Endocrine Disrupting Chemicals and Reproductive Health in Boys and Men

Male reproductive health has declined as indicated by increasing rates of cryptorchidism, i.e., undescended testis, poor semen quality, low serum testosterone level, and testicular cancer. Exposure to endocrine disrupting chemicals (EDCs) has been proposed to have a role in this finding. In utero exposure to antiandrogenic EDCs, particularly at a sensitive period of fetal testicular development, the so-called ‘masculinization programming window (MPW)’, can disturb testicular development and function. Low androgen effect during the MPW can cause both short- and long-term reproductive disorders. A concurrent exposure to EDCs may also affect testicular function or damage testicular cells. Evidence from animal studies supports the role of endocrine disrupting chemicals in development of male reproductive disorders. However, evidence from epidemiological studies is relatively mixed. In this article, we review the current literature that evaluated relationship between prenatal EDC exposures and anogenital distance, cryptorchidism, and congenital penile abnormality called hypospadias. We review also studies on the association between early life and postnatal EDC exposure and semen quality, hypothalamic-pituitary-gonadal axis hormone levels and testicular cancer.


INTRODUCTION
Reports on deteriorating male reproductive health have been published in many countries. Serum testosterone levels and semen quality have been declining (1)(2)(3). In addition, the rates of congenital cryptorchidism, i.e. undescended testis, and testicular germ cell tumors have been increasing (4,5). Exposure to endocrine disrupting chemicals (EDCs) has been proposed to be one of the causes of these adverse trends. This is because these chemicals are ubiquitous, we are exposed to them via food, skin and inhaled air. Environmental EDCs include for instance pesticides, chemicals used in plastic products [like phthalates and bisphenol A (BPA)], in personal care products (like triclosan and parabens), in hydraulic and electronic devices [like polychlorinated biphenyls (PCBs)], chemicals used in clothes (like perfluorinated compounds), flame retardants, solvents, chemicals produced unintentionally as side products in chemical processes (dioxins) and many others (6).
Many experimental and epidemiological studies have supported links between EDC exposures and male reproductive health problems (7). Since development of male reproductive system requires androgens, substances that have antiandrogenic effects can disturb this process and possibly cause male reproductive disorders ( Figure 1). Anti-androgenic chemicals with different mechanisms of actions (inhibition of androgen biosynthesis or receptor antagonism) show accumulative effects (8). This causes the risk that even low concentrations in mixtures can be harmful. Furthermore, non-monotonic dose-response to EDCs has been described (9,10).
It has been proposed that the disruption of fetal testicular development due to, for example, maternal exposure to EDCs, can result in disorders manifested at birth, i.e., congenital cryptorchidism, congenital penile abnormality called hypospadias and reduced anogenital distance (AGD), as well as disorders presented later in life, including poor semen quality, testicular germ cell tumors, and altered reproductive hormone levels. This is the concept of testicular dysgenesis syndrome, TDS (11,12). In addition, some studies have shown associations between postnatal EDC exposures and male reproductive disorders.
We will review the human epidemiological studies that investigated the association between pre-and postnatal EDC exposure (based on environmental chemical concentration measurements in different matrices) and above mentioned male reproductive health indicators (anogenital distance, cryptorchidism, hypospadias, semen quality, reproductive hormone levels in adults and testicular cancer) and were published in English by August 2020 and found in Pubmed. Heavy metals and pharmaceuticals are not included in this review, because medicines have been recently reviewed elsewhere (13) and because the effects of heavy metals are mostly toxic rather than endocrine modulating (14,15). However, we include organotins, because their action is clearly hormonal.

Pesticides
Dichlordiphenyldichloroethylene (p,p'-DDE) is the most persistent congener of DDT. The effects of DDE and DDT may persist even though they were banned in 1970s-1980s (26,27). DDT and p,p'-DDE can accumulate in body fat for many years (half-life of approximately 6 years for DDT and 10 years for p,p'-DDE) (6,27,28). Persistent chemicals include also other organochlorine pesticides, for example lindane, chlordane and heptachlor (25).

PCBs and Dioxins
PCBs were widely used in industrial and consumer products. Even though their use was banned in the 1970s, they still persist in the environment and people continue to be exposed (29). They accumulate in body fat and have a half-life of 1 to 10 years. Humans are exposed to PCBs through ingestion of contaminated food, inhalation or skin contact (29). As mentioned above, dioxins are not produced intentionally, but they are formed as side products and humans are exposed to these persistent chemicals mainly via food of animal origin (30).

Flame Retardants
PBDEs are used as flame retardants and are found in house dust. The major routes of exposure are dust inhalation or ingestion (31,32). They can exert anti-androgenic and estrogenic activity, which potentially leads to male reproductive disorders (16,33). Also polybrominated biphenyls (PBBs) have been used as flame retardants (6).

Perfluorinated Compounds
Perfluorinated compounds (PFCs) are used in industry and consumer products, including surfactants, paints, lubricants and impregnation of clothes, textiles, footwear, furniture and carpets (34). Perfluorooctane sulfonate (PFOS) is the most abundant perfluoroalkyl substances (PFAS) in humans and in environment, followed by perfluorooctanoic acid (PFOA) (35,36). PFOA was used in the production of polytetrafluoroethylene, which is used in non-stick coating cookware (37). Human exposure occurs via inhalation, ingestion and skin contact (38).

Organotins
Organotins have been used widely in industry, e.g., in antifouling paints of boats and ships and they have been observed to have endocrine-disrupting properties and adverse effect on male reproductive health (6,39). Humans are exposed to them via contaminated seafood.

BPA and Other Phenols
Bisphenol A is used in the lining of water supply pipes, aluminum cans, reusable plastic food containers, dental sealants, thermal receipts, medical equipment, and building supplies (41). Humans can be exposed to BPA via ingestion, inhalation or skin contact (42). It can act as a weak agonist of the estrogen receptor by binding to estrogen receptors (ER) ERa and Erb (43,44). It can also act as an androgen receptor antagonist (45,46). It can cause reduced serum follicle-stimulating hormone (FSH), luteinizing hormone (LH) and testosterone levels (47). It can interfere 17a-hydroxylase/17,20 lyase and aromatase, which are important steroidogenic enzymes of Leydig cells (46). In addition, it can cross the placenta from mothers to the fetus, but its concentration in fetal circulation is much lower than in mother and thus, the placenta appears to reduce BPA exposure of the fetus (48). BPA is metabolized in the liver and excreted in urine with plasma half-life of six hours (46). Therefore, the standard method of BPA measurement is analysis of urinary levels (42,49). Bisphenol S was used as a potentially safer substitute for BPA. However, a limited number of studies showed that it also has estrogenic, androgenic, and antiandrogenic effects (50), and therefore it might have adverse reproductive effects in humans. Triclosan is an antimicrobial agent used for instance in personal care products and it is also a phenol (6).

Phthalates
Phthalates are ubiquitous chemicals, which are widely used as plasticizers, a component of polyvinyl chloride (PVC), excipients in some medications, personal care products, solvents or adhesives (51). Humans are exposed to phthalates via ingestion, which is the main route of exposure, inhalation, intravenous administration and through direct skin contact (51). After entering the human body, phthalates are rapidly metabolized into monoesters, which are excreted into urine with a half-life of 12 hours (52,53). Therefore, phthalate measurement from urine results in a higher level than from other biological samples, and urine is the most frequently used sample in epidemiological studies (54).

Parabens
Parabens belong to a group of esters of p-hydroxybenzoic acid. They have antibacterial and antifungal properties, therefore they are used as preservatives in personal care products, cosmetics, foodstuffs and some pharmaceuticals (55)(56)(57). They show weak estrogenic effect in vitro (57). Parabens belong to nonhalogenated phenols (6).

Non-Persistent Pesticides
Non-persistent pesticides include for instance organophosphates, pyrethroids, and carbamates. Some of these chemicals have been shown to have endocrine disrupting effects and may cause male reproductive disorders (58,59).

Anogenital Distance
Anogenital distance has been measured either as anoscrotal distance, i.e., the distance between anus and perineoscrotal junction, or as an anopenile distance, i.e., the distance between anus and cephalad insertion of the penis. Sometimes also the distance from the centre of the anus to the posterior base of the penis was recorded (62). Anogenital distance is considered to be a life-long marker of androgen exposure in the prenatal male programming window (MPW), at least in rats (63,64). In humans, MPW is presumed to be in gestational weeks (GW) 8-14 (63). Prenatal exposure to antiandrogenic EDCs has been associated with short AGD in male rats [reviewed in (65)]. Several human studies have evaluated associations between prenatal EDC exposure and anogenital distance in infant and young boys ( Table 1).
In Table 1, three out of four studies suggested a negative association between PCB or PBDE exposure levels and anogenital distance. BPA/phenol levels were negatively associated with anogenital distance in less than half of the listed studies. Negative associations between pesticide exposure levels (different chemicals included) and anogenital distance in the boys were reported in less than half of the studies. For some chemical groups, only a few human studies have been published so far and it is difficult to draw any conclusions. Differences in results of the studies may be explained by variation in exposure levels, in timing of the sample collection, in matrices analyzed, in the age of the boys at examination, in other factors included in the statistical analysis (e.g., stress), and in metabolites/chemicals analyzed. It also has recently been suggested that human-rodent differences in results concerning associations between prenatal EDC exposure and anogenital distance could be due to species differences in regulation of fetal androgen production (98).

Cryptorchidism
Congenital cryptorchidism, i.e. undescended testis, is one of the most common congenital malformations in newborn boys and prevalences between 1 and 8 percentage have been reported in full term boys in prospective cohort studies (4). Testicular descent from the intra-abdominal position into the scrotum is usually completed by 35th GW [reviewed in (99)]. Proper androgen action is important especially for the last phase of testicular descent, the inguinoscrotal phase (100). Furthermore, the first phase of testicular descent is, at least in mice, dependent on insulin-like peptide 3 (INSL3), a hormone produced by Leydig cells, and estrogens have been shown to downregulate the expression of INSL3 gene (99,101). Therefore, fetal exposure to environmental chemicals with anti-androgenic and estrogenic properties might be associated with cryptorchidism in boys.
For pesticides, several studies have been published, and nine out of 14 studies listed in Table 2 suggested no significant association with the risk of cryptorchidism. All but two studies (one for each group) in Table 2 found no significant association between PCB or phthalate exposure levels and the risk of cryptorchidism. Two out of five studies suggested that PBDE exposure levels are positively associated with the risk of cryptorchidism. For phenols, two out of five studies suggested positive association between BPA exposure levels and the risk of cryptorchidism. For dioxins, perfluorinated compounds, parabens, organotins and solvents, only a few studies have been published so far and it is difficult to draw any conclusions. In a study evaluating simultaneously the risk of cryptorchidism and levels of several congeners of different chemical groups, levels of four PBDEs and octachlorodibenzofuran (OCDF) were significantly higher in the group representing Danish cryptorchid boys when compared with controls (131).
Bonde et al. studied associations between in utero or infant exposure to environmental EDCs and cryptorchidism in a metaanalysis (132). The analysis included studies based on chemical measurements of different biological matrices. No significant association was observed between exposure to environmental EDCs and cryptorchidism, when including eight studies in the analysis (132).

Association Between EDC Exposure and Hormone Levels in Early Life
Some of the above mentioned studies on cryptorchidism or anogenital distance have suggested association between EDC exposure levels and reproductive hormone levels of boys in amniotic fluid, cord blood or in serum samples taken at 3 months of age (79,83,103,106,107,110,114,115,126,128,129). In Danish case-control studies on cryptorchidism, amniotic fluid DEHP and DiNP metabolite and PFOS levels associated positively with amniotic fluid testosterone (T) levels and negatively with amniotic fluid Insulin-like peptide 3 (INSL3) levels (114,115,128). In French case-control studies on cryptorchidism, cord blood levels of BPA correlated negatively with cord blood INSL3 levels (110) and unconjugated BPA levels correlated positively with cord blood T and inhibin B levels (126). Maternal breast milk levels of PCB153, DDE or mBP did not correlate with cord blood INSL3 or T levels   (110). In the Chinese study on anogenital distance, maternal urine BPA levels showed negative associations with boys' cord blood T levels and T/estradiol (E2) -ratio (83). In the Turkish study on anogenital distance in boys, cord blood levels of BPA, phthalates and reproductive hormones were studied (79). BPA levels were positively associated with E2 levels in cord blood, but no other significant associations between chemical and reproductive hormone levels were observed (79). The Danish-Finnish cryptorchidism study evaluated associations between EDC levels in breast milk (106,129) or in placenta (103,106,107) and boy's reproductive hormone levels at 3 months of age. Breast milk phthalate metabolite levels   showed positive associations with boys' Sex hormone-binding globulin (SHBG) levels, LH levels, LH/Free T -ratio, and negative association with boys' Free T levels (129). Breast milk PBDE levels also showed significant positive association with boys' LH levels (106). No other significant associations between PBDE levels in breast milk or placenta and boys' reproductive hormone levels at 3 months were observed (106). Placenta PCB WHO-TEq levels also showed significant positive association with boys' LH levels (only in the Finnish subjects) (103), but no significant association between placenta polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/F) WHO-TEq levels and boys' reproductive hormone levels was observed. Associations between placenta organotin levels and boys' reproductive hormone levels differed between countries; they showed negative associations with LH levels and FSH/Inhibin B -ratio, and positive associations with inhibin B levels in the Finnish data, but in the Danish data, organotin levels in placenta showed negative associations with T levels and T/E2 -ratio (107). These results suggest that EDC exposures may affect except fetal but also postnatal testicular function in boys.

Hypospadias
In hypospadias, penile development is disturbed so that the opening of urethra is situated on the ventral side of the penis, or in the scrotum or perineum (133). Hypospadias is due to failed fusion of penile urethra folds during embryonic weeks eleven to sixteen (134,135). Penile development is dependent on androgens (134). Both genes and environment are thought to have a role in the etiology of hypospadias (136). Four out of eight studies listed in Table 3 have suggested a positive association between pesticide levels and risk of hypospadias. For PCBs and phthalates, none of the few studies suggested significant positive association with risk of hypospadias. Only a few studies have evaluated so far the association between exposure to PBDEs, perfluorinated compounds, and solvents and conclusions are difficult to draw.
In the meta-analysis by Bonde et al, also associations between exposure to environmental EDCs and hypospadias was studied (132). Based on 18 risk estimates no significant association was found (132). No significant link was either found when studying association of hypospadias with specific exposures to DDE (degradation product of pesticide DDT) and PCBs (132).
Some studies evaluated cryptorchid and hypospadias cases in combination. In a Spanish study Arrebola et al. included 29 cases (16 with cryptorchidism, 12 with hypospadias, and one with both disorders) and 60 healthy controls (144). They assessed antiandrogenic activity of placenta samples using total effective xenobiotic burden of anti-androgens (TEXB-AA) as a biomarker, combined with a bioassay-directed fractionation protocol. They found a significant positive association between TEXB-AA levels in fraction 2 and occurrence of genital malformations (144). Another study from Spain compared placenta levels of 16 organochlorine pesticides and total effective xenoestrogen burden between a group of boys with cryptorchidism or hypospadias (n=36) and a group of matched control boys (n=109) (145). Cases had more often measurable level of estrogenicity due to xenoestrogens (TEXBalpha fraction) in their placenta (145). In addition, presence of five pesticides (o,p′-DDT, p,p′-DDT, endosulfan-a, lindane, and mirex) in placenta were associated with an increased risk of birth defects (cryptorchidism or hypospadias) (145). In another study, Fernandez et al. compared placenta levels of BPA, 6 benzophenones and 4 parabens in boys with genital malformations (cryptorchidism or hypospadias, n=28) to those of control boys (n=51) (146). The third tertiles of BPA and propylparaben (PP) levels were associated with significantly increased risk of urogenital malformations, but cryptorchidism and hypospadias were not analyzed separately (146).
Also a study from the USA evaluated cryptorchid and hypospadias cases in combination. Maternal first trimester urinary phthalate metabolite (n=6) levels were not significantly associated with the risk of cryptorchidism or hypospadias (n=5 and n=3, respectively, analyzed together, and n of controls = 334) (147). In a study from Turkey, cord blood BPA, DEHP and MEHP levels were not statistically different in patients (14 out of 100 boys) with either hypospadias, cryptorchidism or retractile testis compared to control boys (79). Another study from the USA evaluated association between in utero exposure to polybrominated biphenyls (PBBs) and cryptorchidism and hypospadias separately and combined (n of all boys = 393) (105). No association was observed in the analyses (105). In the above mentioned metanalysis by Radke et al., the evidence for association between phthalate exposure and cryptorchidism or hypospadias was slight or indeterminate (51).
Small studies on risk factors of hypospadias or cryptorchidism may have limited power to find statistically significant differences. Especially hypospadias is less frequent malformation and therefore inclusion of cases may be challenging. However, some of the studies that remained negative included almost two hundred cases and thus, limited number of cases seems unlikely reason for their negative result. Differences in severity of cases, in exposure levels, in timing of the sample collection, in matrices and statistical analyses may also explain differences in results of different studies.

Semen Quality
Epidemiological studies have reported a global decline in semen quality, particularly in countries of Western origin. In 1992, Carlsen et al. reported a considerable global decline of mean sperm concentration from 113 million/mL in 1938 to 66 million/mL in 1991 (148). This finding is confirmed by later meta-analyses, including a systematic review and meta-regression analysis by Levine et al. in 2017, which showed a decline in sperm concentration globally at a rate of 0.70 million/mL/year from 1973 to 2011 (1). The decrease in sperm concentration and total sperm count was significant only among men in North America, Europe, Australia and New Zealand, which have a population of the European descent, but not in other regions (1). The cause of deteriorating semen quality is still unclear; however, some research supports the role of EDC exposure. Here, we review epidemiological studies that investigated the association between EDC exposure and results from standard semen analysis. We include only studies that reported chemical measurements in biological matrices.  (149). A total of 284 men had maternal serum measured for BPA levels. Serum samples were collected at 18 th and 34 th week of gestation and pooled for the statistical analysis. Maternal serum BPA levels were positively associated with sperm concentration and progressive sperm motility, but not with other semen quality parameters, after adjustment for maternal smoking, duration of sexual abstinence and the presence of varicocele (149). This result did not support the link between BPA and poor semen quality. However, the BPA level was measured in the serum, and not in the urine, which is the standard method of assessment. In addition, BPA levels at the adult age were not measured, and therefore the BPA exposure in adulthood was not determined.

Polychlorinated Biphenyls, dioxins
Some studies have reported an association between dioxins and PCBs and low semen quality (150,151). Guo et al. reported that men born to mothers who had been exposed to PCBs and/or polychlorinated dibenzofurans (PCDFs) during pregnancy (n=12) had higher percentage of sperm with abnormal morphology and lower percentage of total or progressive motility as compared with men who were born to non-exposed women (n=23) (151). In an Italian study, 21 men who had prenatal exposure to dioxins due to a factory accident in Seveso in 1976, were observed to have lower sperm concentration, total sperm count, percentage of progressive sperm motility, and total motile sperm count than the 36 controls (150). This finding supports a link between prenatal exposure to PCBs and/or PCDFs and poor semen quality. In contrast, a study of 176 young men from a pregnancy cohort in Denmark showed that maternal serum ∑PCB and ∑DL-PCB levels collected at 30 th week of pregnancy were not correlated with semen quality of the sons (152).

Phthalates
Hart et al. studied association between prenatal phthalate exposure and reproductive health in adulthood in the above-mentioned Raine study (153). The study showed that pooled maternal serum levels of monoisononyl phthalate (MiNP), sums of DEHP and DiNP metabolites and the sum of high molecular weight phthalates collected at 18 and 36 weeks of pregnancy were negatively associated with testicular volume of the sons in adulthood. Maternal serum MEP levels were negatively associated with semen volume and mono-carboxy-iso-octyl phthalate (MCiOP) levels were negatively associated with progressive sperm motility (153). Axelsson et al. analyzed association between maternal serum levels of DEHPand DiNP metabolites during pregnancy and semen quality of the 112 sons (154). They reported that men who had MEHHP and MCiOP exposure levels in the highest tertile had lower semen volume than those of men in the lowest exposure tertile (154). The results of these studies suggested a potential role of prenatal exposure to phthalates in determination of semen quality. The mechanism of the association between phthalate exposure and poor semen quality in men is unclear. Studies in animals, such as rodents, demonstrated that prenatal phthalate exposure, particularly during masculinization programming window, can disrupt fetal testis development and cause a reduced androgen production. This effect can result in a variety of male reproductive disorders postnatally (63,(155)(156)(157)(158)(159). Fetal testis xenograft into castrate male nude mice showed that serum testosterone did not differ between vehicle and DBP-exposed hosts (52). This finding suggested that human fetal testes exposure to DBP did not impair fetal testicular testosterone production as shown in animal studies (52). However, an increased amount of multinucleated germ cells were observed in the testes exposed to DBP, indicating an adverse effect on spermatogenesis (158). Some animal studies have shown that some phthalate metabolites can act as estrogen receptor agonists by binding to estrogen receptor a or b (160).

Pesticides: DDT and Degradation Products
One case-control study showed that mothers of subfertile men had significantly higher serum p,p'-DDE levels than mothers of the fertile men, which indirectly suggest the link between prenatal exposure to p,p'-DDE and male infertility (161). However, maternal serum DDE levels were measured when the men were in adult age, not during pregnancy. A pregnancy cohort study in Denmark showed that maternal level of p,p'-DDE during pregnancy was not associated with sons' semen quality (152).

Perfluorinated Compounds
A Danish pregnancy cohort study showed a negative association between maternal serum PFOA level during pregnancy and adjusted sperm concentration and total sperm count of the sons at the young adult age (162). There was no significant association between maternal serum PFOS level and semen quality of the sons (162).
In summary, there is a limited number of studies on the association between prenatal exposure to EDCs and semen quality in adulthood. Some studies demonstrated a link between prenatal EDC exposures and poor semen quality, supporting the testicular dysgenesis syndrome (TDS) hypothesis, which stated that prenatal EDC exposure can interfere with fetal testicular development and function and may result in long-term reproductive health problems (11,163). For EDCs with a long half-life, e.g., persistent organic pollutants (POPs), some studies use the concurrent measurement of EDCs in men or their mothers and semen quality, assuming that these EDC levels may reflect exposure since the fetal or infancy period. However, EDC exposures may have continued postnatally, and therefore, the timing of endocrine disrupting effects cannot be clearly identified.
The studies on the association between prenatal exposure to EDCs and semen quality are summarized in Table 4. Owing to a limited number of studies and inclusion of only few birth cohort studies, no conclusions can be drawn at the moment. More birth cohort studies are needed to better illustrate the role of prenatal EDC exposures in poor semen quality.

Postnatal Exposure
There is some evidence to support a relationship between postnatal exposure to some endocrine disrupting chemicals and low semen quality. The studies are summarized in Table 5.

Phenols: Bisphenol A
To date, ten cross-sectional, one case-control and four cohort studies have evaluated the role of BPA exposure in semen quality and they have shown mixed results. All of the studies measured BPA in urine samples, except one study in which plasma and semen samples were analyzed for BPA (179). Most studies showed a negative association between urinary BPA level and sperm concentration and/or total sperm count (164, 166-168, 170, 174, 179). A negative association between seminal BPA, but not plasma BPA levels, and sperm concentration, total sperm count and percentage of morphologically normal sperm was found in one study (179). Urinary BPA levels were negatively associated with sperm motility in some studies (170,177).
In summary, current evidence supports the link between BPA exposure in adulthood and poor semen quality, particularly low sperm concentration, total sperm count and sperm motility.

Flame Retardants
Several studies have evaluated associations between PBDE levels in serum, hair or seminal fluid and semen quality. Most of them suggested negative associations with sperm concentration or sperm motility (184,208,214,234,235). One study including men from three countries found no consistent associations across countries (236).

Phthalates
Evidence supports the link between phthalate exposure in adulthood and poor semen quality. A number of studies have shown negative associations of phthalate metabolite levels with semen quality, particularly sperm concentration (167,186,191,193) and sperm motility (167,186,187,189,191,192,195,197), although two studies showed a positive association between levels of some phthalate metabolites and sperm motility (187,188). Two studies showed a negative association of phthalate metabolite levels with semen volume (196,198,199) and four studies demonstrated a negative association with percentage of morphologically normal sperm (186,193,195,198). Three studies did not show any significant association with semen quality (184,190,194).

Organochlorine Pesticides
Most studies on the association between p,p'-DDE levels and semen quality were conducted in the early 2000s. To date, evidence has supported an association between serum DDT or DDE levels and poor semen quality, particularly reduced sperm motility (210,212,215,237,238). Some studies also reported an association with low sperm concentration (210,211), abnormal morphology (210) and low semen volume (210,211). However, some studies did not show an association with semen quality (161,208,213,229). One study reported positive associations between semen quality and serum levels organochlorine pesticides (214). Another study did not find significant associations between peripubertal serum p,p'-DDE levels and any semen parameters in adult men (209). The role of peripubertal p,p'-DDE exposure on semen quality needs to be studied further. Studies on the association with semen quality and levels of other organochlorine pesticides, including lindane and hexachlorobenzene, are summarized in Table 5.

Other Pesticides
In epidemiological studies, organophosphate exposure is usually assessed by detecting decreased blood, erythrocyte or plasma cholinesterase activity or detecting metabolites of organophosphates, for example dialkylphosphates (DAPs), in urine samples (239,240). Possible relationship between organophosphate         exposure and low semen quality has been shown (200)(201)(202)241), however the number of studies are limited. Three studies showed a negative association between the level of 3-phenoxybenzoic acid (3-PBA), which is a metabolite of pyrethroids, and sperm concentration (203,204,207). However, one study did not report such a finding (205). One study showed that higher pyrethroid metabolite levels were associated with higher percentage of sperm with abnormal morphology, lower sperm concentration, and increasing rate of abnormal computer-assisted semen analysis (CASA) parameters, including lower curvilinear velocity and linearity (206).

Parabens
Four cross-sectional studies investigated relationship between urinary paraben levels and semen quality and showed mixed results (56,(181)(182)(183). One study showed a significant positive association between urinary level of butylparaben (BP) and percentage of morphologically abnormal sperm and a negative association with percentage of sperm motility (56). Another study showed that urinary ethyl paraben (EP) and BP levels were negatively associated with sperm concentrations and urinary BP, EP and methyl paraben (MP) levels were negatively associated with percentage of sperm motility parameters by CASA. Levels of hydroxylated metabolites methyl-protocatechuic acid (OH-Me)P and ethylprotocatechuic acid (OH-EtP) were positively associated with percentage of morphologically normal sperm (183). However, two studies did not show any significant associations between paraben level and semen quality (181, 182). To summarize, there is limited amount of evidence suggesting a link between paraben exposure and semen quality.

Perfluorinated Compounds
To date, four cross-sectional studiestwo from Denmark, one from Faroe Island, and one from the USAhave examined the relationship between PFC exposure in adulthood and semen quality. Three studies did not find any significant associations between serum PFC levels and semen quality (220)(221)(222). Only one study from Denmark showed lower percentage of morphologically normal sperm in men who had high combined PFOA and PFOS levels as compared with those who had low levels (218).

Polychlorinated Biphenyls
Several cross-sectional studies have demonstrated a link between PCB exposure, particularly PCB-153, in adulthood, and low semen quality (213,224,226,228,229,231,233,242,243), particularly low sperm motility (213,224,226,228,229,231,233). In contrast, one Chinese study showed an association with higher progressive sperm motility (238). In addition, this study also showed a positive association between sum of seminal dioxin-like PCB levels and semen volume, a negative association between seminal PCB-66, PCB-105 and sperm concentration, and a positive association between seminal PCB-44 and sperm concentration (238). A study on male partners of couples trying to conceive also reported positive associations between semen quality and serum levels of some PCB congeners (214). Whereas a study on men from a fertility clinic found no significant association between semen quality and serum PCB levels (208).

Early Life Exposure
A limited number of studies have investigated the association between prenatal exposure to EDCs and reproductive hormone levels in adult men. These studies are summarized in Table 6.

Phenols: Bisphenol A
To date, there is no evidence supporting the relationship between maternal BPA exposure and reproductive hormone levels of the sons at the adult age. The pregnancy cohort in Western Australia 'Raine study' found no association of maternal serum BPA with hypothalamic-pituitary-gonadal (HPG) hormone levels of the sons (149).

Phthalates
The Australian Raine study found that serum total testosterone levels of the sons at a young adult age were positively associated with maternal serum levels of serum phthalate metabolites during pregnancy, including MEHP, MiNP, the sum of DEHP and DiNP metabolites, and the sum of high molecular weight phthalates after adjustment for BMI (153). In addition, there was a positive association between maternal serum MiNP levels and FSH levels of the men and between maternal serum DEHP levels and serum LH levels of the men. A negative association between maternal MEHP level and serum LH to testosterone ratio in adult men was also observed. No association between maternal levels of phthalate metabolites during pregnancy and serum inhibin B or estradiol levels in adult men was found (153).
A study in 112 males, aged 17.5-20.5 years, and paired maternal serum samples collected at a mean of 12 weeks of gestation in Sweden demonstrated positive associations between maternal serum levels of MCiOP and mono-(oxo-iso-nonyl) phthalate (MOiNP) and FSH levels of the sons, and between maternal serum levels of two DiNP metabolites [mono-hydroxy-iso-nonyl phthalate (MHiNP) and MOiNP] and LH levels of the sons (154). DEHP metabolite levels in maternal serum were positively associated with total and free testosterone levels of the sons (154).
Results from these two studies suggested the potential longterm effects of prenatal phthalate exposure on the hypothalamicpituitary-gonadal axis. However, more studies are needed to corroborate or refute these findings.

Dioxins
Mocarelli et al. studied reproductive hormone levels of sons born to mothers who were exposed to dioxins during pregnancy due to an accident in Seveso, Italy, and compared them with hormone levels of sons born to mothers who had background exposure. Among breastfed group, 21 sons with maternal dioxin exposure had higher FSH and lower inhibin B levels than 36 sons with maternal background exposure (150). Among the maternal dioxin exposure group, breastfed sons (n=21) had higher FSH and lower inhibin B levels than formula-fed sons (n=18) (150). Among breastfed group, sons born to the exposed mothers had lower semen quality than sons born to the non-exposed mothers. These results suggest that in-utero and/or neonatal exposure to dioxins have a role in germ cell defects.

PCBs, Pesticides (p,p′-DDE), and Perfluorinated Compounds
A birth-cohort study in Denmark showed that maternal serum PFOA level at 30 th week of pregnancy was positively associated with serum FSH and LH level of the sons at the age of 19-21 years (162). There was no significant association between maternal serum PFOS, PCBs, p,p'-DDE levels and serum levels of FSH, LH, testosterone, inhibin B, estradiol or SHBG in the adult sons (152, 162).

Postnatal Exposure
Several studies have investigated the HPG axis hormone levels in adult men in relation to EDC exposure. Many studies examined the association of EDC exposure with testosterone levels. Some studies also evaluated pituitary FSH and LH levels. Only a small number of studies evaluated levels of inhibin B, which is Sertoli cell and germ cell biomarker. Results of the studies are summarized in Table 7.

Phenols: Bisphenol A
Associations between BPA levels and reproductive hormone levels were examined in 14 cross-sectional studies ( Table 7), and they showed variable results. Eleven studies analyzed BPA level in urine samples, two studies analyzed BPA level in serum (252,253), and one study measured BPA level in plasma and seminal plasma (179). An association between BPA level and serum testosterone level was not significant in most studies (164,169,179,(247)(248)(249)(250)(251)(252)(253). Two studies have demonstrated a positive association between BPA level and serum total testosterone level (172,244) and only one study showed a negative association (246). Some studies did not show significant association between BPA and LH levels (169,245,250,251), whereas some showed a significant positive association (164,172,248). Studies on the relationship between BPA and FSH levels have also shown mixed resultsmost studies did not show any significant correlation (164,169,172,247,250,252), while two studies showed a positive association (248,251). Some studies also evaluated inhibin B level, which showed no significant association with BPA level (164,169,172,252,253).

Flame Retardants
Several studies have evaluated associations between flame retardant levels in serum or hair and reproductive hormone levels in adult men. Two studies suggested a positive association with testosterone levels (254,256). One small study suggested a negative association with inhibin B levels (255). In contrast, two large studies found no consistent or significant association between reproductive hormone levels and flame retardant levels (184, 236).

Phthalates
Experimental studies showed that phthalates had a variety of effects on the HPG axis function in male rats, including low FSH and LH levels as well as high or low GnRH and testosterone levels [reviewed in Hlisnıḱová2020 (281)]. Phthalates can also disturb testicular hormone production by altering steroidogenic enzyme activity, including decreased or increased levels of CYP11a1, Hsd3b, Hsd17b enzymes and decreased levels of CYP17a1 enzyme, changes in steroidogenic acute regulatory protein (StAR) amount (281). Maternal serum TCDD 39 men born to mothers who exposed to dioxin following the accident in Seveso, Italy (mean age, 22.5 y) vs 58 comparisons (mean age = 24.6 y)

Italy
Higher FSH and lower inhibin B in the breast-fed exposed group vs breast-fed comparisons Higher FSH and lower inhibin B in the breast-fed exposed group vs formulafed exposed group         Epidemiological studies, most of which were cross-sectional, have shown inconsistent results on the association between phthalate and reproductive hormone levels. Phthalates or phthalate metabolites were measured in urine in most studies (184, 260-262, 264, 265, 267, 282), in serum in three studies (199,263) and in seminal fluid in three studies (193,198,215). Numerous studies showed an association between levels of phthalates or phthalate metabolites and low serum total or free testosterone levels (189,195,258,260,261,264,265,267,268,282,283), and one study also found a concurrent low LH level (195), suggesting an impaired LH secretion as a cause of low testosterone level. Some studies assessed levels of serum inhibin B, which reflects Sertoli and germ cell function and/or number, and they showed that there was no association between phthalate and inhibin B levels (188,199,263,266,284,285), except for a negative association between urinary MiBP levels and serum inhibin B levels which was found in a Chinese study (259).

Polychlorinated Biphenyls
Some studies have demonstrated a negative association between PCB exposure and serum total testosterone levels (179,286). Some studies have shown an association with low free testosterone level (179,228,257,287), which might be due to an associated increased SHBG level in some studies (228,257). Most studies did not show any significant associations with FSH and LH, except for two studies. Lin et al. found a negative association between CB52, CB209 and LH level and a positive association between CB44 and LH level (288), while CB170 level was positively associated with total testosterone levels (288). Petersen et al. reported a positive association between PCB level and serum FSH level (227). Giwercman et al. found no association between PCB level and serum FSH and inhibin B levels in Sweden, Greenland, Poland, and Ukraine, suggesting no disturbance in the hypothalamic-pituitary-Sertoli cell axis (257). A study from China found either no significant association between serum PCB levels and reproductive hormone levels (254). Overall, evidence suggests a link between PCB exposure and disturbed hypothalamic-pituitarygonadal axis in men, particularly low serum testosterone level.

Perfluorinated Compounds
Cross-sectional studies on the link between perfluorinated compound and reproductive hormone levels in adult men have shown inconsistent results. Four studies did not show any significant association (218,219,269,270). Only one study by Joensen et al. showed negative associations with total and free testosterone levels, free androgen index, free testosterone/LH, total testosterone/LH and free androgen index/LH (289). Positive association between serum PFOS and LH was shown in one study (220) and between plasma PFOA and LH in another study (222).

Testicular Cancer
Testicular germ cell tumors (TGCTs) are relatively rare -accounting for about 1% of cancers in men. However, they are the most common cancer in young adult men (290,291). Their prevalence has been increasing in many Western countries (292). The main cause of this adverse trend is still unclear, but it has been proposed that EDCs might have a role (11). Testicular cancer appears to have a fetal origin, although it usually manifests after puberty when gonadotropin stimulation has started (293). Testicular cancer, cryptorchidism and hypospadias have similar prenatal risk factors and men with a history of cryptorchidism or hypospadias have an increased risk of testicular cancer (11).
Most of the studies that investigated the relationship between EDC exposure and testicular cancer used data on self-reported exposures or the occupational history or a history of chemical use without showing the chemical levels. Studies which reported EDC concentrations are scarce. Many studies are case-control studies. In addition, cohort studies evaluating the association between prenatal exposure levels and testicular cancer occurrence are lacking. Therefore, the cause-and-effect relationship is inconclusive.

Early Life EDC Exposure
Even though TGCTs are most commonly diagnosed between the ages of 15-40 years, there is evidence supporting the hypothesis that prenatal exposure to EDCs has a role in the development of testicular cancer. A Swedish study of 44 TGCT case mothers and 45 control mothers found that the concentrations of the sum of PCBs, sum of PBDEs, hexachlorobenzene (HCB), cis-and transnonachlordane and sum of chlordanes were higher in case mothers than in control mothers (294,295), suggesting a link between prenatal exposure to these chemicals and the development of TGCTs. The chemical levels of maternal blood samples were analyzed when the sons were diagnosed with testicular cancer. Chemical measurements were not performed in the blood taken during pregnancy; therefore, the timing of chemical exposure was unclear. However, these findings suggest a link between testicular cancer and chemical exposures, since these organochlorines have very long half-life and can stay in human body for several years.

Pesticides
A nested case-control study of 49 TGCT cases and 51 controls in Norway used pre-diagnostic serum samples, and no significant difference in the levels of oxychlordane, trans-nonachlor, and total chlordanes between the cases and controls was reported (296).
To date, five case-control studies have examined relationship between pre-diagnostic serum levels of p,p'-DDE and TGCTs. Two studies found higher levels of p, p'-DDE in TGCT group than those in the controls. A study among US servicemen (297) and a hospital-based study in Italy showed that the TGCT cases had significantly higher p,p'-DDE levels than those of the controls (298). A Swedish study and a Norwegian study found a tendency to higher serum p,p'-DDE levels among the TGCT cases as compared with controls; however, the difference was not statistically significant (295,296). Another US study did not show an association between TGCT and serum DDE (299).
In summary, studies on the role of prenatal EDC exposure on TGCTs are scarce. Studies evaluating the role of concurrent EDC exposure on TGCTs have shown mixed results. However, significant associations between EDC exposure and testicular cancer have been shown at least in some studies. More studies are needed to further assess these connections.

DISCUSSION
There has been a growing research interest in the potential health risk of EDCs during recent years. Experimental studies support the role of EDC exposure in the occurrence of male reproductive health problems. Results from epidemiological studies are mixed, however, evidence suggests a link between some EDC exposures and adverse male reproductive health. Maternal exposure to some EDCs during pregnancy has, at least in part of the studies, been associated with congenital urogenital anomalies, i.e., cryptorchidism and hypospadias, and low semen quality, altered HPG hormone levels and testicular cancer in adult men. The evidence for the link to the adverse adult male reproductive health is derived from a small number of studies. The association of concurrent exposure to some EDCs in adulthood with low semen quality, low serum testosterone levels and testicular cancer has been reported, although the results are not consistent.
Human studies on the association between exposure to environmental EDCs and male reproductive health are challenging because of a number of factors. First, we are continuously exposed to a mixture of different chemicals, which is different from many experimental studies that evaluated the effect of one chemical at a time. In addition, the level of exposure in animal models can be higher than human exposure in real life. Results from experimental studies are not always repeatable in human studies. Second, the exposure starts already at the embryonic period or even before that, since paternal exposure to environmental and lifestyle factors may change sperm epigenome and recent studies suggest that such changes may be the link between paternal exposures and offspring health (303,304). Furthermore, the critical period for exposure may vary for different reproductive outcomes, since for instance hypospadias is caused by a defect in fetal development of penile urethra, but sperm production capacity is determined by the number of Sertoli cells and these cells divide fast during fetal development but also postnatally and at the beginning of puberty (133,305). Therefore, the timing of exposure measurement may affect the results on the association between EDC exposure and male reproductive health. Third, participant settingsmen from general population, men who had occupational exposure to EDCs, or men who lived in the areas of accidental chemical leakage -also influence the results. Studies on the effects of accidental chemical leakage have usually shown a negative impact on semen quality or male reproductive hormone levels, while studies in men from general population are more likely to show mixed results. Men recruited from an infertility clinic, men from general population and men at a different age possibly show dissimilar association to chemical exposures. In addition, differences in exposure levels between study population may influence the observed associations. Fourth, a cross-sectional study examines the relationship between chemical exposure and semen quality or reproductive hormones at one point of time. For a chemical with a short halflife, chemical measurement at a single point might not reflect the real level of exposure in long-term. In addition, a significant correlation observed in cross-sectional study does not indicate a cause-and-effect relationship. Lastly, studies on the association between prenatal EDC exposures and adult male reproductive outcomes, including semen quality, serum reproductive hormone levels and testicular cancer need long period of follow-up, and are therefore difficult to conduct. In addition, prenatal EDC exposure is also followed by postnatal exposure from birth to adulthood.
More studies on the effects of maternal EDC exposures on the sons' semen quality and reproductive hormone levels, and more results from birth cohort studies would be beneficial. Role of paternal EDC exposure during pre-conception, particularly epigenetic studies, is a topic that needs to be studied further.