| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Immunology |
1 Unit of Human Tumors Immunobiology and 2 Department of Pathology, Istituto Nazionale per lo Studio e la Cura dei Tumori, Milan, Italy and 3 Department of Immunology, Roswell Park Cancer Institute, Buffalo, New York
Requests for reprints: Andrea Anichini, Human Tumors Immunobiology Unit, Department of Experimental Oncology, Istituto Nazionale per lo Studio e la Cura dei Tumori, Via Venezian 1, 20133 Milan, Italy. Phone: 39-0223902817; Fax: 39-0223902630; E-mail: andrea.anichini{at}istitutotumori.mi.it.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
CD4+ T cells recognize complexes of MHC class II molecules with peptides generated from degradation of exogenous or endogenous proteins (3). A wide array of TAAs (see ref. 4 for update) is naturally processed in neoplastic cells and to yield peptides that associate with MHC class II molecules for recognition by CD4+ T cells. Intracellular proteins can access the vacuolar compartment by sequestration in double-membraned organelles named autophagosomes, which then fuse with vesicles of the endosomal/lysosomal compartment where antigen processing and peptide loading onto MHC class II molecules take place (5, 6). Processing of endogenous proteins can also be initiated by cytosolic proteases, and the resulting peptide fragments are then transferred from the cytoplasm to the endosomal/lysosomal compartment by transport mechanisms that require the function of proteins as LAMP-2 and HSC-70 (3).
It has been shown that loss of some APM components, such as TAP1, causes marked MHC class I antigen down-regulation in human cells as well as defective antigen presentation to CTLs (7). Furthermore, qualitative analysis of surgically removed tumors, with immunohistochemical techniques, has shown frequent defects in HLA class I and II and APM components expression in neoplastic cells (8–17), but the evaluation of the potential effect of these defects on clinical outcome has not provided concordant result in all tumors investigated. In fact, in melanoma and head and neck carcinomas, down-regulation or loss of expression of HLA class I or of some APM components is significantly associated with disease progression or with shorter patients' survival (8, 10–13). In contrast, down-regulation or loss of HLA class I antigens has been shown to correlate with better prognosis in colorectal and breast cancers (14, 15). Discordant results have been obtained even by looking at the effect on clinical outcome of HLA class II antigen expression of tumor cells. Thus, in metastatic melanoma, based on a specific pattern of expression of HLA class I and II on neoplastic cells, early results suggested that a high expression of HLA class II antigens may be associated with shorter patients' survival (8). On the other hand, in large B-cell lymphoma and colorectal carcinomas, HLA-DR expression on neoplastic cells is associated with improved survival (16, 17).
In addition, the relationship between quantitative variations in the overall APM component, HLA class I and II expressions, and clinical outcome has not been investigated previously in any human tumor. Here, by taking advantage of a recently developed panel of APM component-specific monoclonal antibodies (mAb; refs. 18, 19) and of appropriate intracellular staining technique (20, 21), we have evaluated APM component and HLA class I and II expressions in a large panel of human melanomas by flow cytometry. Melanoma was selected for this study based on the evidence that immunologic events are involved in the clinical course of this disease (22, 23) and that HLA and APM defects are frequently found in this tumor (8, 10, 13). The results of this quantitative analysis were evaluated for the potential association of the APM and HLA class I and II profiles with relevant clinical variables.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
Monoclonal antibodies. The Delta(Y)-specific mAbs SY-4 and SY-5, the MB-1(X)-specific mAb SJJ-3, the LMP2-specific mAb SY-1, the LMP7-specific mAb HB-2, the LMP10-specific mAbs TO-6 and TO-7, the calnexin-specific mAb TO-5, the calreticulin-specific mAb TO-11, the ERp57-specific mAb TO-2, the tapasin-specific mAb TO-3, the mAb HC-10 that recognizes a determinant expressed on all ß2M-free HLA-B and C heavy chains and on ß2M-free HLA-A10, HLA-A28, HLA-A29, HLA-A30, HLA-A31, HLA-A32, and HLA-A33 heavy chains, the ß2M-specific mAb L368, the HLA-DR, HLA-DQ, and HLA-DP-specific mAb LGII-612.14, and the MHC class I–related molecule B (MICB)–specific mAb SJJ-5 were developed and characterized as described (18, 19, 24–26). The TAP1-specific mAb NOB-1 and the TAP2-specific mAb NOB-2 are secreted by hybridomas derived from the fusion of murine myeloma cells P3-X63-Ag8.653 with splenocytes from BALB/c mice immunized with partial length TAP1 recombinant protein (434-735) and a keyhole limpet hemocyanin–conjugated TAP1 peptide (717-735) and with partial length TAP2 recombinant protein (316-703), respectively, by the strategy described elsewhere (19). FITC-conjugated goat anti-mouse IgG antibody was purchased from Jackson ImmunoResearch (West Grove, PA).
Cell staining and fluorescence-activated cell sorting analysis. Cells were permeabilized and intracellularly stained by sequential incubation with primary and secondary mAbs as described (20, 21). Myeloid dendritic cells generated from CD14+ monocytes (27), after permeabilization, were used as positive controls for the staining by APM component-specific and HLA antigen-specific mAbs. As a positive control for cell permeabilization, melanoma cells were stained with a human vimentin-specific mAb (Cymbus Biotechnology Ltd., Eastleigh, United Kingdom) followed by incubation with FITC-conjugated goat anti-mouse antibody (Jackson ImmunoResearch). The anti-idiotypic mAb MK2-23, used as an isotype-negative control, was developed as described (19). After staining and washing, cells were analyzed for antigen expression by a FACSCalibur cytofluorimeter (Becton Dickinson, Franklin Lakes, NJ). Percentage positive cells and mean fluorescence intensity (MFI), after background subtraction, were recorded.
Immunohistochemical staining. Immunohistochemical staining of formalin-fixed and paraffin-embedded tissue sections with mAbs was done as described previously (23). Sections of melanoma lesions were analyzed as described (23) by acquiring digital images of 10 areas of each tissue section (at x400 magnification) on a Zeiss Axiovert 100 microscope (Carl Zeiss, Oberkochen, Germany). Antigen expression was scored as positive, heterogeneous, weak, and negative when >50% cells were stained with strong intensity, 10% to 50% cells were stained with strong intensity, >10% cells were stained with weak intensity, and <10% cells were stained with weak intensity, respectively.
Data analysis and statistics. Spearman correlation analysis was used to correlate immunohistochemical data with flow cytometry analysis of melanoma cells and to correlate expression of APM components and HLA class I and II antigens in short-term melanoma cell lines. Expression of markers, in terms of percentage positive cells and of MFI, was compared in primary tumors versus metastases by Mann-Whitney test. Correspondence analysis (28) and hierarchical clustering of APM and HLA phenotype were carried out as described previously (27, 29) using the J-Express Pro 2.6 software (http://www.Molmine.com). MFI values were used for hierarchical clustering of APM and HLA profiles. The flow cytometry data were subjected to normalization before carrying out hierarchical clustering because each marker had a distinct range of fluorescence intensity values in the panel of cell lines. To this end, the ratio of the MFI of each marker to the median MFI of the same marker in the panel of 85 cell lines was evaluated. Normalized expression values were then clustered by both tumor and antigen. Survival analysis was done by the Kaplan-Meier method, and survival curves were compared by the log-rank test. The Cox proportional hazards model and the Wald statistic, by Biometric Research Branch Array Tools version 3.3.0, developed by Dr. Richard Simon and Dr. Amy Peng Lam (Biometric Research Branch, Division of Cancer Treatment and Diagnosis, National Cancer Institute, Bethesda, MD) were used to test for association of single antigens with patients' survival. Permutation Ps for significant antigens were computed based on 10,000 random permutations.
|
Results and Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
|
|
|
The results of the correspondence analysis showed similar expression patterns of distinct markers as indicated by clustering of Delta, calnexin, TAP subunits, HLA class I antigens, and ß2M in the biplot (Fig. 2A). To evaluate the statistical significance of the similarities in the patterns of expression of several markers, we then carried out a Spearman correlation analysis of all 136 possible correlations generated by evaluating all combinations of the 17 markers analyzed, two at a time. MFI values were used for this purpose, as in several instances this variable was more discriminative of the phenotypes of the tumors than percentage stained cells. Sixty-seven of the 136 possible correlations were significant (Supplementary Fig. S2). Interestingly, 61 of the significant correlations had a positive coefficient (r) and involved proteasome subunits, TAP subunits, peptide-loading complex chaperones, HLA class I antigens, and ß2M. This finding suggests that several components of the HLA class I APM are coordinately expressed in melanoma cells. This conclusion is corroborated by the lack of significant correlation between level of HLA class I APM components and that of HLA class II antigens, with the exception of a positive and significant correlation between level of HLA class I and II (Supplementary Fig. S2).
Taken together, these results indicate that the APM and HLA class I and II phenotype of a large panel of short-term melanoma cell lines, as evaluated by quantitative assays, is characterized by marked heterogeneous expression of most of the markers analyzed and by coordinated expression of several of the HLA class I APM components.
Hierarchical clustering of the HLA and APM profiles of 85 tumors identifies two clusters of patients characterized by significantly different overall survival. As a first approach to evaluate the potential association of APM component and HLA phenotype of the tumors analyzed with clinical variables, we carried out a hierarchical clustering as described (27, 29). In this analysis, MFI values were used after normalization to the median value of each marker in the whole panel of tumors (see Materials and Methods). The resulting clusters were then compared for association with relevant clinical variables, as clinical data and long-term survival information were available for 84 of the 85 patients investigated. This analysis (Fig. 3A ) identified two main clusters of tumors, (cluster 1, 37 lesions; cluster 2, 48 lesions) and two main sets of antigens (antigen set 1 and antigen set 2; Fig. 3A). No significant differences were found between the two clusters of tumors for all the relevant clinical variables, including (a) stage of progression (primary lesion versus metastasis; P = 0.335), (b) clinical stage of the disease at time tumors were surgically removed (AJCC stage IIIb versus stage IIIc versus stage IV; P = 0.192), and (c) type of therapy administered after surgery (P = 0.364). In contrast, Kaplan-Meier survival analysis and log-rank test showed that patients, whose tumors fell in cluster 1, had a significantly better overall survival, measured from the time of surgical removal of the tumor to death, than patients in cluster 2 (Fig. 3B; P = 0.0063). This significant difference persisted even when survival was measured as time from the initial diagnosis of primary melanoma to death (P = 0.0147; data not shown). The potential association with patients' survival of any of the APM components and HLA antigens was then evaluated by the Cox proportional hazards model. ß2M (permutation P = 0.0175; hazard ratio = 0.467) and HLA class II antigens (permutation P = 0.0164; hazard ratio = 0.699) resulted to be significantly associated with patients' survival. To corroborate the validity of these results, the association of HLA class II molecule expression with patients' survival was further investigated in two sets of experiments. First, HLA class II antigen expression was evaluated by immunostaining with mAb LGII-612.14 in the tumors used to isolate the short-term melanoma cell lines. The 24 lesions available for this purpose were from 12 patients in cluster 1 and from 12 in cluster 2. Most of the lesions from the latter group of patients were either not stained or weakly stained with mAb LGII-612.14 (Supplementary Fig. S3). In contrast, all the tumors from patients in cluster 1 were stained with the mAb LGII-612.14 with a homogeneous or a heterogeneous pattern in agreement with HLA class II expression in the corresponding cell lines. The only exception was the tumor from patient 66 (Supplementary Fig. S3). The short-term cell lines corresponding to these two subsets of lesions differed significantly in their HLA class II antigen expression in terms of both percentage stained cells (mean 76.8% for the 12 cell lines from cluster 1 versus 15.5% for the 12 cell lines from cluster 2; P < 0.0001) and MFI (mean 146.1 for the 12 cell lines from cluster 1 versus 7.3 for the 12 cell lines from cluster 2; P < 0.0001).
Furthermore, the association of HLA class II antigen expression in melanoma cells with patients' survival was evaluated using metastatic lesions from an unrelated group of 52 AJCC stage III or IV melanoma patients (Supplementary Table S2). To this end, tissue sections were stained by immunohistochemistry with mAb LGII-612.14. Based on HLA class II antigen expression, lesions where classified into four subsets (see legend to Supplementary Table S2). Kaplan-Meier survival analysis and log-rank test showed that patients, whose tumors where classified as "positive" or "heterogeneous," had a significantly longer overall survival, measured as time from surgical removal of the metastatic lesion to death, than patients whose tumors were classified as either "weak" or "negative" (Supplementary Fig. S4; P = 0.0115). No significant differences were found between these two groups of patients in terms of clinical stage (i.e., AJCC stage IIIb, IIIc, or IV) at time when tumors were removed (P = 0.471) and type of therapy administered after surgery (P = 0.376).
These results suggest that quantitative variations in the expression level of APM components and HLA class I and II antigens in melanoma cells from AJCC stage III and IV patients may have an effect on the clinical course of the disease. The mechanism underlining this association is likely to be immunologic because patients' survival is associated with the expression in neoplastic lesions of HLA class II antigens and ß2M, which play a key role in the generation of HLA class II antigen-peptide and HLA class I antigen-peptide complexes recognized by CD4+ and CD8+ T cells, respectively. One possibility is that the T-cell–mediated antitumor response may be more effective (at either the priming and/or effector phases) in metastatic lesions with retained or higher levels of expression of HLA class II antigens and ß2M. In fact, defective generation of HLA class II-peptide and HLA class I-peptide complexes might impair the initial activation of antitumor immunity if such phase relies on direct T-cell priming by tumor cells instead of being mediated by dendritic cell–dependent cross-priming (30). Similarly, expansion at tumor site of previously primed T cells and effective recognition of neoplastic cells by effector T cells may be affected by the extent of expression of HLA-peptide complexes by neoplastic cells. Interestingly, by looking at frequency and differentiation phenotype of melanoma-specific T cells from tumor-invaded lymph nodes, we found the highest frequency of tumor antigen-specific T cells (and with a differentiated phenotype) in the lesions that retained expression of the tumor antigens and of the appropriate HLA allele acting as restricting element for such T cells (23). Moreover, recent results have indicated that presence of melanoma-specific T cells at tumor site is a significant prognostic factor predicting improved survival in advanced-stage melanoma (31).
The significant association of HLA class II antigen expression with longer survival, found in this study, agrees with the results obtained in large B-cell lymphomas and colorectal carcinomas (16, 17) but not with previous results obtained by immunohistochemistry in metastatic melanoma patients, where a low expression of HLA class II on neoplastic cells was associated with longer survival (8). However, in the latter study, survival was evaluated by comparing groups of patients defined by differential expression in neoplastic cells of both HLA class I and HLA class II antigens and not only of HLA class II antigens. This may contribute to explain the discrepancy with the current study.
A few of the HLA and APM markers that we evaluated in the present study are progressively down-regulated during tumor progression in melanoma and other tumors (9, 10). In addition, mutations of ß2M gene, resulting in loss of functional protein, have been described in human metastatic melanomas (32). Moreover, expression of HLA class I and TAP proteins has been shown previously to correlate with survival in carcinomas of the head and neck and in primary melanomas (11, 13). Taken together, these results suggest that the HLA and APM component profiles may be relevant to the clinical outcome in different human tumors. In addition, we found that LMP2 and LMP7 immunoproteasome subunits and ERp57 were the markers most frequently not detectable or expressed at very low levels among 85 tumors. In agreement with these findings, reduced expression of LMP proteasomal subunits has been reported in melanoma and other tumors (10, 12, 33). In contrast, most tumors investigated in this study showed frequent, although heterogeneous expression of constitutive proteasomal component Delta, of TAP subunits, and of different components of the peptide-loading complex, such as the chaperones, calnexin, calreticulin, and tapasin. Interestingly, several of these molecules involved in the HLA class I antigen-processing and presentation pathway showed significant correlations in the extent of their expression in melanoma cells. This suggests that HLA class I and APM component expression is, to some extent, coordinated in melanoma cells as proposed also by Giorda et al. (34). The retention of several APM components even in late-stage metastases may be explained even on the basis of the functions that these molecules exert, in addition to antigen-processing and MHC peptide-loading complex. In fact, these molecules do also several relevant functions needed for cell survival, cell cycle progression, and inhibition of apoptosis, such as the quality control of newly synthesized proteins in the ER (35) and the degradation by the proteasome of proteins tagged by the ubiquitin system (36). Thus, two opposing selection forces may contribute to shape the HLA and APM phenotype of neoplastic cells in advanced tumors. On one hand, neoplastic cells may attempt to maintain protein degradation and ER function pathways relevant for cell survival. On the other hand, the function of these pathways will favor the production of MHC-peptide complexes recognized by T cells, thus exposing the tumors to negative selection by the immune system. Interestingly, among the 85 cell lines used in this study, 39 were HLA-A2+, allowing to evaluate these tumors for recognition by HLA-A2-restricted CTLs directed to two different melanoma antigens (Melan-A/Mart-1 and gp100) and to correlate the results with antigen expression and with the HLA and APM phenotype of the neoplastic cells. In this subset of tumors, we found reduced or absent melanoma lysis by Melan-A/Mart-1-specific and gp100-specific CTLs in 13 of 39 and in 17 of 39 melanomas, respectively. Such impaired CTL-mediated recognition was explained by loss/defective expression of the melanoma antigens Melan-A/Mart-1 and gp100 in the neoplastic cells, although some significant correlations with defective APM component expression were also found (data not shown). These results support a model in which "immune editing" of the tumor phenotype (37) during tumor progression may promote selection of those neoplastic cells, whose genes coding for TAAs (not essential for cell survival) were silenced in preference to tumor cells with widespread defects in most of their APM components.
|
Acknowledgments |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
Footnotes |
|---|
A. Anichini and R. Mortarini contributed equally to this work.
Received 3/ 7/06. Revised 4/13/06. Accepted 4/19/06.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResults and DiscussionReferences |
|---|
in situ. Cancer Sci 2006;97:57–63.[CrossRef][Medline]c cytokines. Blood 2006;107:602–9.-chain cytokines. J Immunol 2003;171:2134–41.This article has been cited by other articles:
|
|
 |
|
  M. Lotem, A. Machlenkin, T. Hamburger, A. Nissan, L. Kadouri, S. Frankenburg, Z. Gimmon, O. Elias, I. B. David, A. Kuznetz, et al. Autologous Melanoma Vaccine Induces Antitumor and Self-Reactive Immune Responses That Affect Patient Survival and Depend on MHC Class II Expression on Vaccine Cells Clin. Cancer Res., August 1, 2009; 15(15): 4968 - 4977. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Di Nicola, R. Zappasodi, C. Carlo-Stella, R. Mortarini, S. M. Pupa, M. Magni, L. Devizzi, P. Matteucci, P. Baldassari, F. Ravagnani, et al. Vaccination with autologous tumor-loaded dendritic cells induces clinical and immunologic responses in indolent B-cell lymphoma patients with relapsed and measurable disease: a pilot study Blood, January 1, 2009; 113(1): 18 - 27. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Raffaghello, P. Nozza, F. Morandi, M. Camoriano, X. Wang, M. L. Garre, A. Cama, G. Basso, S. Ferrone, C. Gambini, et al. Expression and Functional Analysis of Human Leukocyte Antigen Class I Antigen-Processing Machinery in Medulloblastoma Cancer Res., June 1, 2007; 67(11): 5471 - 5478. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
, tumors; 
, antigens. Density plot, dotted lines, based on their HLA and APM profiles, the 85 tumors defined six subsets (1-6). Two different anti-LMP10 and two different anti-Delta mAbs (1 and 2) were used. B, HLA and APM profiles of the 85 tumors for each of the markers recognized by the panel of 17 mAbs. P, primary VGP tumor; M, metastasis. Tumors were grouped according to the six subsets identified by correspondence analysis in (A). Gray, results expressed as % positive cells for each marker. 
, tumors belonging to cluster 1 of 
, stage IIIb (light gray); stage IIIc (dark gray); stage IV (black). 
, IFN-
(green); chemotherapy (orange); IFN-