Telomerase gene expression bioassays indicate metabolic activation of genotoxic lower chlorinated polychlorinated biphenyls

Polychlorinated biphenyls (PCBs) are ubiquitously occurring pollutants with different chemical and toxicological properties. In this study we evaluated blood plasma samples of two PCB-exposed cohorts for their ability to alter telomerase (hTERT) gene expression. Blood plasma from PCB-exposed individuals inhibited hTERT expression depending solely on the concentration of lower chlorinated PCBs, with the lowest observed adverse effect level (LOAEL) at a plasma concentration between 0.5 and 2 µg/L of LC PCBs. Individual OH-metabolites derived from the WHO indicator congeners PCB 28 and PCB 101 mimicked these effects on hTERT expression in vitro with high toxicity, including DNA damage. However, by the combination of different OH-metabolites, the bio effective PCB concentration was reduced and the respective effects on hTERT expression could be increased. At a concentration which showed no toxic activity in MTT assay, hTERT inhibition reflected the interference of OH-PCBs with the mitochondrial respiratory chain, which could lead to the production of reactive oxygen species (ROS). As individual OH-metabolites already showed a much stronger inhibition of hTERT gene expression at a lower concentration than their parental compounds, the hTERT gene expression bioassay described in this study seems to indicate metabolic activation of LC PCBs rather than the mere effect of LC PCBs on their own. In summary, this study provides dose-response linkages between effects of lower chlorinated PCBs and their concentrations in human plasma.

. Through autoxidation (autocatalytic oxidation in the presence of oxygen), peroxidases (POX) or the prostaglandin-H synthase (PHS), catechols and hydroquinones can be further modified to ortho-or para-quinones. These reactive arene oxides can covalently bind to biological macromolecules and form protein-, RNA-and DNA-adducts 14 . The electron transfer associated with the oxidation of OH-PCBs can lead to the formation of free oxygen radicals. Genotoxicity, mediated by adduct formation and generation of reactive oxygen species (ROS) is regarded as an important mechanism for the carcinogenic effect of LC PCBs 15 . Several studies have shown that PCBs can influence telomerase (hTERT) and telomere length (TL) in vivo and in vitro. Leukocyte telomere length were prolonged in individuals after long-term exposure to low doses of non orthochlorinated PCBs and DL PCBs [16][17][18] . In contrast to these observations, in vitro studies with individual airborne PCB congeners showed a reduced TL and hTERT enzyme activity in liver and keratinocyte cell lines 13,19 . In our previous experiments 20 , we found a highly significant reduction of age-adjusted TL in peripheral blood lymphocytes of high-level PCB exposed individuals (∑ indicator PCBs up to 236 µg/L plasma 21 ) where a dependency of telomere shortening on the exposure level of LC PCBs could be shown. Telomere shortening in lymphocytes can lead to replicative senescence, a terminal state which correlates with the progressive loss of immunological memory, a phenomenon known as immunosenescence. Accelerated onset of immunosenescence as seen for instance in chronic viral infections, compromises T-cell mediated immunity 22 . Immunosenescence has been related to higher infection rates, cancer and age-related diseases, and these events have also been associated with the exposure to PCBs in the past 2,23,24 . Accelerated telomere shortening in lymphocytes could thus contribute to the immunosuppressive activity of PCBs and explain some of the longterm effects of PCB exposure on the adaptive immune system.
Here we address the question to what extent the inhibition of hTERT, the telomere extending enzyme, can provide specific information about the extent of PCB contamination and whether it can be used to draw conclusions on PCB metabolic activation and PCB induced adverse health effects. We therefore made use of blood plasma samples of two characterized cohorts of PCB exposed individuals and focused on the WHO indicator congeners, 2,4,4′-trichlorobiphenyl (PCB 28) and 2,2′,4,5,5′-pentachlorobiphenyl (PCB 101). Furthermore we extended findings from blood plasma samples in a model cell culture system using hydroxylated metabolites of PCB 28 and PCB 101.

Results
Inhibition of hTERT gene expression by blood plasma of PCB exposed individuals. We have previously shown, that blood plasma samples containing high concentrations of PCB 28 inhibit the expression of telomerase (hTERT), the telomere-extending enzyme 20 . In order to further evaluate these findings in depth we made use of tert + B6B5.1 cells, targeted with a transgenic bacterial artificial chromosome (BAC) reporter containing a Renilla (Rluc) and a Firefly (Fluc) luciferase expression cassette under the control of the hTERT and CRR9 promoter 25 . The resulting ratio of Rluc to Fluc activity has been shown to be a reliable assessment of hTERT promoter activity 25 . When incubated with longitudinally collected blood plasma samples from individuals with a high PCB body burden due to occupational exposure (HELPcB-cohort) 21 , the capability of plasma to inhibit hTERT gene expression in tert + B6B5.1 cells was reduced over time, whereas plasma samples collected in 2011 inhibited hTERT expression by a factor of 7 (mean Rluc/Fluc-ratio = 0.13) as compared to untreated controls (mean Rluc/Fluc-ratio = 0.92), the suppressive effect of plasma was less pronounced in 2015 (mean Rluc/Fluc-ratio = 0.21; 4 fold inhibition as compared to controls) (Fig. 1A). Similar results were obtained when tetanus toxoid stimulated peripheral blood mononuclear cells (PBMCs) were incubated with blood plasma from the HELPcB-cohort and hTERT mRNA levels were assessed by quantitative RT-PCR (qRT-PCR). Here plasma samples collected in 2011 caused a significantly stronger inhibition of hTERT expression levels (mean hTERT/18S rRNA-ratio = 0.1) than the corresponding plasma samples collected in 2015 (mean hTERT/18S rRNA-ratio = 0.16) (Fig. 1B). In contrast, blood plasma collected from a cohort of residents exposed to PCBs via indoor air scarcely influenced hTERT gene expression (mean Rluc/Fluc-ratio = 0.71) as compared to controls (Fig. 1A). Corresponding mean PCB levels in the plasma of the individuals were 1.44 μg/L for the sum of higher chlorinated (HC) PCBs, 0.218 μg/L for the sum of dioxin-like (DL) PCBs and 0.109 μg/L for the sum of lower chlorinated (LC) PCBs (Fig. 1C). With the exception of the sum of LC PCBs, residents therefore did not show significantly elevated blood PCB levels compared to the general population. In contrast, high internal exposures for PCBs have been described for individuals of the HELPcB-cohort 26 . Plasma PCB levels of longitudinally collected samples of this cohort showed no significant difference in the level of HC PCBs at two different time points (mean concentration HC PCB 2011 = 5.55 μg/L vs mean concentration HC PCB 2015 = 5.36 μg/L, corresponding to 3,4% decrease, p = 0.63,), whereas levels of DL PCBs (mean concentration DL PCB 2011 = 1.75 μg/L vs mean concentration DL PCB 2015 = 1.11 μg/L, corresponding to 36,6% decrease; p = 0.13) and LC PCBs (mean concentration LC PCBs 2011 = 0.95 μg/L vs mean concentration LC PCBs 2015 = 0.25 μg/L, corresponding to 73,7% decrease, p = 0.015) decreased over time (Fig. 1C). We therefore conclude that high PCB plasma levels inhibit hTERT gene expression in transformed tert + B6B5.1 cells as well as in antigen-driven lymphoproliferation (Fig. 1B). The inhibition of hTERT expression within plasma samples from the HELPcB-cohort decreases over time, paralleling the decrease in levels of DL and LC PCBs.
To further analyze the suppressive effects of blood plasma from the HELPcB-cohort, we pooled data from longitudinally collected samples and analyzed hTERT expression in plasma dependent on concentrations of LC, HC and DL PCBs. The extent of hTERT inhibition in tert + B6B5.1 cells exposed to plasma was dependent on both the PCB congener profile as well as the concentration of PCBs found in plasma of individuals (Fig. 1D). Whereas the plasma concentration of HC PCBs and DL PCBs had no effect on hTERT expression, a high plasma concentration of LC PCBs led to a more pronounced and significant inhibition of hTERT expression as compared to a low concentration of LC PCBs (Fig. 1D). We therefore conclude that the capability of PCB contaminated plasma from the HELPcB-cohort to inhibit hTERT expression depends on the concentration of LC PCBs and decreases with the elimination of LC PCBs due to biotransformation.  HELPcB-cohort into five groups according to plasma concentrations of LC PCBs ( Fig. 2A). Mean Rluc/Fluc-ratio of each group was statistically compared to the group of individuals without detectable levels of LC PCBs (mean Rluc/Fluc-ratio = 0.19) ( Fig. 2A). hTERT inhibition was significant at a minimum plasma concentration range between 0.5 and 2 μg/L LC PCBs (mean Rluc/Fluc-ratio = 0.1), representing the lowest observed adverse effect level (LOAEL) within the HELPcB-cohort ( Fig. 2A). In a dose-response curve we related the median plasma concentration of LC PCBs in each group to the affected fraction of B6B5.1 cells susceptible to hTERT inhibition. The resulting plasma concentration-response relationship for the five concentrations yielded a typical sigmoidal curve, with the entire range of response data considered (Fig. 2B). These results imply, that hTERT gene expression in tert + B6B5.1 cells allows for discriminating between high and low exposure levels of LC PCBs in human plasma samples.
hTERT inhibition and toxicity by activation of LC PCBs. A variety of PCB toxication processes involve or depend on the metabolism of parent PCBs or their metabolic progeny 6 . We have previously shown that PCB 28 forms four and PCB 101 forms two hydroxylated PCB metabolites in the plasma of PCB exposed individuals 27 . With respect to the inhibition of telomerase gene expression, this could mean that the cumulative effect of hydroxy-metabolites emanating from PCB 28 and PCB 101 contribute to the plasma level-response relationship of LC PCBs in the HELPcB-cohort. Therefore, each OH-PCB (structural formulas in Supplementary Fig. 1) as well as the parent compound was tested at three different concentrations to inhibit hTERT expression in human Jurkat T cells (Fig. 3A). Changes in hTERT expression levels between the control and treated Jurkat T cells were normalized to 18S mRNA levels and then calculated by the 2-ΔΔct method. In parent PCB 28 treated cultures, hTERT expression increased above control levels at all concentrations tested (10-100 µM), whereas the PCB 28 derived To see if OH-PCBs also display cytotoxic activity towards Jurkat T cells, we assessed the influence of PCB 28 and PCB 101 derived OH-PCBs on metabolic activity using MTT assays (Fig. 3B). When cells were treated with OH-derivatives 3-OHCB 28 and 3′-OHCB 28, as well as 3-OHCB 101 and 4-OHCB 101, a concentration of 100 µM reduced metabolic activity below 10% of control values. At a concentration of 50 µM, the inhibition of metabolic activity was variable between congeners and reached a level below 50% of control values. At a concentration of 10 µM, the influence of all congeners on metabolic activity was less pronounced. These results therefore suggest, that depending on concentration, the effects of hydroxylated PCBs on hTERT gene expression and metabolic activity were higher than those of parent PCBs. Assuming that in MTT assays the conversion of tetrazole to formazan depends on the number of viable cells, these results further imply, that the effects of individual OH-PCBs on hTERT gene expression in Jurkat T cells mostly reflect a lower number of living cells. However, it must be emphasized here that the concentrations used for these in vitro experiments are several log levels higher

DNA damage induced by activation of LC PCBs.
Our results on toxicity indicate that the tested OH-PCBs display a higher cytotoxicity than their parental compounds at high concentrations. Based on these findings and reports from others 15 , we asked if OH-PCB induced cytotoxicity is caused by DNA damage that occurred in treated cells. The generation of reactive PCB metabolites, ROS or PCB dysregulated proteins can have a major impact on the integrity of DNA. We therefore assessed the ability of 3-OHCB 28 and 3′-OHCB 28, the major metabolites in the HELPcB-cohort, to induce DNA double strand breaks (DSB). We exposed Jurkat T cells for 48 hours to 3-OHCB 28 and 3′-OHCB 28, after which DNA damage was determined with a neutral COMET assay, which measured double strand breaks (DSB) in individual cells. The olive tail moment was calculated and used to quantify DNA damage. As shown in Fig. 4A, both 3-OHCB 28 and 3′-OHCB 28 induced DSB, with DNA damage from 3′-OHCB 28 increasing dose-dependently. DNA fragmentation correlated especially for 3′-OHCB 28 with the duration of action and increase in concentration. A very early, indirect response to DSB in the cell is the phosphorylation of the histone H2Ax at serine 139 (γH2Ax), which happens immediately after the occurrence of DNA damage and can be detected as foci in the cell nucleus by using a phosphospecific antibody. The confocal microscopy images of Jurkat cells (Fig. 4B) show increased foci in the nuclei upon 3-OHCB 28 and 3′-OHCB 28 incubation. The images show that higher substance concentration correlated with increased focus formation. Since basal DNA damage also occurs in untreated cells, a distinction was made between those with more than two foci per nucleus and those with less than two foci per nucleus. The foci of at least 58 cells were counted for quantification. After differentiation of the cells into those containing more and less than two or five foci, the individual ratios were calculated in each approach. We determined more than five foci per cell to be severe DNA damage. The proportion of these cells is shown in Fig. 4C. Both OH-PCB metabolites caused an increased phosphorylation of H2Ax. The proportion of cells with more than two foci correlated with the concentration of both metabolites. As in the comet assay, 3′-OHCB 28 showed a significant influence on the DNA integrity in Jurkat T cells at a concentration of 10 µM (see Fig. 4C). Both metabolites also led to an increase in the number of cells with more than five foci and thus increased DNA damage. With the densitometric evaluation of immunoblots, we used another method to quantify the formation of phosphorylated H2Ax (Fig. 4D). γH2Ax (15 kDa) was detected in total protein lysates of Jurkat T cells by using a phosphospecific antibody. The relative H2Ax phosphorylation was calculated upon the detection of ß-actin (42 kDa). Subsequently, the relative H2Ax phosphorylations were normalized on the negative control. Jurkat T cells showed a 5.5-fold higher phosphorylation of H2Ax as compared to negative controls when incubated for 48 hrs. with 10 µM 3-OHCB 28. Similar results were found for 3′-OHCB 28. Both metabolites showed a clear

Combination of OH-PCBs inhibits hTERT expression without reduction of metabolic activity.
The concentrations of the OH-PCBs used for the in vitro experiments so far were several orders of magnitude higher than the concentrations measured in the plasma of individuals of the HELPcB-cohort 27 . In addition, when using single OH-PCBs derived from PCB 28 and PCB 101, a spsecific effect on hTERT expression without reduction of metabolic activity was not evident (Fig. 3A,B). We therefore assessed if the combination of the metabolites (OH-PCB mix) is able to potentiate the effects on hTERT gene expression in serial 10x dilutions (Fig. 5A).  are principally capable of matching the results obtained with blood plasma from the HELPcB-cohort ex vivo. As the effects of individual OH-PCBs on hTERT gene expression in Jurkat T-cells mostly reflected a lower number of living cells, we next assessed metabolic activity of Jurkat T cells incubated with the OH-PCB mix as described above (Fig. 5B). By using the OH-PCB mix and its dilutions, a significant inhibition of metabolic activity could be reached at 50 µM and 5 µM concentrations, respectively. Lower OH-PCB mix levels had no effect on metabolic activity in Jurkat T cells. Together with the experiments on inhibition of hTERT expression these results therefore show, that the combination of OH-PCBs inhibits hTERT expression at both toxic and subtoxic concentrations. In addition, a selective inhibition of hTERT at zero metabolic toxicity can be achieved by combining different OH-metabolites. MTT assays are dependent on mitochondrial respiration and indirectly serve to assess the cellular energy capacity of a cell 28 . The overall strong effect of OH-PCBs in the MTT assay prompted us to investigate if LC PCBs could influence the function of mitochondria, e.g. the production of ATP, without reducing metabolic activity.  Fig. 3. Mean ± SD of three different experiments each with a total of three replicates are shown. Statistical analysis was performed by an unpaired, two-tailed Student's t-test. Statistically significant differences are indicated (**P < 0.01; ****P < 0.0001). (B) Jurkat T cells were preincubated for 4 hrs. with 500 nM OH-metabolites mix. ATP production was measured as luminescent signal after adding ATP detection reagent from ToxGlo TM -assay. Mean ± SD of three replicates are shown. Statistical analysis was performed by an unpaired, two-tailed Student's t-test. Statistically significant differences are indicated (****P < 0.0001). (C) hTERT gene expression in the presence and absence of pyruvic acid. Jurkat T cells were treated with 500 nM OH-metabolites mix for 48 hrs. Where indicated, 24 hrs. before treatment start, pyruvic acid was added to the culture medium. hTERT expression was evaluated by RT-PCR as described. Mean ± SD of three different experiments each with a total of three replicates are shown. Statistical analysis was performed by an unpaired, two-tailed Student's t-test. Statistically significant differences are indicated (**P < 0.01). (D) Production of reactive oxygen species (ROS) in Jurkat T cells in the presence of OH-PCB mix. Jurkat T cells preincubated with 4,5-diaminofluorescein-2-diacetate (DAF-2DA) were treated with 500 nM OH-PCB mix (Σ OH-PCB = 615.4 µg/L) for 4 hours and ROS production was measured as fluorescence signal. Mean ± SD of three replicates are shown. Statistical analysis was performed by an unpaired, two-tailed Student's t-test. Statistically significant differences are indicated (*P < 0.05). Jurkat T cells were therefore incubated with a 500 nM (Σ OH-PCB = 615.4 μg/L) OH-PCB mix. With this concentration no effect on metabolic activity could be observed, but the inhibition of hTERT expression was still evident (Fig. 5A). After 1 hour of incubation, ATP production levels had dropped around 50% compared to control treated samples (Fig. 5B), implying that even at a low dose, OH-PCBs interfere with the mitochondrial respiratory chain and ATP synthesis. Since it is not clear, whether mitochondrial dysfunction causes the inhibition of hTERT expression, or whether hTERT inhibition is an independent event in LC PCB toxication, we next assessed if PCB mediated hTERT inhibition and mitochondrial damage are intertwined. Therefore Jurkat T cells were preincubated for 24 hours with 1 mM pyruvic acid, the key intermediate of several energy producing process reactions within mitochondria. After incubating cells for a further 48 hours with the 500 nM OH-PCB mix, hTERT gene expression was evaluated. Interestingly, pyruvic acid prevented the OH-PCB mix induced hTERT inhibition at 500 nM (Fig. 5C). Mitochondrial dysfunction and inhibition of hTERT expression induced by activation of LC PCBs seemed therefore to be part of the same energy stress-induced pathway. For OH-metabolites derived from PCB 28 (3 chlorine atoms), oxidation to quinones or semiquinones is possible. A reactive semiquinone which, in the presence of O 2 , can form a reactive oxygen species (ROS), such as superoxide anions, could thus cause oxidative stress. In addition, OH-PCBs could interfere with the regulation of both intra-mitochondrial ROS production and disposal 10 . For that reason, we took a look at ROS production in the presence and absence of the 500 nM OH-PCB mix. A significant increase in ROS production in the presence of the mixture of OH-metabolites in comparison to the solvent control was observed (Fig. 5D). In summary these results show that similar to dioxin-like coplanar PCBs causing oxidative stress in the brain, PCB 28 and PCB 101 derived OH-PCBs are potent promoters of oxygen radical formation, which together with their derivatives can oxidize and alter important functional macromolecules within the cell.

Discussion
The effect of PCBs on humans is very heterogeneous and depends on various factors such as the congeners involved, cell-types affected, the degree of exposure and genetic predisposition 6 . Low-chlorinated PCBs are more reactive than HC PCB and DL PCB and their metabolism can lead to intermediates (arene oxides) with a higher toxicity than their parent compounds. In addition, LC PCBs are suspected of causing mutations in genetic material 29,30 . Low-chlorinated, volatile PCBs from open systems (e. g. joint sealants) reach the air through permanent emission and can thus be incorporated into the human body by inhalation 31 . In addition, when replaced by recycling companies, improper handling and non-compliance with hygiene regulations can result in a considerable LC PCB impact on employees and their relatives 32 .
In this paper we report a dose-response linkage between the effects of LC PCBs and their concentrations in human plasma. Using hTERT gene expression bioassays we establish the highest doses at which no toxic effects were identified and the lowest doses at which toxic or adverse effects were observed. Usually, the three indicator congeners for LC PCBs (PCB 28, 52, 101) and three indicator congeners for the HC PCBs (PCB 138, 153, 180) are analyzed in human plasma samples. This selection is the established surrogate for estimating the internal PCB burden, but does not focus on toxicological aspects. Until now, plasma concentration-effect relationships for LC PCBs could never be analysed and only biological reference values were being specified. For instance, the German MAK collection on Occupational Health and Safety established a biological tolerance value at the work place (BAT value) at 15 μg/L for the combined plasma levels of the PCB congeners 28, 52, 101, 138, 153, and 180 33 . With the hTERT gene expression bioassay defining a NOAEL at 0.2-0.5 μg/L and a LOAEL between 0.5 and 2 μg/L LC PCB, we believe, that verifiably adverse health effect levels of LC PCBs begin to occur at a concentration of ≥ 0.5 µg/L blood plasma.
Our in vitro experiments using Jurkat T cells suggest, that the oxidation of LC PCBs to reactive OH-intermediates is a prerequisite for the effects on hTERT expression and cellular integrity. In fact, individual OH-metabolites derived from PCB 28 and PCB 101 (OH-metabolites of PCB52 are not defined in humans) already show a much stronger inhibition of hTERT gene expression at a lower concentration than their parental compounds (Fig. 3A). This can be further enhanced by the combination of the metabolites 3-OHCB 28 and 3′-OHCB 28, as well as 3-OHCB 101 and 4-OHCB 101 (Fig. 5A).
As we have observed significant positive correlations between PCB 28 and its metabolites in the plasma of the HELPcB cohort 12 , the hTERT gene expression bioassay described in this study seems to indicate metabolic activation of LC PCBs rather than the mere effect of LC PCBs on their own. Therefore the dependence of the production of reactive OH-metabolites on the level of LC PCB exposure should be taken into account in future dose-response assessments. For instance the concentration of LC PCB at which saturation of the involved cytochrome P450-enzyme begins should be clarified and attention should be paid to the fact, that some OH-PCBs show higher half-lives than their parental PCB 12 . Moreover, CYP450 monooxygenases are in fact characterized by more frequent polymorphisms 34 , which lead to stronger or weaker active or inactive variants of the enzymes. Some PCB exposed individuals therefore might be more rapid converters of LC PCBs than others.
Several bioassays for the detection of PCBs have been described in the literature. Samples suspected to contain polychlorinated dibenzo-para dioxins (PCDD), dibenzofurans (PCDF) or DL PCBs can be screened with EROD activity based and CALUX test systems, respectively. In both bioassays, DL PCBs bind to the cellular aryl hydrocarbon (Ah) receptor 35 . This leads to an induction of the CYP1A1 gene in the EROD assay. The 7-etoxyresorufin-o-deethylase (EROD) formed in this process, which is associated with the CYP1A1 enzyme, converts the added substrate 7-ethoxyresorufin into a measurable fluorescent resorufin. The CALUX DR assay is a reporter gene assay in which a DNA construct consisting of a dioxin response element (DRE) promoter and a downstream luciferase reporter gene have been stably integrated into the genome of the rat liver cell line H4IIE 35 . The observed luciferase expression is directly proportional to the binding of dioxin to the Ah receptor. Whereas EROD and CALUX assays are able to measure contamination within a diverse array of applications (e.g. sediment, soil, water, tissue, food, feedstuff), the hTERT gene expression bioassays described in this study, report promoter activity due to the metabolic activation of a pollutant (LC PCB) to reactive intermediates. Most importantly, these assays are able to generate signals from an amount of toxic OH-PCBs which is too low to induce any measurable effect on metabolic activity (see Fig. 5A). Whereas many in vitro studies on LC PCBs and their metabolites suffer consistently from the very high concentrations used, which are in range of cytotoxic concentrations, this assay is able to detect genotoxic substances at a concentration range that is actually also reached in humans.
With the assays described in this study we add a new tool for the differentiated analysis of PCB congeners and their metabolites. Our results thus contribute to the risk assessment of PCB-related diseases. In addition to the general assessment of the exposure to PCBs and their toxic potency, the knowledge of interindividual differences in metabolizing capacity with respect to PCBs allows to include an individual component of the person concerned in the assessment. It is well established that mammalian telomerase expression is controlled at multiple levels with transcription factor binding and regulation of chromatin playing key roles 36 . It would therefore not come as a surprise that other plasma components of the HELPcB cohort intervene in these regulation mechanisms and therefore increase the inhibitory effect of LC PCBs. Both HC PCB and DL PCB have been shown to interfere with telomerase activity and the telomere complex 19 , which could enhance the effects of LC PCB derived OH-metabolites on hTERT inhibition. In addition, endogenous substances, whose production or release is induced by PCBs, can have an effect on the expression of hTERT. Thus, prostaglandins reduce the activity of telomerase and accelerate cellular ageing 37 . LC PCBs have been shown to increase the production of prostaglandins in vitro and in vivo 38,39 and thus could indirectly inhibit telomerase hTERT gene expression. In fact, prostaglandin levels in the plasma of the HELPcB cohort are increased by a factor of 4 compared to control plasma samples (data not shown).
PCBs have been shown to induce mitochondrial dysfunction and oxidative stress in vitro 40 and upon application in vivo, reduce the surfaces and volume density of mitochondria while maintaining their numbers at a constant rate. As a consequence, the energy supply of mitochondria were reduced, as could be demonstrated by incubating Jurkat T cells with OH-PCBs at a concentration of 500 nM (Fig. 5B). Mitochondrial dysfunction can also be induced by glucose restriction 41 , which modulates hTERT gene expression by DNA methylation changes and/or histone remodeling 42 . Thus, inhibition of hTERT expression by OH-metabolites of LC PCBs could reflect a reduced activity of the mitochondrial respiratory chain. This assumption is supported by the protective effect of pyruvic acid, which was added to the culture medium. Pyruvic acid as a key regulator of metabolism and aerobic energy production, prevented hTERT inhibition by the OH-PCB mix most likely by restoration of oxidative phosphorylation and aerobic glycolysis. This could provide evidence against a mechanism involving covalent binding of OH-PCB derived quinones with cytochrome c, an important electron acceptor within the mitochondrial respiratory chain 43 . In contrast to that, oxidation of the PCB 28 metabolites 3-OHCB 28 and 3′-OHCB 28 to quinones or semiquinones would be a prerequisite for DNA adduct formation as observed for PCB 28 after incubation of isolated DNA with rat microsomes 44 .

Material and Methods
Study groups. Blood samples from the PCB-exposed HELPcB-cohort were collected in longitudinal studies as described previously 27,45 . Blood samples from rresidents exposed to PCBs via indoor air were collected between September 2010 and March 2014 as part of human biomonitoring studies. Exposure to lower chlorinated PCBs within the resident-cohort was verified by air monitoring in respective buildings, which partially exceeded a preventive threshold value of 300 ng/m 3 air. Studies were carried out according to the Helsinki Declaration recommendations, participation was voluntary and all participants provided their written informed consent regarding the donation of blood and urine samples for scientific purposes. An approval by the ethical commission of the RWTH Aachen University is available (Nr. EK 176/11). Details of the internal PCB burden for each congener and the sum of non-dioxin-like PCB for the HELPcB-cohort and the resident-cohort are shown in Supplementary  Tables 1-3 (50 μg/mL). Cells were passaged using Trypsin-EDTA (0.25%) once a week. For incubation with blood plasma from PCB exposed individuals, plasma samples were taken from frozen stocks, thawed and filtrated. Cells were cultured in a standard medium for 48 hrs. in the presence of 20% (vol/vol) human plasma. Plasma samples collected in 2011 and 2015 from the same individual were tested in parallel.

Peripheral blood mononuclear cell (PBMC) separation and lymphoproliferation assay. Culture
and expansion of primary cytotoxic T cells was performed as described previously 20 . Forty milliliters of peripheral blood were collected after informed consent from healthy donors. PBMCs were enriched by density gradient centrifugation at 500 g for 45 minutes at room temperature using Percoll Hypaque. After two rounds of washing with phosphate-buffered saline (PBS), cells were put in a culture with a concentration of 1 × 10 7 cells per 2 mL of medium. The medium used throughout the experiments was RPMI 1640-GlutaMAX ™ -I supplemented with 10% (vol/vol) fetal calf serum (FCS), penicillin (50 U/mL) and streptomycin (50 μg/mL). T cells were stimulated in the presence of interleukin 2 and 7 (IL-2; IL-7; 40 IU/mL; PeproTech Inc, Rocky Hill, CT, USA) and tetanus toxoid (3 μg/mL, Statens Serum Institute, Copenhagen, Sweden). After 5 days of incubation cells were harvested, reseeded and used for experimentation. Comet-assay. For neutral comet assays, Jurkat T cells were treated with parent PCBs or single OH-PCBs.
After treatments 50 μL of a suspension of 10 5 cells/mL were washed in PBS and resuspended in 500 μl of preheated agarose (37 °C). The agarose/cell suspensions were applied to microscopic slides and left for 30 minutes at 4 °C to allow gel polymerisation. After cell lysis using a commercial buffer (Trevigen, Gaithersburg, USA), single cell gelelectrophoresis was performed at 1 V/cm for 1 hour. After precipitation of the DNA with ammonium acetate, samples were fixed in 70% ethanol for 30 minutes and dried. To visualize dsDNA SYBR ™ Green I Nucleic Acid Gel Stain (Invitrogen, Carlsbad, USA) was used. Pictures of at least 50 cells/section were taken using a fluorescence microscope (DMRX, Leica: 450-490 nm BP-filter, 40x magnification). The extent of the genotoxic effect of PCB metabolites was calculated using the ImageJ-based OpenComet software. Statistical analysis. Statistical analysis was done by a Wilcoxon matched-pairs signed rank test for telomerase gene expression assays performed with PCB plasma samples from 2011 and 2015 ( Fig. 1A-C). Analysis of concentration-dependent differences in telomerase gene expression was done by a Mann-Whitney test (Figs 1D and 2A). For all other data sets two-tailed, unpaired Student's t-tests were performed. Where indicated, SCIEnTIFIC RePoRTS | (2018) 8:16903 | DOI:10.1038/s41598-018-35043-w differences were investigated using Kruskal-Wallis followed by a Dunn's post-hoc test for multiple comparisons (Fig. 4A). All statistical tests performed are designated in the respective figure legends.

Data Availability
The data generated and/or analysed during the current study are available from the corresponding author on request.