Environment International 115 (2018) 295–300

Contents lists available at ScienceDirect

Environment International

journal homepage: www.elsevier.com/locate/envint

Circulating levels of perfluoroalkyl substances and left ventricular geometry T of the heart in the elderly ⁎ Ingrid Mobackea, Lars Lindb, Linda Dundera, Samira Salihovicc, P. Monica Linda, a Department of Medical Sciences, Occupational and Environmental Medicine, Uppsala University, Uppsala, Sweden b Department of Medical Sciences, Cardiovascular Epidemiology, Uppsala University, Uppsala, Sweden c Department of Medical Sciences, Molecular Epidemiology and Science for Life Laboratory, Uppsala University, Uppsala, Sweden

ARTICLE INFO ABSTRACT

Keywords: Aims: Some persistent organic pollutants (POPs) such as hexachlorobenzene (HCB) and some polychlorinated Per- and polyfluoroalkyl substances (PFASs) biphenyls (PCBs) have been shown to interfere with myocardial function and geometry. We therefore in- Heart vestigated if also another group of POPs: per- and polyfluoroalkyl substances (PFASs) were associated with Left ventricular geometry alterations in left ventricular geometry. Elderly Methods: 801 subjects aged 70 years were investigated in a cross-sectional study within the scope of the Environmental contaminants Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) study. Eight PFASs were detected Perfluorononanoic acid (PFNA) in > 75% of participants´ plasma by ultra-performance liquid chromatograph/tandem mass spectrometry. Left ventricular geometry was determined by echocardiography. Multivariable linear regression was used to in- vestigate the associations between PFASs and left ventricular geometry of the heart after exclusion of subjects with previous myocardial infarction (n = 72). Results: When adjusting for multiple comparisons, none of the eight PFASs evaluated were significantly related to left ventricular mass. However, perfluorononanoic acid (PFNA), perfluorodecanoic acid (PFDA), and per- fluoroundecanoic acid (PFUnDA) were related to relative wall thickness (RWT) in a negative fashion (p < 0.0021). Besides being inversely related to RWT, PFNA was also positively related to left ventricular end- diastolic volume (LVEDD) (p < 0.0021). These analyses were adjusted for traditional cardiovascular risk fac- tors. Conclusion: In this cross-sectional study, several of the PFASs evaluated, especially PFNA, were related to myocardial geometry: a reduction in relative wall thickness and an increase in left ventricular diameter fol- lowing adjustment for traditional cardiovascular risk factors, suggesting a role for PFASs in cardiac remodeling.

1. Introduction epidemiological studies on the effects of PFASs on the adult heart are lacking. Although a few animal studies on developmental cardiotoxicity It has been known for many years now that per- and polyfluoroalkyl have been performed, studies on humans are needed to confirm the substances (PFASs) can be detected in human and environmental results of animal studies and to help restrict the use of substances ha- samples all over the world, from the Great Lakes in North America to zardous to human health. the Pearl River in China (Lau et al., 2007; Boulanger et al., 2004; So In humans, remodeling of the left ventricle of the heart is associated et al., 2007; Stubleski et al., 2016). PFASs have highly desirable prop- with increased morbidity and mortality in cardiovascular disease erties and are therefore used in a vast number of products such as (CVD), besides being an independent risk factor for death due to any textiles, paper and food packaging, cosmetics, non-stick products and cause (Ghali et al., 1992; Koren et al., 1991). There are a number of electric equipment (Buck et al., 2011). However, knowledge about well-known risk factors for left ventricular hypertrophy (LVH) such as possible adverse health effects from PFASs exposure is scarce, and hypertension, diabetes, hyperlipidemia and obesity (Levy et al., 1988;

Abbreviations: BMI, Body mass index; CVD, Cardiovascular disease; CAR, Constitutive androstane receptor; GFR, Glomerular filtration rate; HCB, Hexachlorobenzene; HDL, High density lipoprotein; IVS, Inter ventricular septum; LDL, Low density lipoprotein; LVEDD, Left ventricular end diastolic diameter; LVM, Left ventricular mass; LVMI, Left ventricular mass index; PFASs, Per- and polyfluoroalkyl substances; PIVUS, The Prospective Investigation Of The Vasculature In Uppsala Seniors; POP, Persistent organic pollutants; PPAR, Peroxisome pro- liferator-activated receptors; PW, Posterior wall thickness; RWT, Relative wall thickness; PXR, Pregnane X-receptor ⁎ Corresponding author at: Department of Medical Sciences, Occupational and Environmental Medicine, Uppsala University, 751 85 Uppsala, Sweden. E-mail addresses: [email protected] (I. Mobacke), [email protected] (L. Lind), [email protected] (L. Dunder), [email protected] (S. Salihovic), [email protected] (P.M. Lind). https://doi.org/10.1016/j.envint.2018.03.033 Received 1 December 2017; Received in revised form 2 March 2018; Accepted 21 March 2018 Available online 03 April 2018 0160-4120/ © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/). I. Mobacke et al. Environment International 115 (2018) 295–300

Saunders et al., 2008; Ho et al., 1993). However, LVH is multicausal, Table 1 and less well-known risk factors can play a significant role in the de- Basic characteristics and major cardiovascular risk factors of the subjects velopment of LVH. Recently, some persistent organic pollutants (POPs), (n = 801). such as hexachlorobenzene (HCB) and some polychlorinated biphenyls Women/men (%) 417/384 (52) (PCBs), have been shown to be associated with increased wall thickness of the left ventricle, ventricular remodeling, and impairments in left High degree of education, n (%) 200 (25) ventricular systolic and diastolic function (Sjoberg Lind et al., 2013a; Current smoking, n (%) 80 (10) Regular physical activity, n (%) 216 (27) Sjoberg Lind et al., 2013b). Previous studies have reported associations Waist circumference (cm) 90.91 ± 11.6 between PFASs exposure and developmental cardiotoxicity in chicken Fasting blood glucose (mmol/l) 5.29 ± 1.5 (Jiang et al., 2012), as well as atherosclerosis in the elderly (Lind et al., Systolic blood pressure, (mm Hg) 149.74 ± 22.53 2017; Jiang et al., 2012). Given these results, and considering that HDL-cholesterol (mmol/l) 1.53 ± 0.43 fi LDL-cholesterol (mmol/l) 3.42 ± 0.86 PFASs are also classi ed as POPs, it is desirable to investigate whether Serum triglycerides (mmol/l) 1.27 ± 0.59 also PFASs affect the geometry of the adult human heart. Therefore, the BMI (kg/m2) 26.96 ± 4.31 aim of the present cross-sectional study was to investigate the asso- Energy consumption per day (kcal) 1883.4 ± 464.9 ciation between circulating levels of PFASs and alterations in left ven- Alcohol consumption, g/day 2.55 ± 2.81 tricular geometry in a 70-year-old population in Sweden in the Pro- Antihypertensive treatment, n (%) 240 (30) PFHpA ng/ml (SD) 0.7 ± 0.7 spective Investigation of the Vasculature in Uppsala Seniors (PIVUS) PFHxS 3.41 ± 3.64 study. L-PFOS 14.9 ± 8.88 PFOA 3.59 ± 1.69 2. Method PFNA 0.8 ± 0.43 PFDA 0.34 ± 0.15 PFOSA 0.14 ± 0.14 2.1. Participants and sampling - PIVUS study PFUnDA 0.31 ± 0.14 LVMI (g/m2.7) 42.48 ± 12.79 Data was collected from the Prospective Investigation of the LVEDD (mm) 46.79 ± 5.35 Vasculature in Uppsala Seniors (PIVUS) study. Participants were RWT 0.44 ± 0.09 70 years old, and just over half were female (52%). For details on Values expressed as n, means ± SD or medians (25th to 75th percentile). − subjects and basic investigation, see (Lind et al., 2005). In total, 1016 The mean concentration of PFASs is given as SD, ng mL 1. High degree of subjects participated in the study. Participants with previous myo- education is defined as > 9 years of education. Regular physical activity is cardial infarction were excluded from this present study (n = 72), since defined as moderate or hard physical activity > 1 time per week. HDL, High a myocardial infarction can alter left ventricular geometry. Following density lipo-protein; LDL, Low density lipo-protein. PFASs, Per- and poly- exclusion (previous heart attack) and deficient quality of ultrasound fluoroalkyl substances; Perfluoroheptanoic acid (PFHpA), Perfluorohexane and information on confounders, the final number of subjects used for sulfonic acid (PFHxS), Linear isomer of perfluorooctane sulfonic acid (L- fl fl statistical calculations was 801. Informed consent was given by all PFOS), Per uorooctanoic acid (PFOA), Per uorononanoic acid (PFNA), fl fl participants, and the study was approved by the Ethics Committee of Per uorodecanoic acid (PFDA), Per uorooctane sulfonamide (PFOSA), Perfluoroundecanoic acid (PFUnDA). Heart geometry: LVMI, Left ventricular Uppsala University. mass indexed for height2.7; LVEDD, Left ventricular end diastolic diameter, Basic characteristics were collected, presented in Table 1. To eval- RWT, Relative wall thickness of the left ventricle. uate alcohol and energy consumption, a seven-day recording of all food and drinking items consumed by the participants was performed. All of LVEDD). To determine the left ventricular mass (LVM), the Penn con- the above baseline information was collected using standardized vention was applied, and thereafter indexed for height2.7 to retrieve left fi methods (self-reported questionnaires lled in at home right before the ventricular mass index (LVMI). A technician unaware of other collected examination, fasting plasma samples for determinations of glucose and data performed the evaluation of the left ventricular data. lipids, and clinical examination). Glomerular filtration rate (GFR) was estimated by measurements of serum creatinine using the MDRD fi (Modi cation of Diet in Renal Disease) formula (Levey et al., 1999). 2.4. Statistics

2.2. Analysis of PFASs To obtain normal distributions of the skewed PFASs, the data was log transformed. To evaluate the relationships between the different Details on the analysis of PFASs have previously been described PFASs and left ventricular geometry, linear regression analysis was used (Salihovic et al., 2013). A total of eight PFASs for which > 75% of the with LVMI, RWT and LVEDD as dependent variables in different models study population showed measurable levels above the lower level of and the PFASs as independent variables in separate analyses for each fl detection were evaluated in the present study: per uoroheptanoic acid PFASs. As a first step the regression model was adjusted for sex only fl (PFHpA), per uorohexane sulfonic acid (PFHxS), linear isomer of per- (same age for all subjects). Thereafter multiple adjustments were made fl fl uorooctane sulfonic acid (L-PFOS), per uorooctanoic acid (PFOA), for classical risk factors for LVH and life-style factors (sex, blood pres- fl fl per uorononanoic acid (PFNA), per uorodecanoic acid (PFDA), per- sure, antihypertensive medication, high density lipoprotein (HDL) and fl fl uorooctane sulfonamide (PFOSA), per uoroundecanoic acid low density lipoprotein (LDL), cholesterol, blood glucose, waist cir- (PFUnDA). These PFASs were measured in serum using ultra-perfor- cumference, triglycerides, body mass index (BMI), education levels, mance liquid chromatography coupled to tandem mass-spectrometry exercise habits, smoking, energy consumption and alcohol intake). As (UPLC-MS/MS). an additional step, the squared form of the PFAS data was included to analyze potential non-linear relationships. Furthermore, an interaction 2.3. Echocardiography term between the PFASs and sex was included to investigate possible differences between the sexes. Stata 14 (Stata, College Station, TX, For details on performance of the echocardiography see Sjoberg USA) was the statistical software used for the calculations. The critical Lind et al. (2013a). The M-mode measurements performed were: in- p-value for significance was set to 0.0021 according to the Bonferroni terventricular septum (IVS), posterior wall thickness (PW), and left correction (0.05/8 PFASs/3 outcomes). ventricular end diastolic diameter (LVEDD). From these values, the relative wall thickness (RWT) was calculated (RWT = (IVS + PW)/

296 I. Mobacke et al. Environment International 115 (2018) 295–300

Table 2 Linear relationships between circulating levels of eight PFASs (all log-trans- Predictive Margins with 95% CIs for LVEDD (mm) formed) and left ventricular mass (LVMI), relative wall thickness (RWT) and left 47 ventricular end-diastolic volume (LVEDD).

Sex-adjusted Multiple adjusted 46

Variable Beta (95% CI) p-Value Beta (95% CI) p-Value 45 LVMI PFHpA −0.5 (−1.64, 0.64) 0.39 −1.11 (−2.05, −0.17) 0.02

PFHxS −1.18 (−2.23, −0.13) 0.03 −0.23 (−1.11, 0.65) 0.61 44 L-PFOS −0.45 (−1.05, 0.96) 0.93 0.12 (−0.73, 0.97) 0.78 Linear Prediction PFOA −1.80 (−3.32, −0.27) 0.02 −0.65 (−1.94, 0.65) 0.33

PFNA 0.85 (−0.62, 2.32) 0.26 0.43 (−0.82, 1.68) 0.50 43 PFDA −2.02 (−4.09, 0.06) 0.06 −1.22 (−3.03, 0.59) 0.19 PFOSA −1.79 (−3.18, −0.39) 0.12 −1.13 (−2.26, 0.01) 0.52 PFUnDA −1.79 (−3.70, 0.12) 0.07 −0.55 (−2.27, 1.17) 0.53 42 -3 -2.5 -2 -1.5 -1 RWT PFNA (ln-transformed, ng/ml) PFHpA 0.003 (−0.004, 0.01) 0.37 −0.001 (−0.01, 0.01) 0.75 PFHxS −0.01 (−0.01, 0.002) 0.13 −0.002 (−0.01, 0.01) 0.52 − − − − L-PFOS 0.01 ( 0.01, 0.001) 0.02 0.01 ( 0.01, 0.03 Fig. 1. Relationship between circulating levels of PFNA (on a ln-scale, ng/ml) − 0.001) and left ventricular end-diastolic volume (LVEDD) illustrated with predictive PFOA −0.01 (−0.02, 0.01 −0.12 (−0.22, 0.03 margins (p = 0.0006). −0.003) −0.001) PFNA −0.01 (−0.02, 0.03 −0.02 (−0.03, −0.01) 0.0018 −0.001) PFDA −0.02 (−0.03, −0.01) 0.003 −0.02 (−0.04, −0.01) 0.0009 Predictive Margins with 95% CIs for RWT PFOSA −0.01 (−0.02, 002) 0.12 −0.004 (−0.01, 0.01) 0.33 PDUnDA −0.02 (−0.03, −0.01) 0.001 −0.02 (−0.35, −0.01) 0.001 .52

LVEDD PFHpA −0.17 (−0.61, 0.26) 0.43 −0.26 (−0.70, 0.18) 0.25 .5 PFHxS −0.13 (−0.53, 0.28) 0.54 0.01 (−0.4, 0.42) 0.96 L-PFOS 0.45 (0.06, 0.83) 0.02 0.47 (0.08, 0.87) 0.02 PFOA 0.29 (−0.29, 0.88) 0.33 0.58 (−0.03, 1.18) 0.06

PFNA 0.95 (0.39, 1.51) 0.0009 1.01 (0.43, 1.58) 0.0006 .48 PFDA 0.49 (−0.31, 1.28) 0.23 0.95 (0.11, 1.79) 0.03 PFOSA −0.21 (− 0.73, 0.31) 0.43 −0.14 (−0.66, 0.38) 0.60 Linear Prediction PFUnDA 0.53 (−0.20, 1.25) 0.16 1.04 (0.25, 1.84) 0.01 .46

The results are adjusted for sex as well as for multiple factors (sex, systolic blood pressure, antihypertensive medication, high density lipoprotein (HDL) and low density lipoprotein (LDL)-cholesterol, blood glucose, waist cir- .44 cumference, triglycerides, body mass index (BMI), education levels, exercise -3 -2.5 -2 -1.5 -1 habits, smoking, energy and alcohol intake). Regression coefficient (Beta) is PFNA (ln-transformed, ng/ml) given with p-values and 95% CI for the following log-transformed PFASs: Perfluoroheptanoic acid (PFHpA), Perfluorohexane sulfonic acid (PFHxS), Fig. 2. Relationship between PFNA (on a ln-scale, ng/ml) and relative wall Linear isomer of perfluorooctane sulfonic acid (L-PFOS), Perfluorooctanoic acid thickness (RWT) illustrated with predictive margins (p = 0.0018). (PFOA), Perfluorononanoic acid (PFNA), Perfluorodecanoic acid (PFDA), Perfluorooctane sulfonamide (PFOSA), Perfluoroundecanoic acid (PFUnDA). Introducing quadratic terms for the PFASs in the models did not 3. Results disclose any significant relationships not found in the linear analyses. Further, no major low-dose effects or inverse U-shaped relationships General characteristics of the participants are presented in Table 1. were disclosed by this non-linear analysis. No significant interactions Table 2 and Figs. 1 and 2 show the results regarding the linear re- between sex and the PFASs regarding the indices of left ventricular lationships for all PFASs and heart measurements, and the results for geometry were found. PFNA levels in relation to LVEDD and RWT. When adjusting for mul- tiple factors (sex, blood pressure, antihypertensive medication, HDL 4. Discussion and LDL-cholesterol, blood glucose, waist circumference, triglycerides, BMI, education levels, exercise habits, smoking, energy and alcohol The results of this cross-sectional study demonstrate that elevated intake), none of the PFASs were significantly related to LVMI. However, levels of several PFASs are associated with increased LVEDD, as well as several of the PFASs evaluated were related to RWT in a negative decreased RWT in an elderly population in Uppsala. These associations fashion, with PFNA (p = 0.0018), PFDA (p = 0.0009) and PFUnDA were mainly found for the longer-chained PFASs. (p = 0.001) being significant in the multiple adjusted model. Additional To our knowledge, this present study is the first to investigate as- adjustment for GFR did only reduce the power of these relationships in sociations between levels of PFASs and left ventricular geometry in a minor fashion, being still significant (now p = 0.013 for PFNA, humans. A previous study using PIVUS data has reported a relationship p = 0.002 for PFDA and p = 0.002 for PFUnDA). between another environmental contaminant, HCB, and abnormal left Those PFASs being inversely related to RWT were also positively ventricular geometry (Sjoberg Lind et al., 2013a). The current study has related to LVEDD. However, PFNA (p = 0.0006) was the only PFAS expanded our knowledge on which POPs that might have adverse ef- with statistically significant associations in the multiple adjusted fects on heart geometry, now including several PFASs. Changes in left models, PFDA (p = 0.03), PFUnDA (p = 0.01). Additional adjustment ventricular geometry are a risk factor for CVD in general, and LVH is an for GFR did not alter these relationships in any substantial way (now established risk factor for heart failure (Ghali et al., 1992; Koren et al., p = 0.009 for PFNA, p = 0.02 for PFDA and p = 0.005 for PFUnDA). 1991). Interestingly, in the PIVUS-cohort, HCB and PFNA seemed to

297 I. Mobacke et al. Environment International 115 (2018) 295–300 have opposite effects on the heart, with HCB being linked to a con- heart geometry. This implies that the PPAR pathway might not solely be centric left ventricular remodeling and increased wall thickness responsible for the observed effects as has been suggested previously (Sjoberg Lind et al., 2013a). This whereas PFNA in this present study (Rosen et al., 2017). Additionally, the majority of mechanistic studies was related to dilation of the left ventricle and a reduction in relative have been made on PFOS and PFOA, therefore many of the potential wall thickness. In the case of HCB and PFASs, the adverse effects on the modes of action might not fully apply for the PFASs for which we ob- heart are potentially antagonistic, illustrating that the task of under- served significant associations in our study. standing the properties and adverse effects of environmental con- Risk factors in the development of cardiac hypertrophy and re- taminants is complex. modeling might be oxidative and nitrosative stress (Elahi et al., 2007). Nitric oxide is an important regulator of vasculature function, together 4.1. Potential mechanisms with arachidonic acid it regulates fatty acid oxidation and smooth muscle contraction. PFASs have been found to induce both oxidative The healthy heart mainly relies on free fatty acids as energy sub- and nitrosative stress in vitro, in vivo and in humans (Wang et al., 2017; strate, but also on free glucose to a smaller extent. In the failing heart Han et al., 2018; Wielsøe et al., 2015). there is an increase in the cardiomyocytes glucose utilization. Quite the Other possible mode of actions for PFASs are through the Bone contrary happens in failing hearts of patients with diabetic cardio- morphogenetic protein-pathway, Constitutive androstane receptor myopathy, whose cardiomyocytes increase their fatty acid uptake and (CAR) and Pregnane X-receptor (PXR), of which most have been shown reduce their glucose utilization. Changes seen in diabetic cardiomyo- to have a possible role either in the development or in the function of pathy are for instance left ventricular hypertrophy and dysfunction the heart in animals and in vitro (Abe et al., 2016; Jiang et al., 2013; Ren (features not unique for the diabetic heart). The Peroxisome proliferator et al., 2009). Another interesting receptor that might me relevant in the activated receptor-α (PPAR-α) is a nuclear receptor and is amongst context is the Aryl hydrocarbon-receptor, which is suggested to have a others known to regulate lipid metabolism. The gene pathways of the role in the development of CVD, and that has been shown to be acti- PPAR family have been implicated in the development of CVD (Mistry vated by PFASs (Lund et al., 2003; Long et al., 2013). and Cresci, 2010). Furthermore, PPAR-α has been reported to regulate Of particular interest regarding the role of PFASs on the heart is the substance preference in the heart of mice, where an overexpression of role of mitochondrial dysfunction. Mitochondrial dynamics and mor- PPAR-α resulted in an increased fatty acid uptake and a reduction in phology can be altered for instance through fission and fusion, pro- glucose utilization in the cardiomyocytes, similarly to the changes seen cesses controlled by several proteins. The role of fission and fusion in diabetic cardiomyopathy (Finck et al., 2002; Finck and Kelly, 2002). proteins seems to differ between the developing and the adult heart Several studies have shown that PFASs have the ability to stimulate (Ong et al., 2013). Furthermore, the movement of the mitochondria in PPAR-α in vitro and in mice (Wolf et al., 2012; Wolf et al., 2010). This the adult heart is much more limited since they are tightly packed, indicates that PFASs could possibly have a role in changes in the heart suggesting that the role of fission and fusion might be less important in similar to those seen in diabetic cardiomyopathy, mediated through the adult heart. However, cardiac ablation of the fusion proteins mi- PPAR-α (Finck et al., 2002; Wolf et al., 2012; Wolf et al., 2010). In tofusins 1 and 2 in the adult heart of mice has shown to result in severe chicken, developmental cardiotoxicity with PPAR-pathway involve- dilated cardiomyopathy (Chen et al., 2011). The interest for PFASs' ment has been seen after exposure to PFOA, with effects such as thin- possible mitochondrial toxicity is continuously gaining more attention. ning of the right ventricle wall, increased left ventricular wall thickness Mitochondrial pathways, together with a number of other possible and increased heart rate (Jiang et al., 2012). Although it is known that pathways such as nuclear lipid hyperaccumulation, may even play an several PFASs are peroxisome proliferators, it is hard to draw any even more important role than the more classical pathways such as conclusions concerning effects on the heart without having studied the activation of PPARs, CAR or PXR (Li et al., 2017). Studies on animals direct effect of PFASs on the PPARs in the adult heart. Furthermore, and humans show that mitochondrial pathways are important in the most studies on PPAR activation have been done with an exposure re- toxic effects of both PFOS and PFOA, possibly involving uncontrolled sulting in higher blood levels of PFASs than observed in our study. It is fission and fusion processes in the mitochondria (Han et al., 2018; Li however intriguing that the substance preference of the heart might et al., 2017; Quist et al., 2015). Rats prenatally exposed to PFOS have play a role in PFASs' cardiotoxicity. been shown to develop alterations in mitochondrial dynamics including The three PFASs in our study with a significant association to heart mitochondrial damage, adenosine triphosphate - production and mi- geometry are all classified as long-chain, since they have 9 carbons tochondria - mediated apoptosis (Xia et al., 2011; Zeng et al., 2015; (PFNA), 10 carbons (PFDA) and 11 carbons (PFUnDA), respectively. Kleszczynski et al., 2009). Furthermore, studies of stem-cell derived Long-chained PFASs have a greater potential for bioaccumulation into cardiomyocytes show that mitochondrial damage is important in PFOS' living organisms, compared to short-chained PFASs, which is most cardiomyocyte toxic effects (Tang et al., 2017). It is hypothesized that probably due to their long half-lives. The importance of chain length of there are 3 main pathways of mitochondrial toxicity caused by PFASs: PFAS has for example been demonstrated in previous studies showing uncoupling of oxidative phosphorylation, induction of mitochondrial that the longer the carbon chain, the greater the effect on PPAR-α (Wolf permeability transition and inhibition of mitochondrial respiration et al., 2012; Kudo et al., 2006; Rosenmai et al., 2018). However, the (Wallace et al., 2013). Just as with PPAR-α it has been noted that length of the carbon chain only seems to increase PPAR-α activity up to PFASs' toxic potency on mitochondria increases with its carbon length, 9 carbons, molecules longer than that have a lower ability of inducing through PFDA (10 carbons) (Wallace et al., 2013). effects on PPAR-α, and in one study PFDA did not seem to activate Since the clearance of the PFASs might be effected by kidney human PPAR-α at all (Wolf et al., 2012). Similarly in a recent study of function and since also left ventricular geometry has been found to be human liver cells, the highest induction of PPAR-α activity was ob- related to kidney function (Dai et al., 2017), we performed additional served after treatment with PFHxA, PFHpA, PFOA, PFNA but also adjustment for GFR, but did not find kidney function to influence the PFDA, moreover, in this study PFUnDA did not seem to activate PPAR-α PFASs vs RWT or PFASs vs LVEDD relationship in any major way given at all while even longer chain alkyls such as PFDoDA and PFTeDA did the other confounders used in the models. (Rosenmai et al., 2018). Due to the similarity in structure amongst many PFASs they are expected to produce their effects in similar ways 4.2. Limitations on PPAR-α and also having additive effects when combined (Wolf et al., 2014). It is therefore somewhat contradictory that the PFASs occurring A limitation to our study is that it was restricted to Caucasians aged at highest levels, see Table 1 (L-PFOS 14.9 ng/ml; PFOA 3.59; PFHxS 70, which means that the results cannot readily be extrapolated to other 3.41) in our study were not the ones with significant associations with ages or ethnic groups. However, given that our results were significant

298 I. Mobacke et al. Environment International 115 (2018) 295–300 after adjusting for several confounders, the risk is low that they are disease. Ann. Intern. Med. 117, 831–836. merely chance findings, but our results need to be replicated in addi- Han, R., HU, M., Zhong, Q., Wan, C., Liu, L., Li, F., Zhang, F., Ding, W., 2018. Perfluorooctane sulphonate induces oxidative hepatic damage via mitochondria-de- tional prospective studies and there is a need for toxicological testing to pendent and NF-kappaB/TNF-alpha-mediated pathway. Chemosphere 191, investigate the pathways responsible for the alterations in heart geo- 1056–1064. metry that were found in the present study. Prospective studies are less Ho, K.K.L., Pinsky, J.L., Kannel, W.B., Levy, D., 1993. The epidemiology of heart failure: fl the Framingham study. J. Am. Coll. Cardiol. 22, A6–A13. prone to reverse causation (the outcome in uence the exposure) than Jiang, Q., Lust, R.M., Dewitt, J.C., 2013. Perfluorooctanoic acid induced-developmental cross-sectional studies, and are therefore of greater interest. cardiotoxicity: are peroxisome proliferator activated receptor alpha (PPARalpha) and bone morphorgenic protein 2 (BMP2) pathways involved? J. Toxic. Environ. Health A – 4.3. Conclusion 76, 635 650. Jiang, Q., Lust, R.M., Strynar, M.J., Dagnino, S., Dewitt, J.C., 2012. Perflurooctanoic acid induces developmental cardiotoxicity in chicken embryos and hatchlings. Toxicology In conclusion, the results of this cross-sectional study demonstrate 293, 97–106. that elevated levels of PFNA, PFDA and PFUnDA were associated with a Kleszczynski, K., Stepnowski, P., Skladanowski, A.C., 2009. Mechanism of cytotoxic ac- tion of perfluorinated acids II. Disruption of mitochondrial bioenergetics. Toxicol. decreased RWT. PFNA was also associated with an increase in LVEDD. Appl. Pharmacol. 235, 182–190. The findings of this present study indicate that PFASs may play a role in Koren, M.J., Devereux, R.B., Casale, P.N., Savage, D.D., Laragh, J.H., 1991. Relation of altered left ventricular geometry, which subsequently can increase the left ventricular mass and geometry to morbidity and mortality in uncomplicated essential hypertension. Ann. Intern. Med. 114, 345–352. risk of later clinically manifested cardiac disorders. Kudo, N., Suzuki-Nakajima, E., Mitsumoto, A., Kawashima, Y., 2006. Responses of the liver to perfluorinated fatty acids with different carbon chain length in male and Disclosure statement female mice:in relation to induction of hepatomegaly, peroxisomal beta-oxidation and microsomal 1-acylglycerophosphocholine acyltransferase. Biol. Pharm. Bull. 29, 1952–1957. The authors have no competing interests to declare. Lau, C., Anitole, K., Hodes, C., Lai, D., Pfahles-Hutchens, A., Seed, J., 2007. Perfluoroalkyl acids: a review of monitoring and toxicological findings. Toxicol. Sci. 99, 366–394. Levey, A.S., Bosch, J.P., Lewis, J.B., Greene, T., Rogers, N., Roth, D., 1999. A more ac- Funding curate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of diet in renal disease study group. Ann. Intern. This study was supported by the Swedish Research Council for en- Med. 130, 461–470. vironment, Agricultural Sciences and Spatial Planning (FORMAS, Levy, D., Anderson, K.M., Savage, D.D., Kannel, W.B., Christiansen, J.C., Castelli, W.P., 1988. Echocardiographically detected left ventricular hypertrophy: prevalence and http://www.formas.se/en/, Grant number 2007-2047). This funding risk factors. The Framingham Heart Study. Ann. Intern. Med. 108, 7–13. source had no role in the study design, in writing the report, or in the Li, K., Sun, J., Yang, J., Roberts, S.M., Zhang, X., Cui, X., Wei, S., Ma, L.Q., 2017. fl decision to submit the paper for publication. Molecular mechanisms of per uorooctanoate-induced hepatocyte apoptosis in mice using proteomic techniques. Environ. Sci. Technol. 51, 11380–11389. Lind, L., Fors, N., Hall, J., Marttala, K., Stenborg, A., 2005. A comparison of three dif- Ethical approval ferent methods to evaluate endothelium-dependent vasodilation in the elderly. The Prospective Investigation of the Vasculature in Uppsala Seniors (PIVUS) Study 25, 2368–2375. The study was approved by the Ethics Committee of Uppsala Lind, P.M., Salihovic, S., Van Bavel, B., Lind, L., 2017. Circulating levels of perfluoroalkyl University, and the participants gave informed consent prior to the substances (PFASs) and carotid artery atherosclerosis. Environ. Res. 152, 157–164. study. Long, M., Ghisari, M., Bonefeld-Jorgensen, E.C., 2013. Effects of perfluoroalkyl acids on the function of the thyroid hormone and the aryl hydrocarbon receptor. Environ. Sci. Pollut. Res. Int. 20, 8045–8056. Acknowledgements Lund, A.K., Goens, M.B., Kanagy, N.L., Walker, M.K., 2003. Cardiac hypertrophy in aryl hydrocarbon receptor null mice is correlated with elevated angiotensin II, endothelin- 1, and mean arterial blood pressure. Toxicol. Appl. Pharmacol. 193, 177–187. This study was funded by the Swedish Research Council for Mistry, N.F., Cresci, S., 2010. PPAR transcriptional activator complex polymorphisms and Environment, Agricultural Sciences and Spatial Planning (FORMAS, the promise of individualized therapy for heart failure. Heart Fail. Rev. 15, 197–207. Grant number 2007-2047) and was written as a degree project at the Ong, S.B., Hall, A.R., Hausenloy, D.J., 2013. Mitochondrial dynamics in cardiovascular – medical program at Uppsala University, at Occupational and health and disease. Antioxid. Redox Signal. 19, 400 414. Quist, E.M., Filgo, A.J., Cummings, C.A., Kissling, G.E., Hoenerhoff, M.J., Fenton, S.E., Environmental medicine (Department of Medical Sciences). 2015. Hepatic mitochondrial alteration in CD1 mice associated with prenatal ex- posures to low doses of perfluorooctanoic acid (PFOA). Toxicol. Pathol. 43, 546–557. References Ren, H., Vallanat, B., Nelson, D.M., Yeung, L.W., Guruge, K.S., Lam, P.K., Lehman- McKeeman, L.D., Corton, J.C., 2009. Evidence for the involvement of xenobiotic- responsive nuclear receptors in transcriptional effects upon perfluoroalkyl acid ex- Abe, T., Takahashi, M., Kano, M., Amaike, Y., Ishii, C., Maeda, K., Kudoh, Y., Morishita, posure in diverse species. Reprod. Toxicol. 27, 266–277. T., Hosaka, T., Sasaki, T., Kodama, S., Matsuzawa, A., Kojima, H., Yoshinari, K., 2016. Rosen, M.B., Das, K.P., Rooney, J., Abbott, B., Lau, C., Corton, J.C., 2017. PPARalpha- Activation of Nuclear Receptor CAR by an Environmental Pollutant Perfluorooctanoic independent transcriptional targets of perfluoroalkyl acids revealed by transcript Acid. [Online]. Available. https://www.ncbi.nlm.nih.gov/pubmed/27832320 profiling. Toxicology 387, 95–107. [Accessed]. Rosenmai, A.K., Ahrens, L., Le Godec, T., Lundqvist, J., Oskarsson, A., 2018. Relationship Boulanger, B., Vargo, J., Schnoor, J.L., Hornbuckle, K.C., 2004. Detection of per- between peroxisome proliferator-activated receptor alpha activity and cellular con- fluorooctane surfactants in Great Lakes water. Environ. Sci. Technol. 38, 4064–4070. centration of 14 perfluoroalkyl substances in HepG2 cells. J. Appl. Toxicol. 38, Buck, R.C., Franklin, J., Berger, U., Conder, J.M., 2011. Perfluoroalkyl and poly- 219–226. fluoroalkyl substances in the environment: terminology, classification, and origins. Salihovic, S., Karrman, A., Lindstrom, G., Lind, P.M., Lind, L., Van Bavel, B., 2013. A rapid Integr. Environ. Assess. Manag. 7, 513–541. method for the determination of perfluoroalkyl substances including structural iso- Chen, Y., Liu, Y., Dorn, G.W., 2011. Mitochondrial fusion is essential for organelle mers of perfluorooctane sulfonic acid in human serum using 96-well plates and function and cardiac homeostasis. Circ. Res. 109, 1327–1331. column-switching ultra-high performance liquid chromatography tandem mass Dai, D., Chang, Y., Chen, Y., Yu, S., Guo, X., Sun, Y., 2017. Gender-specific association of spectrometry. J. Chromatogr. A 1305, 164–170. decreased estimated glomerular filtration rate and left vertical geometry in the Saunders, J., Mathewkutty, S., Drazner, M.H., McGuire, D.K., 2008. Cardiomyopathy in general population from rural Northeast China. BMC Cardiovasc. Disord. 17, 24. type 2 diabetes. Herz Kardiovaskuläre Erkrankungen 33, 184–190. Elahi, M.M., Naseem, K.M., Matata, B.M., 2007. Nitric oxide in blood. The nitrosative- Sjoberg Lind, Y., Lind, L., Salihovic, S., Van Bavel, B., Lind, P.M., 2013a. Persistent or- oxidative disequilibrium hypothesis on the pathogenesis of cardiovascular disease. ganic pollutants and abnormal geometry of the left ventricle in the elderly. J. FEBS J. 274, 906–923. Hypertens. 31, 1547–1553. Finck, B.N., Kelly, D.P., 2002. Peroxisome proliferator-activated receptor α (PPARα) Sjoberg Lind, Y., Lind, P.M., Salihovic, S., Van Bavel, B., Lind, L., 2013b. Circulating signaling in the gene regulatory control of energy metabolism in the normal and levels of persistent organic pollutants (POPs) are associated with left ventricular diseased heart. J. Mol. Cell. Cardiol. 34, 1249–1257. systolic and diastolic dysfunction in the elderly. Environ. Res. 123, 39–45. Finck, B.N., Lehman, J.J., Leone, T.C., Welch, M.J., Bennett, M.J., Kovacs, A., Han, X., So, M.K., Miyake, Y., Yeung, W.Y., Ho, Y.M., Taniyasu, S., Rostkowski, P., Yamashita, N., Gross, R.W., Kozak, R., Lopaschuk, G.D., Kelly, D.P., 2002. The cardiac phenotype Zhou, B.S., Shi, X.J., Wang, J.X., Giesy, J.P., Yu, H., Lam, P.K., 2007. Perfluorinated induced by PPARα overexpression mimics that caused by diabetes mellitus. J. Clin. compounds in the Pearl River and Yangtze River of China. Chemosphere 68, Investig. 109, 121–130. 2085–2095. Ghali, J.K., Liao, Y., Simmons, B., Castaner, A., Cao, G., Cooper, R.S., 1992. The prog- Stubleski, J., Salihovic, S., Lind, L., Lind, P.M., Van Bavel, B., Karrman, A., 2016. Changes nostic role of left ventricular hypertrophy in patients with or without coronary artery in serum levels of perfluoroalkyl substances during a 10-year follow-up period in a

299 I. Mobacke et al. Environment International 115 (2018) 295–300

large population-based cohort. Environ. Int. 95, 86–92. receptor-alpha activation. Toxicology 316, 43–54. Tang, L.L., Wang, J.D., Xu, T.T., Zhao, Z., Zheng, J.J., Ge, R.S., Zhu, D.Y., 2017. Wolf, C.J., Schmid, J.E., Lau, C., Abbott, B.D., 2012. Activation of mouse and human Mitochondrial toxicity of perfluorooctane sulfonate in mouse embryonic stem cell- peroxisome proliferator-activated receptor-alpha (PPARalpha) by perfluoroalkyl derived cardiomyocytes. Toxicology 382, 108–116. acids (PFAAs): further investigation of C4-C12 compounds. Reprod. Toxicol. 33, Wallace, K.B., Kissling, G.E., Melnick, R.L., Blystone, C.R., 2013. Structure-activity re- 546–551. lationships for perfluoroalkane-induced in vitro interference with rat liver mi- Wolf, C.J., Zehr, R.D., Schmid, J.E., LAU, C., Abbott, B.D., 2010. Developmental effects of tochondrial respiration. Toxicol. Lett. 222, 257–264. perfluorononanoic Acid in the mouse are dependent on peroxisome proliferator-ac- Wang, X., Liu, L., Zhang, W., Zhang, J., Du, X., Huang, Q., Tian, M., Shen, H., 2017. Serum tivated receptor-alpha. PPAR Res. 2010 (11 pages). metabolome biomarkers associate low-level environmental perfluorinated compound Xia, W., Wan, Y., Li, Y.Y., Zeng, H., Lv, Z., Li, G., Wei, Z., Xu, S.Q., 2011. PFOS prenatal exposure with oxidative/nitrosative stress in humans. Environ. Pollut. 229, 168–176. exposure induce mitochondrial injury and gene expression change in hearts of Wielsøe, M., Long, M., Ghisari, M., Bonefeld-Jørgensen, E.C., 2015. Perfluoroalkylated weaned SD rats. Toxicology 282, 23–29. substances (PFAS) affect oxidative stress biomarkers in vitro. Chemosphere 129, Zeng, H.C., He, Q.Z., Li, Y.Y., Wu, C.Q., Wu, Y.M., Xu, S.Q., 2015. Prenatal exposure to 239–245. PFOS caused mitochondria-mediated apoptosis in heart of weaned rat. Environ. Wolf, C.J., Rider, C.V., Lau, C., Abbott, B.D., 2014. Evaluating the additivity of per- Toxicol. 30, 1082–1090. fluoroalkyl acids in binary combinations on peroxisome proliferator-activated

300