HLA Class I Allelic Sequence and Conformation Regulate Leukocyte Ig-Like Binding

This information is current as Des C. Jones, Vasilis Kosmoliaptsis, Richard Apps, Nicolas of October 1, 2021. Lapaque, Isobel Smith, Azumi Kono, Chiwen Chang, Louise H. Boyle, Craig J. Taylor, John Trowsdale and Rachel L. Allen J Immunol 2011; 186:2990-2997; Prepublished online 26 January 2011; doi: 10.4049/jimmunol.1003078 Downloaded from http://www.jimmunol.org/content/186/5/2990

Supplementary http://www.jimmunol.org/content/suppl/2011/01/26/jimmunol.100307 http://www.jimmunol.org/ Material 8.DC1 References This article cites 86 articles, 34 of which you can access for free at: http://www.jimmunol.org/content/186/5/2990.full#ref-list-1

Why The JI? Submit online.

by guest on October 1, 2021 • Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2011 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

HLA Class I Allelic Sequence and Conformation Regulate Leukocyte Ig-Like Receptor Binding

Des C. Jones,* Vasilis Kosmoliaptsis,†,‡ Richard Apps,* Nicolas Lapaque,x Isobel Smith,* Azumi Kono,{ Chiwen Chang,* Louise H. Boyle,* Craig J. Taylor,† John Trowsdale,*,1 and Rachel L. Allen‖,1

Leukocyte Ig-like receptors (LILRs) are a family of innate immune receptors predominantly expressed by myeloid cells that can alter the Ag presentation properties of macrophages and dendritic cells. Several LILRs bind HLA class I. Altered LILR recognition due to HLA allelic variation could be a contributing factor in disease. We comprehensively assessed LILR binding to >90 HLA class I alleles. The inhibitory receptors LILRB1 and LILRB2 varied in their level of binding to different HLA alleles, correlating in some cases with specific amino acid motifs. LILRB2 displayed the weakest binding to HLA-B*2705, an allele genetically associated

with several autoimmune conditions and delayed progression of HIV infection. We also assessed the effect of HLA class I Downloaded from conformation on LILR binding. LILRB1 exclusively bound folded b2-microglobulin–associated class I, whereas LILRB2 bound both folded and free H chain forms. In contrast, the activating receptor LILRA1 and the soluble LILRA3 displayed a preference for binding to HLA-C free H chain. To our knowledge, this is the first study to identify the ligand of LILRA3. These findings support the hypothesis that LILR-mediated detection of unfolded versus folded MHC modulates immune responses during infection or inflammation. The Journal of Immunology, 2011, 186: 2990–2997. http://www.jimmunol.org/ uman leukocyte Ag class I direct the functions LILRB1 (ILT2/LIR1/CD85j) and LILRB2 (ILT4/LIR2/CD85d), of both adaptive and innate immunity through their rec- which recognize a wide range of classical and nonclassical HLA H ognition by the TCR and leukocyte receptor complex- class I proteins (8, 9, 14–22). The activating receptor LILRA1 has encoded receptors, which include members of the killer Ig-like been shown to bind HLA-B27, the product of which is associated receptor (KIR) and leukocyte Ig-like receptor (LILR) families. with diseases such as ankylosing spondylitis (23). LILRs expressed on professional APCs can influence adaptive Classical HLA class I proteins are highly polymorphic (24). immune responses (1) by modulating release and co- KIRs, expressed on NK cells and some subsets, recognize stimulatory receptor expression (2–6) subsets of HLA class I alleles. Their binding may be influenced LILRs are termed activating (LILRA) or inhibitory (LILRB) on also by peptide bound to class I (25–30). KIR variation, in con- by guest on October 1, 2021 the basis of their cytoplasmic domains. Inhibitory LILRs possess junction with that in HLA class I, is associated with diseases, a long cytoplasmic tail containing ITIMs (7–9), whereas activa- including viral infections (31, 32), autoimmunity (33, 34), and ting LILRs possess a short cytoplasmic tail and associate with the complications of pregnancy (35). The broad specificity of LILRB1 ε adaptor molecule Fc Rg (10, 11). The putative soluble protein and LILRB2 results from their interaction with the conserved b2- LILRA3 has no known signaling ability (12, 13). The best char- microglobulin (b2m) subunit (LILRB1) and/or the a3 domain of acterized members of the LILR family are the inhibitory receptors HLA class I (20, 36). Despite this, a study of four different HLA class I alleles indicated a range of affinities for LILRB1 and *Immunology Division, Department of Pathology, University of Cambridge, Cam- LILRB2 (21). Such differences in the affinity and avidity of ligand bridge CB2 1QP, United Kingdom; †Tissue Typing Laboratories, Addenbrookes Hos- binding can influence signaling through LILR and the activation pital, Cambridge, CB2 0QQ United Kingdom; ‡Department of Surgery, University of Cambridge, Addenbrooke’s Hospital, Cambridge, CB2 0QQ United Kingdom; xIn- of APCs bearing them (5). Consequently, variations in affinity for stitut National de la Recherche Agronomique, Unite´ Mixte de Recherche, 1319 individual HLA class I alleles might be detected by genetic as- { Micalis, Domaine de Vilvert, F-78352 Jouy-en-Josas, France; Department of Basic sociation of LILR with disease (37). Medical Science and Molecular Medicine, Tokai University School of Medicine, Isehara, Kanagawa 259-1143, Japan; and ‖Centre for Infection, St. George’s Hospital, Although HLA class I molecules usually require association with University of London, London SW17 0RE, United Kingdom b2m and antigenic peptide before they can progress to the cell 1J.T. and R.LA. are cosenior authors. surface (38), they can subsequently dissociate to generate open Received for publication September 17, 2010. Accepted for publication December confomers that lack peptide and/or b2m (39, 40). Unfolded HLA 23, 2010. class I molecules are a feature of activated lymphocytes (41–47) This work was supported by Arthritis Research UK Project Grant 17951, the Well- and their recognition and functions are of growing interest (45, come Trust, the Medical Research Council, and the National Institute for Health Research Cambridge Biomedical Research Centre. 48). The inhibitory receptor LILRB2 can bind to b2m-free forms of both the HLA-B27 allele as well as the nonclassical HLA-G Address correspondence and reprint requests to Dr. Des C. Jones, Immunology Di- vision, Department of Pathology, University of Cambridge, Tennis Court Road, Cam- molecule (20, 23). The activating receptor LILRA1 has also been bridge CB2 1QP, United Kingdom. E-mail address: [email protected] shown to engage b2m-free forms of HLA-B27 (23). The online version of this article contains supplemental material. We sought to perform what we believe to be the first compre- Abbreviations used in this article: FHC, free H chain; KIR, killer Ig-like receptor; hensive study of LILR binding to the products of different HLA LD, linkage disequilibrium; LILR, leukocyte Ig-like receptor; b2m, b2-microglobu- class I alleles using a panel of .90 single Ag beads (SABs) (49). lin; MFI, mean fluorescence intensity; SAB, single Ag bead. We also determined the binding of LILR to “unfolded” or b2m- Copyright Ó 2011 by The American Association of Immunologists, Inc. 0022-1767/11/$16.00 free forms of HLA class I. www.jimmunol.org/cgi/doi/10.4049/jimmunol.1003078 The Journal of Immunology 2991

Materials and Methods Ab blocking assay Cloning of LILR sequences and production of LILR-Fc DNA The HLA class I Abs W6/32 and HC10 were assessed for their ability to constructs block LILRB1, -B2, and -A1 binding to b2-associated and FHC forms of C*0602 expressed on stably transfected 721.221 cells. Concentrated su- Full-length LILRB1, -B2, and -A1 were amplified from macrophage cDNA pernatants containing 0.5 mM LILR-Fc fusion protein were incubated for using the primers listed in Supplemental Table I. RNA extraction and 1 h at 4˚C with PE-labeled F(ab9)2 goat anti-human Fc Ab (109-116-170; subsequent cDNA synthesis were performed as previously described (50). Jackson ImmunoResearch Laboratories) at a concentration of 5 mg/ml. All PCRs were performed using Phusion polymerase (Finnzymes) with the Meanwhile, 1 3 105 cells were incubated for 30 min at 4˚C with either following cycling parameters: 2 min at 98˚C followed by 11 cycles of 98˚C 1 mg W6/32 (Sigma-Aldrich), 10 ml HC10 supernatant, 1 mg IgG2a iso- for 10 s, 68˚C for 30 s, and 72˚C for 60 s, followed by 21 cycles of 98˚C type Ab (Sigma-Aldrich), or W6/32 and HC10 in combination and brought for 10 s, 65˚C for 30 s, and 72˚C for 60 s, followed by 10 cycles of 98˚C to a volume of 25 ml with PBS/2% BSA. Twenty-five microliters of the for 10 s, 60˚C for 30 s, and 72˚C for 60 s. PCRs were carried out on either LILR-Fc/secondary Ab solution was then added to the cell/Ab mixture and MJ Research (Reno, NV) Dyad DNA engines or MJ research PTC-200 incubated for a further 1 h at 4˚C. Cells were washed twice with PBS/2% thermal cyclers. PCR products were HindIII and XbaI (New England BSA followed by a final wash in PBS. The level of LILR-Fc binding was Biolabs) digested and then ligated into the p3xFLAG-CMV-9 vector measured using a FACScan flow cytometer (BD Biosciences). The same (Sigma-Aldrich) using a Rapid DNA ligation (Roche). The full protein procedure was performed on cells previously treated with pH 3.5 citrate coding sequence of LILRA3 was amplified from dendritic cells using buffer (0.131 M sodium citrate, 0.066M sodium phosphate, and 2% BSA) primers NK1076 and NK645 (Supplemental Table I) and cloned into for 30 s at 4˚C followed by two washes with PBS/2% BSA to increase the TOPO vector (Invitrogen) following the manufacturer’s recommended pro- level of HLA class I FHC. tocol. The extracellular coding region of each LILR was subsequently am- Statistical analysis plified from the cloned sequences by PCR using the appropriate primer pairs listed in Supplemental Table I. The extracellular sequence of LILRB3 was The relationship between LILR-Fc binding and level of b2m-associated Downloaded from amplified from a clone supplied by Professor Marco Colonna. PCR products HLA class I (as determined by W6/32) was assessed by non-linear re- 2 were ligated into the Signal plgplus vector containing the human IgG1-Fc gression (GraphPad Prism 5) generating a curve of best fit and R values. domain (R&D Systems) following a HindIII and XbaI restriction digest. Differences in LILR-Fc binding between groups of HLA class I alleles Sequences were verified by cycle sequencing using BigDye Terminator were statistically assessed using a two-tailed Mann–Whitney U test (Graph version 3.1 methodology (Applied Biosystems) and an Applied Biosys- pad 5). LILR MFI values were normalized against level of HLA class I by tems 3730xl DNA analyzer. division using W6/32 (for b2m-associated HLA) or HC10 (for FHC) MFIs prior to statistical analysis. Transfections and production of LILR-Fc fusion proteins http://www.jimmunol.org/ Prior to transfection, HEK293T cells were maintained in RPMI 1640 Results supplemented with 10% FBS and penicillin-streptomycin (50 U/ml) and Differential binding of LILRA1, LILRA3, LILRB1, and LILRB2 incubated at 37˚C with 5% atmospheric CO2. HEK293T cells were transfected with Signal plgplus vector containing LILR-Fc coding se- to HLA class I and allele products quence or an empty Signal plgplus control vector using jetPEI transfection On the basis of sequence similarities with LILRB1, the activating reagent (Polyplus Transfection), following the manufacturer’s recom- mended protocol. Culture media were replaced with fresh Pro293a serum- receptor LILRA1 and the soluble receptor LILRA3 are predicted free medium (Lonza) 24 h posttransfection. Supernatants were harvested to engage with HLA-class I (36). LILRA1 has been shown to bind 5 d posttransfection and concentrated using an Amicon Ultra 15 centrif- HLA B27 (23), but as yet no other alleles of HLA-class I have by guest on October 1, 2021 ugal filter column (Millipore) with a nominal molecular mass limit of 30 been investigated and no binding studies have been reported for kDa. The retentate was then washed twice with 10 ml PBS while on the filter column. LILRA3. Using Fc fusion proteins, we assessed the binding of The concentration of Fc protein was assessed by ELISA. LILRA1, LILRA3, LILRB1, LILRB2, and LILRB3 to beads coated with allomorphs of either HLA class I or HLA class II. MHC SAB binding assay LILRA1, LILRA3, LILRB1, and LILRB2 bound HLA class I but Fc-fusion proteins at a concentration of 1 mM were screened against not HLA class II. No binding of LILRB3 to either HLA class I or LABScreen HLA class I SAB (LS1A04 lot no. 005; One Lambda, Canoga HLA class II was detected (data not shown). The inability of Park, CA) and HLA class II-coated beads (LSM12 lot no. 008; One LILRB3 to bind HLA class I is in accordance with previous Lambda), according to the manufacturer’s standard protocol. The HLA studies and predictions (22, 23, 36). class I SAB panel contained HLA-A (n = 31), -B (n = 50), and -C (n = 16) alleles, the protein sequences of which are provided in the Supplemental SABs were used next to assess the influence of isotypic and al- Materials section (Supplemental Fig. 12). Binding was assessed on lotypic variation of HLA class I on LILR binding. The level of HLA a Luminex LABScan 100 (One Lambda), and the median fluorescence class I on each SAB was determined using the mAbs BBM.1 (anti- intensity (MFI) value was obtained. For each bead, the MFI values of each b2m) and W6/32 (a pan-HLA class I Ab). Their binding profiles LILR-Fc were normalized for background binding by subtracting the MFI correlated strongly (Supplemental Fig. 1). The binding of LILRA1, values of the Fc-negative control. Results for B*3701 and B*4701 were omitted due to their high background binding of LILRB3-Fc and the Fc- LILRA3, LILRB1, and LILRB2 to the product of each HLA class I negative control. allele was compared with that of W6/32 (Supplemental Fig. 2). Of To assess the level of b2m-associated HLA class I on each bead pop- all the receptors tested, LILRB1 binding correlated best with level ulation, the HLA class I-specific mouse IgG mAb W6/32 was used and Ab of HLA class I, as indicated by W6/32 reactivity (Fig. 1). However, binding was detected using PE-conjugated rabbit F(ab9) anti-mouse IgG 2 LILRB1 binding to HLA-A alleles displayed considerable vari- (both from AbD Serotec, Oxford, U.K.). The level of b2m was measured 193 194 using the mouse mAb BBM.1 (Santa Cruz Biotechnology). The presence ability (Fig. 1). HLA-A alleles with Ala and Val (numbers of HLA class I free H chain (FHC) was assessed using the mouse mAbs correspond to the mature protein) were significantly associated with HC10 (51), HCA2 (52) (provided by Prof. Hidde Ploegh, Cambridge, a lower level of LILRB1 binding (Fig. 1, Supplemental Fig. 6B). MA), and L31 (53, 54) (provided by Dr. Patrizio Giacomini, Rome, Italy). These two variable positions correspond to an established binding These FHC-specific Abs vary in their recognition of allelic subsets; alleles that carry Thr or Asn at residue 80 and Arg82 of the mature protein are site of LILRB1 (36). Two further variations, serine at position 207 recognized by HCA2 (55). L31 recognizes alleles with Phe67 or Tyr67 and glutamine at 253, are in almost complete linkage disequilibrium (prominently HLA-C alleles) (54), whereas HC10 binds strongly to alleles (LD) with Ala193 and Val194 and were also strongly associated with 62 carrying Arg (mainly HLA-B and -C alleles) (55). weaker binding (Fig. 1, Supplemental Fig. 6B). Alleles that carried To assess binding of LILR-Fc molecules to HLA class I FHC, b2m was removed from SABs by incubating the beads in pH 3.0 citrate buffer a serine at position 246 also displayed lower binding, although this (0.131 M sodium citrate, 0.066 M sodium phosphate, and 2% BSA) for 2 was less significant (Fig. 1, Supplemental Fig. 6B) and is most min, then washing twice in PBS supplemented with 1% BSA (56). likely due to linkage with the Ala193 and Val194 polymorphisms. 2992 LILR RECOGNITION OF HLA CLASS I

were unable to identify any HLA-A or -B residue positions that correlated with the overall binding pattern of LILRB2. However, the presence of a cysteine at residue 1 (Cys1) and/or an aspartic acid at position 9 (Asp9) of HLA-C correlated with significantly stronger LILRB2 binding (Supplemental Fig. 7). Residue 1 is located on an exposed surface distal from the b2m and LILR binding sites, whereas position 9 is located on the b-pleated sheet of the a1 domain. This residue influences peptide loading within the groove. Variation at positions 1 and 9 of class I did not in- fluence LILRB1 binding (Supplemental Fig. 7). LILRA1-Fc and LILRA3-Fc binding profiles did not correlate with levels of HLA class I on SABs as determined by W6/32 reactivity. Both of these LILRs displayed an overall preference for HLA-C (Fig. 3A,3B, Supplemental Fig. 6A), binding strongest to alleles that carry Asp9 (Supplemental Fig. 7B). Direct com- parison of the HLA class I binding profiles for LILRA1 and LILRA3 revealed a strong correlation (Fig. 3C), suggesting that these two receptors share highly similar HLA recognition sites.

LILRA1, LILRA3, and LILRB2 bind HLA class I FHC Downloaded from FIGURE 1. Binding of LILRB1 to HLA class I SABs. The level of Two members of the LILR family, LILRA1 and LILRB2, have LILRB1 binding correlates with the level of b m-associated HLA class I, 2 previously been shown to bind b m-free forms of the HLA-B27 as assessed using W6/32. Binding to products of a range of alleles from 2 individual HLA loci are as indicated. Binding curves were produced by allele and the nonclassical HLA-G allele (20, 23, 57). It is possible nonlinear regression. LILRB1 binding to HLA-A correlated with variation that FHCs of other HLA class I alleles that are present on at residues 193, 194, 207, 246, and 253 within the a3 region. activated lymphocytes (41–47) may also act as LILR ligands.

Levels of unfolded HLA-class I on SABs were determined using http://www.jimmunol.org/ the Abs HCA2, HC10, and L31, all of which bind unfolded H The presence of Val194, in the absence of the other polymorphisms, chains of allelic subsets of HLA-class I, as described in Materials as found in several HLA-B alleles (such as B*35, B*51, and B*58) and Methods (Supplemental Fig. 3). Binding of LILRA1-Fc, and most HLA-C alleles, did not influence LILRB1 recognition. LILRA3-Fc and LILRB2-Fc positively correlated with HC10 Ab None of the HLA-A polymorphisms correlating with altered staining, unlike LILRB1-Fc (Fig. 4A). To further assess LILR LILRB1 binding appeared to influence LILRB2 recognition (Sup- binding to FHCs of HLA class I, b2m was liberated from HLA plemental Fig. 6C). No other HLA variation correlated with altered class I trimeric complexes using acid treatment. Removal of b2m LILRB1 binding. was confirmed by the almost complete loss of W6/32 binding In contrast to LILRB1, there was a greater degree of variability (Supplemental Fig. 4). The presence of the remaining HLA class I by guest on October 1, 2021 in LILRB2 binding to SABs. This protein displayed strongest bind- FHC was confirmed by Ab staining (Supplemental Fig. 4). Of the ing to HLA-A and weakest binding to a subset of HLA-B alleles, including B*2705 and B5701 (Fig. 2, Supplemental Fig. 6A). We

FIGURE 3. Binding of LILRA1 and -A3 to HLA class I SABs. Binding of LILRA1 and LILRA3 was compared with that of W6/32 (A, B). Both LILRA1 and -A3 displayed a preference for HLA-C (red dots). The FIGURE 2. Binding of LILRB2 to HLA class I SABs. The level of binding patterns of these two LILR correlated strongly (C, left) in com- LILRB2 binding was compared with that of W6/32. The binding results parison with LILRB2 which correlated poorly with LILRA3 binding (C, from all the HLA alleles are displayed together and separated by loci. right). The Journal of Immunology 2993 Downloaded from

FIGURE 4. Differential binding of LILR to folded or unfolded HLA class I molecules. A, Binding of LILRB1, -B2, -A1, and -A3 was compared with the level of FHC on untreated SABs as assessed using the mAb HC10. B, Using SABs, LILR binding of b2m-associated (folded) HLA class I was compared with that of FHC denatured by acid treatment. http://www.jimmunol.org/ receptors tested, LILRA1, LILRA3, and LILRB2 all bound FHCs, LILRA1-Fc (and the negative control LILRB3-Fc) failed to bind whereas LILRB1 binding was abrogated by the loss of b2m (Sup- cells prior to acid treatment. plemental Fig. 5). Interestingly, LILRA1 and -A3 displayed in- The level of LILRB1-Fc binding diminished following acid creased binding to SABs following the removal of b2m, whereas treatment (reflecting the reduction in the level of b2m-associated LILRB2 binding was slightly reduced (Fig. 4B). HLA class I; Fig. 5B) and was completely abrogated by W6/32, whereas HC10 had little influence (Fig. 5A). In contrast, W6/32 Blocking with mAbs W6/32 and HC10 confirms differential had only a moderate effect on LILRB2-Fc binding to acid-treated LILR recognition of alternative forms of HLA class I

cells, whereas HC10 displayed a greater level of blocking (Fig. by guest on October 1, 2021 The Abs W6/32 and HC10 were used to block the interaction of 5A,5C). The blocking of LILRB2 binding by either W6/32 or LILR-Fc proteins to a 721.221-HLA-C*0602 transfectant, both HC10 is consistent with binding of this receptor to both b2m- before and after the partial removal of b2m (using a mild acid associated and FHC forms of HLA class I. The increased level treatment) to assess further the influence of HLA class I confor- of HC10-mediated blocking, following acid treatment, suggested mation on LILR binding. W6/32 dramatically reduced the binding that FHC was the dominant LILRB2 ligand on these cells. of LILRB1 and LILRB2-Fc to untreated cells (Fig. 5A,5C), in LILRA1 binding, which was predominantly to acid-treated cells, accordance with previous reports (12, 58), whereas HC10 had was greatly reduced by HC10 but was unaffected by W6/32 (Fig. a weak effect on LILRB2 and no effect on LILRB1-Fc binding. 5A,5C), suggesting that this receptor interacts with FHC.

FIGURE 5. Differential blocking of LILR binding by W6/32 and HC10 mAbs. W6/32, HC10, and an IgG2a isotype control were assessed for their ability to block LILR binding to 721.221 HLA-C*0602 transfectants both before and after mild acid treat- ment. W6/32 and HC10 were used both singly and in combination. Representative results are shown in A. The effect of acid treatment on the binding of W6/32 (red), HC10 (blue), and isotype mAb (black) are shown in B. Combined results of the blocking ex- periment are displayed in C. LILRA3 binding was not determined as it gave a high level of nonspecific binding to cells. 2994 LILR RECOGNITION OF HLA CLASS I

The ability of HC10 to block LILRB2 and -A1 binding to FHC Variation at residue 9 also appeared to influence the binding of suggests that some LILR contact sites are centered on the peptide- LILRA1, -A3, and -B2 to HLA-B and HLA-C FHC. All three binding groove, as the HC10 binding site is located on the edge of LILRs displayed strongest binding to HLA-B and -C alleles car- the empty peptide-binding groove within the a1 helix. rying Asp9 (namely HLA-B*0801, HLA-C*0602, C*0702, and C*1802) (Fig. 7B, Supplemental Fig. 8). However, the influence of HLA class I variation influences LILR binding of FHC 9 HLA-C Asp was not dependent on the removal of b2m by acid Using acid-treated SABs, LILRB2-Fc binding correlated well with treatment (Supplemental Fig. 7B). overall level of FHC (as assessed with HC10 binding) (Fig. 6A), The presence of a free cysteine at position 1 of HLA-C was also with a preference for the FHC of HLA-A compared with that of significantly associated with higher binding of LILRA1, -A3, and HLA-B and -C (Supplemental Fig. 8). Significantly, LILRA1 and -B2 (Fig. 7C), as it was for LILRB2 binding prior to acid treat- -A3 displayed marked preferential binding to HLA-C FHC (Fig. ment (Supplemental Fig. 7A). Free cysteines enable homodimer 6A, Supplemental Fig. 8). These locus-specific preferences were formation for HLA-G and HLA-B27 (23, 57) and the free cysteine consistent with those found prior to the acid treatment of SABs at position 1 of some HLA-C alleles may be acting in a similar (Supplemental Fig. 6A). The binding patterns of LILRA1 and manner. Dimerization of HLA-C could be important in in vivo LILRA3 to FHC of all HLA class I alleles tested were highly interactions by encouraging the clustering of HLA-C and, in doing comparable (Fig. 6B), providing further evidence that these so, maximize LILR binding, thereby increasing the avidity of the receptors share highly similar binding sites. interaction with the dimeric LILR-Fc molecules. The level of LILRB2 binding to HLA-A correlated with variation Interestingly, most alleles carrying Cys1 are in strong LD with at positions 9, 144, and 145 of the mature protein, as normalized to a polymorphism known as 235C, which is associated with HCA2 (Fig. 7A, Supplemental Fig. 9). This HLA-A variation did delayed onset to AIDS (59). Conversely, most alleles carrying Downloaded from not influence LILRB2 binding to SABs prior to acid treatment Gly1 display strong LD with the variant associated with rapid (Supplemental Fig. 10). Residue 9 is located on the floor of the onset (termed 235T). Consequently, LILRA1 bound significantly peptide-binding groove, whereas 144 and 145 are located on the higher to the FHC products of 235C-linked alleles (p = 0.038, a2 helix. The influence of these polymorphisms on LILR binding Mann–Whitney U test, GraphPad Prism 5) (Supplemental Fig. is further evidence of LILR contact sites outside of the charac- 11). LILRB2 and LILRA3 also displayed preference for 235C- terized b2m and a3 binding regions. linked alleles (p = 0.053) (Supplemental Fig. 11). http://www.jimmunol.org/ None of the polymorphic sites that altered LILR binding of FHC occurred within the recognition sites of the mAbs used for the purpose of normalization, namely HC10, L31, and HCA2, which all bind within the a1 helix. Consequently, it is unlikely that these findings are due to altered mAb binding.

Discussion by guest on October 1, 2021 We assessed the binding of four LILR proteins to a panel of classical HLA class I allelic products. LILR binding was clearly influenced by HLA allelic variation: LILRB1 exhibited a lower affinity for a subset of HLA-A alleles, LILRB2 displayed a lower affinity for several HLA-B alleles, including HLA*B2705, whereas LILRA1 and LILRA3 generally displayed a greater preference for HLA-C alleles, particularly following the removal of b2m by acid treatment. Additionally, HLA allelic variation located within the peptide-loading groove influenced LILR binding of FHC, altering LILRB2 binding to HLA-A alleles, perhaps accounting for the high binding of LILRB2, -A1, and -A3 to HLA-B*0801 FHC. Unusual recognition of individual HLA class I alleles by LILR may alter the overall balance between activating and inhibitory signals within immune cells and subsequently contribute to the development of disease. LILRB2 displayed considerably weaker binding to HLA-B*2705 on untreated SABs. B*2705 is geneti- cally associated with several autoimmune conditions, particularly ankylosing spondylitis (60). Thus, it is possible that the compar- atively weak interaction of HLA B*2705 with the inhibitory re- ceptor LILRB2 may influence immune activity during disease. For example, as LILRs have been shown to regulate cytokine secre- tion, reduced LILRB2-mediated inhibition of APCs could enhance the production of inflammatory such as TNF-a that FIGURE 6. The binding of LILRB2, -A1, and -A3 to acid treated HLA strongly contribute to the immunopathology of ankylosing spon- class I SABs. A, The level of LILR binding was compared with the overall dylitis. Conversely, an increased tendency for activation could be level of acid-treated FHC on each bead as assessed by HC10. LILRB2 binding correlated with levels of FHC, whereas LILRA1 and -A3 dis- of benefit in controlling infection: HLA-B27 is strongly associated played a preference for HLA-C. Only alleles that bound well to HC10 in with delayed onset of AIDS in HIV patients, as is HLA-B*5701 addition to either L31 or HCA2 were analyzed. B, The binding patterns of (61), another allele that displayed weak binding to LILRB2. These LILRA1 and -A3 correlated strongly with each other but not with LILRB2. weaker LILRB2–ligand interactions may promote more effective All SABs were used in this analysis. immune responses against HIV-infected cells. This is in contrast to The Journal of Immunology 2995 Downloaded from

FIGURE 7. HLA class I polymorphism influences LILR binding to FHC on acid-treated SABs. A, Variation at residues 9, 144, and 145 within the peptide-loading groove of HLA-A FHC influenced binding of LILRB2. LILRB2 binding was normalized to HCA2. Significance was calculated using a two-tailed Mann–Whitney U test (*p , 0.05, **p , 0.01, ***p , 0.001; ns = p . 0.05; GraphPad Prism 5). B, Variation at residue 9 of HLA-C FHC influenced LILRB2, -A1, and -A3. LILR binding was normalized to HC10. Similar results were obtained when normalizing against L31 binding (data not http://www.jimmunol.org/ shown). C, Presence of a free cysteine at residue 1 of HLA-C FHC correlated with increased binding of LILRB2, -A1, and -A3. LILR binding was normalized to HC10. Similar results were obtained when normalizing against L31 (data not shown).

HLA-B*3503, an allele associated with rapid onset of AIDS that haplotype that is linked to a number of autoimmune conditions has previously been found to bind LILRB2 more strongly than the and appears to be associated with an exaggerated Th2-type cyto- neutral HLA-B*3501 allele. The presence of HLA-B*3503 results kine response (62, 63). in greater inhibition and reduced dendritic cell responsiveness LILRA1 and LILRA3 displayed highly comparable binding (37). Further work will explore the link between the strength of patterns to HLA class I beads both before and after the removal of by guest on October 1, 2021 LILRB2 interaction with HLA-B27/57 and APC responses. b2m, suggesting that they share similar binding sites. LILRA3 is Variations in LILRB2 binding to HLA-B did not correlate predicted to encode a soluble protein (64, 65) and may therefore significantly with HLA polymorphism, which is in contrast to compete with LILRA1 to provide a negative regulatory effect, as LILRB1 binding patterns of HLA-A. It is possible that this vari- previously shown for a soluble form of LILRB1 (50). LILRA3 is ation resulted from other factors, such as the nature of the peptides not present on all leukocyte receptor complex haplotypes (64). presented by HLA class I and the levels of their stability as FHCs. LILRA3 deficiency has been associated with the development of The influence of bound peptide on LILRB2 binding has been Sjo¨gren’s syndrome and multiple sclerosis (66–68). demonstrated previously (5), which raises the possibility that Our results are consistent with differential LILR recognition of changes in the MHC-bound peptide repertoire in certain disease alternative forms of HLA class I: LILRB1 binding is solely re- states such as infection, malignancy, and cellular stress could lead stricted to b2m-associated HLA class I, LILRB2 is able to bind to altered LILR binding and consequent effects on immune func- both b2m-associated and FHC forms, whereas LILRA1 only tion. recognizes FHC HLA class I. The LILRB1 and -B2 profiles are in LILRA1 and -A3 displayed preference for FHC forms of specific accordance with their crystal structures, which revealed a b2m alleles, most of which belonged to the HLA-C locus (particularly requirement for LILRB1, but not LILRB2, binding (20, 36, 69). C*0602 and C*0702). Interestingly, most of those alleles are in The ability of LILRA1 and LILRB2 to bind FHCs of HLA*B2705 strong LD with a polymorphism termed 235C, located 35 kb has been reported previously (57, 70). LILRB2 also bound FHC upstream of the HLA-C locus (59). This polymorphism is asso- forms of HLA-G and HLA-Cw4 (20). In this study, we show that ciated with increased expression of HLA-C and is genetically recognition of HLA class I FHC is a feature of LILRA1, LILRA3, linked with delayed onset to AIDS (59). Alleles in LD with the and LILRB2. polymorphic form associated with rapid onset (termed 235T) The influence of polymorphism located within the peptide overall displayed weaker binding (Supplemental Fig. 11). How- binding groove on the binding of LILRB2, -A1, and -A3 to FHC ever, this correlation was not complete, as LILRA1 and LILRA3 would be consistent with LILR binding to this domain. This is bound relatively strongly to HLA-C*0702, an allele associated supported by the ability of peptide to alter LILRB2 recognition of with rapid onset. The observed differences in the level of binding HLA-B27 (5) and HC10 to abrogate LILR binding to FHC (the and the resulting modulation of LILRA1 signaling may contribute HC10 binding site is located within the empty peptide-binding to the relative protective or detrimental effects conveyed by some groove). Consequently, the preferential binding displayed by of these alleles in HIV infection. The potential functional impact LILRA1 and -A3 to FHC may be due to the lack of peptide within of high-affinity 235C alleles would be further enhanced by their the groove, rather than an absence of b2m. Crystal structure increased expression. HLA-B*0801 FHC also bound LILRA1 and analysis has solely focused on the interaction between the first and -A3 strongly. HLA-B*0801 occurs on the ancestral MHC 8.1 second Ig domains of LILR and the a3 and b2m regions of the 2996 LILR RECOGNITION OF HLA CLASS I class I complex (20, 36, 69). Additional interaction of the peptide- 12. Borges, L., M. L. Hsu, N. Fanger, M. Kubin, and D. Cosman. 1997. A family of human lymphoid and myeloid Ig-like receptors, some of which bind to MHC binding groove with the third and fourth Ig domains of LILRB2, class I molecules. J. Immunol. 159: 5192–5196. -A1, and -A3 cannot be ruled out, as the crystal structures of these 13. Colonna, M., H. Nakajima, F. Navarro, and M. Lo´pez-Botet. 1999. A novel domains have yet to be determined. family of Ig-like receptors for HLA class I molecules that modulate function of lymphoid and myeloid cells. J. Leukoc. Biol. 66: 375–381. LILR activity on professional APCs can strongly influence 14. Allan, D. S., M. Colonna, L. L. Lanier, T. D. Churakova, J. S. Abrams, adaptive immune responses (2, 3, 6, 71, 72). Consequently, LILR- S. A. Ellis, A. J. McMichael, and V. M. Braud. 1999. Tetrameric complexes of mediated recognition of FHC could represent a novel mechanism human histocompatibility leukocyte (HLA)-G bind to peripheral blood myelomonocytic cells. J. Exp. Med. 189: 1149–1156. of immune regulation by detection of altered conformation of HLA 15. Lepin, E. J., J. M. Bastin, D. S. Allan, G. Roncador, V. M. Braud, D. Y. Mason, class I complexes. Increased cell surface expression of FHC is as- P. A. van der Merwe, A. J. McMichael, J. I. Bell, S. H. Powis, and sociated with the activation of lymphocytes (41–47) and potentially C. A. O’Callaghan. 2000. Functional characterization of HLA-F and binding of HLA-F tetramers to ILT2 and ILT4 receptors. Eur. J. Immunol. 30: 3552–3561. APCs (73). HLA-C is particularly prone to forming FHC due to its 16. Chapman, T. L., A. P. Heikeman, and P. J. Bjorkman. 1999. The inhibitory re- unusual stability in the absence of b2m (54, 74–77), and this form of ceptor LIR-1 uses a common binding interaction to recognize class I MHC molecules and the viral homolog UL18. Immunity 11: 603–613. HLA-C is specifically upregulated during macrophage differentia- 17. Navarro, F., M. Llano, T. Bello´n, M. Colonna, D. E. Geraghty, and M. Lo´pez- tion (78). Increased levels of FHC influence HLA class I cluster- Botet. 1999. The ILT2(LIR1) and CD94/NKG2A NK cell receptors respectively ing on activated T cells (44, 79, 80). This clustering could enhance recognize HLA-G1 and HLA-E molecules co-expressed on target cells. Eur. J. Immunol. 29: 277–283. receptor recognition and consequent cellular activity (81). The 18. Li, D., L. Wang, L. Yu, E. C. Freundt, B. Jin, G. R. Screaton, and X. N. Xu. 2009. notion that increased FHC levels are symptomatic of cellular acti- Ig-like transcript 4 inhibits lipid antigen presentation through direct CD1d in- vation is complemented by the association of inflammatory disease teraction. J. Immunol. 182: 1033–1040. 19. Liang, S., V. Ristich, H. Arase, J. Dausset, E. D. Carosella, and A. Horuzsko. activity with serum b2m levels (e.g., see Ref. 82), presumably 2008. Modulation of dendritic cell differentiation by HLA-G and ILT4 requires Downloaded from arising from the dissociation of the HLA class I complex. the IL-6–STAT3 signaling pathway. Proc. Natl. Acad. Sci. USA 105: 8357–8362. LILR-mediated detection of FHC may not be limited to direct 20. Shiroishi, M., K. Kuroki, L. Rasubala, K. Tsumoto, I. Kumagai, E. Kurimoto, K. Kato, D. Kohda, and K. Maenaka. 2006. Structural basis for recognition of the cell–cell interactions: FHC is the dominant form of HLA class I nonclassical MHC molecule HLA-G by the leukocyte Ig-like receptor B2 on exosomes (83), and it constitutes the major form of soluble (LILRB2/LIR2/ILT4/CD85d). Proc. Natl. Acad. Sci. USA 103: 16412–16417. HLA in serum (shed from cells by the action of metalloproteases) 21. Shiroishi, M., K. Tsumoto, K. Amano, Y. Shirakihara, M. Colonna, V. M. Braud, D. S. Allan, A. Makadzange, S. Rowland-Jones, B. Willcox, et al. 2003. Human (46, 84–86). inhibitory receptors Ig-like transcript 2 (ILT2) and ILT4 compete with CD8 for Our findings demonstrate an influence of allelic variation and MHC class I binding and bind preferentially to HLA-G. Proc. Natl. Acad. Sci. http://www.jimmunol.org/ USA 100: 8856–8861. conformation of HLA class I on LILR binding. Further work is re- 22. Colonna, M., J. Samaridis, M. Cella, L. Angman, R. L. Allen, C. A. O’Callaghan, quired to assess the functional implications in immune responses. R. Dunbar, G. S. Ogg, V. Cerundolo, and A. Rolink. 1998. Human myelomonocytic cells express an inhibitory receptor for classical and nonclassical MHC class I molecules. J. Immunol. 160: 3096–3100. Disclosures 23. Allen, R. L., T. Raine, A. Haude, J. Trowsdale, and M. J. Wilson. 2001. Leu- The authors have no financial conflicts of interest. kocyte receptor complex-encoded immunomodulatory receptors show differing specificity for alternative HLA-B27 structures. J. Immunol. 167: 5543–5547. 24. Parham, P., E. J. Adams, and K. L. Arnett. 1995. The origins of HLA-A,B,C polymorphism. Immunol. Rev. 143: 141–180. References 25. Gumperz, J. E., V. Litwin, J. H. Phillips, L. L. Lanier, and P. Parham. 1995. The 1. Anderson, K. J., and R. L. Allen. 2009. Regulation of T-cell immunity by leu- Bw4 public of HLA-B molecules confers reactivity with by guest on October 1, 2021 cocyte immunoglobulin-like receptors: innate immune receptors for self on clones that express NKB1, a putative HLA receptor. J. Exp. Med. 181: 1133–1144. antigen-presenting cells. Immunology 127: 8–17. 26. Winter, C. C., J. E. Gumperz, P. Parham, E. O. Long, and N. Wagtmann. 1998. 2. Beinhauer, B. G., J. M. McBride, P. Graf, E. Pursch, M. Bongers, M. Rogy, Direct binding and functional transfer of NK cell inhibitory receptors reveal U. Korthauer, J. E. de Vries, G. Aversa, and T. Jung. 2004. Interleukin 10 reg- novel patterns of HLA-C allotype recognition. J. Immunol. 161: 571–577. ulates cell surface and soluble LIR-2 (CD85d) expression on dendritic cells 27. Kollnberger, S., A. Chan, M. Y. Sun, L. Y. Chen, C. Wright, K. di Gleria, resulting in T cell hyporesponsiveness in vitro. Eur. J. Immunol. 34: 74–80. A. McMichael, and P. Bowness. 2007. Interaction of HLA-B27 homodimers with 3. Chang, C. C., R. Ciubotariu, J. S. Manavalan, J. Yuan, A. I. Colovai, F. Piazza, KIR3DL1 and KIR3DL2, unlike HLA-B27 heterotrimers, is independent of the S. Lederman, M. Colonna, R. Cortesini, R. Dalla-Favera, and N. Suciu-Foca. sequence of bound peptide. Eur. J. Immunol. 37: 1313–1322. 2002. Tolerization of dendritic cells by T(S) cells: the crucial role of inhibitory 28. Malnati, M. S., M. Peruzzi, K. C. Parker, W. E. Biddison, E. Ciccone, receptors ILT3 and ILT4. Nat. Immunol. 3: 237–243. A. Moretta, and E. O. Long. 1995. Peptide specificity in the recognition of MHC 4. Tenca, C., A. Merlo, E. Merck, E. E. Bates, D. Saverino, R. Simone, D. Zarcone, class I by natural killer cell clones. Science 267: 1016–1018. G. Trinchieri, C. E. Grossi, and E. Ciccone. 2005. CD85j (leukocyte Ig-like 29. Peruzzi, M., N. Wagtmann, and E. O. Long. 1996. A p70 killer cell inhibitory receptor-1/Ig-like transcript 2) inhibits human osteoclast-associated receptor- receptor specific for several HLA-B allotypes discriminates among peptides mediated activation of human dendritic cells. J. Immunol. 174: 6757–6763. bound to HLA-B*2705. J. Exp. Med. 184: 1585–1590. 5. Lichterfeld, M., D. G. Kavanagh, K. L. Williams, B. Moza, S. K. Mui, T. Miura, 30. Peruzzi, M., K. C. Parker, E. O. Long, and M. S. Malnati. 1996. Peptide se- R. Sivamurthy, R. Allgaier, F. Pereyra, A. Trocha, et al. 2007. A viral CTL es- quence requirements for the recognition of HLA-B*2705 by specific natural killer cells. J. Immunol. 157: 3350–3356. cape mutation leading to immunoglobulin-like transcript 4-mediated functional 31. Zu´n˜iga, J., V. Romero, J. Azocar, D. Terreros, M. I. Vargas-Rojas, D. Torres- inhibition of myelomonocytic cells. J. Exp. Med. 204: 2813–2824. Garcı´a, L. Jime´nez-Alvarez, G. Vargas-Alarco´n, J. Granados-Montiel, Z. Husain, 6. Young, N. T., E. C. Waller, R. Patel, A. Roghanian, J. M. Austyn, and et al. 2009. Protective KIR-HLA interactions for HCV infection in intravenous J. Trowsdale. 2008. The inhibitory receptor LILRB1 modulates the differentia- drug users. Mol. Immunol. 46: 2723–2727. tion and regulatory potential of human dendritic cells. Blood 111: 3090–3096. 32. Carrington, M., M. P. Martin, and J. van Bergen. 2008. KIR-HLA intercourse in 7. Cella, M., C. Dohring, J. Samaridis, M. Dessing, M. Brockhaus, A. Lanzavecchia, ¨ HIV disease. Trends Microbiol. 16: 620–627. and M. Colonna. 1997. A novel inhibitory receptor (ILT3) expressed on monocytes, 33. Suzuki, Y., Y. Hamamoto, Y. Ogasawara, K. Ishikawa, Y. Yoshikawa, macrophages, and dendritic cells involved in antigen processing. J. Exp. Med. 185: T. Sasazuki, and M. Muto. 2004. Genetic polymorphisms of killer cell 1743–1751. immunoglobulin-like receptors are associated with susceptibility to psoriasis 8. Colonna, M., F. Navarro, T. Bello´n, M. Llano, P. Garcı´a, J. Samaridis, vulgaris. J. Invest. Dermatol. 122: 1133–1136. L. Angman, M. Cella, and M. Lo´pez-Botet. 1997. A common inhibitory receptor 34. Jones, D. C., R. S. Edgar, T. Ahmad, J. R. Cummings, D. P. Jewell, J. Trowsdale, for major histocompatibility complex class I molecules on human lymphoid and and N. T. Young. 2006. Killer Ig-like receptor (KIR) genotype and HLA ligand myelomonocytic cells. J. Exp. Med. 186: 1809–1818. combinations in ulcerative colitis susceptibility. Immun. 7: 576–582. 9. Fanger, N. A., D. Cosman, L. Peterson, S. C. Braddy, C. R. Maliszewski, and 35. Hiby, S. E., L. Regan, W. Lo, L. Farrell, M. Carrington, and A. Moffett. 2008. L. Borges. 1998. The MHC class I binding proteins LIR-1 and LIR-2 inhibit Fc Association of maternal killer-cell immunoglobulin-like receptors and parental receptor-mediated signaling in monocytes. Eur. J. Immunol. 28: 3423–3434. HLA-C genotypes with recurrent miscarriage. Hum. Reprod. 23: 972–976. 10. Cao, W., D. B. Rosen, T. Ito, L. Bover, M. Bao, G. Watanabe, Z. Yao, L. Zhang, 36. Willcox, B. E., L. M. Thomas, and P. J. Bjorkman. 2003. Crystal structure of L. L. Lanier, and Y. J. Liu. 2006. Plasmacytoid dendritic cell-specific receptor HLA-A2 bound to LIR-1, a host and viral major histocompatibility complex ILT7-FcεRIg inhibits Toll-like receptor-induced production. J. Exp. receptor. Nat. Immunol. 4: 913–919. Med. 203: 1399–1405. 37. Huang, J., J. J. Goedert, E. J. Sundberg, T. D. Cung, P. S. Burke, M. P. Martin, 11. Nakajima, H., J. Samaridis, L. Angman, and M. Colonna. 1999. Human myeloid L. Preiss, J. Lifson, M. Lichterfeld, M. Carrington, and X. G. Yu. 2009. HLA- cells express an activating ILT receptor (ILT1) that associates with B*35-Px-mediated acceleration of HIV-1 infection by increased inhibitory im- g-chain. J. Immunol. 162: 5–8. munoregulatory impulses. J. Exp. Med. 206: 2959–2966. The Journal of Immunology 2997

38. Groothuis, T. A., A. C. Griekspoor, J. J. Neijssen, C. A. Herberts, and 8.1 ancestral haplotype (A1, B8, DR3) with multiple immunopathological dis- J. J. Neefjes. 2005. MHC class I alleles and their exploration of the antigen- eases. Immunol. Rev. 167: 257–274. processing machinery. Immunol. Rev. 207: 60–76. 64. Torkar, M., A. Haude, S. Milne, S. Beck, J. Trowsdale, and M. J. Wilson. 2000. 39. Gille, A. C., B. Pe´rarnau, P. Mercier, and F. A. Lemonnier. 1990. Serological Arrangement of the ILT cluster: a common null allele of the ILT6 gene analysis of the dissociation process of HLA-B and C class I molecules. Eur. J. results from a 6.7-kbp deletion. Eur. J. Immunol. 30: 3655–3662. Immunol. 20: 759–764. 65. Wilson, M. J., M. Torkar, A. Haude, S. Milne, T. Jones, D. Sheer, S. Beck, and 40. Carreno, B. M., and T. H. Hansen. 1994. Exogenous peptide ligand influences J. Trowsdale. 2000. Plasticity in the organization and sequences of human KIR/ the expression and half-life of free HLA class I heavy chains ubiquitously ILT gene families. Proc. Natl. Acad. Sci. USA 97: 4778–4783. detected at the cell surface. Eur. J. Immunol. 24: 1285–1292. 66. Kabalak, G., S. B. Dobberstein, T. Matthias, S. Reuter, Y. H. The, T. Do¨rner, 41. Schnabl, E., H. Stockinger, O. Majdic, H. Gaugitsch, I. J. Lindley, D. Maurer, R. E. Schmidt, and T. Witte. 2009. Association of immunoglobulin-like tran- A. Hajek-Rosenmayr, and W. Knapp. 1990. Activated human T lymphocytes script 6 deficiency with Sjo¨gren’s syndrome. Arthritis Rheum. 60: 2923–2925. express MHC class I heavy chains not associated with b2-microglobulin. J. Exp. 67. Ordo´n˜ez, D., A. J. Sa´nchez, J. E. Martı´nez-Rodrı´guez, E. Cisneros, E. Ramil, Med. 171: 1431–1442. N. Romo, M. Moraru, E. Munteis, M. Lo´pez-Botet, J. Roquer, et al. 2009. 42. Madrigal, J. A., M. P. Belich, R. J. Benjamin, A. M. Little, W. H. Hildebrand, Multiple sclerosis associates with LILRA3 deletion in Spanish patients. Genes D. L. Mann, and P. Parham. 1991. Molecular definition of a polymorphic antigen Immun. 10: 579–585. (LA45) of free HLA-A and -B heavy chains found on the surfaces of activated B 68. Koch, S., R. Goedde, V. Nigmatova, J. T. Epplen, N. Mu¨ller, J. de Seze, and T cells. J. Exp. Med. 174: 1085–1095. P. Vermersch, T. Momot, R. E. Schmidt, and T. Witte. 2005. Association of 43. Demaria, S., R. Schwab, and Y. Bushkin. 1992. The origin and fate of b2m-free multiple sclerosis with ILT6 deficiency. Genes Immun. 6: 445–447. MHC class I molecules induced on activated T cells. Cell. Immunol. 142: 103– 69. Willcox, B. E., L. M. Thomas, T. L. Chapman, A. P. Heikema, A. P. West, Jr., 113. and P. J. Bjorkman. 2002. Crystal structure of LIR-2 (ILT4) at 1.8 A˚ : differences 44. Matko, J., Y. Bushkin, T. Wei, and M. Edidin. 1994. Clustering of class I HLA from LIR-1 (ILT2) in regions implicated in the binding of the human cyto- molecules on the surfaces of activated and transformed human cells. J. Immunol. megalovirus class I MHC homolog UL18. BMC Struct. Biol. 2: 6. 152: 3353–3360. 70. Allen, R. L., and J. Trowsdale. 2004. Recognition of classical and heavy chain 45. Arosa, F. A., S. G. Santos, and S. J. Powis. 2007. Open conformers: the hidden forms of HLA-B27 by leukocyte receptors. Curr. Mol. Med. 4: 59–65. face of MHC-I molecules. Trends Immunol. 28: 115–123. 71. Suciu-Foca, N., and R. Cortesini. 2007. Central role of ILT3 in the T suppressor 46. Pickl, W. F., W. Holter, J. Sto¨ckl, O. Majdic, and W. Knapp. 1996. Expression of cell cascade. Cell. Immunol. 248: 59–67. Downloaded from LA45 reactive b2-microglobulin free HLA class I a-chains on activated T-cells is 72. Wagner, C. S., L. Walther-Jallow, E. Buentke, H. G. Ljunggren, A. Achour, and regulated by internalization, constitutive and protein kinase C inducible release. B. J. Chambers. 2008. Human cytomegalovirus-derived protein UL18 alters the Tissue 48: 15–21. phenotype and function of monocyte-derived dendritic cells. J. Leukoc. Biol. 83: 47. Pickl, W. F., W. Holter, J. Sto¨ckl, O. Majdic, and W. Knapp. 1996. Expression of 56–63. b2-microglobulin-free HLA class I a-chains on activated T cells requires in- 73. Santos, S. G., S. Lynch, E. C. Campbell, A. N. Antoniou, and S. J. Powis. 2008. ternalization of HLA class I heterodimers. Immunology 88: 104–109. Induction of HLA-B27 heavy chain homodimer formation after activation in 48. Raine, T., and R. Allen. 2005. MHC-I recognition by receptors on myelomo- dendritic cells. Arthritis Res. Ther. 10: R100. nocytic cells: new tricks for old dogs? Bioessays 27: 542–550. 74. Sibilio, L., A. Martayan, A. Setini, E. Lo Monaco, E. Tremante, R. H. Butler, and

49. Apps, R., S. P. Murphy, R. Fernando, L. Gardner, T. Ahad, and A. Moffett. 2009. P. Giacomini. 2008. A single bottleneck in HLA-C assembly. J. Biol. Chem. 283: http://www.jimmunol.org/ Human leucocyte antigen (HLA) expression of primary trophoblast cells and 1267–1274. placental cell lines, determined using single antigen beads to characterize allo- 75. Martayan, A., M. Fiscella, A. Setini, G. Ciccarelli, R. Gambari, G. Feriotto, type specificities of anti-HLA antibodies. Immunology 127: 26–39. A. Beretta, A. G. Siccardi, E. Appella, and P. Giacomini. 1997. Conformation 50. Jones, D. C., A. Roghanian, D. P. Brown, C. Chang, R. L. Allen, J. Trowsdale, and surface expression of free HLA-CW1 heavy chains in the absence of b2- and N. T. Young. 2009. Alternative mRNA splicing creates transcripts encoding microglobulin. Hum. Immunol. 53: 23–33. soluble proteins from most LILR genes. Eur. J. Immunol. 39: 3195–3206. 76. Martayan, A., R. Fraioli, E. Giorda, A. Setini, G. Ciccarelli, L. Delfino, 51. Stam, N. J., H. Spits, and H. L. Ploegh. 1986. Monoclonal antibodies raised G. B. Ferrara, and P. Giacomini. 1999. Biosynthesis of HLA-C heavy chains in against denatured HLA-B locus heavy chains permit biochemical characteriza- melanoma cells with multiple defects in the expression of HLA-A, -B, -C tion of certain HLA-C locus products. J. Immunol. 137: 2299–2306. molecules. Br. J. Cancer 80: 639–649. 52. Seitz, C., B. Uchanska-Ziegler, A. Zank, and A. Ziegler. 1998. The monoclonal 77. Giacomini, P., A. Beretta, M. R. Nicotra, G. Ciccarelli, A. Martayan, C. Cerboni, antibody HCA2 recognises a broadly shared epitope on selected classical as well L. Lopalco, D. Bini, L. Delfino, G. B. Ferrara, et al. 1997. HLA-C heavy chains

as several non-classical HLA class I molecules. Mol. Immunol. 35: 819–827. free of b2-microglobulin: distribution in normal tissues and neoplastic lesions of by guest on October 1, 2021 53. Grassi, F., R. Meneveri, M. Gullberg, L. Lopalco, G. B. Rossi, P. Lanza, C. De non-lymphoid origin and interferon-g responsiveness. Tissue Antigens 50: 555– Santis, G. Brattsand, S. Butto`, E. Ginelli, et al. 1991. Human immunodeficiency 566. virus type 1 gp120 mimics a hidden monomorphic epitope borne by class I major 78. Schaefer, M. R., M. Williams, D. A. Kulpa, P. K. Blakely, A. Q. Yaffee, and histocompatibility complex heavy chains. J. Exp. Med. 174: 53–62. K. L. Collins. 2008. A novel trafficking signal within the HLA-C cytoplasmic 54. Setini, A., A. Beretta, C. De Santis, R. Meneveri, A. Martayan, M. C. Mazzilli, tail allows regulated expression upon differentiation of macrophages. J. Immu- E. Appella, A. G. Siccardi, P. G. Natali, and P. Giacomini. 1996. Distinctive nol. 180: 7804–7817. features of the a1-domain a helix of HLA-C heavy chains free of b2-micro- 79. Bodna´r, A., Z. Bacso´, A. Jenei, T. M. Jovin, M. Edidin, S. Damjanovich, and globulin. Hum. Immunol. 46: 69–81. J. Matko´. 2003. Class I HLA oligomerization at the surface of B cells is con- 55. Sernee, M. F., H. L. Ploegh, and D. J. Schust. 1998. Why certain antibodies trolled by exogenous b2-microglobulin: implications in activation of cytotoxic cross-react with HLA-A and HLA-G: epitope mapping of two common MHC T lymphocytes. Int. Immunol. 15: 331–339. class I reagents. Mol. Immunol. 35: 177–188. 80. Bodna´r, A., A. Jenei, L. Bene, S. Damjanovich, and J. Matko´. 1996. Modification 56. Apps, R., L. Gardner, S. E. Hiby, A. M. Sharkey, and A. Moffett. 2008. Con- of membrane cholesterol level affects expression and clustering of class I HLA formation of human leucocyte antigen-C molecules at the surface of human molecules at the surface of JY human lymphoblasts. Immunol. Lett. 54: 221–226. trophoblast cells. Immunology 124: 322–328. 81. Achdout, H., I. Manaster, and O. Mandelboim. 2008. Influenza virus infection 57. Kollnberger, S., L. Bird, M. Y. Sun, C. Retiere, V. M. Braud, A. McMichael, and augments NK cell inhibition through reorganization of major histocompatibility P. Bowness. 2002. Cell-surface expression and recognition of complex class I proteins. J. Virol. 82: 8030–8037. HLA-B27 homodimers. Arthritis Rheum. 46: 2972–2982. 82. Gottenberg, J. E., M. Busson, J. Cohen-Solal, F. Lavie, K. Abbed, 58. Gonen-Gross, T., H. Achdout, R. Gazit, J. Hanna, S. Mizrahi, G. Markel, R. P. Kimberly, J. Sibilia, and X. Mariette. 2005. Correlation of serum D. Goldman-Wohl, S. Yagel, V. Horejsı´, O. Levy, et al. 2003. Complexes of B lymphocyte stimulator and b2 microglobulin with autoantibody secretion and HLA-G protein on the cell surface are important for leukocyte Ig-like receptor-1 systemic involvement in primary Sjo¨gren’s syndrome. Ann. Rheum. Dis. 64: function. J. Immunol. 171: 1343–1351. 1050–1055. 59. Thomas, R., R. Apps, Y. Qi, X. Gao, V. Male, C. O’hUigin, G. O’Connor, D. Ge, 83. Lynch, S., S. G. Santos, E. C. Campbell, A. M. Nimmo, C. Botting, A. Prescott, J. Fellay, J. N. Martin, et al. 2009. HLA-C cell surface expression and control of A. N. Antoniou, and S. J. Powis. 2009. Novel MHC class I structures on exo- HIV/AIDS correlate with a variant upstream of HLA-C. Nat. Genet. 41: 1290– somes. J. Immunol. 183: 1884–1891. 1294. 84. Haynes, L. D., Y. Bushkin, R. B. Love, and W. J. Burlingham. 2002. Interferon-g 60. Gaston, H. 2006. Mechanisms of disease: the immunopathogenesis of spondy- drives the metalloproteinase-dependent cleavage of HLA class I soluble forms loarthropathies. Nat. Clin. Pract. Rheumatol. 2: 383–392. from primary human bronchial epithelial cells. Hum. Immunol. 63: 893–901. 61. Carrington, M., and S. J. O’Brien. 2003. The influence of HLA genotype on 85. Demaria, S., L. D. DeVito-Haynes, R. D. Salter, W. J. Burlingham, and AIDS. Annu. Rev. Med. 54: 535–551. Y. Bushkin. 1999. Peptide-conformed b2m-free class I heavy chains are inter- 62. Candore, G., D. Lio, G. Colonna Romano, and C. Caruso. 2002. Pathogenesis of mediates in generation of soluble HLA by the membrane-bound metal- autoimmune diseases associated with 8.1 ancestral haplotype: effect of multiple loproteinase. Hum. Immunol. 60: 1216–1226. gene interactions. Autoimmun. Rev. 1: 29–35. 86. Demaria, S., R. Schwab, S. R. Gottesman, and Y. Bushkin. 1994. Soluble b2- 63. Price, P., C. Witt, R. Allcock, D. Sayer, M. Garlepp, C. C. Kok, M. French, microglobulin-free class I heavy chains are released from the surface of activated S. Mallal, and F. Christiansen. 1999. The genetic basis for the association of the and leukemia cells by a metalloprotease. J. Biol. Chem. 269: 6689–6694.