Mouse dead end1 acts with Nanos2 and Nanos3 to regulate testicular teratoma incidence

Spontaneous testicular teratomas (STTs) derived from primordial germ cells (PGCs) in the mouse embryonic testes predominantly develop in the 129 family inbred strain. Ter (spontaneous mutation) is a single nucleotide polymorphism that generates a premature stop codon of Dead end1 (Dnd1) and increases the incidence of STTs in the 129 genetic background. We previously found that DND1 interacts with NANOS2 or NANOS3 and that these complexes play a vital role in male embryonic germ cells and adult spermatogonia. However, the following are unclear: (a) whether DND1 works with NANOS2 or NANOS3 to regulate teratoma incidence, and (b) whether Ter simply causes Dnd1 loss or produces a short mutant DND1 protein. In the current study, we newly established a conventional Dnd1-knockout mouse line and found that these mice showed phenotypes similar to those of Ter mutant mice in spermatogenesis, oogenesis, and teratoma incidence, with a slight difference in spermiogenesis. In addition, we found that the amount of DND1 in Dnd1+/Ter embryos decreased to half of that in wild-type embryos, while the expression of the short mutant DND1 was not detected. We also found that double mutants for Dnd1 and Nanos2 or Nanos3 showed synergistic increase in the incidence of STTs. These data support the idea that Ter causes Dnd1 loss, leading to an increase in STT incidence, and that DND1 acts with NANOS2 and NANOS3 to regulate the development of teratoma from PGCs in the 129 genetic background. Thus, our results clarify the role of Dnd1 in the development of STTs and provide a novel insight into its pathogenic mechanism.


Introduction
Testicular teratomas are tumors that originate from germ cells. A diverse array of cell and tissue types are found differentiating in these tumors: erythrocyte, adipocyte, cartilage, muscle,

Mice
We had previously established a Dnd1_flox mouse line by using TT2 ES cells and maintained it via intercrosses generating Dnd1 flox/flox mice. To establish the Dnd1 conventional knockout mouse line, we crossed these Dnd1 flox/flox mice with Rosa26 +/CreERT2 mice and obtained Dnd1 +/ flox ; Rosa26 +/CreERT2 female mice. After administering 75 mg per kg body weight of tamoxifen, we crossed these female mice with wild-type male mice and obtained Dnd1 +/Δ offspring. The Dnd1 +/Δ mice were backcrossed into the BL6, 129, or MCH strains for at least eight generations. All three mouse strains were purchased from CLEA Japan (CLEA Japan, Inc., Japan). The 129 +Ter /SvJcl strain had been established from an STT-high permissive 129/terSv strain [4] by Dr. Noguchi [5] and provided to and maintained at CLEA Japan, whereas the Dnd1 +/Ter mice had been bred in the animal facility at Shizuoka University by Drs. Tokumoto and Noguchi. Genotyping of Dnd1-flox, Δ, and Ter alleles was performed as described previously [9,14]. The Nanos2 +/LacZ and Nanos3 +/Cre mouse lines were established as previously described [17] and backcrossed into the 129 +Ter /SvJcl strain for at least eight generations.

Histological methods
For histological analysis, the testes and ovaries were fixed with Bouin's solution and embedded in paraffin. After sectioning (6 μm), the samples were stained with hematoxylin and eosin.

Tumor surveys
Four-week-old male mice were surveyed for testicular teratomas. Teratoma incidence was calculated as the percentage of male mice with at least one testicular teratoma. Histological analysis (hematoxylin and eosin staining) was used to confirm any teratomas that were ambiguous at autopsy.

Sperm count
Mature spermatozoa were isolated from the caudal epididymis, as described previously [18]. Briefly, both epididymides from each mouse were minced and incubated in 1 mL warm phosphate-buffered saline (PBS) for 30 min, following which the sperm suspension was fixed in 10% neutral-buffered formalin. The sperm were counted using a hemocytometer. Data have been shown in terms of mean ± standard error of mean (SEM), and the values were statistically analyzed using a Student's t-test.

In vitro fertilization (IVF)
Three-week-old female B6C3F1 mice were purchased from CLEA Japan and superovulated using intraperitoneal injections of 100 μL of CARD HyperOva (Kyudo Co., Ltd., Japan) followed by intraperitoneal injection of 100 μL of 50 units/mL human chorionic gonadotropin (hCG; Aska Pharmaceutical Co., Ltd., Japan) 48 h later. Eggs were recovered 16 to 17 h after hCG injection and placed in a 200 μL CARD MEDIUM (Kyudo Co., Ltd.) drop covered with liquid paraffin (Nacalai Tesque, Inc., Japan) and then used for the experiments 0.5-1 h after preparation. The spermatozoa were collected from the cauda epididymis of more than 12-week-old wild-type or Dnd1 +/Δ male mice of the 129 strain, suspended in 100 μL CARD FERTIUP (Kyudo Co., Ltd.), and incubated for 1 h for capacitation. Capacitated sperm (3 μL) were added to the drop containing eggs. The eggs were washed in an 80 μL mHTF (Kyudo Co., Ltd.) drop three times after 3 h of coincubation and were transferred to 100 μL KSOM (Kyudo Co., Ltd.) 6 h after insemination. The blastocysts were counted 4 days later.

Comparison of the PGC number in male gonads at E11.5
Embryonic stage was determined as E11.5 on the basis of the number of somites, while the sex of the embryos was determined by polymerase chain reaction (PCR) for the Sry gene using the following primer pairs: Sry-F: 5 0 -ggttgcaatcataattcttcc-3 0 and Sry-R: 5 0 -cactcctctgtgacactttag-3 0 .
For section immunostaining, embryonic gonads were fixed with 4% paraformaldehyde overnight at 4˚C and then embedded in paraffin. Whole gonads were sectioned (6 μm) and autoclaved with Antigen Unmasking Solution (Vector Laboratories, Inc., USA). After the samples were subjected to blocking with 5% skim milk in PBS, they were incubated overnight at 4˚C with primary antibodies against deleted in azoospermia-like (DAZL; 1:1,800) [14] and NANOG (1:5,000; IHC-00205, Bethyl Laboratories, Inc., USA) in Can Get Signal immunostain (NKB-501; Toyobo Co., Ltd., Japan). After the samples were washed, they were incubated with Alexa 488-or Alexa 594-conjugated IgG antibodies at 25˚C. The sections were enclosed in Gel/Mount (Biomeda Corp., USA) and observed using fluorescence microscopy (Axio Imager M2, Carl Zeiss, Germany). The number of germ cells in each genital ridge section was assessed by Image J (version 1.50i).
All antibodies used in western blotting analysis were diluted by Can Get Signal Immunoreaction Enhancer Solution (NKB-101; Toyobo Co., Ltd.).

Statistical analysis
The statistical significance of differences in the ratios of testis weight per body weight, ratios of defective tubules, sperm count, and relative DND1 expression were assessed using a two-tailed t-test, whereas the data for genotype distribution in the intercrosses of 129 Dnd1 +/Δ heterozygotes and tumor surveys were analyzed with χ 2 analysis and Fisher's exact test, respectively. A P-value of <0.05 was considered to represent statistical significance. Statistical analysis was performed using Microsoft Excel for Mac (version 16.33) or R (version 3.6.0).

Homozygous for Dnd1-Δ mutant mice were born at Mendelian ratios and developed testicular teratoma in the 129 genetic background
To clarify the effect of Dnd1 loss, we aimed to establish a new conventional Dnd1-knockout mouse line. We had previously generated a Dnd1-flox mouse line to conditionally remove DND1 by Cre recombinase [14]. In the current study, a conventional knockout (Dnd1-Δ) mouse line was generated from the Dnd1-flox mice by Cre-mediated excision of exons 2 and 3, which included the two RRMs of DND1 (S1D and S1E Fig), and was subsequently backcrossed into three different mouse strains: BL6, MCH, or 129. Intercrosses of Dnd1 +/Δ heterozygotes within each strain produced the Dnd1 Δ/Δ homozygotes in accordance with a Mendelian expectation of 1:2:1 in all three mouse strains (Table 1), showing no evidence for embryonic lethality of the homozygous mutants even in the 129 genetic background.
We then examined whether the Dnd1-Δ allele affects the occurrence of testicular teratoma. For this purpose, testes from 4-week-old Dnd1 Δ/Δ mice, Dnd1 +/Δ mice, and their Dnd1 +/+ wild-type control littermates of all three strains were harvested and surveyed for teratomas ( occurrence in male mice from a baseline of 4.2% to 28.8% in Dnd1 +/Δ heterozygotes and 92.6% in Dnd1 Δ/Δ homozygotes, which are ratios similar to those for Ter mutant mice [4,5,16,21]. In contrast, teratomas developed in the testes of approximately 10% of the MCH Dnd1 Δ/Δ male mice, indicating that the MCH strain has low sensitivity to testicular teratoma, while teratomas were not observed in the testes of all three genotypes of the BL6 strain. Collectively, the results indicate that the Dnd1-Δ allele is inherited in accordance with Mendelian ratios, similar to the findings for the Dnd1-Ter allele, and that the ratios of testicular teratoma incidence are similar between the Δ and Ter alleles.

Dnd1 loss caused defects in both spermatogenesis and oogenesis
We also examined the effect of Dnd1 loss on germ cell development. For this purpose, we first compared testes from 4-week-old Dnd1 +/+ wild-type, Dnd1 +/Δ , and Dnd1 Δ/Δ male mice that did not have testicular teratomas. We found that even testes of Dnd1 +/Δ mice appeared to be smaller than those of wild-type male mice, in addition to a clear decrease in testis size noted in Dnd1 Δ/Δ male mice in all three strains (Fig 1A-1C). The ratios of testis weight per body weight of Dnd1 +/Δ male mice were significantly lower than those of wild-type male mice in all three strains, while the ratios of Dnd1 Δ/Δ male mice showed a drastic decrease compared with those of wild-type and Dnd1 +/Δ male mice ( Fig 1D). Histological analyses showed some seminiferous tubules with impaired spermatogenesis in Dnd1 +/Δ testes, while no germ cells were observed in Dnd1 Δ/Δ testes (Fig 1E-1M). We counted the tubules with impaired spermatogenesis in Dnd1 +/Δ testes and found significant increases in the ratio of defective tubules as compared with those of wild-type testes in all three strains, especially in the 129 strain (Fig 1N-1P).
These data indicate that Dnd1 loss results in disappearance of germ cells in adult testes and that spermatogenesis is impaired in a part of the seminiferous tubules, even in Dnd1 +/Δ heterozygous male mice.
We then compared the ovaries of 4-week-old female mice of each genotype among the three strains (Fig 2A-2C) and found that the Dnd1 +/Δ ovaries were as large as the wild-type ovaries in Table 2. Genotype of Dnd1 and the incidence of male mice affected with testicular teratomas in the BL6, MCH, and 129 strains.

BL6
% of affected males Genotype total � unilateral (L) unilateral (R) bilateral all the three strains, in contrast to the findings for testes. Only the Dnd1 Δ/Δ ovaries from the 129 female mice appeared to be slightly larger than those of the other two strains, although the ovaries from Dnd1 Δ/Δ female mice were smaller than those of wild-type and Dnd1 +/Δ female mice in all three strains. Histological analyses showed that a considerable number of oocytes were developing in the ovaries from 129 Dnd1 Δ/Δ female mice, whereas the other two strains had few oocytes in Dnd1 Δ/Δ ovaries (Fig 2D-2L). We then crossed these female mice at 6 weeks of age with MCH male mice and counted the number of offspring to examine their fertility until they reached 30 weeks of age ( Fig 2M). These analyses showed no significant difference between wild-type and Dnd1 +/Δ female mice in each strain; no offspring were born from Dnd1 Δ/Δ female mice in both BL6 and MCH strains. However, the 129 female mice delivered offspring even in the case of Dnd1 Δ/Δ mice, but the number was fewer than those of wild-type and Dnd1 +/Δ female mice, indicating that 129 Dnd1 Δ/Δ female mice were subfertile. Collectively, the data indicate that the spermatogenic and oogenic phenotypes observed in Dnd1-Δ mutant mice were very similar to those previously reported for Ter mutant mice [5].

Dnd1 +/Δ male mice progressively lost fertility because of sperm count decrease and impaired sperm function in the 129 strain
We next checked whether the Dnd1-Δ mutant mice exhibited phenotypes other than those mentioned above. We noted that 129 Dnd1 +/Δ male mice produced fewer offspring than BL6 and MCH strains in their intercrosses. For more in-depth analysis of the phenotype, we crossed 10-week-old Dnd1 +/Δ male mice of each strain with female mice of the MCH strain and determined the litter size. BL6 or MCH Dnd1 +/Δ male mice continuously impregnated the female mice and produced offspring until they were 20-24 weeks old (Fig 3A and 3B). In contrast, female mice that were crossed with 129 Dnd1 +/Δ male mice produced offspring until the male mice reached 10-12 weeks of age, but the number of offspring drastically decreased after 12 weeks of age ( Fig 3C). Subsequently, they stopped giving birth when the male mice were at 20-24 weeks of age; however, copulatory plugs were continuously found in the female mice. Since wild-type 129 male mice could impregnate the female mice even after they were 12 weeks old (Fig 3D), this phenotype was caused by the heterozygous Dnd1-Δ mutant allele only in the 129 genetic background.
To identify the cause of infertility in 129 Dnd1 +/Δ male mice, we compared the ratios of testis weight per body weight of 4-week-old wild-type and Dnd1 +/Δ male mice with those of 12-weekold male mice in each of the three strains. The ratios of Dnd1 +/Δ male mice were lower than those of wild-type male mice at both time points and in all three strains (Fig 3E-3G), consistent with the results shown in Fig 1D. Furthermore, we found that the ratio significantly decreased from 4 to 12 weeks of age in only 129 Dnd1 +/Δ male mice, whereas in the other two strains, the ratio increased even in Dnd1 +/Δ male mice, suggesting testicular growth retardation due to impaired spermatogenesis. We therefore determined the sperm count of 12-week-old wild-type and Dnd1 +/ Δ male mice in each strain and found that the count had significantly decreased in Dnd1 +/Δ male

PLOS ONE
Dnd1 regulates testicular teratoma incidence mice of all the strains. However, the sperm counts of 129 Dnd1 +/Δ male mice decreased to approximately one-fifth of those of wild-type male mice (Fig 3J), whereas the decrease was only approximately one-half in the other two strains (Fig 3H, 3I), suggesting that such a large reduction in sperm count might lead to fertility loss in 129 Dnd1 +/Δ male mice. However, since the sperm count reduction does not necessarily cause fertility loss [22], we further examined whether the sperm generated in 129 Dnd1 +/Δ male mice were functional by performing IVF with sperm from more than 12-week-old 129 wild-type or Dnd1 +/Δ male mice (Fig 3K). Nearly 90% of oocytes developed to blastocysts when sperm from wild-type male mice were used. However, in the case of sperm from Dnd1 +/Δ male mice, the developmental rate drastically decreased to 3.9%, indicating that the sperm were not fully functional in 12-week-old 129 Dnd1 +/Δ male mice. Collectively, the findings indicate that the 129 Dnd1 +/Δ male mice lost fertility, presumably because of a significant decrease in sperm count and impaired function of sperm after 12 weeks of age.
To determine why a large decrease in sperm count and impaired sperm function occurred only in 129 Dnd1 +/Δ male mice after 12 weeks of age, we speculated that the impaired spermatogenesis causing these phenotypes might be attributable to some defect in spermatogonia because Dnd1 is expressed in spermatogonia and is required for the maintenance of these cells [15]. To examine this hypothesis, cells in the testes from 12-week-old wild-type and Dnd1 +/Δ male mice of all three strains were subjected to flow cytometric analysis performed using antibodies against promyelocytic leukemia zinc-finger (PLZF) because we have previously shown that PLZF is expressed in all populations of DND1-positive spermatogonia [15]. These analyses revealed that the number of PLZF-positive spermatogonia significantly decreased in both 129 and BL6 Dnd1 +/Δ male mice (S3A- S3F Fig). However, although BL6 Dnd1 +/Δ male mice showed a slightly greater decrease in the number of PLZF-positive spermatogonia than 129 Dnd1 +/Δ male mice did (S3C and S3F Fig), the decrease in sperm count in BL6 Dnd1 +/Δ male mice was much lesser than that in 129 Dnd1 +/Δ male mice (3H and J). Furthermore, although the number of spermatogonia was unchanged between wild-type and Dnd1 +/Δ male mice in the MCH strain (S3G-S3I Fig), the sperm count significantly decreased in Dnd1 +/Δ male mice (Fig 3I). These data indicate no correlation between the decrease in spermatogonia and the decrease in sperm count in Dnd1 +/Δ male mice.
https://doi.org/10.1371/journal.pone.0232047.g003 loss of fertility observed in 129 Dnd1 +/Δ male mice has not been previously reported in 129 Dnd1 +/Ter male mice. To determine whether the Ter mutant allele leads to the same phenotype for fertility as the Dnd1-Δ mutant allele in the 129 strain, we compared the ratio of testis weight per body weight between 12-week-old wild-type and Dnd1 +/Ter male mice in the 129 strain and found a significant decrease in the ratio to 70% of that of wild-type male mice in the case of Dnd1 +/Ter male mice (Fig 4A). However, the reduction rate was lesser than that of Dnd1 +/Δ male mice since the Dnd1-Δ mutant allele decreased the ratio by almost half (Fig 3G). We therefore compared the sperm count of 12-week-old 129 Dnd1 +/Ter male mice with those of wild-type male mice and found that the number tended to decrease but was not significantly changed as compared with that of wild-type male mice (Fig 4B). We then crossed 129 Dnd1 +/ Ter male mice with female mice of the MCH strain and determined the litter size, as mentioned in Fig 3, to check whether the Ter mutant allele caused progressive loss of fertility, as is the case with 129 Dnd1 +/Δ male mice (Fig 3C). These analyses revealed that 129 Dnd1 +/Ter male mice continuously impregnated the female mice and produced offspring until they were 20-24 weeks old, unlike the 129 Dnd1 +/Δ male mice (Fig 4C). These data thus showed a clear phenotypic difference between Dnd1 +/Δ and Dnd1 +/Ter male mice in the 129 strain, suggesting that either the Dnd1-Δ or Ter mutant allele caused some gene expression change, other than Dnd1 loss, and that such changes generated these phenotypic differences.
In this context, we focused on the Ter mutant allele because it was previously suggested that Ter was not a mutation causing loss of Dnd1 expression, but instead generated a short mutant DND1 protein, called DND1 Ter , consisting of Dnd1 exons 1-2 and a part of exon 3 (S1C Fig) [16,23]. To examine DND1 Ter expression, we first tested whether antibodies against DND1, which we previously generated from a rabbit and three guinea pigs [14], could detect DND1 Ter . For this purpose, wild-type DND1 or DND1 Ter were forcibly expressed in HeLa cells by transfection of 3×Flag-tagged Dnd1 or Dnd1 Ter and then subjected to western blot using four different antibodies against DND1. Our analyses showed that the antibodies generated from guinea pigs #1, #2, and #3 could detect Flag-tagged DND1 Ter at the same level as the anti-FLAG antibody, while the antibody generated from rabbit #1 barely detected it (Fig 4D). We then analyzed the in vivo expression of DND1 Ter in the E15.5 male gonads of 129 Dnd1 +/ Ter embryos by using the antibody against DND1 generated from guinea pig #2. However, western blot analyses revealed that DND1 Ter was undetectable in all five Dnd1 +/Ter embryos (Fig 4E), whereas the DND1 levels decreased to approximately half of those in wild-type embryos (Fig 4F). These results support the possibility that Ter is a loss of Dnd1 mutation, which in turn suggests that the Dnd1-Δ mutant allele causes some defect leading to sperm count decrease and sperm malfunction independently of Dnd1 loss.

Dnd1 genetically interacted with Nanos2 and Nanos3 for suppression of testicular teratomas
We have previously shown that both NANOS2 and NANOS3 interact with DND1 and that these complexes are essential for germ cell development [14,15], leading us to speculate that both NANOS2 and NANOS3 play a vital role with DND1 even in the regulation of testicular teratoma incidence. To examine this hypothesis, we crossed Dnd1-Δ mutant mice with both Nanos2 or Nanos3 mutant mice to establish male mice that were double mutants for Dnd1 and Nanos2 or Nanos3 and subsequently analyzed these mice for testicular teratomas in the 129 genetic background.
In the case of crossing with Nanos2 mutant mice (Figs 5A and S4A), the ratio of affected male mice slightly increased from 2.9% in wild-type male mice to 8.8% in Nanos2 +/LacZ male mice and 11.8% in Nanos2 LacZ/LacZ male mice, indicating that Nanos2 is one of the genes responsible for testicular teratomas, as previously reported [24]. In addition, introducing the heterozygous Dnd1-Δ mutant allele into both Nanos2 +/LacZ and Nanos2 LacZ/LacZ male mice increased the ratio to 38.8% and 80.0%, respectively, indicating a synergistic effect of combining mutations for Dnd1 and Nanos2 since only 28.5% (Table 2) of Dnd1 +/Δ male mice were affected. In the case of crossing with Nanos3 mutant mice (Figs 5B and S4B), the ratio of affected male mice increased from 4.4% in wild-type male mice to 24.2% in Nanos3 +/Cre male mice, indicating that Nanos3 is one of the genes responsible for testicular teratomas, as previously reported [25]. In addition, the male mice that were double mutants for Dnd1 +/Δ ; Nanos3 +/Cre showed a drastic increase in the ratio to 84.3%, indicating a synergistic effect of combining heterozygous mutations for Dnd1 and Nanos3. Based on these data, we conclude that both Nanos2 and Nanos3 interact with Dnd1 in the regulation of testicular teratoma incidence.
Testicular teratomas were not observed in Nanos3 Cre/Cre male mice, even when these mice had a heterozygous Dnd1-Δ mutant allele. To determine why Nanos3 Cre/Cre male mice were not affected, we counted the number of PGCs in Nanos3-Cre mutant male embryos at E11.5 by section immunostaining of gonads with antibodies against DAZL and NANOG and then compared the numbers with those in Dnd1-Δ mutant male embryos. These analyses revealed that PGCs were almost completely absent in Nanos3 Cre/Cre embryos at E11.5 (Fig 5C), which is before the transformation to EC cells. This might be because the PGCs undergo apoptotic cell death [26] or transdifferentiate into somatic cells as observed in other species [27,28] during migration stages in the absence of Nanos3. In contrast, a small number of PGCs were still observed in the male gonads of Dnd1 Δ/Δ embryos (Fig 5D), suggesting that some of these survivor cells eventually acquire pluripotency at the late stages of embryogenesis and then develop to form teratomas after birth.

Discussion
In the current study, we newly generated Dnd1-Δ, a conventional knockout allele of Dnd1 (S1E Fig), and found that the Dnd1-Δ mutant mice showed defects similar to those of Ter mutant mice in spermatogenesis, oogenesis, and incidence of testicular teratomas, with a slight difference in spermiogenesis. In addition, a short mutant DND1 protein, called DND1 Ter , was not detected in E15.5 male gonads from Dnd1 +/Ter embryos in our western blotting analysis (Fig 4D and 4E), whereas the amounts of full-length DND1 in Dnd1 +/Ter embryos decreased to about half of those in wild-type embryos (Fig 4F). On the basis of these data, we suggest that the Ter mutation simply causes loss of Dnd1 expression, as previously reported [9]. However, we cannot rule out the possibility that DND1 Ter is highly expressed in germ cells other than at E15.5 or that DND1 Ter is translated but is undetectable because its expression level is too low. The expression level of DND1 Ter was lesser than that of wild-type DND1 when both proteins were force-expressed in cultured cells (Fig 4D), which was consistent with a previous study suggesting that DND1 Ter was unstable [23]. In either case, DND1 Ter might be involved in the development of testicular teratomas. Alternatively, given that the defects in Dnd1-Ter mutant mice were milder than those in Dnd1-Δ mutant mice (Figs 3J and 4B), it is also possible that DND1 Ter alleviates the phenotypes of Dnd1 loss, thereby leading to the phenotypic differences in sperm count between Dnd1-Δ and Ter mutant mice.
The 129 Dnd1 +/Δ male mice lost fertility after 12 weeks because of sperm count decrease and impaired sperm function, which was not observed in 129 Dnd1 +/Ter male mice (Figs 3 and  4C). If Ter is a loss of Dnd1 mutation as suggested, these phenotypic differences between Dnd1 +/Δ and Dnd1 +/Ter might be caused by sequence deletion in the Dnd1-Δ allele. Generating the Dnd1-Δ allele might collaterally remove regulatory sequence(s) located between two loxP sequences controlling the expression of gene(s) around Dnd1, which might cause changes in the expression of other gene(s), rather than loss of Dnd1. This is because Dnd1 is located in a genomic region where 16 genes are closely placed within 150 kilobase pairs, thereby resulting in sperm count decrease and impaired sperm function. However, it is unclear which gene expression level is affected by the targeted deletion and which of the cells are affected by these gene expression changes occurring with the Dnd1-Δ mutant allele because no study has reported the physiological role of the 15 neighboring genes in spermatogenesis (S5 Fig). Since no correlation was observed between the decrease in spermatogonia and the decrease in sperm count in Dnd1 +/Δ male mice (S3 Fig), the sperm count decrease and impaired sperm function could be attributed to a defect in cells other than spermatogonia.
In the context of phenotypic difference, Zechel et al. reported that, in the 129S1/SvImJ genetic background, homozygous deletion from Dnd1 exon 1 to most of exon 3 (S1B Fig) induced embryonic lethality before E3.5 [16]. In addition, they showed that the incidence of testicular teratomas did not increase in the 129S1/SvImJ Dnd1 +/KO heterozygous mutant male mice although Dnd1 +/Ter male mice showed increased incidence in the same genetic background. These phenotypes are different from those of both Dnd1-Δ and Ter mutant mice. As mentioned above, the deletion from Dnd1 exon 1 to most of exon 3 might collaterally remove regulatory sequence(s) that control the expression of genes placed close to Dnd1, leading to phenotypes different from those in Dnd1-Δ and Ter mutant mice. Alternatively, the phenotypic difference might be caused by the neomycin selector unit because it is widely known that insertion of a pgk-neo cassette or reporter gene into a specific locus might affect the expression of the neighboring genes [29,30]. It is also possible that a de novo mutation occurred during establishment of the Dnd1-KO mouse line and caused changes in the expression of the neighboring genes, resulting in the phenotypic differences. The expression profile and phenotypes of mice with knockout of the genes around Dnd1 would yield valuable information for further understanding the phenotypic differences among Dnd1-Δ, Ter, and Dnd1-KO mutant mice.
We have also shown synergistic increase in testicular teratoma incidence in the male mice that were double mutants for Dnd1 and Nanos2 or Nanos3 (Fig 5A and 5B). In addition to these genetic interactions, given that both NANOS2 and NANOS3 associate with DND1, it is highly plausible that both DND1-NANOS2 and DND1-NANOS3 complexes regulate teratoma incidence in male embryonic germ cells. We have previously shown that DND1 plays a role in loading RNAs onto the NANOS2-CNOT deadenylase complex, leading to degradation of the target RNAs [15]. In addition, in the absence of NANOS2, a part of male embryonic germ cells continuously proliferate and cannot undergo cell-cycle arrest [31], which increases the incidence of transformation of germ cells to EC cells in the 129 strain [24]. These results suggest that the DND1-NANOS2 complex suppresses target RNAs that induce cell-cycle arrest via CNOT deadenylase-mediated RNA degradation, thereby suppressing the transformation of germ cells to EC cells. NANOS3 is highly expressed in migrating and proliferative PGCs but its expression gradually disappears after colonization to male gonads, in accordance with the increase in NANOS2 expression [17,32]. Therefore, the regulatory mechanisms of the DND1-NANOS3 complex in the case of testicular teratoma incidence are still unknown and may differ from those of NANOS2. Comprehensive identification of target RNAs for DND1-NANOS2 and DND1-NANOS3 complexes is required to elucidate the transformation of germ cells to EC cells.