Clonal expansion of T lymphocytes in human melanoma metastases after treatment with a hapten-modified autologous tumor vaccine.

Metastatic melanoma patients treated with an autologous DNP-modified tumor cell vaccine develop inflammatory responses in metastatic tumors characterized by infiltration of CD8+ T cells. To further define this immune response, we analyzed T cell receptor beta-chain variable (TCRBV) region repertoire in biopsy specimens and peripheral blood lymphocytes of six patients. After administration of DNP vaccine, a restricted set of TCRBV gene families was found to be expanded compared with prevaccine metastases. In several postvaccine lesions of one patient, obtained over a 2-yr period, TCRBV14+ T cells were clonally expanded and identical T cell clonotypes could be detected. Two major recurring clones were biased toward the use of TCRBJ1S5. Furthermore, T cell lines derived from two such infiltrated skin lesions and, enriched in TCRBV14+ T cells, displayed HLA-class I-restricted lysis of the autologous melanoma cells. Clonal expansion of T cells was demonstrated in the T cell-infiltrated, postvaccine metastasis of a second patient as well. These results indicate that vaccination with autologous, DNP-modified melanoma cells can expand selected clones of T cells at the tumor site and that such clones are potentially destructive to the tumor.

Evaluation of a "nude" mouse-human tumor panel as a predictive secondary screen for cancer chemotherapeutic agents.

Nine established human melanoma tissue-cultured cell lines heterotransplanted in C57BL/6 "nude" mice were exposed to each of 4 chemotherapeutic agents of known clinical activity against human melanoma. Two of the therapeutic agents, 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) and 5-(3,3-dimethyl-1-triazino) imidazole-4-carboxamide (DTIC), are known to be active against human melanoma; the other two, adriamycin and 5-azacytidine, are known to be inactive. Sterile saline served as a control agent. In 2 cell line heterotransplants, the control tumor spontaneously regressed. Of the 7 cell lines that remained for evaluation, 4 were sensitive to DTIC, 1 was sensitive to BCNU, and none was sensitive to adriamycin or 5-azacytidine. These data indicate that the nude mouse-human tumor model may be a predictive secondary screen for cancer chemotherapeutic agents.

Depletion of suppressor-cytotoxic T-lymphocytes by administration of a murine monoclonal antibody.

We administered anti-Leu 2a, a murine monoclonal antibody to the human suppressor-cytotoxic T-cell subset, to 20 patients with advanced cancer to determine the toxicity and its ability to deplete circulating Leu 2(+) lymphocytes. Four doses were tested, 1, 5, 25, and 100 mg, and the infusion rate varied from 2 to 24 h. Administration of anti-Leu 2a produced rapid (within 4 h) depletion of circulating Leu 2(+) lymphocytes, the magnitude and duration of which were dose dependent: the 1-mg dose caused a less than 50% fall in circulating Leu 2(+) lymphocytes, while the higher doses caused 75-98% depletion of those cells in 10 of 17 patients so treated. The duration of depletion of Leu 2(+) cells was 1-2 days after 5 mg of anti-Leu 2a, 3-4 days after 25 mg, and 4-30+ days after 100 mg. In seven patients (3-5, 3-25, and 1-100 mg), there was persistence of variable numbers of circulating mouse Ig-coated Leu 2(+) cells, possibly indicating impaired capacity to clear these cells. Modulation of the Leu 2a antigen after antibody treatment was not observed. Peak serum levels of anti-Leu 2a were (medians): 5-mg dose = 350 ng/ml, 25-mg dose = 5500 ng/ml, and 100-mg dose = 48,000 ng/ml; murine antibody was detectable in the serum for 7-14 days after the 100-mg dose. All patients had increased titers of antimouse antibody following treatment with peak titers on day 14. The most common toxicity was shaking chills. This occurred within 1.5 h of beginning the infusion, lasted for no more than 30 min, and did not require cessation of the treatment. Anti-Leu 2a administration caused a clinically unimportant, but statistically significant fall in hematocrit (mean = -7%) and platelet (mean = -23%) count and a transient (less than 1 day duration) fall in the total lymphocyte count. Thus, infusion of murine monoclonal anti-Leu 2a can cause substantial depletion of the suppressor-cytotoxic T-cell subset with modest toxicity.

Effect of low dose cyclophosphamide on the immune system of cancer patients: reduction of T-suppressor function without depletion of the CD8+ subset.

Low dose cyclophosphamide (CY) can augment the development of delayed-type hypersensitivity to primary antigens in patients with advanced cancer. In this paper, we have considered the hypothesis that the immunopotentiation is related to reduction of T-suppressor activity. Peripheral blood lymphocytes were collected and cryopreserved from 45 patients with metastatic malignancy before and then 3, 7, and 19 days after administration of CY, 300 mg/m2 i.v. The peripheral blood lymphocytes were tested for generation of concanavalin A-inducible suppressor activity, proliferative response to phytohemagglutinin, and phenotype using monoclonal antibodies to CD4 and CD8. Concanavalin A-inducible suppression was significantly reduced by day 3 and declined progressively through day 19. The mean percentage changes in suppression were: day 3, -23.4 +/- 6.8 (SE) (P less than 0.01); day 7, -33.1 +/- 14.3 (P = 0.052); day 19, -43.1 +/- 10.7 (P less than 0.01). In contrast, CY caused no significant changes in phytohemagglutinin proliferation (mean percentage changes: day 3, -4.7 +/- 6.1; day 7, -15.6 +/- 7.5; day 19, -5.5 +/- 8.1), indicating that the reduction in concanavalin A-inducible suppression was not merely a reflection of a general reduction in peripheral blood lymphocyte function. The total number of circulating lymphocytes was not affected by low dose CY. Moreover, flow cytometric analysis showed no significant changes in the percentage of circulating CD8+ or CD4+ T-cells or in the CD4/CD8 ratio at any time point after CY. Thus, administration of low dose CY to these patients caused impairment of nonspecific T-suppressor function without selective depletion of the CD8+ subset that is generally associated with that function. Several immunoregulatory models that are consistent with these observations are discussed.

Effect of low dose cyclophosphamide on the immune system of cancer patients: depletion of CD4+, 2H4+ suppressor-inducer T-cells.

We studied peripheral blood lymphocytes (PBL) from 42 patients with metastatic melanoma undergoing treatment with cyclophosphamide (CY) plus melanoma vaccine to determine whether CY immunopotentiation could be related to depletion of T-cells that function as inducers of suppression. Every 28 days, the patients were given CY, 300 mg/m2 i.v., followed 3 days later by the intradermal injection of autologous, irradiated melanoma cells mixed with Bacillus Calmette-Guérin. PBL were separated by density gradient centrifugation and cryopreserved until needed for testing. They were stained with monoclonal antibodies directly conjugated to fluorescein isothiocyanate or phycoerythrin and analyzed by two-color flow cytometry. At no time after the initiation of CY plus vaccine were there any significant changes in the percentages of helper-inducer T-cells (CD4+), suppressor-cytotoxic T-cells (CD8+), or the subpopulation of CD8+ cells expressing Leu 15, a marker for suppressor cells. Treatment of melanoma patients with CY plus vaccine resulted in a progressive fall in the proportion of CD4+ T-cells expressing the 2H4 (CD45) antigen, which identifies inducers of suppression. The reduction of CD4+, 2H4+ T-cells did not become apparent until day 28 after the first dose of CY and reached statistical significance only on days 49 (21 days after the second dose) and 105 (21 days after the fourth dose) (mean changes +/- SE: day 49, -5.4 +/- 1.4%, P less than 0.01; day 105, -9.1 +/- 2.2%, P less than 0.01; t test for nonindependent samples). In contrast, the proportion of CD4+ T-cells expressing the antigen 4B4 (CDw29), which are true helper cells, increased slightly, although not significantly, following the institution of CY plus vaccine (mean changes: day 49, +2.9 +/- 2.1%; day 105, +3.6 +/- 2.4%). Similar results were obtained when absolute numbers of circulating cells, rather than percentages, were analyzed. Thus the number of CD4+, 2H4+ T-cells fell from a mean of 395,000/ml on day 0 to 309,000/ml on day 49 (P less than 0.01) to 256,000/ml on day 105 (P less than 0.05). The absolute number of CD4+, 4B4+ cells remained unchanged at the same time points. These changes were not due to progression of metastatic disease, since a comparison of patients with progressive metastases with those who were rendered disease free by surgery showed no significant differences in the reduction of the percentage of CD4+, 2H4+ T-cells.(ABSTRACT TRUNCATED AT 400 WORDS)

Immunological enhancement of leukemia L1210 by Corynebacterium parvum in allogeneic mice.

The effect of Corynebacterium parvum on the immune response of C57BL/6 mice (H-2b) to the allogeneic leukemia L1210 (H-2d) was investigated. Mice were either left untreated or given C. parvum i.v. or i.p. in various dosages. Seven days later they were challenged with 2.5 to 10 X 10(6) live L1210 cells i.p. Control animals almost always rejected the challenge. In contrast, most mice pretreated with either 1.0, 0.5, or 0.25 mg of C. parvum i.v. and 1.0 or 0.5 mg i.p. exhibited enhanced growth of leukemia L1210 as indicated by gross ascites and significantly greater weight gain. This sometimes progressed to the death of the animal, but more often regressed after several days. Spleen cell-mediated cytotoxicity to alloantigens, evaluated in vitro by release of 51Cr from P815Y (H-2d) target cells, was significantly decreased in the mice pretreated with either 1.0 or 0.5 mg of C. parvum i.v. or 0.5 mg of C. parvum i.p. This suppression could not be reversed by reduction of the concentration of macrophages in the spleen cell suspensions. Complement-dependent cytotoxic antibody, measured by release of 51Cr from L1210 cells, was profoundly suppressed in mice pretreated with C. parvum i.v. in dosages ranging from 1.0 to 0.1 mg. These data suggest an immunological basis for the enhanced growth of leukemia L1210 caused by C. parvum at these schedules.

Induction of cell-mediated immunity to autologous melanoma cells and regression of metastases after treatment with a melanoma cell vaccine preceded by cyclophosphamide.

There is considerable evidence in animal tumor systems that antitumor immunity is modulated by suppressor T-lymphocytes, and that the cytotoxic drug cyclophosphamide (CY) can abrogate that suppression. We measured the acquisition of delayed-type hypersensitivity (DTH) to autologous melanoma cells in 19 patients with metastatic malignant melanoma. The patients were treated with an autologous melanoma cell vaccine, either given alone, or given 3 days after the administration of CY, 300 mg/m2 i.v. The DTH responses of CY-pretreated patients were significantly greater than those of control (vaccine only) patients. Thus, after two vaccine treatments, the median DTH responses (mm induration) were as follows: controls, 4 mm; CY pretreated, 11 mm; P = 0.034, Mann-Whitney U test, 2-tailed. Whereas seven of eight CY-pretreated patients developed DTH to autologous melanoma cells of at least 5 mm, only two of seven controls did so (P = 0.034, Fisher's exact test). Two patients had significant antitumor responses to treatment with CY plus vaccine, consisting of complete disappearance of skin metastases and a pulmonary nodule in one, and regression of s.c. and liver metastases in the other. Both patients remain free of melanoma after 42 and 33 mo, respectively.

Flow cytometric determination of the frequency and heterogeneity of expression of human melanoma-associated antigens.

We used flow cytometry to measure the expression of human melanoma antigens on cell suspensions dissociated from metastatic masses. The objective was to study the heterogeneity between tumor samples from different patients and between different tumors excised from a single patient. Fifty-three metastases excised from 34 melanoma patients were analyzed with a panel of nine murine monoclonal antibodies (MOABs). Melanoma cells were stained by an indirect fluorescent method and analyzed on a Coulter EPICS C flow cytometer after gating to exclude tumor-infiltrating leukocytes and dead cells. The most consistently and most strongly expressed antigen was the high-molecular-weight proteoglycan (detected by the MOAB 9.2.27), which was expressed on 95% of the melanoma specimens and by a high proportion of cells within each specimen (mean +/- SE, 79.2 +/- 5.5). However, strong expression of this antigen was limited to melanoma cells that had been dissociated mechanically and was markedly diminished by exposure to collagenase. Culture of collagenase-dissociated tumor cells for 24 to 48 h resulted in reexpression of the antigen. The expression of other melanoma-associated antigens was not affected by collagenase treatment, but for these antigens there was more variability between cells from an individual tumor and between tumors from different patients. The percentage of enzyme-dissociated tumors considered positive for MOAB binding (defined as at least 10% of cells positive) and the mean +/- SE of the percentage of positive cells within a tumor were as follows: MOAB ME-9-61 (antigen, p97) = 84% + (41.2 +/- 5.4%); MOAB ME-20.4 (antigen, nerve growth factor receptor) = 40% + (18.7 +/- 5.1%); MOAB ME-24 (antigen, ganglioside GD3) = 84% + (50.8 +/- 4.8%); MOAB ME-311 (antigen, ganglioside 9-O-acetyl-GD3) = 76% + (42.5 +/- 5.1%); MOAB ME-361 (antigen, mainly ganglioside GD2) = 3% + (1.9 +/- 0.8%); MOAB 3F8 (antigen, ganglioside GD2) = 36% (10.5 +/- 3.8%); MOAB 14G2a (antigen, ganglioside GD2) = 86% + (46.0 +/- 6.7%); MOAB L243 (antigen, HLA-DR) = 56% + (22.5 +/- 5.5%). In 19 cases, we were able to compare the antigenic profiles of two tumors excised from the same patient at different times. Analysis by nonindependent t test showed no significant differences in MOAB binding between the paired tumors. Moreover, linear regression analysis indicated that there was a linear relationship, with a slope approximately = 1, between the percentage of positive cells in Tumor 1 versus Tumor 2.(ABSTRACT TRUNCATED AT 400 WORDS)

Depletion of T-cells with the CD4+CD45R+ phenotype in lymphocytes that infiltrate subcutaneous metastases of human melanoma.

We studied the phenotypes of lymphocytes extracted from 22 specimens of human melanoma, 11 s.c. metastases and 11 lymph node metastases, by two-color flow cytometry. Lymphocytes extracted from s.c. metastases were characterized by a significantly reduced ratio of CD4+ to CD8+ T-cells, as compared with peripheral blood lymphocytes from the same patients. Ten of 11 tumor-infiltrating lymphocytes from s.c. metastases, but only 1 of 11 peripheral blood lymphocytes, had a CD4/CD8 ratio of less than 1.0. This alteration was not observed for lymphocytes obtained from nodal metastases. Furthermore, almost all of the CD4+ T-cells in s.c. metastases expressed the antigen CD29w and were negative for the complementary antigen CD45R. In contrast, the CD29w/CD45R ratio of tumor-infiltrating lymphocytes from lymph node metastases was similar to that of matched peripheral blood lymphocytes. Thus tumor-infiltrating lymphocytes from s.c. metastases have the phenotype associated with true helper or antigen-committed T-cells, which could reflect their sensitization to tumor antigens, while tumor-infiltrating lymphocytes from lymph node metastases may represent merely an expanded residua of lymph node lymphocytes. Since tumor-infiltrating lymphocytes expanded in vitro are being tested as therapy for patients with advanced cancer, these observations may have important therapeutic implications.

Impairment of concanavalin A-inducible suppressor activity following administration of cyclophosphamide to patients with advanced cancer.

We have shown that cyclophosphamide (CY) can augment the development of delayed-type hypersensitivity to a primary antigen in patients with advanced cancer. In the present study, we administered CY (1000 mg/sq m) to 19 patients with advanced, metastatic cancer and monitored the compositional and functional changes in their peripheral blood mononuclear cells. Within 2 days of administration of CY, the lymphocyte count fell significantly (mean decrease = 26.0%) and remained significantly depressed through Day 14 with recovery beginning by Day 21. T- and B-lymphocytes were depleted to about the same degree at each time point. Moreover, there was no selective depletion of the Leu 2(+) (suppressor-cytotoxic) or Leu 3(+) (helper-inducer) subsets of T-lymphocytes. Proliferative responses to mitogens (phytohemagglutinin, concanavalin A, pokeweed mitogen) and to allogeneic cells fell significantly within 1 day of administration of CY and continued to be diminished on Day 2. However, these responses recovered to pretreatment levels by Day 3, and, in some cases, exceeded pretreatment levels on Day 7. Concanavalin A-inducible suppressor activity was also diminished on Day 1 (mean decrease, 23.4%) and Day 2 (mean decrease, 39.2%). However, in contrast to the proliferative responses, suppressor activity continued to be significantly impaired on Day 3 (mean decrease, 31.6%) and only partially recovered by Day 7 (mean decrease, 22.1%). Both concanavalin A-inducible suppression and proliferative responses declined again on Days 14 and 21. Thus, between 3 and 7 days after administration of CY, there appeared to be impairment of nonspecific T-cell-mediated suppressor activity of peripheral blood lymphocytes that was not merely a reflection of impaired lymphocyte function in general. This could account for the augmented delayed-type hypersensitivity responses of CY-treated patients.

Potentiation of human cell-mediated and humoral immunity by low-dose cyclophosphamide.

Although cyclophosphamide (CY) is a potent immunosuppressive drug, under the proper conditions, it can potentiate immune responses as well. In past work, we have shown that administration of a commonly used oncostatic dose of CY (1000 mg/sq m) to patients with advanced cancer 3 days before sensitization with the primary antigen, keyhole limpet hemocyanin (KLH), resulted in augmentation of delayed-type hypersensitivity (DTH) but not antibody response to that antigen. The present study was performed to test the immunopotentiation of a lower dose of CY (300 mg/sq m); animal studies and studies of human lymphocytes in vitro suggested that the lower dose might be more effective. Eighteen patients with advanced metastatic cancer were alternately assigned to one of two groups. Sixteen days before CY, one group received KLH and the other group received 1-chloro-2,4-dinitrobenzene (DNCB). CY 300 mg/sq m was given as an i.v. bolus on Day 0. Three days after CY, the patients received KLH or DNCB, whichever they had not received initially. Blood was drawn for antibody titer, and skin testing was performed 14 days after administration of KLH or DNCB. In addition, skin tests to microbial recall antigens were made 2 days before and 17 days after CY. Pretreatment with low-dose CY resulted in significant augmentation of DTH to KLH; thus, the median DTH responses were: KLH alone, 10 mm; and KLH after CY, 27 mm (p less than 0.01). CY pretreatment also resulted in augmentation of the antibody response to KLH. The median total antibody titers (log2 of reciprocal of dilution) were as follows: KLH alone, less than 1; and KLH after CY, 3 (p less than 0.01). All nine CY-pretreated subjects but only 4 of 9 controls developed measurable anti-KLH antibody titers. CY pretreatment neither augmented nor suppressed the 48-hr challenge reaction to DNCB. Moreover, CY had no effect on DTH responses to the recall antigens, dermatophytin, Candida, and mumps.

Immunization with haptenized, autologous tumor cells induces inflammation of human melanoma metastases.

Twenty-four patients with metastatic melanoma were treated with a novel form of active immunotherapy, autologous tumor cell vaccine conjugated to the hapten, dinitrophenyl. This approach is based on the idea, well established in animal systems, that presentation of tumor antigens in the context of a strongly immunogenic hapten augments the development of immunity to those antigens. After being sensitized to dinitrophenyl, patients were given injections of dinitrophenyl-vaccine every 28 days following pretreatment with low dose cyclophosphamide. The vaccine induced a striking inflammatory response in superficial metastases in 14 of 24 patients, consisting of erythema, swelling, warmth, and tenderness over tumor masses. Immunohistochemistry and flow cytometric analysis of biopsy specimens showed marked infiltration with lymphocytes, the majority of which were CD8+, HLA-DR+ T-cells. These observations suggest that a T-cell-mediated immune response against melanoma-associated antigens was facilitated by the "helper" effect of the anti-hapten response.

Abnormal expression of perlecan proteoglycan in metastatic melanomas.

Abnormal expression of proteoglycans has been implicated in cancer and metastasis primarily because these macromolecules are involved in the control of cell growth and matrix assembly. In this report, we have investigated the expression and immunolocalization of perlecan, a major heparan sulfate proteoglycan of basement membranes and pericellular matrices, in human metastatic melanomas. Twenty-six of the 27 tumor samples showed a significant increase (up to 15-fold) in the perlecan mRNA levels when compared with normal tissue. This change correlated with a vast deposition of perlecan protein core in the pericellular matrix of metastatic melanomas. Furthermore, we have established a relationship between perlecan expression in clonal melanoma cells (70W) stimulated with neurotrophins and their increased invasiveness. Interestingly, perlecan mRNA levels were up-regulated within 10 min of neurotrophin stimulation, indicating that perlecan is an early response gene. This upregulation also occurred prior to heparanase production, suggesting that perlecan expression and its regulation might play a pivotal role in the initial onset of invasion.

Growth and metastasis of fresh human melanoma tissue in mice with severe combined immunodeficiency.

Cryopreserved cell suspensions of freshly excised melanoma metastases from nine patients were injected s.c. into C.B-17 severe combined immunodeficiency (SCID) mice. All 9 tumors grew as s.c. masses and six of nine were successfully transplanted into other SCID mice. Transplant inocula as low as 5 x 10(5) cells resulted in 100% tumor incidence. Moreover, seven of nine tumors metastasized, five from the original s.c. implants and two from transplanted s.c. tumors. Metastases were detected mainly in the lungs but also were found in abdominal viscera (liver, spleen, and pancreas) and thoracic lymph nodes. Flow cytometric analysis showed that expression of a panel of melanoma antigens, melanoma-associated proteoglycan, ganglioside GD3, and ganglioside GD2, was maintained with SCID passage. The original tumor inocula contained a variable percentage of tumor-associated lymphocytes (1-76%). Flow cytometry analysis indicated that these were mainly CD3+ T-cells. However, there was no correlation between the percentage of tumor-associated lymphocytes and the time required for development of a palpable tumor after s.c. injection or the ability to metastasize. These results demonstrate the growth and spontaneous metastasis of fresh human melanoma in SCID mice and suggest that this model could be important for therapeutic and basic biological studies.

Rarity of somatic and germline mutations of the cyclin-dependent kinase 4 inhibitor gene, CDK4I, in melanoma.

Evidence from cytogenetics, multipoint linkage analyses of familial melanoma, and loss of heterozygosity studies of familial and sporadic melanomas support localization of a melanoma susceptibility or tumor suppressor gene at chromosomal region 9p21-23. Recently, the inhibitor of cyclin-dependent kinase 4 (CDK4I; also known as p16INK4, multiple tumor suppressor 1, or CDKN2 gene) has been mapped to 9p21 and shown to be mutated or deleted in a large fraction of cell lines derived from many tumor types, including melanoma, suggesting that this gene could be a melanoma suppressor gene. In order to test for somatic mutations in the CDK4I gene in tumors, DNAs from 30 surgically resected melanomas of both cutaneous and uveal origins were sequenced. No mutations were detected in the coding region of the CDK4I gene, while mutations or deletions were detected in 60% (9 of 15) of the cultured melanoma cell line DNAs. Among presumptive familial cases, nine of which were members of families with one or two other documented melanoma cases, no germline mutations were detected by sequence analysis. A deletion in the second exon of the CDK4I gene was found in one germline allele of a familial melanoma patient from a family with eight affected first degree relatives. These results not only support the suggestion that the CDK4I gene is a familial malignant melanoma gene, they also suggest the presence of another suppressor gene locus within 9p21 which is the target of loss of heterozygosity in sporadic melanomas.

Treatment of metastatic melanoma with an autologous tumor-cell vaccine: clinical and immunologic results in 64 patients.

We treated 64 patients with metastatic melanoma using a melanoma vaccine preceded by low-dose cyclophosphamide (CY), and monitored immunologic effects and antitumor activity. On day 0, the patients were given CY 300 mg/m2 intravenously. Three days later, they were injected intradermally with vaccine consisting of 10 to 25 x 10(6) autologous, enzymatically dissociated, cryopreserved, irradiated (25 Gy) tumor cells mixed with bacillus Calmette-Guérin (BCG). This treatment sequence was repeated every 28 days. Of 40 assessable patients with measurable metastases, five had responses, four complete and one partial, with a median duration of 10 months (7 to 84+ months). In six additional patients, we observed an antitumor response that seems to be peculiar to this vaccine therapy: the regression of metastatic lesions that appeared after the immunotherapy was begun. Delayed-type hypersensitivity (DTH) to autologous, mechanically dissociated melanoma cells that had not been exposed to extraneous antigens, such as enzymes or fetal calf serum, increased significantly following immunotherapy (day 0 v day 49, P less than .001; day 0 v day 161, P less than .001; day 0 v day 217, P = .021). Antitumor responses to the vaccine were strongly associated with DTH, as indicated by three observations: (1) eight of 10 patients who exhibited tumor regression had positive DTH, (2) in postsurgical adjuvant patients, there was a highly significant linear relationship (P less than .001) between the intensity of DTH to autologous melanoma cells and the time to recurrence of tumor, and (3) nine patients who developed DTH to the autologous melanoma cells in their original vaccine developed new metastases that failed to elicit DTH or elicited a much smaller response. In three cases, we were able to excise regressing tumors for histologic examination; such tumors were characterized by an intense infiltration of lymphocytes. This demonstration that an immune response to melanoma-associated antigens can be elicited in cancer-bearing patients provides some basis for optimism about the prospects for developing active immunotherapy that has practical therapeutic value.

Autologous hapten-modified melanoma vaccine as postsurgical adjuvant treatment after resection of nodal metastases.

PURPOSE: To determine whether treatment with an autologous whole-cell vaccine modified with the hapten dinitrophenyl (DNP vaccine) is an effective postsurgical adjuvant treatment for melanoma patients with clinically evident nodal metastases. PATIENTS AND METHODS: Eligible patients had regional nodal metastases that were large enough (> or = 3 cm diameter) to prepare vaccine. Following standard lymphadenectomy, patients were treated with DNP vaccine on a monthly or weekly schedule. RESULTS: Of 62 patients with metastasis in a single lymph node bed (stage III), 36 are alive after a median follow-up time of 55 months (range, 29 to 76); the projected 5-year relapse-free and overall survival rates are 45% and 58%, respectively. Of 15 patients with metastases in two nodal sites, five are alive with a median follow-up time of 73 months. An unexpected finding was the significantly better survival of older patients; the projected 5-year survival of patients greater than 50 versus < or = 50 years was 71% and 47%, respectively (P = .011, log-rank test). The development of a positive delayed-type hypersensitivity (DTH) response to unmodified autologous melanoma cells was associated with significantly longer 5-year survival (71% v 49%; P = .031). Finally, the median survival time from date of first recurrence was significantly longer for patients whose subcutaneous recurrence exhibited an inflammatory response (> 19.4 v 5.9 months; P < .001). CONCLUSION: Postsurgical adjuvant therapy with autologous DNP-modified vaccine appears to produce survival rates that are markedly higher than have been reported with surgery alone. Moreover, this approach has some intriguing immunobiologic features that might provide insights into the human tumor-host relationship.

Interleukin 10 production by human melanoma.

Interleukin 10 (IL-10) has the physiological role of down-regulating cell-mediated immunity. We have recently reported that mRNA for IL-10 was present in most metastatic melanoma tissues. The purpose of this investigation was to determine whether melanoma metastases produce IL-10 protein. Single-cell suspensions were prepared by enzymatic dissociation of 28 lymph node metastases and 7 s.c. metastases and cryopreserved. Of these 35 samples, 30 produced IL-10 after a 24-h incubation (median, 125.1 pg/ml). IL-10 production was slightly diminished after 25 Gy irradiation but almost completely abrogated after modification with the hapten dinitrophenyl. After 7 or 14 days in tissue culture, melanoma cells continued to produce IL-10 but only at about 10% of the levels of freshly dissociated tissues. Moreover, of eight melanoma cell lines established from these cultures, only one produced IL-10 protein. To determine whether IL-10 was produced by melanoma cells or tumor-associated leukocytes, single-cell suspensions were fractionated with anti-CD45 antibody-conjugated magnetic beads. In four of five samples, IL-10 production was increased by depletion of leukocytes, suggesting that the primary source was the melanoma cells themselves. This was confirmed by immunohistochemical staining of cytospin preparations and frozen tissue sections. Finally, 10 of 55 patients with clinically evident metastases showed elevations of circulating IL-10; three patients who had been melanoma-free developed high serum IL-10 levels, concurrent with the appearance of distant metastases. These data indicate that production of IL-10 is characteristic of metastatic melanomas and raise the possibility that this cytokine allows tumors to avoid or to modulate immunological attack.

"Autoregulation" of human neutrophil activation in vitro: regulation of phorbol myristate acetate-induced neutrophil activation by cell density.

Phorbol myristate acetate (PMA, 10(-7) M) activation of adherent neutrophils (PMNs) led to a markedly attenuated release of superoxide anion (O2-) per cell when PMNs were activated at high density (2.85 fmol O2-/PMN at 2 million in 0.1 ml) in comparison with cells activated at low cell density (12.0 fmol O2-/PMN at 250,000 in 0.1 ml). This "autoregulatory" phenomenon was not due to a defect in the superoxide anion assay employed, to a differential adherence of neutrophils at high vs. low density, or to substrate (cytochrome c) or cell stimulus (PMA) limitation. It was associated with an inhibition of apparent NADPH oxidase activity and a leftward shift (toward a lower level of activation) in the activation profile of PMNs (as determined by FACS analysis using PMNs preloaded with 2'7'-dichlorofluorescin diacetate in which H2O2 production results in the production of the fluorescent product 2'7'-dichlorofluorescein intracellularly). Other aspects of the neutrophil activation response including arachidonic acid mobilization, phospholipid metabolism, and perhaps phosphatidylinositol turnover were also attenuated when PMNs were activated at high cell density. Studies with cells in solution, cells treated with cycloheximide, and cells treated with 4,4'-diisothiocyanostilbene-2,2'-disulfonic acid suggest that PMN contact with a surface, neutrophil protein synthesis, and an increased surface expression of the heterodimer CD11b/CD18 on PMNs all were not required for autoregulation. Finally, morphometric and morphologic examination of PMNs activated at low vs. high density revealed histologic and structural correlates associated with the attenuated PMN activation response of cells triggered at high cell density. We conclude that multiple structural and functional aspects of the PMN activation response are modulated by cell density and suggest that this property is important both in the physiologic control of neutrophil activation and in the design of in vitro assays of the neutrophil activation response.

Autologous melanoma vaccine induces inflammatory responses in melanoma metastases: relevance to immunologic regression and immunotherapy.

Human primary malignant melanoma is often accompanied by a host response of infiltrating lymphocytes suggestive of tumor antigen-induced immunity and correlated in some tumors with prognosis. Whereas metastatic melanoma deposits typically are not inflamed and contain relatively few lymphocytes and dendritic immune cells, immunization with autologous melanoma-cell vaccine may induce a clinical inflammatory response associated with mononuclear-cell infiltration. In this study, we characterize immune responses to dermal and subcutaneous melanoma metastases in dinitrophenyl (DNP)-pre-sensitized patients immunized with DNP-conjugated melanoma cells. Patients so treated develop cutaneous delayed hypersensitivity responses to DNP-conjugated autologous mononuclear cells, and approximately one-half show clinical evidence of inflammation and regression of metastases within 2-4 months. Whereas pre-vaccination biopsies of metastatic melanoma failed to reveal significant infiltration by lymphocytes, biopsies obtained after vaccination and coincident with clinical inflammation were markedly infiltrated preponderantly by T cells with a CD8+ phenotype. Clustering of these cells about individual degenerating melanoma cells in a manner analogous to "satellitosis" was a consistent feature of this reaction. Enhanced expression of intercellular adhesion molecule-1 (ICAM-1) and human leukocyte antigen (HLA)-DR by melanoma cells were invariably associated with zones of T-cell infiltration, whereas diminished or absent expression was observed in relatively unaffected regions of tumors. Numerous HLA-DR+, CD4+, CD1-, Leu-1- dendritic cells were also associated with zones of early T-cell infiltration. These data indicate that clinical inflammation and regression of metastatic melanoma induced by autologous melanoma-cell vaccine involves activated T cells with cytotoxic-suppressor phenotype and dendritic cells putatively capable of local antigen presentation. ICAM-1 upregulation on melanoma cells is a likely mediator of ligand interaction between infiltrating T cells and target cells in this model of antigen-induced host anti-tumor response. Structural alterations identified in this setting (e.g., tumor cell satellitosis) may provide additional insight into identifying features of naturally occurring host immune responses to primary cutaneous melanomas.