Prenatal and Neonatal Exposure to Perfluorooctane Sulfonic Acid

May 17, 2012 - Our results indicated that miRNA had little direct regulatory effect on ... it seems that the PFOS-induced synaptic dysfunctions and ch...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/est

Prenatal and Neonatal Exposure to Perfluorooctane Sulfonic Acid Results in Changes in miRNA Expression Profiles and Synapse Associated Proteins in Developing Rat Brains Faqi Wang,† Wei Liu,† Junsheng Ma,† Mingxi Yu,† Yihe Jin,†,* and Jiayin Dai‡ †

School of Environmental Science and Technology, Dalian University of Technology, Key Laboratory of Industrial Ecology and Environmental Engineering, MOE, Dalian 116024, China; ‡ Key Laboratory of Animal Ecology and Conservation Biology, Institute of Zoology, Chinese Academy of Sciences, Beijing 100190, China; S Supporting Information *

ABSTRACT: We previously identified a number of perfluorooctane sulfonic acid (PFOS)-responsive transcripts in developing rat brains using microarray analysis. However, the underlying mechanisms and functional consequences remain unclear. We hypothesized that microRNAs (miRNAs), which have emerged as powerful negative regulators of mRNA and protein levels, might be responsible for PFOS-induced mRNA changes and consequent neural dysfunctions. We used eight miRNA arrays to profile the expression of brain miRNAs in neonatal rats on postnatal days (PND) 1 and 7 with maternal treatment of 0 (Control) and 3.2 mg/kg of PFOS feed from gestational day 1 to PND 7, and subsequently examined six potentially altered synapse-associated proteins to evaluate presumptive PFOS-responsive functions. Twenty-four brain miRNAs on PND 1 and 17 on PND 7 were significantly altered with PFOS exposure (P < 0.05), with miR-466b, -672, and -297, which are critical in neurodevelopment and synapse transmission, showing a more than 5-fold reduction. Levels of three synapse-involved proteins, NGFR, TrkC, and VGLUT2, were significantly decreased with no protein up-regulated on PND 1 or 7. Perfluorooctane sulfonic acid might affect calcium actions during synapse transmission in the nervous system by interfering with SYNJ1, ITPR1, and CALM1 via their targeting miRNAs. Our results indicated that miRNA had little direct regulatory effect on the expression of mRNAs and synapse-associated proteins tested in the developing rat brain exposed to PFOS, and it seems that the PFOS-induced synaptic dysfunctions and changes in transcripts resulted from a combinatory action of biological controllers and processes, rather than directed by one single factor.

■

INTRODUCTION Perfluorooctane sulfonic acid (PFOS) is the ultimate degradation product of many perfluorinated compounds and has been widely used in commercial, industrial, and household applications such as fire-fighting foams, metal surfaces, textiles, polishes, and carpets.1 Recent studies have reported that PFOS can penetrate the brain blood barrier (BBB) and the placental barrier during pregnancy, leading to persistent effects on the developing nervous system such as neurobehavioral dysfunctions, and defects in learning and memory possibly through to adulthood.2−5 While such issues have prompted many recent studies on PFOS neurotoxicity in developing offspring, the underlying mechanisms, especially PFOS-induced molecular actions, are still not thoroughly understood. Our previous study employed microarrays to identify PFOS-responsive transcripts, showing that genes enriched for central nervous system (CNS) development, neurogenesis, neurotransmitter transport, memory, and synaptic transmission were highly responsive to PFOS.5 However, little is known about what biological control © 2012 American Chemical Society

processes lie behind the transcriptional effect of PFOS and what the functional consequences are. Micro-RNAs, a group of 18- to 25-nucleotide-long noncoding RNA molecules, play an important role in fundamental biological and metabolic processes in eukaryotic organisms, because they regulate the stability and translation of large numbers of mRNAs serving varied functions.6,7 Recently, miRNA profiles that respond to chemical treatments have been intensively studied to explore the molecular mechanisms underlying chemically triggered toxicity, and to develop a set of sensitive biomarkers for assessing the health risks of environmental toxicants.7−9 Brain tissue is a major site of miRNA expression,6 with numerous studies reporting on the association of miRNA regulation with brain dysfunctions.7,9,10 Received: Revised: Accepted: Published: 6822

September 28, 2011 May 16, 2012 May 17, 2012 May 17, 2012 dx.doi.org/10.1021/es3008547 | Environ. Sci. Technol. 2012, 46, 6822−6829

Environmental Science & Technology

Article

MicroRNA Labeling Kit (Kreatech, U.S.). To reduce background signals and increase the performance of the microarray, a prehybridization of labeled targets was carried out in accordance with recommended instructions with the blocking buffer (5 × SSPE, 0.1% SDS, 1% BSA), before subsequent complete hybridization using the glass cover slide method (Corning Cover Glass). Microarrays with labeled samples were washed with an excess amount of prewarmed washing solution (2 × SSC, 0.2% SDS) for 5 min at 37 °C, and scanned using a Genepix 4100B laser scanner (Molecular Devices). The obtained microarray images were analyzed using OneArray Studio Analysis Software (Phalanx Biotech, Taiwan) to identify differentially expressed miRNAs (DE miRNAs) after 75% scaling normalization. Based on,5 we considered miRNAs with FC (fold change) > 1.50 or 0.7 as differentially expressed miRNAs. Bioinformatics Analysis. PFOS-Responsive mRNA Targets Prediction. We first predicted potential targets for DE miRNAs (the FC was cut off at 1.5) among mRNAs previously shown to be affected by PFOS exposure using four individual computational methods, namely TargetScan (http://www. targetscan.org/), MirSVR (http://www.microrna.org/), MiRDB (http://mirdb.org/miRDB/), and PicTar (http:// pictar.mdc-berlin.de/), as well as the mRNA expression data.5 The filtering of mRNA targets was set for TargetScan with conserved sites ≥1, for MirSVR with mirSVR_score ≤ −1, for MiRDB with the score ≥60, and for PicTar with PicTar score ≥2. Specifically, PFOS-responsive mRNAs that were copredicted by at least two of the four data sets were considered as target mRNAs and used for subsequent functional analysis of mRNA targets. Evaluation of the Regulatory Effect of Each DE-miRNA on the mRNA Expression Profile. The predicted mRNA target sets on PND 1 and 7 were integrated with mRNA expression data to calculate the regulatory effect-score (REscore) as described in Cheng et al.19 Because TargetScan and MirSVR predicted the most overlapped PFOS-sensitive mRNA targets (See Figure S1, Supporting Information), we used mRNA targets obtained from these two databases to calculate the RE-Score for each DE miRNA with FC ≥ 2. Further, we calculated the difference in regulatory effect score (DRE-score) for each DE miRNA between the control and PFOS treatment groups to evaluate the regulatory effects of DE miRNAs on mRNA expression after PFOS exposure. A brief description of the DRE-score calculations is shown in Text S2, Supporting Information. Distribution of Regulation Values for PFOS-Responsive mRNA Targets. We distributed the regulation values (Rvalues) 20 for PFOS-responsive mRNA targets obtained from the TargetScan and MirSVR databases, respectively, using normal density estimation (Stata/SE 9.0) to see which effect, inhibitory or elevating, was exhibited on global mRNA expression on PND 1 and 7. The regulation values were calculated as described in Ma et al,20 and are shown in Text S3, Supporting Information. Gene Ontology (GO) and Pathway Analysis of PFOSResponsive mRNA Targets. We investigated the potential functions associated with anticorrelated mRNAs supposedly regulated by miRNAs using GO enrichment analysis and the KEGG method. The text files of the anticorrelated PFOSresponsive mRNA targets on PND 1 and 7 were submitted electronically to the web-based tool DAVID Bioinformatics Resources 6.7 (http://david.abcc.ncifcrf.gov/) to determine the

Studies have shown the involvement of miRNAs (e.g., miR-297 and miR-466-b) in synaptic transmission 11 and neuronal development and differentiation,6 which are established phenotypic consequences of PFOS exposure.12−15 We therefore hypothesized that the PFOS induced transcriptional effect and associated neurological dysfunctions would partly involve interaction with brain miRNAs during early development. To test this hypothesis, we examined the effect of PFOS on global miRNAs expression in developing rat brains, particularly on synapse and neurotransmission related miRNAs. This preference was not only because most brain miRNAs are located in close proximity to synapses (dendrite or dendritic spines), playing a critical role in postsynaptic plasticity,16 but that the sensitivity of the synaptic pathway to PFOS has been demonstrated by perturbed synaptic function and synaptogenesis,17 as well as altered levels of synaptic proteins12,14 observed in vitro or in vivo. Analysis of synapse associated miRNA-mRNA-protein interaction will clarify the mechanisms of PFOS-induced neurotransmission toxicity. Thus, we additionally tested the expression levels of six proteins involved in synapses and synaptic transportation in the CNS to test the potential functional consequences of observed molecular changes after PFOS exposure. These proteins were solute carrier family 17 member 6 (VGLUT2), synaptotagmin I (Syt I), synaptotagmin XI (Syt XI), tyrosine kinase receptors B and C (TrkB and TrkC), and nerve growth factor receptor (NGFR). Because most mammalian miRNAs bind to their target mRNAs imperfectly, and tend to repress the protein translation than to degrade mRNAs,18 the selected proteins were not only from those predicted by both miRNA and mRNA array analysis (VGLUT2, Syt I, and Syt XI), but also from those predicted by miRNA array alone (TrkB, TrkC, and NGFR) with the intention to evaluate the contribution of miRNA translational modulation to PFOS-induced toxicity. In addition, we exposed experimental animals to the same model as used previously 13 to determine the association between miRNA and mRNA expressions and relative functional interactions between them in the neuro-pathway..

■

MATERIALS AND METHODS Chemicals, Animals and Administration, and Tissue Collection and RNA Extraction. Total RNAs of six randomly selected parallel samples (three male and three female) were pooled as one replicate of each treatment for subsequent microarray analysis. Each pair of male and female pups were from one litter. Detailed procedures about chemicals, animals and administration, tissue collection, and RNA extraction were performed as per established protocols,5 and can be seen in the Text S1, Supporting Information (SI). MicroRNA Microarray Analysis. MicroRNA expression profiles of brain tissues were generated by applying Mouse & Rat miRNA OneArray (Phalanx Biotech, Taiwan). A total of eight separate arrays were used for the analysis of eight pooled total RNA samples from four treatment groups (Control and PFOS treatments on PND 1 and 7) with two replicates for one treatment. All reagents were optimized for use with Mouse & Rat miRNA OneArray. Each microarray contained 387 mature rat miRNA probes and 105 experimental control probes. Each unique probe had three features, and probes contained 100% Sanger miRBase v15 mouse and rat miRNA content. All procedures were carried out according to the manufacturer’s protocols. Briefly, 2.5 μg total RNA, which had passed the quality control test, was used for RNA labeling with a ULS 6823

dx.doi.org/10.1021/es3008547 | Environ. Sci. Technol. 2012, 46, 6822−6829

Environmental Science & Technology

Article

Table 1. miRNAs with Significant Expression Levels in the Cortex after PFOS Exposure on PND 1 and 7a name PND 1 down-regulated (n = 13) rno-miR-466b rno-miR-672 rno-miR-297 rno-miR-674−3p rno-miR-207 rno-miR-346 rno-miR-466c rno-miR-125b-3p rno-miR-325−3p rno-miR-598−5p rno-miR-1 rno-miR-336 rno-miR-764 up-regulated (n = 11) rno-miR-10b rno-miR-204 rno-miR-25 rno-miR-542−5p rno-miR-181c rno-miR-17−3p rno-miR-126 rno-miR-23a rno-miR-363 rno-miR-668 rno-miR-665 PND 7 down-regulated (n = 15) rno-miR-672 rno-miR-297 rno-miR-466b rno-miR-363 rno-miR-207 rno-miR-674−3p rno-miR-325−3p rno-miR-346 rno-miR-667 rno-miR-668 rno-miR-336 rno-miR-764 rno-miR-1 rno-miR-214 rno-miR-466c up-regulated (n = 2) rno-miR-204 rno-miR-494

normalized intensity (control 1)

normalized intensity (control 2)

normalized intensity (pfos 1)

normalized intensity (pfos 2)

FC 1

FC 2

log2FC

6489.70 810.06 480.34 238.57 385.60 206.09 188.30 458.57 160.07 127.98 127.98 192.17 122.57

6472.50 798.42 467.74 206.09 395.52 187.14 148.48 380.47 220.78 121.41 142.68 173.99 129.92

663.62 122.53 86.64 88.73 166.49 91.49 82.11 210.18 111.21 70.22 75.58 104.74 74.36

676.08 118.65 82.11 80.02 165.52 90.52 80.82 217.65 103.45 78.50 85.06 116.38 78.24

0.10 0.15 0.18 0.37 0.43 0.44 0.44 0.46 0.69 0.55 0.59 0.55 0.61

0.10 0.15 0.18 0.39 0.42 0.48 0.54 0.57 0.47 0.65 0.60 0.67 0.60

−3.27 −2.74 −2.49 −1.40 −1.24 −1.11 −1.05 −0.98 −0.83 −0.75 −0.74 −0.73 −0.73

58.00 537.25 80.04 63.22 81.58 185.40 51.04 220.39 1767.28 142.48 118.32

61.86 534.01 93.96 71.14 89.12 202.61 55.29 238.57 1657.14 157.76 127.98

102.80 884.11 169.40 130.28 145.16 364.67 80.82 473.29 3463.69 311.00 289.02

94.40 901.74 125.43 106.04 159.38 342.68 113.47 428.68 3463.69 318.11 254.75

1.77 1.65 2.12 2.06 1.78 1.97 1.58 2.15 1.96 2.18 2.44

1.53 1.69 1.33 1.49 1.79 1.69 2.05 1.80 2.09 2.02 1.99

1.14 1.07 1.02 0.97 0.87 0.87 0.84 0.81 0.76 0.74 0.72

763.60 439.19 5372.23 2597.21 469.47 231.11 247.31 195.08 220.45 208.94 190.07 119.39 138.58 329.18 132.18

752.50 440.07 5382.99 2602.41 470.41 226.02 258.49 200.81 209.36 220.89 210.69 129.03 136.72 299.93 140.14

123.88 79.66 1015.05 820.56 149.36 82.00 115.39 106.60 123.74 121.83 117.14 71.46 82.00 183.91 79.66

139.40 78.49 938.32 838.02 154.63 74.97 111.87 89.03 110.54 115.39 102.50 70.87 79.66 188.01 82.29

0.16 0.18 0.19 0.32 0.32 0.35 0.47 0.55 0.56 0.58 0.62 0.60 0.59 0.56 0.60

0.19 0.18 0.17 0.32 0.33 0.33 0.43 0.44 0.53 0.52 0.49 0.55 0.58 0.63 0.59

−2.53 −2.48 −2.46 −1.65 −1.63 −1.54 −1.15 −1.02 −0.88 −0.86 −0.83 −0.80 −0.77 −0.76 −0.75

667.32 3593.28

518.69 3600.48

938.32 6063.93

1015.05 5817.91

1.41 1.69

1.96 1.62

0.72 0.72

Note. “Control 1/2” and “PFOS 1/2” represent replicate 1/2 for Control and PFOS group, respectively. “FC 1” and “FC 2” represent fold changes for replicate 1 and 2, respectively. “FC” represents the fold change of mean normalized density of two replicates for each group. Only miRNAs (p < 0.05) with log2FC < −0.7 or >0.7 were presented.

a

biological phenomena involved. Gene ontology for translating lists of differentially regulated genes into functional profiles and pathway for graphical representations of gene interactions (pathways) were both employed for analysis of predicted mRNA data. This objective tool used Benjamini and P value for gene-enrichment analysis. Statistical significance was calculated for each category by P value ≤0.05 and Benjamini value ≤1. Real-Time Reverse Transcription Polymerase Chain Reaction (RT-PCR) for Validation of Potential miRNAs.

Six isolated but not pooled total RNAs (three male and three female) were randomly selected from the RNA samples for miRNA mciroarray analysis to perform validation of altered miRNAs using RT-PCR. The experimental procedure is shown in Text S4, Supporting Information. Western Blot. We performed Western blot analysis as described previously,21 with minor modifications, to determine expression levels of six synapse-associated proteins (VGLUT2, 6824

dx.doi.org/10.1021/es3008547 | Environ. Sci. Technol. 2012, 46, 6822−6829

Environmental Science & Technology

Article

Figure 1. Normal density distribution of regulation values for PFOS-responsive mRNA targets obtained from MirSVR and TargetScan on PND 1 and 7.

between PND 1 and PND 7 (See Figure S2, Supporting Information). This suggested that the brain miRNA expression profile was more sensitive to PFOS exposure than development during early growth. This was also reflected by the hierarchical clustering analysis of miRNAs in the control and PFOS-treated rats on PND 1 and 7, which showed miRNA expression patterns from each group were clustered into two distinct categories, while expressions from the two different time points were classified into one category (See Figure S3, Supporting Information). A total of 24 cortex miRNAs on PND 1 and 7 on PND 7 showed significantly differential expression after PFOS exposure (p < 0.05) (Table 1; see Figure S4, Supporting Information). Computational Prediction of PFOS-Responsive Target mRNAs on PND 1 and 7. The PFOS-responsive mRNAs predicted as targets for DE miRNAs by the four individual computational methods are shown in Table S3, Supporting Information. The number of overlapped mRNA targets within at least two databases over PND 1 and 7 is shown in Figure S1, Supporting Information. We predicted 55 out of 842 mRNAs previously affected by PFOS exposure to be targets for DE miRNAs on PND 1, of which 33 were correlated and 22 were anticorrelated targets. On PND 7, 58 of 634 PFOS-responsive mRNAs were shown to be potential targets for DE miRNAs, of which, 13 were correlated and 45 were anticorrelated. Eight transcript targets were differentially expressed on both PND 1 and 7 (See Table S4, Supporting Information), of which, Syt I, Syt XI, and Slc17a6 were specific to the nervous system. Regulatory Effect of Each miRNA on PFOS-Responsive mRNA Expression Profile According to TargetScan and MirSVR Prediction. As shown in Table S5, Supporting Information, the DRE-Score exhibited that on PND 1, miR-297 and -10b by TargetScan prediction and miR-672 and -466c by MirSVR prediction ranked in the upper quartile (25%),

Syt I, Syt XI, and TrkB and TrkC, and NGFR) (See Text S5, Supporting Information). Statistical Analysis. Data analyses were performed using SPSS software (Ver13.0; SPSS) with statistical significance set at p < 0.05. Univariate analysis of variance (ANOVA) was used for the six proteins (VGLUT2, Syt I, Syt XI, TrkB, TrkC, and NGFR) expression levels to test the main effect of factors and differences among subjects. When results from the overall significance test led to the rejection of the null hypothesis, a posthoc test was performed to determine the influence source. The posthoc test was comprised of least-significant-difference (LSD) or Dunnett’s T3 test when appropriate for pairwise comparisons of means, and by Duncan’s multiple range test and Student−Newman−Keuls (SNK) test for multiple comparisons among groups. All data were expressed as the mean ± Standard Error (SE). Individual values for protein expression analysis were analyzed by the posthoc test.

■

RESULTS Expression Profiles of Rat Brain miRNA after PFOS Exposure on PND 1 and 7. Eight data sets from four study groups were obtained by the miRNA array analysis. Every two sets of replicate data represented one individual experimental group. The Pearson correlation analysis of raw data revealed a significantly high correlation (R > 0.99) between two biological replicates for each group, suggesting good reproducibility of samples (See Table S2, Supporting Information). In total, 243 miRNAs were detected in the rat brain out of 387 currently known mature miRNAs on the array. A significant difference in intensity distribution was observed between PFOS-treated and control rats but not between PND 1 and 7. This was shown as PFOS exposure resulted in significantly expanded distribution of miRNA expression levels on both postnatal days compared to the control, while no distribution divergence was found 6825

dx.doi.org/10.1021/es3008547 | Environ. Sci. Technol. 2012, 46, 6822−6829

Environmental Science & Technology

Article

Figure 2. Comparison of expression changes between microarray and RT-PCR. No standard error for miRNA array data because the individual pooled samples were less than three. “FC” means the fold change in expression level of the miRNA in the PFOS group compared with Control.

miRNAs, miR-672, -207, -297, and -1, which acted as molecular mediators of the tested synapse associated proteins. Results showed that expression variations of miRNAs by RT-PCR analysis were consistent with the array analysis data (Figure 2). Protein Levels of VGLUT2, Synaptotagmin I, Synaptotagmin XI, NGFR, TRK-B, and TRK-C in Rat Cortex After PFOS Exposure on PND 1 and 7. As shown in Figure 4, the NGFR expression level significantly decreased on PND 1 (p < 0.05, LSD) and TrkC and VGLUT2 levels significantly decreased on PND 7 (p < 0.05, LSD and Dunnett’s), with no protein level increasing during the test time. There was a significant time effect [F (1, 72) = 6.84, p < 0.05] in expression levels for all tested proteins. With regard to the time effect in the control and PFOS-treated groups, we observed that PFOS exposure induced a significant time-dependent decreasing trend (p < 0.05, Duncan and SNK) for TrkC and VGLUT2 (Figure 3). To better understand the interactive responses of miRNA, mRNA, and protein in relation to PFOS-induced neurotoxicity, the variation levels in tested proteins, mRNAs, and mediating miRNAs were compared based on present miRNA and previous mRNA data (Table 2) in the same animal model.

respectively, indicating that the identification of significantly regulatory miRNAs between TargetScan and mirSVR was divergent on PND 1. In contrast, on PND 7, two computational methods generated the same top regulatory miRNAs, miR-672 and -297, both of which were shown by either TargetScan or MirSVR to be significantly regulatory on PND 1, indicating that miR-672 and -297 played a dominant role in the regulation of PFOS-induced changes in transcript levels. Of mRNA targets for significantly regulatory miRNAs, only Ypel5, Fgfr3, and Syt I were mRNAs common to both TargetScan and MirSVR methods. Global Regulatory Effect of miRNAs on mRNA Expressions after PFOS Exposure. As shown in Figure 1, the R-value distribution peak was below zero on both PND 1 and PND 7 according to both TargetScan and MirSVR, showing that the targeting miRNAs of PFOS-responsive mRNAs were preferentially down-regulated. Biological Implication of PFOS-Responsive mRNA Targets. On PND 7, the most statistically significant GO terms enriched with PFOS-responsive mRNA targets were learning and memory, potassium ion transport, and transmission of nerve impulse as BP, metal ion binding, cation binding, and potassium ion binding as MF, and plasma membrane part, membrane raft, and synapse as CC, and the statistically enriched pathway was phosphatidylinositol signaling system only (P < 0.05) (See Table S6, Supporting Information). No GO term or pathway related to PFOSresponsive targets was observed on PND 1. The brain-specific gene, syt I, was shown to be involved in five BP terms [transmission of nerve impulse (P = 0.010), cell−cell signaling (P = 0.016), regulation of neurotransmitter levels (P = 0.026), neurotransmitter transport (P = 0.036), and vesicle-mediated transport (P = 0.040)], four MF terms along with Syt XI [metal ion binding (P = 0.002), cation binding (P = 0.003), ion binding (P = 0.004), and lipid binding (P = 0.019)], and four CC terms along with Syt XI [plasma membrane part (P = 0.001), synapse (P = 0.007), synapse part (P = 0.009), and plasma membrane (P = 0.014)] (See Table S6, Supporting Information). Confirmatory Studies on Differentially Expressed miRNAs by Real Time RT-PCR. To validate the microarray data, we used RT-PCR to assay the expression levels of four

■

DISCUSSION In the present study, we attempted to explore the interaction between miRNA and mRNA expression to explain the mechanism of PFOS-induced global changes in mRNA expression. However, we found little negative regulatory effect of miRNA on the mRNA expression profile reflected by the few anticorrelated targets based on the mRNA expression data and no GO terms for anticorrelated genes on PND 1, nor on the expression of tested synapse associated proteins. Nevertheless, our study could not conclude whether other synaptic proteins interacted with their regulatory miRNAs and effected synaptic functions with PFOS exposure. As identified by array analysis, the five miRNAs demonstrating the greatest decrease were miR-466b, -672, -297, -674-3p, and -207, respectively, some of which relate to neural functions or particular diseases.11,22 The synaptoneurosome-involved miRNAs MiR-297 and -466b 11 showed parallel reductions on both postnatal days with a 5.6fold (PND 1) and 5.6-fold (PND 7) change for miR-297, and a 10-fold (PND 1) and 5.6-fold (PND 7) change for miR-466b. These coordinated down-regulation changes in miR-297 and 6826

dx.doi.org/10.1021/es3008547 | Environ. Sci. Technol. 2012, 46, 6822−6829

Environmental Science & Technology

Article

(calmodulin 1), which play their respective roles in the synapse transmission.5,21,22 These three mRNAs were coordinately upregulated on PND 7, accompanied by the extensive negative alterations in their targeting miRNAs (miR-672 for SYNJ1, miR-204,-325-3p, -466c, -466b, and -363 for ITPR1, and miR325-3p and -764 for CALM1), indicating that PFOS might affect the calcium actions during synaptic transmission in the nervous system through interference with SYNJ1, ITPR1, and CALM1 via their targeting miRNAs. This implied function has been confirmed by parallel studies in our lab,21,23 showing that PFOS can accumulate in cultured neurons and elevate calcium concentrations via release of intracellular calcium stores, and the action on IP3Rs served a predominant role and could partially account for the perturbation of calcium homeostasis caused by PFOS. The current results could provide evidence for further study on the mechanisms underlying the actions of PFOS on calcium signaling pathways in the nervous synapse. The only three brain-specific gene targets changed on both PND 1 and 7 were Syt I, Syt XI, and Slc17a6, of which, Syt I was involved in many GO terms on PND 7. Therefore, we chose those three synapse-associated transcripts to examine the changes in their coded protein levels. Further, three synapse associated transcripts not affected by PFOS exposure based on the mRNA data but coregulated by PFOS-responsive miRNAs on both PND 1 and 7 were selected for measurement at the protein level. We observed that Syt I was not only the target of the significantly regulatory miRNA, miR-672, but also enriched for some significant GO terms associated with synapse function. Moreover, Syt I and Syt XI were two of only several mRNA targets that were differentially expressed on both PND 1 and 7. These results suggested a high potential interaction between miR-672 and its targets, Syt I and Syt XI. However, we did not find anticorrelation between miR-672 and Syt I or Syt XI expression on PND 1, although the anticorrelative effect was found on PND 7. The explanation for this may relate to the time lag in the changes between miR-672 and Syt I expression, which rendered the mediating effect of miR-672 observable on PND 7 rather than PND 1, or may relate to other factors that regulate transcription via an indirect pathway and partly cooperate with miRNA-672., Synaptotagmin I, a synaptic protein encoded by Syt I, serves as Ca2+ sensors in the process of vesicular trafficking and exocytosis at the synapse.24 However, we failed to observe any change in the Syt I or Syt XI protein levels on PND 1 and 7, indicating the PFOSinduced changes in Syt I or Syt XI transcripts were not functionally translated. For VGLUT2, TrkC, and NGFR, it seems the PFOS-induced miRNA mediation did not lead to the observed alterations in mRNA and protein expression as miRNA always negatively regulates the posttranscriptional and translational expression of its target gene.18 It was recently reported that all TrkC isoforms, but not TrkA and TrkB, function directly in excitatory glutamatergic synaptic adhesion, and can induce differentiation of functional excitatory presynaptic terminals.25 The neuronspecific glutamate transporter most abundantly expressed during embryonic and early postnatal development, VGLUT2,26 serves to terminate the excitatory signal by recycling glutamate after synapse, and thereby avoids excitotoxicity.27 Our findings showed that TrkC but not TrkB decreased in protein levels with VGLUT2 reduction after PFOS exposure, suggesting a potential interference by PFOS in the glutamate recycling and formation of glutamate excitatory presynaptic terminals. Their targeting miRNAs were miR-297

Figure 3. Effect of PFOS exposure on the protein levels of NGFR, TRK-B, TRK-C, SYN-I, SYN-XI, and VGLUT2 on PND 1 and 7. Six individual brain tissues were sampled for each treatment. A p < 0.05 significantly different in PFOS group compared to control for single postnatal day checked by one way ANOVA; B p < 0.05 significantly different on PND 7 compared to PND 1 for single treat group checked by one way ANOVA. “1C” represents Control group from PND 1; “1T” represents PFOS-treated group from PND 1; “7C” represents Control group from PND 7, “7T” represents PFOS-treated group from PND 7. “RV” means relative value of changes in proteins with PFOS treatment to control.

-466b across the entire experimental period, suggesting the potential deterrence of synapse associated functions by PFOS exposure,17 which may involve the action of miR-297 and -466b. According to the distribution of the R-values for PFOSresponsive targets, it is unlikely that the global regulatory effect of miRNAs on mRNA expression was inhibitory, whereas the up-regulation of those anticorrelated mRNAs with PFOS exposure may be partly attributable to the action of miRNAs. We found that the phosphatidylinositol signaling system was the only statistically significant pathway enriched with anticorrelated PFOS-responsive targets on PND 7. The enriched targets for this pathway were SYNJ1 (synaptojanin 1), ITPR1 (inositol 1,4,5-triphosphate receptor 1), and CALM1 6827

dx.doi.org/10.1021/es3008547 | Environ. Sci. Technol. 2012, 46, 6822−6829

Environmental Science & Technology

Article

Table 2. Summary of PFOS-Induced Changes in Protein Levels, mRNA Levels and Associated miRNA levels on PND 1 and 7a protein

TrkB

PND 1 PND 7 mRNA PND 1 PND 7 miRNA PND 1 PND 7

― ― Ntrk2 ― ― -204* up up

TrkC ― down

-207 down down

down ― Ntrk3 ― ― -297 down down

NGFR ― ―

-325−3p down down

― ― Ngf r ― ― -297 down down

Syt I

Syt XI

VGLUT2

Syt-Ib down up -672 down down

Syt-XIb down up -672 down down

Slc17a6b down down -214 ― down

― down

-466b down down

-542-5p ― up

-1 down down

Note. “Up” means up-regulation compared to Control by PFOS-treated group, p < 0.05 checked by One-Way ANOVA. “Down” means downregulation compared to Control by PFOS-treated group, p < 0.05 checked by One-Way ANOVA. “―” means no significant difference between Control and treated group. bData from Wang et al.5

a

■

and -325-3p for TrkC and miR-214 and miR-1 for VGLUT2, respectively (Table 5, SI). Both MiR-214 and -1 decreased with time and the magnitude of their reduction on PND 7 was the same (miR-1 by 0.59, miR-214 by 0.59). The homochronous and paralleling action of these two miRNAs has also been found in previous CNS studies,27−29 where miR-214 and -1 cooperatively played a protective role in neural differentiation and neurodevelopment, suggesting their potential interaction in neurological functions. The observed decrement in their expression in the present study implies an interference with neurodevelopment by PFOS exposure through miR-1 and -214 interactions, but not through the action on their targets because of the positive variation between miRNA and its targets. Results showed that MiR-297, which functions in the synaptoneurosomes, decreased significantly on both PND 1 and 7,11 suggesting a possible perturbation of synapse-associated functions by PFOS exposure. However, it is unlikely that miR-297 acted via direct regulation of its targets, TrkC and NGFR, due to the lack of negative correlation in their changes. Taken together, the PFOS-induced synaptic dysfunctions are likely partly related to the changes in tested proteins in the current study; however, their interaction with miRNAs is unlikely to be involved. Nevertheless, the dramatic statistically altered miRNAs identified in this study could contribute to the synaptic pathway via interaction with targets other than the tested ones, such as CALM1, ITPR1, and SYNJ1 which might play an intermediary role in the synaptic calcium signaling pathway by the miRNA-mRNA interaction.

■

ASSOCIATED CONTENT

S Supporting Information *

Detailed description and presentation of part of methods, tables and figures that are mentioned in the main document. This material is available free of charge via the Internet at http:// pubs.acs.org.

■

REFERENCES

(1) Renner, R. Growing concern over perfluorinated chemicals. Environ. Sci. Technol. 2001, 35, 154−160. (2) Chang, S. C.; Ehresman, D. J.; Bjork, J. A.; Wallace, K. B.; Parker, G. A.; Stump, D. G.; Butenhoff, J. L. Gestational and lactational exposure to potassium perfluorooctane sulfonate (K+ \PFOS) in rats: Toxicokinetics, thyroid hormone status, and related gene expression. Reprod. Toxicol. 2009, 27, 387−399. (3) Johansson, N.; Fredriksson, A.; Eriksson, P. Neonatal exposure to perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) causes neurobehavioural defects in adult mice. Neuro. Toxicol. 2008, 29, 160−169. (4) Luebker, D. J.; Case, M. T.; York, R. G.; Moore, J. A.; Hansen, K. J.; Butenhoff, J. L. Two-generation reproduction and cross-foster studies of perfluorooctanesulfonate (PFOS) in rats. Toxicology 2005, 215, 126−148. (5) Wang, F. Q.; Liu, W.; Jin, Y. H.; Dai, J. Y.; Yu, W. G.; Liu, X. H.; Liu, L. Transcriptional effects of prenatal and neonatal exposure to PFOS in developing rat brain. Environ. Sci. Technol. 2010, 44, 1847− 1853. (6) Olsen, L.; Klausen, M.; Helboe, L.; Nielsen, F. C.; Werge, T. MicroRNAs show mutually exclusive expression patterns in the brain of adult male rats. PLoS One 2009, 4, e7225. (7) Zhang, B.; Pan, X. RDX induces aberrant expression of microRNAs in mouse brain and liver. Environ. Health. Perspect. 2009, 117, 231−240. (8) Moffat, I. D.; Boutros, P. C.; Celius, T.; Lindén, J.; Pohjanvirta, R.; Okey, A. B. microRNAs in adult rodent liver are refractory to dioxin treatment. Toxicol. Sci. 2007, 99, 470−487. (9) Sathyan, P.; Golden, H. B.; Miranda, R. C. Competing interactions between micro-RNAs determine neural progenitor survival and proliferation after ethanol exposure: Evidence from an ex vivo model of the fetal cerebral cortical neuroepithelium. J. Neurosci. 2007, 27, 8546−8557. (10) Hua, Y. J.; Tang, Z. Y.; Tu, K.; Zhu, L.; Li, Y. X.; Xie., L.; Xiao, H. S. Identification and target prediction of miRNAs specifically expressed in rat neural tissue. BMC Genom. 2009, 10, 214. (11) Konecna, A.; Heraud, J. E.; Schoderboeck, L.; Raposo, A. A.; Kiebler, M. A. What are the roles of microRNAs at the mammalian synapse? Neurosci. Lett. 2009, 466, 63−68. (12) Johansson, N.; Eriksson, P.; Viberg, H. Neonatal exposure to PFOS and PFOA in mice results in changes in proteins which are important for neuronal growth and synaptogenesis in the developing brain. Toxicol. Sci. 2009, 108, 412−418. (13) Wang, F. Q.; Liu, W.; Jin, Y. H.; Dai, J. Y.; Zhao, H. X.; Xie, Q.; Liu, X. H.; Yu, W. G.; Ma, J. S. Interaction of PFOS and BDE-47 coexposure on thyroid hormone levels and TH-related gene and protein expression in developing rat brains. Toxicol. Sci. 2011, 121, 279−291. (14) Zeng, H. C.; Zhang, L.; Li, Y. Y.; Wang, Y. J.; Xia, W.; Lin, Y.; Wei, J.; Xu, S. Q. Inflammation-like glial response in rat brain induced by prenatal PFOS exposure. Neurotoxicology 2011, 32, 130−139. (15) Zeng, H. C.; Li, Y. Y.; Zhang, L.; Wang, Y. J.; Chen, J.; Xia, W.; Lin, Y.; Wei, J.; Lv, Z. Q.; Li, M.; Xu, S. Q. Prenatal exposure to

AUTHOR INFORMATION

Corresponding Author

*Phone: +86-411-84708084; fax: +86-411-84708084; e-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS Funding for this project was provided by the National Natural Science Foundation of China (Grant No. 20837004, 21177020, and 30771772). 6828

dx.doi.org/10.1021/es3008547 | Environ. Sci. Technol. 2012, 46, 6822−6829

Environmental Science & Technology

Article

perfluorooctanesulfonate in rat resulted in long-lasting changes of expression of synapsins and synaptophysin. Synapse 2011, 65, 225− 233. (16) Lugli, G.; Torvik, V. I.; Larson, J.; Smalheiser, N. R. Expression of microRNAs and their precursors in synaptic fractions of adult mouse forebrain. J Neurochem. 2008, 106, 650−661. (17) Liao, C. Y.; Li, X. Y.; Wu, B.; Duan, S.; Jiang, G. B. Acute enhancement of synaptic transmission and chronic inhibition of synaptogenesis induced by perfluorooctane sulfonate through mediation of voltage-dependent calcium channel. Environ. Sci. Technol. 2008, 42, 5335−5341. (18) Pillai, R. S.; Artus, C. G.; Filipowicz, W. Tethering of human Ago proteins to mRNA mimics the miRNA-mediated repression of protein synthesis. RNA 2004, 10, 1518−1525. (19) Cheng, C.; Fu, X.; Alves, P.; Gerstein, M. mRNA expression profiles show differential regulatory effects of microRNAs between estrogen receptor-positive and estrogen receptor-negative breast cancer. Genome Biol. 2009, 10, R90. (20) Ma, L.; Huang, Y.; Zhu, W.; Zhou, S.; Zhou, J.; Zeng, F.; Liu, X.; Zhang, Y.; Yu, J. An integrated analysis of miRNA and mRNA expressions in non-small cell lung cancers. PLoS One 2011, 6, e26502. (21) Liu, X. H.; Liu, W.; Jin, Y. H.; Yu, W. G..; Liu, L.; Yu, H. Y. Effects of subchronic perfluorooctane sulfonate exposure of rats on calcium-dependent signaling molecules in the brain tissue. Arch. Toxicol. 2010, 84, 471−479. (22) Marks, B.; McMahon., H. T. Calcium triggers calcineurindependent synaptic vesicle recycling in mammalian nerve terminals. Curr. Biol. 1998, 8, 740−749. (23) Liu, X.; Jin, Y.; Liu, W.; Wang, F.; Hao, S. Possible mechanism of perfluorooctane sulfonate and perfluorooctanoate on the release of calcium ion from calcium stores in primary cultures of rat hippocampal neurons. Toxicol. In Vitro 2011, 25, 1294−1301. (24) Yoshihara, M.; Littleton, J. T. Synaptotagmin I functions as a calcium sensor to synchronize neurotransmitter release. Neuron. 2002, 36, 897−908. (25) Takahashi, H.; Arstikaitis, P.; Prasad, T.; Bartlett, T. E.; Wang, Y. T.; Murphy, T. H.; Craig, A. M. Postsynaptic TrkC and presynaptic PTPσ function as a bidirectional excitatory synaptic organizing complex. Neuron. 2011, 69, 287−303. (26) Miyazaki, T.; Fukaya, M.; Shimizu, H.; Watanabe, M. Subtype switching of vesicular glutamate transporters at parallel fibre-Purkinje cell synapses in developing mouse cerebellum. Eur. J. Neurosci. 2003, 17, 2563−2572. (27) Wallén-Mackenzie, A.; Wootz, H.; Englund, H. Genetic inactivation of the vesicular glutamate transporter 2 (VGLUT2) in the mouse: what have we learnt about functional glutamatergic neurotransmission? Ups. J. Med. Sci. 2010, 115, 11−20. (28) Chen., H.; Shalom-Feuerstein, R.; Riley, J.; Zhang, S. D.; Tucci, P.; Agostini, M.; Aberdam, D.; Knight, R. A.; Genchi, G.; Nicotera, P.; Melino, G.; Vasa-Nicotera, M. miR-7 and miR-214 are specifically expressed during neuroblastoma differentiation, cortical development and embryonic stem cells differentiation, and control neurite outgrowth in vitro. Biochem. Biophys. Res. Commun. 2010, 394, 921− 927. (29) Liu, N. K.; Wang, X. F.; Lu, Q. B.; Xu, X. M. Altered microRNA expression following traumatic spinal cord injury. Exp. Neurol. 2009, 219, 424−429.

6829

dx.doi.org/10.1021/es3008547 | Environ. Sci. Technol. 2012, 46, 6822−6829