Matters of life and death: how estrogen and estrogen receptor binding to the immunoglobulin heavy chain locus may influence outcomes of infection, allergy, and autoimmune disease

Abstract

Sex hormones are best known for their influences on reproduction, but they also have profound influences on the immune response. Examples of sex-specific differences include: (i) the relatively poor control of influenza virus infections in males compared to females, (ii) allergic asthma, an IgE-associated hypersensitivity reaction that is exacerbated in adolescent females compared to males, and (iii) systemic lupus erythematosus, a life-threatening autoimmune disease with a 9:1 female:male bias. Here we consider how estrogen and estrogen receptor α (ERα) may influence the immune response by modifying class switch recombination (CSR) and immunoglobulin expression patterns. We focus on ERα binding to enhancers (Eμ and the 3’ regulatory region) and switch sites (Sµ and Sε) in the immunoglobulin heavy chain locus. Our preliminary data from ChIP-seq analyses of purified, activated B cells show estrogen-mediated changes in the positioning of ERα binding within and near Sμ and Sε. In the presence of estrogen, ERα is bound not only to estrogen response elements (ERE), but also to adenosine-cytidine (AC)-repeats and poly adenosine (poly A) sequences, in some cases within constant region gene introns. We propose that by binding these sites, estrogen and ERα directly participate in the DNA loop formation required for CSR. We further suggest that estrogen regulates immunoglobulin expression patterns and can thereby influence life-and-death outcomes of infection, hypersensitivity, and autoimmune disease.

Estrogen, a nuclear hormone

Estrogen is a nuclear hormone that is best known for its influence on reproductive organs, but is also critical for the development and function of cardiovascular, skeletal, nervous, and immune systems [1–8]. Estrogen signals at both intra- and extra-nuclear sites [9], but we will focus here on its intra-nuclear functions.

Estrogen is a ligand for estrogen receptors (ER) α and β, which are type I nuclear hormone receptors with DNA binding sites throughout the mammalian genome [9–11]. The receptors are frequently associated with promoters, enhancers, DNA loops, and complex protein-DNA interactions [11–13].

Nuclear hormone receptors have multiple, structurally distinct domains including an N-terminal domain with activation function (AF-1), a central DNA-binding domain (DBD), and a carboxyl-terminal ligand binding domain (LBD) with ligand-dependent function (AF-2). ER binds DNA best as a dimer, but monomer binding is also observed [14, 15]. Receptor functions are regulated by ligand binding (e.g. with estradiol or tamoxifen), DNA targets, and interactions with other transcription factors [11].

Typically, two ER monomers bind cooperatively with a consensus ERE (GGTCANNNTGACC), but binding is promiscuous and ER can also bind non-consensus sites, in some cases tethered by other proteins. In fact, a large fraction of ER binding sites (and other nuclear hormone binding sites) in the mammalian genome do not conform to the consensus motif [11, 16, 17]. Ligand binding directly influences transcriptional regulation by altering ER-DNA conformations and associations with other transcription factors. For example, research using an estrogen-responsive pS2 promoter showed that unliganded ER (apo ER) could bind and repress transcription, and that the addition of estrogen resulted in conformational changes that supported RNA Polymerase II (RNA Pol II) recruitment and improved gene expression [18–21]. Outcomes of interactions between ER and other transcription factors may be enhancing or antagonistic, as molecules may synergize or compete for ligands and DNA binding sites [22].

Benefits and risks of high estrogen levels

In mice and humans, immune responses of males and females differ. For example, studies of rodents and humans show sex-specific differences in antibody isotype expression patterns in sera. Overall, females tend to express higher levels of immunoglobulins compared to males (often including IgM, IgG1, IgG2, and/or IgE [23, 24]). In small animal studies, the treatment of males with estrogen was shown to up-regulate immunoglobulin levels and antigen-specific antibody-producing cells [25, 26]. In humans, estrogen levels correlated with IgG responses toward an influenza virus vaccine [27]. In ex vivo cultures of mouse splenocytes, estrogen increased levels of IgM and IgE [28, 29].

Differences in immunoglobulin levels between the sexes help explain, at least in part, why adult females gain better control of certain infectious diseases compared to males [30, 31]. This benefit for females is not without risk; females suffer more than males from hyper-immune responses toward typically harmless substances or self-antigens [32–36]. Consequences can include life-threatening episodes of allergic asthma or systemic lupus erythematosus (SLE, lupus), particularly among pregnant women [37–43] in whom estrogen levels can soar to >6,000 pg/ml (as compared to the ≤100 pg/ml level typical of an adult male)[44, 45]. In a mouse model, supplemental estrogen was shown to increase antibody levels against self-antigens including cardiolipin [5, 6]. In a separate model, an estrogen inhibitor was shown to reduce levels of allergen-specific IgG1, IgG2a and IgE, as well as clinical disease symptoms [46].

As stated above, the influences of estrogen on immune responses depend on the environment and cross-regulation by other nuclear hormones [47–57]. For example, in a study of C57BL/6 mice, we found that IgG2b levels were higher in females compared to males, but when mice were deficient in vitamin A (a hormone bound by type II nuclear hormone receptors), IgG2b levels rose in male mice and exceeded the IgG2b levels of females [24].

Because estrogen affects many cell types, there are multiple explanations for sex differences in immunity and disease. For example, estrogen may upregulate interleukins that can stabilize B cells [58]. Estrogen also upregulates the expression of activation induced deaminase (AID) in B cells [59, 60]. Below, we focus on ERα binding to the immunoglobulin heavy chain locus to highlight a direct mechanism by which estrogen may influence immunoglobulin production and downstream consequences.

ERα and the immunoglobulin heavy chain locus

B cells initially produce IgM by transcribing V-D-J-Cμ heavy chain gene segments. Upon B cell activation, isotypes can switch (e.g. from IgM to IgE) by positioning V-D-J upstream of a different constant region gene (e.g. Cε). This class switch recombination (CSR) event involves double stranded DNA breaks at switch (S) sites (e.g. Sμ and Sε) upstream of target C genes followed by ligation of one S region to the other and deletion of intervening sequences. CSR can occur either directly or in succession (IgM to IgE or IgM to IgG to IgE) [61–66].

The process of CSR commences with the production of sterile transcripts, initiated in or near donor and recipient S sites [67–70], a process that creates R loops and exposes single stranded DNA to additional enzyme activity. After splicing, sterile transcripts retain a short intervening exon and a C gene exon (e.g. Iμ-Cμ and Iε-Cε). RNA Pol II stalls during transcription and recruits AID, which converts dC to dU [65, 68–70]. Downstream events can include the removal of dU by uracil DNA glycosylase (UNG), DNA cleavage by abasic endonucleases, and ligation of donor and recipient S regions by non-homologous end-joining machinery.

Enhancers that influence the quality and quantity of CSR and antibody expression include Eμ (located upstream of Iμ) and multiple DNase I hypersensitivity sites (including HS1,2) located in the 3’ regulatory region (3’RR) downstream of Cα [71–73]. During CSR, DNA loops integrate enhancers, promoters and S sites [69, 74–76].

We previously questioned whether ERα might directly influence CSR and antibody expression and therefore interrogated the immunoglobulin heavy chain locus for ERE. We then discovered hotspots for ERE in S regions including Sμ and Sε [77]. Using ChIP analyses, we next demonstrated that ERα and RNA Pol II exhibited strikingly similar binding patterns within the locus. Sites of peak binding included enhancer sequences Eμ and HS1,2 of the 3’ RR [67, 78–80]. When supplemental estrogen was added to purified, activated B cell cultures, ERα and RNA Pol II shifted DNA binding positions synchronously [24]. We additionally discovered that in the presence of estrogen, ERα binding was focused not just on consensus ERE, but on sites containing AC-repeats [78, 80] near enhancer regions. It is likely that ERα binding to AC-repeat sequences assists the DNA loop formation that is essential for integrating enhancers, promoters and S sites during CSR. We additionally found that ERE knock-out, either within Eμ or HS1,2 sequences of a B cell line, was sufficient to significantly reduce CSR [78]. The HS1,2 site is of particular interest to us, because it marks a site of polymorphism associated with an unusually high frequency of lupus in humans [81].

Given the differences noted in IgE production and related allergic reactions between males and females (described above), we next focused on IgM and IgE S sites (Sμ and Sε), as these must be cleaved to support the IgM to IgE switch. As a preliminary analysis, we aligned sequences from the previously described ChIP libraries from activated, purified B cells stimulated with lipopolysaccharide (LPS) or LPS plus estrogen (LPS + E) [24, 79]. ERα binding within the Sμ region is shown in Figure 1. This region is a hotspot for ERE (RRYYRNNNTGANY, mapped using IGV software) [77]). In both LPS and LPS + E cultures, there was ERα binding to Eμ (enhancer/promoter) [67, 68], Iμ, and Sμ regions, but ERα appeared to be better focused on the Sμ site in the LPS + E culture. Binding peaks were not always coincident with ERE, but in the LPS + E library, peaks were usually aligned either with an ERE or with an AC-repeat or poly A sequence. In the LPS + E culture, ERα preferentially bound to intron rather than exon regions of Cμ and Cδ. Perhaps ERα influences patterns of mRNA splicing, as DNA conformations are known to instruct RNA splicing and certain heterogeneous nuclear ribonucleoproteins (hnRNPs) are known to bind AC-repeats [82–87].

FIGURE 1. Estrogen receptorα (ERα) binding in and near Sμ.

ChIP-seq libraries were produced from purified B cells one day after stimulation with LPS or LPS + 100 nM estrogen (LPS + E)[24]. Binding patterns were evaluated using IGV software. Positions for J_H, Eμ, Iμ, Sμ, Cμ and a portion of Cδ are shown in murine chromosome 12 (using the mm9 sequence). Estrogen response elements (ERE, RRYYRNNNTGANY), AC-repeats (ACACAC) and poly A (AAAAA) sequences are mapped. Oligonucleotides were mapped in a 5’ – 3’ orientation from left to right, unless labeled ‘rev’, in which case orientations were from right to left.

Next, we examined the Sε region (Figure 2). Again, in the LPS + E library, ERα bound AC-repeats and poly A sequences downstream of the S site. Based on the Sμ result, one might have expected that ERα would also bind Sε in the LPS + E culture, but this was not the case. There was in fact no signal observed within the Sε site. A unique NFKB binding motif (GGGGTTCC) marked the promoter region for sterile Iε-Cε transcription [61, 63, 64, 88, 89], but ERα binding was also absent in this position in the LPS + E library, indicating another estrogen-induced ERα binding shift.

Figure 2. ERα binding in and near Sε.

As described in Figure 1, ChIP-seq libraries were produced from purified B cells one day after stimulation with LPS or LPS + 100 nM estrogen (LPS + E). Positions for Sε, Cε are shown (mm9). A unique NFKB site (GGGGGTTCC) marks the promoter region for Iε-Cε sterile transcripts. ERE, AC-repeats and poly A sequences are mapped as described in the Figure 1 legend.

Presently, it is not clear if the negative binding of ERα in Sε indicates the complete absence of ERα at this site (or a change in configuration that artifactually inhibits ERα immunoprecipitation by the ChIP antibody). It also remains to be determined whether estrogen-driven changes in ERα patterns are transient or durable, and how a positive ERα ChIP signal in Sμ and/or Sε might predict frequencies of CSR and IgE expression. Nonetheless, evidence of estrogen-driven changes in Sμ and Sε switchosomes (proteins associated with S sites) lend credence to the hypothesis that estrogen directly modifies CSR and isotype expression. In addition to its partnership with RNA Pol II [12, 90, 91], ER is also known to recruit topoisomerase IIβ to DNA. This topoisomerase associates with poly (adenosine diphosphate-ribose) polymerase-1 enzymatic activity and generates doubled stranded DNA breaks [92], perhaps supplementing the AID/UNG-mediated cleavage mechanism described above. Further experimentation is encouraged to test hypotheses and provide details of how estrogen and ERα binding to the heavy chain locus influences immunoglobulin output among activated B cells.

May clinical manipulations of estrogen and ERα-DNA binding patterns improve disease outcomes?

Attempts have been made to treat certain diseases by modifying estrogen levels. Patients may receive estrogen supplements (hormone replacement therapy) or inhibitors. Difficulties are associated with each of these treatment options, because treatments are not well focused on a particular cell type or function. Tamoxifen (a selective estrogen receptor modulator, SERM), for example, is used widely as an antagonist to treat breast cancers, but can act as an agonist in other target tissues [9]. In post-menopausal women who suffer an increased predisposition to infections and impaired vaccine-induced immune responses, the effect of estrogen replacement therapy is a topic of continued debate [93–98].

Today, gene therapy offers an opportunity to focus treatments. New technologies provide researchers with a variety of methods for blocking protein interactions with DNA or RNA (e.g. using CRISPR-CAS9, anti-sense oligonucleotides [ASO], and/or DNA decoys) [99–113]. We consider that targeted manipulations of ERα binding within enhanceosomes [13] or switchosomes may (i) improve antibody activities when pathogen-specific responses are weak, or (ii) reduce immune responses in cases of allergic asthma or autoimmunity. Important first targets could include the ERα binding sites in HS1,2 (to modify polymorphisms associated with lupus) and ERα binding sites in Sε (to modify the IgE over-expression associated with allergic asthma).

Highlights.

Males and females suffer differently from pathogens, allergies and autoimmunity.

ERα binds regulatory elements in the immunoglobulin heavy chain locus.

Estrogen alters ERα binding to sequences near Sμ and Sε.

ERα binds ERE, adenosine-cytidine (AC)-repeats and poly adenosine (poly A) sequences.

Estrogen instructs immunoglobulin expression and disease consequences.

Abbreviations:

ER

estrogen receptor

AID

activation induced deaminase

AC-repeat

cytidine-adenosine repeat

S

switch site

RNA Pol II

RNA polymerase II

CSR

class switch recombination

LPS

lipopolysaccharide

LPS + E

culture condition with estrogen added to LPS

V

variable

D

diversity

J

joining

C

constant

SERM

Selective estrogen receptor modulator

hnRNP

heterogeneous nuclear ribonucleoprotein

HS

(DNase I) hypersensitive site

SLE or ‘lupus’

systemic lupus erythematosus

IGV

Integrative Genomics Viewer

CRISPR

clustered regularly interspaced short palindromic repeats

Cas9

CRISPR-associated protein-9 nuclease

ChIP

chromatin immunoprecipitation