Allogeneic gene-modified tumor cells (RCC-26/IL-7/CD80) as a vaccine in patients with metastatic renal cell cancer: a clinical phase-I study.

Despite novel targeted agents, prognosis of metastatic renal cell cancer (RCC) remains poor, and experimental therapeutic strategies are warranted. Transfection of tumor cells with co-stimulatory molecules and/or cytokines is able to increase immunogenicity. Therefore, in our clinical study, 10 human leukocyte antigen (HLA)-A(*)0201(+) patients with histologically-confirmed progressive metastatic clear cell RCC were immunized repetitively over 22 weeks with 2.5-40 × 10(6) interleukin (IL)-7/CD80 cotransfected allogeneic HLA-A(*)0201(+) tumor cells (RCC26/IL-7/CD80). Endpoints of the study were feasibility, safety, immunological and clinical responses. Vaccination was feasible and safe. In all, 50% of the patients showed stable disease throughout the study; the median time to progression was 18 weeks. However, vaccination with allogeneic RCC26/IL-7/CD80 tumor cells was not able to induce TH1-polarized immune responses. A TH2 cytokine profile with increasing amounts of antigen-specific IL-10 secretion was observed in most of the responding patients. Interferon-γ secretion by patient lymphocytes upon antigen-specific and non-specific stimulation was substantially impaired, both before and during vaccination, as compared with healthy controls. This is possibly due to profound tumor-induced immunosuppression, which may prevent induction of antitumor immune responses by the gene-modified vaccine. Vaccination in minimal residual disease with concurrent depletion of regulatory cells might be one strategy to overcome this limitation.

Adoptive tumor therapy with T lymphocytes enriched through an IFN-gamma capture assay.

Successful adoptive T-cell therapy has been demonstrated in viral disease and selected forms of cancer. However, it is limited by the difficulty to efficiently isolate and amplify autologous tumor-reactive T-cell clones. Tetramers of major histocompatibility complex (MHC) class I and peptide have facilitated the characterization of CD8+ T cells specific for tumor-associated antigens. However, for adoptive T-cell therapy, MHC-tetramers have limitations: they require knowledge of tumor antigens, which is often not available; they select T cells with a single specificity, thereby posing risk for selection of tumor escape variants; they do not select for function, so that T cells may be anergic when isolated from cancer patients; and they do not allow the isolation of CD4+ T cells that can be essential for tumor rejection. Because interferon (IFN)-gamma is essential for tumor rejection, we isolated live T cells based on their IFN-gamma production. IFN-gamma secreted by previously activated T cells is retained on the cell surface, allowing their specific isolation and expansion. We show here that IFN-gamma+ but not IFN-gamma- T cells from tumor-immunized mice are cytolytic and mediate tumor rejection upon adoptive transfer. Importantly, tumor-specific T cells can be enriched from lymphocytes infiltrating human renal cell carcinoma by the IFN-gamma capture assay.

Expression of HLA-C molecules confers target cell resistance to some non-major histocompatibility complex-restricted T cells in a manner analogous to allospecific natural killer cells.

Specific HLA molecules have recently been shown to confer target cell resistance to lysis by some CD3- natural killer (NK) cells. For certain NK clones, resistance is governed by two specificities (NK1 and NK2) that are associated with particular HLA-C alleles: in general, target cells expressing Cw1, Cw3, Cw7, or Cw8 are susceptible to NK1 but resistant to NK2 clones, whereas target cells expressing Cw2, Cw4, Cw5, or Cw6 are susceptible to NK2 and resistant to NK1 cells. These two clusters of HLA-C alleles are distinguished by a dimorphism in the alpha 1 helical region, localized at amino acid positions 77 and 80. In this report, we show that highly enriched CD3+/CD56- cytotoxic T cell sublines and CD3-/CD56+ NK sublines derived from the same donor have identical cytolytic specificities when tested against a panel of allogeneic LCL and various HLA-B and -C transfectant cell lines. The lysis pattern of the allogeneic cells appeared to be related to the NK2 specificity for both effector cells: LCL expressing HLA-Cw2, Cw4, Cw5, or Cw6 alleles were lysed, while LCL expressing HLA-Cw1, Cw3, or Cw7 molecules were resistant. Resistance to lysis could be conferred to susceptible target cells by transfection with a Cw*0702 gene, while expression of a Cw*0602 gene did not provide protection. Similar patterns of HLA-C-mediated resistance were also found with two polyclonal T cell lines generated from the peripheral blood lymphocytes of unrelated donors. Thus, major histocompatibility complex (MHC) molecules that induced resistance to particular NK cells also regulated target cell resistance to lysis by these non-MHC-restricted effector T cells. For both types of effector cells, direct binding to HLA-C molecules was necessary to achieve inhibition since preincubation with mAb specific for class I molecules destroyed the protection from lysis of HLA-Cw7 expressing target cells. mAbs specific for CD3 and CD8 molecules had no influence on lysis or inhibition of the NK-like T cells. Formation of MHC complexes with particular peptides did not appear to be essential to confer resistance, since a cell line with defective peptide transporter genes (TAP genes), when transfected with an appropriate HLA-C allele, was as resistant to lysis as HLA-C transfectant lines with normal TAP function. These results suggest that HLA-C molecules may deliver negative regulatory signals to some non-MHC-restricted T cells in a manner similar to that described previously for particular NK cells.

Genetic control of cell-mediated lympholysis in mouse.

H-2 congenic mouse strains were tested in vitro to investigate the genetic control of cell-mediated lympholysis (CML). Combinations were selected such that differences in various segments of H-2 could be examined for their ability to stimulate production of effector cells and to serve as targets for lysis. Particular emphasis was directed towards understanding the roles of LD and SD. SD-region differences are important in the sensitization of effector cells and they also function as strong targets for lysis, or as markers for the CML targets. LD differences are also important for sensitization of cytotoxic effector cells, but they serve only as very weak targets for lysis. Collaboration occurs between LD and SD in generation of CML. The nature of this interaction can be of two types: together LD and SD can produce CML which neither difference alone can stimulate; LD can enhance a CML response stimulated by SD-region differences alone.

In vivo expansion of HLA-B35 alloreactive T cells sharing homologous T cell receptors: evidence for maintenance of an oligoclonally dominated allospecificity by persistent stimulation with an autologous MHC/peptide complex.

The nature of alloantigens seen by T lymphocytes, in particular the role of peptides in allorecognition, has been studied intensively whereas knowledge about the in vivo emergence, diversity, and the structural basis of specificity of alloreactive T cells is very limited. Here we describe human T cell clones that recognize HLA-B35 alloantigens in a peptide-dependent manner. TCR sequence analysis revealed that several of these allospecific clones utilize homologous TCR: they all express TCRAV2S3J36C1 and TCRBV4S1J2S7C2 chains with highly related CDR3 sequences. Thus peptide-specific alloreactivity is reflected in homologous CDR3 sequences in a manner similar to that described for T cells that recognize nominal peptide/self-MHC complexes. The in vivo frequency of this TCR specificity was studied in unstimulated PBL of the responding cell donor who was not sensitized against HLA-B35. The vast majority (approximately 75%) of the VA2S3J36 junctional regions obtained from two samples of PBL, isolated at a 9-yr interval, encode CDR3 identical or homologous to those of the functionally characterized HLA-B35 allospecific T cells. These data are most easily explained by a model of alloreactivity in which persistent or recurrent exposure to a foreign peptide/self-MHC complex led to the in vivo expansion and long-term maintenance of specific T cells that show fortuitous crossrecognition of an HLA-B35/peptide complex and dominate the alloresponse against HLA-B35.

Direct demonstration of an HLA-DR allotypic determinant on the low molecular weight (beta) subunit using a mouse monoclonal antibody specific for DR3.

A murine monoclonal antibody directed against a human B cell surface antigen with the characteristics of HLA-DR is described. The antigen detected is tightly linked to HLA and is correlated with the alloantigen HLA-Dw/DR3. Reactivity with a fraction of Dw/DRw6 cells is also observed. The determinant recognized by this antibody has been shown to be present on the smaller molecular weight beta subunit of the HLA-DR antigen.

The Qa-1 antigenic system. Relation of Qa-1 phenotypes to lymphocyte sets, mitogen responses, and immune functions.

The antiserum (B6 X A-Tlab) anti-A (Tlaa) defines several TL antigens expressed exclusively on thymocytes. When reacted with peripheral lymphocytes, the same antiserum defines another antigenic system, provisionally termed Qa-1. The genotypic disparity distinguishing the recipients and donors in this immunization comprises a section of chromosome 17 extending from a crossover point between H-2D and Tla to a presently unmarked point beyond Tla. Therefore although Qa-1 may constitute a single cell surface component, it is equally probable that the Qa-1 system defines two or more cell surface components determined by genes in this region, each of which may be expressed on a different cell set. Cytotoxicity assays indicate that Qa-1 antigen is expressed on Lyt-1 cells and Lyt-123 cells, and may serve to subclassify these two cell sets; it is not known whether Qa-1+ cells may occur within the small Lyt-23 set. There may be also be a cell set with the phenotype Thy-1--:Qa-1+. Another distinctive feature of the Qa-1 system is the characteristic profile of responses to mitogens exhibited by spleen cell populations from which Qa-1+ cells have been eliminated; in conventional assay of [3H]thymidine incorporation the response to lipopolysaccharide was essentially unchanged, the response to phytohemagglutinin M (PHA-M) was virtually abolished, and the response to concanavalin A (Con A) was reduced by 40%. The third distinctive feature of the Qa-1 system is the characteristic profile of changes which elimination of Qa-1+ cells produces in tests of immune function in vitro: (a) proliferation, measured by [3H]thymidine incorporation, in mixed lymphocyte culture (MLC) with major histocompatibility complex (MHC)-incompatible stimulator cells, was not affected. (b) in tests of cell-mediated cytotoxicity (CMC) of MHC-incompatible target cells, neither the generation nor the effector functions of cytotoxic lymphocytes was affected, implying that Lyt-23 prekiller and killer cells are Qa-1--. (c) primary and secondary responses to SRBC were considerably augmented, suggesting that Qa-1+ cells may be responsible for suppression in this test system. (d) accordingly the suppression of the anti-sheep erythrocyte (SRBC) response normally engendered in spleen cells by culture with SRBC was profoundly reduced by elimination of Qa-1+ cells, either before or after culture. (e) the suppression of the anti-SRBC response normally engendered in spleen cells cultured with Con A was reduced by removal of Qa-1+ cells before but not after culture with Con A. Although analysis is as yet far from complete, the Qa-1 system should already be of considerable value because it distinguishes a population of lymphocytes that is not defined by any other antigenic system, according to three criteria: (a) representation of Qa-1 cells among T-cell sets defined by Lyt phenotypes, (b) the profile of responses to mitogens exhibited by lymphocyte populations depleted of Qa-1+ cells, and (c) the profile of immune responses of lymphocyte populations depleted of Qa-1+ cells.

Cell-mediated lympholysis. Importance of serologically defined H-2 regions.

The cell-mediated lympholytic capability of mouse spleen cells stimulated in mixed lymphocyte culture is related to the major histocompatibility complex genotype on target lymphocytes. The strain combinations AQR-B10. T(6R) and B10.A(4R)-B10.A(2R) that result in significant mixed lymphocyte culture activation do not mediate cell-mediated lympholysis on sensitizing target lymphocytes; serologically defined regions (H-2K and H-2D) are identical within each combination. H-2K or H-2D region disparity alone does not cause cell-mediated lympholysis. However after mixed lymphocyte culture activation as seen with B10.A-B10.T(6R), a target cell bearing only an H-2K region difference from the effector cell is sensitive to cell-mediated lympholysis. Likewise an H-2D region difference is an adequate target after mixed lymphocyte culture activation of the effector cell in the combination B10.A(2R)-B10.D2.

GPI-anchored TIMP-1 treatment renders renal cell carcinoma sensitive to FAS-meditated killing.

The resistance of tumours to immune-mediated lysis has been linked to the biology of matrix metalloproteinases (MMPs), and specifically to the cell surface expression of MMPs by the tumour cell. The endogenous tissue inhibitors of metalloproteinases (TIMPs) exhibit diverse physiological/biological functions including the moderation of tumour growth, metastasis and apoptosis. These biologic activities are mediated in part by the stoichiometry of TIMP/MMP/cell surface protein interactions. A glycosylphosphatidylinositol (GPI) anchor was fused to TIMP-1 to focus defined concentrations of this inhibitory protein on the surface of three renal cell carcinoma (RCC) cell lines (RCC-26, RCC-53 and A498) independently of cell surface protein-protein interactions. Exogenously added TIMP-1-GPI efficiently inserted into the RCC cell membrane and dramatically altered the association of MMPs with the cell surface. TIMP-1-GPI treatment inhibited RCC proliferation and rendered the normally FAS-resistant RCC cells sensitive to FAS-induced apoptosis but did not alter perforin-mediated lysis by cytotoxic effector cells. The increased sensitivity to FAS-mediated apoptosis correlated with an alteration in the balance of pro- and antiapoptotic BCL-2-family proteins. By interfering with the proliferative capacity and inducing sensitivity to immune effector mechanisms GPI-anchored TIMP-1 may represent a more effective version of the TIMP-1 protein for therapeutic strategies.

Identification of naturally processed T cell epitopes from glutamic acid decarboxylase presented in the context of HLA-DR alleles by T lymphocytes of recent onset IDDM patients.

Glutamic acid decarboxylase (GAD) has been defined as a major target antigen in insulin-dependent diabetes mellitus (IDDM). To identify the molecular ligands triggering a T cell response to GAD, a panel of human GAD65-specific T lymphocyte lines was generated from peripheral blood of three recent onset IDDM patients. All lines derived from a patient expressing the high-risk-conferring HLA-DR*0301/ *0401 haplotypes recognized a single epitope localized between amino acid positions 270 and 283 of GAD65, a stretch that is located in close proximity to the homology region shared with Coxsackie virus P2-C protein. All lines with this specificity were restricted to the DRA, B1*0401 product of the DR4 haplotype. Analysis of the GAD-specific T cell response in a second patient homozygous for DR4 haplotypes demonstrated that the same DRA, B1*0401 allele selected T cells specific for a different determinant. The T cell response profile in a third patient showed that DR*1501/ *1601-encoding haplotypes could present at least three different epitopes to GAD65-specific T lymphocytes. One of these epitopes was presented by a DR allele associated with the resistance-conferring DRB1*1501 haplotype. GAD-specific T cell lines could not be isolated from HLA class II-matched normal individuals. Our data reveal that (a) the T cell response to GAD65 is quite heterogenous in recent onset IDDM patients; (b) HLA-DR, not DQ, seems to be the principal restriction element used by T cells present at the onset of the disease; and (c) T cells responding to epitopes containing identical sequences to Coxsackie virus P2-C protein were not detected.

Human renal cell carcinoma antigen-specific CTLs: antigen-driven selection and long-term persistence in vivo.

Renal cell carcinomas (RCCs) are thought to be immunogenic, because cytokine-induced and even spontaneous tumor regression has been observed in a significant number of patients. However, little is known about the nature of immune responses that might lead to tumor regression. We studied naturally arising human T-cell responses against RCC by combining molecular analyses of T-cell receptor (TCR) usage in primary tumors in situ with functional analyses of tumor-infiltrating lymphocytes (TILs) in vitro. TILs of patient 26 that were cultured in vitro showed a human leukocyte antigen (HLA-A*0201)-restricted cytotoxic activity specific for autologous tumor cells. These tumor-derived lymphocytes were dominated by a family of T cells expressing V alpha20- and V beta22-positive TCRs. Their specificity-conferring third complementarity-determining regions were highly homologous with respect to the loop length and selection of particular amino acids in both TCR chains. These characteristics are similar to those reported for antigen-selected murine T cells recognizing immunodominant epitopes of non-self proteins. To evaluate the biological significance of these CTLs in vivo, we analyzed the corresponding TCR transcripts in the cryopreserved tumor material of patient 26 and in a second HLA-A*0201-positive RCC patient whose tumor cells were also lysed by TIL-26. The in situ TIL populations of both patients used related families of highly homologous TCRs, supporting the contention that immunodominant responses directed against a shared tumor-associated antigen occurred in both individuals in vivo. Furthermore, in the absence of overt metastatic disease, the tumor antigen-specific CTLs of patient 26 were shown to persist in the periphery 4 years after removal of the primary tumor. These results demonstrate that antigen-driven T-cell responses specific for spontaneously arising carcinomas developed in these patients and showed long-term persistence, even in the absence of immunotherapy.

Tumor-specific lysis of human renal cell carcinomas by tumor-infiltrating lymphocytes: modulation of recognition through retroviral transduction of tumor cells with interleukin 2 complementary DNA and exogenous alpha interferon treatment.

Two cytotoxic effector cell populations were isolated from a patient with renal cell carcinoma. The tumor-infiltrating lymphocytes comprised a population of highly specific, major histocompatibility complex-restricted, cytotoxic T lymphocytes (CTL). An autologous non-major histocompatibility complex-restricted lymphokine-activated killer (LAK) cell population was generated by culturing the peripheral blood lymphocytes with high doses of recombinant interleukin 2 (rIL-2). The capacity of these two effector cell types to lyse cytokine-modulated autologous tumor cells was compared in vitro. A complementary DNA for rIL-2 was introduced into the tumor cells by retroviral transduction, and tumor cells secreting low doses of rIL-2 were isolated. The CTL recognition of these tumor cells was enhanced, compared to unmodified tumor cells, whereas LAK cell recognition was unchanged or slightly reduced. Pretreatment of tumor cells with exogenous alpha interferon led to an up-regulation of some major histocompatibility complex class I molecules and to slightly better recognition by the CTL; little effect on LAK cell recognition was observed. CTL were found to be 50-150-fold more effective than LAK cells in lysing autologous tumor cell lines or clones modulated with both rIL-2 and alpha interferon. The assessment of a patient's cytotoxic immune capacity directed against genetically modified autologous tumor cells in vitro provides important insight for cytokine-mediated gene therapy of cancer.

HLA-DQ-restricted, islet-specific T-cell clones of a type I diabetic patient. T-cell receptor sequence similarities to insulitis-inducing T-cells of nonobese diabetic mice.

We established a T-cell line and 20 CD4+ T-cell clones from the peripheral blood lymphocytes of a type I diabetic patient using a membrane preparation of a rat insulinoma cell line (beta membrane antigen [BMA]) as a source of antigen. The T-cell line and three selected clones proliferated specifically to stimulation with BMA and to membranes prepared from human islets, rat pancreas, and NOD pancreas, but not to control antigens. Proliferation-inhibition studies using human leukocyte antigen (HLA)-specific monoclonal antibodies revealed HLA-DQw1-restricted recognition of BMA. An analysis of the T-cell receptor (TCR) repertoire of the T-cell line after 8 and 40 days of culture showed a prominent usage of the V alpha 1 and V alpha 12 gene families. Although sequencing of the TCR V alpha and V beta chains of the three clones demonstrated that each used different V alpha and V beta gene segments, structural similarities were found in complementary-determining region 3 (CDR3), the region that is postulated to interact with the peptide component of the TCR ligand. When we compared these TCR sequences with published sequences of disease-inducing T-cells of NOD mice, highly related TCR V beta families were detected. Furthermore, stretches of identical amino acids within the CDR3 region were found between two pairs of human and mouse T-cells. If one considers the species differences in TCR genes and sequence differences in the restriction elements for these cells (HLA-DQ vs. H-2 I-A nod), these sequence similarities are striking and may be useful for pinpointing T-cells of primary importance in the development of disease.

Cellular and molecular analyses of major histocompatibility complex (MHC) restricted and non-MHC-restricted effector cells recognizing renal cell carcinomas: problems and perspectives for immunotherapy.

Renal cell carcinomas belong to the small group of tumors that are able to induce antitumor responses. Here we describe two general types of cytotoxic effector lymphocytes that can eliminate autologous tumor cells and discuss the role that major histocompatibility complex encoded molecules play in governing their specificities. Improved understanding of the cellular and molecular basis of renal cell carcinoma recognition opens new avenues of research with the potential to develop better immunotherapies for patients with metastatic disease.

The use of fusion proteins to study HLA-B27-specific allorecognition.

Potential antigenic regions of the various external domains of the HLA-B27 antigen were expressed as fusion proteins in bacterial hosts and analyzed for their ability to induce humoral and cellular responses. Monoclonal antibodies directed against the proteins recognized monomorphic determinants of denatured HLA-antigens, but not B27-antigens expressed by intact lymphocytes. T-cell proliferation and IL-2 secretion were induced with a fusion protein representing regions of the first and second domains around amino acid residue 114. None of the fusion proteins stimulated cytotoxic T-lymphocytes (CTL) in an HLA-specific manner, although several included those amino acid sequences thought to be important for CTL recognition.

Double suicide gene (cytosine deaminase and herpes simplex virus thymidine kinase) but not single gene transfer allows reliable elimination of tumor cells in vivo.

Suicide genes such as cytosine deaminase (CD) and herpes simplex virus thymidine kinase (TK) encode products that convert nontoxic substances (prodrugs) into toxic metabolites. Suicide gene transfer is currently being used in cancer therapy or can be used as a safety modality. To analyze the reliability of suicide genes as a safety modality for a vaccination study with viable cytokine/B7 gene-modified tumor cells, the individual and combined efficacy of the two suicide genes was compared for in vitro and in vivo cell killing of a murine mammary adenocarcinoma cell line (TS/A). To adapt the system to an in vivo gene delivery situation, bulk cultures cotransfected with the CD and TK gene were used instead of selected clones. In vitro, both CD and TK conferred sensitivity to the respective prodrug but the combined cytotoxic effects of both gene products were always superior. For in vivo analysis BALB/c mice were injected subcutaneously with CD- and TK-modified TS/A cells, treated with prodrugs, and tumor size was evaluated for a period of 100 days. In the in vivo situation the combination of both enzyme/prodrug systems was again most effective. The highest single concentration of 5-FC (500 mg/kg) or GCV (100 mg/kg) was not able to fully protect the animals from developing tumors, whereas a combination of 5-FC (250 mg/kg) and GCV (50 mg/kg) resulted in complete tumor eradication. In nude mice treated in the same way, most CD/TK tumors could not be eliminated. Furthermore, BALB/c mice cured of TS/A-CD/TK tumors developed a systemic tumor immunity against challenge with parental TS/A cells. These findings indicate that reliable tumor elimination by the suicide genes depends on T cells. The cooperative effect of both suicide genes was confirmed in vitro with the human renal cell carcinoma line RCC26. We conclude that TK and CD together, but neither gene alone, act as a safety mechanism for the elimination of tumor cells in a reliable fashion and suggest that a rapid and quantitative antigen release by effective TK- and CD-mediated tumor destruction is necessary for T cell immunity to develop.

HLA-C revisited. Ten years of change.

During the past 10 years knowledge about the interactions between major histocompatibility complex (MHC) class I molecules and the T-cell receptor (TCR) complex of cytotoxic T-cells (CTL) has developed dramatically. But the primary interest, both with respect to structure as well as function, has concentrated on HLA-A and -B molecules because of their high sequence polymorphism and their dominating presence at the cell surface. In contrast, HLA-C molecules seemed to be of only minor importance in the cascade of immune reactions owing to their more limited polymorphism and reduced levels of surface expression. The inability to define a number of antigen specificities had the result that HLA-C molecules were often neglected in studies of immune response, transplantation, and disease association. More recently a new function has been identified for HLA class I molecules where they act as inhibitors of the lytic capacity of natural killer (NK) cells and non-MHC-restricted T-cells. Moreover, the understanding of this novel mode of negative regulation of cytotoxicity was remarkably influenced by HLA-C since these were the first HLA class I molecules found to have such inhibitory potential. With this new inhibitory function serving as an essential component of the immune system, HLA-C molecules can no longer be neglected.

Distribution of cell adhesion molecules (ICAM-1, VCAM-1, ELAM-1) in renal tissue during allograft rejection.

The tissue distribution of cellular adhesion molecules (CAMs) was studied in specimens from 10 normal human kidneys and in 52 biopsies from kidney allografts with cell-mediated rejection. In addition to the vascular presence of ICAM-1, a common finding in normal kidneys, expression of ICAM-1 on tubular cells was observed in 22 graft biopsies. Compared with normal kidneys, where VCAM-1 was present on Bowman's capsules and few proximal tubular cells, a markedly enhanced expression of VCAM-1 in numerous tubuli (including distal tubular segments) was observed in 51 graft biopsies. In 41 graft specimens VCAM-1 appeared also in variable numbers of peritubular capillaries. Infiltrating leukocytes carrying VCAM-1 were observed in 7 grafts. ELAM-1 could not be found in normal kidneys but was restricted to some peritubular capillaries in 29 grafts. Comparable results were obtained with cultured renal tubular cells when stimulated by TNF-alpha. That the induced appearance of adhesion molecules was in fact related to actual cellular synthesis was demonstrated by Northern blot analysis. Thus, little ICAM-1 specific mRNA of 3.4-kb length could be detected in unstimulated cultured renal tubular cells, but hybridization was markedly increased after stimulation with TNF-alpha. A substantial amount of VCAM-1 specific mRNA of 3.2-kb length was present already in unstimulated renal tubular cells. Likewise, synthesis of VCAM-1 mRNA was enhanced by stimulation with TNF-alpha. TNF-stimulated endothelial cells also showed weak synthesis of VCAM-1 mRNA. The results provide further evidence that constitutive and inducible expression of cell adhesion molecules contributes to the process of allograft rejection.

On the peptide model of allorecognition: cytotoxic T lymphocytes recognize an alloantigen encoded by two HLA-linked genes.

The peptide model of allorecognition hypothesizes that alloreactive cytotoxic T lymphocytes recognize peptides associated with major histocompatibility complex (MHC) molecules. This study characterizes an unusual alloantigen that is recognized by two human cytotoxic T-cell lines and may represent a complex formed by the association of a naturally selected peptide with an MHC molecule. Family studies demonstrate that both components of the alloantigen are products of MHC-linked genes. One component is a class I molecule most likely encoded by the HLA-B locus. The second component is encoded by an MHC-linked gene that shows a limited polymorphism; whether it represents the product of a second class I gene or of some other MHC gene remains to be determined. These data provide experimental evidence supporting the peptide model of allorecognition in human beings and indicate that some naturally selected peptides involved in allorecognition may be derived from MHC-encoded proteins.

Noncodominant expression of target antigens recognized by human cytotoxic T lymphocytes.

Human CTL that recognized MHC-controlled determinants distinct from HLA-A. B, C, and D/DR antigens were tested in a family with nine siblings. Segregation analysis of positive CML reactions showed strong lysis of target cells of three HLA-identical siblings, inheriting the b and c MHC haplotypes (blc), but not of the parents or siblings inheriting only one of these haplotypes. Some cytotoxicity was seen against parental target cells, although it seemed to be qualitatively distinct from that directed against the b/c targets. Studies using the competitive inhibition technique showed that cells inheriting only the b or c haplotypes were not effective in decreasing the specific cytotoxicity. Furthermore, simultaneous inclusion of inhibitor cells of both the b and c haplotypes did not hinder the specific cytotoxicity. Induction of CTL recognizing this new determinant did not occur when either antigens of the b or c haplotypes were used as MLC stimulating cells, or when antigens of both haplotypes were presented simultaneously, but on separate stimulating cells, in a three-cell MLC. These results suggest that the target determinant recognized by these unusual CTL is complex; it may be formed through interactions of two surface molecules or by genetic complementation yielding a hybrid antigen.