2.2 The association between glyphosate / glyphosate-based herbicides, endocrine disruption and reproductive toxicity
According to the EOS article:
Romano et al (2010) have shown that a Roundup formulation was a potent endocrine disruptor in male rats and caused disturbances in reproductive development during puberty. Adverse effects (including delayed puberty and reduced testosterone production) were found at and above the lowest dose of 5 mg/kg.
Dallegrave (2007) observed adverse reproductive effects in the male offspring of female rats treated with a Roundup formulation at 50, 150 or 450 mg/kg during pregnancy and lactation. The effects, which occurred in the absence of maternotoxicity, included dose-related decreases in serum testosterone level at puberty, decreased sperm number and daily sperm production in adulthood, an increased percentage of abnormal sperm, and sperm cell degeneration.
Glyphosate active constituent causes sperm damage in rabbits (Yousef et al, 1995).
When administered to rats for two years at 3, 10 and 32 mg/kg bw/d, glyphosate caused testicular tumours (Lankas, 1981). Although the effect did not occur in a second rat carcinogenicity study at 100, 410 and 1060 mg/kg bw/d, EOS argues that effects related to endocrine hormones can be more potent at low doses than higher ones.
Based on the following evidence, EOS proposes that glyphosate and GBHFs cause reproductive toxicity by mechanisms involving endocrine disruption:
Glyphosate-based herbicides perturb hormone levels in female catfish and decrease egg viability (Soso et al, 2007) and mediate anti-androgenic and anti-oestrogenic activity in human cells at concentrations as low as 5.0 ppm (Gasnier et al, 2009).
Roundup reduces production of progesterone in mouse cells in vitro by inhibiting expression of a regulatory protein (Walsh et al, 2000).
Glyphosate disrupts oestrogen-regulated gene expression in human cells (Hokanson et al, 2007) and is toxic to human placental cells, an effect enhanced in the presence of Roundup adjuvants (Richard et al, 2005). Richard et al are said to have shown that Roundup inhibits aromatase (the enzyme responsible for oestrogen production), and proposed this as an explanation for increased premature births and miscarriages reported in female members of farming families using glyphosate (Savitz et al, 1997 and Arbuckle et al, 2001; see previous Section).
Glyphosate and Roundup damage or kill human umbilical, embryonic and placental cells at concentrations below those recommended for agricultural use, and may interfere with human reproduction and embryonic development (Benachour et al, 2007; Benachour and Seralini, 2009).
Between them, the German BVL (for the EU and JMPR), Australian DoHA and US EPA have assessed no fewer than eight single- or multi-generation reproduction studies with glyphosate in rats, most of which involved dietary administration. The various agency evaluations are summarised in Appendix 3. The overall dose range was 3 – ca 1500 mg/kg bw/d. The toxicological end-points examined included oestrus cycling, mating performance, pregnancy rate and gestation length; litter size and sex ratio; the growth rate, attainment of post-natal developmental landmarks and onset of puberty in pups; and histology of the reproductive organs and analysis of sperm and oocytes in adults. If glyphosate was capable of interfering with the sexual development and reproductive performance of either males or females, the studies would have revealed these effects.
There were few indications of reproductive toxicity. In the parental generations, toxicity was seen as depressed bodyweight or bodyweight gain from doses of ca 670 mg/kg bw/d upwards; and, in one study only, histological abnormalities in the salivary glands occurred at >200 mg/kg. Parental NOELs ranged from 10 to ca 700 mg/kg bw/d. Pup bodyweight or bodyweight gain was depressed at >670 mg/kg, while in one study, litter size was reduced at ca 1500 mg/kg bw/d. NOELs in pups varied from 10 to ca 800 mg/kg bw/d. The Australian ADI for glyphosate (0.3 mg/kg bw/d) is based on the three-generation dietary study of Schroeder and Hogan (1981), in which there were no treatment-related effects on the parental of filial generations at the highest dose of 30 mg/kg bw/d.
Lower threshold doses for toxicity were seen with glyphosate trimesium in a two-generation study by Stauffer Chemical Co (1983a, assessed by DoHA, 1991). A NOEL of 7.5 mg/kg bw/d was assigned for parental animals and offspring based on reduced bodyweight gain, food consumption and plasma protein levels in adults and depressed pup bodyweight and relative spleen weight at >40 mg/kg. The only effect on reproductive parameters was a reduction in litter size, which occurred at the highest dose of 100 mg/kg bw/d.
For the EU and JMPR reviews, the BVL also assessed a 13—week US National Toxicology Program study in rats (Chan and Mahler, 1992). Caudal epididymal sperm concentrations declined by ca 20% at 25 000 and 50 000 ppm glyphosate in the diet (calculated glyphosate intake ca 2500 and 5000 mg/kg bw/d). However, all values were within the HC range and no effects occurred on caudal, epididymal and testicular weights, sperm motility, total spermatid heads/testis and total spermatid heads/gram caudal tissue. Compared with controls, oestrus cycle length was prolonged from 4.9 to 5.4 days at 50 000 ppm. The EU and JMPR regarded this finding as having unknown biological significance, if any. An identical study in male and female mice did not find any evidence of reproductive toxicity or endocrine modulation at up to 50 000 ppm in the diet (7500 mg/kg bw/d), the highest dietary concentration tested.
In an unreliable and poorly-reported study, Yousef et al (1995) administered glyphosate orally to male rabbits for six weeks at 1% or 10% of the LD50. The study authors did not identify the dosing interval, or the doses in terms of mg/kg bw. Semen quality was assessed at weekly intervals for six weeks prior to treatment, during the dosing period, and a further six weeks after treatment to study reversibility of effects. Glyphosate was claimed to have caused fully or partially reversible decreases in ejaculate volume, sperm viability and sperm activity. However, the results are likely to have been affected by methodological deficiencies, and effects on sperm concentration and morphology are uninterpretable due to major, unexplained variations over time within the control group.
2.2.2 Evidence of endocrine modulation in other studies
Even though they are not specifically designed to test for endocrine disruption, the short-term repeat-dose, subchronic and chronic in vivo toxicology studies required by the APVMA and other regulatory agencies can detect modulation of endocrine system activity. Chemicals affecting endocrine target sites initiate direct or compensatory biochemical or cellular responses which are observable by assessment of the weight, gross pathology and histopathology of endocrine organs and tissues. In fact, these studies have some advantages over in vitro screening assays, as they assess a variety of endocrine-sensitive endpoints in live animals capable of metabolic activation and/or detoxification of xenobiotic chemicals, and use extended exposure periods encompassing various stages of endocrine development (Williams et al, 2000).
There have been no findings in these subchronic or chronic toxicity studies indicating that glyphosate produces any endocrine-modulating effects. Negative results also were obtained in a dominant lethal mutation study in mice at 2000 mg/kg bw PO (Wrenn, 1980). While this latter test is typically used to assess genetic toxicity, substances that affect male reproductive function through endocrine modulating mechanisms can also produce effects in this type of study (Williams et al, 2000).
2.2.3 Testicular carcinogenicity
A carcinogenicity study by Lankas (1981) has been reviewed by the Australian DoHA (1985), the WHO (1994) and the US EPA (1993). The German BVL did not evaluate this study for the JMPR, but the EU review includes a summary of the WHO assessment. Rats were treated with glyphosate in the diet for 26 months to achieve intakes of ca 3, 10 and 31 mg/kg bw/d in males and 3.4, 11 and 34 mg/kg bw/d in females. The incidence of testicular interstitial (Leydig) cell tumours at termination was 0/15 among controls and 2/26, 1/16 and 4/26 at the low-, mid- and high-doses respectively. The total incidence for all males was 0/50, 3/50, 1/50 and 6/50. The BVL evaluator did not attribute the finding to treatment, noting that Leydig cell tumours are common in ageing rats, that the incidence at 31 mg/kg “only slightly exceeded the historical control range,” and that no such effect had been observed in several more recent rat studies at much higher doses. In the absence of treatment-related effects, the NOEL was set at 31 mg/kg bw/d.
The WHO (1994), US EPA (1993) and DoHA (1985) all agreed that the tumours were not treatment—related because their incidence lay within the HC range. This interpretation was supported by data shown in the Australian assessment, showing that the incidences of Leydig cell tumours in glyphosate-treated rats were not different to those in male controls from concurrent studies at the same laboratory (4/65, 3/11, 3/26, 3/24 and 3/40).
Furthermore, testicular tumours have not occurred in any of the other carcinogenicity studies with glyphosate in rats or mice at doses of up to 4800 and 1200 mg/kg bw/d, respectively. Despite EOS’s claim that endocrine-mediated effects are specifically low dose phenomena, doses of between 4 and 12 mg/kg bw/d (within the range given by Lankas) have failed to cause any testicular effects in two carcinogenicity studies in mice or in three similar studies in rats. Therefore, the weight of evidence does not support EOS’s assertion that glyphosate is a testicular carcinogen.
2.2.4 Effects of glyphosate-based herbicide formulations
Notwithstanding the mainly negative findings on glyphosate in carcinogenicity and reproductive toxicity studies in laboratory animals, the APVMA has initiated an independent assessment of publications cited by EOS, and other relevant articles obtained from the scientific literature. Three of these publications describe studies of the effects of GBHFs on the reproductive physiology of rodents and birds, while the remainder cover experiments in isolated cells. The detailed assessments are presented in Appendix 4.
22.214.171.124 Findings in birds
Oliviera et al (2007) observed a 90% reduction in plasma testosterone levels in sexually mature drakes gavaged orally with Roundup (480 g/L glyphosate isopropylamine salt, no other constituents identified) for 15 days at 5 or 100 mg/kg bw/d. This occurred in conjunction with decreased androgen receptor expression within testicular (Sertoli) cells and histological abnormalities in the testis (reduction in seminiferous tubule epithelium and interstitial tissue), epididymal region, proximal efferent ductules (vacuolisation and increased lipid in the epithelium) and epididymal duct (collapsing and folding). As most of these effects were present in birds receiving the lowest dose of 5 mg/kg bw/d, a NOEL was not demonstrated. The study did not investigate whether there were any associated effects on the behaviour or reproductive performance of the birds, define the mechanism by which the effects occurred, or identify the causative component(s) of the test formulation.
126.96.36.199 Findings in rats
Dallegrave et al (2007) performed a single generation reproduction study in rats with a Roundup product (360 g/L glyphosate and 18% POEA surfactant) at maternal oral doses equivalent to 0, 50, 100 and 450 mg glyphosate/kg bw/d. The test formulation was administered to the dams throughout pregnancy and lactation, until the offspring reached 21 days of age. Male pups were then evaluated when 65 or 140 days old. There was no NOEL because of decreased sperm production, an increased incidence of abnormal sperm, and depression in blood testosterone concentration at and above the lowest dose.
In a post-natal development study, Romano et al (2010) treated weanling rats orally with a Roundup product containing 648 g/L glyphosate isopropylamine salt plus unidentified “inert ingredients”. The doses were 0, 5, 50 and 250 mg glyphosate/kg bw/d, administered from 23 to 53 days of age. Treated males displayed reduced serum testosterone levels and thinning of the seminiferous tubule germinal epithelium, suggesting diminished production of sperm. Male puberty was delayed at 50 and 250 mg/kg. There was no NOEL.
The APVMA’s independent assessment notes that the studies by Dallegrave et al (2007) and Romano et al (2010) appear to have demonstrated evidence of reproductive toxicity. However, both studies are affected by flaws in their design, methodology and / or reporting. Neither research group identified which constituent(s) in the test formulations mediated the reported effects. Also, while there is a biologically plausible association between delayed puberty, deficiency in circulating testosterone level and inhibited sperm production, the studies did not identify the mechanism involved.
The situation is complicated by a pre / post-natal development experiment by Romano et al (2012), which yielded markedly different findings despite using the same rat strain and Roundup product as did the 2010 study. In the 2012 report, reproductive physiology and behaviour were investigated in male rat pups whose mothers had been dosed orally from GD 18 to PND 5, at 50 mg glyphosate/kg bw/d. The pups were then reared without further exposure until evaluation at 60 days of age. Compared to controls, puberty occurred earlier in the test group; serum testosterone, oestradiol and LH concentrations were doubled; sperm production was enhanced; and males showed a greater preference for the company of female rats despite an increase in the delay before mating. Based on these findings, Romano et al concluded that glyphosate is a potential endocrine disruptor.
However, DeSesso and Williams (2012; see Appendix 4), have questioned several aspects of the study’s design and conduct, and observed that the average age and bodyweight of test animals at puberty lay within the range shown by concurrent controls and controls in Romano et al (2010). DeSesso and Williams also note that surfactants likely to be present in the test formulation inhibit steroid production in Leydig (testicular) cells (Levine et al, 2007) and could have affected the study outcome.
188.8.131.52 Findings in vitro
According to the JMPR (2004b), glyphosate had no oestrogenic activity in assays for activation of rainbow trout oestrogen receptors in yeast or vitellogenin production in a trout liver cell culture system (Petit et al, 1997). The incubation concentrations of glyphosate were not given.
A Roundup formulation was reported as having dose-dependently inhibited progesterone synthesis in mouse MA-10 (Leydig tumour) cells (IC50 of 24 µg/mL) (Walsh et al, 2000). The putative mechanism involved preventing the expression of steroidogenic acute regulatory (StAR) protein, a mitochondrial phosphoprotein that transfers cholesterol to cytochrome P450scc, the enzyme that initiates steroid hormone biosynthesis. Glyphosate active constituent, by contrast, had no such effect over the concentration range tested (0–100 µg/mL). However, Levine et al (2007) replicated the effect on progesterone synthesis in the same experimental model using ‘blank’ Roundup formulation (without glyphosate), and demonstrated that inhibition arose from damage to mitochondrial membranes by the surfactant.
In MCF-7 human breast adenocarcinoma (oestrogen sensitive) cells exposed for 18 hours to a GBHF at 0.00023 – 0.23%, significant changes occurred in the activity of three out of 1550 oestrogen-regulated genes. There was a 2.2-fold increase in the activity of HIF1 (which primes cells for the initiation of apoptosis) and ca 50% reductions in expression of CXCL12 (a lymphocyte chemoattractant) and EGR1 (which has a range of activities potentially affecting apoptosis and tumour vascularisation) (Hokanson et al, 2007). However, the study did not demonstrate any alteration of the physiology, survival or growth of the test cells, or establish whether the effects on gene expression would have implications for the survival, development and function of other mammalian cells, tissues, foetuses or adult animals. Furthermore, the formulation component that altered gene expression levels was not identified.
As reported by EOS, a Roundup formulation inhibited aromatase (CYP19, an enzyme which converts androgens to oestrogens) in human plancental cancer (JEG3) cells (Richard et al, 2005; assessed by DoHA, 2005). However, as the DoHA evaluation observed, the use of human placental cancer cells (rather than normal placental cells) was not a valid basis for any conclusion that glyphosate or its products cause reproductive effects in humans, particularly given the weight of evidence from laboratory animals that glyphosate is not a reproductive toxin. Williams et al (2012) have pointed out that the concentrations of Roundup causing aromatase inhibition
(0.2–2.0%) in Richard et al’s study were cytotoxic and much higher than physiologically relevant; by contrast, pure glyphosate had no effect in the assay system at up to 0.8%, the highest concentration tested. The French Ministry of Agriculture and Fish (2005) has also evaluated Richard et al (2005), and concluded that the study was of no value for human health risk assessment.
Roundup formulations also inhibited aromatase in human embryonic kidney (HEK293) (Benachour et al, 2007) and hepatoma (HepG2) cells (Gasnier et al, 2009). By contrast, glyphosate inhibited aromatase weakly or had no effect on its activity. Roundup formulations had anti-oestrogenic activity at human oestrogen receptors (hER) α or β, and anti-androgenic activity at human androgen receptors (hAR) (Gasnier et al, 2009). However, the potencies of Roundup formulations correlated poorly with the concentration of glyphosate they contained; furthermore, glyphosate itself had no anti-oestrogenic activity at hER α or β and, at most, weak anti-androgenic activity at hAR.
Benachour and Seralini (2009) studied the cytotoxicity of glyphosate, its metabolite AMPA, four Roundup products and the surfactant POEA in three human cell lines (umbilical cord vein endothelial [HUVEC] cells, JEG3 and HEK293). Based on inhibition of mitochondrial respiration, the least potent cytotoxin was AMPA, glyphosate had intermediate potency, and POEA was the most potent (the respective EC50s were >40 000, ca 10 000 and 3–30 ppm). All the product concentrates were more toxic than glyphosate alone, having EC50s of 30 – 9000 ppm. Their potency was not dependent on the concentration of glyphosate they contained, suggesting that other formulation components were biologically active. AMPA and POEA caused necrotic cell death, glyphosate caused cell death via apoptosis, while the Roundup formulations mediated cell death via both necrosis and apoptosis.
Cytotoxicity experiments with isolated rat testicular cells in vitro have shown that germ cells are relatively resistant to glyphosate and Roundup Bioforce, Leydig cells are resistant to glyphosate but sensitive to the product at concentrations of >0.10% in solution, and Sertoli cells are sensitive to glyphosate at >0.01% and the product at 0.10% (Clair et al, 2012). Notwithstanding the decreases in circulating testosterone levels observed in vivo, neither the active nor the formulation inhibited 3β-hydroxysteroid dehydrogenase activity (an index of testosterone synthesis) in cultured Leydig cells exposed for 24 hours at up to 0.10%. Testosterone concentration in the cell incubation medium declined by ca 1/3 in response to glyphosate and Roundup at 0.0001%, but not at higher concentrations. There was no explanation for this paradoxical concentration-response relationship.