The endocrine system plays an essential and pervasive role in the regulation of whole body homeostasis. The thyroid system is a classic neuro-endocrine axis, comprising a feedback system between the hypothalamus releasing thyrotropin-releasing hormone (TRH), the pituitary gland releasing thyroid-stimulating hormone (TSH) and the thyroid gland producing and secreting T4 and, to a lesser extent, T3. In the circulation, both hormones are protein bound (in humans to a high-affinity protein called TBG or thyroxine binding globulin) and less than 1% of these hormones circulate as free-T3 (fT3) and free-T4 (fT4). The feedback among the hypothalamus, pituitary, and thyroid maintains blood levels of thyroid hormones within relatively narrow limits [1
]. Thyroid hormones are key players in the control system of our metabolism, with, amongst others, a pivotal role in energy metabolism, heart rate, and emotional stability [2
Thyroid function can be influenced and disrupted by both endogenous and exogenous factors. Some of these factors are not amendable. TSH tends to increase during the lifespan [4
], and is influenced by gender [6
]. Other elements are more modifiable. Several reports have linked active smoking to lower TSH levels [7
], whereas the influence of BMI is not as uniform [10
]. Finally, lipid levels may also be associated with TSH [12
Several man-made chemicals are suspected to have thyroid-disrupting properties [13
]. Amongst them are the polychlorinated biphenyls (PCBs), a group of organochlorine chemicals. Since they are very resistant to extreme temperature and pressure, they were widely used in capacitors, transformers, hydraulic fluids, lubricants and as plasticizers. More than 200 individual PCB congeners have been identified in commercial mixtures, whose chemical and toxicological properties are related to the number and position of the chlorine atoms. Despite the ban on their production in Europe since the 1970s, their ongoing use, chemical stability, resistance to degradation, and lipophilicity has lead to significant bioaccumulation in most compartments of the ecosystem and human tissues [16
]. This bioaccumulation leads to an ongoing human exposure to PCBs through a variety of pathways, dietary intake being the most important [18
PCBs are slowly metabolized by the human body [19
]. The first phase of this biotransformation process is mediated by cytochrome P450 monooxygenase isozymes, resulting in hydroxylated PCB compounds (OH-PCBs), which are persistent too and are slowly eliminated [20
]. Many OH-PCBs are retained in the blood bound to proteins and, as such, have the potential to exert toxic effects [20
Given their structural resemblance to thyroxin [21
], both PCBs and OH-PCBs are under scrutiny for their potential role in thyroid disruption. In animal studies, both circulating free thyroxin (fT4), as well as serum triiodothyronine (fT3) were significantly reduced after exposure to PCB mixtures or individual congeners [22
]. Some animal studies report that serum thyroid-stimulating hormone (TSH) is elevated by PCBs in response to low fT4 [26
], whereas others report essentially no effect of PCB exposure on serum TSH [27
Several studies have examined associations between PCB serum levels and thyroid hormones in humans. However, studies focusing on the link between hydroxylated PCB metabolites and thyroid function are scarce [29
]. At present, taken into consideration the past and current exposure level in a Western population, thyroid function is considered a potential target for endocrine toxicity by persistent chemicals such as PCBs [30
]. Our study aimed at investigating the relationship between serum levels of PCBs and, in particular, hydroxylated PCB metabolites, with clinically available markers of thyroid function in a group of normal weight, overweight and obese adults.
Polychlorinated biphenyls are suspected to disrupt endocrine balance, and in particular thyroid homeostasis. The mechanisms through which PCBs and their hydroxylated metabolites exert thyroid disruptive capacities have been unraveled in recent years, both in vitro
and in animal studies [37
]. PCBs are capable of interfering with thyroid hormone levels at many sites. At the DNA level, PCBs interfere with thyroid hormone receptor-mediated transcription [38
]. Secondly, they affect synthesis associated proteins such as thyroid peroxidase [24
]. In vitro
, PCBs interfere with the sodium/iodide symporter [39
] and compete with native hormone at the levels of transport protein [40
] and act as agonists at the level of the thyroid receptor [42
]. PCBs also enhance the excretion of thyroid hormones in vitro
and in rats [22
]. Finally, in fish, PCBs interfere with several iodothyronine deiodinases [44
]. These interactions offer an clarification for the reduced fT4 levels, but the pathophysiological mechanism(s) causing blunting of a compensatory TSH rise remain largely unexplained. One possible mechanism has been documented in rats, indicating high doses of PCBs can induce impaired TRH-induced TSH release at the pituitary level [45
]. As opposed to data on PCBs, data on the action mechanisms of PCB metabolites are scarce, and are largely derived from in vitro
data with few in vivo
data. Hydroxylated PCB metabolites disrupt thyroglobulin synthesis and excretion [47
] and display a weak affinity for the human thyroid receptor [41
]. In vitro
, inactivation of thyroid hormones is inhibited by OH-PCBs due to limiting of the sulfotransferase activity [49
]. Hydroxylated PCBs also interfere with thyroid hormone receptor-mediated transcription [50
]. Increased T4 accumulation in the liver, inducing decreased serum T4, has been reported [51
]. To date, no interference of PCB metabolites with TRH or TSH secretion has been documented. The lipid solubility of PCB metabolites is slightly lower than that of PCBs, which could explain the mechanistic differences. It should be noted that all of the studies mentioned above describe responses on acutely administered doses of PCBs and PCB metabolites. However, in the human situation, there is a chronic exposure, resulting from both low-level intake and internal exposure. It is unclear if the human body is capable of certain adaptive mechanism, for example by inducing CYP450 enzymes.
Human studies have provided conflicting data on the exact nature of the resulting thyroid profile change. Our study aimed at investigating the relationship between serum levels of PCBs and, in particular, the hydroxylated PCB metabolites, with clinically available markers of thyroid function in an adult population.
To our knowledge, we are the first to report 3HO-PCB118 and 3HO-PCB180 as independent, significant predictors of fT4. 3HO-PCB180 is a metabolite of PCB180 and PCB172, while 3HO-PCB118 can be formed from PCB118, PCB126 and PCB107. PCB118, PCB180 and PCB172 were also measured in our sample, but their levels did not show a significant link with fT4. We did not measure both other mother compounds of 3HO-PCB118. Dallaire reported a negative relation between serum levels of HO-PCB199 and HO-PCB208 and TSH levels in an adult Inuit population, but it should be noted that the concentration of hydroxylated PCB metabolites in their study was about a tenfold higher than in our sample [52
]. In the same study, seven HO-PCBs were associated with decreased total T3 concentration [52
]. Hagmar et al.
measured 4HO-PCB107 and 4HO-PCB187 in a female cohort, yet they do not report on the link between these PCB metabolites and thyroid function [53
]. In Vietnamese women, several PCBs and HO-PCBs correlated with fT4, and PCB74 and PCB118 correlated negatively with TSH; in Vietnamese men only PCB138 and PCB153 correlated significantly with fT4 and no relation with TSH could be detected [54
]. Our findings underscore the need for further research into the endocrine disrupting capacities of hydroxylated PCB metabolites.
When the different PCBs and PCB metabolites were analyzed individually, regression analysis indicates a significantly negative effect of PCB95lw
, and a significantly positive effect of PCB99lw
on fT4 levels. Possibly as a consequence of these divergent influences, and the non-significant effect of most of the measured PCBs, the predictive effect of the sum of all PCBs was not significant in our population. The effect of PCBs on circulating levels of fT4 has been assessed by several previous studies. Free T4 was unaffected by levels of PCBs in several studies [53
]. The study of Hagmar focused on PCB153, while all other studies analyze several PCB congeners. Concentrations of PCBs vary substantially between studies. In the studies focusing on participants with background exposure [52
], concentrations of PCBs in our study are comparable. However, in studies focusing on participants with higher exposure levels, the PCB serum levels are up to an 8 fold higher compared to our study [53
]. In the study of Langer et al.
, fT4 levels were increased in participants with higher PCB levels [37
]. However, it should be noted that the population studied by Langer et al.
consisted of people living in a heavily polluted area. Their PCB levels were an order of magnitude higher than the levels detected in our cohort and only the effect of the sum of all examined PCBs was explored. In this perspective, it is of particular interest that the effect of individual PCB compounds diverged in our study.
Our analyses indicate no effect of either PCBs or their hydroxylated metabolites on the TSH levels in our study sample. This is in line with the majority of the existing studies that did not detect a measurable influence on TSH levels [53
]. The magnitude of PCB serum levels vary considerably between studies, as both groups with background exposure as groups with high exposure were studied. Most studies are limited to one gender [53
]. It is somewhat puzzling that the remaining studies reported both increased and decreased TSH levels [56
]. Four studies reported a decreased TSH with higher levels of PCBs, in 63 occupationally exposed men in the USA [56
], in 89 women who miscarried in Germany [65
], in 464 women and men in an industrially heavily polluted area of Slovakia [66
] and in 545 fish consuming captains in the USA [67
]. In a study by Turyk et al
] in 2445 American citizens with background exposure, influence on TSH levels was affected by age and gender, with a lower TSH in older men, and an elevated TSH in older women. Abdelouahab et al.
detected no effect on TSH levels in Canadian women, while TSH levels were higher in fish consuming men with higher PCB levels [64
When analyzing the different studies in humans, the difficulties in comparability are striking. First, the analytical techniques for PCBs and their metabolites differ, but most methods can accurately detect low levels. Secondly, often the sum of PCB congeners is used for analysis, but the number of measured PCB congeners differs between studies, as does the choice in International Union of Pure and Applied Chemistry (IUPAC) numbers. Thirdly, the analyzed media vary (serum, whole blood). Finally, some authors report wet weight data, while others use lipid-adjusted data. However, PCBs are capable of inducing dyslipidemia, even in the absence of obesity [69
]. Disturbances in thyroid metabolism also give cause for lipid alterations. As such, using lipid-adjusted data to investigate the link between PCBs and thyroid metabolism may inadvertedly cause an overadjustment of reality. This caveat has recently been adverted by La Merril [70
]. In this perspective, the significant influence of lipid adjusted PCB levels that were not reproducible with wet weight data, should be interpreted with caution. Hydroxylated PCB levels are, however, always reported as wet weight, due to their binding to proteins. As PCBs compete with thyroid hormone at the level of the transport protein in vitro
, it is interesting to study this effect in vivo
in men. However, since our study focused on TSH and fT4, without data on total T4 or total T3, this was not feasible. The available data on this topic show divergent result, with several studies reporting no effect on total T4 [53
] or total T3 [55
]. In Canadian and U.S. men with high PCB levels, total T4 is lower [58
], a result that was replicated in a study with 51 U.S. women [62
] and in a mixed gender Spanish study group [63
Additionally, it should also be noted that in vitro and animal testing to study the effects of PCBs is often undertaken using a single PCB congener. However, the complexity of human exposure is extensive, given the vast array of compounds to which humans are exposed simultaneously. The concentrations of endocrine disrupting chemicals are often closely intercorrelated, providing additional difficulty in determining which of the compounds have an effect. In our study, HO-PCB180 was the only PCB metabolite to retain its statistical significance after stepwise multiple regression, while the components PCB95lw and PCB99lw also had statistical significance. As demonstrated by our results, grouping the levels of a certain class of chemicals might obscure the underlying opposing effects of individual compounds.
As mentioned previously, most studies have not been able to detect a link between TSH levels and PCB serum levels. PCBs are capable of inducing an impaired TSH response [46
], but it has not yet been explored if hydroxylated PCBs possess such capacities. However, from a clinical point of view, TSH elevation or reduction is considered to be the hallmark of impaired thyroid function. Even if the distinction between subclinical and overt thyroid disease is somewhat arbitrary because it depends to a considerable extent on the position of the patient’s normal set point for fT3 and fT4 within the laboratory reference range, subclinical hypo- and hyperthyroidism has been associated with increased, predominantly cardiovascular, morbidity [1
]. Moreover, thyroid hormones are thought to exert a complex, though profound impact on several key components of the metabolic syndrome, such as weight control, blood pressure and glucose metabolism [72
]. PCBs have been linked to the pandemic rise of obesity and associated diseases such as diabetes [73
]. The possibility of an additional aggravating effect on these conditions through the thyroid system, certainly merits investigation. So although the effect of PCB and HO-PCB disturbance at the individual level might be small, the impact at population level, and its ensuing burden of disease and health care related costs, should be considered significant.
This study has some limitations. Assessment of ongoing exposure through intake of items such as contaminated food or ingestion of dust has not been performed in this study population. PCB serum levels are known to increase slowly with age and acutely during weight loss [33
]. As this study is cross-sectional, no information was available on past fluctuations in serum levels. If these fluctuations could have an effect on thyroid hormone metabolism, remains unclear. From the current knowledge, the authors cannot rule out that an obese and/or diabetic participant metabolizes PCBs into hydroxylated PCB metabolites differently than a lean and/or non-diabetic participant. Given the cross-sectional design of our study, a causal link between the measured levels of PCB and PCB metabolites and thyroid hormone levels cannot be established. Other chemicals, such as polychlorinated dibenzodioxins or polychlorinateddibenzofurans might also have the in vitro
potency to disrupt thyroid metabolism, but were not studied in our population [13
Our study suggests PCBs and their metabolites are associated with lower levels of fT4, without a simultaneous rise in TSH. Given the limited effects on fT4 and the lack of compensatory TSH rise, it remains uncertain if these changes are clinically relevant. In the future, a prospective, longitudinal study, focusing on TSH, fT4 and fT3 would help to unravel to what extend PCBs and their metabolites are capable of disrupting circulating fT4 and fT3 hormone levels and inhibiting the compensatory rise in TSH normally triggered by fT4 and fT4 decline. It would be interesting to investigate if such inappropriately low TSH is caused by an impaired TRH-response, or caused by a change in local fT4/fT3 signaling at the target tissues.