Aspartic acid homozygosity at codon 57 of HLA-DQ beta is associated with susceptibility to pulmonary tuberculosis in Cambodia

After infection with Mycobacterium tuberculosis, clinical disease usually remains latent, contained by the host immune response. Although polymorphisms of HLA loci have been hypothesized to play a major role in the breakdown of latency, a functional
of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Aspartic Acid Homozygosity at Codon 57 of HLA-DQ    IsAssociated with Susceptibility to Pulmonary Tuberculosis inCambodia 1 Julio C. Delgado,* † Andres Baena, 2 * Sok Thim, § and Anne E. Goldfeld 3 * ‡ After infection with  Mycobacterium tuberculosis , clinical disease usually remains latent, contained by the host immune response.Although polymorphisms of HLA loci have been hypothesized to play a major role in the breakdown of latency, a functional linkhas not been established. Molecular-based HLA-typing methods were used to test the association of sets of HLA alleles encodingan aspartic acid at codon 57 of the HLA-DQ   -chain (HLA-DQ   57-Asp) with susceptibility to tuberculosis in a cohort of 436pulmonary tuberculosis patients and 107 healthy controls from Cambodia. HLA class II null cells were transduced with HLA-DQ  57-Asp or HLA-DQ   57-Ala and evaluated for their ability to bind peptides from two immunogenic  M. tuberculosis  specificproteins, ESAT-6 and CFP-10. In this study, we report a highly significant association between progressive pulmonary tuberculosisand homozygosity for HLA-DQ   57-Asp alleles. The presence of HLA-DQ   57-Asp resulted in a significantly reduced ability tobind a peptide from the central region of the ESAT-6 protein. Furthermore, when this peptide was presented by an HLA-DQ  57-Asp allele, Ag-specific IFN-    production from CD4  T cells from tuberculosis patients was significantly less than when thispeptide was presented by an HLA-DQ-   allele encoding an alanine at codon 57. Multiple genetic loci and ethnic-specific factorsare likely involved in the human immune response to tuberculosis. The data presented here provide a functional explanation fora highly significant association between an HLA polymorphism and tuberculosis in a highly characterized group of patients withsusceptibility to progressive tuberculosis infection in Cambodia.  The Journal of Immunology,  2006, 176: 1090–1097. H uman beings have evolved on a background of infectionand it has long been speculated that genetic variabilityamong populations that confers resistance against infec-tious pathogens is favored by evolution (1). Numerous case-con-trol studies have identified associations between susceptibility orresistance to human  Mycobacterium tuberculosis  infection andclinical disease with polymorphisms of candidate genes (2–7).However, these polymorphisms are of unknown functional impactupon tuberculosis pathogenesis, and several of these candidategene variations are ethnic-specific markers of other linked andunidentified genetic factors that may impact host immune re-sponses to tuberculosis susceptibility or resistance in a givenpopulation (8).The highly polymorphic HLA class II loci encode moleculesresponsible for Ag presentation to CD4  T cells which are criticalin containment of   M. tuberculosis  infection (9–12). Several allelesof the HLA class II DR2 serotype (13–16), and the  HLA- DQB1*0503  allele (17) are associated with susceptibility to pro-gression to clinical tuberculosis after infection in diverse popula-tions. However, despite these multiple association studies, afunctional link between HLA associations, Ag-driven T cell re-sponses, and tuberculosis disease has not been demonstrated.Crystal structures of HLA class II molecules have shown thatpeptides bind to a groove in the HLA class II molecule, and thatAg-binding specificity is determined by pockets formed by poly-morphic side chains (18, 19). HLA-DQ molecules, encoded bypolymorphic HLA-DQ    and   -chain genes, bind peptides withcertain amino acids that are anchored at specific positions withinpeptide-binding pockets termed P1, P4, and P9 in the HLA-DQmolecular groove (20, 21). P9-binding specificity for example, iscritically dependent upon the specific amino acid at codon 57 of the HLA-DQ   -chain (HLA-DQ   57), because this position in-fluences both the charge of pocket 9 and peptide-binding spatialconstraints created by the specific amino acid side chains at thiscodon. Specifically, the P9 pocket has different peptide-bindingspecificities dependent upon whether there is an aspartic acid(HLA-DQ   57-Asp) or a nonaspartic acid at this position (22, 23)(Fig. 1).The presence of an alanine at this position (HLA-DQ   57-Ala),which is encoded by a variety of HLA-DQ alleles, favors the bind-ing of large branched amino acids (24), and is associated withsusceptibility to autoimmune type 1 diabetes in humans (25, 26).By contrast, HLA-DQ   57-Asp, which is also encoded by a num-ber of HLA-DQ alleles, favors the binding of small hydrophobicpeptides (24). Notably, the  HLA-DQB1*0503  allele we previouslydemonstrated to be associated with tuberculosis susceptibility inCambodia (17) encodes an aspartic acid at HLA-DQ   57. This ledus to hypothesize that the presence of an aspartic acid at codon 57of HLA-DQ may influence P9-binding specificity, and thus impactthe functional binding of HLA-DQ and subsequent host immuneresponses to immunogenic  M. tuberculosis -derived peptides.Thus, we performed HLA association studies investigatingwhether sets of alleles that form a similar P9 peptide-binding *CBR Institute for Biomedical Research, Harvard Medical School, Boston, MA02115;  † Department of Pathology and  ‡ Department of Medicine, Brigham and Wom-en’s Hospital, Harvard Medical School, Boston, MA 02115; and  § The CambodianHealth Committee, Phnom Penh, CambodiaReceived for publication August 8, 2005. Accepted for publication October 31, 2005.The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked  advertisement   in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by National Institutes of Health Grants HL67471 (toJ.C.D.) and HL59838 (to A.E.G.). 2 Current address: Department of Microbiology and Immunology, Albert EinsteinCollege of Medicine, Bronx, NY 10461. 3 Address correspondence and reprint requests to Dr. Anne E. Goldfeld, CBR Institutefor Biomedical Research, 800 Huntington Avenue, Boston, MA 02115. E-mail ad-dress: The Journal of Immunology Copyright © 2006 by The American Association of Immunologists, Inc. 0022-1767/06/$02.00  pocket, secondary to shared binding specificities resulting from thepresence of an aspartic acid or a nonaspartic acid at codon 57, wereassociated with susceptibility to tuberculosis. In this study, weshow a highly significant association between homozygosity foralleles encoding HLA-DQ   57-Asp and active tuberculosis dis-ease. Furthermore, as compared with alleles which encode an ala-nine at this position, HLA-DQ   57-Asp binds a peptide from thecentral region of the  M. tuberculosis  immunogenic proteinESAT-6 with 5-fold less affinity and is significantly inferior toHLA-DQ   57-Ala in stimulating effector T cell responses whenpresenting this same peptide. Materials and Methods Study site and subjects The study patients were Cambodian individuals over 14 years old whowere randomly recruited between 1995 and 2002 from the CambodianHealth Committee tuberculosis treatment program in southeastern ruralCambodia (27). The diagnosis of clinical pulmonary tuberculosis was madeas described previously (8). A total of 436 HIV-1-negative pulmonary tu-berculosis patients were recruited in this study. Their mean age was 42.2  14.1 years, and 62.7% were female.Control subjects were recruited in 1995 from tuberculin-positive pa-tients visiting the same hospitals for minor complaints. Based on detailedclinical history, control subjects did not have a history of tuberculosis orcurrent signs or symptoms consistent with tuberculosis. A total of 117tuberculin-positive control subjects with no history of tuberculosis werefollowed up for at least 7 years to verify their control status. Ten controlsubjects who developed clinical tuberculosis disease over the 7-year fol-low-up period were excluded. The mean age of the remaining 107 controlindividuals studied was 37.5    12.9 years, and 56.1% were female. Allindividuals gave appropriate written consent and the study was approvedby the Institutional Review Boards at the CBR Institute for BiomedicalResearch and in Cambodia.All patients and control subjects were screened with tuberculin. Briefly,5 tuberculin units of Tubersol (Aventis-Pasteur) was injected intradermallyin the forearm of patients and controls and was evaluated for induration48 h later by a trained tuberculosis heath care worker (S.T.) under thesupervision of a physician with subspecialty training in infectious disease(A.E.G.). Patients were classified as tuberculin-positive if they had an in-duration of   10 mm as evaluated by the ballpoint method (28). Althoughbacillus Calmette-Gue´rin (BCG) 4 vaccination was not routinely performedin Cambodia before 1990 because of the social breakdown of the country,and thus, is not a usual cause of false-positive tuberculin results in adults,we did however screen all our control subjects and patients for the presenceof a BCG scar to rule out previous BCG vaccination. We note that we havepreviously demonstrated that BCG vaccination or infection by other envi-ronmental mycobacterial species are not common causes of false-positivetuberculin results in this population by correlating Ag-specific IFN-    as-says to  M. tuberculosis  Ags with tuberculin positivity (29).  HLA and gene polymorphism typing  HLA  alleles were identified in PCR-amplified products of exons 2 and 3(  HLA-A ,  -B , and  -C  w) and exon 2 (  HLA-DRB1  and  -DQB1 ) by sequence-specific oligonucleotide probe hybridization as previously described (17). Cell lines The EBV-transformed human B cell line TEM expressing  HLA- DQA1*0104  and  HLA-DQB1*0503  described in the 10th International His-tocompatibility Workshop (November 18–21, 1987, New York, NY) wasobtained from the ASHI Cell Repository (Minneapolis, MN). BLS-1 cells,an HLA class II-null EBV-transformed B cell line generated from the cellsof a patient with bare lymphocyte syndrome (30) was a gift from J. Blum(Indiana University School of Medicine, Indianapolis, IN). Human embry-onic kidney 293 T cells were a gift from R. Van Etten (Tufts-New EnglandMedical Center, Boston, MA). Murine fibroblast NIH-3T3 cells were pur-chased from the American Type Culture Collection. Generation of BLS-1 cell lines expressing HLA-DQ   57-Aspand HLA-DQ   57-Ala BLS-1 cell lines expressing HLA-DQA1*0104 and HLA-DQ   57-Asp or  57-Ala, were generated by retroviral-mediated gene transfer. Briefly, theHLA-DQA1*0104 and HLA-DQB1*0503 cDNAs were isolated fromTEM cells. In addition, a mutagenic primer was used to change the codon(GAC) specifying aspartic acid at position   57 to create an isogenic mol-ecule containing the codon specifying alanine (GCC) at this position. Thewild-type (HLA-DQ   57-Asp) and mutated (HLA-DQ   57-Ala) cDNAswere then cloned into the retroviral vector pLPCX and the HLA-DQA1*0104 cDNA was cloned into the pLXSN vector (BD Clontech). 293T cells (4  10 6 ) were plated in 60-mm plates and the next day, 10   g of the retroviral vector DNA containing the cDNA of interest was cotrans-fected by calcium phosphate precipitation with 5   g of MCV-amphopack,a gift from R. Van Etten. Titers of viral supernatants obtained from thetransfected 293 T cells were determined by transduction of NIH-3T3 cells,and colonies resistant to G-418 sulfate (pLXSN) or puromycin (pLPCX)were generated. All viruses had titers of 2.5–5.0    10 6 G-418 sulfate orpuromycin-resistant PFU/ml. Viral supernatants were then used to infectBLS-1 cells and 500   g/ml G-418 sulfate (pLXSN vector) or 200 ng/mlpuromycin (pLPCX vector) was used for selection. BLS-1 cell lines stablyexpressing HLA-DQ   57-Asp or   57-Ala were verified by FACS with theanti-HLA-DQ SPVL3 mAb (BD Pharmingen). Whole-cell peptide-binding assays ESAT-6, a secreted protein specific to the  M. tuberculosis  complex (31,32), is recognized by a large proportion of tuberculosis patients and tuber-culin-positive controls from a variety of populations, including Cambodia(29), Ethiopia (33), Kuwait (34), and the United States and Germany (35).CFP-10, another  M. tuberculosis -specific secreted protein, has also beenshown to elicit an immune response in tuberculosis patients (36, 37). Thus,peptides from these two proteins allowed us to compare the binding affinityof HLA-DQ   57-Asp and HLA-DQ   57-Ala alleles to peptides from im-munogenic  M. tuberculosis  proteins and their ability to elicit T cellactivation.Peptides spanning the length of the ESAT-6 and CFP-10 proteins weresynthesized by F-moc chemistry at Mixture Sciences. Each peptide was 15aa in length and overlapped the adjacent peptide by 10 residues. Purity(  95%) and identity of each peptide were characterized using mass spec-trometry. A total of 2  10 5 BLS-1 cells expressing HLA-DQA1*0104 andHLA-DQ   57-Asp or   57-Ala were incubated in 96-well plates with 100  M of the biotinylated peptide (VSKMRMATPLLMQAL) derived fromthe CLIP peptide sequence and with increasing concentrations (0.1–100  M) of nonbiotinylated specific competitor ESAT-6 or CFP-10 peptides at37°C for 4 h. At the end of the incubation, the cells were washed thenincubated for 30 min at 4°C with 10   g/ml FITC-avidin (Vector Labora-tories), washed again, followed by a 30-min incubation at 4°C with 10  g/ml biotinylated anti-avidin (Vector Laboratories), washed again andfinally followed by a 30-min incubation at 4°C with 10  g/ml FITC-avidin.Ultimately, the cells were resuspended in FACS buffer and analyzed by 4 Abbreviation used in this paper: BCG, bacillus Calmette-Gue´rin. FIGURE 1.  A model of the peptide-binding groove of an HLA-DQ mol-ecule showing pocket P9 formed by the side chains of amino acids encodedby the HLA-DQ    and    chains. Peptide-binding specificity of the P9pocket depends on the interaction between the invariant amino acid argi-nine at position 76 of the HLA DQ  -chain and whether there is an asparticacid (HLA-DQ  57-Asp) or a nonaspartic acid (HLA-DQ  57-non-Asp) atposition 57 of the HLA DQ   -chain. The figure illustrates that HLA-DQ   alleles encoding an aspartic acid at position 57 (including  HLA- DQB1*0503 , shown in red), result in a P9 pocket that has a neutral chargebecause the negatively charged aspartic acid forms a salt bridge with thepositively charged arginine at position 76 (shown in yellow). The HLA-DQ   alleles encoding HLA-DQ   57-non-Asp are also noted. 1091The Journal of Immunology  FACS. All binding experiments using different concentrations of compet-itor peptide were performed at least six times. The average concentrationat which 50% inhibition of the mean fluorescent intensity was obtainedwith the biotinylated CLIP peptide alone was determined for each nonbi-otinylated competitor peptide by plotting a curve for each peptide exam-ined and extrapolating from the curve the concentration of peptide at which50% inhibition of CLIP binding occurs.  Determination of HLA-DQ   57-Asp and    57-Ala T cell-restricted response to  M. tuberculosis  peptides After appropriate consent was obtained, PBMC were isolated from Cam-bodian individuals positive for  HLA-DQB1*0503  with prior history of tu-berculosis disease but without current clinical or laboratory evidence of active disease. CD4  T cell fractions were separated using paramagneticmethods (Miltenyi Biotec). The purity of each CD4  T cell separation wasanalyzed by FACS and was consistently   95%. The  HLA-DQ   57-Asp and   57-Ala  T cell-restricted response to  M. tuberculosis  peptides werethen studied by IFN-    ELISPOT assays. Briefly, 5    10 4 APC (BLS-1cells expressing  HLA-DQ   57-Asp  or   57-Ala ) were plated in Immuno-bilon-P membrane plates (Millipore) previously coated with IFN-    captureAb (BD Pharmingen). Where indicated, transduced BLS-1 cells werepulsed with 20   g/ml individual peptides for which we previously estab-lished an optimal saturating concentration to minimize variations in load-ing for T cell stimulation in ELISPOT assays (data not shown). Further-more, lower concentrations of peptides (5–15  g/ml) did not produce moresignificant differences in the production of IFN-    by T cells in response topeptide-loaded HLA-DQ   57-Asp- or   57-Ala-transduced BLS-1 cells.The cells were cultured at 37°C for 2 h and, subsequently, 2.5    10 5 CD4  T cells were added to the transduced Ag-presenting BLS-1 cells andincubated at 37°C for 16 h. After addition of IFN-    detection Ab (BDPharmingen), followed by the streptavidin-HRP conjugate (DakoCytoma-tion) HRP-substrate 3-amino-9-ethylcarbazole (BD Pharmingen), IFN-   spots were scored using a Series 1 ImmunoSpot Analyzer (Cellular Tech-nology). All conditions were performed in triplicate. To rule out T cellstimulation from self- or EBV-derived peptides from the EBV-transformedBLS-1 transfectants or the participation of HLA class I-restricted peptidepresentation as the stimulatory sources of IFN-   , we also cocultured CD4  T cells with HLA-DQ   57-Asp- or HLA-DQ   57-Ala-transduced cellswithout peptide or with CD4  T cells matched for HLA class I allelesexpressed on BLS-1 cells and found no detectable IFN-    spots (data notshown).Measurement of IFN-   levels in culture supernatants by ELISA gave thesame pattern of results as did IFN-    ELISPOT assays in our standardiza-tion experiments of the T cell stimulation assays (data not shown). How-ever, because IFN-    ELISPOT assays are a more sensitive measure of Tcell stimulation as compared with ELISA measurement of IFN-    levels inculture supernatants (38, 39), we used the ELISPOT method to determinethe HLA-DQ   57-Asp and   57-Ala T cell-restricted response to each oneof the peptides. Statistical analysis The statistical significance of the difference in the frequency of individualHLA loci between the two groups was calculated by the 2    2 Fisher’sexact test. The frequencies of homozygous HLA-DQ   57 genotypes wereeach compared with the more common heterozygous reference genotypeusing two separate 2    2    2 tests. Normally distributed variables wereanalyzed by the paired  t   test. Levels of significance are reported as two-tailed  p  values. A  p  value   0.05 was considered significant. These dataanalyses were performed with the aid of INSTAT software (GraphPad).Testing for the Hardy-Weinberg equilibrium was performed by use of SAS/Genetics Software. Results  Homozygosity for aspartic acid at HLA-DQ   57 is associated with susceptibility to pulmonary tuberculosis Among genotypes derived from 436 tuberculosis patients in thisstudy, we found that individuals homozygous for HLA-DQ   57-Asp alleles were at a significantly higher risk for developing pul-monary tuberculosis as compared with 107 tuberculin-positivecontrol individuals with no history or symptoms consistent withtuberculosis (  p  0.001; odds ratio, 3.05; 95% confidence interval,1.53–6.07) (Table I). Consistent with our previous study (17), wedetected 41  HLA-DQB1*0503  alleles among tuberculosis patientsand detected none in the 107 tuberculin-positive control individ-uals (  p  0.0002). Furthermore, we found no association betweenalleles of class II  HLA-DR  or class I  HLA-A , -  B , or - Cw  loci withsusceptibility or resistance to pulmonary tuberculosis (data notshown).Because  HLA-DQB1*0503  occurs with higher frequency inAsians in comparison with Caucasian and African-derived popu-lations (40), we also performed the analysis excluding HLA-DQ  57-Asp homozygous genotypes carrying an  HLA-DQB1*0503  al-lele to rule out that a Cambodian-specific HLA population markerwas influencing our finding of HLA-DQ   57-Asp homozygosityand susceptibility to pulmonary tuberculosis. When 26 HLA-DQ  57-Asp homozygous genotypes carrying an  HLA-DQB1*0503  al-lele were excluded from the analysis, the frequency of HLA-DQ  57-Asp homozygosity remained significantly increased in tuber-culosis patients (  p    0.02; odds ratio, 2.35; 95% confidence in-terval, 1.17–4.71), indicating the robustness of the association of HLA-DQ   57-Asp homozygosity and tuberculosis in thispopulation.  HLA-DQ   57-Asp impairs binding of tuberculosis-derived  peptides relative to HLA-DQ   57-Ala To determine the functional impact of this genetic associationupon binding of   M. tuberculosis  peptides and subsequent T cellstimulation, we constructed retroviral vectors expressing eitherHLA-DQ  57-Asp or HLA-DQ  57-Ala using retroviral-mediatedgene transfer into the HLA class II-null BLS-1 cell line. We iso-lated an HLA-DQ-   gene encoding aspartic acid at codon 57 and,using site-directed mutagenesis, changed the aspartic acid to analanine. Thus, these two cell lines expressed an isogenic HLAclass II DQ molecule differing by only a single amino acid atposition 57 of the HLA-DQ   -chain. We confirmed that HLA-DQ  57-Asp and HLA-DQ   57-Ala were successfully transduced andequivalently expressed into the BLS-1 cells by flow cytometricanalysis (Fig. 2  A ) and RT-PCR analysis of the DQ   -chain (Fig.2  B ) with subsequent sequencing of the PCR products (Fig. 2 C  ).To test the binding affinity of ESAT-6 and CFP-10 peptides, wefirst established that HLA-DQ   57-Ala- and HLA-DQ   57-Asp-transduced cells equivalently bound a peptide derived from the Table I.  Comparison of the genotype frequency of the presence or absence of aspartic acid at position 57 of HLA-DQ    chain a Polymorphism GenotypeTuberculosis ( n  436) Controls ( n  107)   2 Test Odds Ratio (95% CI)Number % Number % HLA-DQ   57 Non-Asp/non-Asp 135 31.0 41 38.3  p  0.98 0.98 (0.61–1.54)Non-Asp/Asp 187 42.9 55 51.4 100Asp/Asp 114 26.1 11 10.3  p  0.001 3.05 (1.53–6.07) a HLA-DQ    alleles with an aspartic acid (Asp) at position 57 in the cohort analyzed included:  HLA-DQB1*0301,  0303,  04  (  0401,  0402 ),  0503,  0601,  0602, and  0603.  HLA-DQ  alleles carrying non aspartic acid (non-Asp) at position 57 included:  HLA-DQB1*0201,  0202,  0302,  0501,  0502,  0604,  and  0605.  Frequenciesof the homozygous HLA-DQ   57 genotypes (non-Asp/non-Asp and Asp/Asp) were independently compared with the more common heterozygous reference genotype using a2  2    2 test. The control population was in Hardy-Weinberg equilibrium. CI denotes confidence interval. 1092 ASPARTIC ACID IN HLA-DQ    CODON 57 IS LINKED TO TUBERCULOSIS  CLIP protein (Fig. 3  A ). The CLIP peptide binds newly synthesizedHLA class II molecules and directs them to acidified intracellularvesicles where peptide loading occurs (reviewed in Ref. 41). Thus,our demonstration that the CLIP peptide binds equivalently toHLA-DQ   57-Ala- and HLA-DQ   57-Asp-expressing cells,shown in Fig. 3  A , allowed us to use this peptide to perform com-petition assays with  M. tuberculosis -derived peptides fromESAT-6 and CFP-10. Using increasing concentrations of thesepeptides, we were therefore able to calculate their binding affinitybased upon their ability to inhibit binding of prebound CLIP toHLA-DQ   57-Ala- and HLA-DQ   57-Asp-transduced cells.Increasing concentrations of 31 competitor peptides spanningESAT-6 and CFP-10 were thus incubated with HLA-DQ  57-Ala-or HLA-DQ   57-Asp-transduced cells and their ability to competewith CLIP for binding to the two cell lines was measured (Fig. 3  B ).Strikingly, when the ESAT 46–60  peptide was presented byHLA-DQ   57-Ala-transduced BLS-1 cells, 50% inhibition of CLIP binding was achieved at an  5-fold lower concentration of ESAT 46–60  than when the peptide was presented by HLA-DQ  57-Asp-transduced cells (Fig. 3,  B  and  C  ). We note that six otherpeptides (ESAT 31–45 , ESAT 36–50 , ESAT 41–55 , ESAT 51–65 ,CFP 16–30 , and CFP 41–55 ) minimally competed with CLIP for bind-ing to one or the other of the HLA-DQ molecules (Fig. 3,  B  and  D–I  ). Presentation of ESAT  46–60  by HLA-DQ   57-Asp vs HLA-DQ  57-Ala elicits significantly decreased IFN-    from CD4  T cells To determine the functional impact of this differential binding of   M. tuberculosis -derived peptides, we next isolated CD4  T cellsfrom six former pulmonary tuberculosis patients from Cambodiawho carried an  HLA-DQB1*0503  allele and cocultured theirCD4  T cells with HLA-DQ   57-Asp- or HLA-DQ   57-Ala-transduced cells that had been loaded with one of the seven pep-tides shown to have differential binding to HLA-DQ   57-Ala orHLA-DQ   57-Asp. As a measure of CD4  T cell activation, wecalculated IFN-    levels subsequent to peptide presentation by theHLA-DQ   57-Asp- or HLA-DQ   57-Ala-transduced cells.Presentation of ESAT 31–45 , ESAT 36–50 , ESAT 41–55 , ESAT 51–65 , CFP 16–40 , or CFP 41–55  peptides by HLA-DQ   57-Asp- orHLA-DQ  57-Ala-transduced cells resulted in an erratic pattern of IFN-    production in each of the six patients (Fig. 4), consistentwith the minor differences detected in the binding of these peptidesby HLA-DQ   57-Asp or HLA-DQ   57-Ala molecules. Remark-ably, however, in every case using CD4  T cells from each of thesix patients tested, we detected increased numbers of IFN-   -se-creting cells when they were stimulated with HLA-DQ   57-Ala-transduced cells loaded with ESAT 46–60  as compared with stim-ulation with HLA-DQ   57-Asp-transduced cells loaded withESAT 46–60  (Fig. 4). Furthermore, the difference in mean numberof IFN-    spots per 500,000 CD4  cells observed in CD4  T cellsstimulated with HLA-DQ   57-Ala (106.5    7.9) vs HLA-DQ  57-Asp (101.8  9.8) when the six patients were grouped (  p  0.049), reached the accepted critical value of statistical signifi-cance in biology (  p    0.049). Furthermore, the difference in themean number of IFN-    spots per 500,000 CD4  cells observed inthe CD4  T cells stimulated with HLA-DQ  57-Ala (106.5  7.9)vs HLA-DQ   57-Asp (101.8    9.8) in the six patients reachedstatistical significance (  p  0.049). We note that the presentationof the complete set of peptides spanning the length of the ESAT-6or CFP-10 proteins by the HLA-DQ   57-Ala or HLA-DQ   57-Asp cells elicited a nonstatistically significant increase in the num-ber of IFN-   -secreting T cells when stimulated with HLA-DQ  57-Ala cells as compared with HLA-DQ   57-Asp cells (  p   0.08). Peptides presented by HLA class II molecules are generatedfrom exogenous native proteins and only a few epitopes of specific FIGURE 2.  Expression of HLA-DQ molecules in BLS-1 cells.  A , FACSanalysis of HLA-DQ allelic proteins expressed on BLS-1 cells (HLA classII-null) before and after retroviral-mediated gene transfer as indicated inthe figure. Cells were stained with anti-HLA-DQ SPVL3 mAb demonstrat-ing that the BLS wild-type cells did not express HLA-DQ whereas theHLA-DQ   57-Asp and HLA-DQ   57-Ala did express surface HLA-DQ.  B , RT-PCR analysis of DQ    transcription. First-strand cDNA was syn-thesized from total RNA and the HLA-DQ   -chain was amplified usingPCR demonstrating that HLA-DQ   57-Asp and HLA-DQ   57-Ala tran-scripts were being synthesized in both transduced BLS-1 cells.   -actin wasamplified as an internal control. MW, DNA molecular weight marker. C  , Sequence chromatogram results of the HLA-DQ   -chain transcrip-tion products shown in  B  from the transduced BLS-1 cells demonstrat-ing that the appropriate allele contains either an alanine or an asparticacid at codon 57. 1093The Journal of Immunology  FIGURE 3.  Relatively reduced binding of HLA-DQ   57-Asp to the ESAT 46–60  peptide.  A , Binding of biotinylated CLIP peptide at various concen-trations to HLA-DQ   57-Ala- or   57-Asp-transduced cells. The cells were incubated with increasing concentrations of CLIP peptide and binding wasdetermined by FACS using fluorescent avidin. The dotted line shows the concentration of CLIP peptide (100   M) used for the peptide-binding competitionexperiments. MFI, mean fluorescent intensity.  B , The binding affinities of ESAT-6 and CFP-10 peptides to HLA-DQ   57-Asp- and HLA-DQ   57-Ala-transduced BLS-1 cells. IC 50  represents the concentration of unlabeled competitor  M. tuberculosis  (ESAT-6 or CFP-10) peptide (0.1–100   M) required toinhibit 50% of the binding of indicator CLIP peptide. Synthetic peptides ESAT 6–20 , ESAT 71–85 , CFP 1–15 , and CFP 31–45  were not tested due to poor qualitysynthesis.  C–I  , Binding of ESAT-6 and CFP-10 peptides to HLA-DQ  57-Ala- or  57-Asp-transduced cells at different concentrations expressed as percentbinding of the indicator CLIP peptide. The average concentration of each peptide at which 50% inhibition of CLIP binding occurs at 100   M is indicatedin the figure. 1094 ASPARTIC ACID IN HLA-DQ    CODON 57 IS LINKED TO TUBERCULOSIS
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks