In this study we tested the hypothesis that loss of T cell signaling molecules in metastatic melanoma patients’ T cells may affect differently T cell subsets characterized by distinct TCR variable regions. By a two-color immunofluorescence technique, expression of ζ-chain, lck, and ZAP-70 was evaluated in CD3+ T cells and in three representative T cell subsets expressing TCRAV2, TCRBV2, or TCRBV18. Partial loss of lck and ZAP-70 was found in CD3+ T cells from PBL of most melanoma patients, but not of healthy donors. The extent of ζ-chain, lck, and ZAP-70 loss depended on the TCRV region expressed by the T cells, and this association was maintained or increased during progression of disease. Coculture of patients’ or donors’ T cell with melanoma cells, or with their supernatants, but not with normal fibroblasts or their supernatants, down-modulated expression of ζ-chain, lck, and ZAP-70 in a TCRV region-dependent way. Immunodepletion of soluble HLA class I molecules present in tumor supernatants, but not of soluble ICAM-1, blocked the suppressive effect on T cell signaling molecule expression. T cell activation with mAbs to a single TCRV region and to CD28 led to significant and TCRV region-specific re-induction of ζ-chain expression. These findings indicate that extent of TCR signaling molecules loss in T lymphocytes from metastatic melanoma patients depends on the TCRV region and suggest that tumor-derived HLA class I molecules may contribute to induce such alterations.

Failure to activate an effective antitumor response in cancer patients may result from impaired T cell function, as shown by absence of proliferative response and cytotoxic activity in tumor infiltrating lymphocytes (TIL) (1, 2, 3, 4). Decreased or lost expression of signal transducing molecules is a possible mechanism underlying functional defects in patients’ T cells. Altered expression of ζ-chain has been reported in TIL from renal cell carcinoma (5, 6), in TIL and PBL from colorectal carcinoma (7, 8, 9), and in PBL from patients with advanced cancer of different histological origin (renal cell carcinoma, hepatic colorectal carcinoma metastases, head and neck squamous cell carcinoma, acute lymphoblastic leukemia, and cervical cancer) (10, 11, 12). In melanoma patients, an association between the reduced overall survival and a low level of ζ-chain expression in T cells has been described (13). Concomitant down-modulation of ζ-chain molecule and tyrosine kinases, lck and/or ZAP-70, in PBL or T cell isolated from tumor site was observed in patients with renal cell carcinoma, metastatic melanoma, breast cancer, and in tumor bearing mice (5, 14, 15, 16, 17, 18). In addition, in T lymphocytes from some tumor patients, even when the expression of TCR and CD4/CD8 signal transducing elements was normal, a defective phosphorylation pattern of these molecules was nevertheless found, explaining the altered functionality (19, 20).

Several studies have investigated the mechanism responsible for the down-modulation of these signal transducing molecules. The evidence that in vitro culture of ζ-chain T lymphocytes infiltrating colorectal carcinoma restored the expression of the protein suggested a tumor-derived suppressive effect (21). This hypothesis was supported by experiments showing that coculture of T cells from healthy donors’ PBL with lung tumor cells induced a defective expression of ζ-chain, lck, and ZAP-70, subsequently restored by culture in the presence of IL-2-transduced cells (22). At least two tumor-released suppressive factors have been recently identified from human renal cell carcinoma (23). In PBL from healthy donors these factors induced down-modulation of Janus kinase-3, a tyrosine kinase involved in IL-2 signaling pathway, and to a lesser extent of the tyrosine kinases lck and fyn (23). Defects in signaling molecules can be induced not only by tumor cells, but even by other cells. For example, in tumor bearing mice activated macrophages, or NO derived from these cells, can induce altered expression of CD3 ζ-chain expression (24, 25).

Despite these results, it is still not known whether the mechanism that triggers the loss of signaling molecules (due to the direct T cell tumor interaction or to tumor-derived factors) has similar activity on any T cell or whether some degree of specificity exists, leading to selective impairment of T cell function in distinct T cell subsets. Because T cell specificity is uniquely defined by the molecular structure of the TCR, in this study we looked for evidence of a selective mechanism of induction of signaling defects depending on the TCRV region expressed by the T cells in melanoma patients. The results indicated that in both freshly isolate PBL, as well as in T cells cocultured with melanoma cells or with their supernatants, a striking degree of selectivity in TCR signaling molecule loss could be found or induced depending on the TCRV region expressed. The possible mechanism underlying this process was also investigated.

The characteristics of melanoma patients are summarized in Table I. No patient had received immunosuppressive treatment during 1 mo before PBL isolation for this study. Melanoma cells were isolated from surgical specimens and kept in culture in RPMI 1640 (BioWhittaker, Verviers, Belgium) supplemented with 2 mM glutamine (BioWhittaker), 20 mM HEPES buffer (BioWhittaker), 200 U/ml penicillin (Farmitalia Carlo Erba, Milan, Italy), and 40 μg/ml gentalyn (Schering Plough, Milan, Italy) plus 10% FCS (Biologic Industries, Kibbutz Beit Haemek, Israel). Tumor cells to be used for coculture with T lymphocytes were instead seeded for three days in RPMI 1640 supplemented with 10% pooled human serum. All melanoma cell lines used were analyzed by ELISA test for the absence of Mycoplasma contamination (Mycoplasma detection kit, Boehringer Mannheim, Milan, Italy). Melanoma lines were characterized for expression of HLA Ags by cytofluorometric analysis with mAbs to a monomorphic determinant of all HLA class-I alleles (w6/32, American Type Culture Collection (ATCC), Manassas, VA), to HLA-B,C alleles (4E) (26), to HLA-A2,A68 (CR11.351) (26). and to HLA-A2,A69 (BB7.2, ATCC).

Table I.

Characteristics of melanoma patients

Patient No.Sex/AgeStageCurrent StatusaSurvivalb
F/26 IV DOD 10 
F/71 IV DOD 
F/86 IV DOD 
F/38 IV DOD 
F/41 IV DOD 
M/51 IV DOD 
F/37 IV DOD 
F/46 IIIB AWD  
F/45 IV DOD 16 
10 M/46 IV DOD 
11 F/64 IV DOD 
12 M/35 IV DOD 13 
13 F/65 IV DOD 24 
14 M/79 IV AWD  
15 M/57 IIIA NOD 36 
16 F/46 IIIA AWD  
17 F/51 IV DOD 
18 M/30 IV AWD  
19 F/74 IV AWD  
20 F/61 IIIB DOD 19 
21 M/58 IV DOD 
22 F/56 IIIB DOD 
23 F/59 IV DOD 
24 F/32 IV DOD 
25 F/73 IV AWD  
26 M/67 IV DOD 
27 M/32 IV AWD  
28 F/51 IV AWD  
29 F/50 IIIB NOD 48 
Patient No.Sex/AgeStageCurrent StatusaSurvivalb
F/26 IV DOD 10 
F/71 IV DOD 
F/86 IV DOD 
F/38 IV DOD 
F/41 IV DOD 
M/51 IV DOD 
F/37 IV DOD 
F/46 IIIB AWD  
F/45 IV DOD 16 
10 M/46 IV DOD 
11 F/64 IV DOD 
12 M/35 IV DOD 13 
13 F/65 IV DOD 24 
14 M/79 IV AWD  
15 M/57 IIIA NOD 36 
16 F/46 IIIA AWD  
17 F/51 IV DOD 
18 M/30 IV AWD  
19 F/74 IV AWD  
20 F/61 IIIB DOD 19 
21 M/58 IV DOD 
22 F/56 IIIB DOD 
23 F/59 IV DOD 
24 F/32 IV DOD 
25 F/73 IV AWD  
26 M/67 IV DOD 
27 M/32 IV AWD  
28 F/51 IV AWD  
29 F/50 IIIB NOD 48 
a

DOD, dead of disease; AWD, alive with disease; NOD, no evidence of disease.

b

Survival (in months) after the blood sample used in this study was obtained.

The following mAbs, directed to TCR A or B variable regions, were used: TCRAV2S1 (Serotec, Oxford, U.K.), and TCRBV2S1 and TCRBV18S1 (Immunotech, Marseille, France). In addition, mAb to CD28 (L293, CAMfolio, Becton Dickinson, Sunnyvale, CA), to CD3 (OKT3), to HLA class I (W6/32) (ATCC), and to ζ-chain (TCR-Zeta clone, Coulter, Hialeah, FL) were used. Polyclonal sera were used to detect Zap-70 kinase and lck (Transduction Laboratories, Lexington, KY). Goat anti-mouse IgG and IgM antisera (GAM; Sigma, St. Louis, MO) and affinity isolated goat anti-mouse F(ab′)2 Ig fluorescein conjugate (BioSource International, Camarillo, CA) were also used.

T lymphocytes were isolated from peripheral blood (PBL) or from lymph node metastases (tumor-infiltrated lymph nodes, TAL)3 of melanoma patients admitted for surgery to the Istituto Nazionale Tumori (Milan, Italy) or from PBL of healthy donors. T lymphocytes, either freshly isolated or after coculture with melanoma cells (or with melanoma-derived supernatants), were analyzed for the expression of TCRV regions and signal transduction molecules (ζ-chain, Zap-70 kinase, and lck) by simultaneous membrane and intracytoplasmatic immunofluorescence technique. In the first step, T lymphocytes were stained with anti-CD3 or anti-TCRV region mAb followed by incubation with goat anti-mouse Ig fluorescein conjugate. T lymphocytes were then fixed in phosphate-buffered paraformaldehyde (4% PFA) (Sigma-Aldrich, Milan, Italy). In the second step cell membrane of the T cells was permeabilized by the detergent saponin (0.1%) (saponin from Gypsophyla, Sigma-Aldrich) as described by Sander et al. (27). Abs directed to T lymphocyte signal transduction molecules (ζ-chain, lck, or ZAP-70), conjugated with biotin by Biotin Labeling Kit (Boehringer Mannheim), were then added. After incubation with the transduction molecule-specific Ab, cells were washed in the presence of saponin and incubated with PE-coupled streptavidin (PharMingen, San Diego, CA) diluted in saponin solution. After one wash in HBSS (BioWhittaker), the samples were fixed by formaldehyde (1%) and the analysis was performed by a FACScan cytofluorometer (Becton Dickinson). Control of effective lymphocyte permeabilization was conducted by staining permeabilized cells with a mAb to vimentin (Boehringer Mannheim). For analysis, a logical gate was imposed on forward scatter vs green fluorescence plots to identify either CD3+ or TCRAV+/TCRBV+ T lymphocytes (depending on the mAb used in the first step of the staining procedure). After gating, analysis for expression of signaling molecules (detected by the Ab bound to PE-streptavidin used in the second step) was then performed on green vs red fluorescence dot plots. Thresholds to define negative and positive cells in each dot plots were defined by analyzing permeabilized cells stained with negative and positive controls. For T cell signaling molecule expression, negative controls consisted of permeabilized cells stained only with PE-coupled streptavidin and positive controls consisted of permeabilized cells stained with PE-conjugated anti-vimentin mAb. For CD3 and TCRV region expression, negative controls consisted of cells stained only with goat anti-mouse Ig fluorescein. The results were expressed as % cells lacking T cell signaling molecule among those positive for each TCRV region subset. Cell cycle analysis in patients’ T cells was conducted by two-color immunofluorescence on cells stained first with FITC-labeled anti-TCRBV2 mAb and then, after treatment with paraformaldehyde and ethanol, with propidium iodide (Sigma). A total of 100,000 events were acquired and analysis of cell cycle on TCRBV2+ cells was performed with the aid of the Modfit software (Becton Dickinson).

Melanoma cell lines or normal fibroblasts (2 × 105 cells), after culture for 72 h in a 75-cm2 flask (Corning, Corning, NY) in RPMI 1640 supplemented with 10% pooled human serum at 37°C and 5% CO2, were seeded (5 × 104 cells/wells) in 24-well plates in medium plus 10% pooled human serum. Twenty-four h later T lymphocytes from melanoma patients or from healthy donors were added to tumor cells or to fibroblasts (tumor or fibroblast:lymphocyte ratio was 1:5). To determine the effect of tumor-derived soluble products, T lymphocytes (2.5 × 105 cells) were cultured in the presence of complete medium with or without tumor- or fibroblast-derived supernatant. Tumor- and fibroblast-derived supernatants were produced by seeding 2 × 105 melanoma cells or fibroblasts for 72 h in a 75-cm2 flask in the presence of 10 ml of RPMI 1640 without any serum. In these conditions, recovery of melanoma cells and of fibroblasts after 72 h was similar. Supernatant was then harvested, centrifuged at 3000 × g for 30 min to remove all cells and debris, and then added to lymphocytes at the final dilution of 1:1 with complete medium. After 96 h of incubation, T cells were harvested and the phenotypic analysis for the expression of ζ-chain was performed.

Culture flasks (75 cm2, Corning) were coated with w6/32 mAb (10 μg/ml) for 2 h at 37°C followed by three washes with cold PBS. Tumor- or normal fibroblast-derived supernatants (10 ml) were then incubated at room temperature in flasks precoated with mAb w6/32. After 2 h of incubation the supernatant was transferred to a new w6/32-coated flask and the procedure was repeated four times, each time with 2-h incubation. The same technique was used to deplete supernatants from soluble ICAM-1 molecules by using flasks precoated with mAb 84H10 (28). The depletion of HLA class I and ICAM-1 molecules from supernatants was verified by slot blot technique. Samples were directly blotted onto polyvinylidene difluoride (PVDF) membrane (Hybond-P, Amersham, Arlington Heights, IL) in a trans-blotter apparatus (Bio-Rad, Richmond, CA). The membrane was blocked overnight at 4°C with 5% dry milk in Tris-buffered saline/0.1% Tween 20 (TTBS) and then immunodetected for 2 h with mAb w6/32 or 84H10 diluted 1:1000 in 5% BSA/TTBS. After three washes with TTBS, the membrane was incubated for 1 h with HRP-conjugated anti-mouse mAb (Bio-Rad), diluted 1:5000 in 5% BSA/TTBS, followed by three additional washes with TTBS. Visualization of Ab localization was performed with the enhanced chemiluminescence (ECL) kits (Amersham).

T lymphocytes were stimulated by the “TCR + CD28” protocol as described previously (26). Briefly, 5 × 105 cells, after monocyte removal by 2-h plastic adherence, were seeded in 24-well plates previously coated with anti-TCRBV region mAb (0.2 μg/ml) plus anti-CD28 mAb (1 μg/ml) cross-linked through goat anti-mouse IgG or IgM antisera in RPMI 1640 plus 10% pooled human serum (2 ml/well). The plates were incubated for 96 h at 37°C in 5% CO2, then T lymphocytes were collected and the expression of ζ-chain was evaluated by indirect immunofluorescence. For proliferation assays, flat-bottom 96-well plates were precoated with cross-linked mAb to TCRV regions (0.2 μg/ml) and to CD28 (1 μg/ml). T lymphocytes (5 × 104) were then seeded. The plates were incubated for 144 h at 37°C in 5% CO2. During the last 18 h of incubation, T cells were pulsed with 1 μCi [3H]thymidine/well. (New England Nuclear, Boston, MA); samples were then collected, absorbed onto nitrocellulose paper, and washed using an harvester (Titertek, Flow Laboratories, Costa Mesa, CA). The nitrocellulose filters were dried and counted after liquid scintillation in a beta counter (1205 Betaplate, Wallac, Turku, Finland). The proliferation assay was performed in three replicate wells, and data were expressed as the mean cpm. Results were considered positive when the cpm of lymphocytes in presence of anti-TCRV region plus anti-CD28 mAbs were higher than the mean + 3 SD of the cpm in wells containing lymphocytes in medium alone.

TNF released by T lymphocytes stimulated with anti-TCRAV or TCRBV region mAbs plus anti-CD28 mAb was assayed against the TNF-susceptible fibrosarcoma WEHI 164 clone 13 (29). Supernatants (50 μl) were harvested after 144 h of culture and added to 3 × 104 cells of cell line WEHI 164 clone 13 in 0.15 ml. The extent of TNF-induced cell death of WEHI cells was estimated 24 h later by a colorimetric test (MTT assay) as described by Hansen et al. (30) and analyzed by a microplate spectrophotometer (Easy Reader 400 AT, SLT-Labinstrument, Salzburg, Austria). The test was performed in three replicates. The amount of TNF (U/ml) present in the supernatants was calculated by linear regression analysis of absorbance of the standard curve produced by serial dilutions of recombinant human TNF-α (Eurocetus, Amsterdam, The Netherlands).

PBL from melanoma patients were assayed for the expression of ζ-chain, lck, and ZAP-70 signal transducing molecules by immunofluorescence involving simultaneous staining for both surface and intracytoplasmatic molecules. Representative examples of a patient (Fig. 1, A–C) with normal expression of the three signaling molecules in all CD3+ T cells and of a patient (Fig. 1, D–F) showing an almost complete loss of expression of lck and Zap-70 (Fig. 1, E and F) are shown. This assay was applied to evaluate extent of ζ-chain, lck, and ZAP-70 expression in CD3+ T cells from PBL of 29 stage III and IV melanoma patients and from 14 healthy donors. As shown in Fig. 2, the proportion of ζ-chain cells within the CD3+ T fraction, from most of melanoma patients, was lower then 10%; and only in four cases the ζ-chain T cells were between 20 and 88%. A few cases of defective ζ-chain expression were found even in healthy donors. The most striking defective expression of T lymphocyte signal transduction elements was observed for lck and ZAP-70 kinases in PBL from 15 melanoma patients (Fig. 2). CD3+ T cells from 7 of 15 patients showed decreased expression of lck kinase with a range of 20–40% of negative cells; in 4 patients lck T cells were between 62 and 98%. In contrast, the range of lck cells within the CD3+ fraction was between 0 and 20% in healthy donors. Similar results were obtained when looking at ZAP-70 CD3+ T cells: in 10 of 15 patients >20% of the CD3+ T cells were ZAP-70, whereas all but one healthy donor showed ZAP-70 CD3+ T cells.

FIGURE 1.

Two-color immunofluorescence analysis for expression of ζ-chain, lck, and ZAP-70 in CD3+T cells. PBL from melanoma patients 1 (A–C) and 29 (D–F) were analyzed by two-color immunofluorescence on fixed and permeabilized cells for the expression of CD3 (green fluorescence on the x-axis) and of ζ-chain (A and D), lck (B and E), and ZAP-70 (C and F) (red fluorescence on the y-axis). During analysis a gate on the forward scatter vs green fluorescence plots was adopted to select CD3+ lymphocytes. Data are represented as dot plots with logarithmic scale of the fluorescence intensity on both axes. The quadrants in each dot plot define the thresholds used to evaluate positive and negative cells; these quadrants were established on the basis of T cell staining with negative an positive controls as explained in Materials and Methods. By these criteria the % CD3+ cells lacking any of the three T cell signaling molecules was calculated as follows: % negative cells = [cells in lower right quadrant/(cells in lower right quadrant + cells in upper right quadrant)] × 100.

FIGURE 1.

Two-color immunofluorescence analysis for expression of ζ-chain, lck, and ZAP-70 in CD3+T cells. PBL from melanoma patients 1 (A–C) and 29 (D–F) were analyzed by two-color immunofluorescence on fixed and permeabilized cells for the expression of CD3 (green fluorescence on the x-axis) and of ζ-chain (A and D), lck (B and E), and ZAP-70 (C and F) (red fluorescence on the y-axis). During analysis a gate on the forward scatter vs green fluorescence plots was adopted to select CD3+ lymphocytes. Data are represented as dot plots with logarithmic scale of the fluorescence intensity on both axes. The quadrants in each dot plot define the thresholds used to evaluate positive and negative cells; these quadrants were established on the basis of T cell staining with negative an positive controls as explained in Materials and Methods. By these criteria the % CD3+ cells lacking any of the three T cell signaling molecules was calculated as follows: % negative cells = [cells in lower right quadrant/(cells in lower right quadrant + cells in upper right quadrant)] × 100.

Close modal
FIGURE 2.

Expression of ζ-chain, lck, and ZAP-70 in CD3+T cells from PBL of melanoma patients and healthy donors. PBL of melanoma patients were assayed for expression of ζ-chain (n = 29), lck (n = 15), and ZAP-70 (n = 15). PBL from 14 healthy donors were also phenotyped. The analysis was performed as described in the legend to Fig. 1 by evaluating ζ-chain, lck, or ZAP-70 expression after gating on CD3+ cells. Data expressed as the proportion, in each patient or donor, of CD3+ T cells lacking ζ-chain, lck, or ZAP-70. The p values for the patients vs donors comparisons (Mann-Whitney U test) were 0.1362 (for ζ-chain expression), 0.0004 (for lck expression), and <0.001 (for ZAP-70 expression).

FIGURE 2.

Expression of ζ-chain, lck, and ZAP-70 in CD3+T cells from PBL of melanoma patients and healthy donors. PBL of melanoma patients were assayed for expression of ζ-chain (n = 29), lck (n = 15), and ZAP-70 (n = 15). PBL from 14 healthy donors were also phenotyped. The analysis was performed as described in the legend to Fig. 1 by evaluating ζ-chain, lck, or ZAP-70 expression after gating on CD3+ cells. Data expressed as the proportion, in each patient or donor, of CD3+ T cells lacking ζ-chain, lck, or ZAP-70. The p values for the patients vs donors comparisons (Mann-Whitney U test) were 0.1362 (for ζ-chain expression), 0.0004 (for lck expression), and <0.001 (for ZAP-70 expression).

Close modal

To test whether loss of T cell signal transducing molecules affected equally different T cell subsets, expression of ζ-chain, lck, and Zap-70 was evaluated within three randomly chosen T cell subsets identified for expression of distinct TCRV regions (TCRAV2+, TCRBV2+, and TCRBV18+ subsets). As shown in Fig. 3, in many instances, proportion of ζ-chain cells was markedly different in each of the three TCRV region subset in each patient. For example in patient 5 the proportion of T cells lacking ζ-chain was 7.6, 1.8, and 38.9% in TCRBV2+, TCRAV2+, and TCRBV18+ subsets, respectively. Similar evidence for marked differences in the extent of ζ-chain loss, depending on the TCRV region subset, was evident in many other patients such as patients 4, 12, 13, 15, 20, 21, 23–26, and 28 (Fig. 3). Interestingly, in selected cases, marked defects in ζ-chain expression could be evident within one or more of the three TCRV region subsets, but not when looking at the overall T cells subpopulation stained by anti-CD3 mAb. This was exemplified by patients 1 and 4 (Fig. 3). By contrast, in patients 7, the defective expression of ζ-chain observed on the whole CD3+ population was not seen in any of the three TCRV region subsets, suggesting that in this patient, ζ-chain T cells expressed TCRV regions different from TCRBV2, TCRAV2, and TCRBV18. Furthermore, proportion of ζ-chain cells and of TCRV+ region cells, in each subset, were not related. For example, patients 2 and 5 had a similar proportion of TCRBV18+ T cells (7% of all CD3+ T cells in each patient), but ζ-chain cells (as proportion of all TCRBV18+ cells) were 15.5% in patient 2 and 39% in patient 5 (Fig. 3). However, patients (such as patient 28 in Fig. 3) with a high level of ζ-chain loss in the CD3+ fraction showed a high proportion of ζ-chain loss in all TCRV regions analyzed (Fig. 3), as well as in other TCRBV regions (% ζ cells in patient 28 was between 65 and 90% in TCRBV1+, TCRBV3+, TCRBV5.1+, TCRBV6.1+, TCRBV8.1+, and TCRBV12.1+ subsets; data not shown).

FIGURE 3.

ζ-chain expression in TCRAV2+, TCRBV2+, and TCRBV18+T cells from melanoma patients. PBL from 29 melanoma patients were analyzed for the expression of ζ-chain within the CD3+ T cell fraction, or within the TCRAV2+, TCRBV2+, or TCRBV18+ subsets. Data expressed as percentage of cells lacking ζ-chain within each subset. ∗, Proportion of CD3+, TCRAV2+, TCRBV2+, and TCRBV18+ T lymphocytes in PBL of each patient. The p value for the differential losses of T cell signaling molecules was <0.0001 (two-way ANOVA on the whole data set).

FIGURE 3.

ζ-chain expression in TCRAV2+, TCRBV2+, and TCRBV18+T cells from melanoma patients. PBL from 29 melanoma patients were analyzed for the expression of ζ-chain within the CD3+ T cell fraction, or within the TCRAV2+, TCRBV2+, or TCRBV18+ subsets. Data expressed as percentage of cells lacking ζ-chain within each subset. ∗, Proportion of CD3+, TCRAV2+, TCRBV2+, and TCRBV18+ T lymphocytes in PBL of each patient. The p value for the differential losses of T cell signaling molecules was <0.0001 (two-way ANOVA on the whole data set).

Close modal

The analysis for ζ-chain expression in distinct TCRV region T cell subsets was performed even in two healthy donors showing, respectively, ∼60% and ∼20% ζ-chain negative cells within the CD3+ fraction. In both donors evidence for TCRV region selective ζ-chain loss was obtained. In the first donor ζ-chain negative cells represented 12.9, 78.5, and 61.6% of TCRAV2+, TCRBV2+, and TCRBV18+ cells, respectively. In the second donor, ζ-chain negative cells represented 11.5, 9.1, and 1.7% of TCRAV2+, TCRBV2+, and TCRBV18+ cells, respectively (data not shown).

Analysis of lck (Fig. 4) and ZAP-70 (Fig. 5) expression in TCRBV2+, TCRAV2+, and TCRBV18+ T cells in melanoma patients’ PBL provided further evidence of marked differences in expression of these signal transducing molecules depending on the TCRV region. In many patients, the proportion of T cells lacking lck or ZAP-70, within each TCRV region subset, was much higher than the corresponding defect, in the same subset, for ζ-chain expression (compare, for example, patients 2, 7, and 29 in Figs. 3–5). Taken together, these data indicate that in peripheral blood of metastatic melanoma patients the extent of signal transducing molecule loss shows a significant degree of selectivity depending on the TCRV region expressed by the T cells.

FIGURE 4.

lck kinase expression in TCRAV2+, TCRBV2+, and TCRBV18+T cells in PBL from melanoma patients. PBL from 15 melanoma patients were analyzed for the expression of lck within the CD3+ T cell fraction, or within the TCRAV2+, TCRBV2+, or TCRBV18+subsets. Data expressed as percentage of cells lacking lck in each subset. ∗, Proportion of CD3+, TCRAV2+, TCRBV2+, and TCRBV18+ T lymphocytes in PBL of each patient. N.D., not done. The p value for the differential losses of T cell signaling molecules was <0.0001 (two-way ANOVA on the whole data set).

FIGURE 4.

lck kinase expression in TCRAV2+, TCRBV2+, and TCRBV18+T cells in PBL from melanoma patients. PBL from 15 melanoma patients were analyzed for the expression of lck within the CD3+ T cell fraction, or within the TCRAV2+, TCRBV2+, or TCRBV18+subsets. Data expressed as percentage of cells lacking lck in each subset. ∗, Proportion of CD3+, TCRAV2+, TCRBV2+, and TCRBV18+ T lymphocytes in PBL of each patient. N.D., not done. The p value for the differential losses of T cell signaling molecules was <0.0001 (two-way ANOVA on the whole data set).

Close modal
FIGURE 5.

ZAP-70 expression in TCRAV2+, TCRBV2+, and TCRBV18+ T cells in PBL from melanoma patients. PBL from 15 melanoma patients were analyzed for the expression of ZAP-70 within the CD3+ T cell fraction, or within the TCRAV2+, TCRBV2+, or TCRBV18+subsets. Data expressed as percentage of cells lacking ZAP-70 in each subset. ∗, Proportion of CD3+, TCRAV2+, TCRBV2+, and TCRBV18+ T lymphocytes in PBL of each patient. N.D., not done. The p value for the differential losses of T cell signaling molecules was <0.0001 (two-way ANOVA on the whole data set).

FIGURE 5.

ZAP-70 expression in TCRAV2+, TCRBV2+, and TCRBV18+ T cells in PBL from melanoma patients. PBL from 15 melanoma patients were analyzed for the expression of ZAP-70 within the CD3+ T cell fraction, or within the TCRAV2+, TCRBV2+, or TCRBV18+subsets. Data expressed as percentage of cells lacking ZAP-70 in each subset. ∗, Proportion of CD3+, TCRAV2+, TCRBV2+, and TCRBV18+ T lymphocytes in PBL of each patient. N.D., not done. The p value for the differential losses of T cell signaling molecules was <0.0001 (two-way ANOVA on the whole data set).

Close modal

Expression of ζ-chain, lck, and Zap-70 in PBL from two stage IV melanoma patients was evaluated in blood samples taken 45–60 days apart. During this time interval both patients underwent further clinical progression of disease. In patients 4 and 9, the extent of defect in ζ-chain, lck, and Zap-70 expression in CD3+ T cells augmented markedly from day 0 to day 45 and 60, respectively (Fig. 6). In the same two patients the proportion of cells lacking at least one of these three molecules increased in the TCRBV2+ subset (Fig. 6). Interestingly, patient 4 showed modest changes of ζ-chain, lck, and Zap-70 expression in the TCRAV2+ and TCRBV18+ subsets, whereas patient 9 underwent a marked increase in the proportion of ζ-chain, lck TCRAV2+ and ζ-chain TCRBV18+ T cells (Fig. 6). These data indicate the losses of T cell signaling molecules in peripheral blood are maintained or increased during clinical progression in metastatic melanoma. Furthermore TCRV region selectivity in T cell signaling molecule loss was often maintained during disease progression.

FIGURE 6.

Expression of ζ-chain, lck, and ZAP-70 in T cells of melanoma patients during the evolution of the disease. ζ-chain, lck, and ZAP-70 expression was evaluated in PBL samples isolated from patients 4 and 9 at two different time points (first sample, open bars; second sample, filled bars). Interval between the two samples was 60 days for patient 4 and 45 days for patient 9. Data expressed as proportion of cells lacking ζ-chain, lck, or ZAP-70 in the CD3+ fraction, or within the TCRAV2+, TCRBV2+, or TCRBV18+ subsets.

FIGURE 6.

Expression of ζ-chain, lck, and ZAP-70 in T cells of melanoma patients during the evolution of the disease. ζ-chain, lck, and ZAP-70 expression was evaluated in PBL samples isolated from patients 4 and 9 at two different time points (first sample, open bars; second sample, filled bars). Interval between the two samples was 60 days for patient 4 and 45 days for patient 9. Data expressed as proportion of cells lacking ζ-chain, lck, or ZAP-70 in the CD3+ fraction, or within the TCRAV2+, TCRBV2+, or TCRBV18+ subsets.

Close modal

To evaluate whether tumor cells could have a direct effect on ζ-chain expression, lymphocytes from three patients who did not display marked defects in ζ-chain expression were cultured for 4 days in the presence of autologous melanoma cells, allogeneic melanoma cells, or normal fibroblasts. As shown in Table II, in patient 6, after culture with autologous tumor, but not with autologous fibroblasts, the proportion of CD3+ TAL lacking ζ-chain increased from 18 to 90% and no effect was seen on the TCRAV2+ subset, whereas the ζ-chain cells in the TCRBV2+ subset increased from 2 to 47%. In patients 9 and 29, T cell culture with allogeneic or autologous fibroblasts had no effect on ζ-chain expression, whereas both autologous melanoma and allogeneic tumors induced significant increase in ζ-chain cells in CD3+ T cells. Effect of tumor cells on T cell ζ-chain expression depended on the tumor and on the TCRV-region subset: for example, in patient 9, T cell culture with Me 6/1 and Me 29 tumors affected significantly the extent of ζ-chain expression in the TCRAV2+ but not in the TCRBV2+ subset, whereas the autologous tumor down-modulated ζ-chain expression in both TCRAV2+ and TCRBV2+ T cells (Table II). These data indicate that tumor cells, but not normal cells, induce down-modulation of ζ-chain in T lymphocytes and that evidence of TCRV region selectivity can be observed in these conditions, in agreement with the results observed during clinical progression.

Table II.

Suppressive activity by melanoma cells on the expression of ζ-chain in CD3+ T cells and in TCRAV2+ or TCRBV2+ T cells

PatientLymphocytes fromLymphocytes Cocultured witha% of Cells Lacking ζ-Chain inb
CD3+TCRAV2+TCRBV2+
TAL − 16 
  Medium 18 
  Me 6/1 90 47 
  Patient 6 Fibroblasts 25 
PBL − 
  Medium 
  Me 9 35 58 31 
  Me 6/1 21 18 
  Me 29 25 31 
  Patient 29 Fibroblasts 12 
29 PBL − 
  Medium 
  Me 9 34 30 45 
  Me 6/1 18 18 14 
  Me 29 35 35 23 
  Patient 29 Fibroblasts 11 10 
PatientLymphocytes fromLymphocytes Cocultured witha% of Cells Lacking ζ-Chain inb
CD3+TCRAV2+TCRBV2+
TAL − 16 
  Medium 18 
  Me 6/1 90 47 
  Patient 6 Fibroblasts 25 
PBL − 
  Medium 
  Me 9 35 58 31 
  Me 6/1 21 18 
  Me 29 25 31 
  Patient 29 Fibroblasts 12 
29 PBL − 
  Medium 
  Me 9 34 30 45 
  Me 6/1 18 18 14 
  Me 29 35 35 23 
  Patient 29 Fibroblasts 11 10 
a

Lymphocytes from PBL or TAL of melanoma patients were analyzed before culture (−) and after culture for 96 h with either complete medium alone (Medium) or with melanoma cells or fibroblast at a lymphocyte: tumor or lymphocyte:fibroblast ratio of 5:1. After 96 h the expression of ζ-chain on T lymphocytes was assayed by cytofluorimetric analysis. Melanoma lines Me 6/1, Me 9, and Me 29 are autologous to patients 6, 9, and 29, respectively.

b

Proportion of T cells lacking ζ-chain in either CD3+, TCRAV2+, or TCRBV2+ subsets. Down-modulation of the expression of ζ-chain was considered significant (underlined values) when the percentage of negative cells was at least twice the value measured in T lymphocytes cultured with medium alone. TCRAV2+ T cells represented 5, 9, and 3% of all CD3+ T cells in patient 6, 9, and 29, respectively. TCRBV2+ T cells represented 10, 13, and 7% of all CD3+ T cells in patient 6, 9, and 29, respectively.

To verify whether down-modulation of ζ-chain in T lymphocytes from melanoma patients could be induced even by soluble products in tumor supernatants, PBL or TAL from melanoma patients, or PBL from healthy donors, were analyzed for ζ-chain expression after 96-h culture with medium supplemented with 50% tumor- or fibroblast-derived supernatants. The effect of tumor supernatants on ζ-chain expression was weaker than seen in coculture experiments with tumor cells, however the supernatant from tumor 6/1 increased the proportion of ζ-chain T cells within the TCRAV2+ subset in the autologous setting (from 4 to 14% ζ-chain cells) (Table III). Incubation of TAL from the patient 6 with the autologous melanoma (Me 6/1) supernatant markedly reduced the expression of ζ-chain on TCRAV2+ cells. Me 6/1 and Me 18656 supernatants affected ζ-chain expression in TCRAV2+ and TCRBV2+ subsets of patients 9 and 29. Furthermore, most tumor-derived supernatants affected ζ-chain expression in TCRAV2+ and TCRBV2+ T cells of two healthy donors (Table III). Interestingly, the same tumor-derived supernatant affected differently CD3+ T cells or TCRV region subsets depending on the origin of the lymphocyte population, as shown by comparing the effect of Me 6/1 and Me 18656 supernatants on TAL and PBL from patient 6 and on PBL from the two donors (Table III). Normal fibroblast-derived supernatants failed to induce decrease of ζ-chain expression in lymphocytes from both melanoma patients and healthy donors in both autologous and allogeneic lymphocyte/fibroblast combinations (Table III). Thus, soluble products released by melanoma cells, but not by normal cells, can down modulate T cell ζ-chain.

Table III.

Suppressive effect by melanoma supernatant on the expression of ζ-chain

SubjectLymphocytesSupernatant froma% of Cells Lacking ζ-Chain inb
CD3+TCRAV2+TCRBV2+
Patient 6 PBL − 13 
  Me 6/1 15 14 
  Me 18656 
  Patient 6 Fibroblasts 11 
      
Patient 6 TAL − 18 
  Me 6/1 22 34 11 
  Me 18656 12 16 
  Patient 6 Fibroblasts 
      
Patient 9 PBL − 
  Me 9 10 28 53 
  Me 6/1 15 46 71 
  Me 18656 10 56 49 
      
Patient 29 PBL − 
  Me 29 50 55 31 
  Me 6/1 37 28 27 
  Me 18656 43 30 38 
      
Donor 22 PBL − 15 
  Me 9 39 39 41 
  Me 6/1 71 49 36 
  Me 29 37 53 38 
  Me 18656 60 53 37 
  Patient 6 Fibroblasts 
      
Donor 40 PBL − 10 
  Me 9 30 
  Me 6/1 20 79 15 
  Me 29 16 26 34 
  Me 18656 13 50 17 
  Patient 6 Fibroblasts 
SubjectLymphocytesSupernatant froma% of Cells Lacking ζ-Chain inb
CD3+TCRAV2+TCRBV2+
Patient 6 PBL − 13 
  Me 6/1 15 14 
  Me 18656 
  Patient 6 Fibroblasts 11 
      
Patient 6 TAL − 18 
  Me 6/1 22 34 11 
  Me 18656 12 16 
  Patient 6 Fibroblasts 
      
Patient 9 PBL − 
  Me 9 10 28 53 
  Me 6/1 15 46 71 
  Me 18656 10 56 49 
      
Patient 29 PBL − 
  Me 29 50 55 31 
  Me 6/1 37 28 27 
  Me 18656 43 30 38 
      
Donor 22 PBL − 15 
  Me 9 39 39 41 
  Me 6/1 71 49 36 
  Me 29 37 53 38 
  Me 18656 60 53 37 
  Patient 6 Fibroblasts 
      
Donor 40 PBL − 10 
  Me 9 30 
  Me 6/1 20 79 15 
  Me 29 16 26 34 
  Me 18656 13 50 17 
  Patient 6 Fibroblasts 
a

Lymphocytes from PBL or TAL were cultured in complete medium with or without (−) melanoma- or fibroblast-derived supernatant at 1:1 dilution. After 96 h, the expression of ζ-chain on T lymphocytes was assayed by indirect immunofluorescence followed by cytofluorimetric analysis. Supernatant was produced from four melanoma lines (Me 9, Me 6/1, Me 29, and Me 18656) and from normal fibroblasts.

b

Proportion of T cells lacking ζ-chain in either CD3+, TCRAV2+, or TCRBV2+ subsets. Down-modulation of the expression of ζ-chain was considered significant (underlined values) when the percentage of negative cells was at least twice the value measured in T lymphocytes cultured with medium alone.

Phenotype of tumor lines used in supernatant experiments (Me 6, Me 9, Me 29 and Me 18656) by anti-HLA Abs indicated that overall HLA class I expression (as evaluated by w6/32 mAb) was similar, while a reduced expression of HLA-B,C alleles was detected in Me 29 cells (Fig. 7). Furthermore, one (Me 18656) of two tumors from HLA-A68+ patients had lost expression of HLA-A68 (Fig. 7). These differences did not correlate with the results obtained in supernatant experiments as the tumor with a reduced HLA-B,C expression (Me 29) or the HLA-A68-loss tumor (Me 18656) did not induce a greater or lower loss of ζ-chain in comparison to the other tumors. To directly assess the role of shed HLA class I Ags on induction of ζ-chain, supernatants were immunodepleted by serial rounds of exposure to flasks coated with w6/32 mAb (an mAb that recognizes membrane-bound as well as soluble HLA class I molecules associated with β2-microglobulin). As shown in Fig. 8, HLA class I molecules, present in Me 29 supernatant, but absent from two HLA-class I negative tumor lines (K562 and melanoma Me 9923) (Fig. 8,A), were gradually removed from Me 29 supernatant by four steps of depletion (Fig. 8,A, slots3a–3d). As control, tumor-derived supernatants were immunodepleted of soluble ICAM-1 molecules (Fig. 8,B, slots 3a–3c). Supernatants (depleted or not of HLA or ICAM-1) were then checked for the suppressive activity on ζ-chain expression. When T lymphocytes from patients 9 and 29 were cultured in the presence of HLA-depleted supernatants from Me 9 and Me 29, the suppressive activity on ζ-chain expression was profoundly inhibited or, in some instances, almost completely abrogated on either CD3+ T cells (patient 29) or the two TCRV region subsets (patients 9 and 29, Table IV). Tumor supernatants (from Me 9 and Me 29) depleted of soluble ICAM-1 maintained the suppressive effect on ζ-chain expression of all TCRV region subsets analyzed (Table IV). Depletion of HLA class I molecules, but not of ICAM-1, from tumor supernatant, inhibited the suppressive effect on both lck and ZAP-70 in either CD3+ or TCRAV2+ T cell subsets of a healthy donor (Table V). These data suggest that soluble HLA-class I molecules shed by melanoma cells play a role in TCRV region-related down-modulation of T cell signaling molecules.

FIGURE 7.

HLA class I phenotype of melanoma lines. Tumor cells were typed (empty histograms) by staining with anti-HLA class I (mAb w6/32), anti-HLA-B,C (mAb 4E), HLA-A2,A68 (mAb CR11.351), HLA-A2,A69 (mAb BB7.2), followed by FACS analysis. Filled histograms: control staining with FITC-labeled secondary anti-mouse IgG. Me 6 and Me 18656 were derived from two HLA-A68+ patients, Me 29 was from an HLA-A2+ patient, and Me 9 from an HLA-A1+ patient.

FIGURE 7.

HLA class I phenotype of melanoma lines. Tumor cells were typed (empty histograms) by staining with anti-HLA class I (mAb w6/32), anti-HLA-B,C (mAb 4E), HLA-A2,A68 (mAb CR11.351), HLA-A2,A69 (mAb BB7.2), followed by FACS analysis. Filled histograms: control staining with FITC-labeled secondary anti-mouse IgG. Me 6 and Me 18656 were derived from two HLA-A68+ patients, Me 29 was from an HLA-A2+ patient, and Me 9 from an HLA-A1+ patient.

Close modal
FIGURE 8.

Slot blot analysis of soluble HLA class I and soluble ICAM-1 molecules in tumor supernatants and in immunodepleted supernatants. A, supernatants from HLA class-I negative tumor lines Me 9923P (slot 1), K562 (slot 2), from two HLA class I-positive melanomas Me 29 (slot 3) and Me 9 (slot 4), from HLA-Class-I-positive fibroblasts (slot 5), and supernatants of Me 29 after four subsequent rounds of HLA class I immunodepletion (slots 3a–3d). All samples were directly blotted onto PVDF membrane and immunodetected with w6/32. B, Detection of soluble ICAM-1 with mAb 84H10 in supernatants of Me 9923 (slot 1), K562 (slot 2), Me 29 (slot 3), and Me 29 after three subsequent rounds of ICAM-1 immunodepletion (slots 3a–3c). All supernatants were produced by culturing 2 × 105 normal or neoplastic cells in RPMI 1640 without any serum for 72 h.

FIGURE 8.

Slot blot analysis of soluble HLA class I and soluble ICAM-1 molecules in tumor supernatants and in immunodepleted supernatants. A, supernatants from HLA class-I negative tumor lines Me 9923P (slot 1), K562 (slot 2), from two HLA class I-positive melanomas Me 29 (slot 3) and Me 9 (slot 4), from HLA-Class-I-positive fibroblasts (slot 5), and supernatants of Me 29 after four subsequent rounds of HLA class I immunodepletion (slots 3a–3d). All samples were directly blotted onto PVDF membrane and immunodetected with w6/32. B, Detection of soluble ICAM-1 with mAb 84H10 in supernatants of Me 9923 (slot 1), K562 (slot 2), Me 29 (slot 3), and Me 29 after three subsequent rounds of ICAM-1 immunodepletion (slots 3a–3c). All supernatants were produced by culturing 2 × 105 normal or neoplastic cells in RPMI 1640 without any serum for 72 h.

Close modal
Table IV.

Depletion of soluble HLA class I molecules, but not of soluble ICAM-1, from tumor-derived supernatants, inhibits the suppressive effect on the expression of ζ-chain in PBL from melanoma patients

PatientLymphocytesSupernatant fromaSupernatants Depleted ofb% of Cells Lacking ζ-chain inc
CD3+TCRAV2+TCRBV2+
PBL − − 
  Me 9 − 10 26 50 
  Me 9 HLA-class I 12 11 
  Me 9 ICAM-1 11 24 58 
  Me 29 − 49 57 64 
  Me 29 HLA-class I 21 21 17 
  Me 29 ICAM-1 38 52 60 
       
29 PBL − − 
  Me 9 − 31 34 35 
  Me 9 HLA-class I 15 10 
  Me 9 ICAM-1 28 30 32 
  Me 29 − 53 47 26 
  Me 29 HLA-class I 11 18 
  Me 29 ICAM-1 47 53 21 
PatientLymphocytesSupernatant fromaSupernatants Depleted ofb% of Cells Lacking ζ-chain inc
CD3+TCRAV2+TCRBV2+
PBL − − 
  Me 9 − 10 26 50 
  Me 9 HLA-class I 12 11 
  Me 9 ICAM-1 11 24 58 
  Me 29 − 49 57 64 
  Me 29 HLA-class I 21 21 17 
  Me 29 ICAM-1 38 52 60 
       
29 PBL − − 
  Me 9 − 31 34 35 
  Me 9 HLA-class I 15 10 
  Me 9 ICAM-1 28 30 32 
  Me 29 − 53 47 26 
  Me 29 HLA-class I 11 18 
  Me 29 ICAM-1 47 53 21 
a

Lymphocytes were cultured in complete medium with or without (−) melanoma-derived supernatant at 1:1 dilution. After 96 h, the expression of ζ-chain in CD3+, TCRAV2+, or TCRBV2+ T cells was assayed by cytofluorimetric analysis.

b

Supernatants were depleted or not (−) of either HLA-class I molecules or ICAM-1 molecules as described in the legend to Fig. 8.

c

Proportion of T cells lacking ζ-chain in either CD3+, TCRAV2+, or TCRBV2+ subsets. Down-modulation of the expression of ζ-chain was considered significant (underlined values) when the percentage of negative cells was at least twice the value measured in T lymphocytes cultured with medium alone.

Table V.

Down-modulation of lck and ZAP-70 kinases mediated by tumor-derived supernatant is inhibited by depletion of HLA class I molecules, but not by depletion of soluble ICAM-1

DonorLymphocyteSupernatant fromaSupernatant Depleted ofb% Cells Lackingc
lckZAP-70
CD3+TCRAV2+CD3+TCRAV2+
22 PBL − − 16 11 
  Me 29 − 32 22 19 14 
  Me 29 HLA-class I 18 14 
  Me 29 ICAM-1 32 24 15 13 
DonorLymphocyteSupernatant fromaSupernatant Depleted ofb% Cells Lackingc
lckZAP-70
CD3+TCRAV2+CD3+TCRAV2+
22 PBL − − 16 11 
  Me 29 − 32 22 19 14 
  Me 29 HLA-class I 18 14 
  Me 29 ICAM-1 32 24 15 13 
a

Lymphocytes were cultured in complete medium with or without (−) melanoma-derived supernatant at 1:1 dilution. After 96 h, the expression of lck or ZAP-70 in CD3+ or TCRAV2+ T cells was assayed by cytofluorimetric analysis.

b

Supernatants were depleted or not (−) of either HLA-class I molecules or ICAM-1 molecules as described in the legend to Fig. 8.

c

Proportion of T cells lacking lck or ZAP-70 in either CD3+ or TCRAV2+ T cells. Down-modulation of the expression of lck or ZAP-70 was considered significant (underlined values) when the percentage of negative cells was at least twice the value measured in T lymphocytes cultured with medium alone.

To verify the possibility to rescue expression of ζ-chain in defective T cells, PBL from metastatic melanoma patients showing a marked loss of this molecule were stimulated with a mAb recognizing a single TCRV region (either TCRAV2 or TCRBV2) and a mAb to CD28. This protocol (briefly defined as “TCRV + CD28”) has been previously shown to induce selective proliferation of T cells expressing a single TCRV region (29). After 4 days of culture, the expression of ζ-chain in T cells was assayed. As shown in Table VI, T cell activation by anti-TCRBV2 plus anti-CD28 mAb led in all three patients to a dramatic reduction in the proportion of ζ-chain cells within the TCRBV2+ subset in comparison to fresh T cells, or to T cells cultured in medium alone (patient 28). The effect of the anti-TCRBV2 mAb was subset-specific, as, in the same patients, no effect was seen on the proportion of ζ-chain cells within the TCRAV2+ or TCRBV18+ T cell subsets. Similarly, in one patient (no. 28) restoration of ζ-chain expression was induced in the TCRBV18 subset by activation with anti-TCRBV18 mAb (Table VI), without affecting ζ-chain expression in the TCRAV2+ or TCRBV2+ subset.

Table VI.

T cell activation by mAb to TCRV region plus mAb to CD28 restores ζ-chain expression

PatientLymphocyteAnalysis at Time (h)Activation with mAbs toa% ζ-Chain Negative Cells inb
TCRAV2+TCRBV2+TCRBV18+
28 PBL − 57 (3) 80 (7) 67 (1) 
  96 − 57 (4) 79 (4) 53 (2) 
  96 TCRBV2+ CD28 60 (4) 4 (13) 52 (1) 
  96 TCRBV18 + CD28 65 (2) 83 (1) 20 (19) 
23 PBL − 30 (3) 87 (7) 13 (3) 
  96 TCRBV2+ CD28 45 (5) 21 (30) 16 (4) 
24 PBL − 18 (3) 93 (6) 20 (2) 
  96 TCRBV2+ CD28 23 (2) 17 (11) 24 (3) 
PatientLymphocyteAnalysis at Time (h)Activation with mAbs toa% ζ-Chain Negative Cells inb
TCRAV2+TCRBV2+TCRBV18+
28 PBL − 57 (3) 80 (7) 67 (1) 
  96 − 57 (4) 79 (4) 53 (2) 
  96 TCRBV2+ CD28 60 (4) 4 (13) 52 (1) 
  96 TCRBV18 + CD28 65 (2) 83 (1) 20 (19) 
23 PBL − 30 (3) 87 (7) 13 (3) 
  96 TCRBV2+ CD28 45 (5) 21 (30) 16 (4) 
24 PBL − 18 (3) 93 (6) 20 (2) 
  96 TCRBV2+ CD28 23 (2) 17 (11) 24 (3) 
a

Lymphocytes were cultured without (−) or with the indicated mAbs to TCRBV regions plus mAb to CD28 cross-linked to plastic through goat anti-mouse antiserum.

b

Proportion of T cells lacking ζ-chain in TCRAV2+, TCRBV2+, or TCRBV18+ subsets. Underlined values represent significant reduction in the proportion of ζ-chain T cells in each subset. Numbers in parenthesis represent the proportion of cells positive for TCRAV2, TCRBV2, or TCRBV18, respectively. Fresh (analysis at time 0 h) or cultured (analysis at time 96 h) lymphocytes were characterized for ζ-chain expression by cytofluorimetric analysis.

Although selective outgrowth of some ζ-chain+ TCRBV2+ cells could not be ruled out, the comparison of the proportion of TCRBV2+ T cells (shown in parentheses in Table VI) in fresh and activated lymphocytes suggested that restoration of ζ-chain expression after TCRV + CD28-mediated activation was likely to be due to re-expression of this molecule in ζ-chain TCRBV2+ T cells. in fact, TCRBV2+ T cells in patient 28 represented 7% in fresh lymphocytes and 13% after activation, while at the same time, the proportion of ζ-chain within this TCRV region subset dropped from 80 to 4%. Furthermore, cell cycle analysis in the TCRBV2+ fraction of patient 28 (Table VII), after PBL activation with mAbs to TCRBV2 + CD28 (or with PHA as control), indicated activation of cell proliferation (as revealed by TCRBV2+ cells in S phase), but confirmed that only a minor fraction of the TCRBV2+ cells completed the cell cycle (as indicated by the low proportion of cells in G2 M phase). Interestingly, culture of T cells in medium alone for 96 h did not change the proportion of ζ-chain cells in any of the subsets (Table VI), suggesting that defective expression of this signaling molecule is a permanent loss, unless appropriate activation signals are provided. In addition to restoring ζ-chain expression, stimulation with mAb to TCRV region plus mAb to CD28 led to T cell activation (Fig. 9) as shown by [3H]thymidine incorporation and TNF release by lymphocytes from patients 4, 24, and 28. Taken together, these data suggest that a TCRV region-specific T cell activation protocol can rescue normal expression of TCR transducing elements in single TCRV region subsets from melanoma patients.

Table VII.

Cell cycle analysis of TCRBV2+ T cells activated by anti-TCRBV2 + anti-CD28 mAbsa

Cell Cycle Analysis at Time (h)T Cell Activated with
PHAmAbs to TCRBV2 and to CD28
G1SG2 MG1SG2 M
24 96.3b 3.7 94.5 5.5 
48 74.8 22.7 2.4 77.0 15.8 7.1 
72 65.8 21.2 13.0 84.4 13.0 2.6 
96 75.3 23.0 1.7 88.6 9.5 2.0 
Cell Cycle Analysis at Time (h)T Cell Activated with
PHAmAbs to TCRBV2 and to CD28
G1SG2 MG1SG2 M
24 96.3b 3.7 94.5 5.5 
48 74.8 22.7 2.4 77.0 15.8 7.1 
72 65.8 21.2 13.0 84.4 13.0 2.6 
96 75.3 23.0 1.7 88.6 9.5 2.0 
a

PBL from patient 28 were cultured for 96 h in the presence of mAbs to TCRBV2 and to CD28 or in the presence of PHA (1 μg/ml) as positive control. Cell cycle analysis within the TCRBV2+ subset was carried out by two-color immunofluorescence at 24, 48, 72, and 96 h of culture on cells stained with anti-TCRBV2 mAb and with propidium iodide.

b

Results expressed as proportion of TCRBV2+ T cells in G1, S, or G2 M of the cell cycle.

FIGURE 9.

Proliferative response and TNF release by lymphocytes from melanoma patients after stimulation with mAbs to TCRBV2 and CD28. Proliferative response (upper panels) was evaluated by [3H]thymidine incorporation and TNF release (lower panels) by MTT assay in PBL from three melanoma patients after stimulation for 144 h with anti-TCRBV2 mAbs plus anti-CD28 mAb. Data expressed as mean cpm (upper panels) or as U/ml of TNF (lower panels).

FIGURE 9.

Proliferative response and TNF release by lymphocytes from melanoma patients after stimulation with mAbs to TCRBV2 and CD28. Proliferative response (upper panels) was evaluated by [3H]thymidine incorporation and TNF release (lower panels) by MTT assay in PBL from three melanoma patients after stimulation for 144 h with anti-TCRBV2 mAbs plus anti-CD28 mAb. Data expressed as mean cpm (upper panels) or as U/ml of TNF (lower panels).

Close modal

The results of this study indicate that partial losses of ζ-chain, lck, and ZAP-70 molecules observed in PBL from metastatic melanoma patients do not affect equally T cell subset characterized by distinct TCRV regions. In contrast, the extent of signaling molecule loss showed a marked degree of selectivity depending on the TCRV region expressed by the T cells. When looking at the overall T cell population, stained by anti-CD3, marked loss of ζ-chain was found only in four patients, whereas defective expression of lck and ZAP-70 was much more frequent, being present at some level in CD3+ T cells from most patients. These data suggest that loss of lck and ZAP-70 is more frequent in the overall T cell fraction than ζ-chain down-modulation. However, when we looked at signaling molecule expression in distinct TCRV region subsets, ζ-chain, lck, and ZAP-70 were not expressed at similar levels in comparison to each other and to the overall CD3+ population. This suggested that in melanoma patients the process leading to loss of signaling molecule occurs with a degree of selectivity that depends on the TCR structure (i.e., the TCRV region expressed) of the T cells.

Furthermore, in some patients a high level of defective expression of ζ-chain, lck, or ZAP-70 was evident in the global T cell population as identified by anti-CD3 staining. In these instances, not unexpectedly, the loss of T cell signaling molecules did not show TCRV selectivity, but affected similarly and at high levels all three signaling molecules investigated. Additional evidence indicated that loss of signaling molecules in PBL of these patients is maintained or increased over time during disease progression, while maintaining TCRV region selectivity. The possible relationship between ζ-chain loss and progression is in agreement with previous findings (7), although more recent data in experimental models do not support an increased T cell signaling molecule loss during tumor progression (31).

The results of analysis of T cell signaling molecule expression in patients’ PBL prompted us to evaluate whether direct T cell-tumor interaction or exposure of patients’ lymphocytes to tumor supernatants might contribute to induce the loss of signaling molecules. The results indicated that only neoplastic cells or their supernatants, but not normal cells as fibroblasts, could down modulate T cell signaling molecules. More importantly, tumor cells and tumor-derived supernatants induced a TCRV region selective loss of signaling molecules. Thus, both freshly isolated PBL as well as coculture experiments confirmed the existence of a mechanism that induces selective T cell signaling molecule loss depending on the TCR structure. On the basis of the significant tumor burden present in most patients of the study at time the PBL were isolated, it is possible that even the results observed with fresh uncultured lymphocytes may reflect the effect of interaction in vivo of T cells with tumor cells or with factors release by the neoplastic cells. Interestingly, the effects of tumor cells and of their supernatants were observed in autologous and in allogeneic T cell-tumor combinations as well as on healthy donors lymphocytes. The latter observation indicates that tumor cells and/or tumor-derived factors can induce loss of signaling molecules even in lymphocytes that have never been previously exposed to tumor cells (such as those from healthy donors). In contrast with in vitro experiments, showing that tumor cells or their supernatants could down-modulate all three signaling molecules, analysis of freshly isolated PBL indicated frequent losses of expression of lck and ZAP-70 but less frequent defective expression of ζ-chain. The reasons for this discrepancy are unknown but could reflect differences between in vivo and in vitro at the level of the concentrations of the factors that induce the defect of expression, as well as differential susceptibility of the three signaling molecules to the mechanisms that leads to their loss.

Previous studies have shown that soluble HLA molecules or peptides derived from HLA nonpolymorphic regions have an immunomodulatory effect on T cells that can lead to apoptosis or inhibition of T cell effector function (32, 33). These data, and the observation that signaling molecule loss in patients’ PBL was TCRV region selective, prompted us to evaluate evidence for a role of tumor-derived HLA molecules. The experiments of immunodepletion of tumor supernatants with a mAb (w6/32) that binds specifically soluble HLA class I molecules associated with β2-microglobulin, supported a role of tumor-derived HLA class I molecules in inducing the loss of T cell signaling molecules. Removal of HLA class I molecules, but not of another tumor derived molecule (ICAM-1), markedly reduced and in some instances abolished the suppressive effect of tumor supernatants on signaling molecule expression. These results suggest a model consistent with data indicating that in certain conditions (such as those resulting from partial T cell activation or prolonged TCR engagement) specific interaction of TCRs with MHC/peptide complexes may lead to impairment of TCR transduction pathway and even to induction of tolerance (34, 35, 36, 37). Such a model, based on direct interaction of TCRs with tumor-derived HLA-peptide complexes, can even provide a possible explanation for the observation that T cells expressing the same TCRV region, but isolated from different patients or donors, were not affected equally by the same tumor or tumor supernatants. In fact, pools of T cells expressing the same TCRV region but isolated from distinct individuals, or even from distinct tissues, are not expected to represent the same immune repertoire (in terms of the spectrum of TCR specificities) (38). Therefore, a single set of tumor-derived HLA class I molecules is not expected to interact in the same way and to determine the same effect on distinct T cell pools, although these share the same TCRV region. The comparison of tumor-derived and fibroblasts-derived supernatants indicated that both contained shed HLA molecules, but only the former exerted suppressive effect on T cell signaling molecule expression, even in conditions where the tumor and the fibroblasts were both autologous to the T lymphocytes. This might be due to differences in the amount of soluble HLA complexes shed by normal cells in comparison to neoplastic cells. Alternatively, the nature of the sets of HLA-peptide complexes found in tumor vs fibroblasts supernatants may play a role. Differential gene expression in melanoma cells vs fibroblasts, due to difference in histological origin and to neoplastic transformation, might in fact lead these two cell types to shed significantly different sets of HLA-peptide complexes in terms of both relative abundance of each HLA-class I allele and nature of the peptides bound to each allele. Furthermore, T cell signaling molecule loss was induced both in autologous and in allogeneic tumor supernatant-T cell combinations, and independently from the source of T cells (namely donors or patients). This suggests that both HLA-restricted interactions (in the autologous combination) as well as interactions due to MHC incompatibility (i.e., alloreactivity) may contribute to the effect.

As shown in previous studies (23) tumor-derived factors may contribute to loss of signaling molecules in patients’ T cells. It is thus possible that the final result (i.e., loss of T cell signaling molecules) depends on the activity and concentration not only of soluble HLA class I molecules, but even of additional factors released by neoplastic, but not normal cells.

Although we investigated only three different T cell subsets, it appears likely that the results seen on TCRAV2+, TCRBV2+, and TCRBV18+ T cells may be true for other T cell subsets that can be distinguished on the basis of expression of any TCRAV or TCRBV region. The implication of these findings is that loss of TCR signaling molecules in metastatic melanoma patients may lead to functional impairment of selected T cell subsets depending on the structure of their TCR. Thus, at some point during evolution of the disease, the immune repertoire of a patient may show a heterogeneous pattern of normal and defective T cells. Moreover, according to the results of this study, the set of defective T cells may show several degrees of functional impairment due to the complexity deriving from the nature of signaling molecule loss (ζ-chain, lck, or ZAP-70), the extent of loss and to the structure of the TCR expressed by the defective T cells. Clearly, it is even possible that a fraction of the defective T cells may express TCRs directed to tumor Ags, thus providing a mechanism for possible tumor escape from immune surveillance.

Previous studies have shown that correction of the defects in signal transducing molecule expression in T cells of tumor patients can be achieved by different activation methods including stimulation with anti-CD3 plus anti-CD28 mAb (39) or T cell culture with rIL-2 (10), or with IL-2-transduced tumor cells (14, 22), or even by in vivo IL-2 therapy of melanoma patients (16). However, in those studies it was not assessed whether correction of the defective expression resulted from outgrowth of T cells expressing the signal transducing molecule or re-expression in the T cells initially showing loss of expression. Comparison of the proportion of ζ-chain+ T cells and of TCRV+ T cells before and after activation with the TCRV + CD28 protocol suggested that the reduction in ζ-chain T cells after activation was due to signaling molecule re-expression in ζ-chain T cells. Although outgrowth of some ζ-chain+ T cells could not be ruled out, cell cycle analysis in a selected TCRV region subset showed that only a minor fraction of cells completed the cell cycle. Nevertheless, the observed proliferation (in terms of proportion of T cells entering the S phase) correlated with [3H]thymidine incorporation and cytokine release (TNF), indicating that it was possible to rescue T cell function in response to TCR and costimulatory molecule triggering.

We thank Dr. F. Belli (Department of Surgical Oncology of our Institute) for assistance in patient selection and Dr. M. Maio (C.R.O., Aviano, Italy) for the gift of mAbs. The skilful technical work of A. Molla is gratefully acknowledged.

1

This work was supported in part by funds from the Italy-USA Program on Therapy of Tumors (ISS-Rome) and the Italian Association for Cancer Research (AIRC, Milan, Italy). C.M. was supported by a fellowship from AIRC. P.P. was supported by a fellowship from the Fondazione Italiana per la Ricerca sul Canciro (FIRC) (Milan, Italy).

3

Abbreviation used in this paper: TAL, tumor-infiltrated lymph nodes.

1
Miescher, S., T. L. Whiteside, S. Carrel, V. Von Fliedner.
1986
. Functional properties of tumor infiltrating and blood lymphocytes in patients with solid tumors: effects of tumor cells and their supernatants on proliferative responses of lymphocytes.
J. Immunol.
136
:
1899
2
Miescher, S., M. Stoeck, L. Qiao, C. Barras, L. Barrelet, V. Von Fliedner.
1988
. Proliferative and cytolitic potentials of purified human tumor-infiltrating T lymphocytes: impaired response to mitogen-driven stimulation despite T-cell receptor expression.
Int. J. Cancer
42
:
659
3
Yoshino, I., T. Yuno, M. Murata, T. Ishida, K. Sugimachi, G. Kimura, K. Nomoto.
1992
. Tumor-reactive T-cells accumulate in lung cancer tissues but fail to respond due to tumor cell-derived factor.
Cancer Res.
52
:
775
4
Alexander, J. P., S. Kudoh, K. A. Melsop, T. A. Hamilton, M. G. Edinger, R. R. Tubbs, D. Sica, L. Tuason, E. Klein, R. M. Bukowski, J.H. Finke.
1993
. T-cell infiltrating renal cell carcinoma display a poor proliferative response even though they can produce IL-2 and express IL-2 receptors.
Cancer Res.
53
:
1380
5
Finke, J. H., A. H. Zea, J. Stanley, D. L. Longo, H. Mizoguchi, R. R. Tubbs, R. H. Wiltrout, J. J. O’Shea, S. Kudoh, E. Klein, et al
1993
. Loss of T-cell receptor ζ chain and p56lck in T-cells infiltrating human renal cell carcinoma.
Cancer Res.
53
:
5613
6
Tartour, E., S. Latour, C. Mathiot, N. Thiounn, V. Mosseri, I. Joyeux, C. Dubois D’Enghien, L. Renshiang, B. Debre, W. H. Fridman.
1995
. Variable expression of CD3-ζ chain in tumor-infiltrating lymphocytes (TIL) derived from renal-cell carcinoma: relationship with TIL phenotype and function.
Int. J. Cancer
63
:
205
7
Matsuda, M., M. Petersson, R. Lenkei, J. L. Taupin, I. Magnusson, H. Mellstedt, P. Anderson, R. Kiessling.
1995
. Alterations in the signal-transducing molecules of T cells and NK cells in colorectal tumor-infiltrating, gut mucosal and peripheral lymphocytes: correlation with the stage of the disease.
Int. J. Cancer
61
:
765
8
Nakagomi, H., M. Petersson, I. Magnusson, C. Juhlin, M. Matsuda, J. L. Taupin, E. Vivier, P. Anderson, R. Kiessling.
1993
. Decreased expression of the signal-transducing ζ chains in tumor-infiltrating T-cells and NK cells of patients with colorectal carcinoma.
Cancer Res.
53
:
5610
9
Cabrera, T., A. Collado, M. A. Fernandez, A. Ferron, J. Sancho, F. Ruiz-Cabello, F. Garrido.
1998
. High frequency of altered HLA class I phenotypes in invasive colorectal carcinomas.
Tissue Antigens
52
:
114
10
Farace, F., E. Angevin, J. Vanderplancke, B. Escudier, F. Triebel.
1994
. The decreased expression of CD3 ζ chains in cancer patients is not reversed by IL-2 administration.
Int. J. Cancer
59
:
752
11
Kono, K., M. E. Ressing, R. M. P. Brandt, C. J. M. Melief, R. K. Potkul, B. Andersson, M. Petersson, M. Kast, R. Kiessling.
1996
. Decreased expression of signal-transducing ζ chain in peripheral T cells and natural killer cells in patients with cervical cancer.
Clin. Cancer Res.
2
:
1825
12
Rossi, E., E. Matutes, R. Morilla, K. Owusu-Ankomah, A. M. Heffernan, D. Catovsky.
1996
. ζ Chain and CD28 are poorly expressed on T lymphocytes from chronic lymphocytic leukemia.
Leukemia
10
:
494
13
Zea, A. H., B. D. Cutri, D. L. Longo, W. G. Alvord, S. L. Strobl, H. Mizoguchi, S. P. Creekmore, J. J. O’Shea, G. C. Powers, W. J. Urba, A. C. Ochoa.
1995
. Alterations in T cell receptor and signal transduction molecules in melanoma patients.
Clin. Cancer Res.
1
:
1327
14
Salvadori, C. B., A. Gansbacher, A. Pizzimenti, K. I. Zier.
1994
. Abnormal signal transduction by T cells of mice with parental tumor is not seen in mice bearing IL-2 secreting tumors.
J. Immunol.
153
:
5176
15
Noda, S., T. Nagata-Narumiya, A. Kosugi, S. Narumiya, C. Ra, H. Fujiwara, T. Hamaoka.
1995
. Do structural changes of T cell receptor complex occur in tumor-bearing state?.
Jpn. J. Cancer Res.
86
:
383
16
Rabinowich, H., M. Banks, T. E. Reichert, T. F. Logan, J. M. Kirkwood, T. L. Whiteside.
1996
. Expression and activity of signaling molecules in T lymphocytes obtained from patients with metastatic melanoma before and after interleukin 2 therapy.
Clin. Cancer Res.
2
:
1263
17
Correa, M. R., A. C. Ochoa, P. Ghosh, H. Mizoguchi, L. Harvey, D. L. Longo.
1997
. Sequential development of structural and functional alterations in T cells from tumor-bearing mice.
J. Immunol.
158
:
5292
18
Kurt, R. A., W. J. Urba, J. W. Smith, D. D. Schoof.
1998
. Peripheral T lymphocytes from women with breast cancer exhibit abnormal protein expression of several signaling molecules.
Int. J. Cancer
78
:
16
19
Wang, Q., J. Stanley, S. Kudoh, J. Myles, V. Kolenko, T. Yi, R. Tubbs, R. Bukowski, J. Finke.
1995
. T cells infiltrating non-Hodgkin’s B cell lymphomas show altered tyrosine phosphorylation pattern even though T cell receptor/CD3-associated kinases are present.
J. Immunol.
155
:
1382
20
Fargnoli, M. C., R. L. Edelson, C. L. Berger, S. Chimenti, C. Couture, T. Mustelin, R. Halaban.
1997
. Diminished TCR signaling in cutaneous T cell lymphoma is associated with decreased activities of ZAP70, Syk and membrane-associated Csk.
Leukemia
11
:
1338
21
Mulder, W. M. C., E. Bloemena, M. J. Stukart, J. A. Kummer, J. Wagstaff, R. J. Scheper.
1997
. T cell receptor-ζ and granzyme B expression in mononuclear cell infiltrates in colon carcinoma.
Gut
40
:
153
22
Guarini, A., L. Riera, A. Cignetti, L. Montacchini, M. Massaia, R. Foa.
1997
. Transfer of interleukin-2 gene into human cancer cells induces specific antitumor recognition and restores the expression of CD3/T-cell receptor associated signal transduction molecules.
Blood
89
:
212
23
Kolenko, V., Q. Wang, M. C. Riedy, J. O’Shea, J. Ritz, M. K. Cathcart, P. Rayman, R. Tubbs, M. Edinger, A. Novick, et al
1997
. Tumor-induced suppression of T lymphocyte proliferation coincides with inhibition of Jak3 expression and IL-2 receptor signaling.
J. Immunol.
159
:
3057
24
Aoe, T., Y. Okamoto, T. Saito.
1995
. Activated macrophages induce structural abnormalities of the T cell receptor-CD3 complex.
J. Exp. Med.
181
:
1881
25
Otsuji, M., Y. Kimura, T. Aoe, Y. Okamoto, T. Saito.
1996
. Oxidative stress by tumor-derived macrophages suppresses the expression of CD3 ζ chain of T-cell receptor complex and antigen-specific T-cell responses.
Proc. Natl. Acad. Sci. USA
93
:
13119
26
Maccalli, C., C. Farina, M. L. Sensi, G. Parmiani, A. Anichini.
1997
. TCR β chain variable region-driven selection and massive expansion of HLA-class I-restricted antitumor CTL lines from HLA-A*0201+ melanoma patients.
J. Immunol.
158
:
5902
27
Sander, B., J. Andersson, U. Andersson.
1991
. Assessment of cytokines by immunofluorescence and the paraformaldehyde-saponin procedure.
Immunol. Rev.
119
:
65
28
Makgoba, M. W., M. W. Sanders, G. E. Luce, M. Dustin, T. A. Springer, E. A. Clark, P. Mannoni, S. Shaw.
1988
. ICAM-1, a ligand for LFA-1-dependent adhesion of B, T and myeloid cells.
Nature
331
:
86
29
Espevick, T., J. Nissen-Meyer.
1986
. A highly sensitive cell line, WEHI 164 clone 13, for measuring cytotoxic factor/tumor necrosis factor from human monocytes.
J. Immunol. Methods
95
:
99
30
Hansen, M. B., S. E. Nielsen, K. Berg.
1989
. Re-examination and further development of a precise and rapid dye method for measuring cell growth/cell kill.
J. Immunol. Methods
119
:
203
31
Levey, D. L, P. K. Srivastava.
1995
. T cells from late tumor-bearing mice express normal levels of p56lck, p59fyn, ZAP-70, and CD3 ζ despite suppressed cytolytic activity.
J. Exp. Med.
182
:
1029
32
Zavazava, N., M. Kronke.
1996
. Soluble HLA class I molecules induce apoptosis in alloreactive cytotoxic T lymphocytes.
Nat. Med.
9
:
1005
33
Zavazava, N..
1998
. Soluble HLA class I molecules: biological significance and clinical implications.
Mol. Med. Today
4
:
116
34
Sloan-Lancaster, J., A. S. Shaw, J. B. Rothbard, P. M. Allen.
1994
. Partial T cell signaling: altered phospho-ζ and lack of ZAP70 recruitment in APL-induced T cell anergy.
Cell
79
:
913
35
Lee, J. E., M. B. Cossoy, L. A. Chau, B. Singh, J. Madrenas.
1997
. Inactivation of lck and loss of TCR-mediated signaling upon persistent engagement with complexes of peptide: MHC molecules.
J. Immunol.
159
:
61
36
Loftus, D. J., P. Squarcina, M. B. Nielsen, C. Geisler, C. Castelli, N. Ø dum, E. Appella, G. Parmiani, L. Rivoltini.
1998
. Peptides derived from self-proteins as partial agonists and antagonists of human CD8+ T-cell clones reactive to melanoma/melanocyte epitope MART127–35.
Cancer Res.
58
:
2433
37
Tang, B., L. K. Myers, E. F. Rosloniec, K. B. Whittington, J. M. Stuart, A. H. Kang.
1998
. Characterization of signal transduction through the TCR-ζ chain following T cell stimulation with analogue peptides of type II collagen 260–267.
J. Immunol.
160
:
3135
38
Sensi, M. L., G. Parmiani.
1995
. Analysis of TCR usage in human tumors: a new tool for assessing tumor-specific immune responses.
Immunol. Today
16
:
588
39
Renner, C., S. Ohnesorege, G. Held, S. Bauer, W. Jung, J. P. Pfitzenmeier, M. Pfreundschuh.
1996
. T cells from patients with Hodgkin’s disease have a defective T-cell receptor ζ chain expression that is reversible by T-cell stimulation with CD3 and CD28.
Blood
88
:
236