The Common R71H-G230A-R293Q Human TMEM173 Is a Null Allele

This article is featured in In This Issue, p.555

Early detection of invasive pathogens is achieved by germline-encoded innate immune sensors. TMEM173 encodes an endoplasmic reticulum–associated molecule MPYS (also known as MITA and STING) (1–3). MPYS is a cytosolic sensor for cyclic dinucleotides (CDNs) including bacterial CDNs, cyclic di-AMP (CDA), cyclic di-GMP (CDG), and mammalian CDN 2′5′-3′5′-cyclic GMP-AMP (2′3′-cGAMP) generated during cytosolic DNA sensing (4–6). Consequently, MPYS is critical for host defense against DNA viruses (7), RNA viruses (7, 8), intracellular bacteria (9, 10), and extracellular bacteria (11, 12) in mice. MPYS also plays a key role in the development of auto-inflammatory diseases in mice (13–15) and STING-associated vasculopathy with onset in infancy in humans (16, 17). Last, there are ongoing efforts to develop MPYS/STING-targeting immunotherapy for cancer and infectious diseases (11, 18–22).

We first showed that human TMEM173 gene has significant heterogeneity (23). We identified R232 of TMEM173, not H232, as the most prevalent allele (wild type, WT) in the human population (23). However, we found that only ∼50% of Americans are R232/R232 (23). We further identified HAQ, which contains three nonsynonymous single-nucleotide polymorphisms (SNPs), R71H-G230A-R293Q, as the second most common human TMEM173 allele and estimated that ∼3% of Americans are homozygous for HAQ (23). Transiently overexpressing HAQ in 293T cells leads to >90% decrease of type I IFN production, the hallmark function of MPYS/STING (23). 293T cells stably transfected with HAQ also have decreased response to CDN stimulation (6, 24). In this study, we examined the endogenous function of the HAQ allele using human cells from homozygous HAQ carriers and the knock-in mouse expressing a mouse equivalent of the HAQ allele (mHAQ). We discovered, unexpectedly, that the HAQ TMEM173 has decreased protein expression (∼90%) and did not respond to CDN stimulation in vivo and in vitro.

The linearized targeting vector (Supplemental Fig. 3A), which covers ∼10 kb of the genomic region in MPYS locus on mouse chromosome 18, was transfected into JM8A3.N1 embryonic stem cells originated from the C57BL/6 strain, followed by the selection for neomycin-positive and diphtheria toxin–negative clones. Targeted clones were screened by PCR. Positive embryonic stem clone was subjected to the generation of chimera mice by injection using C57BL/6J blastocysts as the host. The male chimeras (chimerism >95% determined by coat color) were mated with C57BL/6J female mice for germline transmission. Successful germline transmission was confirmed by PCR sequencing (Supplemental Fig. 3B). The heterozygous mice were bred to Actin-flpase mice [The Jackson Laboratory, B6.Cg-Tg(ACTFLPe)9205Dym/J] (Supplemental Fig. 3A) to remove the neo gene and make the HAQ-MPYS knock-in mouse. Animals were generated at the National Jewish Health Mouse Genetics Core Facility. Animal care and handling was performed according to institutional animal care and use committee guidelines.

For all experiments, 6- to 12-week old mice, both males and females, were used. MPYS^−/− mice (Tmem173^<tm1Camb>) have been described previously (25). All mice were on a C57BL/6 background. Mice were housed and bred in the Animal Research Facility at Albany Medical College and the University of Florida. All experiments with mice were performed following the regulations and approval of the Institutional Animal Care and Use Committee from Albany Medical College or the University of Florida.

The following reagent was obtained through BEI Resources, National Institute of Allergy and Infectious Diseases, National Institutes of Health: Streptococcus pneumoniae Family 1, Clade 2 pneumococcal surface protein A (PspA UAB055) with C-Terminal Histidine Tag, Recombinant from Escherichia coli, NR-33178.

Human TMEM173 genotype information was obtained from the 1000 Genomes Project (Phase III, http://browser.1000genomes.org/index.html). Human B cells with the corresponding TMEM173 genotypes were obtained from Coriell Cell Repositories (https://catalog.coriell.org/) and cultured in RPMI 1640 with 15% FCS, 2 mM l-glutamine, 37°C under 5% CO₂. Information related to TMEM173 gene expression was obtained from the Genotype-Tissue Expression (GTEx) project (http://www.gtexportal.org/home/).

For CDA (cat#vac-cda; Invivogen), CDG (cat#vac-cdg; Invivogen), and 2’3′-cGAMP (cat#vac-cga23; Invivogen) activation, human B cells were harvested and suspended (5 × 10⁶ cells/ml) in transporter buffer (26) (110 mM potassium acetate, 5 mM sodium acetate, 2 mM magnesium acetate, 1 mM EGTA, 2 mM DTT, 20 mM HEPE pH 7.3 and protease-inhibitor mixture (cat# B14011; BioTool)) with 10 μg/ml digitonin (cat# 300410; Calbiochem) in the presence or absence of CDNs (10 μg/ml). Cells were cultured at 37°C for 10 min in 24-well plates. Afterward, cells were harvested and resuspended in the human B cell culture medium at 5 × 10⁶ cells/ml, and cultured with or without CDNs (10 μg/ml) for 5 h. Human IFN-β was measured in cell supernatant by ELISA (cat#41415; PBL Bioscience).

To measure IRF3 nuclear translocation, cells were harvested at the end of 5 h incubation. Nuclear fraction was isolated as previously reported (27) and run on a 10% Mini-PROTEAN TGX gel (CAT#456-1035; BioRAD). Abs used for western blot were IRF3 Ab (cat# 43025; CellSignaling), α-rabbit IgG-HRP (cat#7074s; CellSignaling), α-mouse IgG-HRP (cat#7076s; CellSignaling), cyclophilinB Ab (cat#76952s; CellSignaling), Tubulin Ab (cat#200301-880; Rockland), and rabbit anti-MPYS polyclonal Ab (3).

To measure IRF3 phosphorylation, cells were harvested at the end of 5 h incubation, lysed in the radioimmunoprecipitation assay (RIPA) buffer as previous reported (27), run on a 10% Mini-PROTEAN TGX gel (CAT#456-1035; BioRAD), and probed for anti-p-IRF3 (s396) (cat# 4D4G; CellSignaling).

For RpRp-ssCDA (cat#c118-001; Biolog) activation, cells were suspended in human B cell medium at 5 × 10⁶ cells/ml. Then 5 μg/ml RpRp-ssCDA was added directly into a medium for 5 h. Afterward, IRF3 activation was examined as above.

Human B cells (1.2 × 10⁶) were harvested and lysed in 350 μl of RLT sample buffer with 40 μM DTT. Total RNA was extracted using the RNeasy Plus Mini kit (cat#74134; Qiagen) and reverse-transcribed using the high capacity reverse transcription kit (Applied Biosystems). Quantitative PCR (Q-PCR) was carried out on a StepOnePlus instrument (Applied Biosystems) using the following primers and probes: human αActin (Fwd: 5′-TCACCCACACTGTGCCCATCTACG-3′,Rev: 5′-CAGCGGAACCGCTCATT GCCAATG-3′) and SYBER-Green human TMEM173 (Assay ID: qHsaCID0010565, cat# 10025636; BioRad). Gene expression was normalized to Actin expression and relative expression of the respective gene in untreated cells.

Total RNA was extracted from human B cells using the RNeasy Plus Mini kit (cat#74134; Qiagen). Total cDNA was made using the Superscript IV First-Strand Synthesis System (#18091050; Invitrogen). Full-length human TMEM173 gene (1379 bp) was amplified with following primers: TMEM173-For: 5′-TTGGCTGAGTGTGTGGAGTC-3′; TMEM173-Rev: 5′-CAGTCCAGAGGCTTGGAGAC-3′. Human GAPDH primers (# RDP39; R&D Systems) were used to amplify the GAPDH cDNA as a control.

Bone marrow (BM) cells were cultured in RPMI 1640 (cat#11965; Invitrogen) with 10% FBS, 2 mM l-glutamine, 1 mM sodium pyruvate, 10 mM HEPES buffer, 1% nonessential amino acids, 50 μM 2-ME, 1% Pen/Strep, 20 ng/ml GM-CSF (cat# RP0407M; Kingfisher) or 20 ng/ml M-CSF (cat# RP0462M; Kingfisher). The medium was changed at day 3 and 6. At day 6, cells (1 × 10⁶) were transferred to a 24-well plate with fresh medium. Cells were activated at day 7 with 10 μg/ml CDA, CDG, 2’3′-cGAMP or 5 μg/ml Rp-Rp-ssCDA in culture directly. Mouse IFN-β was measured in culture supernatant after 5 h by ELISA (cat#42410; PBL Bioscience). Separately, BM-derived macrophages (BMDM) and BM-derived dendritic cells (BMDC) were activated with 5 μg/ml HSV DNA (cat# tlrl-hsv60n; Invivogen) and Vaccinia virus DNA (cat# tlrl-vav70n; Invivogen) transfected with lipofectamine 2000 (27) and mouse IFN-β was measured in culture supernatant after 5 h by ELISA. Alternatively, BMDC were activated with heat-killed S. pneumoniae (HKSP) (10⁸ CFU/ml) (cat# tlrl-hksp; Invivogen), LPS from Salmonella (25 ng/ml) (cat# L7261; Sigma), Imiquimod (4 ng/ml) (cat# tlrl-imqs; Invivogen), or CpG-ODN2395 (8 ng/ml) (cat# tlrl-2395; Invivogen). Mouse TNF-α and IFN-β were measured in culture supernatant after 5 h by ELISA.

Mice were intranasally (i.n.) administered 5 μg 2’3′-cGAMP (cat#vac-cga23; Invivogen), then sacrificed after 5 h by CO₂ asphyxiation (11). Lungs were perfused with cold PBS. The harvested lungs were washed once with PBS, then stored in 0.7 ml tissue protein extraction reagent (cat#78510; Thermo Scientific) containing protease inhibitors (cat#11836153001; Roche) at −80°C. Later, the lung was thawed on ice and homogenized with Minilys (Precellys, 5000 RPM for 30 s) using a Precellys lysing kit (cat# KT03961; Precellys). Lung homogenates were transferred to a 1.5 ml tube and spun at 14,000 g for 30 min at 4°C. The supernatant was collected and analyzed for cytokine production.

Cytokine concentrations were measured by ELISA kits from eBioscience. The ELISA kits used were IL-5 (cat#88-7054), IL-12/p70 (cat#88-7921), IL-13 (cat#88-7137), IL-17A (cat#88-7371), TNF-α (cat#88-7324), IFN-λ (cat#88-7284), and IFN-γ (cat#88-7314).

Groups of mice (four per group) were vaccinated i.n. with 5 μg 2’3′-cGAMP adjuvanted PspA (2 μg; BEI Resources) or PspA alone (11). Mice were immunized twice at a 14 d interval. For intranasal vaccination, animals were anesthetized using isoflurane in an E-Z Anesthesia system (Euthanex, Palmer, PA). PspA, with or without 2′3-cGAMP, was administered in 20 μl saline. Sera, bronchoalveolar lavage fluid, and nasal washes were collected 14 d after the last immunization. The PspA-specific Abs were determined by ELISA. Secondary Abs used were anti-mouse IgG1-HRP (cat#1070-05; Southern Biotech), anti-mouse IgG2C-HRP (cat#1079-05; Southern Biotech), and anti-mouse IgA-HRP (cat#1040-05; Southern Biotech). To determine the Ag-specific Th response, splenocytes from PspA or 2′3-cGAMP + PspA immunized mice were stimulated with 5 μg/ml PspA for 4 d in culture. Th1, Th2, and Th17 cytokines were measured in the supernatant by ELISA.

A group of mice (4 mice per group for the Ab experiment and 10 mice per group for the survival experiment) was i.m. administered with 0.125 μg of Pneumovax 23 (cat#7002681601; Merck) in 50 μl Ultrapure PBS (cat#K812; Amresco) or PBS alone. Blood was collected before and after immunization at the indicated time. Anti-pneumococcal polysaccharide type 2 (PPS2) and pneumococcal polysaccharide type 3 (PPS3) IgM were determined by ELISA. The following reagents were used: PPS3 (ATCC 31-X, ID # 61810463), PPS2 (ATCC 500-X, ID# 63406999), 1× ELISA assay diluent (REF # 00-4202-43; eBioscience), and goat anti-Mouse IgM HRP (cat # 1020-05; SouthernBiotech). One month after immunization, mice were challenged (i.n) with S. pneumoniae (A66.1 strain, serotype 3, ∼10⁶ CFU in 50 μl PBS). Animal health was monitored for 8 d.

The study was approved by the Ethics Committee of the Charité University Medicine Berlin, and participants gave informed written consent. Genomic DNA from individuals was isolated and genotyped for analysis of the HAQ haplotype. Three individuals carrying the HAQ haplotype in homozygosity and three carrying R232 (WT) TMEM173 were identified.

DNA from buccal swabs was extracted using a DNA mini kit (Qiagen). Genotyping was performed by PCR using fluorescence-labeled hybridization FRET probes and melting curve analysis employing the LightCyler 480TM (Roche Diagnostics). Genotyping of the TMEM173 SNP R71H was carried out using the following primer and probe: F-primer (rs11554776 S) 5′-GGAGTGACACACGTTGG-3′, R-primer (Rs11554776 A) 5′-GCCTAGCTGAGGAGCTG-3′, probe (rs11554776 C): 5′-CTGGAGTGGA-3′-XI-5′-TGTGGCGCAG-3′-PH. Primer and probes for the TMEM173 SNP R293Q were as follows: F-primer (rs7380824 F): 5′-ACCCTGGTAGGCAATGA-3′, R-primer (rs7380824 R): 5′-GCTTAGTCTGGTCTTCCTCTTAC-3′, sensor probe (rs7380824 C): 5′-CCTCAAGTGTCCGGCAGAAGAGTT-3′-FL, anchor probe (Anc rs7380824): 640-5′-GGCCTGCTCAAGCCTATCCTCCCGG-3′-PH.

Peripheral blood samples were drawn in 50 ml EDTA-coated syringes, and PBMCs were isolated by density gradient centrifugation using sterile-filtered Histopaque-1077 (Sigma-Aldrich). Cells were plated at a density of 1.2 × 10⁶ cell/well in a 24-well format and stimulated with 0.4 or 2 μg/well of Rp, Rp-ssCDA.

Total RNA was isolated from PBMCs lysates using the PerfectPure RNA Cultured Cell Kit (5prime) and reverse-transcribed using the high capacity reverse transcription kit (Applied Biosystems). Q-PCR was carried out on an ABI 7300 instrument (Applied Biosystems) using the following primers and probes: ifnb: F-primer: 5′-CCAACAAGTGTCTCCTCCAAATT-3′, R-primer: 5′-GTAGGAATCCAAGCAAGTGTAGCT-3′, probe: FAM-5′-TGTTGTGCTTCTCCACTACAGCTCTTTCCA-3′-TAMRA. Analysis of tmen173 expression was performed with a TaqMan gene expression assay Hs00736958_m1 (Applied Biosystems). Gene expression was normalized to GAPDH expression and relative expression of the respective gene in untreated cells.

All data are expressed as mean ± SEM. Statistical significance was evaluated using Prism 5.0 software to perform a Student t test (unpaired, two-tailed) for comparison of mean values.

We previously estimated that ∼3% of Americans are HAQ/HAQ (23). To expand this to other ethnic groups, we extracted TMEM173 genotype data from the 1000 Genomes Project (phase III). Among the five ethnic groups defined in the 1000 Genomes Project, we found that HAQ/HAQ is most common in East Asians (∼16.07%), followed by South Americans (∼7.78%), South Asians (∼6.75%), and Europeans (∼2.78%) (Table I). Surprisingly, no homozygous HAQ individual was found among Africans (Table I). Instead, ∼4.39% of Africans are AQ/AQ (G230A-R293Q), which is not found in non-Africans (Table I). We concluded that the human TMEM173 gene has not only great heterogeneity but also shows significant population stratification.

To study the function of HAQ, we obtained EBV-transformed human B cells from homozygous HAQ individuals identified in the 1000 Genomes Project. These cells are distributed by National Human Genome Research Institute Repository at the Coriell Institute. They express B cell surface markers IgM and HLA-DR (Supplemental Fig. 1A). Notably, these cells also express B cell activation markers CD80, CD86, and CD69 (Supplemental Fig. 1A). The expression level of these markers is similar between R232 (WT) and HAQ B cells (Supplemental Fig. 1A).

We next examined MPYS expression in these cells. Surprisingly, we found that the homozygous HAQ B cells from different ethnic groups have very low MPYS protein (Fig. 1A–D). No MPYS protein was detected in the cell debris (Supplemental Fig. 1B) excluding the possibility that HAQ protein somehow may be insoluble or aggregate.

We have been using this rabbit anti-MPYS Ab since we initially identify MPYS in 2008, and its specificity has been well documented in the literature (3, 9, 23, 25). Nevertheless, to exclude the possibility that our anti-MPYS Ab may not recognize the HAQ MPYS, we cloned the HAQ TMEM173 transcript from the homozygous HAQ human B cells and expressed it in 293T cells, which lack the endogenous MPYS expression (23). Our anti-MPYS Ab staining showed a similar expression of the HAQ and R232 of MPYS in the 293T cells (Supplemental Fig. 1C), which indicated that our anti-MPYS Ab recognizes the HAQ of MPYS is as good as the R232 of MPYS. We thus concluded that the low MPYS staining in the HAQ B cells (Fig. 1A−D) is indeed an indication of low MPYS protein expression.

We also compared the MPYS level in these 293T transfectants with our human B cells. We found that the endogenous MPYS level in human B cells is ∼50-fold lower than that of 293T transfectants (Supplemental Fig. 1C), which suggested that overexpressing TMEM173 in 293T cells likely masked the expression difference between the endogenous R232 and HAQ of TMEM173.

MPYS senses natural CDNs, including the bacterial CDN CDA, CDG, and mammalian CDN 2′3′-cGAMP (4, 5, 28, 29). We hypothesized that HAQ/HAQ cells would not respond to these CDNs due to low MPYS expression.

Directly adding CDN to the human B cell cultures did not activate these cells (Fig. 1E, 1G, Supplemental Fig. 2A, 2C, 2E, 2F, no digitonin). To deliver CDN into the cytosol, human B cells were reversibly permeabilized with digitonin in the presence of 2′3′-cGAMP. Activation of MPYS by CDNs leads to phosphorylation and nuclear translocation of IRF3 and subsequently IFN-β production. In Spanish and Chinese samples, 2′3′-cGAMP activates IRF3 translocation and IFN-β production in R232/R232 individuals but not the HAQ/HAQ B cells (Fig. 1E–H). Furthermore, 2′3′-cGAMP did not induce IRF3 phosphorylation in the HAQ/HAQ cells from the Chinese, Spanish, British, or Italian samples (Fig. 1I). Similar observations were made in HAQ/HAQ samples in response to CDA and CDG (Supplemental Fig. 2A–F).

CDN stimulation also activates MPYS-dependent NF-κB signaling (1, 27, 30). However, we found that these B cells have constitutively activated NF-κB as indicated by the presence of nuclear RelA and RelB (Supplemental Fig. 1D). CDN activation, which increases nuclear IRF3, did not further increase nuclear RelA or RelB (Supplemental Fig. 1D). This is consistent with our observation that these B cells have an activated phenotype (Supplemental Fig. 1A).

Recently, a synthetic CDN, RpRp-ssCDA, was shown to activate all major human TMEM173 variants overexpressed in 293T cells (18, 31). We next examined this synthetic CDN in human HAQ/HAQ B cells. First, we found, surprisingly, that RpRp-ssCDA can activate human B cells in a medium without the need of permeabilization (Fig. 2A, 2C, 2E, 2G). Second, HAQ/HAQ cells from British, Italian, Spanish, and Chinese individuals are all defective in IRF3 activation and IFN-β production in response to RpRp-ssCDA (Fig. 2A–H). Lastly, PBMC from three German HAQ/HAQ individuals also had a defective IFN-β response to RpRp-ssCDA compared with the R232/R232 individuals (Fig. 2I). RpRp-ssCDA can induce Type I IFN production in the Goldenticket mouse, which lacks detectable MPYS/STING protein (18); this may explain the residual IFN-β by RpRp-ssCDA in some samples. We concluded that homozygous HAQ B cells are defective in response to the synthetic CDN RpRp-ssCDA.

The H232 of MPYS has a low binding affinity for CDNs (6). We found that MPYS expression in H232/H232 human B cells is similar to R232/R232 whereas the HAQ/H232 B cells have decreased MPYS expression likely due to the presence of the HAQ allele (Fig. 1A–D). We next examined their CDN responses. We found that both the homozygous H232 and the HAQ/H232 B cells did not have IRF3 nuclear translocation and IFN-β production in response to 2′3′-cGAMP, CDA, CDG or RpRp-ssCDA (Fig. 1E–H, Fig. 2, Supplemental Fig. 2). We concluded that, similar to the homozygous HAQ B cells, homozygous H232, and the HAQ/H232 B cells are also defective in response to CDN.

Mouse and human MPYS proteins are 82% homologous (2). To understand the in vivo significance of the HAQ of TMEM173, we generated an mHAQ mouse. This knock-in mouse contains a mouse equivalent of the HAQ mutations: C71H, I229A, and R292Q (Supplemental Fig. 3A). The presence of these three mutations was confirmed by sequencing (Supplemental Fig. 3B). Similar to the human HAQ B cell, we found that MPYS expression is also decreased in mHAQ spleen B cells (Fig. 3A), which suggested that the mHAQ mouse recapitulates the main feature of the human HAQ.

The establishment of the mHAQ mouse allowed us to examine HAQ expression and function beyond the B cells. Indeed, we found that MPYS expression is decreased in mHAQ spleen T cells (Fig. 3A) and mHAQ BM Ly6C^hi monocytes (Fig. 3B). The defect is more pronounced in the mHAQ lung (Fig. 3C) and mHAQ BMDC (Fig. 3D) where MPYS expression is not detectable. Notably, in naive BMDM, IFN-γ differentiated M1 macrophage or IL-4 differentiated M2 macrophage, MPYS expression is similar for WT and mHAQ mice (Fig. 3E), which may indicate a macrophage-specific regulation of MPYS protein expression.

Human HAQ B cells are nonresponsive to CDNs (Fig. 1, Supplemental Fig. S2). We next examined CDNs response in BMDC and BMDM from the mHAQ mouse. As expected, both mHAQ BMDM and BMDC did not produce IFN-β in response to CDA, CDG, 2′3′-cGAMP, or RpRp-ssCDA (Fig. 4A, 4B). The mHAQ BMDM and BMDC also did not make IFN-β in response to transfected HSV DNA or Vaccinia virus DNA (Fig. 3A, 3B). As a control, BMDC from mHAQ mice had similar TNF-α (Fig. 3F) and IFN-β (Fig. 3G) production as the WT mice when stimulated with TLR2 ligand HKSP, TLR4 ligand LPS, TLR7 ligand imiquimod, and TLR9 ligand CpG-ODN2395.

We next examined the in vivo CDN responses in the mHAQ mouse. Intranasal administration of CDN elicits rapid cytokine production in the lung, which is important for the mucosal adjuvant activity of CDNs (11). We found that intranasal administration of 2′3′-cGAMP did not elicit lung production of TNF-α, IL-12p70, IFN-γ, or IFN-λ in the mHAQ mouse (Fig. 4C). We further examined the mucosal adjuvant activity of 2′3′-cGAMP in the mHAQ mouse (32). As expected, 2′3′-cGAMP did not induce Ag-specific Ab or Th response in the mHAQ mouse (Fig. 4D, 4E). We concluded that the mHAQ mouse does not respond to CDN in vivo and in vitro.

The CDNs-MPYS/STING activation in B cells is required for polysaccharide-based vaccine activity such as Pneumovax 23 (33). Because mHAQ mice do not have a functional CDNs-MPYS pathway (Fig. 4), we hypothesized that Pneumovax 23 would not be effective in the mHAQ mouse. Indeed, upon i.m. Pneumovax 23 immunization, the mHAQ mice have lower anti-PPS3 IgM (Fig. 5A, 5B) and anti-PPS2 IgM (Fig. 5C, 5D) production than the WT mice at days 14 and 21.

To examine the protective immunity of Pneumovax 23 in the mHAQ mouse, we challenged vaccinated mice with the A66.1 strain, an invasive strain of Streptococcus pneumoniae. Consistent with the Ab results, Pneumovax 23 protected WT mice from the A66.1 S. pneumococcus infection (Fig. 5E) but not the mHAQ mouse (Fig. 5F). We concluded that Pneumovax 23 is not effective in the mHAQ mouse.

We next ask why the human HAQ allele has low MPYS protein expression. We examined the TMEM173 mRNA level in the homozygous HAQ B cells. Surprisingly, we found that HAQ B cells from Sri Lankan Tamil, Colombian, Japanese, and Italian individuals, which have low MPYS expression (Fig. 6A), all have ∼40% lower TMEM173 mRNA than their R232/R232 ethnic controls (Fig. 6B). Semiquantitative PCR also show that the full-length human TMEM173 transcript (∼1.4 kb) is decreased in homozygous HAQ B cells compared with their R232 counterparts (Fig. 6C). We concluded that human HAQ B cells decreased TMEM173 transcript.

We next asked if the homozygous HAQ individuals have decreased TMEM173 transcript in tissues other than B cells. To answer that, we mined data from the GTEx database, which compiles data on human gene expression related to genetic variations. We focused on the SNPs that affect the TMEM173 transcript. We found that all three HAQ SNPs, rs11554776(R71H), rs78233829(G230A), rs7380824(R293Q), are associated with decreased TMEM173 transcript with highly significant p values as low as 10⁻¹⁹, 10⁻²², and 10⁻²³ (Table II). Furthermore, this decreased TMEM173 transcript in HAQ individuals can be found in non–B cell dominant tissues such as in the artery, fibroblasts, lung, thyroid and esophagus (Table II). Thus, homozygous HAQ individuals have decreased TMEM173 in tissues other than B cells.

The common human HAQ TMEM173 allele was first identified and characterized by us in 2011 (23). We characterized the HAQ as a loss-of-function TMEM173 allele because it loses >90% of the ability to stimulate IFN-β production when transiently overexpressed in the 293T cells, a hallmark function of MPYS/STING (23). In 2013, Diner et al. (5) found that the THP-1 cell, a human monocytic cell line originated from a Japanese individual (34), has the HAQ of TMEM173. However, it is not clear if the THP-1 cells are homozygous or heterozygous for HAQ. Also in 2013, Yi et al. (24) found that 293T cells stably expressing the HAQ can respond to CDN, albeit weaker than the R232 of TMEM173. Using 293T cell stable transfectants to study HAQ function, nevertheless, could be misleading. This is because: 1) the level of MPYS is 50-fold higher in the 293T cell transfectants than the endogenous MPYS (Supplemental Fig. 1C); and 2) human HAQ cells have a decreased TMEM173 transcript (Fig. 6). This feature of the HAQ is lost when expressing the HAQ cDNA in the 293T cells. In the current report, we used homozygous HAQ human B cells from multiple ethnic groups, which showed that homozygous HAQ B cells have very low MPYS expression compared with the R232 B cells and do not respond to CDN in vitro. Furthermore, PBMCs from three homozygous HAQ Germans have a decreased CDN response compared with the R232 German. Lastly, an HAQ knock-in mouse has decreased MPYS expression and did not respond to CDN in vitro and in vivo. Thus, HAQ is indeed a loss-of-function human TMEM173 allele likely due to its extremely low protein expression.

Two other TMEM173 genotypes, HAQ/H232 and H232/H232, also did not respond to CDN (Figs. 1, 2). The underlying molecular mechanisms are likely different. Unlike the HAQ/HAQ, the H232/H232 B cells have similar MPYS expression as the R232/R232 B cells. Previous studies found that the H232 of MPYS does not bind CDN (K_d ∼5.3 μM) as well as the R232 of MPYS (K_d ∼0.11 μM) (6). Consequently, the H232 is severely defective in response to CDN stimulation when expressed in 293T cells (6, 29). Here, we verified this observation in our homozygous H232 human B cells (Fig. 1). The HAQ/H232 B cells have the HAQ allele, which contributes to its low MPYS expression (Fig. 1), and the nonfunctional H232 allele. Together, they lead to the unresponsiveness to CDN in the HAQ/H232 B cells.

It is worth noting that the HAQ/HAQ, HAQ/H232, and H232/H232 genotypes are made up of ∼10% Europeans and ∼31% East Asians (Table I). This is significant because there is tremendous interest in developing MPYS/STING-targeting immunotherapies for cancers and infectious diseases (18, 19, 21, 22, 35). It will be especially challenging to develop MPYS/STING-targeting immunotherapies for the homozygous HAQ individuals because of their extremely low MPYS protein expression. Indeed, we have showed that the licensed pneumococcal vaccine Pneumovax 23 is less effective in the mHAQ mouse than in WT mice (Fig. 5). Fu et al. (18) did show that the synthetic CDN RpRp-ssCDA activates PBMCs similarly in HAQ/HAQ and R232/R232 donors. However, the PBMC was from a single HAQ/HAQ donor (18). In this study, we used three homozygous German HAQ individuals and found they were defective in response to RpRp-ssCDA (Fig. 2I). Future development of MPYS-targeting immunotherapies must adopt the concept of personalized medicine.

Surprisingly, we found that the synthetic CDN RpRp-ssCDA is membrane permeable. Natural CDNs have two phosphate groups, preventing them from directly passing through the cell membrane. To activate MPYS, which is inside cells, investigators have to use transfection or membrane permeabilizing reagents to deliver CDN to the cytosol. We previously showed that, in vivo, only pinocytosis-efficient cells such as macrophage and dendritic cells can directly take up CDG and be activated (11). The observation that RpRp-ssCDA is cell permeable makes it a very attractive CDN to direct activate cells that are not pinocytosis efficient, such as B cells.

Finally, the null phenotype of the HAQ allele is likely a result of the decreased TMEM173 transcript and the amino acids changes (R71H-G230A-R293Q) in the HAQ protein (Fig. 6D). The MPYS protein level is down ∼60% in the mHAQ knock-in mouse. Previously, an I199N change in the mouse TMEM173 gene led to a complete loss of STING/MPYS protein expression (36). Thus, amino acid changes in MPYS protein can impact its expression. MPYS expression can also be regulated at a transcriptional level. Mouse and human TMEM173 genes have conserved STAT1 binding sites (37). Type I and II treatments increase mouse and human TMEM173 expression via a STAT-1 dependent mechanism (37). Nevertheless, treating homozygous human HAQ B cells with IFN-γ did not restore MPYS protein expression (data not shown). Further studies are needed to reveal the mechanisms by which human TMEM173 expression is controlled on the transcriptional and posttranscriptional level.

In summary, we found that human HAQ, the second most common TMEM173 allele, is a null allele. The mouse model of HAQ, the mHAQ knock-in mouse, is not protected by Pneumovax 23. Future studies are needed to determine the impact and mechanisms by which HAQ, as a loss-of-function common TMEM173 allele, influences human diseases, and medicines. Our mHAQ knock-in mouse will be especially valuable in this endeavor.

The authors have no financial conflicts of interest.