Inflammatory-linked changes in CpG island methylation of three opioid peptide genes in a rat model for pain.

Expression of the opioid peptide genes proopiomelanocortin (Pomc), proenkephalin (Penk), and prodynorphin (Pdyn), in immune cells plays a key role in endogenous pain control. In a rat model of painful unilateral paw inflammation, we isolated cells from popliteal lymph nodes and evaluated the role of CpG island C5-methylation on the transcriptional activation of those genes. Using methylated DNA immunoprecipitation, we sorted gDNA into methylated (me) and non-me fractions and then determined the CpG island methylation status of each fraction via quantitative Real Time-PCR (qRT-PCR). In silico analysis by MethPrimer software identified one CpG island in Pdyn and three each in Pomc and Penk. No substantial changes in C5-methylation of any gene were observed. In conclusion, the CpG island methylation status does not seem to be a key regulator of opioid gene activation in immune cells during peripheral tissue inflammation.

Introduction

The expression of endogenous opioid peptides in peripheral tissue can play an important role in pain control [1-3]. In acute inflammation, immune cells express the three opioid peptides beta-endorphin (END), Met-enkephalin (MENK) and Dynorphin A (DynA) [4-10]. Upon their release, antinociception is triggered through binding of the opioid peptides to their corresponding receptors on peripheral sensory nerve endings [11,12].

The opioid peptides END, MENK and DynA are encoded by proopiomelanocortin (Pomc), proenkephalin (Penk) and prodynorphin (Pdyn), respectively. The main expression of Pomc occurs in the pituitary gland where it is regulated e.g. by the transcription factors TPIT and PITX1 [13-15]. Pdyn and Penk are also predominantly expressed in the brain. In addition, all endogenous opioids are synthesized site-specifically in various organs [16-19].

So far, research has been focusing on transcriptional regulation of opioid peptide genes in the brain or on the liberation of opioids from immune cells, but little is known about the transcriptional regulation of opioid gene expression in immune cells.

All three opioid genes are, under certain circumstances, transcriptionally regulated by methylation [20-25]. The most common form of DNA methylation is the addition of a methyl group to the C5 position of cytosine within a CpG dinucleotide which mostly occurs outside CpG islands [26]. The latter is defined as a short interspersed DNA stretch containing increased amounts of CpG dinucleotides with a GC content > 0.5 and an observed vs. expected CpG ratio > 0.6 [27].

Earlier studies showed an upregulation of opioid gene expression in popliteal lymph nodes (LNs) after peripheral inflammation [10,28,29]. To investigate a possible transcriptional regulation via methylation, this paper focusses on total CpG island methylation of gDNA of Pomc, Penk and Pdyn in isolated CD45RA± cells of naïve and inflamed popliteal LNs.

Materials and methods

Animal housing and induction of inflammation

Animal housing and experiments were carried out after approval by the local animal care committee (Landesamt für Gesundheit und Soziales, Berlin; H0370/09) and strictly met the guidelines of the International Association for the Study of Pain [30]. Male WISTAR rats (Janvier Labs) were kept in cages lined with ground corncob bedding with standard laboratory rodent chow and tap water available ad libitum. A light cycle of 12/12 hours light/dark, room temperature of 22°C and humidity between 40–60% were maintained. Under brief anesthesia with Forene® (Abbvie) rats received an intraplantar injection of 0.15 ml Freund’s Complete Adjuvant, Modified, Mycobacterium butyricum (CFA) (Merck Millipore) into the right hind paw. After 96 h of inflammation, animals were killed by an overdose of inhalation anesthesia (Forene®). Naïve animals were left untreated.

Preparation of CD45RA+- cells from LNs

In case of CFA-injected animals, the popliteal LN of the right hind paw, and in case of naïve animals, both popliteal LNs and the cervical LNs were dissected and kept on ice in RPMI 1640 medium (Gibco). Cell suspensions were prepared by homogenizing LNs with scalpel blades and plastic pestles. Remaining cell clumps were excluded using 70 and 40 μm cell strainers. Cells were counted using a Neubauer Chamber, centrifuged and resuspended in 80 μl cold RPMI 1640 medium per 107 cells. Magnetic cell sorting was performed using rat CD45RA MicroBeads, MS or LS columns, the VarioMACS Separator and the MACSmix Tube Rotator (all Miltenyi Biotec) following the manufacturer’s protocol and using RPMI 1640 medium instead of buffer. Collected cells from flow-through and eluate fractions were pelleted at 450 x g for 10 min (4°C), washed with PBS (Gibco) and stored in 1.5 ml tubes at– 80°C until further use.

gDNA isolation and shearing

The isolation of gDNA from CD45RA+ or CD45RA- cells was performed by adding 700 μl lysis buffer (50 mM Tris-HCl; pH 7.5); 100 mM EDTA; 100 mM NaCl; 1% SDS) and 0.5 μg/μl proteinase K to a frozen cell pellet, followed by incubating at 55°C until the pellet was lysed. The lysate was ultrasonicated for 2 x 20 s at 70% power using the SONOPULS Ultrasonic homogenizers (BANDELIN) before centrifuging for 15 min at 2500 x g. The supernatant was transferred to a fresh 2 ml tube and NaCl was added to a final concentration of 1 M. For gDNA precipitation 1 vol of isopropanol was added, mixed by inverting the tube and centrifuged for 15 min at 2500 x g. The supernatant was discarded and the pelleted gDNA was washed twice with 70% EtOH (15 min; 4°C; 2500 x g). The gDNA pellet was dried at room temperature and dissolved in TE buffer (10 mM Tris; 1 mM EDTA; pH 8).

Methylated DNA immunoprecipitation

For gDNA shearing the gDNA was diluted in TE buffer to a final concentration of 0.1 μg/μl and 300 μl were ultrasonicated to a mean size of about 500 bp. The size of DNA was verified on 1% agarose gels using electrophoresis. Methylated DNA immunoprecipitation (MeDIP) was performed using the MethylCap kit and the Diagenode Magnetic Rack (both Diagenode) according to the manufacturer’s manual. For elution of the methylated (me) gDNA fraction only High Elution Buffer was used. To purify DNA after MeDIP the QIAquick PCR Purification Kit (Qiagen) was utilized following the manufacturer’s instructions. To elute DNA the columns were loaded twice with 50 μl PCR grade H2O. All gDNA samples were stored at 4°C.

CpG island prediction

To predict CpG island positions for Pomc, Penk and Pdyn the MethPrimer software [31] was used. The following Ensembl sequences [32] were analyzed: ENSRNOT00000016976 (Pomc); ENSRNOT00000089318 (Penk); ENSRNOT00000037576 (Pdyn).

qRT-PCR

For qRT-PCR the Mastercycler® ep realplex, twin.tec 96 real-time-PCR Plates, Masterclear® real-time-PCR Films (all Eppendorf) and the universal EXPRESS SYBR® GreenER™ qPCR Supermix (Life Technologies) were used with the following PCR program: 2 min 50°C; 2 min 95°C; 15 s 95°C; 10 s 61°C; 20 s 78°C; 50 cycles. An additional melting-curve-analysis was performed. A standard 20 μl reaction contained 1 x SYBR Mix, 0.2 μM primer and 2 μl gDNA/PCR grade H2O. Primers (TIB MOLBIOL) and sequences (forward (F)/ reverse (R)) are listed in Table 1. Rat histone H2B type 1-A (TSH2B) and glyceraldehyde-3-phosphate dehydrogenase (Gapdh) primers (Diagenode) served as positive control for me and non-me gDNA, respectively. Relative transcript levels were sustained by subtracting the threshold cycle (Ct) value of the total input from the corresponding Ct value of the gDNA-IP. Differences in gDNA levels were calculated according to the expression 2-Δ.

Table 1

CpG islands in rat opioid genes and corresponding primers/ sequences.

gene	CpG island	genomic region	primer name	oligonucleotide sequence
Pomc	CpG1	promoter	Pomc CpG1FPomc CpG1R	5’-GCACTTTCCAGGCACATCT-3’5’-CTTCTGCAACGCAACAAGC-3’
	CpG2	exon 1 and intron 1	Pomc CpG2FPomc CpG2R	5’-GCGACAGGTAAGGGTGTCTC-3’5’-CAGAACGCCAGTCTGCATTA-3’
	CpG3	mostly exon 3	Pomc CpG3.1FPomc CpG3.1RPomc CpG3.2FPomc CpG3.2R	5’-GTTTCCAGGCAACGGAGAT-3’5’-AGCACTGCTGCTGTTTCTGG-3’5’-GAGTTCAAGAGGGAGCTGGA-3’5’-GGAAGTGCTCCACCCGATAG-3’
Penk	CpG1	promoter	Penk CpG1.1FPenk CpG1.1RPenk CpG1.2FPenk CpG1.2R	5’-CCTTCGGTTTGGGGCTAAT-3’5’-AAAGAAGGCAAGTGTGAGGTG-3’5’-ACTTTCTCGGGGTTCCTCAT-3’5’-GCTTAGGGCAAGATTCTGGA-3’
	CpG2	end of exon 1 and intron 1	Penk CpG2FPenk CpG2R	5’-AGCCAGGACTGCGCTAAAT-3’5’-TTGGCAGAACAGTCCCTCAT-3’
	CpG3	exon 2	Penk CpG3FPenk CpG3R	5’-CTATGGGGGCTTCATGAGAA-3’5’-TCCGAGGGTAGAGACTCAGC-3’
Pdyn	CpG1	exon 4	Pdyn CpG1FPdyn CpG1R	5’-AAACGCTATGGGGGCTTC-35’-ACCGAGTCACCACCTTGAAC-3’

Statistics

The Shapiro-Wilk normality test followed by Friedman Test and Dunn’s multiple comparison test were performed to evaluate statistical significance within a group of naïve or CFA samples. To compare between naïve and CFA samples the Mann-Whitney U test was performed. p values <0.05 were considered significant. Data were expressed as means ± SD. Statistics were performed using GraphPad Prism version 5.02 for Windows (San Diego, California, USA).

Results

CpG island detection in rat opioid genes

Using the MethPrimer software for identifying CpG islands within the three opioid peptide coding rat genes, three CpG islands were found in each Pomc and Penk, and one in Pdyn (Fig 1A) (Table 1).

Fig 1

CpG island determination in rat opioid genes and methylation status of Pomc CpG islands in CD45RA± LN cells of naïve and CFA-treated rats.

Schematic structure (a) of Pomc, Penk and Pdyn including exons (E1-3, highlighted in grey), calculated CpG islands (CpG1-3, highlighted in black) and qRT-PCR amplicon positions (├┤). For Pomc CpG1 (b) and Pomc CpG2 (c) one amplification region was selected each, and two were selected for Pomc CpG3 (3.1 and 3.2) (d+e) using qRT-PCR. Means ± SD; */# P < 0.05; **/## P < 0.01; *** P < 0.001; Friedman test with Dunn’s multiple comparison post hoc test (* non-me vs. me) and MWU test (# CFA vs. naïve); n = 7 (a, b, c+d naïve), n = 8 (c+d CFA).

Methylation of Pomc

Under all conditions the majority of DNA fragments was found in the non-me fractions (albeit not always statistically significant) and CFA treatment resulted in decreased amounts of me DNA (Fig 1B–1E). The latter differences were statistically significant in CD45RA- cells (CpG1+3) (Fig 1B, 1D and 1E). CFA treatment significantly increased the number of amplicons in the non-me fraction of CpG3.2 in CD45RA+ cells (Fig 1E). Differences between CD45RA- and CD45RA+ cells were not observed.

Methylation of Penk

The majority of DNA for Penk in CpG1+2 was not methylated (Fig 2A–2C). CFA treatment resulted in small (but statistically significant) decreases of me DNA in CpG1+2 of CD45RA+ cells and in CpG2 of CD45RA- cells (Fig 2B and 2C). At CpG3 mainly me DNA was found (Fig 2D) with no differential expression resulting from CFA treatment. No differences between CD45RA- and CD45RA+ cells were observed.

Fig 2

Methylation status of Penk CpG islands in CD45RA± LN cells of naïve and CFA-treated rats.

For Penk CpG1 two amplification regions 1.1 (a) and 1.2 (b) were selected, and one region each for Penk CpG2 (c) and Penk CpG3 (d) were chosen for qRT-PCR. Means ± SD; */# P < 0.05; **/## P < 0.01; Friedman test with Dunn’s multiple comparison post hoc test (* non-me vs. me) and MWU test (# CFA vs. naïve); n = 7 (a, c, d, b naïve) and n = 8 (b CFA).

Methylation of Pdyn

DNA in the Pdyn CpG island almost exclusively amplified in the me DNA fraction under all conditions (Fig 3A). In naïve cells, this difference was statistically significant. No differences in methylation status between CFA-treated and naïve LN or between CD45RA- and CD45RA+ cells were found.

Fig 3

Methylation status of Pdyn CpG island in CD45RA± LN cells of naïve and CFA-treated rats and MeDIP controls.

Relative DNA amounts of the Pdyn CpG island DNA in each gDNA fraction were measured via qRT-PCR (a). gDNA separation via MeDIP was verified with control primers for Gapdh (non-me) (b) and rTSH2B (me) (c). Means ± SD; # P < 0.05; **/## P < 0.01; *** P < 0.001; Friedman test with Dunn’s multiple comparison post hoc test (* non-me vs. me) and MWU test (# CFA vs. naïve); n = 7 (a); n = 19 (b CFA); n = 18 (c CFA); n = 14 (b+c naïve).

MeDIP controls

The control amplification of the CpG site in Gapdh (Fig 3B) resulted in a nearly exclusive enrichment in the non-me DNA fraction, whereas for TSH2B (Fig 3C) the accumulation occurred only in the me DNA fraction. No differences in methylation status between CFA-treated and naïve LN or between CD45RA- and CD45RA+ cells were found.

Discussion

Painful paw inflammation has been shown to increase the expression of Pomc mRNA and its protein product END in immune cells of LNs draining an inflamed hind paw. In 2007, Sitte et al. [10] showed the increased END production in LN cells after paw inflammation and the expression of several splice variants of Pomc mRNA. Maddila et al., 2016 [33] showed that mRNA expression of Pomc was upregulated in LN B-cells after inducing paw inflammation and that mRNA and protein expression of the POMC processing enzymes PC1 and PC2 also increased shortly thereafter. After 24 h of inflammation, a significant increase of END in LN B-cells was confirmed. To further elucidate the underlying mechanisms, the present study focused on DNA methylation, a common means of transcriptional regulation. Since not only END but also ENK and DynA were shown to contribute to endogenous pain control through immune cells [4,8] all three opioid peptide genes were investigated here.

To analyze DNA methylation not only the promoter regions were examined but also exons and introns because alterations in methylation pattern of 5´-upstream regions can also influence transcription [34-36]. This has already been shown for POMC by Kuehnen et al., 2012 [37], who found a hypermethylation at the boundary of intron 2 and exon 3 associated with obesity. Also PENK expression was shown to be regulated by methylation in the first exon in vitro [38].

Several forms of methylation are known, including 5-methylcytosine, 5-hydroxymethylcytosine or the rather unfrequent N6-methyladenosine. Since 5-methylcytosine is the most prevalent and the regulation of all three opioid genes has been associated with C5-(de)methylation, it was chosen for this study [23,25,39]. The in silico identification of CpG islands in all opioid peptide genes (Fig 1A) also suggests epigenetic regulation.

Methylation of CpG islands is generally associated with gene silencing whereas demethylation usually correlates with gene activation [40,41]. We therefore expected DNA hypomethylation in LN cells following CFA treatment. Our results show that there are no drastic differences in methylation of opioid peptide genes with or without paw inflammation, or between CD45RA+ and CD45RA- LN cells in both conditions. Several CpG sites in Pomc (Fig 1B–1D) and Penk (Fig 2B and 2C) show a small but statistically significant decrease in the amount of me DNA comparing CFA-treated and naïve LN cells. Taking into account that in all cases the majority of amplified DNA was found in the non-me DNA fraction, those differences do not seem to be of biological relevance. This suggests that epigenetic regulation via methylation is not particularly prominent. However, a striking difference in the methylation status comparing the different CpG islands of Penk has to be noted. Finding the upstream CpG islands generally unmethylated but the downstream CpG3 highly methylated gives a hint that DNA methylation does have a certain impact on Penk expression.

The following limitations need to be considered: We examined the general CpG island C5-methylation status of the three major opioid peptide genes using MeDIP. The MeDIP assay investigates defined DNA regions for cumulations of C5-methylation, but it is impossible to look at single CpG dinucleotides or other methyl-associated DNA-modifications such as e.g. 5-hydroxymethylcytosine (5-hmC). Thus, only those transcriptional regulation processes can be identified, which are mediated by general C5-(de)methylation of an entire DNA segment including several CpG sites. However, transcriptional regulation can also be initiated by few or single CpG sites [42,43]. To obtain information about single CpG methylation, two common tools, DNA bisulfite modification or methyl-sensitive restriction enzymes could be used for future analysis, keeping in mind that both strategies also have limitations. For example, not every CpG site binds restriction enzymes. Bisulfite treatment would result in single CpG resolution. However, DNA degradation, high GC-contents and 5-hydroxymethylation hinder analysis [44]. The MethylCap kit used here is limited by the quality of DNA and the MethylCap protein. However, both are rather unlikely to influence our results. The DNA preparation was checked using agarose gels and, next to the positive controls (Gapdh and TSH2B), both me and non-me DNA sections were identified in the opioid peptide genes, thus supporting the functionality of the assay and of the MethylCap protein in particular. A false contribution of hemimethylated DNA to the results is also unlikely, since the MethylCap protein has low affinity for hemimethylated DNA [45] and mainly non-me regions were identified. PCR experiments were conducted using specific primers, designed to each span the region with the greatest CpG appearance. For CpG-islands longer than 300 bp, two amplification regions were chosen, in both cases showing the same trend in the C5-methylation status.

Conclusion

In conclusion, CpG island C5-methylation is apparently not the key regulator of transcriptional activation of opioid peptide genes in immune cells from rats with acute paw inflammation. The MeDIP assay is an appropriate tool to analyze the general methylation status of specified genomic regions, as demonstrated by the opposing results for different CpG islands. Nevertheless, our results cannot provide a general prediction without further experiments. Investigations of single C5pG methylation, other methylation-based DNA modifications and CpG island shores in different stages of inflammation are necessary to fully elucidate DNA methylation as a regulating element of opioid peptide expression in immune cells.

Data sets for Figs 1–3.

The table lists the individual values for each bar in Figs 1–3 including the values for means ± SD.

(PDF)

Click here for additional data file.