T-cell receptor gene expression in tumour-infiltrating lymphocytes and peripheral blood lymphocytes of patients with nasopharyngeal carcinoma.

The T-cell receptor (TCR) repertoire expression of tumour-infiltrating lymphocytes (TILs) from 19 nasopharyngeal carcinoma (NPC) biopsies was compared with those of lymphocytes from 18 control nasopharyngeal biopsies. mRNA was extracted from these lymphocytes and the cDNA transcribed. A panel of 18 V alpha- and 21 V beta-specific primers was used to detect the TCR gene use from cDNA. The use of V alpha and V beta genes was restricted in TILs compared with lymphocytes from biopsies. The frequencies of V alpha 2, V alpha 3, V alpha 9, V alpha 10, V alpha 11, V alpha 13, V alpha 14, V alpha 15, V beta 11, V beta 15 and V beta 20 were decreased and the frequencies of V alpha 10 [Pc = 0.04; relative risk (RR) = 0.05], V alpha 11 (Pc = 0.02; RR = 0.07), V alpha 13 (Pc = 0.002; RR = 0), V alpha 14 (Pc = 0.04; RR = 0.05), V beta 14 (Pc = 0.001; RR = 0.03) and V beta 20 (Pc = 0.001; RR = 0.03) remained significantly reduced after correction for the number of families typed. The frequency of V alpha 17 was higher in NPC biopsies than in NPC PBLs (P = 0.05), and the frequency of V beta 15 was lower in NPC biopsies than in NPC PBLs (P = 0.02). The frequencies of V alpha 17 and V alpha 18 in HLA-B46+ patients were significantly lower (P = 0.009; P = 0.044) than in B46+ controls. The results suggest that the restriction of TCR gene use in NPC patients may be important in NPC pathogenesis.

The T-cell receptor (TCR) repertoire expression of tumour-infiltrating lymphocytes (TILs) from 19 nasopharyngeal carcinoma (NPC) biopsies was compared with those of lymphocytes from 18 control nasopharyngeal biopsies. mRNA was extracted from these lymphocytes and the cDNA transcribed. A panel of 18 VaE-and 21 VA-specific primers was used to detect the TCR gene use from cDNA. The use of VcE and VP genes was restricted in TILs compared with lymphocytes from biopsies. The frequencies of Ve2, VO3, Vx9, VaO0, VaEl, Va13, Va14, Va15, V011, VP14, VP15 and VP20 were decreased and the frequencies of ValO [P, = 0.04; relative risk (RR) = 0.05], Val1 (P, = 0.02; RR = 0.07), Va13 (P, = 0.002; RR = 0), VaEl4 (P= 0.04; RR = 0.05), VP14 (P =O0.001; RR = 0.03) and V)2 (P =O0.001; RR = 0.03) remained significantly reduced after correction for the number of families typed. The frequency of Val7 was higher in NPC biopsies than in NPC PBLs (P = 0.05), and the frequency of VP15 was lower in NPC biopsies than in NPC PBLs (P= 0.02). The frequencies of Val7 and Va18 in HLA-B46' patients were significantly lower (P = 0.009; P = 0.044) than in B46 + controls. The results suggest that the restriction of TCR gene use in NPC patients may be important in NPC pathogenesis.
Keywords: TCR; TIL; PBL; NPC; frequency; usage Most mature T lymphocytes specifically recognise antigens presented on MHC molecules through the T-cell receptor (TCR). Analysis of the tumour-infiltrating lymphocytes (TILs) may help to clarify their fole in tumour cell destruction and to achieve a better understanding of the cellular and molecular basis of lymphocyte-tumour interaction.
Nasopharyngeal carcinoma (NPC) is an epithelial tumour characterised by a marked lymphocytic infiltration (Shanmugaratnam et al., 1979>. HLA-B46 and -B58 haplotypes are associated with NPC (Chan et al., 1983). These same alleles are also associated with autoimmune diseases in the Chinese (Chan et al., 1978(Chan et al., , 1981(Chan et al., , 1993Lee et al., 1983), suggesting that these HLA alleles may be associated with abnormal immune functions. Since HLA molecules present foreign and self peptides to T cells via TCR, the restricted HLA alleles may result in a restriction of antigen/MHC combinations as well as a restriction of TCR expression. The association of HLA and TCR in autoimmune diseases has been well documented Wucherpfennig et al., 1990Wucherpfennig et al., , 1992Ben-Nun et al., 1991;Davies et al., 1991Davies et al., , 1992Sioud et al., 1991Sioud et al., , 1992Sottini et al., 1991;Giegerich et al., 1992;Martin et al., 1992;Sumida et al., 1992;Utz et al., 1993). Recent studies have also demonstrated the association between TCR and some malignancies, such as melanoma (Nitta et al., 1990(Nitta et al., , 1991aBennett et al., 1992;Weidmann et al., 1993), glioma, medulloblastoma (Nitta et al., 1991b) and pulmonary and renal carcinomas (Bennett et al., 1992), suggesting that these malignant neoplasms may have stimulated a specific T-lymphocyte response through antigen recognition by the TCR. NPC is also associated with the Epstein-Barr virus (EBV), with high antibody titres to various EBV antigens and low Tcytotoxicity levels Moss et al., 1983). These abnormal immune responses to EBV may result from inappropriate T-cell responses to HLA/EBV peptide combinations. TCR gene polymorphism in NPC has been shown (Chen and Chan, 1994), and we are further investigating whether there are particular TCR uses or lack of in TILs in NPCs, especially in those with HLA-B46 and -B58.
The present study summarises an analysis of TCR variable gene family expression in peripheral blood lymphocytes Correspondence: SH Chan Received 7 December 1994; revised 20 February 1995; accepted 23 February 1995 (PBLs) and TILs in Singaporean Chinese NPC patients and controls using the polymerase chain reaction (PCR) technique. The results showed that there were differences in TCR gene expression between NPC patients and controls: NPC patients had lower frequency of expression of some TCR gene families in PBLs and TILs compared with controls.

Biopsies
Surgical biopsies were obtained from 37 untreated patients suspected of having NPC seen for the first time at a major ENT out-patient clinic. From subsequent histopathology results, there were 19 NPC patients and the 18 biopsynegative patients served as controls. Patients and controls were Chinese from Singapore. The biopsies were collected in sterile medium and processed within 4 h.

Preparation of TILs
Fresh biopsies were teased with two pairs of forceps in a 3.5 cm Petri dish containing 1 ml of RPMI-1640. The released cell suspension was transferred to a 15 ml tube and washed once with 10 ml of RPMI-1640. The cells were resuspended and cultured (37C with 5% carbon -dioxide) in RPMI-1640 with 10% heat-inactivated pooled human serum (from blood donors) and 15 u ml-' interleuk-in 2 (IL-2) (Boerhinger Mannheim). TILs were cultured for 1-2 weeks and the medium (with IL-2) replaced every other day. Cells were harvested by washing three times in 15 ml of phosphate-buffered saline (PBS)-glucose and the cell pellet frozen immediately at -70C. The frozen cells were thawed for RNA extraction in batches.
Peripheral blood samples and preparation of PBLs Peripheral blood samples were also obtained from the patients and controls at the time of the biopsy. Heparinised blood was mixed with an equal volume of PBS-glucose and centrifuged over a Ficoll-Hypaque density gradient for 20 min at 2000 r.p.m. Cells from the interface were collected and washed twice with PBS-glucose. The peripheral blood lymphocytes were cultured and harvested the same way as 118 TILs. HLA typing was also performed on the separated peripheral blood before culture.
Preparation of RNA mRNA from biopsy lymphocytes and PBLs was prepared by using the QuickPrep Micro mRNA Purification Kit (Phar-macia). Briefly, l06-i0' cells were extracted in a buffered solution containing a high concentration of guanidinium thiocyanate (GTC) and then diluted 3-fold with elution buffer. The supernatant clarified by centrifugation was transferred to a microcentrifuge tube containing oligo(dT)-cellulose. After 3 min, during which time the poly(A) RNA bound to the oligo(dT)-cellulose, the tube was centrifuged at 12 000 r.p.m. for lOs. The pelleted oligo(dT)-cellulose was then washed five times with high-salt buffer and twice with lowsalt buffer. The oligo(dT)-cellulose slurry was transferred to a MicroSpin column. Polyadenylated RNA was eluted with elution buffer, prewarmed at 65°C and precipitated in the presence of one-tenth the volume of potassium acetate and 2.5 volumes of 100% ethanol for a minimum of 2 h at -20°C.
cDNA synthesis An aliquot of 1 -2 iLg of mRNA was used for the synthesis of single-strand cDNA in a final volume of 40 ,l, with 50 mM Tris-HCl, 75 mM potassium chloride, 3 mM magnesium chloride, 10 mM dithiothreitol (DTT), 0.5 mM dNTP (Promega), 80 units of RNAsin (Promega), 0.2 iLg of oligo(dT) (New England Biolabs) and 400 units of SuperScript reverse transcriptase (Gibco BRL). The reaction mixture was incubated for 1 h at 37TC and heated at 95°C for 5 min.

PCR
A 1 gil volume of single-stranded cDNA was amplified using either a Va-specific and C4 primers or a VP-specific and Cp primers at a final concentration of 0.5 .LM in each reaction.
Sequences of individual primers are listed in Table I (Nitta et  al ., 1991a;Sottini et al., 1991). The size of amplified products ranged from 250 to 450 bp. Oligonucleotides were synthesised (Biosynthesis, USA) and the amplification was performed with 1 unit of Taq polymerase (Perkin Elmer) on a DNA thermal cycler (Perkin Elmer). The PCR cycle profile was denaturation at 95'C for 1 min, annealing of primers at 55°C (a-chain) or 50°C (a-chain) for 1 min and extension of reaction at 72°C for 1 min for 35 cycles. PCR products were separated on 1.4% agarose gels. Expression of Va or VP genes was considered positive when a correct size band (250-450 bp) was visualised after ethidium bromide staining and when the amplified products were positively hybridised with Caor Cp-specific oligonucleotide probe on Southern blots.

Southern blot analysis
Ten microlitres of amplified products was electrophoresed in a 1.4% agarose gel for 40min and transferred on nylon membrane (Amersham Aylesbury, UK) as described by Southern (1975 with 2 x SSPE at room temperature for 5 min twice. The filters were blocked with skimmed milk for 30 min and then incubated with alkaline phosphatase (AP)-conjugated antidigoxigenin antibody for 30 min in the presence of 0.1 M Tris and 0.15 M sodium chloride. The colour was developed in the presence of 5-bromo-4-chloro-3-indolyl phosphate (Xphosphate) and nitroblue tetrazolium (NBT).

Statistic analysis
The frequencies of VcE or VP expression in patients and controls were analysed using the X2 and Fishers' exact test.
To minimise the possibility of obtaining significant differences by pure chance, the P-value was multiplied by the number of Va (18) or VP (21) families tested to give the corrected P (PJ value. Relative risks (RR) were the crossproducts of the cells in 2 x 2 tables and the 95% confidence limits (CL) calculated.

Optimisation ofprimers
To analyse TCR Vax and VP repertoire expression in this study, PCR amplification using 18 Vm-and 21 VP-specific primers was performed. To test and optimise the primers, mRNA from PBLs of normal individuals was isolated and primer specificities were further confirmed by hybridisation of a Dig-dUTP-labelled Ca or Ci oligonucleotide probe (data not shown).
TCR Va gene expression in TILs and control biopsies TILs and lymphocytes from control biopsies were isolated and cultured for 1-2 weeks and mRNA was extracted to prepare cDNA transcript as a template for the PCR. Overall, of the 18 different Va gene families, the average number of Va genes used was 15.3 in control biopsies (n = 18) and 9.6 in NPC TILs (n = 17). TILs also showed less Va repertoire expression than lymphocytes from control biopsies. The 18 families of Va genes were not equally represented and not all the VaE genes were detectable in all subjects, indicating that some TCR genes were not being expressed (Figure 1). NPC TILs showed considerable individual variation in TCR expression. Most of the TIL samples did not express the complete range of Va families and lacked as many as 13 gene families. When the frequency of specific TCR families was analysed, there were lower frequencies of Va2, VW3, Vx9, ValO, VEll, Vaz13, Vcz14 and Vl.5 genes in TILs compared with those of control biopsies. Of these, Vz10, Vall, Vx13 and VE14 showed significantly lower frequency of expression even after correction for the number of families typed (Table  II). None of the TCR families was expressed more frequently in TTLs than in control biopsies.
T cells that recognse and respond to a specific antigenic peptide by activation and prolferation are considered to be clonally restrcted and to express a imited number of TCR genes (Innide and Whiteside, 1993). The complexity of TCR use in the anti-tumourresonse may result from the involvement of multiple m-and A-chain regions in response to a single antigenic determinant or may reflect mutiple antigenic determinantsexp on tumour cells. Our previous study on rition f ent length polymorphism (RFLP) of TCR genes indicated that there is polymorphism of TCR genes in NPC patients (Chen and Clha, 1994). In the present study, TCR gene expression in  , 1991), and culture with IL-2 reduced rather than ehanced the degree of rstiction. It is possible that T cells with cerain TCRs fail to grow in the presece ofIL-2.
However, T cells with these TCRs grew well in control cultures. On the other hand, immunosuppressed T cells may behave differently, and it is possible that su ed T cells with certain TCRs may not grow in the presence of IL-2, and we are exploring this possibility. Whether the deletion of these TCR genes leads to failure of T-cell recognition of tumour peptide/HLA complexes resulting in the escape of NPC tumours from immune surveillance, or whether NPC results in the deletion of specific TCR genes also need to be investigated We found no specific TCR gene use in NPC, but rather a specific deletion of certain TCR genes, and this finding together with HLA restriction may be important in the pathogenesis of NPC.