Monoclonal antibodies to 125 kd and 95 kd proteins on human melanoma cells: comparison with other monoclonal-defined melanoma antigens

Two monoclonal antibodies produced by hybridomas were identified by an indirect 125I-protein A binding assay that define cell surface antigens expressed on cultured human melanoma cells but not on autologous lymphoblastoid cells. The first antibody, 705F6 (an IgG2b immunoglobulin), bound to 14/14 melanoma lines, 6/9 carcinomas and sarcomas, 7/7 gliomas and neuroblastomas, 2/2 fetal cell lines, 0/8 lymphoblastoid cell lines, and weakly to 2/4 leukemia lines. The second monoclonal antibody, 436G10 (IgG1), reacted with 10/14 melanomas 5/13 carcinomas and sarcomas, 2/7 gliomas and neuroblastomas, and weakly with the fetal cells, but not with the leukemic or lymphoblastoid cell lines. Comparison of 705F6 and 436G10 with 28 other monoclonal antibodies from different laboratories identified several with similar binding patterns to a panel of tumor and nontumor cell lines. Crossblocking of 125I-labeled 436G10 was not observed by R23, I12 or L10 antibodies. However, 705F6 was completely blocked by monoclonal 376.96S, showing that these two antibodies bind to the same antigenic determinant. The 705F6 antibody immunoprecipitated a 95 kd (kilodalton) membrane protein and the 436G10 antibody bound a 125 kd protein from 125I-labeled melanoma cells. The broad distribution of these two proteins on melanomas and other solid tumors suggests that they define common oncodevelopmental antigens expressed on proliferating cells.     

Immunity to melanoma associated antigens detected by both leukocyte adherence inhibition and antibody dependent cellular cytotoxicity assays

Two different assays, leukocyte adherence inhibition (LAI) and antibody dependent cellular cytotoxicity (ADCC) have been used to measure the immune responses of 69 melanoma patients, 116 patients with other tumours and 64 normal controls to a number of melanoma and control antigens. Using the LAI test, melanoma patients were significantly more reactive (43-69% positive) than normal controls (7-32% positive) to membrane extracts from 3 to 6 melanoma cell lines, and 4 to 6 extracts of melanoma biopsy specimens. However, the proportion of patients with other tumours reacting with 3 of these extracts was similar to melanoma patients. Melanoma patients were more reactive to the melanoma extract than to extracts of normal skin, normal muscle and 2 breast cancer cell lines. ADCC tests were used to detect anti-melanoma antibodies in patient sera. Preferential reactivity by melanoma patients was detected towards only 2 of the 6 melanoma cell lines tested--53% of 55 melanoma sera reactive to LiBr compared with 28% of 79 sera from patients with other tumours and 17% of 29 normal sera (P less than 0.05); 42% of 31 melanoma sera reactive to MM127 compared with 18% of 22 sera from other tumour patients (n.s.) and 5% of 20 normal sera (P less than 0.005). Melanoma patients tested against a number of melanoma cell lines by ADCC or antigen extracts by LAI generally reacted with one or more, but not all of them. Thus, incomplete cross-reactivity between different melanomas was observed. There was no correlation between results of the same patient in the 2 tests.     

Expression of heterophile, Paul-Bunnell and Hanganutziu-Deicher antigens on human melanoma cell lines

Expression of heterophile antigens was studied on 6 human melanoma cell lines. Paul-Bunnell and Hanganutziu-Deicher (H-D) antigens, but not Forssman antigen, were demonstrated on these cell lines. H-D antigen was also demonstrated on melanoma cells isolated from primary and metastatic lesions. Evidence was also presented that HLA class I but not class II (DR and DQ) molecules are expressed on these cell lines. H-D antibodies of IgG class were detected in 10 of 23 (42%) melanoma patients suggesting the possibility that H-D antigen might have been expressed in an immunogenic form in the patients.     

Analysis of the NIH workshop monoclonal antibodies to human melanoma antigens

Reagents exchanged at the 2nd workshop on monoclonal antibodies (MoAb) to human melanoma antigens were analyzed using both serological and immunochemical assays. The analysis by laboratories participating in the workshop of our MoAb 225.28S, 345.134S, 376.96S, 465.12S, and 763.24TS reaffirms our own analysis of these reagents in that (1) they all react with the majority of melanoma cell lines tested and (2) the reactivity of MoAb 225.28S and 763.24TS is much more restricted than that of MoAb 345.134S 376.96S, and 465.12S. Our serological analysis revealed that the majority of workshop reagents reacted with cultured melanoma cells. Immunochemical analysis of these monoclonal antibodies allowed for their division into three groups according to the molecular weights of the antigens recognized in immunoprecipitation experiments, greater than 100 kd, 80-100 kd, and DR antigens. Further analysis of the first two groups of monoclonal antibodies by immunodepletion and antibody binding inhibition assays revealed that MoAb 9.2.27, 225.28S, and 763.24TS recognize distinct determinants with a heterogeneous distribution on subpopulations of a high molecular weight melanoma associated antigen. MoAb 376.96S and 705.F6 recognize either the same or spatially close determinant(s).     

Surface antigens of human melanoma cells defined by monoclonal antibodies. I. Biochemical characterization of two antigens found on cell lines and fresh tumors of diverse tissue origin

Monoclonal antibodies were produced against melanoma cell lines derived from patients presently disease free. Based on tissue distribution of binding, the antibodies obtained could be used to classify melanoma surface antigens into groupings similar to those obtained from studies of autologous and xenogeneic antibodies. Three antibodies (21.43, 15.75 and 15.95) reacted with the immunizing tumor but not with peripheral blood lymphocytes or a lymphoblastoid cell line derived from the tumor donor. When tested on a broader cell panel, antibody 21.43 reacted only with melanoma lines, while antibodies 15.75 and 15.95 reacted with carcinoma cell lines as well as with the majority of melanoma cell lines. Antibody 15.95 precipitated a 49000 dalton glycoprotein from both melanoma and carcinoma cells. Antibody 15.75 precipitated a 74 000 dalton glycoprotein from the surface of melanoma and carcinoma cell lines which is also found on freshly isolated tumor cells from melanoma and carcinoma metastases in vivo.     

Reactivity of monoclonal antibodies against human leucocyte antigens with lymphocytes of non-human primate origin

The phylogenetic distribution of antigens present on human lymphocytes was investigated by incubating human or simian cells with murine anti-human monoclonal antibodies and then determining the level of reactivity with a radiolabelled anti-murine IgG reagent. The monoclonal antibodies used were specific for a T-cell antigen, lymphoid and lymphoid:myeloid antigens, Ia antigens, and beta 2 microglobulin. The cells examined included B- and T-lymphoblastoid cell lines and fresh peripheral blood lymphocytes separated by sheep erythrocyte rosetting into T-cell and non T-cell fractions. Results of these studies showed that the antibodies gave complete cross-reactivity with gorilla and chimpanzee cells while B-cell lines of orangutan origin had lost lymphoid and beta 2 microglobulin markers. Gibbon cells and cells of Old World and New World monkeys reacted strongly only with monoclonal antibodies against Ia antigenic determinants. These Ia antigens were found on the non T-cell fraction of fresh peripheral lymphocytes, on B-cell lines and on some virus induced T-cell tumour lines. Immunoprecipitation analysis using the anti-Ia antibodies showed a degree of molecular diversity on owl monkey and marmoset cells compared to the Ia antigens associated with human cells.     

Monoclonal antibodies to human malignant mesothelioma

Murine monoclonal antibodies were used to identify tumor-cell membrane antigens on a new human mesothelioma cell line. Hybridomas were constructed by fusing SP2/0 mouse myeloma cells with spleen cells from Balb/C mice immunized by the human mesothelioma cell line MT-1. Hybridoma antibody was detected in 55/672 microculture wells that reacted to these MT-1 tumor cells by an indirect¹²⁵I-protein A binding assay. Six cultures produced antibody binding selectively to the MT-1 tumor cells but not to a human lymphoblastoid cell line. These six hybridomas were cloned: three were IgG and three were IgM antibodies. One monoclonal, MAb 45, reacted with 4 of 7 human mesothelioma cell lines but with only 1 of 11 carcinomas, 1 of 3 sarcomas, 4 of 11 melanomas, and 0 of 5 lymphoid lines. The other five monoclonals had a much broader cross-reactivity. Using an immunoperoxidase technique, MAb 45 bound to mixed-type malignant mesotheliomas but not to normal lung and pleura. The specificity of MAb 45 for diffuse mesotheliomas and the low cross-reactivity with carcinomas and normal adjacent tissues suggest that this monoclonal may be clinically useful.     

Production and characterization of monoclonal antibody to a melanoma specific glycoprotein

An immunogen consisting of a 4M urea extract derived from human melanoma cells (M14), that was devoid of HLA-A,B,C, HLA-DR antigens and fibronectin was adsorbed to lens culinaris lectin-Sepharose 4B and used to immunize mice for production of monoclonal antibody to a melanoma-specific glycoprotein. Screening for hybridomas secreting antibodies to melanoma associated antigens was facilitated by use of a solid phase target antigen of chemically defined medium of melanoma cells (CDM). Use of these procedures allowed us to select 40 hybridomas secreting antibody which recognized determinants on melanoma cells not found on lymphoid cells. Further characterization of one of these hybridomas, 9.2.27, indicated that the antibody it secreted recognized a 240K dalton glycoprotein found on all melanoma cell lines tested but not on carcinoma, lymphoid, or fibroblastoid cultures. These results demonstrate the utility of soluble antigen preparations devoid of strongly immunogenic non tumor-specific molecules in the elicitation of tumor specific antibody. Preliminary results suggest that immunogens of this kind are superior to intact melanoma cells for production of tumor specific hybridomas.     

Monoclonal antibodies in cell-mediated cytotoxicity against human melanoma and colorectal carcinoma.

Hybridoma-derived monoclonal anti-melanoma antibodies and anti-colorectal carcinoma antibodies were found to mediate in vitro antibody dependent cell-mediated cytotoxicity (ADCC) reactions against melanoma and colorectal carcinoma cells, respectively. The antigen(s) detected in ADCC on melanoma cells maintained for more than one hundred passages in tissue culture were also found on two recently established melanoma cell lines. These antigens were not detected on skin fibroblasts of the same patients from whom the melanomas were obtained. The ADCC reactivities of anti-melanoma and anti-colorectal carcinoma antibodies were found to be specific for melanoma cells and colorectal carcinoma cells, respectively.     

Antigen expression of metastasizing and non-metastasizing human melanoma cells xenografted into nude mice

In order to study differences in antigen expression related to the different stages of the process of metastasis of human melanoma cell lines, we determined the expression pattern of a series of well-characterized genes in a set of human melanoma cell lines with different metastatic behavior in nude mice. This set included non-metastatic (IF6, 530), sporadically metastatic (M14, Mel 57), and frequently metastatic (BLM, MV3) cell lines after subcutaneous inoculation. To study the phenotype of these cell lines both the cultured cells and representative samples of local tumors at the inoculation site and their metastases in the lungs were immunostained with a panel of monoclonal antibodies directed against melanocytic differentiation or progression antigens. Although most cell lines (IF6, 530, M14 and Mel 57) showed HLA-DR expressionin vitro, these antigens were lacking in all xenografted lesions studied with exception of the 530 cell line. 530 Xenografts, however, showed a dramatic down-regulation of HLA-DR compared with the cell linein vitro. The same phenomenon was seen with respect to ICAM-1 expression. The expression of all other antigens studied in xenografts, both in subcutaneous tumors and in lung lesions, was in general comparable to that in the melanoma cell linesin vitro, with exception of the 530 cell line. In all melanoma cell lines except 530 the degree of intra- and interlesional heterogeneity regarding the expression of all antigens studied was limited. Remarkably, comparison of the immunophenotype of the frequently metastasizing (BLM, MV3) and the sporadically (M14, Mel 57) or non-metastasizing (IF6, 530) cell lines showed that the two frequently metastasizing cell lines had marked expression of the progression antigens VLA-2 and epidermal growth factor receptor, and lack of expression of the differentiation antigen NKI-beteb. These findings warrant further studies on the role of these antigens in the process of metastasis of human melanoma cells in nude mice.     

Functional characteristics and differential expression of class II DR,DS, and SB antigens on human melanoma cell lines

Although most cultured melanoma cell lines express DR Class II molecules, many of these do not also express the DS (MB) Class II molecules as detected by a monoclonal antibody specific for DS. Cells lacking either DR or DS molecules or both could only be induced to express DR antigens in rare cases by combined incubations with azacytidine and Interleukin-2 conditioned medium, although the expression of DR molecules on fibroblasts or U937 monocytes could more easily be induced under the same culture conditions. Melanoma cells expressing DR antigens could function in antigen presentation for the histocompatibility antigens themselves and for DR specific presentation of TNP determinants to allogeneic T-cells sensitized to TNP modified lymphocytes and showing restriction in their responses to the specificity of the DR molecules expressed on the original, autologous senzitizing cells. DR positive melanoma cells could not, however, be demonstrated to function in the presentation of any of the soluble antigens tested. All DR positive melanoma cells also expressed SB antigens, but these were not detected on DR negative melanoma cells. These studies collectively indicate that the expression of Class II histocompatibility antigens on diverse cell types is subject to differential regulatory control and is associated with differences in their functional activities.     

A panel immunoblot using co-incubated monoclonal antibodies for identification of melanoma cells

Antigen expression in melanoma is heterogeneous. Immunophenotyping using a panel of monoclonal antibodies may facilitate immunotherapy. An immunoblot procedure was developed to detect antigens in melanoma cells. Numerous monoclonal antibodies were tested to determine if (1) antigens were detected after transfer to membranes, (2) single bands or discrete multiple bands were obtained, (3) co-incubation of multiple monoclonal antibodies had no interference, and (4) banding patterns were non-overlapping. Antigens were selected based upon their association with melanoma and the availability of respective monoclonal antibodies. Antigens were melanoma antigen recognized by T-cells (MART-1), tyrosinase, tyrosinase-related protein 1 (TRP-1), S100, vimentin, glycoprotein 130 (gp130), a carcinoembryonic antigen (CEA)-like marker, KBA-62 and NKI-C3. Actin positive controls could be assessed simultaneously. Test samples were separated by polyacrylamide gel electrophoresis in a 4-15% polyacrylamide gradient, transferred to polyvinylidine fluoride membrane, blotted using a Fast-Blot apparatus (Pierce), and developed using diaminobenzidine/metal. Melanoma cell lines were immunophenotyped using this panel immunoblot, and were compared to a standard control and to non-melanoma cells. Up to four antigens could be detected simultaneously in a single lane of the immunoblot, using a single test sample of greater than 100000 cells.     

Reactivity spectrum of 30 monoclonal antimelanoma antibodies to a panel of 28 melanoma and control cell lines

Thirty monoclonal antibodies from eight laboratories exchanged after the First Workshop on Monoclonal Antibodies to Human Melanoma held in March 1981 at NIH were tested in an antibody-binding radioimmunoassay using a panel of 28 different cell lines. This panel included 12 melanomas, three neuroblastomas, four gliomas, one retinoblastoma, four colon carcinomas, one lung carcinoma, one cervical carcinoma, one endometrial carcinoma, and one breast carcinoma. The reactivity pattern of the 30 monoclonal antibodies tested showed that none of them were directed against antigens strictly restricted to melanoma, but that several of them recognize antigenic structures preferentially expressed on melanoma cells. A large number of antibodies were found to crossreact with gliomas and neuroblastomas. Thus, they seem to recognize neuroectoderm associated differentiation antigens. Four monoclonal antibodies produced in our laboratory were further studied for the immunohistological localization of melanoma associated antigens on fresh tumor material. In a three-layer biotin-avidin-peroxidase system each antibody showed a different staining pattern with the tumor cells, suggesting that they were directed against different antigens.     

Growth and dissemination of human malignant melanoma cells in mice with severe combined immune deficiency.

BACKGROUND: The severe combined immune deficiency (SCID) mouse is lacking mature B and T lymphocytes and may be permissive for human tumor growth and metastasis. EXPERIMENTAL DESIGN: SCID mice received human melanoma cells of diverse origins including: 2 established cell lines, 4 early passage cell lines, and fresh or cryopreserved cells obtained directly from 9 patient biopsies. They were introduced into SCID mice via intraperitoneal, subcutaneous and intravenous injections. RESULTS: Tumor growth occurred with each of the 15 melanoma specimens for a take rate of 100% considering cell source. In addition, 60% of the 102 total mice injected displayed tumor growth in at least one site. The most consistent tumor growth (77%) occurred after intraperitoneal injection. Tumors developed in 41 and 48% of mice injected subcutaneously and intravenously, respectively. The mice developed both local tumor growth with palpable tumor nodules at injection sites and hematogenous and/or lymphatic dissemination to multiple sites in the abdominal and thoracic cavities. The number of metastases per animal averaged 16.3 and the number per organ ranged from 1 to 38. Melanotic and amelanotic tumor nodules obtained from a single patient retained their original characteristics with regard to melanin production after passage in the SCID mouse. The appearance of the human melanoma cells in SCID mouse tissues ranged from implants on the organ capsule to frank parenchymal organ involvement and vascular invasion. Some small foci of tumor were only detected using immunohistochemistry with monoclonal antibodies against the S-100 and HMB-45 to melanoma-related antigens. CONCLUSIONS: We conclude that the SCID mouse consistently supports growth, invasion, and metastatic spread of human melanoma cells, including specimens obtained from fresh patient biopsies. The SCID mouse will serve as a relevant in vivo model for studying the biology of human malignant melanoma and screening new therapeutic agents.     

DR antigens on melanoma cells: analysis with monoclonal antibodies

Two monoclonal antibodies (691-13-17 and 37-7) can precipitate DR antigens from radioiodinated, detergent solubilized melanoma cells. However, after complete depletion of lysates with one antibody (691-13-17) alpha-like chains can be still be precipitated from some melanoma cells by the other antibody (37-7). Antibody 37-7 also precipitates an additional distinct antigen from SK-MEL 37 but not from any five other melanoma cell lines.     

Tissue specificity of anti melanoma monoclonal antibodies analyzed on cell lines

Twenty-eight monoclonal antibodies (MoAbs) from the NIH melanoma exchange program were analyzed in a binding assay for reactivity with a panel of 22 cell lines which included melanomas, carcinomas, fibroblasts, and lymphomyeloid cells. In this test, most MoAbs showed reactivity with a wide range of cell lines. The MoAbs were also tested for binding with freshly removed tumor cells and normal peripheral lymphocytes.     

Attempts to optimize active specific immunotherapy for melanoma.

During the past 5 years, we have been conducting clinical trials with a therapeutic melanoma vaccine (melanoma "theraccine"). Mechanical lysates of two melanoma cell lines chosen for their complementary characteristics were combined with the adjuvant DETOX and injected subcutaneously on weeks 1, 2, 3, 4 and 6 for one or two courses, and then monthly in patients with objective clinical responses. Of 109 patients, 22 (20%) have had objective clinical regression of tumor masses, with 5% complete responses. Ten patients have lived more than a year. Eight of the 10 are still alive, five of whom have lived more than 3 years. It was not necessary to achieve complete remissions to cause an increase in survival, and most of the long-surviving patients have one or more (stable) residual nodules. The pace of the disease process has clearly been slowed in those individuals. A rise in the level of cytotoxic T lymphocyte precursors in the blood (pCTL) has correlated with clinical response. Only one patient without such a rise in pCTL has had a response, and assays in that patient were considered unreliable. Both CD4+ and CD8+ CTL have been cloned from the blood of immunized patients. Both types of CTL killed a number of melanoma cell lines, but not other types of tumor or normal cells (lymphoblasts and melanocytes). CD8+ CTL have not been restricted to killing the autologous melanoma. MHC restriction by the HLA-A2 locus was identified. CD4+ CTL were not restricted only by Class II HLA antigens. Many CD4+ clones killed HLA Class II-negative melanomas, and we were able to block cytotoxicity of a particular clone with either anti-HLA Class I or anti-Class II MHC monoclonal antibodies, or both. An association of clinical response to the theraccine with certain HLA phenotypes, notably HLA-C3, -A2 (and the cross-reactive HLA-A28), B12 (and the related alleles (HLA-B44 and -B45) and perhaps DR4, particularly when combinations of those alleles were present, was suggested by our analysis of 70 patients. It is possible that this simply indicates the sharing of MHC antigens between the immunizing melanomas and the patient's melanoma. However, these MHC molecules may be important in their own right in presenting melanoma-associated antigens in CTL in vivo. Subtractive hybridization of mRNA from lung squamous carcinoma cells from cDNA of the M-1 melanoma cell line has yielded several DNA sequences unique to melanoma. Those are now being analyzed for possible immunogenicity, with cytotoxicity by CTL from immunized patients as the major criterion.(ABSTRACT TRUNCATED AT 400 WORDS)     

A simple microassay for detection of antibodies to fetal calf serum and related antigens and its application to the serological definition of human tumor antigens.

The use of cultured cells for the production and testing of anti-tumor antisera is complicated by the fact that cultured cells can adsorb macromolecules derived from the serum supplement used in the culture medium onto their surface membrane. The need to monitor antisera for the presence of this unwanted reactivity has led us to adapt the mixed hemadsorption assay for the detection of antibodies to antigens present in fetal calf serum (FCS), the most widely used serum supplement. Our improved microassay involves the fixation of FCS to the bottom of plastic microplates with glutaraldehyde for the production of an antigen substrate which is stable for at least 16 weeks. The microassay is simple, reproducible, and capable of detecting nanogram quantities of FCS antigens.Anti-FCS reactivity is detectable with this microassay in approximately 90% of sera from both normal individuals and melanoma patients. Similarly, anti-tumor antisera raised in monkeys, rabbits and patients with melanoma against cultured human melanoma material contain high titers of anti-FCS reactivity. But this reactivity can be completely removed by absorption with insolubilized FCS. Xenoantisera raised against melanoma cells grown in human adult serum still show strong reactivity against FSC. This is due to immunologic cross-reactivity between FCS and normal human serum antigens.     

Augmentation of anti-HLA-E antibodies with concomitant HLA-Ia reactivity in IFNγ-treated autologous melanoma cell vaccine recipients*

HLA-E expressed on the surface of melanoma cells and shed into circulation are known to inhibit killing of tumor cells by binding to CD94/NKGA2 receptors on cytotoxic T- and NKT cells. Interferon (IFN)-γ is known to promote HLA-E over-expression on the cell surface and shedding. The shed HLA-E heavy chain may expose cryptic epitopes to elicit antibodies (Abs). The anti-HLA-E Abs may bind to shed HLA-E or to the tumor cell surface to block its interaction with CTL/NKT cells. This is the basis for a melanoma cell vaccine that will generate anti-HLA-E Abs. The objective of this study was to characterize the antibody response and characterize the cross-reactivity of the antibodies produced in melanoma patients immunized with autologous melanoma cells treated with IFNγ. Anti-HLA-E murine mAbs and serum anti-HLA-E Abs in healthy individuals were known to react with HLA-Ia alleles, which is attributed to the presence of peptide sequences shared between HLA-E and HLA-Ia. Therefore, pre- and post-immune (weeks 4 and 24) serum Abs reacting to both HLA-E and HLA-Ia alleles were measured by multiplex Luminex^®-based immunoassay. To ascertain whether the reactivity of the serum Abs to HLA-Ia was due to anti-HLA-E Abs, the shared-peptides were used to inhibit anti-HLA-E and HLA-Ia reactivities. The level of anti-HLA-E IgG in sera has increased post-immunization from its pre-immune level. Concomitantly, the HLA-Ia reactivity of the sera was also augmented. The reactivity of both anti-HLA-E Abs and HLA-Ia were inhibited by the shared-peptides. The HLA-Ia reactivity of the anti-HLA-E Abs in patients’ sera is similar to the HLA-Ia reactivity of the anti-HLA-E mAbs and anti-HLA-E Abs in normal sera. The results establish the immunogenicity of HLA-E and also ascertain that the HLA-Ia reactivity of the anti-HLA-E Abs is due to shared-peptide epitopes.     

Anti-lymphocyte antibodies in plasma of HIV-1–infected patients preferentially react with MHC class Il-negative T cells and are linked to antibodies against gp41

It has previously been shown that HIV-infected patients develop anti-lymphocyte antibodies. The relationship between anti-lymphocyte antibodies and antibodies against different viral antigens is unknown, and it remains controversial whether some lymphocyte subpopulations are targeted preferentially. We have set out using three-colour flow cytometry to measure antibodies against different lymphocyte subsets. Staining with anti-human immunoglobulin and two MoAbs was performed to characterize the immunoglobulin toad of different lymphocyte subsets. Comparison was done between patients'antibody reactivity against HlV-1 antigens and anti-lymphocyte antibodies. We were able to demonstrate the presence of anti-lymphocyte antibodies in approximately 75% of the HIV-infected patients (n = 78) (healthy controls were all negative). MHC class Il-negative T cells showed a stronger reaction with anti-lymphocyte antibodies than B cells or MHC class Il-positive T cells. Patients with antibodies against CD4 lymphocytes showed a significantly higher antibody reaction with the retroviral antigen gp41 than patients without these antibodies. An association between anti-lymphocyte antibodies and antibody reactivity against other HIV-1 antigens was not noticed. In conclusion, anti-lymphocyte antibodies in HIV-I-infected patients show a preferential reactivity with T cells which lack expression of MHC class II molecules. There is an increased antibody reactivity against gp41 in patients with anti-CD4 * T cell antibodies. The association hints at a specific origin of anti-lymphocyte antibodies in HIV-I -infected patients due to cross-reactivity with viral epitopes or network phenomena. These anti-CD4 cell antibodies could be of interest in the clinical course of HIV infection.