Recent evidence has linked human phthalate exposure to abnormal reproductive and hormonal effects. Phthalates are plasticizers that confer flexibility and transparency to plastics, but they readily contaminate the body and the environment. In this study, timed pregnant CD1 outbred mice were treated with di-(2-ethylhexyl) phthalate (DEHP) from Embryonic Day 7 (E7) to E14. The subsequent generation (F1) offspring were then bred to produce the F2, F3, and F4 offspring, without any further DEHP treatment. This exposure scheme disrupted testicular germ cell association and decreased sperm count and motility in F1 to F4 offspring. By spermatogonial transplantation techniques, the exposure scheme also disrupted spermatogonial stem cell (SSC) function of F3 offspring. The W/WV recipient testes transplanted with F3 offspring germ cells from the DEHP-treated group had a dramatically lower percentage of donor germ cell-derived spermatogenic recovery in seminiferous tubules when compared to the recipient testes transplanted with CD1 control germ cells. Further characterization showed that the major block of donor germ cell-derived spermatogenesis was before the appearance of undifferentiated spermatogonia. Interestingly, the testes transplanted with the F3 offspring germ cells from the DEHP-treated group, when regenerated, replicated testis morphology similar to that observed in the testes from the F1 to F3 offspring of the DEHP-treated group, suggesting that the germ cell disorganization phenotype originates from the stem cells of F3 offspring. In conclusion, embryonic exposure to DEHP was found to disrupt testicular germ cell organization and SSC function in a transgenerational manner.
Plastics are an important component of the world we live in today, with their production reaching 245 million tons in 2006 and increasing each year . One class of chemical component of plastics is phthalate plasticizers that confer flexibility and transparency to plastics. Phthalate esters are found in a wide range of consumer products, from food packaging and personal care products to medical devices and pharmaceuticals. Phthalate esters are not linked covalently to the plastics, so they are highly susceptible to leaching into the environment. Of the different phthalate esters, the most abundantly produced is di-(2-ethylhexyl) phthalate (DEHP), manufactured at an annual rate of 2 million tons .
Humans and animals are exposed to phthalates via multiple routes, including ingestion, inhalation, absorption by dermal contact, and the parenteral route . In the body, DEHP is metabolized into its active monoesters in the intestines and liver by esterases and further degraded to more water-soluble metabolites for excretion. Seven phthalate metabolites in urine samples were measured from 2540 individuals participating in the National Health and Nutrition Examination Survey of 1999–2000 . Mono-ethyl phthalate was in all samples, mono-butyl phthalate in 99% of the samples, and mono-ethylhexyl phthalate (MEHP) in 78% of the samples. The median phthalate exposure in adults was estimated to be 8.2 μg/kg/day and in toddlers 25.8 μg/kg/day . Women were found to have two to four times higher levels of phthalate exposure compared to men . Younger children have a significantly higher exposure than older children . Infants around 1 yr of age were found to be particularly at high risk because of their behavioral patterns of feeding and chewing and of crawling on carpets inhaling phthalate dust . Moreover, the National Toxicity Program had the most serious concern for patients in the neonatal intensive care unit, with an exposure rate as high as 6 mg/kg/day . These neonates undergo blood transfusions and extracorporeal membrane oxygenation, medical procedures that involve DEHP-containing polyvinylchloride medical devices.
Human phthalate exposure is becoming an increasing public health concern, with researchers linking phthalates to a reduction in sperm motility and chromatin damage , disruption of fetal germ cell development , disruption of Leydig cell development and testosterone levels [7, 8], and a reduction in anogenital distance (AGD) [9, 10]. The prevailing thought is that phthalates can induce testicular dysgenesis syndrome (TDS), which originates from disrupted development of fetal testes, manifesting postnatally as cryptorchidism, impaired spermatogenesis, low sperm count and damage, hypospadias, and/or testicular cancer .
Compared to rather limited information on the effects of phthalates on humans, phthalates (DEHP and di(n-butyl) phthalate) have been extensively studied in rodents. Similar to humans, phthalates are considered antiandrogens, causing TDS-like syndrome in rodents, but not by interacting with the androgen receptor [1213–14]. For example, in utero and lactational exposure [1516–17] or in vitro exposure  has resulted in abnormal Leydig cell clusters and decreased testosterone levels. Lowered testosterone levels are thought to account for reduction in AGD, incomplete preputial separation, increased areola/nipple retention, cryptorchidism, penile malformations, sperm count and sexual inactivity, abnormal testicular morphology, and inhibition of sexually dimorphic central nervous system development [12–14, 19, 20]. In contrast, for some phthalate responses, the mechanism of how phthalate acts may be different in rats from that in the mouse. For example, in utero exposure or in vitro neonatal exposure of phthalates has increased the number of multinucleated gonocytes in rats [17, 21] and mice , but this response has not been linked to lowered fetal testosterone levels in mice [22, 23], whereas associations between lowered testosterone levels and multinucleated gonocytes has been made in rats [17, 21, 24]. Additionally, the effect of phthalates on testosterone levels may vary with developmental age .
Recently, in utero exposure to compounds including vinclozolin, bisphenol A, dioxin, polycyclic aromatic hydrocarbon benzo(a)pyrene, polychlorinated biphenyls, a mixture of pesticides (permethrin and the insect repellent N,N-Diethyl-meta-toluamide), a mixture of plastics (bisphenol A and phthalates), dioxin and a hydrocarbon mixture, and high-energy diet have been found to induce reproductive disorders in a transgenerational manner in rats [252627–28] and in mice [293031–32]. Even in human populations, there is accumulating evidence of transgenerational nutritional or endocrine disruptor effects on the F2 offspring [333435–36]. For example, Dutch famine in 1944–1945 has been associated with increased F2 offspring neonatal adiposity and poor health in later life . Similarly, both grandsons and granddaughters of women exposed in utero to diethylstilbestrol between the late 1940s and early 1970s have a higher incidence of birth defects , grandsons have an increased occurrence of hypospadias , and preliminary data on the granddaughters suggest a higher incident of ovarian cancer .
We sought to determine if maternal exposure to DEHP from Embryonic Day E7 (E7) to E14 of pregnancy exerts transgenerational effects on the testis and spermatogonial stem cells (SSCs) of the offspring. The treatment window was chosen to include the time of primordial germ cell migration around E8.5, global DNA methylation pattern changes, DNA demethylation in both male and female germ cells around E10.5–E13.5 and remethylation in male germ cells around E14.5, and sex determination at E12.5 . Female remethylation occurs in a broader range of time during embryo and neonatal development . This window of treatment is thought to be critical for epigenome changes that can be transmitted to multiple generations. We found that germ cell associations in the testis were disturbed and sperm counts and motility declined in multiple generations, and, importantly, the SSC function of F3 offspring was impaired in a transgenerational manner.
MATERIALS AND METHODS
Crl:CD1(ICR) (CD1) outbred mice were obtained from Charles River Laboratories (Wilmington, MA) or an in-house vivarium. WBB6F1/J-KitW/KitW-V (W/Wv) mice were obtained from an in house vivarium or purchased from Jackson Laboratories (Bar Harbor, ME). Animal experimentation and procedures were approved by the Washington State University Institutional Animal Care and Use Committee. Animals were treated humanely in accordance with the standards of care outlined by the NIH Guide for the Care and Use of Laboratory Animals.
DEHP Treatment and Breeding to Obtain F1–F4 Offspring
Postnatal Day (P) 60–64 (P60–64) CD1 females were continuously paired with age-matched CD1 males for generating offspring. Females were examined twice daily for the presence of vaginal plugs, the presence of which was set as E0. Twenty vaginal plug-positive dams were gavaged from E7 to E14 of gestation with DEHP (500 mg/kg body weight/day) (Sigma-Aldrich Chemical Co., St. Louis, MO) in 200 μl of superrefined corn oil (Croda Inc., Edison, NJ) (DEHP-treated group) or gavaged with 200 μl of corn oil alone (vehicle control group). A review  citing the comparison of the measurement of phthalate metabolites in amniotic fluids in rodents  and humans  indicated that the dose range including a 300 mg/kg/day dose in animal studies is relevant to understanding TDS in humans and the animal dosages are not excessively high as originally thought. The treated gestating dams were considered the F0 generation (F0) dams, and the pups born from the F0 dams were labeled as the F1 generation (F1) offspring. The F1 offspring were used in three breeding paradigms to follow inheritance through the maternal, paternal, and double-cross lines. Paternal lines were generated by breeding males from the treated groups (both DEHP-treated and vehicle control) with age-matched nontreated CD1 females. Maternal lines were generated by breeding females from the treated groups with age-matched nontreated CD1 males. For the double-cross lines, males were bred to females, both from the same treated group. Each paradigm was repeated to produce F2 to F4 offspring. No inbreeding was performed.
Embryonic Survival and Weaning Rates
The number of implantation sites on uterine horns, representing the number of embryos embedded in the uterine wall, was determined from the F0 dams at the day of pup weaning using a stereomicroscope (Leica MZ 9.5; Leica, Wetzlar, Germany). The litter size was noted on P1, the body weights were measured at P10, and the number of pups successfully reared to weaning was noted on the day of weaning (P21).
Neonatal and Pubertal Period Analysis
Anogenital distance (AGD) was measured on neonatal pups at P1 by two independent researchers. The distance from the urethral opening to the ventral opening of the anus was measured in millimeters under a stereomicroscope (Leica MZ 9.5). Pups at P10 were anesthetized and nipple positions were determined by measuring the distance of each nipple to the umbilical stump. The nipples were numbered in a pattern beginning with the right foremost cranial nipple labeled nipple 1, the right axillary nipple labeled nipple 2, and continuing down to the rightmost caudal nipple labeled nipple 5. The order was reversed on the left side with the leftmost caudal nipple labeled nipple 6, the inguinal nipple labeled nipple 8, and continuing up to the left foremost cranial nipple, labeled nipple 10. Pubertal onset occurs around the time when the penis detaches from the prepuce in male pups and was determined by retracting the prepuce manually using gentle pressure beginning at P20 onward.
Testicular Tissue Collection and Determination of Testicular Abnormalities
Male mice were killed at P30 and P120 for F1–F4 generations and P360 for F4 generation and their testes were collected, fixed in Bouin solution for 1–5 h or 4% paraformaldehyde (PFA) with 0.25% glutaraldehyde for 4–6 h at 4°C, and embedded in paraffin. Four-micrometer-thick serial sections were cut and mounted on poly-l-lysine slides. To examine testicular morphology, sections were deparaffinized, stained with hematoxylin and eosin, and examined using a Leitz DMRB microscope (Leica). Digital images were obtained by using a Magnafire digital camera (Optronics, Goleta, CA). To determine the percentage of abnormal tubules present in the testes, tubules having moderate to severe germ cell disorganization and the total number of tubules were counted in testicular sections. To obtain good sampling across the entire testis, a minimum of four sections spaced 50 μm apart were counted. This was repeated for at least three animals.
Detection of Apoptosis by TUNEL Assay
Testicular sections fixed in 4% PFA with 0.25% glutaraldehyde were used for the detection of apoptotic cells by TUNEL assay (Promega Corporation, Madison, WI). This assay identifies fragmented DNA by the terminal deoxynucleotidyl transferase (TdT)-mediated incorporation of fluorescein-12-dUTP at the 3′ hydroxyl end of DNA. For a positive control, tissue sections were pretreated with deoxyribonuclease I, and for a negative control, tissue sections were not incubated with TdT. Apoptotic cells were visualized by green fluorescence using a microscope. Tubules with four or more apoptotic cells (representing apoptotic tubules) and the total number of tubules were counted to quantify the results.
Sperm Count and Determination of Sperm Motility
Sperm was collected from the caudal epididymis of F1–F4 male offspring at P120. The excised epididymides were cut in 2-mm intervals (five cuts) and placed in F12 media (1 ml) containing 0.1% bovine serum albumin prewarmed to 37°C and incubated for 15 min to facilitate sperm transmigration from the epididymis. Fresh sperm were loaded into a hemacytometer and counted. Each chamber was counted twice and averaged. A prewarmed glass slide was loaded with fresh sperm, coverslipped, and examined by phase contrast light microscopy to determine sperm motility. Sperm motility was scored in five randomly chosen fields for two samples from each animal. The number of sperm exhibiting forward and progressive activity and the total number of sperm were counted to obtain the percentage sperm motility. Sperm clumps, detached heads, and free tails were not counted.
Three F3 male offspring at P24–P31 from the DEHP-treated group in the double-cross line provided the donor germ cells for spermatogonial transplantation. Germ cells from two age-matched untreated CD1 male mice served as controls. As described , the testes were detunicated, tubules were digested with enzymes, and single-cell suspensions were obtained. The donor germ cells were counted using a hemacytometer and resuspended at a concentration of 107 cells/ml in minimal essential media α with 0.03% trypan blue dye (Invitrogen Corporation, Carlsbad, CA). Donor germ cells were transplanted into testes of at least three W/WV mice per donor animal. A small hole was made in the connective tissue enclosing the efferent ducts of the recipient mouse and a glass needle containing 7 μl (70 000 cells) of the donor germ cell suspension was inserted into the rete testis through the efferent ducts for the injection. After injection, the animals were allowed to recover for 70 days. By this time, all stages of germ cells, including meiotic and postmeiotic germ cells, were represented in the recipient tubules injected with control SSCs.
Posttransplantation Recovery Analysis
The recipient WWV testicular tissues and noninjected WWV testicular tissues were fixed in Bouin solution for 2 h and embedded in paraffin, and 4-μm-thick serial sections were cut. To determine the donor-derived germ cell recovery, the tubules containing round or elongated spermatids, more than four sections separated by 50 μm per testis from at least three animals, were stained with hematoxylin and eosin and examined using a Leitz DMRB microscope, and digital images were obtained using a Magnafire digital camera. To further determine the germ cell types present in the recovered tubules, immunohistochemistry or immunofluorescence was performed using antibodies for DDX4, γH2AX, STRA8, and DMRT1.
Immunohistochemistry and Immunofluorescence
For immunohistochemistry, after a sodium citrate antigen retrieval step, tissue sections were incubated overnight with rabbit anti-STRA8 peptide antibody (1:2500 dilution) or rabbit anti-DDX4 peptide antibody (1:300 dilution; Abcam, Cambridge, MA), washed in PBS, and incubated with biotinylated anti-rabbit IgG (1:300 dilution; Vector Laboratories, Burlingame, CA), followed by streptavidin-peroxidase reaction and incubation with 3,3′-diaminobenzidine tetrahydrochloride (Invitrogen Corporation). For immunofluorescence, the tissue sections were further permeabilized with 0.2% Triton X-100 in PBS for 45 min. The rabbit polyclonal anti-DMRT1 antibody (1:300 dilution) incubation was followed by anti-rabbit IgG conjugated to Alexafluor-555 (1:1000 dilution; Invitrogen Corporation) secondary antibody. The mouse monoclonal anti-γH2AX peptide antibody (1:750 dilution; Millipore Corporation, Billerica, MA) incubation was followed by the use of Mouse on Mouse Basic Kit (Vector). Coverslips were mounted with ProLong Gold antifade reagent with 4′,6-diamidino-2-phenylindole (Invitrogen Corporation) on sections and examined by fluorescent microscopy (Leitz DMRB microscope).
Statistical significance was determined by a pooled t-test or one-way ANOVA, followed by pairwise comparison of the means using the Student t method (JMP 9; SAS Institute Inc., Cary, NC). The P value of <0.05 was considered significant. We made a note of P values between 0.05 and <0.1 in the text.
DEHP Effect on Embryo Survival and Weaning Rates in CD1 Outbred Mice
We established the effect of a 500 mg/kg/day dose of DEHP between E7 and E14 on the embryo survival of the CD1 outbred strain. This window of treatment for CD1 mice had not been conducted previously. The number of implantation sites and litter sizes were determined for the F0, F1, and F2 dams. We found that the E7–E14 exposure of DEHP did not significantly affect the number of implantation sites in the DEHP-treated F0 dams (Fig. 1A). This was expected, as the treatment started 7 days after implantation. However, the litter size of the DEHP-treated F0 dams compared to vehicle control dams was different at P = 0.0667, showing a trend of 17.2% decrease in the litter size for DEHP-treated F0 dams (Fig. 1B). After birth, there was no significant difference at P < 0.05 in the wean rate of F1 pups between the DEHP-treated dams (n = 6) and vehicle-treated dams (n = 3). The neonatal survival of F1 offspring was 100%. For F2 and F3 embryo and neonatal survival, there were no significant differences in the number of implantation sites, the litter size of dams, and the wean rate of pups between the DEHP-treated and vehicle-treated groups for the paternal, maternal, and double-cross lines at P < 0.05 (Table 1). Thus, the embryo survival rate returned to normal for F2 and F3 embryos, indicating that the embryo survival rate is not transgenerational.
Implantation sites, litter size, and weaned rate for F1 and F2 dams.a
Pooled t-tests were conducted, and P values for the DEHP group to the control group is not significant at P < 0.05.
M, maternal line; P, paternal line; DC, double-cross line.
CO, corn oil vehicle control.
DEHP Effect on Anogenital Distance, Body Weight, Nipple Retention, and Pubertal Onset
Fetal and lactational DEHP exposure that includes late embryonic stage E16–E20 has been found to decrease male AGD and increase the incidence of males retaining nipples/areolas in rats [12–14]. To examine if mouse embryos exposed to DEHP from E7 to E14 have similar alterations, the AGD at P1 and nipple/areola retention at P10 were examined. The F1 male offspring at P1 from the DEHP-treated group exhibited a significant decrease (P < 0.0001) in AGD when compared to their vehicle control male counterpart (Fig. 1C) or nontreated control CD1 pups (data not shown). The average AGD for the vehicle control group was 1.90 mm and for the DEHP group was 1.67 mm at P1. Moreover, the body weights of F1 male offspring at P10 from the DEHP group were greater (average of 7.81 g) than those of the vehicle-treated group (average of 5.63 g) (Fig. 1D). Additionally, we detected alopecia/xerosis in 83.4% of F1 offspring in the DEHP group (n = 31), whereas no incident was observed for the control group (n = 17). However, the decrease in male AGDs, the increase in body weights, and the occurrence of alopecia/xerosis were not passed to F2 and F3 offspring from the maternal, paternal, and double-cross lines, evaluated on 21–43 male offspring for each breeding line/treatment group.
Supernumerary nipples were observed with a 16.1% frequency in F1 male pups at P10 from the DEHP-treated group (n = 31) compared to no supernumerary nipples in the vehicle-treated (n = 17) male pups. The retained nipples corresponded to nipple positions 3, 4, and 8 (as described in Materials and Methods). Examining the positions individually, 9.7% of pups had supernumerary nipples at both positions 3 and 4, and 12.9% of pups had supernumerary nipples at position 8. However, the incidence of retained nipples was not transmitted to F2 and F3 offspring males (n = 21–43) for any breeding line or treatment group.
To determine if DEHP exposure disrupted pubertal onset, the penis detachment from the prepuce was examined in males starting at P20. The average day of pubertal onset was P24.7 in the controls. The F1 offspring from DEHP-exposed dams had significantly delayed (P < 0.05) pubertal onset compared to controls (Fig. 2A). For F2 and F3 offspring, no significant difference at P < 0.05 was detected in pubertal onset between offspring of the DEHP-exposed and vehicle-exposed dams in the maternal line (Fig. 2, B and C). In contrast, the F2 and F3 offspring from the DEHP-exposed group had a significant delay (P < 0.05 to < 0.0001) in pubertal onset in the paternal and the double-cross lines compared to their vehicle counterparts (Fig. 2, B and C). These results show that DEHP decreased AGD, increased nipple retention, increased body weights, and caused alopecia/xerosis in F1 males, but these changes remained only in F1 offspring and did not transmit to F2 and F3 offspring. In contrast, male pubertal onset was delayed in F1 offspring and in F2 and F3 offspring of the paternal and double-cross lines, although the delay of pubertal onset was on average less than a day, relatively small for an outbred strain.
DEHP Increased Abnormal Seminiferous Tubule Morphology in F1–F4 Offspring
Hematoxylin and eosin staining of testicular cross sections of F1–F3 offspring at P120 in the DEHP-treated group for the maternal, paternal, and double-cross lines revealed morphological abnormalities. The weights of testes from the control and DEHP group animals at P120 in the F2 and F3 generations were not significantly different at P < 0.05 (data not shown). For the vehicle control group, the testes of F1 (Fig. 3, A, C, and E), F2 (Fig. 4, A and E), F3 (Fig. 4, I and M), and F4 (Fig. 4, Q and R) offspring at P120 showed normal germ cell associations in the seminiferous tubules, with the different stages of the seminiferous epithelium cycle represented. In contrast, testes of F1 (Fig. 3, B, D, and F), F2 (Fig. 4, B–D and F–H), and F3 (Fig. 4, J–L and N–P) offspring at P120 from the DEHP-treated group in all three breeding lines had moderate to severe testicular abnormalities. On average, 27.1% (the range in the F2 generation was 18.8–34.2%, and in the F3 generation was 22.3–41.0%) of their seminiferous tubule sections had abnormalities in the F2 and F3 generations such that they could not be assigned to specific stages of the seminiferous epithelium. Tubules with moderate and severe abnormalities were easily identifiable, with one or more characteristics such as massive cell sloughing into the center of tubules (Fig. 3D and Fig. 4, D, N, and P), no lumen (Fig. 3, D and F, and Fig. 4, D, F, G, H, K, N, and P), a complete loss of germ cell organization (Fig. 3F and Fig. 4, F, G, H, K, and N), and large vacuoles from apparent bulk germ cell loss (arrow, Fig. 4, C, J and N). Also, a small number of tubules had areas where layers of germ cells were missing (oblong circle, Fig. 4, G, J, L, and O). The germ cells missing were round spermatids (Fig. 4, j‘ and l‘), leptotene (Fig. 4o‘) and pachytene spermatocytes (Fig. 4g‘). Likewise, testes of F4 offspring at P120 from the DEHP-treated group in the double-cross line (Fig. 4, S and T) had testicular morphologies similar to those from F1–F3 offspring from all three breeding lines.
To determine if the testicular abnormalities observed at P120 were progressively worse with aging, we examined the testes from F4 offspring at P360 in the double-cross line. Similar to testes from the P120 mice, seminiferous tubules of animals in the corn oil group had distinguishable stages of the seminiferous epithelium cycle (Fig. 5, A, D, and G). In contrast, testes from mice at P360 in the DEHP group showed abnormal testicular morphologies, much more severe than those observed at P120. Abnormal tubules showed large vacuoles (arrows, Fig. 5, E, F, and I), loss of germ cell organization (Fig. 5, B and H), and presence of multinucleated cells (Fig. 5, B and F, box and insets b‘ and f‘), which are thought to be precursors to cells that may become cancerous. Some severely abnormal tubules contained only a few layers of germ cells (Fig. 5I) and some had masses of unknown cell types (Fig. 5C). Epididymides of F4 offspring from the DEHP group showed round germ cells and multinucleated germ cells (Fig. 5, K and L), which should not be present in epididymides in normal condition.
To compare the frequency of testicular abnormalities, the percentages of moderate to severe abnormal seminiferous tubules were determined for the DEHP-treated and vehicle-treated groups. The testes from F1 offspring at P120 of the DEHP-exposed group had a statistically significant increase of 5.0-fold in the percentage of abnormal tubules compared to the testes from the vehicle control group (Fig. 6A). In F2 and F3 offspring at P120, the testes from the three breeding lines of the DEHP-exposed group were found to have significant changes in the percentage of abnormal tubules compared to their vehicle counterparts (Fig. 6, B and C), although the range of fold changes (2.8- to 3.2-fold) was lower than the 5.0-fold change observed for testes from F1 offspring.
In F4 offspring at P360, we found a 6.3-fold increase in abnormal tubules from the DEHP group compared to the vehicle control (Fig. 6D), which is more than the 2.8- to 3.2-fold increases observed for F2 and F3 offspring at P120, suggesting that testicular germ cell disorganization had increased with aging. F4 offspring at P360 had 66.6% of their tubules with moderate to severe testicular abnormalities compared to 27.1% of tubules in F2 and F3 offspring at P120. These results clearly show that the DEHP exposure of F0 dams induced testicular abnormalities in offspring of multiple generations (F1–F4), and that the frequency of testicular germ cell disorganization and abnormalities was higher in F1 compared to F2 and F3 generations and the frequency and severity increased with aging.
Germ Cell Apoptosis Did Not Contribute to F3 Offspring Phenotypes at P120
To determine if germ cell apoptosis contributed to the germ cell loss and germ cell disorganization in F1–F3 offspring at P120, TUNEL analyses were performed on the testes derived from DEHP- and corn oil vehicle-treated groups for all three breeding lines. The majority of the TUNEL-positive cells in the control group were located near the basement membrane (Fig. 7A), whereas the TUNEL-positive cells in the DEHP group were located near the basement membrane and also one layer above (Fig. 7, B and C). There were also tubules containing sloughed germ cells that did not have any positive stain (Fig. 7C, arrow). The apoptotic tubules with four or more TUNEL-positive cells were counted. There were no statistically significant differences at P < 0.05 for the F1 offspring from the DEHP-treated group (Fig. 7D) or the F2 and F3 offspring from the DEHP-treated group of all three breeding lines (Fig. 7, E and F). Thus, for DEHP treatment of F0 dams between E7 and E14 in mice, the germ cell disorganization observed in the F1, F2, and F3 offspring at P120 was independent of germ cell apoptosis. However, we cannot rule out that apoptosis did not occur earlier in the neonatal developmental window. Systematic developmental studies are necessary to determine if apoptosis contributed to DEHP-affected phenotypes.
DEHP Decreased Sperm Count and Motility in F2 and F3 Offspring
To determine whether sperm count and motility changed transgenerationally, epididymal sperm from the F1, F2, F3, and F4 offspring at P120 of DEHP- and vehicle-treated groups were analyzed. There was a significant decrease in sperm count, on average 7.8% for F2, F3, and F4 offspring in all three lines (Fig. 8, B–D), except for the decrease in the sperm count for the maternal line for the F2 offspring, which was not significant at P < 0.05, but was close to significant (P = 0.0594). The F1 generation sperm counts had a decrease of 27.0%, more than 3-fold higher than those of the F2–F4 generations (Fig. 8A). This is not unexpected, because both germ cells and somatic cells of F1 embryos were directly exposed to DEHP. In terms of sperm motility, sperm from F1–F4 offspring of vehicle-treated F0 dams had a forward progressive motility at a frequency of 78.7% on average, whereas sperm from F1–F4 offspring of DEHP-treated F0 dams exhibited a forward progressive motility at a frequency of 53.0% on average, a decrease in motility compared to controls.
Stem Cell Function of F3 Offspring Is Impaired
Recent reports that phthalates decreased the proliferation of the SSC line [41, 42] led to the investigation of SSC function. Germ cells from F3 offspring (P24 to P31) of the DEHP-treated group in the double-cross line and nontreated CD1 control mice were used as SSC donors and transplanted into W/WV recipient mice, which have no endogenous germ cells undergoing meiosis. Germ cells from F3 offspring were chosen for studying the transgenerational effects because they were not directly exposed to DEHP, whereas the germ cells in F2 offspring were derived from the primordial germ cells in the F1 offspring that were directly exposed to DEHP. The weights of testes from P30 animals in the control and DEHP groups from the F3 generation were not significantly different at P < 0.05 (data not shown).
The hematoxylin and eosin-stained testis cross sections from the W/WV recipients injected with nontreated control CD1 germ cells revealed the presence of seminiferous tubules containing germ cells (Fig. 9A), a 65.1% donor germ cell-derived recovery (Fig. 9J). In contrast, the testis cross sections of W/WV recipients injected with germ cells from F3 offspring of the DEHP-treated group displayed a dramatic decrease in tubules containing donor-derived germ cells (Fig. 9B, arrows), a 4.7% recovery (7.3% of the control) (Fig. 9J). Anti-DDX4 antibody immunostaining (examples shown in Fig. 9, D–F) and anti-γH2AX antibody immunostaining (examples shown in Fig. 9, G–I) and counting (Fig. 9, K and L) confirmed that there are fewer tubules with meiotic and postmeiotic germ cells. DDX4 is found in spermatogonia and spermatocytes, as well as in spermatids, staining more than two layers of germ cells ; γH2AX is localized to preleptotene, leptotene, zygotene, and early pachytene spermatocytes in the testis . As an endogenous negative control, noninjected W/WV testes were used. Noninjected W/WV testes did not contain any donor-derived germ cells (Fig. 9C) and anti-γH2AX (Fig. 9I) immunostains did not reveal any immune-positive cells.
To further examine if the recovered tubules had a complete spermatogenesis, the recovered tubules were counted for the presence of round or elongated spermatids. In the W/WV recipients injected with control CD1 germ cells, 53.3% and 55.6% of the tubules contained round and elongated spermatids, respectively (Fig. 9, M and N). In contrast, there was a dramatic decrease in round and elongated spermatids in the W/WV recipients injected with F3 offspring germ cells from the DEHP-treated group, with only 4.5% (8.4% of the control) of the tubules containing round spermatids and 1.8% (3.2% of the control) of the tubules containing elongated spermatids (Fig. 9, M and N).
Stem Cells from F3 Offspring of the DEHP-Treated Group Have Decreased Colonization
To determine if there was any difference in the number of undifferentiated spermatogonia in the W/WV recipient testes, anti-DMRT1 antibody was used. Anti-DMRT1 strongly immunostains undifferentiated spermatogonia (white arrows in Fig. 10, A, B, and C) and lightly stains differentiated spermatogonia and Sertoli cells . Tubules containing the brightest-staining undifferentiated spermatogonia were counted in the W/WV recipients injected with control germ cells, in F3 germ cells from the DEHP group, and in the noninjected W/WV testes. The W/WV control recipients injected with untreated CD1 donor cells had 58.6% of tubules with undifferentiated spermatogonia (Fig. 10I), whereas the W/WV recipients injected with F3 DEHP donor cells had 12.2% of tubules with undifferentiated spermatogonia (20.8% of the control). Thus, these results indicate that a large decrease (∼79%) in stem cell colonization compared to the controls.
The recipient testes were further examined for tubules that contained differentiating spermatogonia and early meiotic spermatocytes. The antibody against stimulated by retinoic acid gene 8 (STRA8) was chosen to examine cell types (represented by darkly staining brown nuclei), differentiating spermatogonia (A1–A4, Int, and B) and early meiotic spermatocytes (preleptotene and leptotene) . The W/WV control recipients injected with untreated CD1 donor cells had 46.3% of tubules with differentiated spermatogonia and early meiotic spermatocytes expressing STRA8, whereas 8.9% of the tubules (19.5% of the control) in the recipient testes injected with F3 germ cells from the DEHP-treated group had STRA8-positive cells (Fig. 10J). Together, these results indicate that there was no further decrease in donor-derived spermatogenic cells after the initial large decrease in stem cell colonization and proliferation observed with the DMRT1 immunostaining. However, there was a second large drop in donor-derived spermatogenic cells from the beginning of meiosis (19.5% STRA8-positive cells) to the round spermatid stage (Fig. 9, 8.4% of control tubules with round spermatids).
Germ Cell Disorganization Is an Inherent Property of F3 Stem Cells from DEHP-Treated Group
Interestingly, when the tubules from the W/WV recipients transplanted with the F3 offspring germ cells were examined, they had germ cell disorganization (Fig. 11, B, C, E, F, H, and I) very similar to the germ cell disorganization observed in testes from the F3 and F4 offspring of the DEHP-treated group at P120 and P360 (Figs. 4 and 5). The tubules appeared to lack layers of germ cells, were disorganized, and had areas where vacuoles were observed (Fig. 11B) or contained multinucleated cells (Fig. 11, C, F, and I, arrows). On the other hand, the tubules from the recipients injected with untreated CD1 germ cells had normal morphology with readily observable normal stages of seminiferous epithelium (Fig. 11, A, D, and G).
DEHP is a known testicular toxicant and its effects on the rodent testis of F1 offspring after direct embryonic and lactational exposure have been widely documented [12–14, 17, 19, 20, 24]. We undertook this study to investigate the reproductive effects of DEHP exposure to F0 dams on the testis of progeny over multiple generations (F1–F4). We are reporting for the first time that the disruption of testicular function is not limited to the F1 offspring, but that abnormal testicular function persists in F2–F4 offspring. We found that DEHP exposure to F0 dams of CD1 outbred mouse strain at the E7–E14 window of gestation induced disruption of SSC colonization and proliferation and germ cell associations/organization in multiple generations, and, interestingly, the germ cell disorganization phenotype was an inherent property of SSCs of F3 offspring.
Specifically, we identified in the testes from the offspring at P120 of the DEHP-exposed dams increased testicular germ cell disorganization and a loss of identifiable stages of the seminiferous epithelial cycle (germ cell disorganization) in the F1–F4 generations. These acquired phenotypic traits were inherited through the maternal, paternal, and double-cross lines. The abnormal tubule morphology at P120 was observed on average at a frequency of 27.1% of total tubules in F2 and F3 offspring, and consisted of tubules with no lumen, tubules with a massive sloughing of germ cells blocking the lumen, tubules with missing germ cell types or layers of germ cells, and tubules with a complete loss of germ cell organization. It is not insignificant that these disorganized tubules were observed in F1–F4 offspring, because highly organized architecture of germ cell associations in the seminiferous epithelium is required for efficient and quantitative spermatogenesis, providing continuous production of sperm throughout the lifetime .
All germ cell types were present in the disorganized tubules and only a 7.8% decrease in sperm counts was detected for F2–F4 offspring at P120 in the paternal and double-cross lines. In the F1 generation, the decrease of 27.0% in sperm number was more than 3-fold higher than those for the F2–F4 generations. This higher percentage decrease in the F1 generation is not surprising, because germ cells and somatic cells in the F1 generation are directly exposed to DEHP. These sperm results are consistent with higher tubule abnormalities observed in the F1 generation (5.0-fold), comparing the control and DEHP-exposed group, and with tubule abnormalities (2.8- to 3.2-fold) persisting in the F2 and F3 generations.
Moreover, the testicular abnormalities were found to be progressively worse with aging (120–360 days), with an increased frequency (66.6% vs. 27.1%) and severity. The seminiferous tubule sections from F4 offspring at P360 had abnormal germ cell arrangements as seen in tubules from F4 offspring at P120, but also tubules with the germ cell layers missing and tubules with multinucleated cells. Although the length of treatment and treated rodent species were different, the abnormal testicular morphologies at P360 were similar to morphologies reported in previous studies, in which rat embryos were treated for a longer time, in utero and by lactation (for example, E8–P17 or E6–P21) [14, 19, 20]. However, our studies were unique in that the reproductive life span of offspring from the DEHP group seemed to decrease compared to that of the controls in a transgenerational manner. An age-related progressive increase in the incidence of the abnormal testicular phenotypes could be due to an age-related decline in Sertoli cell function in the testis from F3 offspring. Previously, it was shown that Sertoli cells were a target of phthalates , and a recent review indicated that the stem cell niche function dictated by testicular somatic cells declines with aging .
We hypothesize that germ cell disorganization phenotypes may result from the loss of molecular cues for cell-cell interactions important for setting up the germ cell associations of the seminiferous epithelium cycle. Consistent with this, it was reported previously that MEHP disrupted junctional complexes between Sertoli cells, and between Sertoli cells and germ cells, in C57BL/6 mice at P28 . This paper also showed that MEHP-mediated germ cell detachment from Sertoli cells was not dependent on germ cell apoptosis, similar to our results. Further systematic investigation is required to determine the molecular cues altered by in utero DEHP treatment that can control the proper cell-cell interactions in the testis of F3 offspring.
Recently, albeit in a cell line, it was shown that MEHP decreased the SSC proliferation in the C18-4 SSC-derived cell line by disrupting glial cell-derived neurotrophic factor (GDNF) signaling [41, 42]. Using the spermatogonial transplantation assay, we found a dramatic decrease (92.7%) in the tubules containing donor-derived meiotic/postmeiotic germ cells in the recipients injected with F3 germ cells from the DEHP-exposed group compared to the recipients injected with control germ cells. Moreover, using the markers of undifferentiated and differentiated spermatogonia, DMRT1 and STRA8, respectively, we detected that the first major block in the donor-derived spermatogenesis (79% decrease) was before the appearance of undifferentiated spermatogonia, perhaps in SSC colonization and initial proliferation of undifferentiated spermatogonia. In normal neonatal testis, the migration of gonocytes to the basement membrane to become stem cells may be affected by phthalates. Supportive of this, gonocytes at P3 in rats were shown to be sensitive to MEHP in organ cultures [21, 51]. Also supportive is a recent report that MEHP decreased the expression of β1-integrin expression in the testis , and the loss of function of β1-integrin signaling in germ cells and Sertoli cells has been linked to a loss of SSC homing to the stem cell niche in the testis . However, we cannot rule out that the number of stem cells in the germ cell population that we transplanted was also less in F3 offspring in the DEHP group compared to the control. Further analysis is needed to sort out different possibilities. Nonetheless, these results are important in that in utero DEHP exposure in F0 dams affected the germ line at the level of the SSC function transgenerationally. Our analyses further indicated that there was a second block going through meiosis, from 19.5% of tubules containing differentiated spermatogonia and early meiotic spermatocytes to 8.4% of tubules containing round spermatids, which also should be followed in future analysis.
Moreover, it was remarkable that the recovered tubules from F3 germ cells of the DEHP group after transplantation had testicular morphologies replicating what was observed in the abnormal tubules at P120 and P360. These results showed that transgenerationally acquired testicular phenotypes observed in F3 offspring after the embryonic exposure to DEHP of the F0 dams were derived from the changed property of SSCs of F3 offspring. These findings are similar to observations made after transplantation of GDNF-  and WNT4-overexpressing  SSCs, where the phenotypes observed in the original donor testes was recapitulated in the transplanted testes. In the latter case, similar to our experiments, they also showed that the progressive germ cell disorganization phenotype observed in the donor knockout testis was due to the loss of SSC function.
Our results showed that in utero exposure of DEHP to F0 dams reduced male AGDs, increased male nipple retention, and delayed pubertal onset in the F1 generation. This is similar to the results obtained in rats after embryonic and lactational exposure to DEHP [12–14, 19, 20]. Because fetal testicular testosterone synthesis is initiated at E12.5 in mice , DEHP could have changed fetal induction of steroidogenesis and changed AGD and nipple retention in F1 offspring. Neither decreased AGD nor male nipple retention traits observed in F1 offspring were transmitted to the F2 and F3 offspring. Interestingly, similar results were observed for in utero treatment of rats with antiandrogen flutamide . Flutamide induced testicular degeneration and decreased sperm counts in F1 offspring, but the effects were not transmitted to F2 and F3 offspring, which led the investigators to conclude that an antiandrogenic mechanism was not transgenerational.
The exact molecular mechanism by which DEHP is able to disrupt the testis function transgenerationally remains unknown. DNA methylation changes that can be transmitted to the next generation as an epigenetic mechanism are of particular interest to us because the window of treatment (E7–E14) of this study included the time of global DNA demethylation for both male and female germ cells and remethylation of male germ cells . DEHP could interfere with global demethylation and this could leave some DNA with methyl groups still attached. DEHP has been found to increase the methylation status of testicular genes and expression of DNA methyltransferases after E12.5 to E19 exposure . Additionally, multiple lines of evidence strongly suggest that germ line epigenetic changes are inherited to future progeny generations in reproductive disorders, neuronal diseases, and cancer [27, 30, 5960–61].
In conclusion, we report that DEHP-induced testicular germ cell disorganization and impaired SSC function passed from generation to generation and that the germ cell disorganization phenotypes were inherent to SSCs. Further work is needed to determine exactly what genes are methylated after embryonic exposure to DEHP, but not in controls, and how germ line epigenetic changes may be responsible for SSC colonization and establishment and maintenance of germ cell associations in future generation. The transgenerational effect of DEHP and other phthalates on humans is completely unknown, but evidence is mounting that links the levels of DEHP and its metabolites to reproductive disorders in those directly exposed and their offspring [5, 9, 10]. From this perspective, continued molecular work with DEHP and other endocrine disruptors is needed to shed new light on the potential risks of these compounds to disrupt reproductive outcomes over multiple generations.
The authors thank Dr. Jannette Dufour (Texas Tech University) and Dr. Qien Yang (Washington State University) for their valuable input and critical reading of our manuscript. We thank Kelsey Brooks for her technical assistance. We also thank Dr. Michael Griswold (Washington State University) for the gift of the anti-STRA8 antibody and Dr. David Zarkower (University of Minnesota) for the gift of the anti-DMRT1 antibody.