affiliations: Bayer, West Haven, Connecticut (R.H.); Medimmune, Gaithersburg, Maryland (J.M.P.); and Elan Pharmaceuticals, Princeton, New Jersey (L.W.).
Reprints or correspondence: Dr. Robert Yarchoan, National Institutes of Health, Bldg. 10, Rm. 10S255, MSC 1868, 10 Center Dr., Bethesda, MD 20892-1868 .
Kaposi sarcoma (KS) is a vascular neoplasm that occurs simultaneously at multiple sites. KS lesions are characterized by spindle cells that are nearly all infected by KS-associated herpes virus (KSHV), also called human herpesvirus 8 [1, 2]. KSHV is an essential, although not sufficient, etiologic agent for KS. It also is an etiologic agent for primary effusion lymphoma and multicentric Castleman's disease (MCD).
Available evidence suggests that KS results from factor-driven hyperproliferation, possibly with late transformation. Although most spindle cells are latently infected by KSHV, a small percentage express lytic genes [2]. Several lytic KSHV genes are analogs of human cytokines (e.g., viral interleukin-6 [vIL-6]) or angiogenic factors [1] that can stimulate spindle cells. Another lytic gene, ORF74, up-regulates vascular endothelial growth factor (VEGF). Thus, it seemed plausible that antiherpes therapy might be effective against KS. Additional support for this approach came from anecdotal reports of KS responding to antiherpes drugs [3, 4] and from studies showing that treatment of patients with AIDS with ganciclovir or foscarnet reduced the risk of KS [5, 6]. Cidofovir (CDV) was found to be one of the most active drugs against KSHV, with IC50 values of 0.056.3 M (0.0151.98 g/mL) [79]. With this background, we initiated a trial of CDV in patients with KS.
Patients and methods. Patients with KS, with or without human immunodeficiency virus (HIV) infection, were eligible if they had 5 measurable lesions without antecedent local therapy. Patients had to have a Karnofsky performance index 70 and be free of acute life-threatening conditions. Patients also had to have relatively intact renal and other organ function. Patients infected with HIV had to have a CD4+ cell count 50 cells/mm3 and be either on stable antiretroviral therapy or off therapy for 1 month. Systemic or topical anti-KS therapy or antiherpes therapy (except acyclovir cream) were not allowed within 4 weeks of the start date. Patients with severe KS were excluded.
Patients were administered 2 weekly CDV intravenous doses of 5 mg/kg/dose, followed by 5 mg/kg/dose every other week. Two grams of probenecid were administered orally 3 h before CDV was administered, and 1 g was administered orally 2 and 8 h after each dose of CDV. Patients receiving zidovudine did not take that drug on the days they received CDV. Otherwise, changes in antiretroviral therapy were avoided whenever possible. Because of the long intracellular activity of CDV, end of therapy was considered to be 2 weeks after the last dose.
Patients were assessed before doses of CDV and 1 month after the end of therapy. Toxicity was graded according to the National Cancer Institute common toxicity criteria (vers. 1.0). Assessment of tumor extent (T) at entry was made by use of the criteria established by the AIDS Clinical Trials Group (ACTG) [10]. HIV-infected patients also were stratified on the basis of immune status and systemic illness. Periodic photographs were taken to document the changes in the KS lesions. KS response assessments were made by use of a minor modification of the ACTG criteria, as described elsewhere [10, 11].
For HIV-infected patients, plasma HIV-1 mRNA levels were measured by use of the ultrasensitive Roche Amplicor monitor kits, with a lower limit of detection of 50 virions/mL. Six-millimeter punch biopsy specimens of cutaneous tumors were obtained at baseline and periodically to assess pathology, KSHV gene activation by in situ hybridization (ISH), and latency-associated nuclear antigen (LANA) by immunohistochemical staining [2, 12]. Virus loads of KSHV and Epstein-Barr virus (EBV) were determined in peripheral blood mononuclear cells (PBMC) and in plasma by use of a quantitative real-time polymerase chain reaction assay, as described elsewhere [13]. Serum KSHV vIL-6 was measured use of a specific ELISA [14]. Serum human IL-6 was measured by use of the hIL-6 Quantikine kit (R&D Systems). Plasma VEGF levels were measured by use of the CytElisa assay (Cytimmune Sciences).
The study was designed as a standard 2-stage, phase II study (with 15 patients in the first step and 10 in the second) and to have an 80% probability of detecting a 30% response rate. However, patients became increasingly reluctant to enroll after learning the initial results, and the protocol was terminated after 7 patients had entered. The virus load of KSHV and EBV in PBMC at entry and at the end of treatment were compared by use of the Wilcoxon signed-rank test. Values were logarithmically transformed before analysis. The relationship between entry CD4+ cell counts and time to progression on CDV was assessed by use of the Spearman's correlation coefficient. All P values are 2-sided.
Results. Seven male patients with KS were entered, of whom 2 had classical KS, and 5 had HIV-associated KS. All had poor-prognosis KS on the basis of tumor involvement (T1), and 6 patients had >50 lesions. Of the 5 HIV-infected patients, 1 had poor-prognosis KS on the basis of immune status and 2 had poor-prognosis KS on the basis of systemic disease. 1 HIV-infected patient had MCD. Four of the 5 HIV-infected patients were receiving a stable regimen of 3 antiretroviral drugs for a median of 65.5 weeks (range, 24140 weeks). The fifth had an HIV RNA level <5000 copies/mL and a CD4+ cell count of 1041 cells/mm3 without receiving antiretroviral therapy. No patient changed antiretroviral therapy during study.
fig.ommitted
Table 1. Patient characteristics at entry.
CDV was well tolerated for a median of 11 weeks (range, 530 weeks). All clinical toxicities thought to be possibly related to CDV were grade 1. The maximum proteinuria was 1+ (3 patients), and no patient had elevated creatinine levels. All 7 patients had progressive KS while receiving CDV (25% increase in the no. of total lesions, no. of nodular lesions, or size of 5 marker lesions, confirmed by reassessment in at least 1 week). Median time to progression was 8.1 weeks (range, 527 weeks). No patient had a partial or complete tumor response (95% CI for response, 041). There was no correlation between CD4+ cell count and time to progression (r = 0.29; P = .53). In 4 patients, CDV was continued for 314 weeks after tumor progression, but there was no evidence of late regression.
In 2 patients, the virus load in KSHV in PBMC was undetectable at entry and at the end of therapy. The KSHV virus load increased in 4 patients and decreased in 1 (P = .19 for the 7 patients). There was a strong trend upward in PBMC-associated EBV, with the median values at entry and the end of therapy being 133 and 935 copies/106 PBMC, respectively (P = .078). Plasma KSHV was either undetectable or at the limit of detection (45 copies/mL) at entry and at the end of therapy for 7 patients. Plasma EBV was undetectable in 6 patients throughout therapy but, in the patient with MCD, increased from <45 to 375 copies/mL.
KS skin biopsy specimens were obtained from 5 patients at entry and the end of therapy; an additional biopsy specimen was obtained from 3 patients at week 7 or 8. Biopsy specimens from different lesions were obtained before and after therapy. There were no consistent changes by light microscopy. Three of the entry biopsy specimens showed expression of T0.7 (latent gene) by ISH, and there was no clear change during therapy . Similarly, there was no change in the staining for LANA (latent gene) or in ISH for in v-cyclin (latent gene) in the 3 patients studied. T1.1 (early lytic gene) expression was observed at some point in 2 of 5 patients studied, but there were no consistent changes. Major capsid protein (MCP; late lytic gene) was examined in 3 patients. In 2, no expression was seen at any time, whereas, in 1, there was a decrease in the number of positive cells from 9 per slide at entry to 4 per slide at week 8 to 3 per slide at week 11 (time of progression.
fig.ommitted
Figure 1. Expression of latency-associated nuclear antigen (LANA) (A and B); T0.7, a latent gene (C and D); T1.1, an early lytic gene (E and F); and major capsid protein, a late lytic gene (G and H), in Kaposi sarcoma biopsy specimens from a representative patient. A and B, Cells expressing LANA show nuclear staining by immunohistochemistry. CH, In situ hybridization was used to identify cells expressing the relevant genes; these cells were visualized by silver grains that have developed in the photographic emulsion coating the specimen. Panels A, C, E, and G are from a biopsy specimen obtained at entry, whereas panels B, D, F, and H are from a biopsy specimen obtained at the end of therapy (week 11). A 50-m bar is shown in panel A.
Serum vIL-6 was detectable at entry only in the patient with MCD. In that patient, vIL-6 decreased from 5027 pg/mL at entry to 933 pg/mL at week 11. There was, however, no change in MCD tumor size by computer-assisted tomography. Otherwise, vIL-6 was detected only transiently in 1 other patient at week 7. Serum human IL-6 was detectable at entry in all patients (median, 4.63 pg/mL; range, 0.749.16 pg/mL). The last sample measured on therapy was higher in 6 of 7 (median, 4.94 pg/mL; range, 0.5632.96 pg/ml; change not significant). Plasma VEGF was detectable in all patients at entry (median, 7.4 pg/mL; range, 5.89.6 pg/mL), and there was no trend during treatment (median, 7.14 pg/mL; range, 5.4620.19 pg/mL in the last samples). In 2 of 5 HIV-infected patients, the HIV load remained undetectable or just above the limit of detection. Two others had a decrease (0.43 and 0.78 log10 copies/mL), whereas the fifth had an increase (0.17 log10 copies/mL). There was no consistent change in the CD4+ cell count.
Discussion. We found that each of 7 patients with either classical or AIDS-related KS had progressive disease while receiving CDV. It is noteworthy that 4 patients also were receiving HAART for HIV-associated KS. Although the data must be interpreted in light of the small number of patients, they suggest that this CDV regimen has little or no activity in patients with established KS.
Why was no KS effect seen? It has been learned that the KSHV genes that promote spindle-cell growth are early lytic genes [1]. CDV inhibits late lytic genes but not latent or early lytic genes. Thus, the growth of spindle cells stimulated directly or indirectly by early lytic KSHV gene products might have been unaffected by the treatment.
Another question is whether CDV had an anti-KSHV effect. It was noteworthy that the virus load of KSHV in PBMC increased in 4 of 5 patients in which it was detectable. However, antiviral drugs do not inhibit the replication of herpesvirus episomal DNA in latently infected dividing cells, because this relies on cellular DNA polymerase, and the lack of a decrease of KSHV or EBV in PBMC does not demonstrate a lack of antiherpesvirus effect. We attempted to assess KSHV replication and expression of late lytic genes. Unfortunately, the viral load of plasma KSHV was at or below the limit of detection. Also, only 1 of the patients had expression of the late lytic gene MCP in KS as assessed by ISH. Although subsequent biopsies in this patient showed a decrease, these data are too limited to conclude that an anti-KSHV effect was seen.
CDV was used because of its potent anti-KSHV activity in vitro. Several lines of evidence suggest that the dose schedule used was likely to be active against KSHV. CDV administered in this regimen is effective in patients with cytomegalovirus (CMV) infection. However, CDV has been reported to be somewhat less active against KSHV than against CMV, perhaps in part because it is phosphorylated relatively poorly in lymphocytes used to study it in vitro [79, 15]. CDV also has relatively poor activity against herpes simplex viruses 1 and 2 in vitro but still has been reported to be clinically active against these viruses. Taken together, this evidence suggests that CDV at the dose schedule used was reasonably likely to have activity against KSHV. However, we were not able to verify that this was the case.
The study does not provide a proof of principle for the treatment of KS with anti-herpes therapy. However, it does not disprove a role for specific antiviral therapy in the treatment of virus-induced tumors, including KS. Virus-encoded genes play an important role in the pathogenesis of a number of tumors, and these offer potential targets for tumor therapy. The development of such therapy will be aided by our increased understanding of the role of various viral genes in tumor pathogenesis.
Acknowledgments
We thank the patients who volunteered for this study; Darlene Rowe of Gilead Sciences; Scott Whitcup of the National Eye Institute; Stephen Straus of the National Institute of Allergy and Infectious Diseases; Seth M. Steinberg, Betsy Read-Connole, and Ellen Feigal of the National Cancer Institute (NCI); Randy Stevens, Michael Baseler, Wendell Miley, and David Waters of Science Applications International; Kathleen M. Wyvill, the other research nurses, and the data management team of the HIV and AIDS Malignancy Branch, NCI; the medical staff of the Medical Oncology Program of the NCI; and the nursing, pharmacy, social work, and medical staff of the National Institutes of Health Clinical Center.
References
1.
Antman K, Chang Y. Kaposi's sarcoma. N Engl J Med 2000; 342:102738.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.