Cholesterol is the precursor to all steroid hormones, such as estrogens, androgens, and corticosteroids 1. In fish, as in other vertebrate species, cholesterol is obtained by dietary intake, release from intracellular stores or by de novo synthesis 2. Cholesterol absorbed at the intestines is esterified with free fatty acids to form hydrophobic cholesterol esters (CEs), which are transported in the plasma in association with lipoproteins to sites of metabolism or storage 34. At the tissues, receptor-mediated lipoprotein endocytosis delivers cholesterol to the intracellular environment for immediate use or re-esterification for intracellular storage 5. While exogenous cholesterol is obtained in this way by most steroidogenic tissues (such as the adrenal and ovary), some tissues (such as the testis) differentially utilize de novo synthesized cholesterol as a substrate for steroid biosynthesis 26. The definitive mechanisms of cholesterol metabolism in fish are believed to be highly similar to mammals, although comprehensive information on lipid dynamics in teleost species is limited. Cholesterol biosynthesis is believed to proceed by the same pathway in fish as in higher vertebrates 7. A number of reviews on animal 689 and fish 10 lipoprotein dynamics indicate structure is similar, although TG content is elevated at the expense of cholesterol esters in teleost lipoproteins. Fish plasma is considered hypercholesterolemic relative to higher vertebrates, with 2–6× higher circulating cholesterol concentrations a normal physiological condition in fish species 11. The traditional detrimental affects associated with high plasma cholesterol such as coronary lesions and plaque formation are notably absent in fish species, presenting an interesting and stark contrast to the human condition.

De novo cholesterol synthesis begins with the production of acetyl CoA via acetate or citrate. Acetyl CoA is subsequently transformed to 3-hydroxy-3-methylglutaryl-CoA (HMG CoA), mevalonate, squalene, lanosterol and ultimately cholesterol via a number of enzymatic transformations (for a detailed summary see 12). The newly synthesized cholesterol (Figure 1A) can then enter tissue-specific biosynthetic pathways such as steroidogenesis, or be incorporated into plasma membranes or esterified for intracellular storage. The relative contribution of de novo cholesterol synthesis to the total cholesterol pool in endocrine organs such as the gonads is generally not known. Ovarian tissue in mammals has been shown to utilize de novo TG and cholesterol synthesis during the pre-ovulatory phase of gonadal development, doing so preferentially over lipoprotein uptake 213. Testis cholesterol is preferentially sourced from de novo synthesis, while the adrenal and ovary appear to revert to endogenous synthesis only if circulating concentrations are limiting 2. Preferences of lipoprotein-derived or de novo synthesized cholesterol by fish endocrine tissues are presumably similar to mammalian tissues, however, direct studies on fish species are lacking.

The phytosterol β-sitosterol (β-sit) very closely resembles cholesterol, differing only by an ethyl group on carbon 24 (Figure 1B). β-sit disrupts normal endocrine system function in fish by decreasing steroidogenic biosynthetic capacity 14 and disrupting plasma cholesterol concentrations and mitochondrial translocation to the steroidogenic pathway 1516. β-Sit has estrogenic properties as evidenced by the induction of the normally quiescent vitellogenin gene in male fish 17. The capacity for plant sterols to affect de novo cholesterol synthesis is known to occur in the human disorder sitosterolemia, a condition where abnormally high concentrations of plant sterols accumulate in the plasma and tissues. Sitosterolemic patients experience impaired whole body de novo cholesterol synthesis by down-regulation of key synthetic enzymes 1819, and this sensitivity suggests de novo cholesterol synthesis in other species may also be impaired by phytosterol exposure. Previous studies have identified changes in cholesterol availability following phytosterol exposure 12, however, a lack of information on preferred substrate (exogenous or de novo synthesized cholesterol) in the gonads impedes further studies to determine mechanisms of endocrine disruption.

The present study examined endogenous cholesterol synthesis in the gonads of male and female goldfish exposed to β-sit and 17β-estradiol (an estrogenic control; Figure 1C) to determine if de novo cholesterol synthetic capacity was impaired relative to non-exposed fish. Further, the relative contribution of de novo cholesterol synthesis to the reproductive steroidogenic pathway is unknown in fish; therefore, this study also aimed to assess the contribution of de novo synthesis to the total cholesterol substrate pool.

All chemicals were purchased from Sigma-Aldrich (Oakville, ON, Canada) unless otherwise specified.

The de novo cholesterol biosynthetic pathway begins with the acquisition of acetyl-CoA, but commitment to cholesterol synthesis occurs subsequent to acetyl-CoA production. TG biosynthesis also utilizes acetyl-CoA as a substrate, potentially diverting substrate from the de novo cholesterol synthetic pathway 27. TG is the primary metabolic energy storage molecule in fish where upwards of 80% of total body lipid composition is present as TG 7. The current study suggests acetyl-CoA is preferentially directed towards TG synthesis over cholesterol in early to mid-recrudescing gonads of male and female goldfish. Ovarian tissue preferentially uses lipoprotein-derived cholesterol as a steroidogenic substrate 6, therefore, a large de novo cholesterol biosynthetic capacity was not expected in female gonads. The higher ¹⁴C-TG enrichment in ovarian tissue is consistent with an increased TG demand during reproductive development; growing oocytes incorporate high concentrations of TG to provide metabolic fuel for developing embryos 28. Similar studies on trout (Salmo gairdneri) have shown high acetate incorporation towards TG synthesis during later gonadal development 45. In testicular tissue it is believed that de novo-derived cholesterol is the primary substrate for steroidogenesis 2, therefore, more substantial ¹⁴C-acetate incorporation towards cholesterol was predicted. In contrast, ¹⁴C was enriched 6-fold and 3-fold higher towards TG and FFA than towards cholesterol and CEs. This higher level of de novo TG synthesis was consistent in all male treatment groups.

While fish testes generally have low lipid content that varies with season and reproductive stage 7, the high plasma T concentrations in male control fish suggest T synthesis (and thus steroidogenic substrate) was not impaired at the reproductive stage in the present study. The ¹⁴C-acetate incorporation data suggest testis tissue may direct acetyl-coA towards TG formation when cholesterol availability to steroidogenesis is not limiting. Plasma cholesterol concentrations were above 200 mg/dL for both sexes in all treatment groups, suggesting cholesterol and steroidogenic capacity were not limited at the point of circulatory uptake or de novo synthesis of cholesterol. The absence of a reduction in plasma T in the E₂ treated fish and its significant decrease in β-sit exposed animals provides evidence of a unique, non-estrogenic mechanism of β-sit endocrine effects. Additionally, the depression of FFA synthesis by both β-sit and E₂ treatment demonstrates common effects on lipid dynamics in exposed fish. Previous studies have established that β-sit changes plasma cholesterol dynamics and has endocrine effects distinct from E₂ 29. The regulation of steroidogenesis is a multifaceted feedback system among the hypothalamus, pituitary and the gonads, known as the HPG axis, therefore impairment of function can occur at multiple levels. MacLatchy et al. 28 demonstrated that β-sit does not alter plasma luteinising hormone (LH) concentration, while E₂ interacts with the HPG axis extensively 303132. In particular, E₂ and FSH are involved in regulating lipid accumulation in the ovaries of salmon (Oncorhynchus kisutch), thereby demonstrating a potential for E₂ lipidemic effects 33. The increase in plasma cholesterol seen currently in the male β-sit treatment group and an absence in the E₂ treatment group, demonstrates independent effects of β-sit on plasma cholesterol concentration, likely caused by a mechanism distinct from an E₂ feedback system. Evidence of β-sit interacting with lipid synthetic enzymes in humans 18 would suggest disruption likely occurs at the level of cholesterol biosynthesis. Further studies examining lipid synthetic enzyme kinetics will provide good insight towards understanding the interaction of β-sit and E₂ with the cholesterol synthetic pathway.

The goldfish used in the current study were undergoing early to mid-recrudescence, based on preceding acclimation conditions, observations during sampling, GSI values and plasma T concentrations. Reproductive development in goldfish is highly regulated by changes in environmental parameters such as photoperiod and water temperature 34 and the laboratory environmental conditions for this study were designed to initiate recrudescence (15–16°C; 14 h L:10 h D). Fish experience low circulating cholesterol concentrations during sexual maturation, and it is known plasma lipids are variable and influenced by reproductive state 10. Testis lipid content varies with season and, therefore, reproductive development 7. It is thus likely that sensitivity to lipid-altering chemicals may also vary with reproductive state, presenting contradicting results dependent on the specific physiology and lipid metabolism of the particular gonadal stage at the time of exposure. This appears to be the case with β-sit, which affects endocrine endpoints such as plasma T and cholesterol concentration differently in males and females at different reproductive states and exposure durations. For example, in a long-term β-sit exposure covering a 5-month period of gonadal development and starting prior to the initiation of recrudescence, there were no effects on plasma T or cholesterol concentration by β-sit during recrudescence 35. However, at the initiation of a subsequent cycle of gonadal development, in the same 5-month exposure, male fish experienced significant reductions in plasma T 36. In the current study, plasma cholesterol concentrations increased in the β-sit treatment group, an effect in contrast to previous reports 1415, indicating β-sit can also have variable effects on plasma lipids. It is, therefore, critical to interpret endocrine effects of β-sit exposure with conscience to the reproductive stage at the time of exposure and sampling. Further, the complex and intricate nature of the regulation of steroidogenesis by the HPG axis makes the interpretation of plasma hormone data challenging. The reduction in plasma T in male fish exposed to β-sit but not E₂ likely results from different sensitivities of the HPG axis to estrogenic control at different stages of the reproductive cycle; male and female fish were not sensitive to down-regulation of T synthesis by E₂ in the present study. The impact of β-sit exposure on plasma T would, therefore, appear to be caused by non-estrogenic mechanisms, likely related to limited cholesterol availability to the steroidogenic pathway 16. Steroidogenic acute regulatory protein (StAR) has been identified as a mitochondrial cholesterol transporter 37, and β-sit has been shown to reduce StAR mRNA abundance in male goldfish 36. Given that de novo cholesterol synthesis was unaffected by β-sit exposure and plasma cholesterol concentrations were not decreased, it is highly possible that cholesterol delivery to the steroidogenic pathway is impaired rather than intracellular cholesterol availability.

Male and female goldfish responded differently to β-sit and E₂ exposure; plasma testosterone and cholesterol concentrations were unchanged by β-sit in female fish. Few studies have examined the responses of both male and female fish to E₂ or β-sit exposures. Similar steroid responses in male and female goldfish (male GSI = 3.9 – 4.9%, female GSI = 4.5 – 6.2%) to varying concentrations of sitosterol have been reported 29; both sexes experience significant reductions in circulating T in response to β-sit, but not E₂. In contrast, the present research found male fish to be more sensitive to reductions in plasma T following β-sit exposure than females. These differences in plasma hormone response between studies is likely at least partly due to differences in gonadal stage at the time of the exposures (current experiments, male GSI = 0.7–1.78%, female GSI = 1.9 – 2.8%) and, therefore, a different physiological environment in which exogenous chemicals interact with gonadal tissue. Further, circulating E₂ is a normal physiological reproductive stage-dependent phenomenon in female fish 38, therefore feedback systems and metabolic responses to exogenous E₂ exist in females that are not normally activated or required in male fish. In the present study, however, male fish were more sensitive to β-sit than female fish, and generally neither plasma T or cholesterol were sensitive to E₂ in either sex. It appears there are sex differences in sensitivity to β-sit with regard to plasma hormones and cholesterol and these varying sensitivities are likely related to reproductive stage and the associated endogenous regulation of gonadal development.

A quantitative, physiologically-relevant indication of steroidogenic output can be determined using gonadal in vitro incubation methods that measure steroidogenic output by gonadal tissue 39. Reported rates of T production in goldfish testis range from 1 pg/g (unknown GSI) 40 to 20 pg/mg (GSI 2.2%), with the latter corresponding to a plasma T concentration of 3.0 ng/mL 14. In vitro steroid biosynthetic capacity was not measured in the current study due to limited gonadal tissue availability; however, control male plasma T concentrations were comparable to those reported by MacLatchy & Van Der Kraak 28. Therefore, an estimate of the cholesterol requirements for T production in the male control group of the present study can be calculated using the in vitro biosynthetic capacity reported by MacLatchy and Van Der Kraak (20 pg/mg over an 18-h period). Cholesterol is used for additional cellular functions in the testis (storage, plasma membrane integrity, synthesis of other steroid hormones), therefore, any estimation of cholesterol requirements based solely on T production will be highly underestimated. Given this limitation, an estimate of 2.9 × 10^-10 moles of cholesterol were required to produce 20 pg of T (using a 1:1 testosterone: cholesterol stoichiometry). In the current study, control male fish incorporated 1.47 × 10^-9 moles of acetate, or 8.8 × 10¹⁴ molecules towards cholesterol synthesis during the 18-h incubation period. Given an 18:1 stoichiometry (18 molecules of acetate for each cholesterol synthesized) 41, this equates to 8.2 × 10^-11 moles of cholesterol synthesized de novo. Additional studies have measured gonadal tissue cholesterol concentrations from recrudescing male goldfish and report 22.6 mg cholesterol/g tissue 35. This equates to 5.8 × 10^-5 mol/g, which if corrected to 25 mg of testis tissue as used in the current study, estimates 1.5 × 10^-6 mol of endogenous cholesterol is present in each de novo testis sample. Therefore, the de novo contribution (8.2 × 10^-11) is very small relative to the total cholesterol pool (1.5 × 10^-6). However, the estimated de novo synthetic capacity is within an order of magnitude of the estimated steroidogenic needs for T synthesis (2.9 × 10^-10) and could make a significant contribution to the steroidogenic substrate pool.

Identification of lipid classes was limited to cholesterol, TG, FFA and CE; it is highly likely that ¹⁴C-acetate was incorporated to phospholipids and monoacyl- and diacyl-glycerols as well 26. Weigand and Idler 45 examined acetate incorporation to various lipid classes in ovarian tissue of trout and found that while polar lipids were favoured when GSI was small, TG was preferentially synthesized in fish with larger GSI. In catfish (Heteropneustes fossilis), plasma and ovarian phospholipid concentrations followed the same trends as cholesterol and TG throughout the reproductive cycle 42 and high plasma phospholipid concentrations have been documented during maturation in goldfish 43. Given such indications that dominant gonadal lipid classes can vary throughout the reproductive cycle, it should be noted that broad conclusions on gonadal lipid synthesis are limited to only the neutral lipid classes in the current study.

The high ¹⁴C-acetate incorporation towards FFA in control fish was absent in the corresponding male β-sit and E₂ treatment groups. Phytosterols detrimentally affect cholesterol metabolism in the human disease sitosterolemia by inhibiting endogenous cholesterol biosynthesis 19, therefore, inhibitory impacts of sitosterol exposure on fish were expected. The first enzyme subsequent to the formation of acetyl Co-A in the cholesterol biosynthetic pathway, acetoacetyl CoA thiolase, is down-regulated in sitosterolemic patients 18 and phytosterols have been shown to down-regulate numerous other cholesterol biosynthetic enzymes in mammals 44. The current study did not examine enzyme kinetics. The three week β-sit exposure prior to gonadal tissue incubation with ¹⁴C-acetate would serve to establish physiological conditions similar to those in sitosterolemia, where tissues were exposed to higher than normal concentrations of plant sterols. Surprisingly, there were no differences in ¹⁴C-enrichment of cholesterol or CEs in male or female fish, indicating no changes in de novo cholesterol synthetic capacity among treatment groups. The lower ¹⁴C-acetate incorporation to FFA in male β-sit- and E₂-treated gonads suggests both substances have the capacity to down-regulate endogenous FFA synthesis in gonadal tissue. Further, the similar effect by both substances demonstrates the capacity of β-sit to act with estrogenic properties in vivo. Therefore, while the current study provides evidence that β-sit has non-estrogenic endocrine disrupting behaviour, as evidenced by decreased plasma T, it also has effects similar to estrogen in terms of its effects on FFA synthesis in the gonads. There is no evidence in the present study to suggest β-sit has effects on de novo cholesterol biosynthesis at this reproductive stage for either sex; there is no variation in cholesterol synthesis in the gonads of fish exposed to β-sit or E₂.

The present study clearly demonstrates that ovarian and testicular tissues have comparable capacities for de novo cholesterol synthesis at early to mid-recrudescence, and neither β-sit nor E₂ treatment disrupts gonadal cholesterol biosynthesis in goldfish at this reproductive stage. Further, differences between β-sit and E₂ exposure effects were demonstrated, suggesting the mechanisms of action of β-sit occur at points outside the HPG axis to yield effects on steroid biosynthesis and lipid metabolism.

RLS originated the detailed experimental design, some of the study concepts and methodology; carried out the experimental exposures, steroid and lipid analyses, statistical analyses; and drafted the manuscript. MD participated in the infrared spectroscopy validation of lipid types. DLM initiated the original study concepts, participated in design of the study and helped to draft the manuscript. All authors read and approved the final manuscript.