【作者】 张义峰；

【导师】 祝凌燕；

【作者基本信息】 南开大学 ， 环境科学， 2013， 博士

【摘要】 由于长期生产和使用,全氟化合物(perfluoroalkyl substances, PFASs)在环境中广泛存在并在世界各国人体样本中检出。全氟辛烷磺酸(perfluorooctanesulfonate, PFOS)和全氟辛酸(perfluorooctanoate, PFOA)是环境中检出频率最高的两种PFASs. PFASs具有持久性和生物累积性,可能导致多种不良健康结果。历史上主要有两种生产方式用来生产PFASs及其前体物,即调聚法(telomerization)和电氟化法(electrochemical fluorination, ECF)早在1947和1949年,电氟化法就应用于生产PFOA和PFOS,该方法的最终产品一般含有稳定比例的直链(约70%PFOS和80%PFOA)和支链(约30%PFOS和20%PFOA)异构体。而由杜邦公司于1970年开发了调聚法生产的产品通常是纯的直链结构。由于PFOS和PFOA异构体的环境行为、药代动力学、毒性等具有一定的差异,研究人体中PFASs异构体的组成和清除效率对于追溯人体暴露来源和评估健康风险具有重要意义。本文首次应用一种新的PFASs异构体分析方法对来自我国典型城市(石家庄、邯郸和天津)的229名志愿者提供的血清样本进行分析。目标是分析我国人体中的PFASs以及PFOS和PFOA异构体,阐明性别和年龄对典型异构体累积的影响,评估这些人群中的PFOS和PFOA来源。随后,为了评估PFASs以及PFOS和PFOA异构体在人体中的肾排泄能力,对来自石家庄和邯郸的86对血样和尿样进行了分析,对肾清除率和半衰期进行了估算。最后,对30对来自石家庄的夫妇提供的60个成对全血和血清样本进行分析,用于研究PFASs以及PFOS和PFOA异构体在全血和血清中的分布。在石家庄和邯郸的129个血清样本中,PFOS (平均值33.3ng/mL)是最主要的PFAS,依次是全氟己烷磺酸(PFHxS,2.95ng/mL), PFOA (2.38ng/mL)和全氟壬酸(PFNA,0.51ng/mL)。PFOS的平均水平比近些年北美地区的人群还要高。石家庄市区(59.0ng/mL)和邯郸市区(35.6ng/mL)人群中PFASs的总浓度显著(p<0.001和p=0.041)高于石家庄农村地区人群(24.3ng/mL);青年女性组主要PFASs的浓度比男性、老年女性组都低。石家庄和邯郸(n=129)人群中的全氟烷基磺酸(perfluoroalkane sulfonates, PFSAs)比天津(n=100)人群中高,而天津人群中全氟羧酸(perfluoroalkyl carboxylates, PFCAs)则高于石家庄和邯郸人群。这些暗示着不同地区PFASs不同的暴露源和暴露途径。石家庄和邯郸人群中的直链PFOS平均比例(n-PFOS)仅占48.1%,比天津人群低(59.2%),都低于历史上主要PFOS生产商的工业品(约70%直链)。此外,在229个样本中,n-PFOS比例随着PFOS总浓度的增加而显著下降。所得数据支持之前的推论：i)人体血液中高比例支链PFOS异构体是PFOS前体物(PreFOS)的生物标记物；ii)高暴露水平人群中PFOS来源于高浓度的PreFOS不等比例的暴露。石家庄和邯郸人群血清中的直链PFOA (n-PFOA)比例为96.1%,明显高于工业品(约75-80%直链),但是低于2007-2008年的美国人群(接近100%直链),这说明我国仍然存在ECF方式生产的PFOA的使用。建立了高灵敏度的分析尿液中PFASs及异构体的方法,并首次在尿液中发现了多种PFASs。大多数PFASs在尿液中的浓度与血液中的浓度有较强的正相关性(p<0.05),因此尿液可以用于监测人体中PFASs的暴露水平。短链的PFCAs比长链的更容易排泄,而同等碳链长度的PFCAs比PFSAs排泄更快。但是,PFOS (C8)比PFHxS (C6)排泄更快。在PFOS和PFOA的异构体中,除一个PFOS的支链异构体(lm-PFOS)外,各主要支链异构体都比直链异构体更易排泄。尿液中的直链全氟辛基磺酰胺(perfluorooctane sulfonamide, PFOSA)(中位值,84%)显著(p<0.001)高于血液中(65%),说明直链PFOSA相对其支链异构体优先通过尿液排泄,这与PFOS和PFOA的异构体肾排泄方式相反。通过一室药代动力学模型估算PFASs的生物半衰期从0.5±0.1年(5m-PFOA)到90±11年(lm-PFOS)。肾排泄是PFOA的主要排泄方式,但是对于PFOS和PFHxS(以及一些其他长链的PFCAs)来说,月经或者其他排泄途径可能对整体排泄有贡献。除了PFHxA和PFNA外,大多数PFASs在全血和成对血清中有很好的正相关性(n=60, r=0.49-0.87,p<0.001). PFHxA在血清中的浓度显著(p<0.001)低于成对全血；PFOA (p=0.105)和PFNA (p=0.237)没有显著差异；对于更长碳链的PFCAs,全氟癸酸(perfluorodecanoate, PFDA)和全氟十一酸(perfluoroundecanoate, PFUdA)在血清中的浓度显著(p<0.001)高于全血。这说明短链的PFCA更倾向于分布于血细胞中。与PFHxA (C6)和PFOA(C8)不同,尽管有相同的碳链长度,PFHxS (C6)和PFOS (C8)在血清中的浓度显著(p<0.001)高于全血,说明官能团(羧酸和磺酸基团)在血清和全血分配中也起着重要作用。血清/全血分配率的中位数如下,PFHxS (1.67), PFOS (1.68), PFHxA (0.34), PFOA (1.24), PFNA (0.85), PFDA (1.39)和PFUdA (1.60)。对于PFOS的异构体,除iso-PFOS外,其他单支链的PFOS异构体的分配率中位数接近或者大于2(1.94～2.69);n-PFOS,iso-PFOS和2-PFOS的分配率中位数都低于2(1.40-1.54)。对于PFOA的异构体,支链的PFOA分配率中位数都低于1,n-PFOA的为1.29。说明PFOS和PFOA异构体在血清和全血中分布具有较大的差异。把全血中的PFASs浓度直接乘以2转为血清中的浓度,会导致绝大多数的PFASs高估。采用血细胞比容计算的比率(女性1.69和男性1.85)去转化数据,对于PFSAs(PFHxS和PFOS)和长碳链的PFCAs(PFUdA)是合适的,但是对于其他的PFASs也会对结果产生偏高估算。更多还原

【Abstract】 After decades of production and applications, perfluoroalkyl substances (PFASs) are ubiquitous present in the environment and have been widely detected in human samples. The most commonly detected PFASs in human samples are perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA). PFASs are persistent and bioaccumulative and display adverse effects to human health. Historically, two major manufacturing methods, namely telomerization and electrochemical fluorination (ECF), have been used to produce PFASs and their precursors. ECF was used to produce PFOA beginning in1947and PFOS and its precursors in1949. This process was known to yield a complex, yet rather consistent, mixture of linear (ca.70%for PFOS,80%for PFOA) and branched isomers (ca.30%for PFOS,20%for PFOA) in final products. However, the telomerization method, which was developed by Dupont in the1970s, produces typically pure linear isomer. The isomers of PFOS and PFOA display different environmental behaviors, pharmacokinetics and toxicity. Thus, it is of great importance to study the distribution and elimination effciency of isomers for human exposure source analysis and health risk assessment.In the present thesis, a new isomer-specific PFAS analytial method was developed for the first time to analyze229serum samples collected in three typical cities (Shijiazhuang, Handan and Tianjin) in North China. The objectives were to quantify PFASs and the isomer profiles of PFOS and PFOA in general Chinese people, to elucidate the impacts of gender and age on the isomer-specific accumulation pattern, and to try to evaluate the sources of PFOS and PFOA exposure in these populations. Subsequently, to estimate the renal excretion of PFASs and isomers of PFOS and PFOA, paired blood and urine samples (n=86) from adults in Shijiazhuang and Handan were analyzed, and rates of renal clearance and half-lives were estimated. At last, to study the ditribution of PFASs and the isomers of PFOS and PFOA in whole blood and serum, sixty paired samples from thirty couples in Shijiazhuang were analyzed. Among the129serum samples from Shijiazhuang and Handan, total PFOS (∑PFOS, mean33.3ng/mL) was the predominant PFAS followed by perfluorohexanesulfonate (PFHxS,2.95ng/mL), total PFOA (∑PFOA,2.38ng/mL), and perfluorononanoate (PFNA,0.51ng/mL). The level of∑PFOS appeared to be higher than in people from North America in recent years. The mean concentrations of∑PFASs in the participants living in urban Shijiazhuang (59.0ng/mL) and urban Handan (35.6ng/mL) were significantly higher (p<0.001and p=0.041, respectively) than those living in the rural district of Shijiazhuang (24.3ng/mL). The young female sub-population had the lowest∑PFAS concentrations compared with males and older females. The mean concentration of perfluoroalkane sulfonates (PFSAs) in people (n=129) from Shijiazhuang and Handan was higher than in people (n=100) from Tianjin, while, people in Tianjin displayed higher level of perfluoroalkyl carboxylates (PFCAs) than in Shijiazhuang and Handan. These suggest that the people in different regions may have different exposure sources of PFASs.The proportion of linear PFOS (n-PFOS) inFOS was only48.1%for people in Shijiazhuang and Handan, which is lower than people in Tianjin (59.2%). They are both much lower than what was present in technical PFOS from the major historical manufacturer (ca.70%linear). Moreover, the proportion of n-PFOS decreased significantly with increasing∑PFOS concentration in the229serum samples. The results may have two implications:i) high content of branched PFOS isomers in serum is a biomarker of exposure to PFOS-precursors (PreFOS), and ii) that people with the highest∑PFOS concentrations are exposed disproportionately to high concentrations of PreFOS. On average, linear PFOA (n-PFOA) contributed96.1%of∑PFOA in serum samples from Shijiazhuang and Handan, significantly higher than in technical PFOA (ca.75-80%linear), but lower than in Americans (close to100%) in2007-2008, suggesting ECF PFOA was still be used in China.A newly sensitive isomer-specific method was developed to analyze PFASs in human urine, which permitted the detection of many PFASs in human urine for the first time. The levels of most PFASs in urine correlated positively with their levels in paired blood (p<0.05). Thus, urine could be used to monitor PFASs exposure in humans. In general, shorter PFCAs were excreted more efficiently than longer ones, and PFCAs were excreted more efficiently than PFSAs of the same carbon chain-length. However, PFOS (a C8compound) was excreted more efficiently than PFHxS (a C6compound). Among PFOS and PFOA isomers, major branched isomers were more efficiently excreted than the corresponding linear isomer except1m-PFOS. The proportion of linear perfluorooctane sulfonamide (PFOSA) was significantly higher (p<0.001) in urine (median84%) than in blood (65%), demonstrating that linear PFOSA is preferentially excreted via urine relative to branched isomers; opposite to the findings for renal excretion of PFOS and PFOA isomers. A one-compartment model was used to estimate the biological elimination half-lives of PFASs. Among all PFASs, the estimated arithmetic mean with standard error elimination half-lives ranged from0.5±0.1years (5m-PFOA) to90±11years (1m-PFOS). Renal excretion was the major elimination route for PFOA, but for PFOS and PFHxS (and possibly other long-chain PFCAs) menstruation and other routes of excretion likely contribute to overall elimination.For most PFASs, the levels in the whole blood correlated positively with the levels in paired serum samples (n=60, r=0.49-0.87,p<0.001) except PFHxA and PFNA. PFHxA in serum samples was significantly lower (p<0.001) than in the paired whole blood samples, while, for PFOA (p=0.105) and PFNA (p=0.237), no significantly differences were found, but for the longer chain-length PFCAs, perfluorodecanoate (PFDA) and perfluoroundecanoate (PFUdA) in serum samples was significantly higher (p<0.001) than in the paired whole blood samples. These suggested that shorter chain-length PFCAs preferrd to distribute in the blood corpuscles. Unlike PFHxA (C6) and PFOA (C8), althrough kept the same chain-length, PFHxS (C6) and PFOS (C8) in serum samples were significantly higher (p<0.001) than in the paired whole blood samples, which implied functional groups (carboxyl group and sulfonic group) also played an important role in the ditribution of serum and whole blood.The median serum/whole blood distribution ratios were showed as follows, PFHxS (1.67), PFOS (1.68), PFHxA (0.34), PFOA (1.24), PFNA (0.85), PFDA (1.39) and PFUdA (1.60). For PFOS isomers, the median serum/whole blood distribution ratios for all the monomethyl branched PFOS were close to or greater than2(1.94～2.69) except iso-PFOS, while, the median ratios for n-PFOS, iso-PFOS and2-PFOS were similar and lower than2(1.40～1.54). For PFOA isomers, the median ratios for branched PFOA were lower than1, and was1.29for n-PFOA. These suggested the isomers distribution in serum and whole blood were different. When transfer PFASs concentration from whole blood samples to serum samples, using the factor2would lead to overestimation for most PFASs except some branched PFOS isomers. Using haematocrit calculated ratios (1.69for female and1.85for male) to transfer, would be suitable for PFSAs (PFHxS and PFOS) and long chain-length PFCA (PFUdA), but also overestimation for other PFASs.更多还原