Immune selection of hot-spot beta 2-microglobulin gene mutations, HLA-A2 allospecificity loss, and antigen-processing machinery component down-regulation in melanoma cells derived from recurrent metastases following immunotherapy

Scanty information is available about the mechanisms underlying HLA class I Ag abnormalities in malignant cells exposed to strong T cell-mediated selective pressure. In this study, we have characterized the molecular defects underlying HLA class I Ag loss in five melanoma cell lines derived from recurrent metastases following initial clinical responses to T cell-based immunotherapy. Point mutations in the translation initiation codon (ATG-->ATA) and in codon 31 (TCA-->TGA) of the beta(2)-microglobulin (beta(2)m) gene were identified in the melanoma cell lines 1074MEL and 1174MEL, respectively. A hot-spot CT dinucleotide deletion within codon 13-15 was found in the melanoma cell lines 1106MEL, 1180MEL, and 1259MEL. Reconstitution of beta(2)m expression restored HLA class I Ag expression in the five melanoma cell lines; however, the HLA-A and HLA-B,-C gene products were differentially expressed by 1074MEL, 1106MEL, and 1259MEL cells. In addition, in 1259MEL cells, the Ag-processing machinery components calnexin, calreticulin, and low m.w. polypeptide 10 are down-regulated, and HLA-A2 Ags are selectively lost because of a single cytosine deletion in the HLA-A2 gene exon 4. Our results in conjunction with those in the literature suggest the emergence of a preferential beta(2)m gene mutation in melanoma cells following strong T cell-mediated immune selection. Furthermore, the presence of multiple HLA class I Ag defects within a tumor cell population may reflect the accumulation of multiple escape mechanisms developed by melanoma cells to avoid distinct sequential T cell-mediated selective events.     

Lack of correlation between levels of MHC class I antigen and susceptibility to lysis of small cellular lung carcinoma (SCLC) by natural killer cells

Small cellular lung carcinoma (SCLC) cell lines are susceptible to lysis by NK cells. SCLC, normally negative for MHC class I Ag, were rendered positive for HLA-A and -B Ag by two methods: treatment with IFN-gamma or transfection with HLA class I genes. Exposure to IFN-gamma induced high levels of class I Ag and reduced susceptibility to NK-mediated lysis. However, transfection with either HLA-A2, HLA-B27, or HLA-B27 with beta 2m did not result in reduced susceptibility to NK cells. These transfectants expressed amounts of HLA class I Ag comparable to those in IFN-gamma-treated, untransfected cells. Transfection with the beta 2m gene or plasmid alone neither influenced levels of surface class I Ag nor resulted in reduced susceptibility to lysis by NK cells. Thus, the effects of IFN-gamma on NK susceptibility can be dissociated from the induction of class I Ag.     

The selection of tumor variants with altered expression of classical and nonclassical MHC class I molecules: implications for tumor immune escape

Tumor immune escape variants can be identified in human and experimental tumors. A variety of different strategies are used by tumor cells to avoid recognition by different immune effector mechanisms. Among these escape routes, alteration of MHC class I cell surface expression is one of the mechanisms most widely used by tumor cells. In this review we focus our attention on the T-cell immune selection of MHC class I–deficient tumor variants. Different altered MHC class I phenotypes that originate from multiple molecular mechanisms can be identified in human tumors. MHC-deficient tumor clones can escape T-cell immune responses, but are in theory more susceptible to NK-cell–mediated lysis. In this context, we also review the controversial issue of the aberrant expression of nonclassical HLA class I molecules, particularly HLA-G, in tumors. This expression may be relevant in tumor cells that have lost the capacity to interact with NK inhibitory receptors—namely, those tumor cells with no HLA-B or HLA-C expression. Most published studies have not analyzed these possibilities and do not provide information about the complete HLA-A, HLA-B, or HLA-C molecule profiles of the tumors studied. In contrast, HLA-E has been reported to be expressed in some tumor cell lines with very low HLA-A, HLA-B, and HLA-C expression, suggesting that HLA-E may indeed, in some cases, play a role by inhibiting NK lysis of cells that otherwise would be destroyed by NK cells. Finally, we provide evidence that the status of the immune system in the tumor-bearing animal is capable of defining the MHC profile of the tumor cells. In other words, MHC class I–negative metastatic colonies are produced in immunocompetent animals, and MHC class I–positive colonies in T-cell immunodeficient individuals.This article forms part of the Symposium in Writing Tumor escape from the immune response, published in Vol. 53.     

Phenotypical and functional characterization of the CD8+ T cell repertoire of HLA-A2.1 transgenic, H-2KbnullDbnull double knockout mice.

Homozygous HLA-A2.1 transgenic H-2KbnullDbnull double knockout (KO) mice were created. Their potential to develop HLA-A2. 1-restricted cytolytic responses was compared with that of their classical transgenic counterparts, which still express H-2Kb, Db molecules. On cell surfaces, both strains express similar amounts of chimeric (alpha 1 alpha 2 domains of human, alpha 3 cytoplasmic domains of mouse) HLA-A2.1 molecules in noncovalent association with mouse beta 2-microglobulin. Compared with mice that are totally deprived of histocompatibility class Ia molecules (H-2KbnullDbnull double KO), the expression of HLA-A2.1 in transgenic/double KO mice resulted in sizeable increase in the periphery of CD8+ T cells with a normally diversified TCR repertoire. A biased education in favor of HLA-A2.1, ascribable to the absence of H-2 class Ia molecules, was evidenced in these transgenic/double KO mice by their improved capacity to mount HLA-restricted cytolytic responses, regardless of whether they were virally infected or injected with synthetic epitopic peptide. HLA class I transgenic, H-2 class Ia KO mice should represent useful animal models for the preclinical evaluation of vaccine formulations aiming at the induction of HLA class I-restricted CTL responses.     

Expression and release of HLA-E by melanoma cells and melanocytes: potential impact on the response of cytotoxic effector cells

HLA-E are nonclassical MHC molecules with poorly characterized tissue distribution and functions. Because of their capacity to bind the inhibitory receptor, CD94/NKG2A, expressed by NK cells and CTL, HLA-E molecules might play an important role in immunomodulation. In particular, expression of HLA-E might favor tumor cell escape from CTL and NK immunosurveillance. To address the potential role of HLA-E in melanoma immunobiology, we assessed the expression of these molecules ex vivo in human melanoma biopsies and in melanoma and melanocyte cell lines. Melanoma cell lines expressed no or low surface, but significant intracellular levels of HLA-E. We also report for the first time that some of them produced a soluble form of this molecule. IFN-gamma significantly increased the surface expression of HLA-E and the shedding of soluble HLA-E by these cells, in a metalloproteinase-dependent fashion. In contrast, melanocyte cell lines constitutively expressed HLA-E molecules that were detectable both at the cell surface and in the soluble form, at levels that were poorly affected by IFN-gamma treatment. On tumor sections, a majority of tumor cells of primary, but a low proportion of metastatic melanomas (30-70 and 10-20%, respectively), expressed HLA-E. Finally, HLA-E expression at the cell surface of melanoma cells decreased their susceptibility to CTL lysis. These data demonstrate that HLA-E expression and shedding are normal features of melanocytes, which are conserved in melanoma cells of primary tumors, but become dependent on IFN-gamma induction after metastasis. The biological significance of these findings warrants further investigation.     

Peptide-induced modulation of target cell sensitivity to natural killing.

We have previously shown that the capacity of class I molecules to confer resistance to NK in transfected target cells maps to the Ag-binding site (ABS) of the HLA class I structure. Here we examine the effect of peptide (reagents specific for the ABS) pretreatment on the NK sensitivity of class I+ target cells. Synthetic peptides (10-17 amino acids in length) were used to pretreat C1R target cells expressing either no serologically detectable HLA-A, B class I molecules, or C1R transfectants expressing individual HLA-A or -B locus class I molecules. In each case in which the class I allele had previously been shown to directly bind a given peptide, peptide-pulsing of target cells resulted in increased sensitivity to NK-mediated conjugation and cytolysis. The NK susceptibility of C1R target cells expressing no HLA-A, B class I molecules or the nonprotective HLA-A2.1 or HLA-A2M70 mutant class I molecules was unaffected by pretreatment with HLA-A2-binding peptides. These results support the intimate involvement of the HLA class I ABS and potentially ABS-bound peptides in determining target cell sensitivity to NK. Furthermore, these findings form the basis of an effective screening procedure for discerning peptide class I allele-specific interactions.     

Amyloid precursor-like protein 2 association with HLA class I molecules

Amyloid precursor-like protein 2 (APLP2) is a ubiquitously expressed protein. The previously demonstrated functions for APLP2 include binding to the mouse major histocompatibility complex (MHC) class I molecule H-2K^d and down regulating its cell surface expression. In this study, we have investigated the interaction of APLP2 with the human leukocyte antigen (HLA) class I molecule in human tumor cell lines. APLP2 was readily detected in pancreatic, breast, and prostate tumor lines, although it was found only in very low amounts in lymphoma cell lines. In a pancreatic tumor cell line, HLA class I was extensively co-localized with APLP2 in vesicular compartments following endocytosis of HLA class I molecules. In pancreatic, breast, and prostate tumor lines, APLP2 was bound to the HLA class I molecule. APLP2 was found to bind to HLA-A24, and more strongly to HLA-A2. Increased expression of APLP2 resulted in reduced surface expression of HLA-A2 and HLA-A24. Overall, these studies demonstrate that APLP2 binds to the HLA class I molecule, co-localizes with it in intracellular vesicles, and reduces the level of HLA class I molecule cell surface expression.     

Cell-surface expression and alloantigenic function of a human nonclassical class I molecule (HLA-E) in transgenic mice

We have introduced the gene (E*01033) encoding the heavy chain of the human nonclassical MHC class I Ag, HLA-E, into the mouse genome. Two founder mice carry a 21-kb fragment, the others bear an 8-kb fragment. Each of the founder mice was mated to mice of an already established C57BL/10 transgenic line expressing human beta2-microglobulin (beta2m). Cell surface HLA-E was detected on lymph node cells by flow cytometry only in the presence of endogenous human beta2m. However, HLA-E-reactive mouse CTL (H-2-unrestricted) lysed efficiently the target cells originating from HLA-E transgenic mice without human beta2m, showing that the HLA-E protein can be transported to the cell surface in the absence of human beta2m, presumably by association with murine beta2m. Rejection of skin grafts from HLA-E transgenic mice demonstrates that HLA-E behaves as a transplantation Ag in mice. HLA-E transgenic spleen cells are effective in stimulating an allogeneic CTL response in normal and human classical class I (HLA-B27) transgenic mice. Furthermore, results from split-well analysis indicate that the majority of the primary in vivo-induced CTL recognizes HLA-E as an intact molecule (H-2-unrestricted recognition) and not as an HLA-E-derived peptide presented by a mouse MHC molecule, although a small fraction (ranging from 4 to 21%) of the primary in vivo-induced CTL is able to recognize HLA-E in an H-2-restricted manner. Based on these observations, we conclude that HLA-E exhibits alloantigenic properties that are indistinguishable from classical HLA class I molecules when expressed in transgenic mice.     

Striking paucity of HLA-A, B, C and beta 2-microglobulin on human neuroblastoma cell lines

Monoclonal antibodies to beta 2-microglobulin (beta 2m), and to the native two-chain molecule, were used to assess the expression of the HLA-A, B, C molecules on human neuroblastoma-derived cell lines. In radioimmuno-, cytotoxic, and microscopic assays, employing fresh and fixed cells, neuroblastoma cells show at best weak activity as compared to glial or lymphoid cells. In binding inhibition assays, neuroblastoma extracts were 200- to 1800-fold less efficient in inhibiting the antibodies than were glial or lymphoid extracts. Immunoprecipitation and SDS-PAGE analysis confirmed that a beta m-like chain is synthesized by the neuroblastoma cells, but the HLA chain could not be visualized by this technique. HLA-A, B, C and beta 2m levels are known to vary among tissues and cell lines. Yet the magnitude of the differences between the neuroblastoma and lymphoid lines is much greater than the reported differences in expression between some of these same lymphoid lines and many other nonlymphoid malignant or nonmalignant cell types. Metastatic neuroblastoma tumor in bone marrow also showed weak HLA-A, B, C activity, with the cells appearing negative in microscopic assays. Possible clinical implications are discussed.     

HLA-E: strong association with beta2-microglobulin and surface expression in the absence of HLA class I signal sequence-derived peptides

The nonclassical class I HLA-E molecule folds in the presence of peptide ligands donated by the signal sequences of permissive class I HLA alleles, with the aid of TAP and tapasin. To identify HLA-E-specific Abs, four monoclonals of the previously described MEM series were screened by isoelectric focusing (IEF) blot and immunoprecipitation/IEF on >30 single-allele class I transfectants and HLA-homozygous B lymphoid cells coexpressing HLA-E and HLA-A, -B, -C, -F, or -G. Despite their HLA-E-restricted reactivity patterns (MEM-E/02 in IEF blot; MEM-E/07 and MEM-E/08 in immunoprecipitation), all of the MEM Abs unexpectedly reacted with beta(2)-microglobulin (beta(2)m)-free and denatured (but not beta(2)m-associated and folded) HLA-E H chains. Remarkably, other HLA-E-restricted Abs were also reactive with free H chains. Immunodepletion, in vitro assembly, flow cytometry, and three distinct surface-labeling methods, including a modified (conformation-independent) biotin-labeling assay, revealed the coexistence of HLA-E conformers with unusual and drastically antithetic features. MEM-reactive conformers were thermally unstable and poorly surface expressed, as expected, whereas beta(2)m-associated conformers were either unstable and weakly reactive with the prototypic conformational Ab W6/32, or exceptionally stable and strongly reactive with Abs to beta(2)m even in cells lacking permissive alleles (721.221), TAP (T2), or tapasin (721.220). Noncanonical, immature (endoglycosidase H-sensitive) HLA-E glycoforms were surface expressed in these cells, whereas mature glycoforms were exclusively expressed (and at much lower levels) in cells carrying permissive alleles. Thus, HLA-E is a good, and not a poor, beta(2)m assembler, and TAP/tapasin-assisted ligand donation is only one, and possibly not even the major, pathway leading to its stabilization and surface expression.     

HLA-E and HLA-G expression on porcine endothelial cells inhibit xenoreactive human NK cells through CD94/NKG2-dependent and -independent pathways

Human NK cells contribute a significant role to host defense as well as xenogeneic cytotoxicity. Previous studies using human 721.221 cell line have shown that peptides derived from the leader sequence of the HLA-G binds and up-regulates the surface expression of HLA-E molecules, which was considered to consequently provide negative signals to human NK cells. However, the direct role of HLA-G in inhibiting human NK cells remains controversial. In this study, we showed that the expression of HLA-G or HLA-E in porcine endothelial cells directly protected sensitive porcine cells from human NK cell-mediated xenogeneic cytotoxicity. Ab blocking assays using F(ab')2 of the HLA class I-specific mAb PA2.6 indicated that the protection was directly mediated by the expression of HLA-G and HLA-E on the porcine cells. The HLA-E-mediated protection was blocked by anti-human CD94 Ab. In addition, the engagement of HLA-E lead to the phosphorylation of the CD94/NKG2 complex and the recruitment of SH2 domain-containing protein phosphatase 1 (SHP-1) to the complex. Therefore, HLA-E protected porcine cells from xenoreactive human NK cells through a CD94/NKG2-dependent pathway. In contrast, HLA-G inhibited human NK cells in the absence of CD94/NKG2 phosphorylation or SHP-1 recruitment, and the inhibition was not blocked by anti-CD94 Ab. Therefore, HLA-G protected porcine cells from human NK cells through a CD94/NKG2-independent pathway. These results demonstrated that both HLA-E and HLA-G could directly inhibit human NK cells in the absence of other endogenous HLA class I molecules. These results also have practical implications in preventing xenograft rejection mediated by human NK cells.     

Dependence of elevated human leukocyte antigen class I molecule expression on increased heavy chain, light chain (beta 2-microglobulin), transporter associated with antigen processing, tapasin, and peptide

Human leukocyte antigen (HLA) class I molecule expression was investigated by DNA-mediated gene transfer. Cell surface expression was increased up to 75% by transfection of HLA-A2 or HLA-B8 heavy chain genes but not genes encoding light chains (beta(2)-microglobulin (beta(2)m)), transporter associated with antigen processing (TAP), or tapasin. Interferon (IFN) treatment further increased expression of transfected heavy chains, suggesting that IFN inducible molecules support heavy chain expression. IFN induces beta(2)m, TAP, and tapasin mRNAs. Transfected heavy chain expression increased upon cotransfection with genes encoding TAP1 and TAP2 but not individual TAP subunits, beta(2)m, or tapasin. Tetracycline inducible heavy chain gene expression was also increased by IFN treatment or TAP cotransfection, suggesting that IFN-induced TAP supports heavy chain maturation. Expression of a mutant that does not interact strongly with TAP, HLA-A2-T134K, was also increased by IFN. Inhibition of TAP-dependent peptide transport by ICP47 reduced heavy chain expression. Expression of HLA-A2, but not HLA-B8, was restored in ICP47 cells by HLA-A2-binding (IP-30) signal peptides. However, these peptides did not further increase transfected HLA-A2 expression, suggesting that peptide availability does not limit heavy chain expression in the absence of ICP47. These results suggest that cytokine-induced TAP supports maturation of HLA class I molecules through combined chaperone and peptide supply functions.     

HLA class I expression in metastatic melanoma correlates with tumor development during autologous vaccination

Our knowledge of the mechanisms underlying tumor-specific immune response and tumor escape has considerably increased. HLA class I antigen defects remain an important tumor escape mechanism since they influence the interactions between tumor cells and specific T and NK cells in the course of malignant disease. We have studied here HLA class I expression in six subcutaneous metastases obtained from a melanoma patient immunized with an autologous melanoma cell vaccine (M-VAX). We report in this paper that HLA class I antigen expression on these metastatic lesions strongly correlated with the course of the disease. The three metastases that were partially regressing at the time of their excision showed a strong HLA class I expression, whereas the progressing ones showed a very weak or negative staining with most of the anti-HLA class I mAbs used. Real-time quantitative PCR of the samples obtained from microdissected tumor tissue revealed a significant difference in the mRNA levels of HLA-ABC heavy chain and beta2m between the two types of metastases, i.e., lower levels in progressing metastases and high levels in regressing ones, confirming the immunohistological findings. This is, to our knowledge, the first report where the clinical outcome of different HLA class I positive and negative melanoma metastases can be clearly correlated with the regression and progression of the disease, respectively.     

Trophoblast cell line resistance to NK lysis mainly involves an HLA class I-independent mechanism

The lack of classical HLA molecules on trophoblast prevents allorecognition by maternal T lymphocytes, but poses the problem of susceptibility to NK lysis. Expression of the nonclassical class I molecule, HLA-G, on cytotrophoblast may provide the protective effect. However, the class I-negative syncytiotrophoblast escapes NK lysis by maternal PBL. In addition, while HLA-G-expressing transfectants of LCL.721.221 cells are protected from lymphokine-activated killer lysis, extravillous cytotrophoblast cells and HLA-G-expressing choriocarcinoma cells (CC) are not. The aim of this work was therefore to clarify the role of HLA class I expression on trophoblast cell resistance to NK lysis and on their susceptibility to lymphokine-activated killer lysis. Our results showed that both JAR (HLA class I-negative) and JEG-3 (HLA-G- and HLA-Cw4-positive) cells were resistant to NK lysis by PBL and were equally lysed by IL-2-stimulated PBL isolated from a given donor. In agreement, down-regulating HLA class I expression on JEG-3 cells by acid treatment, masking these molecules or the putative HLA-G (or HLA-E) receptor CD94/NKG2 and the CD158a/p58.1 NKR with mAbs, and inducing self class I molecule expression on JAR cells did not affect NK or LAK lysis of CC. These results demonstrate that the resistance of CC to NK lysis mainly involves an HLA class I-independent mechanism(s). In addition, we show that the expression of a classical class I target molecule (HLA-B7) on JAR cells is insufficient to induce lysis by allospecific polyclonal CTL.