Estrogen Reverses HDAC Inhibitor-Mediated Repression of Aicda and Class-Switching in Antibody and Autoantibody Responses by Downregulation of miR-26a

Estrogen contributes to females' strong antibody response to microbial vaccines and proneness to autoimmunity, particularly antibody-mediated systemic autoimmunity, in females. We have hypothesized that this is due to estrogen-mediated potentiation of class switch DNA recombination (CSR) and somatic hypermutation (SHM). As we have shown, estrogen boosts AID expression, which is critical for both CSR and SHM, through upregulation of HoxC4, which together with NF-κB critically mediates Aicda (AID gene) promoter activation. We contend here that additional regulation of Aicda expression by estrogen occurs through epigenetic mechanisms. As we have shown, histone deacetylase inhibitors (HDIs) short-chain fatty acid (SCFA) butyrate and propionate as well as the pharmacologic HDI valproic acid upregulate miRNAs that silence AID expression, thereby modulating specific antibody responses in C57BL/6 mice and autoantibody responses in lupus-prone MRL/Faslpr/lpr mice. Here, using constitutive knockout Esr1−/− mice and B cells as well as conditional knockout Aicdacre/creEsr1flox/flox mice and B cells, we showed that the HDI-mediated downregulation of Aicda expression as well as the maturation of antibody and autoantibody responses is reversed by estrogen and enhanced by deletion of ERα or E2 inhibition. Estrogen's reversion of HDI-mediated inhibition of Aicda and CSR in antibody and autoantibody responses occurred through downregulation of B cell miR-26a, which, as we showed, targets Aicda mRNA 3′UTR. miR-26a was significantly upregulated by HDIs. Accordingly, enforced expression of miR-26a reduced Aicda expression and CSR, while miR-26a-sponges (competitive inhibitors of miR-26a) increased Aicda expression and CSR. Thus, our findings show that estrogen reverses the HDI-mediated downregulation of AID expression and CSR through selective modulation of miR-26a. They also provide mechanistic insights into the immunomodulatory activity of this hormone and a proof-of-principle for using combined ER inhibitor-HDI as a potential therapeutic approach.


INTRODUCTION
Estrogen plays an important role in boosting the production of mature antibodies and autoantibodies (1,2). This provides an explanation for the greater antibody response to microbial vaccines and infections as well as the greater incidence of antibody-mediated autoimmunity, such as systemic lupus erythematosus, in females (3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17). Like protective antibodies against pathogenic microorganisms, pathogenic autoantibodies in humans and mice with systemic autoimmunity are generally class-switched and hypermutated, strongly suggest a role for estrogen in modulation of immunoglobulin (Ig) class switch DNA recombination (CSR) and somatic hypermutation (SHM). Ig CSR recombines S-S regions, generally Sµ with Sγ, Sα, or Sε, located upstream of exons of constant heavy-chain (C H ), thereby encoding a new C H region that endows an antibody with new biological effector functions (18). SHM introduces mainly point-mutations in Ig V(D)J segments, thereby providing the structural substrate for antigen-mediated selection of B cell mutants with higher affinity B cell receptors (BCRs). CSR and SHM are initiated by activation-induced cytidine deaminase (AID, encoded by Aicda in mice and AICDA in humans), which is expressed in B cells in a differentiation stage-specific fashion (19)(20)(21). As a potent DNA mutator, AID must be tightly regulated to prevent offtargeting effects, which can result in mutations in non-Ig genes, genomic instability, interchromosomal translocations and cellular neoplastic transformation (21).
Epigenetic mediators influence gene expression without modifying the genomic sequence. As we have suggested, such mediators, including DNA methylases, histone posttranslational modifiers, such as methyltransferases and acetyltransferases and non-coding RNAs, such as microRNAs (miRNAs), modulate B cell functions. They interact with genetic programs to regulate B cell functions, such as CSR, SHM and plasma cell differentiation, thereby informing the antibody response (1,2,22). We have shown that in addition to DNA methylation and histone acetylation in the Aicda locus, select miRNAs also provide an important mechanism for modulation of AID expression. miRNAs likely play important roles in B cell development, peripheral differentiation, and autoimmunity (2,(23)(24)(25). In B cells, miR-155, miR-181b, and miR-361 repress Aicda expression, while miR-30a and miR-125b repress Prdm1 expression-Prdm1 is the gene that encodes Blimp1, the plasma cell differentiation master transcription factor (23,24). By binding to the evolutionarily conserved miRNA target sites in the 3 ′ UTR of Aicda and Prdm1 mRNAs, these miRNAs cause degradation of the mRNA transcripts and/or inhibit their translation (2,26). As we have also shown, the expression of Aicda-or Prdm1targeting miRNAs can be modulated by histone deacetylase (HDAC) inhibitors (HDIs), including short-chain fatty acids Abbreviations: AID, activation-induced cytidine deaminase; CSR, class switch DNA recombination; ER, estrogen receptor; ERE, estrogen response element; GC, germinal center; HDI, histone deacetylase inhibitor; IgH, immunoglobulin heavy chain locus; miRNA, microRNA; NP-CGG, 4-hydroxy-3-nitrophenyl acetylconjugated chicken γ-globulin; SHM, somatic hypermutation.
Estrone (E1), estradiol (E2), estriol (E3), and estetrol (E4) are the four major naturally occurring estrogens in females. E2 is the predominant estrogen in terms of both absolute circulating levels and activity, and, as we contend here, plays an important role in antibody and autoantibody responses. We have previously shown that estrogen activates AID by upregulating the homeodomain transcription factor HoxC4, which interacts with a conserved HoxC4/Oct binding site in the AICDA/Aicda promoter (21,27,28). At the transcriptional level, we have shown that estrogen-estrogen receptor (ER) complexes bind to three cooperative evolutionarily conserved estrogen response elements (EREs) in the HOXC4/HoxC4 promoter and synergize with the signaling of CD154 or LPS and IL-4 to up-regulate HoxC4 expression, thereby inducing AID and CSR (28). ERs (ERα and ERβ, encoded by Esr1 and Esr2, respectively) are liganddependent transcription factors that modulate gene transcription through recruitment to the target gene locus-ERα and ERβ may play different role in B cell differentiation and function (29). ER signaling can contribute to epigenetic changes (30). Indeed, estrogen is known to induce histone modifications, including histone acetylation, phosphorylation, and methylation, at the ER target gene promoters through dynamic interactions with histone modifying enzymes (30). Estrogen has been shown to suppress or stimulate miRNA expression and post-translationally regulate protein expression in human breast cancer cells, endometrial cells, mouse uterus cells, and rat mammary gland (31,32). In ER + breast cancer cells, estrogen reduces miR-26a expression (33).
Here, we addressed the role of estrogen in post-transcriptional regulation of AID expression and CSR in mouse B cells in vivo and in vitro in the presence of HDIs VPA, butyrate and propionate using Esr1 −/− mice/B cells, conditional knockout Aicda cre Esr1 fl/fl mice we generated by crossbreeding Aicda cre mice with Esr1 fl/fl mice, as well as anti-estrogen drugs, including fulvestrant (a selective ER degrader, SERD) and Letrozole (an aromatase inhibitor that also inhibits endogenous estrogen synthesis). As epigenetic modifiers, SCFA HDIs inhibit Aicda expression and CSR through upregulation of select B cell miRNAs that silence Aicda, including miR-26a and miR-125a, in antibody and autoantibody responses (25). We found that estrogen abrogated the HDI-mediated inhibition of AID expression and CSR through ERα. Using Aicda cre Esr1 fl/fl mice and Esr1 −/− B cells, we showed that activated B cell-specific ERα deletion and SCFAs synergized to decrease AID expression and class-switched antibody responses. The combined impact of estrogen and VPA on class-switched antibody responses was assessed in normal mice antibody response and in the autoantibody response of lupus-prone MRL/Fas lpr/lpr mice. Further, we analyzed how estrogen affected the role of HDIs as epigenetic modifiers, and found that ERα bound to ER-binding cis-elements in the miR-26a CTDSPL and CTDSP1 host gene promoters, thereby inhibiting the expression of such a miRNA. Thus, estrogen/ERα provides an additional layer of epigenetic regulation of AID expression, as mediated by miR-26a that targets Aicda mRNA 3 ′ UTR.  (36), which carry loxP sequences in the Esr1 gene flanking exon 3 that encodes a conserved zinc finger type DNA binding domain, were obtained from Dr. J.-A. Gustafsson (Karolinska Institutet, Sweden). In BAC transgenic Aicda cre mice, the bacterial cre recombinase gene was introduced in lieu of Aicda exon 1 in a supplementary Aicda locus and under the control of the Aicda promoter/enhancers within the BAC transgene (35). Aicda cre Esr1 +/+ and Aicda cre Esr1 fl/fl mice were generated by crossbreeding Aicda cre with Esr1 fl/fl mice. Esr1 −/− mice, and their Esr1 +/+ littermates were generated by breeding Esr1 −/− mice. These mice were viable and fertile, and show no gross abnormalities. All mice were housed under pathogen-free conditions, and provided with autoclaved food and deionized water. The Institutional Animal Care and Use Committees of the University of Texas Health Science Center, San Antonio approved all animal protocols.

SCFA HDAC Inhibitors, Estrogen, and Estrogen/ER Inhibitors
For in vivo studies, VPA sodium salt (VPA, Sigma-Aldrich) was dissolved in drinking water at 0.8% w/v (HDI-water). This yields a stable VPA serum level (400-600 µM) in mice, comparable to the serum concentration in humans under long-term VPA treatment (300-900 µM) (23,37,38). For SCFA treatment, sodium butyrate (140 mM, Sigma-Aldrich) and tributyrin emulsion (20.0 mg/ml), a prodrug of natural butyrate, which is a triglyceride containing three butyrate moieties that has low toxicity levels and is rapidly absorbed and hydrolyzed to butyrate by plasmatic esterases were added to the drinking waters (SCFA-water). Drinking water containing VPA or butyrate and tributyrin at the above concentration was always well-accepted by the mice. Exogenous estrogen E2 was made 1.0 µM by adding 40 mM E2 stock (dissolved in 100% ethanol) to drinking water. To inhibit ERα receptor and estrogen synthesis, some mice were injected subcutaneously every other day with fulvestrant (500 µg per injection) and letrozole (5 µg per injection). For in vitro experiments, VPA (250 or 500 µM), sodium butyrate (butyrate, 500 µM), sodium propionate (propionate, 2000 µM) and/or E2 (10, 30, or 100 nM) were directly diluted in culture medium.

NP 16 -CGG Immunization and Titration of Total and NP-Binding IgM and IgG1
Female C57BL/6, Esr1 −/− and Esr1 +/+ mice (8-12 weeks of age) were given water containing nil or VPA ad libitum 1 week before NP 16 -CGG (average 16 molecules of 4-hydroxy-3-nitrophenyl acetyl coupled to 1 molecule of chicken γ-globulin; Biosearch Technologies) immunization until end of the experiments. For NP 16 -CGG immunization, the mice were injected i.p. with 100 µg of NP 16 -CGG in 100 µl of alum (Imject R Alum, Pierce). Serum was collected 10 days later for titration of circulating total and NP-binding IgM and IgG1 using enzyme-linked immunosorbent assays (ELISAs), as we described (27,(39)(40)(41). Female Aicda cre Esr1 fl/fl and Aicda cre Esr1 +/+ mice were given plain water or SCFA-water containing 140 mM sodium butyrate and 20.0 mg/ml tributyrin starting at the age of 5 weeks and i.p injected with NP 16 -CGG in 100 µl of alum at the age of 8 weeks. The mice were boost injected with NP 16 -CGG in PBS 3 weeks later. Serum total and NP 4 -binding IgM and IgG1 were measured by ELISA 7 days after boost injection.

Luciferase 3 ′ UTR Reporter Assays
The 3 ′ UTRs of Aicda mRNA (NM 009645.2, National Center for Biotechnology Information) was PCR amplified using Phusion DNA polymerase (New England BioLabs.) from spleen B cell cDNA and cloned into the pMIR-REPORT miRNA Expression Reporter Vector System (Invitrogen), which allows for analysis of 3 ′ UTR-mediated regulation of firefly luciferase activity. The mutant (mut) Aicda 3 ′ UTR containing point mutations in the seed sequence of miR-26a target site was generated by PCRbased mutagenesis of the Aicda 3 ′ UTR pMIR-REPORT vector. The sequence of constructs was confirmed by two independent sequencing reactions. Reporter constructs were cotransfected with the pRL-TK vector (Promega), which drives constitutive expression of Renilla reniformis luciferase, into mouse CH12F3 B cells by electroporation (250 V and 900 ) with a Gene Pulser II (Bio-Rad). Transfected CH12F3 B cells were then stimulated with CD154 (1 U/ml), IL-4 (5 ng/ml), and TGF-β (2 ng/ml) to induce AID expression and CSR to IgA) in the presence nil, E2 (30 nM), butyrate (500 µM) or butyrate (500 µM) plus E2 (30 nM) for 48 h. The abilities of butyrate and E2 to modulate reporter activity were determined by firefly luciferase activity and normalized to Renilla luciferase activity, according to the manufacturer's instructions, using the Luc-Pair TM Duo-Luciferase HS Assay Kit (GeneCopoeia).

Quantitative RT-PCR (qRT-PCR) of mRNAs and miRNAs
For quantification of mRNA, pri-miRNA, germline I H -C H transcripts, post-recombination Iµ-C H transcripts and mature V H DJ H -C H transcripts, RNA was extracted from 0.2 to 5 × 10 6 cells using either Trizol R Reagent (Invitrogen) or RNesy Plus Mini Kit (Qiagen). Residual DNA was removed from the extracted RNA using gDNA eliminator columns (Qiagen). cDNA was synthesized from total RNA with the SuperScript TM III First-Strand Synthesis System (Invitrogen) using oligo-dT primer. Transcript expression was measured by qRT-PCR with the appropriate primers (Supplemental Table 1) using a Bio-Rad MyiQ TM Real-Time PCR Detection System (Bio-Rad Laboratories) to measure SYBR Green (IQ TM SYBR R Green Supermix, Bio-Rad Laboratories) incorporation with the following protocol: 95 • C for 15 s, 40 cycles of 94 • C for 10 s, 60 • C for 30 s, 72 • C for 30 s. Data acquisition was performed during 72 • C extension step. Melting curve analysis was performed from 72 to 95 • C. For quantification of mature miRNA transcripts, RNA was extracted from 0.5 to 5 × 10 6 cells using miRNeasy R Mini Kit (Qiagen) and then reverse-transcribed with miScript II RT Kit (Qiagen) using the miScript HiSpec buffer. A Bio-Rad MyiQ TM Real-Time PCR Detection System was used to measure SYBR Green (miScript SYBR Green PCR Kit; Qiagen) incorporation according to manufacturer's instructions. Mature miRNA forward primers (Supplemental Table 1) were used at 250 nM in conjunction with the Qiagen miScript Universal Primer and normalized to expression of small nuclear/nucleolar RNAs Rnu6/RNU61/2, Snord61/SNORD61, Snord68/SNORD68, and Snord70/SNORD70. The Ct method was used for data analysis of qRT-PCR experiments.

Lupus Mice: Autoantibodies, Pathology and Disease
MRL/Fas lpr/lpr mice were started on HDI-water ad libitum at 6or 17-weeks of age, or were on untreated water throughout their life and scarified when moribund. Anti-nuclear antibody (ANA) and anti-dsDNA antibody titers were determined in sera. For ANA assays, sera were serially diluted in PBS (from 1:40 to 1:160), incubated on antinuclear Ab substrate slides (HEp-2 cell-coated slides, MBL-BION) and detected with a 1:1 mixture of FITCanti-IgG1 and FITC-anti-IgG2a mAbs (R19-15; BD Biosciences). Images were acquired with a 40× objective on an Olympus CKX41 fluorescence microscope. Anti-dsDNA IgG and IgG2a antibody titers were measured in sera of MRL/Fas lpr/lpr mice by ELISA as previously described (39). Titers were expressed in relative units (RUs), defined as the dilution factor needed to reach 50% of binding saturation and calculated using Prism? software (GraphPad Software, Inc.).

Statistical Analyses
All statistical analyses were performed using Excel (Microsoft) or Prism R software. Differences in Ig titers, CSR and RNA transcript expression were analyzed with Student's paired (in vitro) and unpaired (in vivo) t-test assuming twotailed distributions.

HDI-Mediated Inhibition of the Class-Switched Antibody Response Is Reversed by Estrogen
We have shown that while HDIs, including the branched SCFA VPA and the four-carbon SCFA butyrate, dampen CSR/SHM and, therefore, the maturation of specific antibody responses, estrogen enhances CSR/SHM, thereby promoting maturation of antibody responses (23,45). To define the impact of estrogen on HDI-mediated modulation of the antibody response, we treated female C57BL/6 mice with nil, HDI (in drinking water) only, HDI plus E2 (in drinking water) or HDI plus E2 (in drinking water), fulvestrant and letrozole (i.p. injection), and injected FIGURE 1 | Estrogen reverses HDI-mediated inhibition of the antibody response. Female C57BL/6 mice were given plain or HDI-water starting at the age of 8 weeks and treated with nil, E2 (added to drinking water) or fulvestrant (Ful) and letrozole (Let) (i.p. injection) 1 day before injection with NP 16 -CGG. (A) Titers of NP 32 -binding IgG1 and high-affinity NP 4 -binding IgG1 (RU) in serum 10 days after NP 16 -CGG injection, as were measured by ELISA. Each symbol represents an individual mouse (n = 5). (B) Surface IgG1 expression in spleen B220 + PNA hi GC B cells 10 days after NP 16 -CGG injection. Data are representative of three independent experiments. Each symbol represents an individual mouse (n = 3). (C) Expression of Aicda transcripts in B220 + spleen B cells measured by real-time qRT-PCR. Expression was normalized to Gapdh expression and depicted as relative to the expression in B cells isolated from mice treated with nil, set as one. *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant (unpaired t-test). Data are mean ± SEM of three independent experiments. these mice with NP 16 -CGG, which preferentially induces IgG1 to NP, 1 day after the treatment was started. Consistent with our previous findings (23), HDI reduced the IgG1 response to NP, including total IgG1, NP 32 -binding and high-affinity NP 4 -binding IgG1 ( Figure 1A). The HDI-mediated inhibition of NP-specific IgG1 production was reversed by E2, but greatly enhanced (in the presence of exogenous E2) by fulvestrant and letrozole. The reciprocal alteration of NP-specific IgG1 response by E2 or fulvestrant and letrozole was associated with a similar alteration in the proportion of IgG1 + PNA hi germinal center B cells and Aicda expression (Figures 1B,C). Thus, consistent with our previous findings, SCFA HDIs inhibited Aicda expression, CSR and the antigen-specific class-switched antibody response (2,23,24). This inhibition was reversed by E2, but greatly enhanced by blocking ERα and inhibiting estrogen production.

HDI-Mediated Inhibition of Autoantibody Response in Lupus MRL/Fas lpr/lpr Mice Is Reversed by Estrogen
Epigenetic dysregulation can result in aberrant antibody responses, such as those associated with autoimmunity or allergy, or neoplastic transformation. The pathogenesis of autoimmune diseases, including lupus, can be traced to both genetic susceptibility and epigenetic modifications arising from exposure to the environment, dietary intake, and hormonal factors (46). Consistent with our previous findings (23,25), HDI also reduced autoantibody production and autoimmunity in lupusprone mice, as reflected by reduced skin lesion, anti-nuclear antibody (ANA), and anti-dsDNA IgG2a titers in MRL/Fas lpr/lpr mice treated with HDIs for 4 weeks (Figures 2A-C). Like the antigen-specific antibody response, the HDI-mediated inhibition of autoantibody response was reversed by E2, as shown by worse skin lesions and significantly increased ANA and anti-dsDNA IgG2a in MRL/Fas lpr/lpr mice treated with HDI together with E2 as compared to the mice treated with HDI only. By contrast, the HDI-mediated attenuation of skin lesion and production of ANA and anti-dsDNA IgG2a autoantibody was enhanced by fulvestrant and letrozole (Figures 2A-C). Like VPA, SCFA butyrate reduced autoantibody production in MRL/Fas lpr/lpr mice ( Figure 3A). It also reduced IgG2a + CD138 + plasmablasts and CD19 − CD138 + plasma cells in MRL/Fas lpr/lpr mice ( Figure 3B). As in VPA-treated mice, such SCFA butyratemediated reduction of IgG2a-producing B cells and plasma cells as well as autoantibodies was enhanced by fulvestrant and letrozole. Thus, in autoimmune lupus-prone mice, the HDI-mediated inhibition of the autoantibody response was   and Aicda cre Esr1 +/+ mice were given plain water (Nil) or SCFA-water containing 140 mM sodium butyrate and 20.0 mg/ml tributyrin starting at the age of 5 weeks, and i.p injected with NP 16 -CGG (in alum) at the age of 8 weeks. The mice were boost injected with NP 16 -CGG (in PBS) 3 weeks later. Circulating total and NP 4 -binding IgM and IgG1 were measured by ELISA 7 days after boost injection. Each symbol represents an individual mouse (n = 3 per group). Data are mean ± SEM of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001, ns, not significant (unpaired t-test).
impaired by estrogen but enhanced by blocking ER and reducing estrogen production.

Activated B Cell-Specific ERα Deletion Synergizes With SCFA HDIs to Further Reduce the Class-Switched Antibody Response
HDI-mediated inhibition of Aicda and CSR was significantly reversed by E2 in Esr1 +/+ B cells but not in Esr1 −/− B cells-E2, HDI, or HDI plus E2 did not alter viability of either Esr1 +/+ or Esr1 −/− B cells. Further, Esr1 deletion, as in Esr1 −/− mice, also reduced the antigen-specific antibody response, as indicated by the reduced high affinity NP 4 -binding IgG1 in such mice immunized with NP 16 -CGG ( Figure 4A). Consistent with our previous findings (23), HDI significantly decreased NP 4 -binding IgG1 in Esr1 +/+ mice, and did so to antigenspecific IgG1 titers even lower in Esr1 −/− mice as compared to their Esr1 +/+ littermates, likely due to lack of impact by endogenous estrogen. Thus, estrogen, through ERα, reverses HDI-mediated inhibition of Aicda, CSR and the antigen-specific antibody response.
To further define the role of B cell E2-ERα signaling in the modulation of SCFA HDI-mediated inhibition of the antibody response, we treated female Aicda cre Esr1 fl/fl and Aicda cre Esr1 +/+ mice with SCFA HDI butyrate and immunized these mice with NP 16 -CGG. Consistent with the enhancing role of E2-ERα in Aicda expression and CSR (28), circulating total IgG1 and NP 4binding IgG1, but not IgM, were reduced in Aicda cre Esr1 fl/fl mice as compared to their Aicda cre Esr1 +/+ counterparts ( Figure 4B). Similar to the reduction in class-switched antibodies we showed in C57BL/6 mice and autoantibodies in lupus-prone MRL/Fas lpr/lpr mice (Figures 1-3) (25), butyrate treatment reduced total and NP 4 -binding IgG1 in Aicda cre Esr1 +/+ mice. Further, like in the HDI treated Esr1 −/− mice, the reduction of total and NP 4 -binding IgG1 by butyrate was even greater in Aicda cre Esr1 fl/fl mice, in spite of preservation of normal levels of total and NP 4 -binding IgM. Thus, B cell ERα deletion synergizes with SCFA to reduce the class-switched antibody response.

Estrogen Reverses SCFA HDI Inhibition of Aicda Expression and CSR Through ERα
We have shown that HDIs inhibit antibody and autoantibody responses through inhibition of Aicda expression and CSRas well as Prdm1 expression and plasma cell differentiation (2,23,24). We have also shown that estrogen enhances Aicda induction and upregulated CSR (45). To assess the ability of E2 to reverse the SCFA-mediated inhibition of Aicda expression and CSR, we set up in vitro B cell stimulation cultures including nil or different doses of E2 (10 and 30 nM). The minimal amount of E2 contained in commercial FBS did not significantly affect the total amount of E2 in our cell culture medium-FBS from major vendors, such as Hyclone and ThermoFisher, contains 1.23-1.53 nM E2, which translates into 0.123-0.153 nM E2 in RPMI medium containing 10% FBS (47). As shown by a previous study of ours, E2 at concentration of <5 nM does not significantly alter AID expression and CSR (28). We, therefore, used RPMI medium containing 10% FBS in most in vitro experiments. Consistent with our previous findings, upon stimulation with LPS plus IL-4, HDI VPA, butyrate and propionate reduced CSR to IgG1 and IgE, as shown by reduced surface IgG1 + B cells and/or post-recombination Iµ-Cγ1 or Iµ-Cε transcripts (Figures 5A,B). Such CSR inhibition was associated with reduction of Aicda expression, as shown by qRT-PCR analysis, and was reversed by E2, which per se upregulated Aicda and CSR ( Figure 5B). However, unlike Aicda expression and CSR, inhibition of Prdm1 expression and plasma cell differentiation by HDI butyrate was not reversed by E2 (Figures 5C,D). Thus, estrogen impairs SCFA HDI-mediated inhibition of Aicda expression and CSR, thereby restoring CSR to pre-inhibition levels.
SCFA HDIs butyrate and propionate also inhibit Aicda expression and CSR in B cells stimulated by LPS plus IL-4 or CD154 plus IL-4 (25). In these B cells, the inhibition of Aicda expression and CSR was also reversed by E2 (Figures 6A,B). The E2-mediated reversion of SCFA HDImediated inhibition of Aicda expression and CSR was impaired by ERα deletion in Esr1 −/− B cells. In such Esr1 −/− B cells, Aicda expression and CSR to IgG1 were marginally lower than Esr1 +/+ B cells, and were reduced to an even lower level by butyrate or propionate. Esr1 knockout or treatment with butyrate and/or E2 did not alter cell viability, as shown by 7-AAD staining ( Figure 6C). Thus, deletion of ERα as in Esr1 −/− B cells impairs the ability of estrogen to reverse the HDI-mediated inhibition of Aicda expression and CSR.

Estrogen Reverses SCFA HDI-Mediated Upregulation of Aicda-Targeting miR-26a
As we showed by qRT-PCR and miRNA-Seq, HDI upregulate B cell miR-155, miR-181b, and miR-361, which silence AID by targeting Aicda 3 ′ UTR, as well as miR-125a, miR-92b, miR-26a, and miR-103, which, as we predicted, also target Aicda 3 ′ UTR (23, 24)-Aicda 3 ′ UTR has two mR-181b target sites ( Figure 7A). To define whether the expression of Aicda-targeting miRNA(s) can be modulated by estrogen in B cells induced to undergo CSR, we stimulated B cells with LPS plus IL-4 in the presence of nil, E2, butyrate or butyrate plus E2 for 60 h, before analyzing the expression of miRNAs. Consistent with our previous findings, miR-155, miR-181b, miR-361, miR-125a, miR-92b, and miR-26a were all upregulated by butyrate ( Figure 7B). Among such miRNAs, however, miR-26a was the only highly expressed miRNA that was significantly reduced by E2, in the presence or absence of butyrate ( Figure 7B)-miR-125a was also reduced by E2, but due to its low basal expression level, it did not likely play a significant role in the modulation of Aicda expression. As we have shown, miR-26a is highly expressed in B cells induced to undergo CSR and plasma cell differentiation, and it is one of the most abundant miRNAs expressed in B cells after stimulation by LPS plus IL-4 (24). In these B cells, the average RPKs of miR-26a, as measured by miRNA-Seq, was 418.2 ( Figure 7C), which was more than 15 folds higher than the average RPMs of all the miRNAs analyzed (24).
To further define the role of miRNAs in SCFA-and E2-mediated modulation of Aicda expression, we generated luciferase reporter constructs containing Aicda 3 ′ UTR. Such constructs were used to transfect mouse CH12F3 B cells that were induced to undergo CSR at a high rate. Like in primary B cells, in such cells, the Aicda-targeting miRNAs were upregulated by butyrate (25), and miR-26a was downregulated by E2 ( Figure 7D). In CH12F3 B cells transfected with the Aicda 3 ′ UTR reporter construct, the luciferase reporter activity was reduced by butyrate but increased by E2. Further, E2 reversed the butyrate-induced downregulation of luciferase reporter activity ( Figure 7E).
To define whether miR-26a indeed directly targets Aicda 3 ′ UTR to silence Aicda expression thereby reducing CSR, we constructed a luciferase reporter vector containing the Aicda 3 ′ UTR sequence that was mutated to destroy the miR-26a target site. In CH12F3 B cells transfected with such a mutant Aicda 3 ′ UTR reporter construct and cultured with CD154 plus TGFβ, the luciferase reporter activity was more than 60% higher than that in CH12F3 B cells transfected with the reporter construct including unmutated Aicda 3 ′ UTR (Figure 8A). In fact, such luciferase reporter activity was comparable to that recorded in CH12F3 B cells transfected with the reporter construct including the unmutated Aicda 3 ′ UTR and cultured with CD154 plus TGF-β in the presence of E2. Unlike in CH12F3 B cells transfected with unmutated Aicda 3 ′ UTR luciferase reporter construct, in CH12F3 B cells transfected with the mutated Aicda 3 ′ UTR luciferase reporter construct, the inhibition of luciferase reporter activity by butyrate was not reversed by E2. Thus, estrogen reverses SCFA HDI-mediated Aicda upregulation likely through downregulation of miR-26a which target Aicda 3 ′ UTR.

AID and CSR Are Inhibited by miR-26a Enforced Expression and Boosted by miR-26a Sponge
To further define the role of miR-26a in the modulation of Aicda expression and CSR, we enforced expressed this miRNA by transducing LPS-preactivated B cells with retroviral vector pMSCV-PIG-miR-26a-2 (43) expressing GFP and miR-26a or the control pMSCV-PIG vector expressing GFP but not a miRNA. The transduced B cells were cultured with LPS plus IL-4 for 72 or 96 h, respectively, before analyzing CSR to IgG1. pMSCV-PIG-miR-26a-2 retroviral vector-transduced B cells showed a significant reduction in Aicda expression and CSR to IgG1 (p = 0.003) as compared to pMSCV-PIG retroviral vector-transduced B cells (Figures 8B,C). To further demonstrate the role of miR-26a in the modulation of Aicda expression and CSR, we constructed a pMG retroviral vector expressing a miR-26a sponge, capable of depressing the miR-26a target. The miR-26a sponge expression vector was generated by inserting a DNA fragment containing five miR-26a binding sequences (miR-26a sponge, Figure 8D) into the pMG retroviral vector. In B cells transduced with pMG-miR26a-sponge retroviral vector, expression of the miR-26a sponge resulted in an increased expression Aicda transcript and AID protein, and consequent CSR (Figures 8E-G). Thus, miR-26a significantly modulates Aicda expression and CSR.
ERα Is Recruited to the Promoter of miR-26a Host Genes CTDSPL and CTDSP1, Directly or Indirectly Through Multiple cis-Elements, to Modulates miR-26a Expression E2 can downregulate as many protein encoding genes and miRNA host genes as it upregulates (31,32,48). After binding to nuclear ERs, E2-bound ERs function primarily as transcription factors through a classical pathway by interacting specifically with EREs in the promoter of estrogen-responsive genes to elicit transcriptional activation or repression. E2-ER-dependent FIGURE 7 | Estrogen impairs HDI mediated upregulation of Aicda-targeting miR-26a expression. (A) Alignment of miR-181b, miR-26a, miR-125a, miR-155, miR-103, miR-361, and miR-92b with their target sites in the 3 ′ UTR of Aicda mRNA (Aicda 3 ′ UTR has two mR-181b target sites). (B) SCFA HDI butyrate upregulates Aicda-targeting miR-155, miR-181a, miR-361, miR-92b, miR-125a, and miR-26a, while estrogen downregulates miR-26a and reverses SCFA-mediated upregulation of this miRNA. B cells were stimulated with LPS plus IL-4 and treated with nil, E2 (30 nM), butyrate (500 µM) or butyrate (500 µM) plus E2 (30 nM) for 60 h. Expression of miRNAs was analyzed by qRT-PCR, normalized to expression of small nuclear/nucleolar RNAs Rnu6, Snord61, Snord68, and Snord70, and depicted as relative to the expression of this miRNA in B cells treated with nil, set as 1. Data are mean ± SEM from three independent experiments. (C) Expression of miR-125a and miR-26a in B cells stimulated with LPS plus IL-4 for 60 h, as analyzed by miRNA-Seq and depicted as RPK. Data are means ± SEM of three independent experiments. (D) miR-26a is upregulated by butyrate and downregulated by E2 in CH12F3 B cells. CH12F3 cells were stimulated with CD154 plus IL-4 and TGF-β, and treated with nil, E2 (30 nM), butyrate (500 µM) or butyrate (500 µM) plus E2 (30 nM) for 60 h. Expression of miRNA-26a was analyzed by qRT-PCR, normalized to expression of Snord70, and depicted as relative to the expression of this miRNA in CH12F3 cells treated with nil, set as 1. Data are mean ± SEM from three independent experiments. (E) SCFA and E2 modulate miRNAs that target Aicda 3 ′ UTR. Luciferase activity in CH12F3 cells transfected with Aicda 3 ′ UTRs (cloned into pMIR-REPORT luciferase reporter vector) after a 24-h treatment with nil, E2 (30 nM), butyrate (500 µM) or butyrate (500 µM) plus E2 (30 nM). Luciferase activity was measured 1.5 h after transfection. Transfection efficiency was controlled by normalizing to signal from co-transfected R. reniformis luciferase vector. Luciferase activity is depicted as relative to values in B cells cultured with nil, set as 100. Data are mean ± SEM from three independent experiments. *p < 0.05, **p < 0.01, ns, not significant (unpaired t-test).
regulation of gene expression can also function through a nonclassical pathways involving protein-protein interaction of DNAbound ER with other DNA-bound transcriptions factors (49). E2induced activation of complexes consisting of ER and Specificity protein (Sp) and Activating protein-1 (AP-1, c-Jun, and c-Fos) involves pathways that modulate expression of many genes (49)(50)(51). E2-ERαs can bind to DNA directly through ERE (or half ERE) or indirectly by association with an intermediary transcription factor such as AP-1 or Sp1. Transcription of genes that depend on AP-1 can be repressed by E2-ERs (52)(53)(54)(55). E2 has been reported to repress gene transcription through ERα-Sp1, which recruits the transcriptional repressor-HDAC3 complex (56). We identified three and five EREs in the promoters of two host genes encoding miR-26a, CTDSPL and CTDSP1, respectively, as well as 10 AP-1 sites in each gene (Figure 9). Three AP-1 sites in the CTDSPL promoter and five AP-1 sites in the CTDSP1 promoter overlap with EREs. We also identified 11 and 6 Sp1-binding GC box elements in promoter regions of CTDSPL and CTDSP1, respectively, further suggesting that E2-ERαs negatively regulate CTDSPL and CTDSP1 expression. Accordingly, recruitment of ERα to both the CTDSPL and CTDSP1 promoters was significantly increased by E2 in a dose-dependent fashion (Figure 10A), and the estrogen-mediated reduction of miR-26a expression was impaired in Esr1 −/− B cells, in the presence or absence of butyrate (Figure 10B). It has been suggested that in ER + breast cancer cells, E2-ERα can indirectly reduce miR-26a expression by upregulating c-MYC, which has been reported to suppress the expression of many miRNAs (33,43). This mechanism, however, may not play a significant role in E2mediated miR-26a downregulation in B cells, as indicated by the unaltered c-Myc expression upon E2 treatment in CH12F3 B cells, in which miR-26a was inhibited by E2 and unaltered or even reduced c-Myc expression upon treatment of different doses of E2 in primary B cells (Figures 7D, 10C,D). Thus, estrogen reduces the expression of miR-26a, a silencer of Aicda, likely through both the classical and the non-classical ERαsignaling pathways. Luciferase activity in CH12F3 cells transfected with wild-type or mutated Aicda 3 ′ UTRs (cloned into pMIR-REPORT luciferase reporter vector) after a 24-h treatment with Nil, E2 (30 nM), butyrate (500 µM) or butyrate (500 µM) plus E2 (30 nM). Luciferase activity was measured 1.5 h after transfection. Transfection efficiency was controlled for by normalizing to signal from co-transfected R. reniformis luciferase vector. Luciferase activity is depicted as relative to values in B cells cultured with nil, set as 100. Data are mean ± SEM from three independent experiments. (B) Enforced expression of miR-26a reduced Aicda. B cells isolated from C57BL/6 mice were transduced with pMSCV-PIG-miR-26a-2 retroviral vector expressing GFP and miR-26a or empty pMSCV-PIG retroviral vector that expression GFP, and cultured for 48 h. Expression of Aicda was analyzed by real-time qRT-PCR, normalized to Gapdh, and depicted as relative to the expression of Aicda in B cells transduced with empty pMSCV-PIG retroviral vector, set as 1. Data are mean ± SEM from three independent experiments. (C) Enforced expression of miR-26a reduced CSR. Proportions of surface IgG1 + B cells among empty pMSCV-PIG or pMSCV-PIG-miR-26a-2 retroviral vector-transduced (B220 + GFP + ) B cells were analyzed by flow cytometry 96 h after transduction. Data are from one representative of three independent experiments. (D) Sequence of miR-26a sponge which contains five repeats of miRNA antisense binding sites (indicated by red boxes); sequences complimentary to mature miR-26a are underlined. (E,F) Expression of miR-26a sponge increased Aicda expression. B cells were transduced with pMG-miR-26a-sponge retroviral vector (expressing GFP and miR-26a sponge) or empty pMG retroviral vector (expressing GFP), and cultured for 48 h. Expression of Aicda was analyzed by real-time qRT-PCR, normalized to Gapdh, and depicted as relative to the expression of Aicda in B cells transduced with empty pMSCV-PIG retroviral vector, set as 1. Data are mean ± SEM from three independent experiments (E). (F) AID and β-Actin proteins as analyzed by immune-blotting. (G) Expression of miR-26a sponge increased CSR. Proportions of surface IgG1 + B cells among empty pMG or pMG-miR-26a-sponge retroviral vector-transduced (B220 + GFP + ) B cells, as analyzed by flow cytometry 96 h after transduction. Data are from one representative of three independent experiments. *p < 0.05, **p < 0.01, ns, not significant (unpaired t-test).

DISCUSSION
Estrogen plays an important role in the generation of classswitched and mutated antibodies and autoantibodies, thereby contributing to the greater response to microbial vaccines as well as the greater incidence of antibody-mediated autoimmunity, such as in systemic lupus erythematosus, in females (4, 6-16, 28). Indeed, we have shown that estrogen enhances AID expression by activating the HOXC4/HoxC4 promoter and inducing HoxC4, which is a critical Aicda promoter activator  together with NF-κB (27,28). HOXC4/HoxC4 activation by estrogen-ER complexes occurs by binding to three evolutionarily conserved and cooperative EREs in the HOXC4/HoxC4 promote (27,28). As induced by CD40 or TLR and IL-4 signaling, HOXC4/HoxC4 synergizes with NF-κB to induce AID and CSR/SHM (27,28). As we have shown, SCFA HDIs function as epigenetic modulators of T-dependent and T-independent classswitched and somatically hypermutated antibody responses as well as autoantibody responses by upregulating select miRNAs that target the 3 ′ UTR of Aicda and Prdm1 mRNAs (2,(23)(24)(25).
Here, we demonstrated that estrogen reverses the SCFA HDImediated modulation of antibody and autoantibody responses (2,(23)(24)(25). This effect is mediated to a great extent by estrogen downregulation of miR-26a, which, as we showed, is upregulated by SCFA HDI and downregulates Aicda transcripts by targeting Aicda 3 ′ UTR.
By binding to estrogen receptors ERα and/or ERβ, estrogen is involved in the regulation of multiple physiological processes. ERα and ERβ are ligand-inducible transcription factors that share common structural and functional domains and activate or repress gene expression in estrogen-responsive tissues (57). In addition to modulating many protein-coding genes, estrogen modulates expression of non-coding RNAs, including several miRNAs, such as miR-26a (31,32). In the classical pathway of estrogen function, E2 diffuses into the cell and binds to ERs, which located mainly in the nucleus. After binding E2, ERs form dimers that bind to inverted palindromic ERE sequences in the regulatory regions of estrogen-responsive genes (57). In the non-classical pathway of estrogen-mediated gene expression, E2-ER-dependent regulation of gene expression occurs through protein-protein interaction of DNA-bound ER with other DNAbound transcriptions factors, such as AP1 and Sp1 (49,58). Indeed, transcription of many genes that depend on AP-1 can be repressed by ERs activated by E2 (52)(53)(54)(55). In addition, E2 has also been suggested to repress gene transcription through ERα-Sp1, which recruits the transcriptional repressor-HDAC3 complex (56). The presence of multiple putative EREs, AP-1 motifs and Sp1-binding GC boxes in the promoter regions of miR-26a host genes CTDSPL and CTDSP1 suggests that the E2-mediated regulation of miR-26a expression involves both classical and nonclassical ER-action pathways possible resulting in overall net inhibition effect of miR-26a expression. c-Myc would provide an additional mechanism for E2-mediated downregulation of miR-26a. In ER + breast cancer cells, E2-ERα can indirectly reduce miR-26a expression by upregulating c-MYC, which has been reported to inhibit the expression of several miRNAs (33,43). Estrogen has been shown to induce c-Myc expression through an upstream enhancer activated by ER and AP-1 (59), and c-MYC can regulate the expression of miR-26a by directly binding to the promoters of miR-26a host genes (CTDSPL and TDSP1). However, this mechanism may not play a major role in E2-mediated miR-26a downregulation in B cells, as indicated by the unaltered or even reduced c-Myc expression upon E2 treatment of such cells. In addition to positive regulation of c-Myc through ERα-mediated activation of c-Myc promoter and enhancer regions, estrogen also promotes a negative regulatory pathway of c-Myc expression through miRNAs, thereby making the overall effect of estrogen on c-Myc expression a result of the balance of its positive and negative effects. Tipping such a balance would depend on the dose of estrogen. The E2 amount used in this study possibly promoted a stronger negative than positive effect on the regulation of c-Myc in primary mouse B cells.
Males and females display differential immunological responses to foreign-and self-antigens (7,12,60). Antibody responses to infections as well as bacterial and viral vaccines, such as influenza, hepatitis A, hepatitis B, rubella, measles, diphtheria, tetanus, Brucella, and rabies vaccines, are generally more vigorous in females than males (5,(61)(62)(63)(64)(65)(66)(67)(68)(69)(70)(71)(72)(73)(74)(75). Consistent with these differences, females show lower incidence of clinical disease than males after vaccination with influenza (76)(77)(78)(79), hepatitis A (80,81), and hepatitis B (82) vaccines. In addition, women vaccinated with half a dose of influenza vaccine achieved comparable antibody titers to males vaccinated with full dose (83). Genes on the X chromosome regulate immune function and play an important role in modulating sex differences in immune responses, and sex hormones interact with genetic and environmental factors to control immune responses (12,84). In general, estrogen displays an immune enhancing activity whereas testosterone, the primary male sex hormone, has a suppressive effect on the immune system. Testosterone has recently been shown to enhance the association of GPR174, an X chromosomeencoded G-protein coupled receptor, with Gai in response to its ligand CCL21 and suppress B cell positioning toward the follicular center and germinal center formation (85). This may lead to a weaker antibody response in males. The effect of estrogen on the immune system can differ dependent on the level of estrogens, cell type, cell activation state, microenvironment, and the experimental context (9). Estrogen can have profound effects on B cell differentiation, activity, function, and survival (7,9,17). Our results further extend our previous findings and further emphasize an important role for estrogen in the regulation of AID expression and, therefore, the maturation of the antibody response.
Sex differences in immune responses also result in differential susceptibility of males and females to autoimmune diseases. A majority of autoimmune diseases, such as systemic lupus erythematosus, Sjögren syndrome, thyroid diseases, scleroderma, and myasthenia gravis, occur predominantly in females (12). Autoantibody response and autoimmune diseases are regulated by genetic, epigenetic, infectious factors, as well as external and internal environmental triggers, such as hormones and the microbiome. It is well-accepted that estrogens contribute to the immunopathogenesis of lupus, based on the tendency for disease onset in women during the child-bearing years, as well as increased relative risk of SLE onset and disease flares in women who were administrated exogenous estrogens (86). Many SLE patients show more symptoms before menstrual periods and/or during pregnancy, when estrogen incretion is high. Further, patients with Klinefelter's syndrome (XXY) characterized by hypergonadotrophic hypogonadism are prone to develop SLE (87). Finally, male-to-female transgenders have been reported to develop SLE after a sexual-reassignment operation and long-term treatment with exogenous estrogens (88)(89)(90).
Estrogen plays an important physiologic role in B cell development and function (4,17). E2 can induce a genetic program that alters B cell survival and activation in a B cell intrinsic fashion and thus skews the naïve immune system toward autoreactivity (3). E2 has been shown to enhance B cell differentiation, increase the numbers of autoantibodyproducing cells and overall autoantibody production (91,92). E2 enhances anti-ds DNA antibody and IgG, IgM production by PBMCs from SLE patients (93). In lupus-prone NZB/WF1 mice, E2 administration resulted in an increased level of anti-C1q and anti-dsDNA autoantibodies, leading to accelerated glomerulonephritis and disease progression (94). Conversely, ERα deficiency in NZB/WF1 mice reduced anti-dsDNA IgG autoantibodies, attenuated glomerulonephritis, and increased survival (95). Progesterone, which in general has effect opposite to estrogen, reduced pathogenic anti-dsDNA IgG2a and glomerular IgG accumulation (96).
We have shown that HDIs upregulate the expression of select miRNAs that inhibit Aicda or Prdm1 by increasing histone acetylation of the host genes of such miRNAs (23,24). Two RNase III proteins, Drosha and Dicer are critical for miRNA biogenesis (97). Multiple acetyltransferases are involved in acetylating Drosha proteins, thereby preventing their degradation by inhibiting ubiquitination. For instance, acetylation of Drosha on the N-terminus inhibits Drosha degradation by ubiquitination, thereby increasing the stability of this protein (98). Drosha acetylation can be increased by HDI. Therefore, SCFA HDIs can increase Drosha protein level by enhancing its stability and, thereby, enhancing miRNA biogenesis. Conversely Droshamediated miRNA processing can be inhibited by ERα (99). This would provide a further potential mechanism by which HDIs upregulate miRNA expression, particularly miR-26a, and how estrogen downregulate this miRNA.
HDIs, including SCFAs butyrate and propionate, reduce both Aicda and Prdm1 expression (2,(23)(24)(25)(26). In the current studies, estrogen reversed HDI-mediated inhibition of Aicda expression, but did not significantly reverse HDI-induced reduction of Prdm1 expression and plasma cell differentiation. Prdm1 expression and plasma cell differentiation entail, among other modifications, DNA hypomethylation (100). DNA methylation is mediated by DNA methyltransferases (DNMTs), which catalyze the transfer of a methyl-group to DNA, and Tet dioxygenases, which oxidates 5-methylcytosine in DNA to 5hydroxylmethylcytosine (5hmC), thereby mediating active DNA demethylation (101). Interestingly, miR-26a has been shown to modulate the expression of Tet2 (102,103), the most abundant Tet in B cells (our unpublished data). Estrogen-mediated reduction of miR-26a would have the potential to upregulate Tet2, thereby leading to increased Prdm1 expression and plasma cell differentiation (104). However, as miR-26a is also predicted to target DNMT3B (105), the possible activation of DNMT3B upon miR-26a reduction by estrogen might have overridden a potential active DNA demethylation caused by an increase in Tet2 expression. It is also possible that estrogen, through ERα, enhances expression of Ets1 (106), a negative regulator of Prdm1, thereby overriding the potentially activating effect of estrogen-ER on plasma cell differentiation. Finally, in addition to Aicda, miR-26a has been shown to target Pten (phosphatase and tensin homolog) gene (107), a lipid phosphastase, which directly antagonizes PI3K activity by dephosphorylating PIP 3 to regenerate PI (4, 5) P 2 . B cell-specific Pten deficiency results in impaired CSR and enhanced plasma cell differentiation (108,109). It is possible that the estrogen-mediated reduction of miR-26a in B cells also resulted in an upregulation of Pten, which might also contribute to CSR enhancement by estrogen.
SCFAs, such as butyrate and propionate, generated by gut commensal bacteria through digestion of dietary fibers, are among the most abundant endogenous HDIs in the body. The gut microbiota is a critical regulator of homeostasis by protecting against pathogens while maintaining immune tolerance to allergens. Commensal bacteria can modulate host immune responses through metabolite-dependent mechanisms (110). Butyrate and propionate are among the most abundant commensal bacterial metabolites with immunoregulatory functions. These endogenous SCFA HDIs play an important role in preventing or dampening dysregulated antibody responses, which can lead to generation of autoantibodies or allergic IgE, while maintaining a certain level of protective antibodies against microbial pathogens (25,111,112). Recent studies have suggested an important interaction between estrogen and gut microbiota (113,114), which would act together to regulate immune responses. Gut commensal bacteria can be influenced by sex hormones including E2, which has been suggested to impact bacterial community composition (114). As we showed here, E2 can also modulate the effect of SCFAs, the major metabolites of gut commensal bacteria. The gastrointestinal microbiome profoundly impacts on various immune responses, including B cell maturation, activation and IgA antibody responses (115,116). In the gut, which contains high concentration of butyrate and propionate, IgA is produced extensively. IgA can bind to the mucus layer covering epithelial cells and form a barrier capable of neutralizing pathogens before they reach the epithelial cells. By opposing the SCFA HDI-mediated downregulation of the antibody response, estrogen may be important for ensuring a physiologic high level of IgA in the gut. Indeed, the intestine is one of the major locations of estrogen metabolism and the intestine maintains a relatively high concentration of estrogen, including E2 (117,118). Conversely, by opposing the SCFA HDI-mediated inhibition of class-switched autoantibody response, high level of estrogen, such as that in pre-ovulation women, may lift the homeostatic protection of gut microbiota from autoantibody production and autoimmunity. Thus, by showing that estrogen reverses the SCFA HDI-mediated downregulation of AID expression and CSR through selective modulation of miR-26a, our finding suggests that the "safeguard" function of SCFAs in dampening antibody overproduction and preventing emergence of autoantibodies can be reversed by estrogen. This may provide an explanation for the female bias in autoantibody-mediated autoimmune diseases, such as lupus, and the greatly increased incidence of food allergies in female adults (119).

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

ETHICS STATEMENT
The animal study was reviewed and approved by the Institutional Animal Care and Use Committees of the University of Texas Health Science Center, San Antonio.

AUTHOR CONTRIBUTIONS
PC conceived and designed the study, planned the experiments, analyzed all primary data, created all figures, and wrote the manuscript. TS, YX, ZQ, JI, DC, and ZX performed experiments. HZ designed the study, planned, analyzed data, created all figures, and wrote the manuscript.