New aspects of activin biology in epididymal function and immunopathology

The activins (A and B) and their binding protein, follistatin, play crucial roles in development, immunoregulation and inflammation throughout the body. In the male reproductive tract of the mouse, activin A and B production is largely confined to the initial segment and proximal caput of the epididymis and the efferent ducts, under normal conditions, with very low expression in the corpus, cauda and vas deferens. However, activin A protein is present throughout the epididymis and vas deferens and is largely associated with the epithelium and interstitial macrophages. Conversely, the activin‐binding protein follistatin is produced in the distal epididymis, with very high expression in the vas deferens. Activin activity in the distal tract is inhibited by follistatin, and the activin–follistatin balance is important for regulating coiling of the duct during epididymal development. In further experiments, as described in this report, in situ hybridisation was used to localise activin A mRNA principally to cells in the periductal zone and interstitium in the efferent ducts and proximal caput. Activin B mRNA, on the other hand, was localised to periductal cells in the efferent ducts and proximal epididymis and, most notably, to epithelial cells in the initial segment. Activin A is implicated in the regulation of mononuclear phagocyte function and immune responses in the caput and stimulates the expression of the key immunoregulatory protein, indoleamine 2,3‐dioxygenase in this region. Activin A production in the corpus and cauda increases dramatically during bacterial epididymitis in mice, promoting inflammation and fibrosis and causing damage to the epithelium and obstruction of the epididymal duct. Consequently, it appears that the activin–follistatin axis is crucial for maintaining normal epididymal structure and function, but disruption of this balance during inflammation has deleterious effects on male fertility. Follistatin has therapeutic potential in ameliorating the proinflammatory and profibrotic effects of activins.


ACTIVINS AND THEIR INHIBITORS-BACKGROUND
Activins are members of the transforming growth factor-β (TGFβ) superfamily of cytokines.They are disulphide-linked dimers of the β subunits of the gonadal hormones, inhibin A and B, which are heterodimers of a common α subunit with an inhibin-βA or -βB subunit that form inhibin A or inhibin B, respectively. 1[4] The inhibins regulate follicle-stimulating hormone (FSH) secretion by the anterior pituitary, and the activins were initially isolated for their capacity to oppose the action of inhibin by inducing FSH release by the pituitary. 2,3However, subsequent research established that the activins exert their biological effects by interaction with specific activin receptors, and it is the inhibins that are antagonists of this receptor binding. 1,5Activins C, D and E are homologous to activins A and B, but do not appear to have inherent biological activity and, such as the inhibins, act chiefly as signalling antagonists of activins A and B by preventing the binding of activins A and B to their receptors. 6tivins are produced by epithelial, mesenchymal and immune cell types in multiple organs and are also involved in regulating a broad range of molecular and cellular events, including cell proliferation, differentiation, apoptosis and even embryonic development. 7,8Crucially, activin is an important regulator of immune cell development, inflammation and immunity. 9,10Activin B has been less well studied than activin A, particularly in regard to immunoregulation, but has similar activities in so far as they have been investigated.
Activins A and B and the TGFβs (TGFβ1-3) have overlapping and complementary functions within the immune system, particular regulating mononuclear phagocyte (macrophages, dendritic cells) and lymphocyte function. 11They have both proinflammatory and antiinflammatory actions, which are dictated by the local immune cell and cytokine environment. 12The activins and the TGFβs principally act via the same intracellular signalling pathway involving the Smad2/3/4 transcription factor complex, but exert their effects through separate, albeit similar, membrane receptor complexes.Thus, activins bind to a type 2 activin receptor and recruit a type 1 activin receptor kinase to initiate signalling, whereas the TGFβs act via their own specific type 1 and 2 receptors.This receptor specificity determines the cellular sites of action of each ligand.The most crucial difference, however, is that the activins are loosely bound to their precursors and are inherently active when released from the cell, unlike TGFβs, which need to be proteolytically activated at their site of action. 13e principal endogenous inhibitor of activin activity in most tissues is the 'activin-binding protein' , follistatin.5][16] This protein is widely expressed and can irreversibly inhibit all the biological actions of the activins. 12,17,18The inhibins and follistatin do not inhibit TGFβ1 or TGFβ2 activity, although TGFβ3 can bind to follistatin, at least in vitro. 19Notably, follistatin and other endogenous or engineered antagonists of activin A are able to inhibit inflammation and fibrosis in a number of disease states and have been investigated as potential therapeutic agents in several inflammatory and fibrotic diseases. 12,20e role of activins in testicular development, function and pathology has been well studied.During foetal life, activin A is a critical modulator of testicular germ cell development and Sertoli cell proliferation. 21Activin imbalance can lead to testicular pathologies, including testicular tumours, spermatogenic arrest and infertility. 22,23e biology of the activins and their functional roles in epididymal function have historically received less attention, as outlined previously. 24is review will primarily focus on developments in epididymal biology and pathology because the publication of that earlier review, but also includes recent unpublished data that more precisely localise the sites of activin production in the caput, indicating that activins A and B may play complementary, yet different, roles in the epididymis.

STRUCTURAL, FUNCTIONAL AND IMMUNOLOGICAL SEGMENTATION OF THE EPIDIDYMIS
Crucially, the functional environment of the epididymis is not homogeneous, with sperm maturation occurring during transit through the caput and corpus and extended storage of mature spermatozoa in the cauda.The epididymis is also segmented both physically and functionally, with clear connective tissue septa further separating the caput, corpus and cauda into discrete segments in the rat and mouse. 25ysical segmentation is less obvious in the human epididymis, but functional segmentation has been reported. 26idence that the immune environment of the epididymis is also highly regionalised goes back to early studies comparing different immune cell populations in the murine caput, corpus and cauda 27 and clinical evidence that susceptibility to anti-sperm antibody development in men with obstructive azoospermia, congenital absence of the vas or vasectomy increased with increasing distance of the lesion from the testis. 28[31][32] In the cauda, interstitial monocyte-like macrophages and lymphocyte subsets, typically responsible for cellular immunity, are much more prominent.Comparable data from other species are relatively limited but tend to be consistent with the observations from murine studies.
3][34] This highly regionalised response is consistent with the observation that the resident immune cells in the caput are largely tolerogenic, while the majority in the cauda are inflammatory.This suggests that the immunoregulatory environment of the A-C) and protein measured by ELISA (D-F) in the testis, epididymal caput (segments 1−5), corpus (segments 6−7), cauda (segments 8−10) and vas deferens in adult wild-type C57Bl6 mice (mean ± SEM, one-way ANOVA).Groups with different letter notations differ significantly, that is, p < 0.05.D.vas, distal vas deferens; P.vas, proximal region of vas deferens; Vas, entire vas deferens.Data have been re-drawn from Wijayarathna et al. 38 caput is responsible for establishing tolerance against the autoantigenic spermatozoa from the testis as they transit through the caput, while the cauda is primed to respond to ascending pathogens.

ACTIVIN EXPRESSION IN THE EPIDIDYMIS-RECENT STUDIES
Our discovery that activins were highly expressed in the murine caput epididymis and gradually declined towards the cauda and vas deferens and that follistatin showed an opposing pattern [35][36][37][38] led to the hypothesis that these proteins play a significant role in the regional functions of the epididymis.Inhba and Inhbb, the genes encoding activins A and B, respectively, were highly expressed in the caput epididymis when measured by qRT-real time quantitative (PCR), with much lower expression in the testis and other regions of the epididymis and vas (Figure 1).Conversely, follistatin-encoding Fst mRNA expression was confined to the cauda and, especially, the vas deferens. 35,37,38Protein levels of activin B measured by enzyme-linked immunossorbent assay (ELISA) and immunohistochemistry tended to mirror the mRNA expression pattern, and follistatin was likewise distributed towards the distal regions of the reproductive tract. 37,38In contrast to the Inhba mRNA expression pattern, activin A protein levels were very high throughout the epididymis before declining in the vas. 37,38Activin A and B immunostaining was predominantly found in the cytoplasm of epithelial cells, particularly those within the proximal epididymis, and in small numbers of interstitial cells with macrophage morphology. 36Activin A, but not activin B, was also detected within the sperm head throughout the epididymal lumen.Follistatin staining was associated with the apical cytoplasm and luminal contents of the epididymal duct and vas deferens, including spermatozoa, and the smooth muscle layer of the vas deferens.Significantly, expression of follistatin mRNA in the vas deferens was several-fold higher than that in any other tissue examined in the adult male mouse. 39sed on the activin A immunohistochemistry results, we hypothesised that epididymal epithelial cells, but especially principal cells of the caput, were the main sites of activin A production in the epididymis.
Upon closer inspection using qRT-PCR, we established that both Inhba and Inhbb were more abundant in the proximal (segments 1−3) than in the distal caput (segments 4−5) (Figure 2).Inhba expression in the
Consequently, we investigated the localised expression of the activins at the mRNA level in the male reproductive tract to more precisely identify the cellular sites of production.Epididymal sections (5 μm in thickness, fixed in 10% formalin) from 56-and 25-day-old mice were analysed by in situ hybridisation using the RNAscope 2.5 HD Brown Assay (Advanced Cell Diagnostics) (in-depth method detailed in Supporting Information).Unexpectedly, Inhba was much more highly expressed in cells surrounding the efferent and epididymal ducts and interstitium than in the epithelial cells themselves, and Inhbb displayed a similar, albeit more proximally distributed, expression pattern (Figure 3A-H).Most notably, Inhbb, but not Inhba, was highly expressed in the epithelium of the initial segment (segment 1).Overall, Inhba was most highly expressed in the efferent ducts and epididymal segment 2, and Inhbb was most highly expressed in the efferent ducts and epididymal segment 1 (Table 1).This study establishes, for the first time, that the efferent ducts also produce both forms of activin.The distribution and localisation pattern of activins A and B as well as follistatin mRNA and protein were simillar in both the 56 day old adult and the immature 25-day-old mouse (unpublished data), indicating indicating that these regionalised expression patterns are established before the appearance of mature sperm in the epididymal lumen.
It is apparent that the cellular sites of activin production determined from mRNA expression and the distribution of the proteins measured by ELISA and immunohistochemistry provide different information.
While the activin A protein is found within epithelial cells throughout the epididymis and is especially localised to the apical cytoplasm and surface of the more distal regions of the epididymis (Figure 3I,J), the evidence suggests that most of this protein is actually produced in the proximal caput and efferent ducts.The production and localisation of activin B show a similar but more confined distribution.These results highlight a particular complication for investigation of activin biology, that there can be a mismatch between the major site of activin A (and activin B) production and the sites where the protein accumulates. 39e functional basis for this is still unclear, but presumably involves the movement of the secreted proteins and their binding to various carrier proteins, especially follistatin, and to their target tissues.Follistatin binds with very high affinity to activins in biological fluids, and this initiates binding of the activin-follistatin complex to proteoglycans on the surface of cells, which activates removal of the complex by an intracellular lysosomal degradation pathway. 42Crucially, both the free and bound (i.e., inactivated) activins are detected by the antibodies typically used in activin immunoassays and immunohistochemistry until they are processed by their target cells.

REGULATION OF ACTIVIN PRODUCTION BY THE TESTIS
In order to investigate the role of the testis in regulating activins in the caput epididymis, the efferent ducts were ligated close to the testicular capsule in adult mice. 43This procedure prevents passage of spermatozoa and other testicular, or lumicrine, secretions and has been shown to profoundly inhibit epithelial function in the initial segment. 44ferent duct ligation reduced activin A immunostaining in the caput epithelium, and in the initial segment in particular, but overall Inhba mRNA levels were unaffected.In contrast, Inhbb and a number of other genes closely associated with the initial segment were reduced substantially.The macrophage-specific gene, Cx3cr1, which is highly expressed on intraepithelial macrophages in the epididymis, was also reduced in expression.It appears that testicular lumicrine secretions regulate activin expression in the initial segment but may play only a minor role in the functional regionalisation of the remainder of the epididymis and vas deferens.
The role of testicular androgens in regionalisation of epididymal function is not clear but could involve differential expression of the androgen receptor and 5α-reductase within the epididymis.Direct evidence that activin is or is not regulated by androgens in the epididymis is lacking; however, activin A and B levels in mice lacking the inhibin α-subunit (Inha −/− ), which is profoundly androgen deficient, were not dramatically different from normal wild-type mice in the epididymis or vas deferens. 37There was no significant association between Inhba or Inhbb transcript levels in the caput and testosterone in the efferent duct ligation mouse model, but they did show an association with 5αreductase expression. 43The data suggest that local androgen action could play a role, but this requires further investigation.

THE PHYSIOLOGICAL AND PATHOPHYSIOLOGICAL ROLES OF THE ACTIVINS IN THE EPIDIDYMIS
Activin A plays an important regulatory role in the structure of the epididymis, and the balance of activin and follistatin affects the anterior-posterior identity of the male reproductive tract. 36Activin A is an essential paracrine regulator of morphogenesis of the Wolffian duct into the highly convoluted epididymal duct. 45The degree of coiling is related to the dosage of Inhba.An increase in activin A activity in transgenic mice deficient in the follistatin 288 variant is associated with increased coiling of the proximal vas deferens in the adult. 36Activin B has similar effects to activin A in regulating epididymal morphology but appears to be less potent. 38ile the activins themselves have immunoregulatory activity, they also regulate expression of the tryptophan-metabolising enzyme indoleamine 2,3-dioxygenase (IDO), which is a critical regulator of tolerance mediated by T reg lymphocytes. 46Expression of the gene encoding IDO, Ido1, is regulated by Smad2/3/4, and mice deficient in Ido1 develop inflammation, although not autoimmunity, in the epididymis. 47In adult transgenic mice with altered activin or follistatin levels, there is a direct correlation between activin activity and Ido1 expression in the caput epididymis. 24,36Although the TGFβs can also regulate Ido1 expression via Smad2/3/4, it may be significant that constitutive IDO production is highest in the caput epididymis, where activins are most highly expressed, whereas TGFβ1-3 show the highest expression more distally in the corpus and/or cauda. 48erall, the relative contributions of activins and TGFβs to immunoregulation in the caput epididymis remain to be established.
On the one hand, the inherent activity of activins without the need for activation suggests that they could play a more immediate or constitutive role than TGFβs.On the other hand, deletion of the type 2 TGFβ receptor in dendritic cells caused a loss of tolerance and spontaneous autoimmune responses in the adult murine epididymis, 49 whereas there are no reports of autoimmune infertility in activin receptorknockout mice.Nonetheless, it seems reasonable to conclude that activins and TGFβs all contribute to endogenous immunoregulatory mechanisms in the caput epididymis.
In contrast to its immunoregulatory role in the caput epididymis, activin A appears to play a key role in induction and exacerbation of inflammation and fibrosis in the cauda, where endogenous activin levels are normally blocked by follistatin. 50Activin A gene expression and protein production increased dramatically in the cauda, but only slightly in the caput, of murine epididymides infected in vivo with uropathogenic and non-pathogenic strains of Escherichia coli.Notably, activin A also induced fibrotic responses in organ cultures of the cauda epididymis, and this effect was blocked by addition of recombinant follistatin 288.Indeed, one reason for the very high levels of follistatin production in the cauda and vas may be to limit downstream proinflammatory effects of endogenous activin produced in the caput under normal conditions.
The therapeutic potential of blocking proinflammatory and profibrotic effects of activin A using follistatin, as well as synthetic activin antagonists, has been investigated. 12,516][57] Multiple studies suggest that short non-coding RNA molecules such as microRNAs (miRNAs) regulate TGFβ1-mediated fibrosis in many organs, including the lung, heart, liver and kidney. 58,59wever, no studies have evaluated if miRNAs mediate the profibrotic activity of activin A. The cauda epididymis is significantly enriched with miRNAs, with several of these unique to this epididymal region. 60 I G U R E 4 Summary diagram of the immunological environment coinciding with the activin-follistatin production gradient along the mouse epididymis under normal conditions.Local constitutive production of activin in the proximal epididymis is associated with immunoregulation and tolerance, while follistatin production is associated with a capacity for cellular immunity and inflammation in the distal epididymis.However, activin A has both immunoregulatory and proinflammatory actions, and infection and immune activation induce a large increase in production of activin A by the cauda, which has a proinflammatory effect.miRNAs can block translation or induce degradation of target mRNA and are promising therapeutic targets because they can modulate multiple biological pathways. 58,59Current treatment guidelines for epididymitis discourage the use of anti-inflammatory drugs that reduce fibrosis because dampening the immune response may prevent bacterial clearance. 61Therefore, identifying new targets for therapeutic intervention in epididymitis is an urgent unmet need.Defining the roles of miRNAs in epididymal fibrosis and studying their interactions with activins and follistatin, will expand our knowledge on the mechanisms of fibrosis in the male reproductive tract, and potentially provide epididymal-specific therapeutic targets.

SUMMARY AND CONCLUSIONS
The activins and the activin-binding protein, follistatin display a highly regionalised pattern of expression in the epididymis, indicating roles in immunoregulation in the proximal epididymis (efferent ducts and caput), possibly acting in concert with the TGFβs, and in inflammation of the distal epididymis (cauda and vas) (Figure 4).The basis for these divergent responses of the different regions of the epididymis to activin is presumably related to the very different immunological environment of the regions of the epididymis maintained by the resident immune cell populations in each region.The fundamental regulation of the regionalisation of the epididymal immune environment remains to be determined.However, it is the inflammatory role of activin in the cauda that indicates activin-blocking therapies, in conjunction with anti-inflammatory therapies, hold considerable promise for preventing and/or repairing damage and infertility caused by inflammation of the male tract.

F
I G U R E 2 Comparison of mRNA expression of Inhba (A) and Inhbb (B) in the proximal caput (epididymal segments 1−3) and distal caput (segments 4−5) of six adult wild-type C57Bl6 mice measured using RT-qPCR (mean ± SEM, Student's t-test, *p < 0.05, ****p < 0.0001, reference gene: Rplp0).The expression pattern at day 25 (not shown) was qualitatively similar in all respects to that in the adult.proximalcaput was at least twice as high as that in the distal caput, while Inhbb was almost undetectable in the distal caput at both 25 and 56 days of age (methods detailed in Supporting Information).However, an attempted conditional knockdown of activin A by crossing mice expressing cre-recombinase under control of the promoter for the Defb41 gene, specifically expressed by principal cells in the proximal caput (courtesy of Professor Petra Sipilä, Institute of Biomedicine, University of Turku, Finland),40 and an Inhba-floxed mouse41 had marginal effects on activin mRNA or protein levels in the caput epididymis (unpublished data).TA B L E 1Relative expression levels of Inhba and Inhbb mRNA in the efferent ducts and caput epididymis detected by in situ hybridisation.

F I G U R E 3
Cellular localisation of Inhba (A, C, E, and G) and Inhbb (B, D, F, and H) mRNA by in situ hybridisation, and activin A by immunohistochemistry (I and J) in the adult mouse epididymis.(A and B) Proximal epididymis; (C and D) efferent ducts (ED); (E and F) initial segment (S1); (G and H) caput segment 2 (S2); (I) caput segment 2; (J) corpus segment 6. Arrows indicate activin A-labelled interstitial cells with macrophage morphology.Arrowheads indicate strong activin A staining on the apical cytoplasm and surface of epithelial cells.Insets: Negative controls.Four mice were examined.The localisation pattern at day 25 (not shown) was qualitatively similar in all respects to that in the adult.