Active Immunotherapy with Ultraviolet B-Irradiated Autologous Whole Melanoma Cells Plus DETOX in Patients with Metastatic Melanoma1

Vol. 4, 6/9-627, Marc/i 1998 Clinical Cancer Research 619

Active Immunotherapy with Ultraviolet B-irradiated Autologous Whole Melanoma Cells plus DETOX in Patients with Metastatic Melanoma1

Omar Eton,2 Dmitri D. Kharkevitch, fall-off heralded progressive disease. Late induction of a weak DTH reaction to AT cells was observed in eight pa. Mary Ann Gianan, Merrick L Ross, Kyogo Itoh, tients. Active immunotherapy with UVR-AT + DETOX had Michael W. Pride, Cherrie Donawho, modest but definite clinical activity in advanced melanoma. Antonio C. Buzaid, Paul F. Mansfield, The induction of both PBMC cytotoxicity and DTH reactiv- Jeffrey E. Lee, Sewa S. Legha, Carl Plager, ity to AT cells supported a specific systemic immune effect of Nicholas E. Papadopoulos, Agop Y. Bedikian, treatment, although the former more closely followed dis- Robert S. Benjamin, and Charles M. Balch ease course in this study. Departments of Melanoma/Sarcoma Medical Oncology [0. E.. M.A.G.,A.C.B.,S.S.L.,C.P..N.E.P..A.Y.B.,R.S.B.]. INTRODUCTION Surgical Oncology [D. D. K., M. I. R.. P. F. M., J. E. L.], and Malignant cells must elude any host immune surveillance Immunology [M. W. P.. C. D.]. The University of Texas M. D. to grow and overcome the host. Hewitt et a!. ( I ) have shown that Anderson Cancer Center, Houston, Texas 77030: Department of Immunology. Kurume University School of Medicine. Kurume, spontaneous tumors in aging mice do not induce a detectable Fukuoka 830, Japan [K. I.]: and City of Hope National Medical host immune response. Nevertheless, the hypothesis that those Center and Beckman Research Institute, Duane, California 91010 same cells that escape immune surveillance in vivo might be [C.M.B.] manipulated in vitro to become more immunogenic is a basis for tumor immunotherapy. Methybcholanthrene and UVB-induced tumors, when excised and reimplanted in mice, have resulted in ABSTRACT varying degrees of specific immune protection (2-5). In pa- Our objective was to determine the clinical activity, tients, tumor burden, immunosebection (6), tolerance induction toxicity, and immunological effects of active immunotherapy (7-I 0), and biophysical barriers ( 1 1 ) all present challenges to using UVB-irradiated (UVR) autologous tumor (AT) cells developing effective antitumor immunotherapy programs. plus adjuvant DETOX in metastatic melanoma patients. Efforts to circumvent these obstacles in the 1990s are Eligibility included nonanergic patients fully recovered after focused on identifying specific factors that either drive or inhibit resection of 5 or more grams of metastatic melanoma. Treat- antitumor immunity (12). However, while the relative impor- ment consisted of intradermal injections of iO UVR-AT tance of these factors continues to evolve almost on a daily plus 0.25 ml of DETOX every 2 weeks x 6, then monthly. basis, cruder vaccine preparations from the patients’ tumors Peripheral blood mononuclear cells (PBMCs) were har- continue to have a role in providing a plethora of characterized vested for cytotoxicity assays, and skin testing was per- and as yet unidentified targets. These vaccines have resulted in formed for delayed-type hypersensitivity (DTH) determina- objective response rates under 20%, including responses in tions before the first, fourth, seventh, and subsequent visceral organ metastases consisting of billions of tumor cells treatments. Forty-two patients were treated, 18 in the adju- (13-15). Such clinical activity supports the continued develop- vant setting and 24 with measurable disease. Among the ment of active immunotherapy in the multidisciplinary antican- latter group, there were two durable responses in soft-tissue cer armamentarium. Viable autobogous whole tumor cells tested sites and in a bone metastasis. Treatment was well tolerated. in melanoma vaccine preparations have some advantages: (a) Thirty-five patients were assessable for immunological pa- the relative safety of intradermab injection of autobogous rather rameters; 10 of these patients, including the 2 responders, than ablogeneic material; (b) the potential for inducing as yet demonstrated early induction of PBMC cytotoxicity against uncharacterized immune responses to altered self-antigens in AT cells that persisted up to 10 months on treatment before common with the patient’s tumor; and (c) the diminution of falling to background levels. In five of seven patients, the some artifacts of vaccine preparation. Disadvantages include: (a) the difficulty in obtaining and preparing individualized vaccines; (b) the limited source material; and (c) the inability to control for intervaccine variation important in clinical trials. Received 6/12/97: revised 12/5/97: accepted 12/15/97. At M. D. Anderson Cancer Center, Hostetler et a!. (16-18) The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked demonstrated that UVR3 (280-320 nm) of the weakly immu- advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

I This work was supported in part by NIH Grant RR-551 1.

2 To whom requests for reprints should be addressed, at The University of Texas, M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, 3 The abbreviations used are: UVR, UVB irradiation; AV, antigenic Box #77, Houston, TX 77030. Phone: (713) 794-4633: Fax: (713) 794- variant; PBMC, peripheral blood mononuclear cell; DTH, delayed-type 1934. hypersensitivity; AT, autologous tumor.

nogenic melanoma cell line K1735P (C3H/HeN; MTV) re- required to avoid immunomodubatory agents (e.g., corticoste- sulted in the generation of a high frequency of highly immuno- roids, nonsteroid anti-inflammatory agents, antihistamines, es- genic AVs. Immunizing syngeneic mice with AV clones trogens, high doses of vitamins, and concurrent specific antitu- induced protective immunity in in vivo challenge assays against mon therapy). All sites of clinically detectable disease were the immunizing AV clone and the untreated parental tumor line objectively evaluated at baseline using computed tomography, but not against unrelated tumors (16-18). The parental tumor magnetic resonance imaging, sonography, photographs, and, for did not induce such protective immunity. These AV clones grew s.c. disease, multiple investigator assessment. in immunosuppressed mice but not in immunocompetent mice, Vaccine Preparation. Fresh tumor was collected at the and they induced cytotoxic/helper I-lymphocyte activity that time of surgery from the frozen section laboratory and frag- cross-reacted with parental tumor cells (18). mented by slicing. Aliquots of 106 cells each were frozen for Based on these data, it was hypothesized that the poly- immunoassays. To maximize yield of viable tumor cells for cbonab population of UVR-parental cells might provide a mul- vaccine preparation, the bulk of tumor was dissociated using titude of immunogenic AVs for active immunotherapy without colbagenase type I (2 mg/mb) and type IV DNase (0.4 mg/ml; requiring individual AV clones to be isolated. Umezu et al. (19) Sigma Chemical Co., St. Louis, MO; Ref. 25). These enzymes developed conditions for UVR of human melanoma and renal can alter the immunogenicity of the resulting cell preparation cell carcinoma cells to optimize inducing cytotoxic activity of (26-29). In control flow cytometry experiments, Itoh et a!. (26) PBMCs against irradiated- and nonirradiated parental cells. Us- did not find enough significant alteration to obviate the use of ing in vivo immunization/challenge and DTH experiments, Pride the enzymes; furthermore, incubation of the tumor cells for 24 h et a!. (20, 2 1) demonstrated that UVR of the K! 735-M2 cell line before vaccine administration, as described below, resulted in greatly increased its immunogenicity. Pride et a!. (21) also reexpression of specific suppressed antigens (26, 29). found that UVR not only up-regulated expression of MHC The dissociated cells were washed in sterile HBSS with classes I and II molecules but also induced expression of murine gentamicin and resuspended in equal volumes of HBSS and B7.l on Kl735-M2 melanoma. chilled 10% DMSO + 4% human serum albumin. Aliquots Mycobacteriab preparations are useful adjuvants in active containing 1.5-2 x i0 viable tumor cells were stored under immunotherapy (22). In melanoma, Bacillus Calmette-Gu#{233}rin liquid nitrogen. For some specimens, an aliquot of fresh tumor induces responses in up to 90% of locally injected and 15% of was also put into culture to establish a tumor cell line. adjacent lesions but has had no effect on survival in Phase III Twenty-four h before patient arrival, an abiquot of tumor trials and can cause disseminated infection. DETOX (Ribi Im- cells was thawed in a 37#{176}Cwater bath, washed twice, resuspended in 5 ml of PBS in a 10-cm Petni dish, and exposed to munoChem Research, Inc. , Hamilton, MI) is a novel adjuvant available in a lyophilized oil droplet emulsion consisting of both UVR (2.2 J/m2/s) for 30 s using an FS4O sunlamp (Westing- cell wall cytoskeleton from Mycobacterium phlei and mono- house Electric Corp., Bboomfield, NJ; Refs. 18 and 19). Incu- phosphoryl lipid A from Salmonella minnesota R595. DETOX bation in macrophage-/serum-free medium (Life Technologies, can augment I cell and humoral immunity. In the murine Inc.) for 24 h resulted in >80% cell viability. Cell suspensions experiments by Pride et a!. (20), DETOX was effective in were checked for bacterial contamination and for Mycoplasma significantly augmenting immunoprotection and DTH. In pa- contamination using the Gene-Probe rapid detection system tients, DETOX has been used only as an adjuvant, and dosing (Gene Probe, Inc., San Diego, CA). After incubation, cells were has been empiric because its effects have been more host than detached by scraper, washed, resuspended in 0.75 ml of PBS, dose rebated (23, 24). and X-irradiated (20,000 cGy) to prevent propagation in vivo We undertook the present study to determine the clinical and insure sterility (30, 31). Concurrently, DETOX (each vial activity and toxicity of active immunotherapy with UVR-AT stored at 2-7#{176}Cconsisted of 500 p.g of cell wall cytoskeleton, 50 cells + DETOX in patients with metastatic melanoma. We used pg of monophosphoryl lipid A, 2% squalene, and 0.4% Tween serial monitoring of PBMC cytotoxicity and DTH reactions 80 in 2X normal saline) was reconstituted by repeated aspiration against AT cells with controls to detect and track any evidence with 0.5 ml of sterile water. In a total volume of 1 ml, approx- of a specific systemic immune response to local intradermal imateby 10 viable replication-deficient UVR autologous mela- therapy during the course of vaccine treatment. noma cells in PBS were mixed with 0.25 ml of DETOX just before each treatment (24). Treatment Plan. Active immunotherapy was adminis- PATIENTS AND METHODS tered at M. D. Anderson Cancer Center by intradermal injection

Patients. Patients > 15 years with metastatic melanoma into the anteromedial deltoid and femoral regions near intact pathologically confirmed at M. D. Anderson Cancer Center and lymph node basins, every 2 weeks X 6 and then every 4-6 from whom more than S g of nonnecrotic tumor had been weeks. Treatment ended with progression of disease (25% resected as a part of routine management were candidates for growth) or exhausted supply of vaccine material. Response was treatment. Patients were enrolled only after full postoperative assessed initially at 6 weeks and then at 8-12-week intervals recovery and with: a Karnofsky performance status >70%; life during long-term follow-up. Toxicity was graded according to expectancy >3 months; DIH reaction positive to foreign anti- National Cancer Institute guidelines. gens (see below); results of complete blood counts, serum Cytotoxicity Assays. For uniformity, all cytotoxicity as- chemistry studies, immunogbobulins (IgG, IgA, and 1gM), com- says were performed by a single investigator (D. K.) throughout plement (C3, C4, and CHSO), and I-cell subsets in the clinically the study. At baseline and before the fourth, seventh, and sub- normal range; and signed informed consent forms. Patients were sequent vaccinations, PBMCs were isolated by Ficoll-Paque

Tab le I Pretreatme nt patient and tumor characteristics

n % of total Indicator lesions Adjuvant setting All patients 42 Age (yr): median, 52: range, 27-81 Male 24 57 Prior treatment None 15 36 Biotherapy 14 33 Chemotherapy ± biotherapy 21 50 Regional perfusion 2 5 Karnofsky performance status 90% 36 86 80% 6 14 Normal serum level of LDH” 38 90 Normal serum level of albumin 40 95

Tumor stage and sites

All patients 42 24 18 AJCC stage III N2 (regional) 6 1 5 AJCC stage IV (disseminated) 36 23 (55%) 13 (3 1%) Soft-tissue only 16 8 8 Visceral metastases 20 15 5 One visceral organ 12 8 4 Lung alone 7 5 2 Brain 6 5 1 Liver 4 3 1 Spleen 4 2 2 Adrenal 2 1 1 Gastrointestinal 2 0 2

C, LDH, lactate dehydrogenase: AJCC, American Joint Committee on Cancer.

gradient (Pharmacia AB, Uppsala, Sweden) from 30 ml of fresh treatment was defined as > 10% specific lysis and at beast 2-fold heparinized peripheral blood at room temperature (22). These higher specific bysis than at baseline. PBMCs were assayed fresh and were also cryopreserved until Antibody-blocking Experiments. To confirm and char- sufficient specimens could be obtained over 2 or more months acterize any positive cytotoxicity results detected, blocking ex- of treatment to perform a single experiment, thereby minimizing periments were performed using antibodies to CD3, CD4, CD8, interassay variability. Four-h ‘Cr-release assays were used to DR. DP, DQ (Becton Dickinson, Sunnyvale, CA), and class I detect cytotoxic activity, as described previously (32). Cryopre- MHC (W6/32; DAKO, Copenhagen, Denmark). Monoclonal served, uncultured autobogous and albogeneic melanoma cells, antibodies were mixed with ‘Cr-labeled AT targets in a volume renal cell carcinoma cells, cultured tumor cell lines, and natural of 50 p.1 each for 1 h at 4#{176}Ctotry to abrogate lysis. Effector cells killer-sensitive tumor line K562 were used as the targets in these were added at a 40: 1 E:T ratio, and cytotoxicity was measured assays. Frozen aliquots of uncultured tumor cells were thawed as described above. Controls were murine IgG and medium on the day of the assay; adherent cell lines were trypsinized. alone. Target cells were labeled with 50 p.Ci/million cells of DTH Assays. Intradermal skin testing on the vobar Na25tCrO4 (Amersham Corp., Arlington Heights, IL) at 37#{176}C surface of the forearm was performed at baseline against: for 1 h. After washing twice, labeled cell suspensions were foreign recall antigens (Merieux Skintest Multitest Device: passed through a cotton column to remove cell aggregates and tetanus, diphtheria, strep, tuberculosis, Candida, trichophy- debris. Labeled target cells at a concentration of 5000 cells/well ton, and proteus); l0 X-irradiated AT, and i05 autobogous in triplicate were added to effector lymphocytes at each of three PBMCs (negative control). AT and PBMC skin tests were E:T ratios, 40: 1, 20: 1, and 10: 1, in 96-well, round-bottomed repeated before the fourth, seventh, and subsequent vaccina- plates. The plates were incubated for 4 h at 37#{176}C,after which tions. A positive DTH reaction consisted of induration over 2 100 p.1 of supematant/well was harvested, and radioactivity was mm in diameter at 48 h. Hypoergy to foreign antigens using measured in a gamma counter. In all experiments performed, the the Merieux skin-test device was defined as two or fewer spontaneous release did not exceed 20% of the maximum positive DTH reactions or sum of induration diameters under release. Cytotoxicity assays were also performed with ad- 1 cm. In this study, patients were not asked to undergo punch herent melanoma cell targets in 96-well, flat-bottomed plates. biopsy of skin test sites to detect subclinical evidence of a The percentage of specific lysis was calculated as: (Xa x)/ DTH reaction. (Xmax x) X 100, where Xa 5 the mean ‘ Cr release at a Statistical Considerations. Complete or partial re- specified E:I ratio; x is the mean spontaneous 51Cr release (no sponse required disappearance or >50% decrease, respec- effector cells); Xmax 5 the maximum 5tCr release by target cells tively, in the sum of the products of largest diameters of all treated with 0. 1 N HC1. Positive cytotoxicity as a result of indicator lesions lasting at beast 1 month. Stable disease

- -.

Fig. 1 Clinical responses to active immunotherapy using autologous UVR-melanoma cells + DETOX. A, partial response in bone 6 months from the start of treatment; this patient also had a complete response in s.c. nodules. B, single view of some of several dozen in-transit lower-extremity metastases which, after an initial flare at 10 weeks, disappeared over a period of 17 months from the start of treatment; surgery to harvest more cells for vaccine preparation resulted in this patient’s formal response being characterized as mixed to vaccine and complete to vaccine plus surgery.

required less than 25% change in the sum of the products of RESULTS diameters lasting at least 8 weeks. Survival from start of Between September 1992 and May 1995, 42 patients re- treatment was plotted by the method of Kaplan-Meier and ceived a median of eight treatments (range, 3-21) with UVR- evaluated using the bog-rank test. PBMC cytotoxicity and autobogous melanoma cells plus DETOX. Pretreatment charac- DTH reaction to AT were assessable only after six or more teristics are summarized in Table I . All patients were assessable treatments. A modification of Simon’s design for Phase II for clinical response and toxicity and 35 for biological response. clinical trials was used; assuming a 10% significance level Seven patients had disease progression before completing the and 90% power, I 2 patients were enrolled in stage I and 23 first six treatments. additional patients in stage II (32). Induction of PBMC 24 patients with indicator cytotoxicity in 6 of 35 patients would support further inves- Clinical Responses. Among lesions, two had major responses and six showed stabilization in tigation, assuming that response rates of 30 and 10%, respec- tively, would and would not be of interest. Enrollment would soft-tissue metastases. One 39-year-old woman treated previ- have terminated early if none of 12 patients achieved a ously with chemo/biotherapy underwent resection of painful s.c. clinical response or if 6 of 9 patients treated in an adjuvant nodules for vaccine preparation. Her indicator lesions consisted setting had progressive disease within 3 months. of nine s.c. nodules, each 1 cm or smaller, and a large sympto-

40 A 93 consistently had a large DTH reaction (3 cm) to vaccine but never to the AT control. .‘ 30 Survival. The median survival of the stage IV patients #{149}0 was 16 months from start of treatment, 20 months from initial 20 diagnosis of stage IV disease. Review of pretreatment prognos- 0 tic factors revealed a high frequency of favorable factors (nor- a 10 mal serum levels of lactate dehydogenase; Refs. 33 and 34), Before 0 I! metastases in soft tissue sites alone, or metastases in one visceral Vaccine #: I 4 IlLI7 8 9 10 11 12 LAK organ such as lung (33), in this study. Months: 3 6 9 12 Toxicity. Grade I toxicities were limited to local irrita-

- 40:1 I tion at the vaccine injection site and in 10% of patients. with - 20:1 I malaise lasting 24-48 h. There was no incidence of skin ulcer- 10:1] ation or other grades I1-IV toxicities. Palpable residual granu- 40 B bomas at the injection sites measuring <2 cm became less

‘ 30 apparent over weeks to months. In one-half of the patients, U routine evaluation by an ophthalmologist yielded no abnormal- 20 ities attributable to treatment. Five of 15 patients (33%) purified

0 protein derivative negative at baseline converted to purified a 10 protein derivative positive (asymptomatic) using the Merieux InjI Icil skin-test device, possibly a result of repeated exposure to DE- Before 0 - Li I lOX (or to the anergy panel). Vaccine #: I 4 7 9 10 11 12 14 15 1617 18 LAK Months: 0 3 6 9 12 15 Detection of PBMC Cytotoxicity against Autologous Tumor. During the course of treatment, 10 of 35 assessable Fig. 2 Mean PBMC cytotoxicity over time against autologous tumor at three E:T ratios for each of the two responders (A and B) shown in Fig. patients had an early rise in mean PBMC cytotoxicity against I. L4K, bymphokine-activated killer activity (positive control). AT cells within 6 weeks, all from a negative baseline of under 10% specific lysis (Table 2). Lytic activity remained detectable up to 10 months, later falling back to background bevels in 7 of the 10 patients. In five patients, the fall-off coincided with matic lytic bone metastasis that had progressed radiographically progression of disease within 3 months. For the three patients 1 year after radiation therapy. Four days after her first treatment, without fall-off, autologous material was insufficient to continue she noted soreness at all tumor sites. Three skin nodules initially bong-term treatment and monitoring. enlarged by less than 20% with minimal inflammatory changes, Table 3 illustrates the individual spectrum of specificity for three remained stable, and three regressed; however, by the PBMC cytotoxicity. The selective rather than general positivity seventh treatment (at 4 months), all s.c. lesions had completely in the cytotoxicity assay results against a panel of allogeneic regressed as had her bone pain, with subsequent confirmation of tumor cells, and K562 does not support a non-specific effect of major radiological response (Fig. 1A). Nine months after her last vaccine preparation, such as the use of collagenase. In four treatment, she developed new s.c. nodules and recurrence in patients, the level of lysis of K562 cells was minimal, indicating bone (time to progression, 21 months). From a negative base- that the cytotoxicity was not mediated by natural killer cells. line, her unstimubated PBMCs demonstrated persistently posi- Table 4 shows that lysis could be abrogated by anti-CD3, tive cytotoxicity against AT cells (Fig. 2A). She had an unusu- anti-class I, and anti-CD8 but not by anti-class II and anti-CD4 ally large DTh reaction (4 cm with 10 cm of surrounding antibodies, indicating that the observed cytolytic activity was inflammation) at the first vaccine intradermal injection site, restricted by class I and was mediated by CD8 + I cells. smaller with subsequent treatments. No DTH reaction to the i0 Univariate analysis indicated that patients with stage IV AT control was detected during 1 year of treatment. disease and positive cytotoxicity had both a prolonged median The second responder was a 63-year-old man with exten- time to progression and overall survival duration compared with sive, progressing, lower-extremity, in-transit nodules, despite those with negative cytotoxicity (Table 5 and Fig. 3). Although regional and systemic chemo/biotherapy and radiation therapy. time bias was minimized because PBMC cytotoxicity was al- Resected symptomatic lesions were used for vaccine prepara- ways detected early, the positive cytotoxicity group also had a tion. Innumerable in-transit nodules, 0.5-2 cm in caliber, were lower disease burden, confounding the results of the analysis. used as indicator lesions. By 10 weeks, most lesions had flat- DTH. Among 42 nonanergic patients, 21 (50%) were tened or remained stable; however, a minority appeared more hypoergic to foreign recall antigens at baseline, a finding that prominent, although by <20%. Because of early resection of did not correlate with any pre- or posttreatment factors. Baseline these lesions to make more vaccine, this was strictly a “mixed” hypoergy was evident in 4 of the 10 patients with positive response. However, after 19 treatments for >20 months, to- PBMC cytotoxicity results (Table 5). gether with surgery, he ultimately achieved a complete response Among the 35 assessable patients, 8 (23%) developed a bate lasting 5 months (Fig. 1B). From a negative baseline and ex- DIH reaction to iO AT cells (and no reaction to PBMC tending through the 16th treatment, his unstimulated PBMCs control); indurations under I cm first appeared after a median of demonstrated cytotoxicity against AT cells (Fig. 2B). Fall-off in 8 courses (4-15 courses). Neither responder and only three cytotoxicity coincided with progressive soft-tissue disease. He patients with positive PBMC cytotoxicity also had a DTH

Table 2 Ten patients with specific lysis > 10% and rising over 2 X baseline during the course of treatment

E:T I” 4” 7” 8” 9” ioa 11” 12” 13” 14” 15” 16” Patient 7b 0” 1b 4/’ _t 6” 11” 12” 14” 16” 17b 40:1 8 14 21 48 49 31 27 55 51 2 0 20:1 4 14 21 28 34 22 47 8 2 3 10:1 12 3 19 29 22 15 22 6 2 3 2 40:1 4 15 14 15 20 10 9 2 5 20:1 5 13 17 5 17 6 6 I 3 10:1 4 10 3 0 5 8 5 0 1 3 40:1 6 22 23 28 30 31 47 16 10 5 4 0 20:1 0 15 23 17 23 23 15 19 7 4 2 0 10:1 0 10 15 8 2 5 15 I 1 4 3 0 4 40:1 0 10 0 6 21 24 30 23 15 9 0 20:1 1 9 1 5 23 12 17 11 10 8 0 10:1 0 9 0 5 15 2 2 0 8 0 3 5 40:1 2 6 10 15 15 10 13 20:1 5 0 5 5 10 9 4 10:1 1 0 0 I 7 3 0 6 40:1 4 24 18 38 46 15 1 7 3 0 0 3 20:1 0 17 3 21 18 8 0 1 3 2 0 0 10:1 0 8 0 11 13 8 0 0 2 1 0 0 7 40:1 8 9 12 11 16 12 8 12 13 5 20:1 5 10 8 0 6 14 1 2 1 3 10:1 0 0 3 0 0 5 3 0 0 6 8 40:1 0 1 2 16 20 16 23 18 28 20:1 I 0 1 6 14 11 12 12 16 10:1 0 0 2 2 7 1 2 8 7 9 40:1 2 19 24 25 13 26 36 27 20:1 0 7 18 19 4 17 19 17 10:1 0 1 6 3 4 3 4 5 10 40:1 1 0 0 27 30 28 8 0 20:1 1 3 1 23 33 17 0 0 10:1 0 3 0 9 10 5 0 0

‘I Vaccine number.

I, Months.

Table 3 SpecifIcity of positive PBMC cytotoxicity in four patients heavy infiltration by histiocytes and I-cells but not by B cells (data not shown). Cytotoxicity against targets

Patient Autologous Allogeneic Renal cell NK”-sensitive no. melanoma melanoma carcinoma K562 DISCUSSION

I + + - - Among 24 patients with indicator lesions, intradermal ac-

4 + - - - tive immunotherapy with UVR-AT + DETOX resulted in two 10 + ± ± ± major responses in soft-tissue sites and in a bone metastasis, 11 + + - - supporting clinically meaningful systemic activity, albeit in “ NK, natural killer. <20% of patients with 95% confidence. This is no different from the results reported in other vaccine trials and does not support using UVR alone to enhance the clinical activity of autologous whole-cell vaccines plus adjuvant. As in other such reaction to AT cells. Skin testing using UVR-AT was not trials, the overall survival data were encouraging but likely performed. Contaminants in preparation could have resulted in reflect selection bias. The present American Joint Committee on a high frequency of false-positive DTH reactions to both AT and Cancer criteria for stage IV disease only distinguish between PBMC control skin tests; this was not observed. No tests, such soft-tissue and visceral metastases in patients having a median as punch biopsy, were done to evaluate for false-negative overall of survival of 8 months. Newer, more effective stratifi- results. cation parameters for clinical trials using pretreatment serum The size of the DTH induration to the vaccine itself (UVR- levels of lactate dehydogenase and albumin (33-34) and the

AT + DETOX), measuring primarily response to adjuvant, was presence of visceral disease in zero-to-one organ versus more- above the average and median values in 7 of the 10 patients with than-one organ (33) have identified prognostic groups with positive cytotoxicity, including the two responders and the three median survivals of over 12 months versus <6 months. Until a patients with a DTH reaction to AT. No consistent trend in the regimen with a high response rate is found, stratification might size of the DTH reaction to vaccine was observed in individuals help determine which active immunotherapy trials warrant con- or in population groups over time. Biopsy and staining of a firmation of “encouraging” survival data in a Phase III trial 1-month-old indurated injection site in one patient revealed setting.

Table 4 Antibody inhibitio n of PBMC cyto toxicity against aut ologous melanoma

Patient Media no. alone IgG anti-CD3 anti-CD8 anti-class I anti-CD4 anti-DR anti-DP

1 50” 56 14 13 12 54 55 52 4 47 53 0 1 1 45 46 48 11 46 43 7 4 6 42 41 45

a Percentage of cytotoxicity.

Table 5 Characteristics of patients with positive and negative I : H1 , detection of PBMC cytotoxicity against autologous tumor

PBMC c ytotoxicity 0.8

Positive Negative z C 0.6 Patients (n = 35) 10 25 Male 8 14 C Median age (yr) 53 52 0.4 4---+ - Q. Prior treatment PBMC Cytotoxicity Total Dead None 50% 36% Biotherapy alone 10% 8% 0.2 Positive 9 3

Chemotherapy ± biotherapy 40% 56% Negative 20 12 P a 0.009 Tumor stage and site(s):

AJCC stage III I 5 0 12 24 36 48 AJCC stage IV 9 (90%) 20 (80%) MONThS Adjuvant 6 6 Soft-tissue only or one visceral organ 8 13 Fig. 3 Kaplan-Meier survival curves for 29 assessable patients with Two or more visceral organs 1 1 stage IV melanoma and similar pretreatment prognostic features dis- Major objective response 2 0 tinguished by the presence or lack of detectable PBMC cytotoxicity Median time to progression (mo) 18 3 against autologous tumor. Median survival (mo) All patients 28+ 16 Stage IV patients 28+ 14 DTH (Table 4). The fall-off of detection of PBMC cytotoxicity con- Hypoergy to foreign antigens 40% 68% current with tumor progression in some patients suggested in- Positive reaction to 10 AT 30% 20% duction of tolerance. It must be noted, however, that the periph- Larger than average reaction to vaccine 70% 38% a AJCC, American Joint Committee on Cancer. eral blood is only one pharmacological compartment (36), and this study did not account for efficient trafficking of cytotoxic I cells to metastatic deposits and any resulting immunoselection (37). With regard to laboratory tests that might also guide such The induction of a DTH reaction to AT cells was not determination, none to-date has proven reliable, in part because observed in clinical responders. Subclinical reactivity could of the dearth of correbatable clinical responses and in part have been evaluated further by punch biopsy of skin test sites; because of the inability to demonstrate reproducible dose- this would require an informed consent from the patients. The response relationships in a heterogeneous patient population. prolonged time to first clinically apparent DTH reaction to AT Posttreatment laboratory results cannot be stratified prospec- cells (median of 8 courses) introduced a substantial “time bias” tively in Phase I or II studies because of time bias and other precluding developing this test for prognostic purposes (35). factors that complicate any analysis of such data (35). Among all DTH assays performed, only above average DTH The serial detection of PBMC cytotoxicity against AT in reaction to vaccine, largely a response to the adjuvant DETOX, 10 of 35 assessable patients (Tables 2 and 5) supported induc- correlated favorably with the clinical course of disease. The tion of a specific systemic immune response to intradermal study was too small to evaluate the effect of concurrent detec- treatment, a primary end point of this study. There are no data tion of both PBMC cytotoxicity and DTH reactivity to AT cells, to permit a comparison of the frequency of induction of positive observed in only three patients. cytotoxicity with that using a non-UVR-AT + DETOX vaccine Demonstration of UVR-induced expression of B7. 1 on because no patient received such treatment in this pilot study. human melanoma, as reported recently on murine melanoma Among the stage IV patients, almost uniformly with favorable (21) as well as on CT-26 (colon carcinoma) and HCA (hepato- clinical pretreatment prognostic features, detection of PBMC cellular carcinoma) cell lines,4 could support using UVR to cytotoxicity against AT cells correlated significantly with sur- study the antigen-presenting capability of tumor cells without vival (P = 0.009; Fig. 3). Different patterns of specificity were resorting to gene transfection. Besides the ongoing character- observed, suggesting that the positive results were not likely to be an artifact of the assays themselves (Table 3). Also, MHC class I-restricted CD8+ I cells and not natural killer cells were the primary mediators of this activity in the patients tested 4 M. Pride, unpublished data.

ization of the positive cellular immune responses as presented 10. Wiertz, E. J., Tortorella, D., Bogyo, M., Yu, J., Mothes, W., Jones, above and in cytokine elaboration experiments, future efforts T. R.. Rapoport. T. A., and Ploegh, H. L. Sec6l-mediated transfer of a membrane protein from the endoplasmic reticulum to the proteasome for will be geared to testing cytokines that drive TH’ immune destruction. Nature (Lond.), 384: 432-438, 1996. responses (25) or that recruit dendritic cells or other professional 11. Jam, R. K. Barriers to drug delivery in solid tumors. Sci. Am., 271: antigen-presenting cells to the vaccine injection site (15). 58-65. 1994. Suto and Srivastava (38) has been able to extract from 12. Cheson, B. D., Phillips. P. H., and Sznol, M. Trials using tumor autobogous tumor cells heat shock protein-96, a chaperon that vaccines. Oncology (Basel), 11: 81-90, 1997. complexes with the cells’ unique repertoire of soluble immuno- 13. Berd, D.. Maguire, H. C., Jr., McCue, P., and Mastrangelo, M. J. genic peptides (HSPPC-96). He has shown that HSPPC-96 can Treatment of metastatic melanoma with an autologous tumor cell vac- elicit antigen-specific CILs. Treatment with HSPPC-96 could cine: clinical and immunologic results in 64 patients. J. Clin. Oncol., 8: provide an array of autobogous tumor-specific peptide targets for 1858-1867, 1990. CIL activation without the need to characterize each peptide or 14. Morton, D. L., Foshag, L. J.. Hoon, D. S., Nizze, J. A., Wanek, to exclude patients based on human leukocyte antigen pheno- L. A.. Chang. C., Davtyan, D. 0.. Gupta. R. K.. Elashoff, R., and Irie, type. This may prove to be a far more exploitable and versatile R. F. Prolongation of survival in metastatic melanoma after active specific immunotherapy with a new polyvalent melanoma vaccine. Ann. strategy for immunotherapy than trying to convert an autologous Surg.. 216: 463-482, 1992. tumor cell that has already eluded host immune surveillance into 15. Leong, S. P. L., Enders-Zohr. P., Zhou, Y. M., Allen, R. W., Jr.. a better antigen-presenting cell. Autologous tumor HSPPC-96 Saagebiel, A. B., Glassberg, A. B., and Hayes, F. A. Active specific could be used in the future to prime dendritic cells activated or immunotherapy with GM-CSF as an adjuvant to autologous melanoma expanded in vitro with cytokines such as granubocyte/macro- vaccine in metastatic melanoma. Proc. Am. Soc. Clin. Oncol. Annu. phage colony-stimulating factor, interleukin 4, or flt3 bigand. Meet., 15: 437. 1996. No active immunotherapy program will be optimized, how- 16. Hostetler, L. W., Ananthaswamy, H. N., and Kripke. M. L. Gener- ever, until methods are developed in individual patients to ation of tumor-specific transplantation antigens by UV radiation can occur independently of neoplastic transformation. J. Immunol.. /37: circumvent immunoselection (6), local secretion of tolerizing 2721-2725. 1986. molecules by tumors (7-10), and biophysicab barriers that make I 7. Hostetler, L. W., and Kripke. M. L. Proportion of antigenic variants metastases sanctuary sites from the immune system ( I I). induced by iii vitro UV irradiation differs in clones derived from a single tumor. J. Immunol., /40: 666-670, 1988. 18. Hostetler. L. W., Romerdahl, C. A., and Kripke. M. L. Specificity ACKNOWLEDGMENTS of antigens on UV radiation-induced antigenic tumor cell variants meas- We thank Dr. Margaret Kripke for helpful suggestions and thought- ured in vitro and in vivo. Cancer Res., 49: 1207-1213, 1989. ful review of the manuscript. 19. Umezu, Y., Augustus. L. B.. Seito. D.. Hayakawa, K.. Ross, M. I., Eton. 0.. Swanson, D. A., and Itoh, K. Increase in the ability of human cancer cells to induce cytotoxic T lymphocytes by ultraviolet irradiation. REFERENCES Cancer Immunol. Immunother., 37: 392-399, 1993. 1. Hewitt, H., Blake, E., and Walder, A. A critique of the evidence for 20. Pride, M. L., Donawho, C. K., and Kripke, M. L. Active specific active host defense against cancer based on personal studies of 27 immunotherapy using UVB-irradiated whole melanoma cell vaccine. murine tumors ofspontaneous origin. Br. J. Cancer, 33: 241-259, 1976. Abstract 52 1 1, 9th International Congress of Immunology (San Fran- 2. Gross, L. Intradermal immunization of C3H mice against a sarcoma cisco), 1995. that originated in an animal of the same line. Cancer Res.. 3: 326-333, 21. Pride, M. L., Donawho, C. K.. and Kripke, M. L. UVB irradiation 1943. of tumor cells enhances the immunogenicity of melanoma cell vaccines 3. to methylcholanthrene- Prehn, R. T., and Main, J. M. Immunity by augmenting cell surface expression of MHC class I and II and B7-l. induced sarcomas. J. Natl. Cancer Inst., 18: 769-778, 1957. Proc. Am. Assoc. Cancer Res., 38: 630, 1997. 4. Klein, G., Sjogren. H.. Klein, E.. and Hellstrom, H. E. Demonstration 22. Shinomiya. Y.. Harada, M.. Kurosawa. S., Okamoto. T.. Terao, H.. of resistance against methylcholanthrene induced sarcomas in the pri- Matsuzaki, G., Shirakusa, T., and Nomoto, K. Anti-metastatic activity mary autochthonous host. Cancer Res., 20: 1561-1572. 1960. induced by the in vito activation of purified protein derivative (PPD)- 5. Kripke, M. L. Antigenicity of murine skin tumors induced by ultra- recognizing ThI type CD4+ T cells. Immunobiology, /93: 439-455, violet light. J. Natl. Cancer Inst., 53: 1333-1336, 1974. 1995. 6. Jager, E., Ringhoffer, M.. Altmannsberger, M., Arand, M., Karbach, 23. Schultz, N., Oratz, R., Chen, D., Zeleniuch-Jacquotte. A., Abeles, J., Jager, D., Ocsch, F., and Knuth, A. Immunoselection in i’iio: loss of G., and Bystryn. J. C. Effect of DETOX as an adjuvant for melanoma MHC class I and melanocyte differentiation antigen expression in met- vaccine. Vaccine, /3: 503-508, 1995. astatic melanoma. Abstracts 521 and M16, Cancer Vaccines, Cancer Research Institute (New York). 1996. 24. Mitchell, M. S., Harel, W., Kempf R. A.. Hu, E., Kan-Mitchell, J., Boswell, W. D.. Dean, G.. and Stevenson, L. Active-specific immuno- 7. Suzuki, T., Tahara, H., Narula, S., Moore, K. W., Robbins, P. D., and therapy for melanoma. J. Clin. Oncol.. 8: 856-869, 1990. Lotze, M. T. Viral interleukin 10 (IL-b), the human herpes virus 4 cellular IL-b homologue, induces local anergy to allogeneic and syn- 25. Itoh, K., Platsoucas, C. D., and Balch, C. M. Autologous tumor- geneic tumors. J. Exp. Med., 182: 477-486, 1995. specific cytotoxic T lymphocytes in the infiltrates of human metastatic melanoma. Activation by interleukin 2 and autologous tumor cells, and 8. Mizoguchi, H., O’Shea, J. J., Longo, D. L., Loeffler, C. M., McVicar, involvement oftheTcell receptor. J. Exp. Med., /68: 1419-1441, 1988. D. W., and Ochoa, A. C. Alterations in signal transduction molecules in T lymphocytes from tumor-bearing mice. Science (Washington DC). 26. Itoh, K., Tilden, A. B., and Balch, C. M. Interleukin-2 activation of 258: 1795-1798, 1992. cytotoxic T lymphocytes infiltrating into human metastatic melanomas. 9. Hahne, M., Rimoldi, D., Schroter, M., Romero, P., Schreier, M., Cancer Res., 46: 301 1-3017, 1986. French, L. E., Schneider, P., Bornand, T., Fontana. A., Lienard, D.. 27. Mulder, W. M., Koenen, H., van de Muysenberg. A. J.. Bloemena, Cerottini, J., and Tschopp, J. Melanoma cell expression of Fas (Apol/ E.. Wagstaff. J.. and Scheper. R. J. Reduced expression of distinct T-cell CD95) ligand: implications for tumor immune escape. Science (Wash- CD molecules by collagenase/DNase treatment. Cancer Immunol. Im- ington DC), 274: 1363-1366. 1996. munother., 38: 253-258, 1994.

28. Abuzakouk, M., Feighery. C.. and O’Farrelly. C. Collagenase and 34. Sirott, M. N., Bajorin, D. F., Wong, G. Y., Tao, Y., Chapman, P. B., dispase enzymes disrupt lymphocyte surface molecules. J. Immunol. Templeton. M. A., and Houghton, A. N. Prognostic factors in patients Methods, 194: 21 1-216. 1996. with metastatic malignant melanoma: a multivariate analysis. Cancer 29. Berd, D., Herlyn, M., Koprowski, H., and Mastrangelo. M. J. Flow (Phila.), 72: 3091-3098, 1993. cytometric determination of the frequency and heterogeneity of expres- 35. Anderson, J. R., Crowley, J. J., and Propert, K. J. Interpretation of sion of human melanoma-associated antigens. Cancer Res., 49: 6840- survival data in clinical trials. Oncology, 5: 104-1 14, 1991. 6844. 1989. 36. Zhu, H., Melder, R. J., Baxter, L. T., and Jam, R. K. Physiologically 30. Bystryn. J. C., Jacobson, S., Harris, M. N., Roses, D. F., Speyer. J. L.. based kinetic model of effector cell biodistribution in mammals: impli- and Levin, M. Preparation and characterization of a human polyvalent cations for adoptive immunotherapy. Cancer Res.. 56: 3771-3781, melanoma antigen vaccine. J. Biol. Response Modif.. 5: 21 1-224, 1986. 1996. 3 1 . Mather, J. P.. and Sato, G. H. The growth of mouse melanoma cells in hormone supplemented serum-free medium. Exp. Cell Res., /20: 37. Economou, J. S., Belldegrun, A. S., Glaspy. J.. Toloza, E. M., 199-200. 1979. Figlin. R.. Hobbs, J.. Meldon. N., Kaboo, R., Tso, C. L.. Miller, A., La. R.. McBride, W., and Moen, R. C. In vito trafficking of adoptively 32. Simon. R. Optimal two-stage designs for Phase II clinical trials. Control. Clin. Trials, 10: 1-10, 1989. transferred interleukin-2 expanded tumor-infiltrating lymphocytes and peripheral blood lymphocytes. Results of a double gene marking trial. 33. Eton, 0., Legha. S. S.. Moon. T. E.. Papadopoulos. N. E.. Plager, J. Clin. Invest., 97: 515-521, 1996. C., Burgess. A. M., Bedikian, A. Y.. Buzaid, A. C., Ring. S.. Dong. Q.. Glassman. A. B., Balch, C. M.. and Benjamin, R. S. Prognostic factors 38. Suto. R.. and Srivastava, P. K. A mechanism for the specific for survival of patients treated systemically for disseminated melanoma. immunogenicity of heat shock protein-chaperoned peptides. Science J. Clin. Oncol., in press. 1998. (Washington DC), 269: 1585-1588, 1995.

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 1998 American Association for Cancer Research. Active immunotherapy with ultraviolet B-irradiated autologous whole melanoma cells plus DETOX in patients with metastatic melanoma.

O Eton, D D Kharkevitch, M A Gianan, et al.

Clin Cancer Res 1998;4:619-627.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/4/3/619

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/4/3/619. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.