Identification of a Genetic Variation in ERAP1 Aminopeptidase that Prevents Human Cytomegalovirus miR-UL112-5p-Mediated Immunoevasion

Article Identification of a Genetic Variation in ERAP1 Aminopeptidase that Prevents Human Cytomegalovirus miR-UL112-5p-Mediated Immunoevasion Graphical Abstract Highlights dHCMV miR-112-5p targets ERAP1, preventing the presentation of viral epitopes to CTLs dThe rs17481334 variant G preserves ERAP1 from miR-UL1125p-mediated degradation dHCMV-infected GG cells are more efficiently lysed by HCMVpeptide-specific CTLs dDecreased HCMV seropositivity is detected among multiple sclerosis GG patients Authors Paolo Romania, Loredana Cifaldi, Benedetta Pignoloni, ..., Franco Locatelli, Mauro D’Amato, Doriana Fruci Correspondence [email protected] In Brief Romania et al. identify a single-nucleotide polymorphism in ERAP1 gene, which represents a resistance mechanism to HCMV miR-UL112-5p-based immune evasion strategy, with potential implications for individual susceptibility to viral infection and other diseases. Romania et al., 2017, Cell Reports 20, 846–853 July 25, 2017 ª2017 The Author(s). http://dx.doi.org/10.1016/j.celrep.2017.06.084 Cell Reports Article Identification of a Genetic Variation in ERAP1 Aminopeptidase that Prevents Human Cytomegalovirus miR-UL112-5p-Mediated Immunoevasion Paolo Romania, 1 Loredana Cifaldi, 1 Benedetta Pignoloni, 2 Nadia Starc, 1 Valerio D’Alicandro, 1 Ombretta Melaiu, 1 Giuseppina Li Pira, 1 Ezio Giorda, 3 Rosalba Carrozzo, 4 Monika Bergvall, 5 Tomas Bergstro ¨m, 6 Lars Alfredsson, 7,8 Tomas Olsson, 9 Ingrid Kockum, 9 Ilkka Seppa ¨la ¨, 10 Terho Lehtima ¨ki, 10 Mikko A. Hurme, 11 Hartmut Hengel, 12 Angela Santoni, 2 Cristina Cerboni, 2 Franco Locatelli, 1,13 Mauro D’Amato, 14,15,16 and Doriana Fruci 1,16,17, * 1 Paediatric Haematology/Oncology Department, Ospedale Pediatrico Bambino Gesu `, IRCCS, 00146 Rome, Italy 2 Department of Molecular Medicine, Istituto Pasteur-Fondazione Cenci Bolognetti, ‘‘Sapienza’’ University of Rome, 00161 Rome, Italy 3 Unit of Flow Cytometry, Ospedale Pediatrico Bambino Gesu `, IRCCS, 00146 Rome, Italy 4 Unit of Muscular and Neurodegenerative Diseases, Laboratory of Molecular Medicine, Ospedale Pediatrico Bambino Gesu `, IRCCS, 00146 Rome, Italy 5 Department of Biosciences and Nutrition, Karolinska Institutet, 171 77 Stockholm, Sweden 6 Department of Infectious Diseases, Section for Clinical Virology, Institute of Biomedicine, University of Gothenburg, 41345 Go ¨teborg, Sweden 7 Institute of Environmental Medicine, Karolinska Institutet, 171 77 Stockholm, Sweden 8 Centre for Occupational and Environmental Medicine, Stockholm County Council, 171 77 Stockholm, Sweden 9 Department of Clinical Neuroscience and Center for Molecular Medicine, Karolinska Institutet, 171 77 Stockholm, Sweden 10 Department of Clinical Chemistry, Fimlab Laboratories and Finnish Cardiovascular Research Center-Tampere, Faculty of Medicine and Life Sciences, University of Tampere School of Medicine, 33014 Tampere, Finland 11 Department of Microbiology and Immunology, FimLab Laboratories and Faculty of Medicine and Life Sciences, University of Tampere School of Medicine, 33014 Tampere, Finland 12 Institute of Virology, Medical Center, and Faculty of Medicine, University of Freiburg, 79104 Freiburg, Germany 13 University of Pavia, 27100 Pavia, Italy 14 Department of Medicine Solna, Karolinska Institutet, 171 77 Stockholm, Sweden 15 BioDonostia Health Research Institute San Sebastian and IKERBASQUE, Basque Foundation for Science, 48013 Bilbao, Spain 16 These authors contributed equally 17 Lead Contact *Correspondence: doriana.[email protected] http://dx.doi.org/10.1016/j.celrep.2017.06.084 SUMMARY Herein, we demonstrate that HCMV miR-UL112-5p targets ERAP1, thereby inhibiting the processing and presentation of the HCMV pp65 495-503 peptide to specific CTLs. In addition, we show that the rs17481334 G variant, naturally occurring in the ERAP1 30UTR, preserves ERAP1 from miR-UL1125p-mediated degradation. Specifically, HCMV miRUL112-5p binds the 30UTR of ERAP1 A variant, but not the 30UTR of ERAP1 G variant, and, accordingly, ERAP1 expression is reduced both at RNA and protein levels only in human fibroblasts homozygous for the A variant. Consistently, HCMV-infected GG fibroblasts were more efficient in trimming viral antigens and being lysed by HCMV-peptide-specific CTLs. Notably, a significantly decreased HCMV seropositivity was detected among GG individuals suffering from multiple sclerosis, a disease model in which HCMV is negatively associated with adultonset disorder. Overall, our results identify a resistance mechanism to HCMV miR-UL112-5p-based immune evasion strategy with potential implications for individual susceptibility to infection and other diseases. INTRODUCTION Human cytomegalovirus (HCMV) is a b-herpesvirus that causes widespread, persistent infection, and it is often lethal in subjects with impaired immunity (Stern-Ginossar et al., 2012). HCMV has evolved multiple strategies to interfere with innate and adaptive immune responses. Virtually, every single step within the MHC class I antigen processing and presentation pathway is targeted by viral proteins (van de Weijer et al., 2015), thus indicating that recognition of viral peptides by CD8 + T cells is a key event in the elimination of HCMV-infected cells. In addition to viral proteins, HCMV also deploys virally encoded microRNAs (miRNAs) to manipulate host immune responses. Currently, 24 mature miRNAs encoded by HCMV have been identified to target both viral and host genes involved in immune defense, chromatin remodeling, cell cycle regulation, apoptosis, signal transduction, and vesicular trafficking (Grey et al., 2007, 2010; Hook et al., 2014a, 2014b; Kim et al., 2015; Lee et al., 2012; Tirabassi et al., 2011). In particular, three viral 846 Cell Reports 20, 846–853, July 25, 2017 ª2017 The Author(s). This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). miRNAs have been found to affect components of the immune system: miR-UL112-1 targets MICB, a stress-induced ligand of the natural killer (NK) cell activating receptor NKG2D, which is crucial for NK cell-mediated killing of viral-infected cells (Stern-Ginossar et al., 2007); miR-UL148D targets the chemokine CCL5, which limits viral infection by recruiting immune cells to the site of infection (Kim et al., 2012); and miR-US4-1 targets ERAP1, a key component of the antigen processing that trims precursors into peptides of the correct length to bind MHC class I molecules, resulting in an impaired production of many viral epitopes and a reduced anti-viral CD8 + T cell response (Kim et al., 2011). Studies in ERAP1-deficient mice have demonstrated that ERAP1 is required for presentation of several viral epitopes, and its absence affects anti-viral CD8 + T cell responses (Blanchard et al., 2010; Firat et al., 2007; Saveanu et al., 2005; Yan et al., 2006). In humans, DNA sequence variation in ERAP1 has been associated to genetic-risk effects in a number of immune-mediated diseases, including ankylosing spondylitis, psoriasis, Behc¸ et’s disease, and multiple sclerosis (MS) (Stratikos et al., 2014). Functional studies have established that these variants affect ERAP1 activity, resulting in changes in the repertoire of antigenic peptides (Reeves et al., 2014), which is critical for shaping innate and adaptive immune responses (Cifaldi et al., 2011, 2015; James et al., 2013). The importance of ERAP1 trimming for immune recognition of virus-infected cells is demonstrated by several findings, including the evolution of escape mutations impairing ERAP1 activity and resulting in decreased CD8 + T cell response (Draenert et al., 2004; Tenzer et al., 2009). Functional genetic variations in miRNA processing machinery, genes, and target sites, named miR-single-nucleotide polymorphisms (miR-SNPs), have attracted special attention due to their involvement in several types of cancer (Ryan et al., 2010). These variants can affect the transcription of target genes, pre-miRNA processing, and/or modulate miRNA-mRNA interactions by affecting miRNA binding to target sites. It is therefore possible to envisage a complex scenario where ERAP1 functional polymorphisms may be relevant to individual susceptibility to both HCMV infection and (directly or indirectly) autoimmunity, through gene-environment (exposure) interactions. In this study, we demonstrate that HCMV miR-UL112-5p specifically targets ERAP1, thereby inhibiting the processing and presentation of the HCMV pp65 495-503 peptide to specific cytotoxic CD8 + T cells. In addition, we found a naturally occurring variant in the 30UTR of the ERAP1 gene (SNP rs17481334), which prevents ERAP1 targeting and degradation by HCMV miR-UL112-5p. RESULTS rs17481334 G Disrupts miR-UL112-5p Binding to ERAP1 mRNA 30UTR Recent deep-sequencing analyses of small RNAs from HCMVinfected human fibroblast cells led to the identification of novel viral miRNA precursors and refinement of miRNA annotations (Meshesha et al., 2012; Stark et al., 2012). In particular, sequences obtained for miR-US4-1 (now miR-US4-5p) resulted in being shifted by 5 bp at the 50end (Figure S1A), thereby complicating the interpretation of the seed-mediated ERAP1 30 UTR targeting described so far (Kim et al., 2011). Regardless of the correct sequence of miR-US4-1, ERAP1 protein decreased during infection with wild-type HCMV, whereas it remained unchanged in cells infected with a mutant HCMV strain carrying a defective processing of the miR-US4-1 primary transcript, suggesting a key role for that region of ERAP1 30UTR in viral-mediated immune evasion. An A-to-G polymorphism mapping to the 30UTR of the ERAP1 mRNA (SNP rs17481334), and corresponding to one of the mutated nucleotides involved in the study by Kim et al. (Figure S1A), was predicted to disrupt the consensus sequence for miR-US4-1 binding. Because rs17481334 did not match the sequence of mature miR-US4-5p, we looked for other HMCV miRNAs targeting the region encompassing this genetic variant. The RNAhybrid algorithm (Rehmsmeier et al., 2004) predicted the miR-UL112-5p to directly target ERAP1 30UTR in that position (Figure S1B). This interaction was validated in vitro using ERAP1 30UTR dual-luciferase assays in 293T cells ectopically expressing miR-UL112-5p or a mutated version of miR-UL1125p with a single base substitution corresponding to the SNP (Figure S1B). A significant decrease in luciferase activity was observed for the cells expressing the rs17481334 A variant of ERAP1 30UTR and the wild-type miR-UL112-5p, as well as for those expressing the rs17481334 G variant and the mutated miR-UL112-5p (Figure 1A). These data suggest that miRUL112-5p directly targeted ERAP1 in the consensus site and that this interaction is affected by the SNP. To test whether this differential binding could modulate ERAP1 expression, endogenous ERAP1 mRNA and protein levels were measured in fibroblasts isolated from individuals homozygous for ERAP1 rs17481334 A or G alleles equally transfected with wild-type or mutated miR-UL112-5p (Figure S1C). In these experiments, miR-UL112-5p induced 57% downregulation of ERAP1 mRNA expression in fibroblasts from the rs17481334 AA individual, but had no effect in rs17481334 GG fibroblast cells. Conversely, the mutated miR-UL112-5p had no effect in AA cells, whereas it induced 40% downregulation of ERAP1 mRNA expression in rs17481334 GG cells (Figure 1B). Of note, ERAP1 mRNA was reduced at similar levels in fibroblasts from both individuals by a siRNA targeting a coding region of ERAP1 (siERAP1) (88% and 91% in rs17481334 AA and GG fibroblasts, respectively) (Figure 1B). Virtually identical results were obtained in immunoblot experiments under the same conditions (Figure 1C). HCMV Infection Fails to Downregulate ERAP1 Expression in rs17481334 GG Fibroblasts Fibroblasts from rs17481334 AA and GG homozygous individuals were infected with the HCMV laboratory strain AD169 at two different multiplicity of infections (MOI). At 3 days post-infection (dpi), fibroblasts were infected at a similar level, as evaluated by expression of viral immediate-early antigens (IE1 and IE2) and MHC class I downregulation (Figure 2A). HCMV-infected cells sorted for reduced MHC class I expression (Figure 2B; Figure S2) showed a reduced overall amount of ERAP1 protein reaching a maximal decrease of 60% in fibroblasts Cell Reports 20, 846–853, July 25, 2017 847 from AA individuals at MOI 5, as compared to uninfected cells (Figure 2C). Conversely, no changes of ERAP1 protein expression were detected in HCMV-infected fibroblasts from the GG individual (Figure 2C), suggesting that the presence of G at the rs17481334 30UTR site prevents ERAP1 downregulation during HCMV infection. Interestingly, the overall amount of ERAP1 protein remained unchanged at different time points in cells infected with a modified HCMV AD169 strain (HCMVDUL112) deleted in the UL114 gene encompassing the miR-UL112-5p (Stern-Ginossar et al., 2007)(Figure 2D). rs17481334 Affects the ERAP1-Dependent HCMV pp65 495-503 CD8 + T Cell Epitope Presentation during HCMV Infection To directly demonstrate the functional relevance of rs17481334 during HCMV infection, we tested the capability of fibroblasts from AA and GG homozygous individuals to present the HLAA*0201-restricted immunodominant HCMV pp65 495-503 epitope, which is one of the well-known viral epitopes requiring ERAP1 activity for its in vitro generation (Urban et al., 2012). Thus, we isolated fibroblasts from AA and GG homozygous HLAA*0201 + individuals expressing similar levels of ERAP1 and cell surface HLA-A2 molecules, with a similar ability to process and present HCMV pp65 495-503 peptide to specific cytotoxic T lymphocytes (CTLs) (Figures S3A–S3G). Fibroblasts with these features were infected with wild-type HCMV or HCMVDUL112 at MOI 1, and their susceptibility to HLA-A*0201-restricted/ pp65 495-503 -specific CTLs was evaluated at 3 dpi. Although fibroblasts from either genotype were similar in relation to IE1/2 expression and MHC class I downregulation (Figure 3A), cells from the GG individual infected with wild-type HCMV and expressing unchanged levels of ERAP1 (Figure S3H), were killed more efficiently than those from the AA individual, at all effector/ target ratios tested (Figure 3B). This different susceptibility to CTL-mediated killing appeared to depend on miR-UL112-5p targeting of ERAP1, because it was not evident in fibroblast cells from AA and GG individuals infected with HCMVDUL112 (Figure 3B; Figure S3H). Consistently, GG fibroblasts stably knocked down for ERAP1 were killed less efficiently than control cells when infected with wild-type HCMV (Figures 3C and 3D; Figure S4). rs17481334 GG Genotype Associates with HCMV-Seronegative Profile in MS Patients The potential impact of the rs17481334 polymorphism in hostvirus interactions was studied by looking at HCMV serology data in relation to genotype, considering the GG homozygous combination the most likely to show biological effects. The G allele occurs at minor allele frequency (MAF) of 0.121 in Caucasians, with GG carriers representing 3% of the general population (ensembl.org). Inspection of genome-wide association study (GWAS) data from a Finnish general population previously studied in relation to HCMV seropositivity (Kuparinen et al., 2012) did not disclose any significant difference for rs17481334 GG homozygotes (data not shown). However, as shown in Figure 4, the same analysis performed in individuals suffering from MS, a condition where HCMV infection has been proposed to play a protective role (Sundqvist et al., 2014; Waubant et al., 2011), detected 5-fold decreased odds of HCMV seropositivity among GG patients compared to all other individuals (p = 0.021; odds ratio [OR] = 0.193 [95% confidence interval (CI), 0.041–0.913]). Figure 1. rs17481334 Variant G Abolishes HCMV miR-UL112-5p Binding to the 30UTR of ERAP1 (A) Luciferase assays of 293T cells transfected with dual-luciferase reporter vector bearing the rs17481334 A or G variant of ERAP1 30UTR and either a control scrambled sequence (miR-scr), miR-UL112-5p, or mutated miRUL112-5p (Mut miR-UL112-5p). Data are representative of five independent experiments. (B and C) Real-time qPCR (B) and representative immunoblotting analysis (C) of ERAP1 expression in fibroblasts with AA or GG genotypes, transfected with miR-scr, miR-UL112-5p, Mut miR-UL112-5p, or siERAP1. In (C), an ERp57 antibody (Ab) is used for normalization. Densitometric analysis of ERp57normalized ERAP1 values of three independent experiments is shown below. Mean ±SD; *p < 0.05, **p < 0.001, ***p < 0.0001. 848 Cell Reports 20, 846–853, July 25, 2017 DISCUSSION In this study, we demonstrated that HCMV miR-UL112-5p targets ERAP1 and that the rs17481334 G variant naturally occurring in the 30UTR of the ERAP1 gene preserves ERAP1 from miR-UL112-5p-mediated degradation. Accordingly, ERAP1 expression was reduced by miR-UL112-5p overexpression both at RNA and protein levels in human fibroblasts from AA individuals, but not in fibroblasts from GG individuals. Consistent Figure 2. rs17481334 Controls ERAP1 Expression in HCMV-Infected Fibroblasts (A) Representative flow-cytometric analysis of MHC class I cell surface expression and intracellular IE1/IE2 (IE1/2) viral protein expression in fibroblasts with AA or GG genotypes infected with HCMV AD169 strain at the indicated MOI, 3 days after infection. The percentage of cells in each quadrant is indicated. (B) Gating strategy used to sort AD169-infected cells based on the viral-mediated MHC class I downregulation. MHC class I expression in uninfected and infected cells is shown in light and dark gray histograms, respectively. (C) Representative immunoblotting analysis of ERAP1 and MHC class I expression in uninfected (U) or HCMV AD169-infected fibroblasts with AA or GG genotypes. Densitometric analysis of ERp57-normalized ERAP1 values relative to uninfected fibroblasts from three independent experiments is shown. (D) Representative immunoblotting analysis of ERAP1 in fibroblasts with AA genotype, either uninfected (U) or infected with wild-type HCMV AD169 (WT) or HCMVDUL112 (DUL112) at MOI of 5 at 24, 48, and 72 hr after infection. Densitometric analysis of ERp57-normalized ERAP1 values relative to uninfected fibroblasts from three independent experiments is shown. Mean ±SD; *p < 0.05, **p < 0.001. with these results, AA fibroblasts infected with the wild-type HCMV, but not with the mutant miR-UL112-5p knockout virus, showed a lower amount of ERAP1 protein. HCMV-infected GG fibroblasts were more efficient in trimming viral antigens to generate immunogenic epitopes and more susceptible to elimination by HCMV-peptide-specific CTLs. Hence, these data demonstrate that the rs17481334 GG genotype hampers the HCMV miR-UL112-5p-based immunoevasion strategy. HCMV has evolved to establish lifelong latent infections in immunocompetent individuals, with periodic and spontaneous reactivation. This asymptomatic coexistence is reflected by a prompt response of HCMV-specific T cells that dominate the memory compartment of exposed individuals, accounting for approximately 10% of total CD8 + T cells (Sylwester et al., 2005). Conversely, HCMV infection in individuals with either an immature or compromised immune system (e.g., preor post-natal infants and transplant recipients, leukemia, or HIV-infected patients, respectively) can cause different diseases, the severity of which depends on the degree of immunosuppression (Halenius and Hengel, 2014). For a virus to survive in an immunocompetent host, it is essential to avoid the presentation of viral epitopes to CD8 + T cells Cell Reports 20, 846–853, July 25, 2017 849 (Halenius et al., 2015). In this context, HCMV has evolved several strategies to target key components in the MHC class I antigenprocessing pathway, including ERAP1 (Kim et al., 2011). We attempted to explore this possibility with our pilot study of HCMV seropositivity in relation to ERAP1 rs17481334 genotype. Unsurprisingly, no significant finding could be detected at the level of the general (Finnish) population, although we observed a significant association when we investigated the relationship between GG genotype and HCMV serology using MS to ‘‘model’’ these interactions. Given the variant allele frequency and the size of the cohort examined, the obtained results, although limited to small number of individuals, were statistically significant. Although the true (if any) impact of GG genotype on HCMV infection, and eventually MS onset, can only be assessed via large-scale longitudinal epidemiological studies taking into account individual viral exposure, it is noteworthy that we observed a 5-fold decreased likelihood (OR = 0.193) of seropositivity among GG individuals with MS (that is, in a disease context where HCMV infection has been proposed to play a protective role) compared to other patients. Hence, although lacking mechanistic relation to causality, these results are intriguing because they suggest that important information may be gained by assessing multiple factors at once and open up novel lines of investigation, where the interplay between HCMV exposure, genotype-driven immune responses, and ultimately disease (autoimmunity/cancer) can be explored toward tailored preventive approaches. Consistently, Li et al. (2011) identified a novel link between HCMV infection and essential hypertension, in which miR-UL112 was independently associated with an increased risk of hypertension. Of note, ERAP1 is also known to play a role in the regulation of blood pressure through its involvement in the renin-angiotensin system (Hattori et al., 2000). Further studies are warranted to confirm and extend these findings, also in relation to the presence of multiple SNP-expression quantitative trait loci (eQTLs) at the ERAP1 locus, including rs17481334 where the G variant appears to associate with lower ERAP1 expression (https://gtexportal.org/). In addition to the pp65 495–503 epitope, ERAP1 is known to trim many other immunogenic HCMV epitopes (Kim et al., 2011). Experiments in the ERAAP / mouse model have demonstrated that ERAP1 controls the magnitude of CD8 + T cell response during murine cytomegalovirus (mCMV) infection (Blanchard et al., 2010). Although neither all HCMV nor mCMV-derived epitopes require ERAP1 activity for their generation, recent mass spectrometry analyses have demonstrated that ERAP1 has a Figure 3. rs17481334 Affects pp65 495-503 Epitope Presentation (A) Representative flow-cytometric analysis of cell surface MHC class I expression and intracellular IE1/2 viral protein expression of fibroblasts with AA or GG genotypes both uninfected (U) or infected at MOI 1 with wild-type HCMV AD169 (WT) or HCMVDUL112 (DUL112), 3 days after infection. The percentage of cells in each quadrant is indicated. (B) 51 Cr-release assay of HCMV-infected fibroblasts of (A) co-cultured with HLA-A2-restricted/pp65 495-503 -specific CTLs at the indicated effector/target (E:T) ratio. (C) Representative flow-cytometric analysis of cell surface MHC class I and intracellular IE1/2 viral protein expression of shERAP1or shCTRL-transduced GG fibroblasts infected with wild-type HCMV at MOI 1, 3 days after infection. The percentage of cells in each quadrant is indicated. (D) 51 Cr-release assay of HCMV-infected GG fibroblasts of (C) co-cultured with HLA-A2-restricted/pp65 495-503 -specific CTLs at the indicated effector/target (E:T) ratio. All data are representative of three independent experiments. Mean ±SD; *p < 0.05, **p < 0.001. 850 Cell Reports 20, 846–853, July 25, 2017 significant influence on the overall peptidome (Lo ´pez de Castro et al., 2016; Nagarajan et al., 2016). In conclusion, we demonstrate that a variant in the 30UTR of ERAP1, a key component of MHC class I antigen processing, impairs miR-UL112-5p-mediated mechanism of HCMV immunoevasion, with potential implications for the control of the viral infection itself and, eventually, indirect effects on the risk of HCMV-associated diseases. EXPERIMENTAL PROCEDURES 30UTR Luciferase Assay HEK293T cells were co-transfected with 1.6 mg of pEZX-MT01 vectors and 80 nmol/L miR-UL112-5p, mutated miR-UL112-5p (Mut miR-UL112-5p), or scrambled control (Ambion) using Lipofectamine 2000 (Invitrogen). After 48 hr, luciferase activity was measured with a LucPair Duo-Luciferase Assay Kit (Genecopoeia). Firefly luciferase activity was normalized to Renilla luciferase activity. Transfection, Viral Infections, and Peptide Pulsing Fibroblasts were transfected with 200 nM of either miR-scramble, miR-UL1125p, Mut miR-UL112-5p (Ambion), or a siRNA targeting a coding region of ERAP1 (Sigma-Aldrich) (Table S1) with Lipofectamine 2000. As control, cells were transfected with the same amount of an oligo-FITC (Invitrogen) in order to assess the efficiency at 24 hr post-transfection by fluorescence-activated cell sorting (FACS) analysis. Fibroblasts were infected with HCMV strains at 80%–90% confluence in serum-free medium. After 2–3 hr of virus adsorption at 37C, fresh growth medium was replaced (day 0). At 3 dpi, cells were harvested and analyzed. In some experiments, fibroblasts were infected either with a lentivirus containing pp65-GFP and sorted for GFP expression, or with a non-target shRNA control vector (SHC002), or the ERAP1 shRNA (TRCN0000060542) (Sigma-Aldrich) and selected with 2 mg/mL puromycin for 3 days. For peptide pulsing, cells were seeded in 96-well flat-bottom plates in triplicate and pulsed with 5 mg/mL pp65 495-503 peptide. After 2 hr, fresh growth medium was replaced. Proliferation Assay and Chromium-Release Assay For proliferation assays, infected or peptide-pulsed cells were seeded in 96-well flat-bottom plates in triplicate and incubated with HLA-A2-restricted/ pp65 495-503 -specific CD8 + T cells. After 72-hr incubation at 37C, cells were pulsed with 3 H-thymidine (1 mg Ci/well; specific activity, 6.7 Ci/mmol; PerkinElmer) and harvested after 18 hr. 3 H-thymidine incorporation was measured by standard procedures with Microbeta Trilux 1450 instrument (PerkinElmer). For cytotoxic activity, 51 Cr-labeled infected or peptide-pulsed target cells (5 310 3 per well) were mixed with HLA-A2-restricted/ pp65 495-503 -specific CD8 + T cells in 96-well plates in triplicate at various ratios and incubated at 37C. After 5 hr of incubation, supernatant was collected, and the 51 Cr release was measured by a TopCount NXT bdetector (PerkinElmer). All experimental groups were analyzed in triplicate, and the specific lysis was determined as follows: 100 3(experimental release spontaneous release)/ (total release spontaneous release). Spontaneous release lower than about 5% of the total release was observed in all assays. MS Patient Cohort Samples from 949 MS patients with ages between 16 and 70 years from the Epidemiological Investigation of Multiple Sclerosis (EIMS) were included in this study. EIMS is a nationwide population-based case-control study of incident cases of MS in Sweden (Hedstro ¨metal.,2011). All of the cases fulfilled the McDonald criteria for MS (Polman et al., 2005) and were examined and diagnosed by a neurologist at one of 30 recruiting centers in Sweden. Anti-HCMV serological status was determined in plasma using ELISA as described previously (Sundqvist et al., 2014). The HCMV antigen was prepared from embryonic fibroblasts infected with strain AD169 until complete cytopathic effect. Human plasma samples were analyzedin duplicates, ata dilution of 1/200. Alkaline phosphatase-conjugated, purified F(ab0) 2 fragment goat anti-human IgG was used as conjugate (Jackson ImmunoResearch). The plates were read at 405 and 650 nmin a Multiskan FC reader (ThermoFisherScientific). Cutoff was calculated by adding 0.2 optical density (OD)units to the mean absorbance value of160 runs of negative control samples. Plasma samples with OD values at cutoff were run in 2-fold dilutions (1/100–1/12,800) and then judged as positive or negative. Ethics Statement Ethical approval for generating and analyzing human fibroblasts from skin biopsies was obtained from the Institutional Review Board of the Hospital Bambino Gesu `. Written informed consent was obtained from adult patients, or the guardians of patients who were children. The EIMS study has been approved by the Regional Ethical Review Board in Stockholm, Sweden (http://www.epn.se). Written informed consent from all study participants or their parents was obtained. Investigations were carried out according to guidelines from the Declaration of Helsinki. Statistical Analysis Data are presented as mean, and error bars indicate the SD. Digital images of western blots were analyzed by ImageJ (https://imagej.nih.gov/ij/index.html), and statistical significance was assessed by the two-way ANOVA test. c 2 test was used to analyze the different distribution of rs17481334 genotype frequency in HCMV-positive and -negative Swedish MS cases. SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures, four figures, and one table and can be found with this article online at http:// dx.doi.org/10.1016/j.celrep.2017.06.084. AUTHOR CONTRIBUTIONS D.F., F.L., and M.D. designed the study. D.F. supervised the work. P.R., L.C., B.P., N.S., V.D., O.M., G.L.P., E.G., and C.C. performed the experiments. R.C. and H.H. provided critical reagents. M.B., T.B., L.A., T.O., I.K., I.S., T.L., M.A.H., and M.D. performed analysis of multiple sclerosis. P.R., L.C., C.C., M.D., and D.F. analyzed and interpreted the data. P.R., M.D., F.L., and D.F. wrote the manuscript. A.S. and C.C. edited the manuscript. ACKNOWLEDGMENTS We thank Dr. Stefano Petrocchi for kindly providing genomic DNA and Prof. Bodo Plachter and Peter van Endert for pcDNA6-pp65 plasmid and ERAP1 antibody, respectively. This work was supported by grants from the Italian Ministry of Health (PE-2011-02351866 to D.F.; RF-2010-2316606 to F.L.), the Associazione Italiana Ricerca sul Cancro (AIRC) (18495 to D.F. and Special Project 5x1000 9962 to F.L.), PRIN 2010 (to F.L.), PRIN 2015-W729WH (to C.C.), and Cordon de Vie (to F.L.). The EIMS study was supported by grants from the Swedish Research Council, the Swedish Research Council for Health and Working Life and Welfare, and the Knut and Alice Wallenberg Foundation. Figure 4. rs17481334 Genotype and HCMV Serology in MS Patients Serological anti-HCMV Ab profiles are reported for MS patients stratified according to rs17481334 genotype. Cell Reports 20, 846–853, July 25, 2017 851 Received: August 16, 2016 Revised: February 6, 2017 Accepted: June 28, 2017 Published: July 25, 2017 REFERENCES Blanchard, N., Kanaseki, T., Escobar, H., Delebecque, F., Nagarajan, N.A., Reyes-Vargas, E., Crockett, D.K., Raulet, D.H., Delgado, J.C., and Shastri, N. (2010). Endoplasmic reticulum aminopeptidase associated with antigen processing defines the composition and structure of MHC class I peptide repertoire in normal and virus-infected cells. J. Immunol. 184, 3033–3042. Cifaldi, L., Lo Monaco, E., Forloni, M., Giorda, E., Lorenzi, S., Petrini, S., Tremante, E., Pende, D., Locatelli, F., Giacomini, P., and Fruci, D. (2011). Natural killer cells efficiently reject lymphoma silenced for the endoplasmic reticulum aminopeptidase associated with antigen processing. Cancer Res. 71, 1597– 1606. Cifaldi, L., Romania, P., Falco, M., Lorenzi, S., Meazza, R., Petrini, S., Andreani, M., Pende, D., Locatelli, F., and Fruci, D. (2015). ERAP1 regulates natural killer cell function by controlling the engagement of inhibitory receptors. Cancer Res. 75, 824–834. Draenert, R., Le Gall, S., Pfafferott, K.J., Leslie, A.J., Chetty, P., Brander, C., Holmes, E.C., Chang, S.C., Feeney, M.E., Addo, M.M., et al. (2004). Immune selection for altered antigen processing leads to cytotoxic T lymphocyte escape in chronic HIV-1 infection. J. Exp. Med. 199, 905–915. Firat, E., Saveanu, L., Aichele, P., Staeheli, P., Huai, J., Gaedicke, S., Nil, A., Besin, G., Kanzler, B., van Endert, P., and Niedermann, G. (2007). The role of endoplasmic reticulum-associated aminopeptidase 1 in immunity to infection and in cross-presentation. J. Immunol. 178, 2241–2248. Grey, F., Meyers, H., White, E.A., Spector, D.H., and Nelson, J. (2007). A human cytomegalovirus-encoded microRNA regulates expression of multiple viral genes involved in replication. PLoS Pathog. 3, e163. Grey, F., Tirabassi, R., Meyers, H., Wu, G., McWeeney, S., Hook, L., and Nelson, J.A. (2010). A viral microRNA down-regulates multiple cell cycle genes through mRNA 50UTRs. PLoS Pathog. 6, e1000967. Halenius, A., and Hengel, H. (2014). Human cytomegalovirus and autoimmune disease. BioMed Res. Int. 2014, 472978. Halenius, A., Gerke, C., and Hengel, H. (2015). Classical and non-classical MHC I molecule manipulation by human cytomegalovirus: so many targets—but how many arrows in the quiver? Cell. Mol. Immunol. 12, 139–153. Hattori, A., Kitatani, K., Matsumoto, H., Miyazawa, S., Rogi, T., Tsuruoka, N., Mizutani, S., Natori, Y., and Tsujimoto, M. (2000). Characterization of recombinant human adipocyte-derived leucine aminopeptidase expressed in Chinese hamster ovary cells. J. Biochem. 128, 755–762. Hedstro ¨m, A.K., Sundqvist, E., Ba ¨a ¨rnhielm, M., Nordin, N., Hillert, J., Kockum, I., Olsson, T., and Alfredsson, L. (2011). Smoking and two human leukocyte antigen genes interact to increase the risk for multiple sclerosis. Brain 134, 653–664. Hook, L., Hancock, M., Landais, I., Grabski, R., Britt, W., and Nelson, J.A. (2014a). Cytomegalovirus microRNAs. Curr. Opin. Virol. 7, 40–46. Hook, L.M., Grey, F., Grabski, R., Tirabassi, R., Doyle, T., Hancock, M., Landais, I., Jeng, S., McWeeney, S., Britt, W., and Nelson, J.A. (2014b). Cytomegalovirus miRNAs target secretory pathway genes to facilitate formation of the virion assembly compartment and reduce cytokine secretion. Cell Host Microbe 15, 363–373. James, E., Bailey, I., Sugiyarto, G., and Elliott, T. (2013). Induction of protective antitumor immunity through attenuation of ERAAP function. J. Immunol. 190, 5839–5846. Kim, S., Lee, S., Shin, J., Kim, Y., Evnouchidou, I., Kim, D., Kim, Y.K., Kim, Y.E., Ahn, J.H., Riddell, S.R., et al. (2011). Human cytomegalovirus microRNA miRUS4-1 inhibits CD8 + T cell responses by targeting the aminopeptidase ERAP1. Nat. Immunol. 12, 984–991. Kim, Y., Lee, S., Kim, S., Kim, D., Ahn, J.H., and Ahn, K. (2012). Human cytomegalovirus clinical strain-specific microRNA miR-UL148D targets the human chemokine RANTES during infection. PLoS Pathog. 8, e1002577. Kim, S., Seo, D., Kim, D., Hong, Y., Chang, H., Baek, D., Kim, V.N., Lee, S., and Ahn, K. (2015). Temporal landscape of microRNA-mediated host-virus crosstalk during productive human cytomegalovirus infection. Cell Host Microbe 17, 838–851. Kuparinen, T., Seppa ¨la ¨, I., Jylha ¨va ¨, J., Marttila, S., Aittoniemi, J., Kettunen, J., Viikari, J., Ka ¨ho ¨nen, M., Raitakari, O., Lehtima ¨ki, T., and Hurme, M. (2012). Genome-wide association study does not reveal major genetic determinants for anti-cytomegalovirus antibody response. Genes Immun. 13, 184–190. Lee, S.H., Kalejta, R.F., Kerry, J., Semmes, O.J., O’Connor, C.M., Khan, Z., Garcia, B.A., Shenk, T., and Murphy, E. (2012). BclAF1 restriction factor is neutralized by proteasomal degradation and microRNA repression during human cytomegalovirus infection. Proc. Natl. Acad. Sci. USA 109, 9575–9580. Li, S., Zhu, J., Zhang, W., Chen, Y., Zhang, K., Popescu, L.M., Ma, X., Lau, W.B., Rong, R., Yu, X., et al. (2011). Signature microRNA expression profile of essential hypertension and its novel link to human cytomegalovirus infection. Circulation 124, 175–184. Lo ´pez de Castro, J.A., Alvarez-Navarro, C., Brito, A., Guasp, P., Martı ´n-Esteban, A., and Sanz-Bravo, A. (2016). Molecular and pathogenic effects of endoplasmic reticulum aminopeptidases ERAP1 and ERAP2 in MHC-I-associated inflammatory disorders: towards a unifying view. Mol. Immunol. 77, 193–204. Meshesha, M.K., Veksler-Lublinsky, I., Isakov, O., Reichenstein, I., Shomron, N., Kedem, K., Ziv-Ukelson, M., Bentwich, Z., and Avni, Y.S. (2012). The microRNA transcriptome of human cytomegalovirus (HCMV). Open Virol. J. 6, 38–48. Nagarajan, N.A., de Verteuil, D.A., Sriranganadane, D., Yahyaoui, W., Thibault, P., Perreault, C., and Shastri, N. (2016). ERAAP shapes the peptidome associated with classical and nonclassical MHC class I molecules. J. Immunol. 197, 1035–1043. Polman, C.H., Reingold, S.C., Edan, G., Filippi, M., Hartung, H.P., Kappos, L., Lublin, F.D., Metz, L.M., McFarland, H.F., O’Connor, P.W., et al. (2005). Diagnostic criteria for multiple sclerosis: 2005 revisions to the ‘‘McDonald Criteria’’. Ann. Neurol. 58, 840–846. Reeves, E., Colebatch-Bourn, A., Elliott, T., Edwards, C.J., and James, E. (2014). Functionally distinct ERAP1 allotype combinations distinguish individuals with Ankylosing Spondylitis. Proc. Natl. Acad. Sci. USA 111, 17594– 17599. Rehmsmeier, M., Steffen, P., Hochsmann, M., and Giegerich, R. (2004). Fast and effective prediction of microRNA/target duplexes. RNA 10, 1507–1517. Ryan, B.M., Robles, A.I., and Harris, C.C. (2010). Genetic variation in microRNA networks: the implications for cancer research. Nat. Rev. Cancer 10, 389–402. Saveanu, L., Carroll, O., Lindo, V., Del Val, M., Lopez, D., Lepelletier, Y., Greer, F., Schomburg, L., Fruci, D., Niedermann, G., and van Endert, P.M. (2005). Concerted peptide trimming by human ERAP1 and ERAP2 aminopeptidase complexes in the endoplasmic reticulum. Nat. Immunol. 6, 689–697. Stark, T.J., Arnold, J.D., Spector, D.H., and Yeo, G.W. (2012). High-resolution profiling and analysis of viral and host small RNAs during human cytomegalovirus infection. J. Virol. 86, 226–235. Stern-Ginossar, N., Elefant, N., Zimmermann, A., Wolf, D.G., Saleh, N., Biton, M., Horwitz, E., Prokocimer, Z., Prichard, M., Hahn, G., et al. (2007). Host immune system gene targeting by a viral miRNA. Science 317, 376–381. Stern-Ginossar, N., Weisburd, B., Michalski, A., Le, V.T., Hein, M.Y., Huang, S.X., Ma, M., Shen, B., Qian, S.B., Hengel, H., et al. (2012). Decoding human cytomegalovirus. Science 338, 1088–1093. Stratikos, E., Stamogiannos, A., Zervoudi, E., and Fruci, D. (2014). A role for naturally occurring alleles of endoplasmic reticulum aminopeptidases in tumor immunity and cancer pre-disposition. Front. Oncol. 4, 363. Sundqvist, E., Bergstro ¨m, T., Daialhosein, H., Nystro ¨m, M., Sundstro ¨m, P., Hillert, J., Alfredsson, L., Kockum, I., and Olsson, T. (2014). Cytomegalovirus seropositivity is negatively associated with multiple sclerosis. Mult. Scler. 20, 165–173. 852 Cell Reports 20, 846–853, July 25, 2017 Sylwester, A.W., Mitchell, B.L., Edgar, J.B., Taormina, C., Pelte, C., Ruchti, F., Sleath, P.R., Grabstein, K.H., Hosken, N.A., Kern, F., et al. (2005). Broadly targeted human cytomegalovirus-specific CD4+ and CD8+ T cells dominate the memory compartments of exposed subjects. J. Exp. Med. 202, 673–685. Tenzer, S., Wee, E., Burgevin, A., Stewart-Jones, G., Friis, L., Lamberth, K., Chang, C.H., Harndahl, M., Weimershaus, M., Gerstoft, J., et al. (2009). Antigen processing influences HIV-specific cytotoxic T lymphocyte immunodominance. Nat. Immunol. 10, 636–646. Tirabassi, R., Hook, L., Landais, I., Grey, F., Meyers, H., Hewitt, H., and Nelson, J. (2011). Human cytomegalovirus US7 is regulated synergistically by two virally encoded microRNAs and by two distinct mechanisms. J. Virol. 85, 11938–11944. Urban, S., Textoris-Taube, K., Reimann, B., Janek, K., Dannenberg, T., Ebstein, F., Seifert, C., Zhao, F., Kessler, J.H., Halenius, A., et al. (2012). The efficiency of human cytomegalovirus pp65(495-503) CD8 + T cell epitope generation is determined by the balanced activities of cytosolic and endoplasmic reticulum-resident peptidases. J. Immunol. 189, 529–538. van de Weijer, M.L., Luteijn, R.D., and Wiertz, E.J. (2015). Viral immune evasion: lessons in MHC class I antigen presentation. Semin. Immunol. 27, 125–137. Waubant, E., Mowry, E.M., Krupp, L., Chitnis, T., Yeh, E.A., Kuntz, N., Ness, J., Chabas, D., Strober, J., McDonald, J., et al.; US Pediatric MS Network (2011). Common viruses associated with lower pediatric multiple sclerosis risk. Neurology 76, 1989–1995. Yan, J., Parekh, V.V., Mendez-Fernandez, Y., Olivares-Villago ´mez, D., Dragovic, S., Hill, T., Roopenian, D.C., Joyce, S., and Van Kaer, L. (2006). In vivo role of ER-associated peptidase activity in tailoring peptides for presentation by MHC class Ia and class Ib molecules. J. Exp. Med. 203, 647–659. Cell Reports 20, 846–853, July 25, 2017 853