Abstract

Objective

   Previous studies have revealed the association of lead (Pb) exposure
   with obesity. DNA methylation alteration has been suggested to be one
   of the regulatory mechanisms of obesity. We aimed to explore whether Pb
   exposure is related with weight gain and DNA methylation alteration.

Methods

   Male adult 8 week Wistar rats were divided into 5 groups: the normal
   chow diet (NCD); the NCD+0.05%Pb; the NCD+0.15%Pb; the NCD+0.45%Pb and
   the high fat diet. Rats were exposed to different dosages of Pb through
   drinking water for 21 weeks. Body weight, fasted blood glucose level,
   fasted insulin level, homeostasis assessment of insulin resistance
   (HOMA-IR) index and lipid profile were detected. Intra-peritoneal
   glucose tolerance test (IPGTT) was constructed to evaluate the glucose
   tolerance. Lipid accumulation of liver was detected and liver DNA
   underwent whole genome bisulfite sequencing.

Results

   The NCD+0.05%Pb group had significantly greater weight, HOMA-IR and
   triglycerides, and lower glucose intolerance than the NCD group (P
   <0.05). This group also showed hepatic lipid accumulation. These
   metabolic changes were not observed in the other two Pb dosage groups.
   Furthermore, DNA hypermethylation extended along pathways related to
   glucose and lipid metabolism in NCD+0.05%Pb group.

Conclusion

   Pb exposure resulted in dose-specific weight gain in adult Wistar rats,
   accompanied by alteration of DNA methylation.

Introduction

   The prevalence of overweight and obesity has drawn worldwide attention
   in all geographic areas during last two decades. In USA, the combined
   prevalence of overweight and obesity had risen from 14.5% to 35.1% in
   adults [[48]1], which is similar to growth seen in China from 14.6% to
   32.3%, due to economic boom and lifestyle changes [[49]2]. However,
   environmental factors are attributed to the rising prevalence of
   obesity, too. Endocrine disrupting chemicals, such as heavy metals,
   have been confirmed to be associated with obesity [[50]3].

   Lead (Pb) is a kind of heavy metal widely used in gasoline and
   lead-acid battery industries [[51]2]. Pb pollutants may be released as
   air pollution or waste mixtures into soil and waterways in the process
   of manufacturing, which were then up-taken through food, water and air
   [[52]4]. Lead poisoning mainly caused brain and neurology defects,
   namely cognitive disorder, movement and coordination impairment,
   hearing and visual disturbance [[53]5].Recent studies have suggested Pb
   as an endocrine disrupting chemical. Pb is not only associated with
   late-onset puberty [[54]6, [55]7] but also acts on
   hypothalamic-pituitary-adrenal (HPA) axis, thus causing high
   stress-related cortisol levels [[56]8–[57]10]. In addition, high bone
   turnover is related with Pb exposure [[58]11].Furthermore, a lot of
   studies have investigated the relationship between early Pb exposure
   with BMI and obesity, and the results were inconclusive. Some
   epidemiology as well as animal studies revealed that maternal or
   adolescent lead exposure caused prolonged obesity that persisted into
   adults[[59]4, [60]5, [61]12, [62]13], whereas others indicated mother
   lead level was associated with low birth weight[[63]14, [64]15].
   However, little has concentrated on the effects of adult Pb exposure on
   obesity. Our previous investigation from the cross-sectional study
   revealed positive relationship between blood lead level (BLL) and
   obesity in adults [[65]2]. Thus, it is essential to establish an animal
   model to explore the underlying mechanisms.

   There is a growing recognition of the impact of environmental factors
   such as heavy metals, diet and stress on the epigenetic regulation of
   gene expression, especially in early life or at maternal stage
   [[66]16]. DNA Methylation is one of the most common epigenetic events.
   Some research has confirmed that both obesity and type 2 diabetes were
   accompanied by DNA methylation changes of some metabolic-related genes
   [[67]17–[68]19]. In one study, various levels of maternal lead exposure
   resulted in the fur color changes of offspring as well as methylation
   level changes of some gene locus and imprinted genes that determined
   fur color [[69]4].

   In our study, we constructed different levels of leaded water on adult
   rats, to observe the association of Pb and weight gain and other
   metabolic parameters. Furthermore, using liver samples, we performed a
   genome-wide methylome analysis to observe DNA methylation signatures.

Materials and Methods

Animal experiments

   Twenty-five male 8-week-old Wistar rats were purchased from SLAC
   Laboratories, SIBS, Shanghai, China. Animals were housed at an ambient
   temperature of 22 ± 2°C and maintained under a normal 12 hours
   light/dark cycle and allowed access to food and water ad libitum. After
   one week of acclimatization, rats were randomly assigned to two groups
   with one (N = 20) conventionally fed with normal chow diets (NCD;
   containing 10% fat) and the other (N = 5) with high fat diets (HFD;
   containing 40% fat), the former group was then divided into 4
   sub-groups: NCD (N = 5); NCD+0.05%Pb (N = 5); NCD+0.15%Pb (N = 5);
   NCD+0.45%Pb (N = 5). Leaded water was made by dissolving Lead (Ⅱ)
   acetate trihydrate ((CH3COO) 2Pb.3H20) (Sangon biotech, shanghai) in
   distilled water. Animal treatment lasted for 21 weeks. Body weight and
   FBG were measured every two weeks at the overnight fasting condition.
   For the IPGTT, the animals were intra-peritoneally injected with
   glucose of 2g/kg body weight (Sigma) after 12 h of fasting, and blood
   glucose levels were measured at baseline and at time points of 15min,
   30min, 60min, 90min, 120min after glucose injection by an electronic
   glucometer (Terumo, Tokyo, Japan). Animals were sacrificed by
   anesthesia, bones were obtained immediately and bone lead concentration
   was detected with inductively coupled plasma atomic emission
   spectroscopy. All the animal experiments were carried out in accordance
   with the Guidelines for Care and Use of Laboratory Animals of Shanghai
   Laboratory Animal Center (SLAC), Chinese Academy of Sciences, Shanghai,
   China. All the protocols were approved by the Institutional Animal Care
   and Use Committee of SLAC (No. 2011–007). We have made all efforts to
   minimize animal suffering.

Measurement of insulin and biochemical indexes

   Tail blood was separately collected from the caudal vein following
   overnight fasting at the end of the experiment for detection of insulin
   and biochemical profiles. Alanine aminotransferase (ALT), aspartate
   transaminase (AST), triglycerides (TG), total cholesterol (TC), free
   fatty acid (FFA) and low-density lipoprotein cholesterol (LDL-C) levels
   were detected with Siemens Dimension MAX (Siemens Healthcare
   Diagnostics Inc); FSI was assessed with ELISA kits (Shibayaji, Japan);
   HOMA-IR index was calculated according to the formula: FBG*FINS/22.5.

Oil red staining of liver

   After the rats were sacrificed, the right lateral lobule of the livers
   was removed and subsequently fixed in phosphate-buffered 10% formalin.
   The fixed liver was embedded in paraffin blocks and were sliced and
   stained with oil red to evaluate lipid droplets.

Genomic DNA isolation and WGBS library construction

   Two samples were randomly selected from control group as well as
   0.05%Pb group, separately. [70]Genomic DNA was isolated and purified
   from 25 mg of frozen liver tissue with the DNeasy Tissue Kit (Qiagen,
   Germany, cat. no. 51306) according to the manufacturer's protocol. DNA
   concentration was assessed using a ND1000 spectrophotometer (NanoDrop
   Technology), and DNA quality was assessed by electrophoresis using a 1%
   agarose gel. 3ug of genomic DNA were broken into fragments by Covaris
   S2 system (Covaris, MA) for 52 seconds with 20% duty cycle, level 5
   intensity and 200 cycles per burst. Fragmented DNA were purified by
   Ampure XP Beads (Beckman Coulter, CA) and the fragments were
   end-repaired, and a single A nucleotide was added to the 3' ends of the
   fragments in preparation for ligation to a methylated adapter
   (Illumina, CA) with a single-base T overhang. The ligation products
   were purified and size-selected (300-400bp) using agarose gel
   electrophoresis (Qiagen Minelute Gel Extraction Kit). DNA was modified
   with sodium bisulfite to convert unmethylated [71]cytosine to
   [72]uracil using the Zymo Gold methylation kit (Zymo Research, CA)
   according to the manufacturer's protocol and then purified. The
   purified converted DNA was amplified with PfuTurbo Cx Hotstart DNA
   Polymerase (Agilent Technologies, CA) and was purified again using
   AMPure XP beads. Library quality was monitored using the Agilent 2100
   Bio-Analyzer (Agilent) and KAPA Library Quantification Kit (Kapa
   Biosystem). Paired-end sequencing (2×100 bp) was then carried out using
   the Illumina Hi-Seq 2000.

Whole-genome methylation analysis

Identification of methylated cytosines

   After sequencing, raw data were filtered into clean data by removing
   pollution reads, low quality reads and Adapters. FASTqc software was
   used to assess sequencing accuracy of the clean data. The sequencing
   error rates were controlled lower than 5%. Clean data were mapped to
   the reference genome of rats using in silico bisulfite conversion
   algorithm with Bismark software. Reads with methylated position were
   reported when sequencing depth was ≥1. We used the bulk fractional
   methylation of mitochondria DNA to measure the rate of false positives.
   The methylation level of an individual cytosine was calculated from the
   number of sequenced cytosines divided by the total read depth, i.e.
   (mC)/(mC+non-mC). The methylation level of a region, named “bulk
   methylation level”, which was calculated from (mC)/(mC+non-mC). mC and
   non-mC referred to the number of methylated and unmethylated cytosines,
   respectively, in this region. The cytosine methylation level covered by
   more than 5 reads were used to calculate DNA methylation levels of
   whole genome and specific genetic elements such as TTS, TSS, CDS,
   transposon element etc. TSS was defined as 2000bp of upstream and
   downstream region of gene start sites. TTS was defined as 2000bp of
   upstream and downstream region of gene termination site. Methylation
   signal intensity values of whole genome were drawn in a circular view
   using CIRCOS visualization software.

Identification and enrichment of differentially methylated regions (DMRs) of
genes

   To identify genes with significant differences in DNA methylation
   between two groups, we locked in TSS regions of genes. Biological
   replicates of each group were merged, and afterwards genomic sites with
   methylation cytosine covered by more than 2 reads were analyzed with
   DMR caller package of R software. P values were calculated by Fisher’s
   exact test, and then FDR correction was performed. Cytosines with FDR
   (<0.05) and methylation differences of at least 0.05 for CG contexts
   were identified as differentially methylated regions. Single reads of
   potential DMRs were analyzed using the Integrative Genomic Viewer (IGV)
   Browser. We processed KEGG pathway enrichment of different methylated
   genes and calculated the significance of every KEGG pathway with
   hypergeometric distribution testing method. A small P value indicated
   differential gene enrichment in specific KEGG pathway.

Statistics analysis

   Data were presented as mean ± SEM. The overall differences between
   treatment groups were compared using one-way analysis of variance
   (ANOVA), followed by Tukey’s multiple comparison test, and LSD, with or
   without repeated measures. P < 0.05 was considered to be statistically
   significant. GraphPad Prism, Version 5 (GraphPad Software) was used for
   all statistical analyses.

Results

Pb was deposited in femur of rats drinking leaded water for 21 weeks

   [73]Table 1 reveals the cumulative bone Pb deposition in each group. Pb
   was scarcely detectable in femur of both NCD and HFD group. However,
   significant Pb accumulations were observed in Pb exposure groups (P
   <0.05). The mean Pb content was 56.25±7.47 ng/mg in NCD+0.05%Pb group,
   which was tripled in 0.15%Pb (160.50±14.15 ng/mg) exposure group.
   Interestingly, Pb content in NCD+0.45%Pb group (189.30±18.22 ng/mg) did
   not exhibit the same increasing pattern.

Table 1. Bone Pb levels in rats (ng Pb/mg dry wt of femur) at 21 weeks.

   NCD NCD+0.05%Pb NCD+0.15%Pb NCD+0.45%Pb HFD
   Bone lead, ng/mg ≤1 56.25±7.47[74]^a 160.50±14.15[75]^a
   189.30±18.22[76]^a ≤1
   [77]Open in a new tab

   Data are expressed as mean ± SEM. Pb was determined in thigh bone by
   inductively coupled plasma atomic emission spectroscopy.

   ^a P< 0.05 vs. NCD by one-way ANOVA.

0.05%Pb resulted in weight gain

   Weight gains ([78]Fig 1A) were observed in all groups. Rats of HFD
   accumulated significantly more weight than that of NCD at the end of 21
   weeks (P <0.05). Both the NCD+0.15%Pb and NCD+0.45%Pb groups exhibited
   less weight gain in comparison with NCD group (P <0.05). Notably, only
   exposure to 0.05%Pb contributed to increased weight compared with the
   NCD control group (P <0.05).

Fig 1. Body Weight and Fasted glucose levels of rats after Pb exposure.

   [79]Fig 1
   [80]Open in a new tab

   (a) Body weight. (b) Fasted glucose level. Data are expressed as mean ±
   SEM. * P<0.05 vs. NCD by one-way ANOVA.

0.05%Pb resulted in glucose intolerance and insulin resistance of rats

   Though no significant difference in fasted blood glucose level ([81]Fig
   1B) was observed across different groups, in order to evaluate glucose
   homeostasis, we further conducted IPGTT at the 20^th week of the study.
   The curves of HFD group remained at the top during the 120 minutes
   ([82]Fig 2A). Meanwhile in the NCD groups, 0.05%Pb resulted in higher
   peaking glucose level at the 15 min time point compared to controls. To
   quantify glucose tolerance, we calculated the area under curve (AUC)
   for all treatment groups. Exposure to Pb increased the average AUC in
   NCD groups ([83]Fig 2B). Among the NCD groups, the NCD +0.05%Pb group
   showed a significant higher AUC than NCD control (P < 0.05), which was
   similar to the AUC of HFD. The AUCs of the other two Pb dosages (0.15%
   and 0.45%) were not significantly larger than that of the NCD control
   group.

Fig 2. Glucose intolerance and insulin resistance were observed after Pb
exposure.

   [84]Fig 2
   [85]Open in a new tab

   (a) IPGTT. * P<0.05 HFD vs. NCD; # P<0.05 NCD+0.05%Pb vs. NCD by
   repeated measures data ANOVA. (b) AUC of IPGTT. (c) Fasted serum
   insulin levels. (d) HOMA-IR indexes. Data are expressed as mean ± SEM.
   (b), (c), (d)* P<0.05 vs. NCD by one-way ANOVA.

   Moreover, a significantly elevated fasted insulin level (FSI) was
   observed in the NCD+0.05%Pb group ([86]Fig 2C) (P <0.05), which was
   even higher than in the HFD group (P <0.05). The HOMA-IR index
   exhibited the same patterns ([87]Fig 2D).

0.05%Pb led to biochemical changes

   Alanine aminotransferase (ALT) and aspartate aminotransferase (AST)
   were detected to evaluate liver function. The NCD+0.45%Pb group showed
   slightly higher ALT and AST level than others, indicating high Pb
   dosage exposure might result in liver damage ([88]Fig 3A and 3B),
   though the differences did not reach significance. The NCD+0.05%Pb
   group exhibited significantly higher TG level ([89]Fig 3C) (P <0.05)
   compared to other NCD groups. This group also had higher low-density
   lipoprotein cholesterol (LDL-C) than NCD control group (P <0.05)
   ([90]Fig 3E). Differences were not observed in serum total cholesterol
   (TC) and free fatty acid (FFA) levels ([91]Fig 3D and 3F).

Fig 3. Biochemical changes after Pb exposure.

   [92]Fig 3
   [93]Open in a new tab

   (a)Serum ALT levels; (b) Serum AST levels; (c) Serum TG levels; (d)
   Serum TC levels; (e) Serum LDL-C levels; (f) Serum FFA levels. * P<0.05
   by one-way ANOVA.

0.05%Pb led to lipid accumulation of liver

   Oil red staining showed excessive lipid droplets in the hepatocytes of
   rats fed with high-fat diet ([94]Fig 4). Meanwhile, increased hepatic
   triglyceride contents were observed in 0.05%Pb treated rats compared to
   the NCD control rats.

Fig 4. Oil red staining of liver.

   [95]Fig 4
   [96]Open in a new tab

   Lipid accumulation was observed in rats of HFD group as well as exposed
   to 0.05%Pb. The magnification is 400×.

Whole genome DNA methylation alteration was observed in 0.05%Pb exposed rats

   The WGBS data has been uploaded to GEO(Series
   record:[97]GSE89919).Significant metabolic changes were observed in the
   NCD+0.05%Pb group, thus we conducted the whole genome bisulfite
   sequencing of liver samples in both the NCD+0.05%Pb group and the NCD
   control group. As was observed in the CIRCOS map ([98]Fig 5A), the
   NCD+0.05%Pb group had hypermethylation at the whole genome level,
   compared to NCD. The density distribution curves revealed that
   unmethylated CG sites were reduced and 100% mCG sites were increased
   after 0.05%Pb treatment ([99]S1 Fig). Furthermore, average DNA
   methylation levels were all elevated in specific genetic functional
   elements such as gene, intergenic, intron, extron, CpG Island as well
   as repeated elements ([100]S2 Fig). The profiling analysis revealed the
   overall trend of DNA methylation in different regions of gene,
   transcription start site (TSS) and transcription terminal site (TTS)
   sections ([101]Fig 5B–5D). All the three sections showed increased
   methylation levels at overall length, although there were small
   differences between two samples of the same group.

Fig 5. DNA methylation files of liver genome for NCD+0.05%Pb group compared
to NCD group.

   [102]Fig 5
   [103]Open in a new tab

   (a) The circus map; (b)The profiling of TSS region; (c) The profiling
   of gene body region; (d) The profiling of TTS region.

Differentially methylated genes associated with glucose and lipid metabolism
in 0.05%Pb exposed rats

   KEGG pathway enrichment analysis was employed to identify important
   pathways that were altered in Pb exposure group compared to control
   group ([104]Fig 6A). Among the top 20 pathways with altered DNA
   methylation were metabolic pathways, fatty acid elongation pathway as
   well as TCA cycle pathway.

Fig 6. DNA methylation enrichment and specific genes with altered DNA
methylation for NCD+0.05%Pb group compared to NCD group.

   [105]Fig 6
   [106]Open in a new tab

   (a) The top 10 KEGG pathways changed in NCD+0.05%Pb group compared to
   NCD group. P values were derived from Fisher’s Exact Test. (b) The heat
   map of specific genes related with glucose and lipid metabolism
   (FDR<0.25).

   Difference analysis in gene methylation aimed at TSS region indicated
   that a total of 2733 genes experienced DNA methylation changes, with
   2320 hypermethylated genes and 413 hypomethylated genes ([107]S1
   Table). A heatmap of different methylated genes related to lipid and
   glucose metabolism were employed ([108]Fig 6B). DNA hypermethylation
   was observed in genes related to fatty acid transport and synthesis
   (LDLR, FATP, Glyctk, DGAT1 and DGAT2) and fatty acid oxidation (CPT-2,
   MCAD, CYP4A, ACO). In addition, genes related to VLDL-TG assembling
   (MTTP and APO-CⅢ, APO-, APO-Ⅳ) and the transcription factor such as LXR
   and SREBP-1c were hypermethylated. Some glucose metabolism related
   genes such as PKA and AKT were hypermethylated. PEPCK (1.45 multiple)
   and G6pase (1.18 multiple) related with gluconeogenesis were also
   hypermethylated, as well as cholesterol metabolism genes such as CYP51
   and ACAT.

Discussion

   In our research, we found that 0.05%Pb water exposure brought about
   weight gain and insulin resistance as well as glucose intolerance in
   the NCD rats, accompanied by elevated serum TG as well as hepatic lipid
   accumulation, which were not seen in the other two dosages (0.15% and
   0.45%) of NCD groups. Furthermore, compared to the NCD control group,
   we found the DNA hypermethylation extended along pathways related to
   glucose and lipid metabolism in the NCD+0.05%Pb group. These results
   indicated that 0.05%Pb may induce abnormal metabolic status through DNA
   methylation that regulated important genes involved in lipid and
   glucose metabolism.

   As a kind of environmental endocrine disruptor, Pb mimics or disrupts
   hormone function at low doses in ways not predicted by high-dose
   studies [[109]20]. Previous animal studies had found that Pb exposure
   played both dose-specific and sex-specific roles in body weight and
   glucose dys-metabolism, however, most of their studies focused on
   prenatal Pb exposure, and the effective dose and conclusions were
   discrepant. In one study, prenatal exposure to 16ppm and 32ppm lead
   water caused weight gain in male offspring of mice; moreover, body fat
   and FSI level as well as HOMA-IR level of 16ppm male offspring were
   significantly increased. In contrast, weight gain and body composition
   changes as well as glucose intolerance were not seen in female
   offspring [[110]12]. In addition, a linear increase in mean weight with
   increasing pre-weaning lead exposure in male offspring was observed in
   a study constructed by Faulk et al. [[111]4].

   The effects of postnatal Pb exposure on metabolism changes were
   inconclusive. Our previous epidemiology study revealed a positive
   relationship between BLL and BMI in China [[112]2]. In another
   longitudinal study in Boston, the chronic lead exposure in childhood
   may result in obesity that persists into adulthood [[113]13]. Whereas
   Both Miguel A. Padilla et al. and Franco Scinicariello et al. reported
   that BLL associated with lower body weight in adults based on data from
   NHANES [[114]1, [115]21]. In another study, no association was found
   between BLL and BMI in adults [[116]22]. The conflicting results may be
   due to ethnic variations as well as different BLL of the study
   population. In one animal study, gestational exposure from 27 to 109ppm
   of leaded water resulted in late-onset obesity of offspring, which was
   not observed in the same-dosage treated postnatal groups[[117]5]. In
   another animal study constructed by Eric Beier et al. [[118]11],
   postnatal 50ppm leaded water ingestion had no effect in either body
   weight or fat composition of mice, whereas it caused higher fasted
   glucose level and elevated leptin levels.

   Gestational exposure has more detrimental effects on offspring,
   probably due to high efficiency of placenta absorption; thus postnatal
   Pb exposure in rats may achieve equivalent effect as in prenatal ones
   only when increasing the exposure dosage. Therefore, higher
   concentration basis and gradients were set in our experiment compared
   to previous studies [[119]5, [120]11].

   In our experiment, weight gain was observed in the NCD rats received
   0.05% lead acetate, which was partly in accordance with our previous
   epidemiology study [[121]2] Glucose tolerance was also impaired when
   exposed to Pb, which can be resulted from defects in insulin secretion
   and/or insulin sensitivity. Both fasting insulin and glucose levels
   were elevated in the NCD+0.05%Pb group, indicating excessive hepatic
   output of glucose and/or insensitivity of peripheral tissues especially
   liver, which could also be verified by HOMA-IR index. Therefore, it can
   be speculated that lead may disrupt glucose homeostasis, somewhat,
   through induction of insulin resistance. Mostafalou et al.
   [[122]23]also found that 0.05% Pb level of drinking water could lead to
   glucose intolerance and insulin resistance in rats. Whereas, overweight
   itself was a contributing factor to hepatic lipid content and insulin
   resistance as well as glucose intolerance, all of which could be
   observed in 0.05% Pb received groups, thus it was hard to illustrate
   whether lead acted mainly as an “obesogen” or directly devoted to
   insulin resistance.

   DNA methylation is an important style of epigenetic regulation
   [[123]24]. Previous studies have concentrated on the relationship
   between Pb exposure and the methylation changes of brain. Developmental
   Pb exposure not only altered DNA methylation in both specific gene
   level and global genome level in brain of Alzheimer’s disease model
   [[124]25–[125]28], but also could affect the expression of DNA
   methyltransferase 1(DNMT1) and DNA methyltransferase 3a (DNMT3a) in
   hippocampus [[126]29]. In addition, Pb exposure could enhance the mean
   DNA methylation level and epigenetic drift of several imprinted genes
   like IGF-2, Igf2r, H19 in both cell model and animal tissues [[127]30,
   [128]31], although others revealed that increased blood or patella lead
   levels were associated with hypomethylation of LINE-1 in both cell
   model and population study[[129]32, [130]33]. Furthermore, Arko et al.
   found that Pb exposure during pregnancy affects the DNA methylation
   status of the fetal germ cells, which led to altered DNA methylation in
   grandchildren’s neonatal dried blood spots [[131]34].

   The liver is a critical organ for the regulation of whole body energy
   homeostasis due to its central role in lipid and glucose metabolism, as
   well as its close connection via the portal vein to nutrient uptake in
   the intestine [[132]17]. Epigenetic mechanisms may partly explain the
   pathogenesis and development of metabolic diseases, as previous studies
   has found altered DNA methylation accompanied by altered
   transcriptional levels of liver in obesity and type 2 diabetes.
   Therefore, we performed a global wide analysis of DNA methylation in
   liver to explain the mechanism of metabolic changes resulting from
   adult Pb exposure for the first time. The bisulfite sequencing results
   revealed a hepatic hypermethylation of whole genome of 0.05% Pb exposed
   rats compared to unexposed ones, which may be explained by changed
   expression or activity of proteins controlling DNA methylation
   [[133]18]. In addition, a lot of genes involved with glucose and lipid
   metabolism were observed with apparent DNA methylation alterations at
   the TSS regions [[134]35]. PKB/AKT pathway was essential in regulating
   lipid and glucose metabolism in liver. Hypermethylation of AKT may
   result in gene repression and thus inhibiting this pathway, which could
   lead to reduced glucose uptake and glycogen synthesis but increased
   gluconeogenesis, thus promoting insulin resistance of liver [[135]17].
   The hypermethylation of genes involved in fatty acid oxidation such as
   CPT-2, MCAD, and ACO could result in reduced fatty acid oxidation rate
   and in turn elevated lipid synthesis in liver. MTTP and lipoprotein
   like APO-CⅢ, APO-Ⅴ, and APO-Ⅳ were important in VLDL-assembling and TG
   transportation. Hypermethylation of these proteins may give rise to
   reduced TG output and a further accumulating of TG in liver [[136]36].

   Bone lead level is a more reliable biomarker of long-term Pb exposure,
   while blood lead level reveals recent exposure [[137]15]. Thus, we
   evaluated Pb accumulation by detecting bone lead level instead of blood
   lead level, which was differed from most previous studies [[138]12,
   [139]21]. Whereas, it would be more rigorous to detect both BLL and
   bone lead level. Furthermore, the lead level of NCD+0.45% group did not
   cause corresponding increase in bone lead levels, which may due to the
   precipitation of the highest concentration of lead in the water.

   Our study has some merits. Firstly, for the first time, we performed
   whole genome DNA methylation analysis of liver in both Pb exposure rats
   and control rats, in order to explore the potential mechanism of
   metabolic changes after exposure to lead. Secondly, our study
   concentrated on adult stage but not prenatal or developmental stage.
   This made good sense in our daily life, because the metabolic disorder
   and altered DNA methylation of adults may have adverse effects on the
   next generation.

   In conclusion, we found that chronic 0.05%Pb exposure results in
   dose-specific insulin resistance and weight gain. These metabolic
   disorders may be induced through altered methylation of genes related
   with metabolism. Further studies are warranted to explore the
   alteration of specific gene expression and its association with DNA
   methylation in both animal and cellular levels.

Supporting Information

   S1 Fig. The density distribution curves of methylated CG sites.

   (a)(b) NCD+0.05%; (c)(d) NCD.

   (TIF)
   [140]Click here for additional data file.^ (236KB, tif)
   S2 Fig. Average DNA methylation levels in specific Genetic functional
   elements.

   (a)(b) NCD+0.05%; (c)(d) NCD.

   (TIF)
   [141]Click here for additional data file.^ (234.3KB, tif)
   S1 Table. Differential methylated genes.

   proportion1: NCD; proportion2: NCD+0.05%Pb.

   (XLS)
   [142]Click here for additional data file.^ (870.5KB, xls)

Acknowledgments