Multiple aspects of male germ cell development and interactions with Sertoli cells require inositol hexakisphosphate kinase-1

Inositol hexakisphosphate kinase-1 (IP6K1) is required for male fertility, but the underlying mechanisms have been elusive. Here, we report that IP6K1 is required for multiple aspects of male germ cell development. This development requires selective interactions between germ cells and Sertoli cells, namely apical ectoplasmic specialization. Spermiation (sperm release) requires tubulobulbar complexes. We found that the apical ectoplasmic specialization and tubulobulbar complexes were poorly formed or disrupted in IP6K1 KOs. Deletion of IP6K1 elicited several aberrations, including: 1, sloughing off of round germ cells; 2, disorientation and malformation of elongating/elongated spermatids; 3, degeneration of acrosomes; 4, defects in germ-Sertoli cell interactions and 5, failure of spermiation. Eventually the sperm cells were not released but phagocytosed by Sertoli cells leading to an absence of sperm in the epididymis.

In mice, spermatogenesis has been classified into 12 stages according to the arrangement of different germ cells that are developmentally synchronized in the seminiferous epithelium. In addition, the differentiation of round spermatids into mature, motile spermatozoa is divided into 16 steps. At step 16 mature, motile sperm are released (spermiation) corresponding to stage VII to VIII in spermatogenesis 20 .
Developing round and elongating/elongated spermatids must maintain stable attachments with Sertoli cells to prevent sloughing of immature germ cells from the seminiferous epithelium, which may result in infertility 21 . The apical ectoplasmic specialization (ES) is a unique adhesion junction formed between elongating/elongated spermatids and Sertoli cells, and is crucial for attachment of spermatids as well as their movement and orientation during spermatogenesis 22 . Disruption of germ cell-Sertoli cell interactions 23 or spermatid polarization 24 leads to spermatogenic defects and infertility. The molecular composition of apical ES includes integral membrane proteins, adaptor proteins, signaling proteins and cytoskeletal proteins. However, little is known about the regulation of apical ES dynamics 22,25 .
The apical tubulobulbar complex (TBC) is another linkage between maturing spermatids and Sertoli cells, formed by cytoplasmic processes extending from spermatids into the neighboring Sertoli cells 26 . In mice apical TBCs are transiently formed at spermatogenic stage VII prior to sperm release and function to remove excess cytoplasm from spermatids 27 and to remove adhesion junctions between spermatids and Sertoli cells 26,28 . The apical TBC is comprised of a long proximal tubule followed by a bulb, then a short distal tubule and a coated pit 26 . Actin filaments are highly enriched in apical TBCs 29 . Defects of apical TBCs caused by chemical treatment or genetic mutation result in spermiation failure [30][31][32] .
In the present study we explored the roles of IP6K1 in male germ cell development. The elongating/elongated spermatids of IP6K1 KO were disoriented and malformed. The acrosomes of IP6K1 KO spermatids were also degenerated during differentiation. Defective apical ES and malformed apical TBCs were apparent in the IP6K1 mutants. Eventually, the defective sperm cells failed to release and were engulfed by Sertoli cells at stage VII and stage VIII in the seminiferous epithelium of KO mice, which can account for the sperm deficits.

Results
Deletion of IP6K1 leads to male infertility caused by scarce and malformed sperms. Deletion of IP6K1 in mice was shown to elicit male infertility 1,11 . In agreement with this finding, we observed a complete absence of offspring from IP6K1 deleted males, whereas fertility appeared normal in IP6K1 KO females (Fig. S1a). In contrast to the lack of fertility of IP6K1 KO males, mice with deletion of IP6K2, IP6K3 or double knockout of IP6K2 and IP6K3 displayed normal fertility (Fig. 1a). We also examined the sexual behavior and ability of mice to mate, which appeared to be normal in the IP6K1 KO males (Fig. 1b).
Reproductive behavior and spermatogenesis is regulated by levels of reproductive hormones. However, plasma levels of FSH, LH and testosterone were normal in IP6K1 KO male mice ( Fig. S1b-d). Thus, the fertility defect of IP6K1 KO male mice does not arise from hormonal deregulation.
We collected motile sperm cells from epididymides and observed very little cellular content and no motile sperm cells in the IP6K1 KO mice (Fig. 1c). We observed severed sperm tails in the IP6K1 KO preparations with some residual DNA surrounded by sperm tails (Figs 1d and S1e).

Deletion of IP6K1 causes sloughing of round germ cells.
We first looked at the gross morphology of the reproductive organs. Testes of the IP6K1 KOs were reduced in size (Fig. S2a,b), and testicular weight was diminished by about 20 percent (Fig. S2b). We then looked at the microscopic morphology. In testes of IP6K1 KOs, spermatids were more loosely attached to Sertoli cells than for wild-type animals (Fig. 2a), and round spermatids were dislodged (Fig. 2b). The sloughing of germ cells in IP6K1 KO seminiferous tubules was confirmed by the presence of round cells in the cauda epididymides of IP6K1 mutant mice (Fig. 2c). Thus, deletion of IP6K1 affected interactions between germ cells and Sertoli cells.
We examined sperm in testes by staining sperm tails for acetyl-α-tubulin, which revealed a virtually total loss of mature sperms in IP6K1 KOs (Fig. 2d). In both caput and cauda epididymides of wild type animals, the lumens were typically filled with sperm cells. The lumens of IP6K1 KO epididymides, however, were devoid of sperm, whereas levels in IP6K2 and IP6K3 KOs were comparable to wild-type animals (Fig. S2c,d).
Immunofluorescence staining of testes revealed abundant IP6K1 localized to the germ cells of wild-type mice, which was completely absent in IP6K1 KOs (Fig. S4a). We noted major differences in the cellular localizations of IP6K1, IP6K2 and IP6K3 in testes (Fig. S4b). IP6K1 was ubiquitously expressed in germ cells; IP6K2 occurred in primary spermatocytes, while the expression of IP6K3 was low and only evident in step 12 spermatids (Fig. S4b). IP6K1 was expressed in elongating/elongated spermatids (Fig. S4c,d), implying that structural/functional defects in these cells in mutant mice stems from the loss of IP6K1 (Fig. 3). We explored a potential loss of Sertoli and Leydig cells in the IP6K1 KOs. Staining of Sertoli cells for GATA4 and of Leydig cells with calretinin revealed relatively normal numbers of these cell types (Fig. S5a,b). IP6K1 deletion causes disorientation of spermatids and failure of spermiation. We examined cellular morphology in the testes by Toluidine blue staining (Fig. 4). Disorientation, defined as a deviation of more than 90 degree from pointing to basement membrane perpendicular 33 , and abnormal morphology of elongating/elongated spermatids were pronounced in the mutants (Fig. 4a-c). Dislodged round spermatids were ScienTific RepoRtS | (2018) 8:7039 | DOI:10.1038/s41598-018-25468-8 evident in the lumens of seminiferous tubules of IP6K1 KOs (Fig. 4b,c and e). Most striking was the substantial engulfment of sperm by Sertoli cells (Fig. 4d-f). In wild type mice, the sperm were released at stage VII-VIII, and the lumens were full of sperm (Fig. 4d). However, sperm were not released in IP6K1 KOs. Instead, they were "eaten" by Sertoli cells (Fig. 4d-f). The phagocytosis of sperm cells was prominent at stage VII ( Fig. 4d) and stage VIII (Fig. 4e), when they should be released. The engulfed sperm cells were also observed at stage IX (Fig. 4f).
To further assess mechanisms underlying the abnormalities in IP6K1 KOs, we evaluated different steps of spermatid differentiation by electron microscopy (Fig. 5). Development of round spermatids was relatively normal in IP6K1 KOs ( Fig. 5a-f). The most notable abnormalities occurred in the disposition of elongating/elongated spermatids at steps 9-16 ( Fig. 5g-l). Disorientation and malformation of spermatids was marked at steps 9-16 ( Fig. 5g-l), indicating that spermatids lose cell polarity in IP6K1 KOs. At steps 12-16 in the KOs (Fig. 5j-l), the structure of spermatid heads was abnormal or improperly formed. The step 16 spermatids were not released but eventually engulfed by Sertoli cells in IP6K1 KO stage VII and VIII seminiferous tubules (Fig. 6a). Severed sperm tails were seen in stage IX of the seminiferous epithelium of IP6K1 KOs (Fig. 6b).

IP6K1 deletion disrupts cell polarity and causes malformation of acrosome.
We examined cell polarity of the spermatids. In IP6K1 KO seminiferous tubules, the elongating spermatids lost orientation (Fig. 6c), and some elongating spermatids showed double acrosome granules (Fig. 6d), indicating that cell polarity was disrupted. We examined the structure of elongating/elongated spermatids in IP6K1 KOs and found malformations impregnate wild type females. By contrast, IP6K2 KO males (n = 6), IP6K3 KO males (n = 6) and IP6K2/IP6K3 double KO males (n = 6) are able to produce progeny. (b) IP6K1 KO male mice (n = 7) were set to mate with wild type female mice, mating plugs were found in six of 7 wild type female mice. Six pairs of WT male and female mice were set as controls. (c,d) Sperm cells were isolated from epididymides. (c) Brightfield microscopy shows severed sperm tails (arrows) but no motile sperm cells in IP6K1 KOs. Scale bar 50 μm. (d) Immunostaining of acetyl-α-tubulin for sperm tails. Nuclei were stained by DAPI. Severed sperm tails (arrow) and sperm heads coiled by sperm tails (arrowhead) were seen in the IP6K1 mutants. Scale bar 20 μm.

IP6K1 deletion disrupts apical ectoplasmic specialization.
We examined the interaction between elongating/elongated spermatids and Sertoli cells, known as the apical ectoplasmic specialization (ES) (Fig. 7a-e). The actin filament bundles that align perpendicular to the Sertoli cell plasma membrane were a characteristic feature of the apical ES in the WT preparations ( Fig. 7a-e). At step 8 of spermatid differentiation, the spermatid nuclei made contact with the cell surface. The apical ES was formed between the spermatids and Sertoli cells of WT mice. In IP6K1 KOs, however, no characteristic apical ES was observed at step 8 (Fig. 7a). Apical ES was observable in step 10-16 spermatids of the IP6K1 KOs. However, the structures were disrupted. The apical ES was not well formed in step 10-12 spermatids (Fig. 7b,c), appeared to be collapsing in step 14 spermatids of IP6K1 KOs (Fig. 7d), and collapsed in step 16 spermatids of the IP6K1 mutants (Fig. 7e).
The staining of F-actin on testes showed that in WT seminiferous tubules, sperm cells covered with abundant actin filaments. In IP6K1 KOs, however, the cells were only partially associated with actin bundles (Fig. 7f,g). IP6K1 deletion disrupts apical tubulobulbar complexes. Sperm release requires normal functioning of apical tubulobulbar complexes [30][31][32] . We examined the morphology of apical tubulobulbar complexes (TBCs) in the IP6K1 mutants (Fig. 8a). The apical TBCs in the IP6K1 KOs were malformed with absence of proximal tubules and swelling of the bulbs (Fig. 8a). Thus, the defective apical TBCs lead to spermiation failure (Fig. 4d,e) and excess residual cytoplasm in IP6K1 KO spermatids (Fig. 8b) 27,32 .

Discussion
In the present study we have developed insights into the role of IP6K1 in male germ cell development. The infertility of IP6K1 KO mice is evidently caused by deficits in both quantity and quality of sperm. We observed several defects: 1, sloughing of round germ cells; 2, disorientation and malformation of elongating/elongated spermatids; 3, degeneration of acrosomes; 4, disrupted apical ES and apical TBCs; 5, failure of spermiation in the IP6K1 KO mice. Eventually sperm were phagocytosed by Sertoli cells, a mechanism which may fully account for the observed sperm deficit 23,24 . Our results are consistent with previous reports 1, 11 , including: the expression of IP6K1 in germ cells; complete loss of fertility in IP6K1 KO males but normal fertility in females; normal testosterone levels in IP6K1 KO males; reduced size and weight of testes in IP6K1 KO mice; absence of mature spermatozoa, and premature release of round cells in IP6K1 KO epididymides; abnormal morphology and incomplete nuclear condensation in elongating/elongated spermatids. Germ cells and Sertoli cells interact both physiologically and structurally, processes that are fundamental to male fertility 23,34 . The apical ES is a dynamic structure, which normally organizes and orients the spermatids within the seminiferous epithelium. Fragmentation or malformation of the apical ES is associated with infertility 22,35,36 . The apical ES comprises multiple cytoskeleton proteins, adaptor proteins and signaling proteins 22 . One of the most typical features of the apical ES is the array of actin filament bundles that lie perpendicular to the Sertoli cell plasma membrane and are sandwiched between the cisternae of the endoplasmic reticulum and the Sertoli cell plasma membrane 37,38 . Emerging evidence has shown that junctional dynamics of the apical ES are supported by F-actin and the microtubule based cytoskeleton 25 . In IP6K1 KOs, the apical ES were not well formed or maintained, the actin bundle of the apical ES was collapsed in later stages of spermatid development. IP6K1′s actions often are mediated by binding with other proteins 5,7,[39][40][41] . We have shown that IP6K1 interacts with α-actinin 5 , which regulates junctions of Sertoli cells and germ cells as well as maintaining the apical ES and spermiation 42 . These interactions with α-actinin may mediate the disrupted apical ES and spermiation in the IP6K1 mutants. It is also possible that IP6K1 binds with junctional proteins that affect the Sertoli-germ cell interactions. Apical ES is required for the orientation of spermatids in the seminiferous tubules. The abnormal positioning of spermatid heads of IP6K1 KOs within the epithelium is likely due to defects in the apical ES 43 . The apical ES also plays a role in shaping the spermatid head 44,45 . The malformation of elongated spermatids in IP6K1 KOs might be the result of defective apical ES and apical TBCs. Thus, the observed aberrations in IP6K1 KOs, such as dislodged round spermatids and malformation of elongating/elongated spermatids, might reflect defective interactions between spermatids and Sertoli cells. On the other hand, the acrosomal and nuclear shape changes in IP6K1 KOs may also derive from deficits of the Golgi complex. Acrosomal formation requires active trafficking from the Golgi apparatus 46 . IP6K1 regulates dynein-dependent trafficking pathways, and cells lacking IP6K1 display defects in Golgi maintenance 47 .
Apical TBCs develop at spermatogenic stage VII, and reflect interactions between maturing spermatids and Sertoli cells. Components of apical TBCs include actin filaments, actin binding proteins, adhesion molecules and endocytic proteins 29 . Apical TBCs remove cytoplasm from spermatids and eliminate adhesion junctions between spermatids and Sertoli cells 26,27,32 . Defective apical TBCs fail to remove spermatid-Sertoli cell adhesions, which leads to spermiation failure [30][31][32]48 . In IP6K1 KO mice the apical TBCs were malformed with absence of proximal tubules and swelling bulbs, which may account for spermiation failure.
In IP6K1 KOs, the interactions between spermatids and Sertoli cells were not well formed and prematurely collapsed. Deletion of IP6K1 elicited malformation of elongating/elongated spermatids. This degeneration would also eventually lead to engulfment of the spermatids by the Sertoli cells, and thereby contribute to spermiation failure 18,19 .
IP6K1 regulates endocytic trafficking 49 . Disruption of endocytosis in EHD4 KO mice results in male infertility similar to that of IP6K1 KO mice 50 . Accordingly, the deficits of male germ cell development in IP6K1 KOs may reflect defects of protein trafficking in membranes of the apical ES and apical TBCs.
5-IP7 affects cell-cell adhesion and adhesion-dependent signaling 3,6 . It is uncertain whether the male infertility in IP6K1 KOs is dependent on 5-IP7 or IP8, both products of IP6K1. The IP6K inhibitor TNP does not induce infertility in mice, which may reflect its inability to penetrate the blood-testis barrier 51 .
In this study, we isolated motile sperm cells after euthanizing mice with CO 2 . This may not be an optimal way to obtain sperm for motility assessment, as CO 2 inhibits sperm motility 52 . Nonetheless, the deficits of IP6K1 KO sperm cells were so striking that they are unlikely to derive primarily from CO 2 treatment.
Non-obstructive azoospermia is a common form of male infertility. Specific molecular mechanisms underlying the majority of these cases have been elusive. Our present findings, as well as the observations of Bhandari and colleagues 11 indicate that loss of IP6K1 affects multiple steps of sperm development. Thus IP6K1 may be a marker for male infertility with prognostic value.
Animals. IP6K1, IP6K2 and IP6K3 KO mice (C57BL/6) were generated by heterozygous breeding 2,10 . The wild-type and IP6K1 KO mice were littermates. IP6K2/IP6K3 double knockout mice were generated by breeding IP6K2 KOs with IP6K3 KOs. Animal breeding and procedures were conducted in strict accordance with the NIH Guide for Care and Use of Laboratory Animals and were approved by the Johns Hopkins University Committee Mice sexual behavior. Male mouse sexual behavior was evaluated by examining the female's face and anogenital region. For analysis of fertility after normal mating, male mice were housed with females for 3 months. Vaginal plugs, pregnancies, and pups were recorded. Copulatory ability was determined by the appearance of vaginal plugs.

Isolation of motile sperm cells and spermatids.
Male mice were euthanized by CO 2 . Motile sperm cells were obtained by gently squeezing the excised epididymides; cells were collected in Krebs-Ringer solution. For collecting spermatids from testes, the testes were minced and forcefully mixed; the dislodged spermatids were collected in Krebs-Ringer solution.
Immunofluorescence staining. Wild-type and IP6K1 KO mice were euthanized, then perfused and fixed with 4% (wt/vol) paraformaldehyde. The sections were cut 30 μm thick. The slides were blocked with 10% goat serum for 10 min, then incubated with primary antibodies (1:100 dilution) at 4 °C overnight. The fluorescence labeled secondary antibody (1:500 dilution) was incubated for 1 h at room temperature. Nuclei were counterstained with DAPI. Slices were mounted with ProLong Gold Antifade Mountant. Pictures were taken with a confocal microscope (Zeiss LSM 700, NIH Grant# S10 OD016374).
Hematoxylin and eosin staining. Wild-type and IP6K1 KO mice were euthanized, then perfused and fixed with 4% (wt/vol) paraformaldehyde. Testes and epididymides were excised and embedded in paraffin. Sections were cut 5 μm thick and stained with hematoxylin and eosin following standard procedures.
Blood FSH, LH and testosterone. Blood samples from wild type and IP6K1 KO mice were collected.
The plasma concentrations of FSH and LH were determined by using a mouse FSH ELISA kit (Neobiolab) and a mouse LH ELISA kit (Neobiolab). The plasma concentrations of testosterone were determined by using a testosterone ELISA kit (Enzo Life Sciences).
Transmission electron microscopy and Toluidine blue staining. Wild-type and IP6K1 KO mice were euthanized, then perfused and fixed with buffer containing 2% (wt/vol) glutaraldehyde, 2% (wt/vol) paraformaldehyde, 100 mM sodium cacodylate and 3 mM MgCl 2 , pH 7.4. Testes were excised and post-fixed in 1.5% potassium ferrocyanide reduced 2% osmium tetroxide in 100 mM sodium cacodylate with 3 mM MgCl 2 for 2 h on ice in the dark. After a brief rinse in 100 mM maleate buffer containing 3% sucrose, samples were placed in 2% uranyl acetate in maleate/sucrose for 1 h at 4 °C with slow rocking in the dark. Samples were then dehydrated through a graded series of ethanol, transferred through propylene oxide and were embedded in Eponate 12 Resin and cured at 60 °C for two days.
For transmission electron microscopy, sections were cut 80 nm thick and stained with uranyl acetate followed by lead citrate. Grids were viewed on a Hitachi 7600 TEM operating at 80 kV and digital images captured with an XR50 (5 megapixel) CCD camera.
For Toluidine blue staining, sections were cut 0.5 μm thick and stained with 1% toluidine blue and 1% sodium borate for 1 min. Statistical analysis. Quantitative data are expressed as means ± SEM. Data were analyzed by unpaired Student's t-test. P < 0.05 was considered statistically significant. For non-quantitative data, results were representative of at least 3 independent experiments. Data Availability. All data generated or analysed during this study are included in this published article (and its Supplementary Information files).