Associations Of Perfluoroalkyl Substances With Breast Cancer Incidence Risk And Mediation Role Of Mitochondrial DNA Copy Number:Prospective Case Cohort Study

Breast cancer is the most commonly diagnosed cancer and the leading cause of cancer deaths among females worldwide,which can be caused by exposure to environmental pollutants.With the development of urbanization and industrialization,synthetic chemicals,such as perfluoroalkyl substances（PFASs）,are increasingly used in industrial and daily necessaries,resulting in emerging environmental pollution.Perfluorinated carboxylic acids（PFCAs）and perfluorinated sulfonic acids（PFSAs）,the two representative PFASs,are widely detected and extensively studied.Toxicological experiments have revealed the developmental toxicity,immunotoxicity,and hormone disrupting effects of PFASs.Perfluorooctanoic acid（PFOA）has been classified as a possible human carcinogen.In vitro experiments have found that PFOA exposure could induce the malignant transformation of normal mammary epithelial cells,suggesting the mammary gland as a potential target organ for PFOA.Epidemiological studies have shown that PFOA,perfluorononanoic acid（PFNA）,perfluorodecanoic acid（PFDA）,PFOS,and perfluorohexane sulfonic acid（PFHx S）are significantly associated with breast cancer.However,the prospective evidences lacked,and no previous studies have investigated the mixture effect of multiple PFASs co-exposure on breast cancer incidence risk.Besides,the underlying mechanisms in the potential associations of PFASs with breast cancer remained unclear.Animal experimental studies have reported that PFASs exposure could cause oxidative damage and result in alteration of mitochondrial DNA copy number（mt DNAcn）.Mitochondria are devoid of histones and have limited DNA repair capability,thus mitochondria DNA are sensitive to environmental contaminants.However,epidemiological evidences of mt DNAcn alterations induced by PFAS exposures lacked.Previous researches have revealed that mt DNAcn was closely related to various adverse health outcomes,while its association with breast cancer incidence risk was inconsistent.Whether mt DNAcn can serve as a bridge in the process of PFAS-induced breast carcinogenesis remained to be elucidated.Therefore,we performed a breast cancer case cohort study of 1203 females within the Dongfeng-Tongji cohort,determined their baseline plasma levels of four PFCAs[PFOA,PFNA,PFDA,and perfluoroheptanoic acid（PFHp A）]and two PFSAs（PFOS and PFHx S）,and assessed the associations of PFASs with breast cancer incidence risk.Further,we measured levels of peripheral blood mt DNAcn,evaluated its associations with PFASs,and explored the potential intermediate roles of mt DNAcn.The current study comprises the following three parts.Part Ⅰ.Associations of PFASs with incident risk of breast cancer:a case cohort studyObjectives:To explore the associations of individual PFAS and PFASs co-exposure with breast cancer incidence risk in a case cohort study.Methods:A total of 41 129 participants were recruited from Dongfeng Motor Corporation,including 27 009 retirees enrolled in 2008 and 14 120 retirees newly enrolled in 2013.After excluding males,participants with cancers or insufficient blood samples,and those who were lost in the follow-up,the remaining 18 387 females consisted of the female base cohort.From the base cohort,5%participants were randomly selected as the sub-cohort population（n=990,including 13 incident breast cancer cases）.The sub-cohort subjects and female breast cancer cases outside the sub-cohort（n=213）comprised the current breast cancer case cohort study participants.Baseline plasma levels four PFCAs（PFOA,PFNA,PFDA,and PFHp A）and two PFSAs（PFOS and PFHx S）were determined by ultraperformance liquid chromatography-tandem mass spectrometry（UPLC-MS/MS）.Weighted Cox proportional hazard regression models were constructed to analyze the association between each plasma PFAS and incident risk of breast cancer.To attenuate potential reverse causation,sensitivity analyses were conducted by excluding new breast cancer cases within the first year of follow-up（n=22）.Considering the differences in breast cancer etiology before and after menopause,we further analyzed the above associations among women with different menopausal status.In addition,quantile g-computation approach was applied to explore the mixture effect of PFASs co-exposure and the relative contribution of each PFAS by using“qgcomp”R package.Results:Among the six determined PFASs,PFOS had the highest median plasma concentration（10.48 ng/m L）,followed by PFOA（1.21 ng/m L）and PFNA（0.81 ng/m L）,while PFHp A had the lowest median level（0.019 ng/m L）.Among all study participants,each 1-unit increase in ln-transformed PFOA and PFHp A was associated with a separate 35%and 19%higher incident risk of breast cancer[HR（95%CI）=1.35（1.02,1.77）and 1.19（1.02,1.38）,P=0.033 and 0.028,respectively],and such associations remained significant in the sensitivity analyses[HR（95%CI）=1.42（1.06,1.89）and 1.19（1.02,1.39）,P=0.017and 0.033,respectively].Considering the limited number of premenopausal breast cancer cases（n=21）,we only investigated the associations of PFASs with postmenopausal breast cancer.Each 1-unit increase in ln-transformed PFOA and PFHp A was associated with a separate 33%and 22%higher incident risk of postmenopausal breast cancer[HR（95%CI）=1.33（1.01,1.76）and 1.22（1.03,1.45）,P=0.042 and 0.024,respectively];and a borderline significant association was observed between PFNA and postmenopausal breast cancer incidence[Q4 vs.Q1,HR（95%CI）=1.53（0.96,2.45）,P=0.075,P_trend=0.051].None associations were observed for PFDA,PFOS,or PFHx S（all P>0.05）.Quantile g-computation analysis observed a 20%increased incident risk of breast cancer along with each simultaneous quartile increase in all ln-transformed PFCA concentrations[HR（95%CI）=1.20（1.01,1.42）,P=0.038],with PFOA,PFNA,and PFHp A accounting for 56%,24%,and 20%of the positive mixture effect,respectively.Besides,per quartile increase in all ln-transformed PFCA concentrations was associated with 22%increased risk of postmenopausal breast cancer[HR（95%CI）=1.22（1.02,1.46）,P=0.033],with PFOA,PFNA,and PFHp A accounting for 41%,28%,and 31%of the positive effect,respectively.None significant mixture effects were observed for PFASs or PFSAs co-exposure on breast cancer risk（all P>0.05）.Conclusions:Increased plasma PFOA and PFHp A were associated with elevated incidence risk of breast cancer.PFCAs co-exposure has mixture effect on breast cancer incidence risk,which are mainly explained by PFOA,PFNA,and PFHp A.Part Ⅱ.Associations between exposure to PFASs and mt DNAcnObjectives:To explore the associations between plasma PFASs and peripheral blood mt DNAcn among the case cohort study participants.Methods:Participants in the breast cancer case cohort study comprised the current study population（n=1203）.UPLC-MS/MS was used to detect their baseline plasma concentrations of PFOA,PFNA,PFDA,PFHp A,PFOS,and PFHx S;real-time fluorescence quantitative PCR was applied to determine the baseline peripheral blood mt DNAcn levels.General linear regression model was fitted to explore the association of each PFAS with peripheral blood mt DNAcn;quantile g-computation approach was applied to assess the mixture effects of PFASs on mt DNAcn levels by using“qgcomp”R package.Results:Each 1%increase in plasma PFHp A was associated with a 0.027%increase in peripheral blood mt DNAcn[β（95%CI）=0.027（0.005,0.048）,P=0.015];after further adjusting for ratio of platelet to white blood cell count,the above association remained significant[β（95%CI）=0.030（0.008,0.052）,P=0.006];and in sensitivity analysis restricted to the sub-cohort population（n=990）,the above significant association still existed[β（95%CI）=0.023（0.001,0.045）,P=0.047].Compared to participants with the lowest PFNA（Q1,<0.54 ng/m L）,those in the Q2 subgroup（0.55-0.79 ng/m L）had significantly higher mt DNAcn[β（95%CI）=0.078（0.011,0.146）,P=0.023];after further adjusting for ratio of platelet to white blood cell count,participants within the Q3（0.80-1.06 ng/m L）and Q4（≥1.07 ng/m L）subgroups also had significantly higher mt DNAcn[β（95%CI）=0.074（0.006,0.143）and 0.092（0.022,0.162）,P=0.034 and 0.010,respectively].None significant associations of the other four PFASs with mt DNAcn were found（all P>0.05）.Quantile g-computation analysis observed none significant mixture effects of PFASs,PFCAs,or PFSAs co-exposure on mt DNAcn（all P>0.05）.Conclusions:Plasma PFNA and PFHp A were positively associated with mt DNAcn levels,but none significant associations were observed for PFOA,PFDA,PFOS,PFHx S,or PFASs co-exposure.PartⅢ.The mediation role of mt DNAcn in associations between PFASs and incident breast cancerObjectives:To assess the association of peripheral blood mt DNAcn with incidence risk of breast cancer,and to investigate whether mt DNAcn could mediate the effects of PFASs on incident breast cancer.Methods:Plasma PFASs concentrations and peripheral blood mt DNAcn were determined for all subjects in the breast cancer case cohort study（n=1203）by using UPLC-MS/MS and real-time fluorescence quantitative PCR,respectively.Weighted Cox regression model was fitted to estimate the association of peripheral blood mt DNAcn levels with incidence risk of breast cancer.Sensitivity analysis was further performed by excluding new breast cancer cases within the first year of follow-up.Besides,mediation analyses were applied to reveal the potential intermediate roles of mt DNAcn by using“%mediation”SAS macro.Results:Compared to participants with the lowest mt DNAcn level（Q1,<0.87）,those with the highest mt DNAcn level（Q4,≥1.52）had significantly increased incidence risk of breast cancer[HR（95%CI）=2.92（1.69,5.04）,P<0.001].Considering the limited number of premenopausal breast cancer cases,we only investigated the associations of PFASs with postmenopausal breast cancer and observed significant association[HR（95%CI）=3.26（1.81,5.86）,P<0.001].After further adjusting for ratio of platelet to white blood cell count,the above associations remained significant（all P<0.05）.In the sensitivity analyses of excluding new breast cancer cases within the first year of follow-up,the above significant association still existed[Q4 vs.Q1,breast cancer:HR（95%CI）=2.56（1.44,4.53）,P=0.001,postmenopausal breast cancer:HR（95%CI）=3.01（1.18,7.65）,P=0.021].Besides,mt DNAcn could partly mediated the association between PFHp A and breast cancer[Indirect effect,HR（95%CI）=1.02（1.00,1.04）,P=0.040],but no significant mediation effect of mt DNAcn was observed in association between PFHp A and postmenopausal breast cancer[Indirect effect,HR（95%CI）=1.02（0.99,1.03）,P=0.145].None significant intermediate role of mt DNAcn was found in associations of PFOA and PFNA with breast cancer or postmenopausal breast cancer,either（all P>0.05）.Conclusions:Peripheral blood mt DNAcn was positively associated with increased breast cancer incidence risk,and partly mediated the association between PFHp A and breast cancer.The above results provided a research clue for exploring the potential mechanism of PFASs induced breast cancer.