Genetic Associations and Functional Characterization of M1 Aminopeptidases and Immune-Mediated Diseases

REVIEW Genetic associations and functional characterization of M1 aminopeptidases and immune-mediated diseases

N Agrawal and MA Brown

Endosplasmic reticulum aminopeptidase 1 (ERAP1), endoplasmic reticulum aminopeptidase 2 (ERAP2) and puromycin-sensitive aminopeptidase (NPEPPS) are key zinc metallopeptidases that belong to the oxytocinase subfamily of M1 aminopeptidase family. NPEPPS catalyzes the processing of proteosome-derived peptide repertoire followed by trimming of antigenic peptides by ERAP1 and ERAP2 for presentation on major histocompatibility complex (MHC) Class I molecules. A series of genome-wide association studies have demonstrated associations of these aminopeptidases with a range of immune-mediated diseases such as ankylosing spondylitis, psoriasis, Behçet’s disease, inﬂammatory bowel disease and type I diabetes, and signiﬁcantly, genetic interaction between some aminopeptidases and HLA Class I loci with which these diseases are strongly associated. In this review, we highlight the current state of understanding of the genetic associations of this class of genes, their functional role in disease, and potential as therapeutic targets.

Genes and Immunity (2014) 15, 521–527; doi:10.1038/gene.2014.46; published online 21 August 2014

INTRODUCTION and apoptosis. Aminopeptidases hydrolyze a polypeptide’s It has long been established that the ‘seronegative’ diseases terminal amino-acid peptide bond, leading to the degradation ankylosing spondylitis (AS), psoriasis, Behçet’s disease and of their target substrate and release of amino acids which can be inﬂammatory bowel disease (IBD) share a tendency to co- recycled to form other proteins in the body. segregate in families, as well as various clinical and genetic In mammals, the M1 aminopeptidase family consists of ﬁve features, such as (with the exception of IBD) association with HLA integral transmembrane proteins: aminopeptidase A, thyrotropin- Class I loci, and association with axial spondyloarthritis. Genome- releasing hormone-degrading ectoenzyme, aminopeptidase N, wide association studies are identifying increasing evidence of ERAP1 and Leucyl/Cystinyl Aminopeptidase (LNPEP); and four non- broader genetic similarity between these diseases, notably with transmembrane enzymes: aminopeptidase B, NPEPPS, Leukotri- shared associations in the IL23R pathway, and more recently in ence A4 and ERAP2.6 This review focuses on the aminopeptidases associations with genes encoding M1 aminopeptidases, such as ERAP1, ERAP2 and NPEPPS that have been genetically associated endoplasmic reticulum aminopeptidase 1 (ERAP1), endoplasmic with human disease. reticulum aminopeptidase 2 (ERAP2) and NPEPPS (puromycin- ERAP1 is a type II integral membrane enzyme, whereas sensitive aminopeptidase). This review discusses the genetics of ERAP2 and NPEPPS are non-transmembrane enzymes. ERAP1 this class of genes and their association with immune-mediated and ERAP2 share ~ 49% sequence homology, LNPEP and ERAP1 diseases (IMDs), their functional role in disease and promise as ~ 44%, whereas NPEPPS shares ~ 40% sequence homology to therapeutic targets. the three aminopeptidases. The protein sequence alignment (Supplementary Figure S1) reveals the presence of a GAMEN and HELAH motifs that are conserved in six aminopeptidases except M1 AMINOPEPTIDASE FAMILY thyrotropin-releasing hormone-degrading ectoenzyme, amino- The M1 aminopeptidase family is ubiquitously expressed in tissue peptidase B, leukotrience A4 and excluding aminopeptidase A types, and within cells is found mostly in cytoplasm, endoplasmic (non-conserved HELAH motif). The rooted-phylogenetic tree reticulum (ER) and as membrane components. M1 aminopepti- (Supplementary Figure S2) of the nine aminopeptidases reveals dases belong to MA clan1,2 that are primarily zinc-dependent close genetic homology between ERAP1, ERAP2 and LNPEP enzymes in which two of the zinc ligands are the histidines in the whereas NPEPPS is far distant from the three proteins. Similarly, highly conserved motif: His-Glu-Xaa-Xaa-His (‘HEXXH’) and gluta- APB and LTA4H are structurally and genetically but not mate are a catalytic residue. They also have highly conserved functionally related.7 The mouse homolog of ERAP1 has 85% GAMEN motifs essential for enzyme activities. Tertiary structures sequence homology with human ERAP1. ERAP2 is absent in have been solved for several M1 aminopeptidases,3–5 and all have rodents and the mouse homolog of NPEPPS has 93% sequence the catalytic active site located between the domains I and II. homology to its human counterpart. M1 aminopeptidases are also known to be critically involved in the ERAP1 and ERAP2 in particular are known to trim peptides that regulation of many cellular processes such as cell cycling, growth are destined for major histocompatibility complex (MHC) class I

University of Queensland Diamantina Institute, Translational Research Institute, Princess Alexandra Hospital, Woolloongabba, Queensland, Australia. Correspondence: Professor MA Brown, University of Queensland Diamantina Institute, Translational Research Institute, Princess Alexandra Hospital, Ipswich Road, Woolloongabba, Brisbane, Queensland 4102, Australia. E-mail: [email protected] Received 2 June 2014; accepted 3 June 2014; published online 21 August 2014 Enzyme-based targeted therapies for IMDs N Agrawal and MA Brown 522

Table 1A. Immune-mediated disease associated ERAP1, ERAP2 and NPEPPS polymorphism

Gene SNP Chr. Pos. Risk allele MAF Coding change Functional activity rel. to wt protein Disease

ERAP1 rs30187 96124330 T 0.33 K528R Reduced activity AS, Psoriasis ERAP1 rs17482078 96118866 C 0.22 R725Q Reduced activity AS T Q725R No change Behçet’s disease ERAP1 rs27524 96101944 A 0.36 —— Psoriasis ERAP1 rs10050860 96122210 C 0.77 D575N No change AS ERAP1 rs27044 96118852 C 0.65 Q730E Reduced activity AS ERAP2 rs2549782 96231000 T 0.5 N392K Reduced activity AS, Psoriasis ERAP2 rs2248374 96235896 A 0.45 — Loss of mRNA and absence of ERAP2 protein AS, IBD ERAP2 rs2549794 96244549 C 0.35 —— IBD ERAP2 rs2910686 96252589 T 0.42 —— AS NPEPPS rs9901869 45575206 G 0.49 —— AS Abbreviations: ERAP1, endoplasmic reticulum aminopeptidase 1; ERAP2, endoplasmic reticulum aminopeptidase 2; NPEPPS, Puromycin-sensitive aminopeptidase; SNP, single-nucleotide polymorphism; MAF, minor allele frequency; AS, Ankylosing Spondylitis; IBD, inﬂammatory bowel disease. Non- synonymous (ns) and synonymous (s) changes in the strands of ERAP1, ERAP2 and NPEPPS are shown. Chromosome position and Amino acid (AA) residue numbering are from Human ERAP1 Isoform 2 (Accession No. Q9NZ08-2), Human ERAP2 Isoform 1 (Accession No. Q6P179-1), Human NPEPPS Isoform 1 (Accession No. P56537-1).

Table 1B. Linkage disequilibrium (r2) between marker pairs across ERAP1/ERAP2 genotyped in Wellcome Trust Case Control 2 controls9

rs27524 rs17482078 rs10050860 rs30187 rs2549782

rs27524 96101944 1.00 0.01 0.14 0.95 0.13 rs17482078 96118866 0.01 1.00 0.07 0.01 0.05 rs10050860 96122210 0.14 0.07 1.00 0.15 0.79 rs30187 96124330 0.95 0.01 0.15 1.00 0.12 rs2549782 96231000 0.13 0.05 0.79 0.12 1.00 Abbreviations: ERAP1, endoplasmic reticulum aminopeptidase 1; ERAP2, endoplasmic reticulum aminopeptidase 2.

antigen presentation. Although there are reported associations of Since the discovery of the association of ERAP1 variants with AS, the genes encoding these enzymes with pre-eclampsia,8,9 cervical genome-wide significant association of ERAP1 variants rs27524 cancer10,11 and type 1 diabetes,12 these findings have not (carried on the same haplotype as rs30187) have also been achieved genome-wide significance and have not all replicated reported with psoriasis31 and rs30187 with Behçet’s disease.17,18 in other larger data sets, and are therefore of uncertain status. Whether or not the rs10050860-rs17482078 haplotype is also Genetic association of ERAP1 and ERAP2 is, however, definite with associated with these diseases has not been reported. While the several ‘seronegative’ IMDs including AS,13–15 psoriasis,16 Behçet’s ERAP1 association with psoriasis is concordant (same haplotype disease,17,18 and IBD19–21 (discussed further below). It is remark- with the same direction of association), in Behçet’s disease the able to note that their genetic analyses have revealed that the association is discordant, with the minor allele being a risk factor natural polymorphisms involving altered DNA sequence and for Behçet’s disease, whereas it is protective for AS and psoriasis. reduced protein function in these genes result in protection from Given the strong association of AS with the HLA Class I allele IMDs, raising the possibility that their inhibition could represent a HLA-B27 and the functional role of ERAP1 in trimming peptides novel class of therapeutic drug treatment for these diseases. before HLA Class I incorporation, Evans et al. tested whether ERAP1 variants show evidence of gene–gene interaction with HLA-B27.23 This was demonstrated in three separate case–control GENETIC ASSOCIATIONS OF M1 AMINOPEPTIDASES WITH cohorts, making this the first replicated example of gene–gene DISEASE interaction in any common human disease. Association with The first reported genome-wide significant association of any ERAP1 was only observed in HLA-B27-positive cases. This finding aminopeptidase with disease was with single-nucleotide poly- has recently been extended with the demonstration that the morphisms (SNPs) in ERAP1 with the disease AS.22 This finding has rs30187 variant interacts with HLA-B40, another AS-associated subsequently been widely replicated, including in Canadians, HLA-B allele, such that rs30187 is associated with AS among Europeans, Hungarians, Portuguese, Taiwanese, Han Chinese and HLA-B27-negative, HLA-B40-positive cases, but not among Koreans.23–30 Two distinct disease-associated haplotypes have HLA-B27-negative, HLA-B40-negative cases. Gene–gene inter- been identified. Of these two haplotypes, the more strongly action has also now been shown in psoriasis between HLA-Cw6 associated is tagged by the nsSNP rs30187. Haplotype evolution and rs2754431 (Table 1A and Table 1B), and between HLA-B51 and studies indicate that rs30187 is the primary disease-associated rs30187 in Behçet’s disease. As discussed below, this strongly SNP on this haplotype. The second haplotype carries two nsSNPs suggests that the four HLA Class I alleles involved in these rs10050860 (Asp575Asn) and rs17482078 (Arg725Gln).23 The gene–gene interactions operate in their respective diseases by linkage disequilibrium between these two SNPs is very strong, similar mechanisms. precluding distinction on genetic grounds alone as to which is the Because of the tight linkage disequilibrium across the chromo- key AS-associated variant. The minor alleles for each of these SNPs some 5q15 locus at which ERAP1, ERAP2 and LNPEP are encoded, are protective for AS; in combination this protective effect is quite and the strong linkage disequilibrium between this locus and HLA strong, with reduction in AS risk in HLA-B27-positive homozygous Class I in AS and psoriasis, distinguishing association of one gene carriers of ERAP1 protective variants of 3- to 4-fold.23 from another has required large sample sizes and careful

Table 2. List of nine M1 aminopeptidases and their known inhibitors

Proteins Diseases Inhibitor [Ref]

ERAP1 AS, Psoriasis, IBD, Behçet’s disease None reported43 ERAP2 AS, Psoriasis, IBD None reported43 LNPEP Alzheimer’s disease, Memory loss, cognitive impairment, Ischemic heart HFI41964 disease Aminopeptidase N Hypertension, angiogenesis and tumour metastasis Bestatin34 Aminopeptidase A Hypertension RB15037 Thyrotropin-releasing hormone-degrading ectoenzyme Brain/spinal injuries and central nervous system disorders CPHNA39 NPEPPS None implicated so far Tosedostat33 Aminopeptidase B Inflammatory disorders, malignant tumours and type II diabetes L-Lysinethiol65 Leukotriene A4 hydrolase Myocardial infarction and stroke DG-05136 Abbreviations: ERAP1, endoplasmic reticulum aminopeptidase 1; ERAP2, endoplasmic reticulum aminopeptidase 2; NPEPPS, Puromycin-sensitive aminopeptidase; AS, Ankylosing Spondylitis; IBD, inflammatory bowel disease. conditional analyses. In AS, considering either HLA-B27-negative LNPEP enzymes are speculated to be involved in cytosolic cases or controls (in which no ERAP1 association is observed processing of these proteasome-derived peptides. Such peptides overall), or conditioning for the association of ERAP1, strong are then subsequently transported into ER by the TAP protein association is seen broadly across the ERAP2 and LNPEP genes. where fine editing of peptides to 8–9 residues by ERAP1 takes Genetically, it has not been possible to determine which of the place,36 this being the optimal length of binding to MHC class I SNPs on this associated haplotype are truly disease associated, molecules and HLA-B27. ERAP1 is known to cleave substrates although as one of the SNPs on the haplotype results in a null more efficiently with a hydrophobic C-terminal residue. The well- allele of ERAP2, these functional data demonstrate that this SNP at processed antigenic peptide repertoires are then presented by least is the key associated SNP in ERAP2. Whether there is residual HLA-B27 to CD8+T cells to instigate an immune response. association in LNPEP is unclear. Similar conditional analyses in fi Nine splice variants of ERAP1 gene are known so far, with no psoriasis have con rmed independent association of the same association to any disease. Strong cis − effects by SNPs on ERAP1 ERAP2 haplotype as is associated with AS. This haplotype is also expression are well established.37 Harvey and coworkers38 associated both with IBD and with juvenile idiopathic arthritis.32 demonstrated that there is a strong correlation of strength of The recent International Genetics of AS Consortium Immuno- chip study13 extended these observations, demonstrating associa- association of ERAP1 variants with AS with their effects on ERAP1 tion of chromosome 17q21 with disease, an area encoding expression. This is consistent with ERAP1 variants affecting AS risk – through effects on the gene’s expression. NPEPPS. No evidence of gene gene interaction was observed 3,39 between this gene and HLA-B alleles. As the Immunochip only Crystal structures of ERAP1 are resolved to 3A resolution. The – – typed a sparse set of SNPs at this locus, it is not yet clear what the protein has four domains, spanning residues 1 254, 255 527/529, key associated variant there is. This locus has not yet been 528/530–613/614 and 614/615–941, respectively. ERAP1 shows associated with any other disease. two distinct conformations: an open and a closed one (Figure 2). The structure3,39,40 reveals extensive domain movements, including an active site closure as well as three different open THERAPEUTIC TARGETING OF M1 AMINOPEPTIDASES conformations,3 thus providing critical insights into the ERAP1 Aminopeptidase inhibition has previously been demonstrated catalytic processes. Domain I is located on above domain II and as a promising therapeutic target in a number of malignant includes binding residues for the N-terminal substrate, such as 33 conditions. All the known inhibitors for the six M1 aminopepti- Gln181 and Glu183, whose mutations affect enzyme’s activity and dases except ERAP1, ERAP2 and NPEPPS are listed (Table 2). specificity.41 Domain II is the catalytic site whereas Domain III Considering IMD associated aminopeptidases, there are no potent connects the catalytic domain with the large C-terminal domain, 34 inhibitors of ERAP1 or ERAP2 reported to date. Tosedostat has and in closed state seals off the active site. A large cavity between fi 33 recently been identi ed as a potent inhibitor of NPEPPS but it domains II and IV is presumably a substrate-binding site that could has not yet been trailled in any immune mediated disease. accommodate different length peptides for catalysis. The residues that are protective in IMDs are depicted in the FUNCTIONAL CHARACTERIZATION OF DISEASE-ASSOCIATED three-dimensional crystal structure of ERAP1 (Figure 2). The ERAP1 AMINOPEPTIDASES SNP rs30187 changes a highly conserved residue (Arg528Lys), ERAP1 dramatically reducing the enzyme catalytic activity. The amino- acid residue 528 is located near the entrance of the substrate ERAP1 is commonly known as ER aminopeptidase associated with pocket and may contribute to the substrate-binding affinity antigen processing or ER aminopeptidase1 (ERAP1) to denote its with the enzyme. Arg725Gln (rs17482078) can affect substrate localization and presumed function as antigen processing fi molecules and involvement in regulation of the immune response. sequence or speci city because they are exposed on the inner ERAP1 is a zinc metallopeptidase with typical motifs characteristic surface of the C-terminal cavity and hence reduces enzyme for members of the M1 aminopeptidase family as described activity. As mentioned above, it has not been possible yet to above. ERAP1 exists in tissues that express higher concentrations determine which of the two polymorphisms of joint SNPs of MHC class I molecules such as liver, lungs, spleen and thymus. rs10050860/rs17482078 (coding for Asn575/Gln725) are protective ERAP1 expression is also induced by interferon-γ.35 The lives of in AS, or if both are contributory to disease risk. In vitro studies of proteins begin with transcription and translation (Figure 1). In a ERAP1 function have suggested that the rs10050860 (Asn575) cell, the complex repertoire of endogenous proteins is initially enzyme activity is unaffected9 when the 575 position is mutated, degraded in the cytoplasm by the multi-unit proteasome complex, making it more likely Gln725 is the key functional variant on this generating peptide fragments up to 25 aa in length. NPEPPS and haplotype rather than Asn575.

CD8-T cell

Nucleus Transcription TCR

HLA Class I

Translation

Golgi Endogenous protein

β m ERAP1 1 2 ERAP2 2 HLA class I multi-unit Trimmed peptides proteasome complex

TAP proteasome- 1 Ankylosing spondylitis, Psoriasis, derived Behçet’s disease peptides 2 Ankylosing spondylitis, Inflammatory NPEPPS 3 Bowel disease, Psoriasis LNPEP 3 Ankylosing spondylitis

Figure 1. Classical MHC Class I biological pathway involving M1 aminopeptidases that are genetically associated with immune-mediated diseases.

Figure 2. Crystal structure of human ERAP1, bound to the zinc aminopeptidase inhibitor, Bestatin. (a) Open form of ERAP1 (pdb no. 3MDJ). (b) Closed form of ERAP1 (pdb no. 2XDT). The various domains of ERAP1 are colored in ribbon representation as follows: I: blue, II: green, III: orange and IV: yellow. Three non-synonymous SNPs at K528, D575 and Q725 position are associated with AS and Psoriasis (in red). Q725 increases the risk for Behçet’s disease but it is protective for AS, suggesting that alterations in the peptide repertoire of different disease-associated MHC-I molecules induced by the same mutation may have opposite effects on their respective diseases. Molecular model images were generated using the Pymol software (Schrodinger, Camberley, UK).

Genes and Immunity (2014) 521 – 527 © 2014 Macmillan Publishers Limited Enzyme-based targeted therapies for IMDs N Agrawal and MA Brown 525 haplotype as rs2248374 (r2 = 0.90 in 1000 Genomes June 2010 data) (Tables 1A and B), and is therefore rarely translated.

NPEPPS NPEPPS, also known as puromycin-sensitive aminopeptidase or cytosol alanyl aminopeptidase, is another zinc metallopeptidase which hydrolyzes N-terminal amino acids from its peptide substrates. NPEPPS is expressed in most tissues as a cytosolic aminopeptidase, but a membrane-associated form has also been identified in the brain.49,50 The enzyme has broad substrate specificity but prefers basic and hydrophobic residues in the substrate.51 NPEPPS is known to hydrolyze physiological endogenous peptides such as dynorphins and enkephalins, and contributes to the degradation of these peptides in the brain.52 NPEPPS is also known to degrade the tau protein, which accumulates and polymerizes in some neurodegenerative diseases.53 To date, no crystal structure of NPEPPS bound in any complex has ever been reported, and thus its structural features are unknown. There is very little knowledge on the effect of mutations on NPEPPS functional activity. There are handful of mutagenesis studies which predicted that mutating the catalytic glutamate at 309 position converts the NPEPPS enzyme into an inactive binding Figure 3. Crystal structure of human ERAP2, Closed form of ERAP2. protein, and constitute evidence that this residue is essential for The various domains of ERAP2 are colored in ribbon representation catalysis whereas the effect of mutating conserved tyrosine 394 as follows: I: blue, II: green, III: orange and IV: yellow. One non- results in complete loss of catalytic activity of the NPEPPS.54 synonymous SNP at N392 position is associated with AS, IBD and fi Crohn’s disease (in red) (pdb no.4E36). Molecular model image was Genome-wide signi cant association is seen at synonymous generated using the Pymol software. SNP rs9901869 (Tables 1A and B) of NPEPPS at chromosome 17q21, however there are no functional variants known so far.13 ERAP2 Further studies of NPEPPS are required to determine the key ERAP2 is also a member of zinc-dependent M1 aminopeptidases disease-associated variants and their effects on NPEPPS function. and shares about 50% of sequence identity with other members Functional effect of ERAP1 variants on HLA Class I properties.In of M1 aminopeptidase family. Unlike its homolog ERAP1, there 2011,23 a study reported that polymorphisms of ERAP1 involved in have been limited in vivo studies of ERAP2 due to its absence in peptide trimming and altered enzyme activity before HLA class I rodent models, for example, mouse, rat, rabbit and guinea pig.42,43 44 and peptide presentation only affected AS risk in HLA-B27-positive At the molecular level, ERAP2 has been shown to have individuals. The findings provided initial evidence that HLA-B27 functional preferences for peptide − substrate sequences,45 the 44 operates in AS through a mechanism involving aberrant proces- processing of epitopes that are poorly trimmed by ERAP1. ERAP2 sing of antigenic peptides involving ERAP1. In the same year, the is known as a highly homologous aminopeptidase that is first crystal structure of ERAP13 was determined in both open and colocalized with ERAP1 but presents distinct specificity for the closed form which provided the first snapshots along the ERAP1 N-terminal residue of the peptide substrate. There is some catalytic pathway. The crystal structures revealed extensive evidence that ERAP2 can form heterodimers with ERAP1, which domain movements, including an active site closure as well as have more efficient trimming activity jointly compared with ERAP1 three different open conformations, thus providing critical insights homotrimers or ERAP2 homodimers.46 into the ERAP1 enzyme’s catalytic cycle. The crystal structure of ERAP2 has been resolved to 3.0-3.3 Å In 2012,55 a study of surface expression of HLA-B27 complexes resolution. The protein shares overall similar domain homology on peripheral blood mononuclear cells of AS patients revealed with ERAP1 and has four domains (Figure 3), spanning residues that AS patients had lower HLA expression levels than normal 1–271, 272–546, 547–647, 648–960, respectively; only the closed individuals, depending on the presence of different ERAP1 form of ERAP2 has been crystallized (Figure 3) so far. functional variants. The studies demonstrated that ERAP1 suppres- ERAP1 and ERAP2 are jointly responsible for trimming peptides sion due to genetic polymorphism led to a B27-subtype specific to the optimum length for recognition by the immune system, response resulting in increased intracellular free heavy chain and thus have an important role in the regulation of the immune surface peptide HLA, and affected cells expressing HLA- B*2704 or B*2705 but not B*2706 or B*2709. These findings confirmed that response. There is strong association seen across ERAP2 with AS, fl – IBD and psoriasis, and by purely genetic means it is difficult to uctuations in the ERAP1 HLA-B27 interactions resulted in variable peptide presentation on the HLA surface, and provide a distinguish the key disease-associated variants. However, the potential explanation for the absence of genetic association of the synonymous SNP rs2248374 AS- and IBD-protective G allele has HLA-B27 subtypes B*2706 and B*2709 with AS. been shown to lead to skipping of a standard exon 10 splice site, In another study,55–59 the hypothesis that ERAP1 variant-related extending exon 10 but leading to premature termination of structural alterations affect the repertoire of antigenic peptides, transcription, as the extended exon 10 contains new stop codons stability and function bound to the HLA-B27 molecules was tested. not found in the wild-type transcripts. The mRNA produced is then The studies showed that AS-associated ERAP1 polymorphisms and subjected to nonsense-mediated decay, and no ERAP2 protein is altered catalytic activity among HLA-B27-positive individuals had translated. This in turn leads to reduced MHC class I surface major effects on the HLA-B27 peptidome. A detailed peptide expression.47 A further SNP rs2549782 (Asn392Lys) in the coding analyses revealed longer peptides to be more abundant in region of ERAP2 gene is also associated with AS, IBD and psoriasis the presence of AS-protective polymorphisms, consistent with susceptibility.48 This SNP is nearly always found on the same diminished enzymatic activity. Moreover, HLA-B27 ligands

© 2014 Macmillan Publishers Limited Genes and Immunity (2014) 521 – 527 Enzyme-based targeted therapies for IMDs N Agrawal and MA Brown 526 predominant in carriers of AS-protective ERAP1 polymorphisms CONFLICT OF INTEREST showed a tendency toward residues that were more susceptible to The authors declare no conflict of interest. ERAP1 trimming. This study demonstrated that ERAP1 variants affect both the quantity of antigenic peptide presented by HLA- B27, and the nature of presented peptides both in terms of length REFERENCES and sequence. It was further postulated that the functional 1 Rawlings ND, Salvesen G. Handbook of Proteolytic Enzymes. Elsevier, London, UK, interaction between ERAP1 and HLA-B27 results in altered balance 2013. 2 Rawlings ND, Barrett AJ. Evolutionary families of peptidases. Biochem J. 1993; 290: between epitope generation and destruction among ERAP1 – variants. Such alteration is both variant and peptide specific, and 205 218. 3 Kochan G, Krojer T, Harvey D, Fischer R, Chen L, Vollmar M et al. Crystal it depends on the N-terminal flanking and P1 residues of natural 57 structures of the endoplasmic reticulum aminopeptidase-1 (ERAP1) reveal the HLA-B27 ligand. molecular basis for N-terminal peptide trimming. Proc Natl Acad Sci 2011; 108: 60 In 2014, an independent group analyzed the effects of ERAP1 7745–7750. on the HLA-B27 peptide repertoire and peptide presentation to 4 Nguyen TT, Chang SC, Evnouchidou I, York IA, Zikos C, Rock KL et al. Structural cytotoxic T lymphocytes and also showed that ERAP1 directly basis for antigenic peptide precursor processing by the endoplasmic reticulum alters the repertoire of peptides and peptide presentation. ERAP1 aminopeptidase ERAP1. Nat Struct Mol Biol 2011; 18:604–613. silencing increased the proportion of longer peptides (11–13 mer) 5 Birtley JR, Saridakis E, Stratikos E, Mavridis IM. The crystal structure of human bound to HLA-B27 and changed the nature of the P1 amino acid endoplasmic reticulum aminopeptidase 2 reveals the atomic basis for distinct – of peptides bound to HLA-B27. Overall, the findings confirmed a roles in antigen processing. Biochemistry 2012; 51:286 295. 6 Albiston AL, Ye S, Chai SY. Membrane bound members of the M1 family: more critical role of ERAP1 in determining the length and sequence of than aminopeptidases. Protein Pept Lett 2004; 11: 491–500. peptides bound and presented by the AS-associated HLA-B27. 7 Fukasawa KM, Fukasawa K, Harada M, Hirose J, Izumi T, Shimizu T. Aminopepti- Moreover, the studies demonstrated that both ERAP1 silencing dase B is structurally related to leukotriene-A4 hydrolase but is not a bifunctional and the presence of a protective ERAP1 allele (nsSNP rs30187 enzyme with epoxide hydrolase activity. Biochem J 1999; 339: 497–502. (Arg528Lys)), reduced functional presentation of a naturally 8 Saito S, Sakai M, Sasaki Y, Nakashima A, Shiozaki A. Inadequate tolerance occurring HLA-B27 epitope to CD8+ T cells. These collective induction may induce pre-eclampsia. J Reprod Immunol. 2007; 76:30–39. findings from various researchers strongly suggest that functional 9 Vanhille DL, Hill LD, Hilliard DD, Lee ED, Teves ME, Srinivas S et al. A novel ERAP2 perturbations in ERAP1 polymorphisms have a major impact on haplotype structure in a Chilean Population: implications for ERAP2 protein expression and preeclampsia risk. Mol Genet Genomic Med 2013; 1:98–107. HLA Class I expression, and peptide handling and stability. This 10 Mehta AM, Jordanova ES, Kenter GG, Ferrone S, Fleuren GJ. Association of antigen suggests that ERAP1 inhibition could potentially be used as a processing machinery and HLA class I defects with clinicopathological outcome in logical therapeutic target both in AS and in other ERAP1-HLA Class cervical carcinoma. Cancer Immunol Immunother 2008; 57:197–206. I-associated diseases. 11 Mehta AM, Jordanova ES, Corver WE, van Wezel T, Uh HW, Kenter GG et al. 61 Using ERAP − / − B27-transgenic mice, Akram et al recently Single nucleotide polymorphisms in antigen processing machinery component demonstrated that the HLA-B27 immunodominant influenza ERAP1 significantly associate with clinical outcome in cervical carcinoma. Genes nucleoprotein epitope (NP383-391) is initially made as an Chromosomes Cancer 2009; 48:410–418. N-terminally extended 14-mer, which is then trimmed by ERAP 12 Fung EY, Smyth DJ, Howson JM, Cooper JD, Walker NM, Stevens H et al. 61 fl − − Analysis of 17 autoimmune disease-associated variants in type 1 diabetes before presentation. In uenza-infected ERAP / mice had fi reduced HLA Class I cell surface expression, reduced cytotoxic identi es 6q23/TNFAIP3 as a susceptibility locus. Genes Immun 2009; 10: 188–191. T lymphocyte against the epitope, and consistent changes in their 13 Cortes A, Hadler J, Pointon JP, Robinson PC, Karaderi T, Leo P et al. Identification CD8 T-cell repertoire. Using a 14-mer N-terminally extended of multiple risk variants for ankylosing spondylitis through high-density NP383, ERAP-replete but not ERAP-deficient mice were able to genotyping of immune-related loci. Nat Genet 2013; 45: 730–738. mount an immune reaction, confirming that ERAP is critical in 14 Reveille JD. The genetic basis of ankylosing spondylitis. Curr Opin Rheumatol 2006; cleavage of this N-terminally extended peptide to enable its 18:332–341. presentation by B27. In contrast, no such differences were 15 Brown MA. Genetics of ankylosing spondylitis. Curr Opin Rheumatol 2010; 22: observed with HLA-B7 expressing mice, although it is very likely 126–132. that the B7 immunodominant NP epitope (NP418-426) is also 16 Genetic Analysis of Psoriasis Consortium & the Wellcome Trust Case Control Consortium 2, Strange A, Capon F, Spencer CC, Knight J, Weale ME et al. initially produced as an N-terminally extended peptide, as A genome-wide association study identifies new psoriasis susceptibility loci and otherwise NP418-426 would be predicted not to be transported an interaction between HLA-C and ERAP1. Nat Genet 2010; 42: 985–990. into the ER by the TAP protein. The absence of any effect on B7 17 Kirino Y, Bertsias G, Ishigatsubo Y, Mizuki N, Tugal-Tutkun I, Seyahi E et al. presentation of NP418 by ERAP knockout suggests that other Genome-wide association analysis identifies new susceptibility loci for Behçet's aminopeptidases cleave the NP418 epitope to enable its disease and epistasis between HLA-B*51 and ERAP1. Nat Genet 2013; 45: 202–207. presentation. It has also been shown that in B7/B27 co- 18 Gül A. Genetics of Behçet's disease: lessons learned from genomewide association expressing mice, that there is partial negative selection of studies. Curr Opin Rheumatol 2013; 26:56–63. B27/NP383-reactive T cells, termed immunodominance.62 Human 19 Guerini FR, Cagliani R, Forni D, Agliardi C, Caputo D, Cassinotti A et al. A functional studies have recently shown that HLA-B7 is protective for AS,63 variant in ERAP1 predisposes to multiple sclerosis. PLoS ONE 2012; 7: e29931. 20 Christodoulou K, Wiskin AE, Gibson J, Tapper W, Willis C, Afzal NA et al. Next and it will be fascinating to compare the T-cell repertoires of generation exome sequencing of paediatric inflammatory bowel disease patients B27-positive cases heterozygote for B7 or other HLA-B alleles to identifies rare and novel variants in candidate genes. Gut 2013; 62: 977–984. see if these murine findings translate. 21 Franke A, McGovern DP, Barrett GC, Wang K, Radford-Smith GL, Ahmad T et al. Genome-wide meta-analysis increases to 71 the number of confirmed Crohn's disease susceptibility loci. Nat Genet 2010; 42:1118–1125. CONCLUSION 22 Wellcome Trust Case Control Consortium, Australo-Anglo-American Spondyloar- In this review, the critical role of ERAP1, ERAP2 and NPEPPS thritis Consortium (TASC), Burton PR, Clayton DG, Cardon LR, Craddock N et al. aminopeptidases have been shown by their multiple genetic and Association scan of 14,500 nonsynonymous SNPs in four diseases identifies – functional associations with HLA Class I and IMDs. Understanding autoimmunity variants. Nat Genet 2007; 39: 1329 1337. 23 Evans DM, Spencer CC, Pointon JJ, Su Z, Harvey D, Kochan G et al. Interaction the physiological processes underlying antigenic peptide genera- between ERAP1 and HLA-B27 in ankylosing spondylitis implicates peptide tion in relation to MHC class I presentation and regulation of handling in the mechanism for HLA-B27 in disease susceptibility. Nat Genet 2011; adaptive immune responses may pave the way for developing 43:761–767. new enzyme-based targeted therapies for IMDs and also other 24 Li C, Lin Z, Xie Y, Guo Z, Huang J, Wei Q et al. ERAP1 is associated with ankylosing diseases such as cancer and viral infections. Spondylitis in Han Chinese. J Rheumatol 2011; 38: 317–321.

Genes and Immunity (2014) 521 – 527 © 2014 Macmillan Publishers Limited Enzyme-based targeted therapies for IMDs N Agrawal and MA Brown 527 25 Pazar B, Safrany E, Gergely P, Szanto S, Szekanecz Z, Poor G. Association 46 Saveanu L, Carroll O, Lindo V, Del Val M, Lopez D, Lepelletier Y et al. Concerted of ARTS1 gene polymorphisms with ankylosing spondylitis in the Hungarian peptide trimming by human ERAP1 and ERAP2 aminopeptidase complexes in the population: the rs27044 variant is associated with HLA-B*2705 subtype endoplasmic reticulum. Nat Immunol 2005; 6:689–697. in Hungarian patients with ankylosing spondylitis. J Rheumatol 2010; 37: 47 Andres AM, Dennis MY, Kretzschmar WW, Cannons JL, Lee-Lin SQ, Hurle B et al. 379–384. Balancing selection maintains a form of ERAP2 that undergoes nonsense- 26 Pimentel-Santos FM, Ligeiro D, Matos M, Mourao AF, Sousa E, Pinto P et al. mediated decay and affects antigen presentation. PLoS Genet 2010; 6: Association of IL23R and ERAP1 genes with ankylosing spondylitis in a Portuguese e1001157. population. Clin Exp Rheumatol 2009; 27:800–806. 48 Evnouchidou I, Birtley J, Seregin S, Papakyriakou A, Zervoudi E, Samiotaki M et al. 27 Australo-Anglo-American Spondyloarthritis C, Reveille JD, Sims AM, Danoy P, A common single nucleotide polymorphism in endoplasmic reticulum amino- Evans DM, Leo P et al. Genome-wide association study of ankylosing spondylitis peptidase 2 induces a specificity switch that leads to altered antigen processing. identifies non-MHC susceptibility loci. Nat Genet 2010; 42:123–127. J Immunol 2012; 189: 2383–2392. 28 Szczypiorska M, Sanchez A, Bartolome N, Arteta D, Sanz J, Brito E et al. ERAP1 49 McLellan S, Dyer SH, Rodriguez G, Hersh LB. Studies on the tissue distribution of polymorphisms and haplotypes are associated with ankylosing spondylitis the puromycin-sensitive enkephalin-degrading aminopeptidases. J Neurochem susceptibility and functional severity in a Spanish population. Rheumatology 1988; 51:1552–1559. (Oxford) 2011; 50: 1969–1975. 50 Dyer SH, Slaughter CA, Orth K, Moomaw CR, Hersh LB. Comparison of the soluble 29 Davidson SI, Liu Y, Danoy PA, Wu X, Thomas GP, Jiang L et al. Association of STAT3 and membrane-bound forms of the puromycin-sensitive enkephalin-degrading – and TNFRSF1A with ankylosing spondylitis in Han Chinese. Ann Rheum Dis 2011; aminopeptidases from rat. J Neurochem 1990; 54:547 554. fi 70: 289–292. 51 Johnson GD, Hersh LB. Studies on the subsite speci city of the rat brain 30 Davidson SI, Wu X, Liu Y, Wei M, Danoy PA, Thomas G et al. Association of ERAP1, puromycin-sensitive aminopeptidase. Arch Biochem Biophys 1990; 276: – but not IL23R, with ankylosing spondylitis in a Han Chinese population. Arthritis 305 309. Rheum 2009; 60: 3263–3268. 52 Safavi A, Hersh LB. Degradation of dynorphin-related peptides by the puromycin- 31 Strange A, Capon F, Spencer CC, Knight J, Weale ME, Allen MH et al. A genome- sensitive aminopeptidase and aminopeptidase M. J Neurochem 1995; 65: – wide association study identifies new psoriasis susceptibility loci and an inter- 389 395. action between HLA-C and ERAP1. Nat Genet 2010; 42:985–990. 53 Kudo LC, Parfenova L, Ren G, Vi N, Hui M, Ma Z et al. Puromycin-sensitive ami- 32 Hinks A, Cobb J, Marion MC, Prahalad S, Sudman M, Bowes J et al. Dense gen- nopeptidase (PSA/NPEPPS) impedes development of neuropathology in hPSA/ TAU(P301L) double-transgenic mice. Hum Mol Genet 2011; 20: 1820–1833. otyping of immune-related disease regions identifies 14 new susceptibility loci for 54 Thompson MW, Govindaswami M, Hersh LB. Mutation of active site residues juvenile idiopathic arthritis. Nat Genet 2013; 45:664–669. of the puromycin-sensitive aminopeptidase: conversion of the enzyme into a 33 Krige D, Needham LA, Bawden LJ, Flores N, Farmer H, Miles LE et al. CHR-2797: catalytically inactive binding protein. Arch Biochem Biophys 2003; 413:236–242. an antiproliferative aminopeptidase inhibitor that leads to amino acid deprivation 55 Haroon N, Tsui FW, Uchanska-Ziegler B, Ziegler A, Inman RD. Endoplasmic in human leukemic cells. Cancer Res 2008; 68:6669–6679. reticulum aminopeptidase 1 (ERAP1) exhibits functionally significant interaction 34 Hitzerd SM, Verbrugge SE, Ossenkoppele G, Jansen G, Peters GJ. Positioning of with HLA-B27 and relates to subtype specificity in ankylosing spondylitis. Ann aminopeptidase inhibitors in next generation cancer therapy. Amino Acids 2014; Rheum Dis 2012; 71: 589–595. 46: 793–808. 56 Alvarez-Navarro C, López de Castro JA. ERAP1 structure, function and pathoge- 35 Saric T, Chang SC, Hattori A, York IA, Markant S, Rock KL et al. An IFN-gamma- netic role in ankylosing spondylitis and other MHC-associated diseases. Mol induced aminopeptidase in the ER, ERAP1, trims precursors to MHC class Immunol 2013; 57 (1): 12–21. I-presented peptides. Nat Immunol 2002; 3: 1169–1176. 57 García-Medel N, Sanz-Bravo A, Van Nguyen D, Galocha B, Gómez-Molina P, 36 York IA, Chang SC, Saric T, Keys JA, Favreau JM, Goldberg AL et al. The ER Martín-Esteban A et al. Functional interaction of the ankylosing spondylitis- aminopeptidase ERAP1 enhances or limits antigen presentation by trimming associated endoplasmic reticulum aminopeptidase 1 polymorphism and HLA-B27 epitopes to 8-9 residues. Nat Immunol 2002; 3: 1177–1184. in vivo. Mol Cell Proteomics 2012; 11:1416–1429. 37 Ouyang C, Smith DD, Krontiris TG. Evolutionary signatures of common human cis- 58 Martín-Esteban A, Gómez-Molina P, Sanz-Bravo A, López de Castro J. Combined regulatory haplotypes. PLoS ONE 2008; 3: e3362. effects of ankylosing spondylitis-associated ERAP1 polymorphisms outside the 38 Harvey D, Pointon JJ, Evans DM, Karaderi T, Farrar C, Appleton LH et al. Investi- catalytic and peptide-binding sites on the processing of natural HLA-B27 ligands. gating the genetic association between ERAP1 and ankylosing spondylitis. Hum J Biol Chem 2014; 289: 3978–3990. – Mol Genet 2009; 18: 4204 4212. 59 Alvarez-Navarro C, López de Castro J. ERAP1 in ankylosing spondylitis: genetics, 39 Nguyen TT, Chang SC, Evnouchidou I, York IA, Zikos C, Rock KL et al. Structural biology and pathogenetic role. Curr Opin Rheumatol 2013; 25: 419–425. basis for antigenic peptide precursor processing by the endoplasmic reticulum 60 Chen L, Fischer R, Peng Y, Reeves E, McHugh K, Ternette N et al. Critical role – aminopeptidase ERAP1. Nat Struct Mol Biol 2011; 18:604 613. of endoplasmic reticulum aminopeptidase 1 in determining the length and 40 Stratikos E, Stern LJ. Antigenic peptide trimming by ER aminopeptidases--insights sequence of peptides bound and presented by HLA-B27. Arthritis Rheumatol 2014; – from structural studies. Mol Immunol 2013; 55:212 219. 66:284–294. 41 Goto Y, Tanji H, Hattori A, Tsujimoto M. Glutamine-181 is crucial in the enzymatic 61 Akram A, Lin A, Gracey E, Streutker CJ, Inman RD. HLA-B27, but not HLA-B7, fi activity and substrate speci city of human endoplasmic-reticulum aminopepti- immunodominance to Influenza is ERAP dependent. J Immunol 2014; 192 (12): dase-1. Biochem J 2008; 416:109–116. 5520–5528. 42 Gibbs RA, Weinstock GM, Metzker ML, Muzny DM, Sodergren EJ, Scherer S et al. 62 Akram A, Inman RD. Co-expression of HLA-B7 and HLA-B27 alleles is associated Genome sequence of the Brown Norway rat yields insights into mammalian with B7-restricted immunodominant responses following influenza infection. Eur J evolution. Nature 2004; 428:493–521. Immunol 2013; 43: 3254–3267. 43 Haroon N, Inman RD. Endoplasmic reticulum aminopeptidases: Biology and 63 Cortes A, de Bakker P, Brown MA. Fine-mapping major histocompatibility complex pathogenic potential. Nat Rev Rheumatol 2010; 6:461–467. variation associated with ankylosing spondylitis susceptibility. In American 44 Ascher DB, Polekhina G, Parker MW. Crystallization and preliminary X-ray College of Rheumatology. Arthritis Rheum 2013; 12: 1706. diffraction analysis of human endoplasmic reticulum aminopeptidase 2. Acta 64 Albiston AL, Morton CJ, Ng HL, Pham V, Yeatman HR, Ye S et al. Identification and Crystallogr Sect F Struct Biol Cryst Commun 2012; 68: 468–471. characterization of a new cognitive enhancer based on inhibition of insulin- 45 Zervoudi E, Papakyriakou A, Georgiadou D, Evnouchidou I, Gajda A, Poreba M regulated aminopeptidase. FASEB J 2008; 22: 4209–4217. et al. Probing the S1 specificity pocket of the aminopeptidases that generate 65 Ocain TD, Rich DH. L-lysinethiol: a subnanomolar inhibitor of aminopeptidase B. antigenic peptides. Biochem J 2011; 435:411–420. Biochem Biophys Res Commun 1987; 145: 1038–1042.

Supplementary Information accompanies this paper on Genes and Immunity website (http://www.nature.com/gene)