Altered testicular cell type composition in males of two outbred mouse lines selected for high fertility

Recently we described two outbred mouse lines which have been selected for high fertility. These mouse models doubled the number of offspring per litter.


| INTRODUC TI ON
Fertility research largely depends on informative animal models.
Worldwide more than 1000 mouse models exist, which show a fertility phenotype. Among them, approximately 99% show a decreased fertility phenotype. By contrast, only ~ 1% of mouse models showing a fertility phenotype are associated with the annotations "enhanced fertility" and/or "increased litter size" (data extracted from Mouse Genome Informatics, MGI-www.infor matics.jax.org).
Approximately a dozen transgenic or knockout inbred mouse models exist showing an increased fertility phenotype due to single gene alterations, most of them showing an increased female fertility.
One example is the Bcl-2 knockout mouse. Due to the genetic inactivation of the Bcl-2 gene affected females show a decreased ovarian somatic cell apoptosis followed by an enhanced folliculogenesis and a higher ovulation rate. 1,2 An increased ovulation rate obviously helps in order to deliver more offspring per litter. To our knowledge, only two mouse lines have been described so far showing increased litter sizes based on the male phenotype. In a first example, mice carrying a targeted gene deletion of RBM44, an intracellular bridge protein which is highly expressed in pachytene and secondary spermatocytes, distinguished by an increased epididymal sperm count.
In vivo mating experiments reveal increased pups per litter which is based on the male side whereas females are inconspicuously. 3 The second example is a mouse line characterized by a disruption of CD46, known as a membrane cofactor involved in complement system function. Whereas again females are inconspicuous, spermatozoa from male CD46 knockout mice show an accelerated acrosome reaction. 4,5 Also in this case, higher litter sizes have been reported based on mating with CD46 knockout males. However, it should be noted that physiological consequences of increased litter sizes are moderate in those mouse models showing a single gene alteration described so far, regardless of whether the genetic modification is relevant for male or female fertility. Only moderately elevated litter sizes of 10%-20% have been reported.
In contrast to classical inbred mouse models, several attempts have been performed to select for higher fertility due to outbreeding in different species including mice. 6 In a two-factorial breeding experiment, we recently described that the fertility phenotype mainly depends on the female gender. 16 The rationale of the present study was to clarify whether longterm selection of a primarily female-derived trait of high fertility (increased litter size) also leads to alterations on the male side. We described differences in physiological and behavioural parameters as well as in endocrine and molecular markers between males of the FLs compared to control bucks. 14,15,17 Since we noticed elevated Leydig cell-specific gene transcription in FL1 bucks 18 but neither in FL2 nor in control (Ctrl) males, we were interested to address this question in more details. In the present study, we further wanted to investigate the cell type composition in testis as well as analysing sperm motility parameters under stress conditions in order to precede the characterization of the male phenotype of mice selected for high fertility.

| Animals
All procedures were performed in accordance to national and international guidelines and approved by our institutional board

| Flow cytometry
For flow cytometric experiments animals with an age of 12 to 15 weeks were euthanized by CO 2 inhalation (n = 6-10 per line).
Testes were cleaned from fat tissue, the tunica albuginea was removed, and parenchyma was dissociated in 5 mL 1 × PBS/1% BSA Statistical analysis was performed with one way ANOVA followed by Dunnett's multiple comparison test for each cell type.

| Histomorphometry
The testes have been fixed in Bouin's solution over night at 4°C,

| Tissue weight
Organ weights have been determined right after extraction from the animal. Paired organs were weighed together. Statistical differences have been analysed by Dunnett's multiple comparison test.

| Sperm motility analysis -Sperm motility stress test
The sperm motility analysis was accomplished by CASA (computerassisted sperm analysis) using SpermVision (Minitube) as described previously. 14 Prior to a 5-hour thermal stress procedure along with CASA measurements, 12-week-old males (n = 10 per line) were euthanized by CO 2 inhalation. To obtain spermatozoa the cauda epididymis was extracted, cleaned and minced (five cuts) in freshly prepared 300 µL sperm-suitable M199 media (M7528, Sigma-Aldrich) and incubated for 5 minutes at 37°C for sperm release. To exclude tissue remnants the suspension was filtered using 30 µm mesh. For generating stress conditions, the sperm suspension was continuingly exposed to 37°C for a 5 hours thermal stress period.
Every hour the current sperm motility characteristics were analysed by applying a fraction of 3 µl to 37°C tempered chamber slides (20 micron, Leja). Sperm motility parameters represent an average of 8 defined chamber partitions were viewed and measured by CASA system (Minitube). For hyperactivation experiment spermatozoa were solved in PBS and hyperactivation has been induced by PBS containing 2 mM CaCl 2 , 25 mM NaHCO 3 and 3 mg/mL BSA 21 and measured at t = 0 min and t = 30 min. Statistical analysis was carried out by two-way ANOVA using the software package GraphPrism 5.01 (GraphPad Software).

| Elevated Leydig cell population in FL1
We recently described the characterization of two mouse lines (FL1 and FL2) which have been selected for high fertility. Although these mice have been primarily selected for a female-derived trait of fertility (increased litter size and elevated birth weight), we noticed distinct alterations also on the male side. Beside altered testicular gene expression patterns and endocrine variations, we also observed distinct physiological characteristics eg in behavioural assays. 17,18 Since we have previously detected an increased diploid cell population in males of the FL1 line 18  in Ctrl and FL2 males, respectively ( Figure 2B). Even so, the number of Leydig cells almost doubled in FL1 males the Cyp17 protein load per cell is slightly decreased ( Figure 2C).

| Size of gonads
To test whether these alterations at the single cell level might also be reflected by different sizes of reproductive organs, we measured the weights of testis, prostate (ventral and lateral) and seminal vesicles. As shown in Figure 3A , we did not notice any differences in testis weight (sum of weights of left and right testis). The prostate weight showed only a tendency of being slightly increased in males of the fertility lines (FL1: +28%, FL2: +21%, Figure 3B). However, we detected significantly elevated weights of seminal vesicles in bucks of FL1 and FL2 (FL1: +51%, FL2: +44%) compared to Ctrl males ( Figure 3C).

| Histomorphometry
To further analyse morphological characteristics of testis, we measured the diameter of tubulus seminiferi, the area of interstitium as well as the area of tubulus seminiferi from Ctrl, FL1 and FL2 males.
We did not notice significant differences between FL1 and Ctrl bucks ( Figure 4). By contrast, we found alterations in FL2 compared to Ctrl males. Diameter of tubulus is reduced, area of tubulus is increased whereas area of interstitium is decreased in FL2 compared to Ctrl males ( Figure 4A-C). We also counted the abundance of round spermatids (St) and Sertoli cells (SC) in all three lines. We noticed slightly elevated round spermatid numbers in FL2 compared to FL1 and Ctrl ( Figure 4D). Since the number of Sertoli cells remained unaffected the Sertoli cell index (St/SC ratio) was increased in FL2 ( Figure 4D,E).

| CASA
Previous sperm motility analysis did not reveal significant differences between lines at t = 0 minute at standard conditions; however, we observed a significant decrease in thermal stability in FL2 spermatozoa compared to Ctrl spermatozoa over a 5 hours observation window. 14 To test whether this phenomenon is also true for FL1 spermatozoa, we applied the thermal stress experiment for FL1 spermatozoa. Again, we noticed no difference in sperm motility parameter between Ctrl and FL1 bucks at t = 0 minutes ( Figure 5) in agreement with previous data. 18 However, after thermal stress we noticed a significant better performance of FL1 spermatozoa compared to Ctrl spermatozoa. This effect is obvious in percentages of motile spermatozoa (+10% in FL1 compared to Ctrl at t = 5 hours, Figure 5A,B) as well as in various sperm velocity parameters (eg VCL: +10 µm/s elevated pace compared to Ctrl at t = 5 hours, Figure 5D). We also

| D ISCUSS I ON
The cell-type composition in testis is highly dynamic during development due to a complete lack of mature germ-cell stages prior to the onset of puberty. [23][24][25] Also in adults within one species cell-type composition might differ eg in cases where dominant and philopatric males live together in one mouse colony. 26 Surprisingly, we noticed differences in testicular cell type composition in (single caged) males of FL1 compared to males of Ctrl and FL2 (Figure 1). We observed an elevated percentage of diploid cells in FL1 bucks. Further analysis identified these elevated cell population as Leydig cells due to the expression of Leydig cell marker Cyp17 (Figure 2A,B). Approximately

1/4 of the diploid cell population within the testis correspond to
Leydig cells in rodents 24 which is also the order of magnitude in Ctrl and FL2 (27% and 24%, respectively, Figure 2B). By contrast, this cell portion is increased to 51% in FL1 males ( Figure 2B). Of note, the amount of Cyp17 load per positive cell is decreased in FL1 compared to Ctrl and FL2 ( Figure 2C). Elevated numbers of Leydig cells in FL1 are not accompanied by significantly increased intestinal area ( Figure 4B) and are not associated with Leydig cell hyperplasia and decreased fertility as it has been described for several knockout mouse models. [27][28][29] Accelerated gene transcription of Leydig cell markers in FL1 has already been reported in previous work. 15,18 Based on this observation, one might suggest elevated testosterone concentrations in FL1 males compared to Ctrl and FL2 males.
However, we did neither observe enhanced testosterone, 4-androstenedione, 3α-androstanediol, DHEA nor increased DHT concentrations in FL1 males. 16 Thus, a raised Leydig cell number in conjunction with augmented expression of Leydig cell markers relative to total testicular transcripts (or total testis weight, respectively) did not reflect increased serum androgen concentrations in FL1 bucks. Since we noticed slightly elevated progesterone and corticosterone levels in FL1, 16 the effect of unaffected steady-state androgen concentrations might be explained by an accelerated metabolic turnover rate in these animals. In line with this assumption, we noticed altered steady-state transcript levels of steroid hormone clearance enzymes (such as Sult1e1) in holistic gene expression analysis from FL1 testis. 15 Beside the responsibility for sperm production, the testis is as an endocrine organ and the hormonal demand influences its weight. An elevated androgen activity would point towards an increased testicular weight, which we did not observe. Testis weights were shown to remain unaffected in FL1 compared to FL2 and Ctrl males. However, we could detect the weight of seminal vesicles known that size and/or weight of accessory sex glands are more or less directly related to androgens levels. 30,31 Hence, according to our study findings, we cannot exclude elevated androgen activities in FL bucks. However, the testes weights of the Dummerstorfer outbred lines are in the upper range compared to different tissue weights, which were described in a systematic study of eight different inbred lines ( Figure 3A) and. 32 Furthermore, analysis of tissue weights might be biased by differing body weights of the animals at the time point of analysis. As we described recently males of the FL2 lines are generally heavier compared to Ctrl and FL1 males 16 as well as compared to most inbred mice. 32 Thus normalized to body weight, FL2 males appear to have rather smaller testis compared to FL1.
Histomorphometric analysis also revealed differences in FL2 compared to FL1 and Ctrl testis. The diameter of tubuli seminiferi and the area of interstitium are smaller, whereas area of tubuli seminiferi is greater in FL2 testis compared to FL1 and Ctrl males. By contrast, we did not observe statistically significant differences  between FL1 and Ctrl bucks (Figure 4). Since we also observed an elevated Sertoli cell index (St/SC ratio) in FL2 ( Figure 4D), which might point towards a higher Sertoli cell efficiency and a higher spermatogenic efficiency in FL2 compared to FL1. 33 St/SC ratios identified in our mouse lines (between 6 and 8, Figure 4D) are rather smaller compared to reports from other mouse strains (between 8 and 11) as well as compared to most other mammals. [34][35][36] Unfortunately, we cannot correlate spermatid/Sertoli cell indices from our lines with sperm production rates since we prepared epididymal spermatozoa in the present study,however, a negative correlation between tubulus diameter and spermatid/Sertoli cell index (as observed for FL2, Figure 4A,D) has also been described in cats. 37 We recently described a decreased successful mating rate of 82% in FL2, 14 which is relatively low compared to wild type mice (98%) 38 as well as compared to FL1 and Ctrl bucks (94% and 95%, respectively). 15 In this context it should be noted that successful mating rates of commonly used inbred mouse strains such as C57BL/6J, CBA/J, A/J and BALB/cJ are with 55%-85% much lower. 39 Figure 5) and compared to FL2 spermatozoa. 14 In a hyperactivation assay of spermatozoa from all three lines, we noticed an elevated percentage of hyperactivated spermatozoa already in uninduced spermatozoa from FL1 and FL2 compared to Ctrl animals. However, this already elevated proportion of hyperactivated spermatozoa could not be increased further by the bicarbonate/BSA treatment ( Figure 5G). We recently described that the fertility phenotype of the mice is mainly depending on the female gender in a two-factorial breeding experiment using pure-bred animals. 16 In the present study as well as in previous studies, we demonstrated that selection for a primarily female-derived trait of high fertility also leads to complex alterations on the male side. Some of these (subtle) changes seem to be beneficial for adequate fertility, while others tend to have opposite effects (see above). Therefore, we cannot claim from these data that male fertility is generally compromised by selection for a primary female feature of increased fertility, as suggested recently. 40 Taken together, this data indicates that males as well as females of the two high fertile mouse lines-FL1 and FL2-developed different strategies in order to maintain a high-fertility phenotype.
Furthermore, there seems to be no direct link between fertility of both sexes. High fertility on female side does not necessarily mean elevated fertility in males within the same line. Further experiments will focus on the identification of pathways that are differentially regulated in females of the high-fertility lines and their possible impact in defining the male phenotype.

ACK N OWLED G EM ENTS
The authors would like to thank Christian Plinski, Petra Reckling,

CO N FLI C T O F I NTE R E S T
All authors declare that they have no conflict of interests.

AUTH O R S' CO NTR I B UTI O N S
MM: conceived study, performed experiments, analysed data, carried out statistical analysis; AS, CL, HM: performed experiments, analysed data, carried out statistical analysis; ML: conducted animal breeding, analysed data; JS: conceived study, interpreted data; JW: conceived study, analysed data, interpreted data, drafted the manuscript. All authors have read and approved the manuscript.