This article is mentioned in:

Introduction

Poly(ADP-ribose) polymerases (PARPs) comprise a superfamily containing 17 members in humans (1). Numerous PARP-1, -2, -14 and tankyrases have previously been shown to mediate inflammatory reactions (2–6), and PARP-1 has been associated with inflammatory or oxidative stress-related skin pathologies (7–9). PARP-1 is an important pro-inflammatory mediator with important roles in the maturation of immune cells (10–12), activation of pattern recognition receptors (13–19), activation of pro-inflammatory transcription factors (6), and reactive species (20).

The contact hypersensitivity (CHS) reaction is a T-cell mediated, delayed type hypersensitivity reaction that is elicited by low molecular weight molecules, such as oxazolone (OXA) (21,22). Upon sensitization, infiltrating cells and keratinocytes produce free radicals, pro-inflammatory cytokines, adhesion molecules and matrix metalloproteinases in a PARP-1 and nuclear factor-κB-dependent manner, resulting in the induction of inflammation (6,7,20,23–25). PARP-1 expression inhibition, or knock-out, has been shown to reduce the severity of experimental CHS (7,26). The present study aimed to characterize whether alterations of eicosanoid and docosanoid metabolism and signaling may contribute to the anti-inflammatory effects of PARP-1 expression ablation.

Materials and methods

Materials

OXA and the other chemicals used in the present study were obtained from Sigma-Aldrich (St. Louis, MO, USA), unless stated otherwise.

Animal experiments

All of the experiments performed on mice were approved by the Local Ethical Committee (9/2008/DE MÁB) and were performed according to EU and national guidelines. PARP-1+/+ and PARP-1−/− mice, derived from heterozygote-to-heterozygote breeding, were kept in the animal facility of the Life Science Building, University of Debrecen (Debrecen, Hungary) (27). A CHS model was constructed as described by previous methods (7). Briefly, the mice were randomized into four groups: PARP-1+/+ vehicle sensitized (control), PARP-1−/− vehicle sensitized (control), PARP-1+/+ OXA sensitized (CHS) and PARP-1−/− OXA sensitized (CHS) groups (n=5/5/6/5). Sensitization was carried out on the shaved abdominal wall skin of the mice, with an administration of 100 μl 2% OXA solubilised in acetone: olive oil (4:1; CHS groups), or 100 μl vehicle of acetone:olive oil (4:1; control groups). Seven days after sensitization, all of the mice were challenged with 100 μl 0.5% OXA, which was applied to the epidermis of the shaved back. After 24 h, the mice were sacrificed and the skin was harvested.

Histology and microscopy

Hematoxylin-eosin staining and immunohistochemistry was performed on paraffin-fixed 7 μm tissue sections, as described previously (7).

High performance liquid chromatography-electrospray tandem mass spectrometry (HPLC-ESI-MS-MS) analysis

The procedure for the HPLC-ESI-MS-MS analysis The analysis of free fatty acids, eicosanoids and docosanoids in the skin was performed, according to previously described methods (28), with minor modifications.

Sample preparation

The whole analytical sample preparation procedure used was based on an established method used for retinoid quantification (29). Briefly, 50 mg skin biopsy, was treated with 150 μl acetonitrile and the skin samples were cut into small pieces with scissors, on ice. If less than 50 mg of skin biopsy was present, water was added to yield 50 mg of sample weight. The mixtures were agitated for 3 min, and the precipitated protein was centrifuged at 16060 g (13,000 rpm), 4°C for 6 min. A total of 130 μl resulting supernatant was mixed with 10 μl isotope supernatant mix and evaporated in Eppendorf reaction vials (Eppendorf, Hamburg, Germany) using an Eppendorf concentrator at 30°C for ~60 min, until the sample volume was ~10 μl. The Eppendorf concentrator was vented with argon, in order to prevent degradation of eicosanoids and docosanoids. The dried extract was resuspended in ~25 μl HPLC solvent A [64.3% water (Chromasolv Plus; Sigma-Aldrich, Budapest, Hungary), 35.5% acetonitrile (Merck KGaA, Darmstadt, Germany) and 0.2% formic acid (Sigma-Aldrich, Hungary)] to yield 35 μl. The sample was then vortexed (15 sec), agitated (3 min) and transferred into micro injection inserts vials (Waters, Budapest, Hungary). The glass vials containing 35 μl extract were transferred into brown screw top vials with PTFE/silicone septa (Waters), and placed into the pre-cooled (15°C) autosampler of the Waters 2695XE separation module.

Chromatographic system

The HPLC system consisted of a Waters 2695XE separation module (Waters) including a gradient pump, autosampler, degasser and a heated column compartment. A MS-MS detector, with an ESI ionizing option, (Micromass Quattro Ultima PT; Waters, Elstree, UK; a gift from Biosystems Int., France) was used as the detector. The system was controlled using MassLynx software (Waters).

HPLC conditions

The eluents were degassed in the Waters 2695XE separation module prior to mixing, and passed through an in-line filter (1–2 μm; Knauer, Berlin, Germany) before reaching the analytical column (LiChroCART, 125×2 mm; Superspher 100, RP-18, endcapped; Merck KgaA), which was embedded in the column compartment. A multilinear gradient was formed from solvent A (as mentioned previously) and solvent B (methanol; Merck KGaA). The gradient consisted of the following steps: 0.0 min 20% B, 3.0 min 20% B, 5.0 min 60% B, 15.0 min 100% B, 15.9 min 100% B and 16.0 min 5% B. The flow rate was adjusted to 0.4 ml/min and the column was heated to 40°C. From the same biological extract, 10 μl was used for each HPLC analysis. This step was performed twice, using the same HPLC conditions, and two different MS-MS analysis options were conducted for better resolution and quantification of the various analytes.

MS options

The Micromass Quattro Ultima PT was controlled using MassLynx software. Argon was used with an inlet pressure of 0.8 bar. ESI (electro spray ionization source; Waters) was vented by nitrogen, which was continuously produced by a nitrogen generator (Peak Scientific NM30 Nitrogen generator; Peak Scientific, Billerica, MA, USA) including a compressor (Waters), the inlet flow was set at 3.6×10−3 mbar.

Multiple reaction monitoring settings

ESI, with a negative ESI-setting, was performed with the HPLC eluent, following the ion source temperature of 85°C. The desolvation gas flow was 780 l/h, the desolvation temperature was 400°C, the cone gas flow was 10 l/h, the capillary current was 3 μA, and the cone voltage was 50 V. The aperture voltage was set at 0 V and the range-finder lens voltage was set at 35 V (for one) and 0.2 V (for two). The analyzer settings were LM1 resolution, 14.5; HM1 resolution, 14.5; ion energy, 10.7; entrance, −1; collision, 0 (collision parameters were set for each substance at the MS-method parameters); exit, 2; LM2 resolution, 14.5; HM2 resolution, 14.5; ion energy, 25.0 and multiplier energy, 650 V.

Multiple reaction monitoring settings for polyunsaturated fatty acids (PUFA), eicosanoids/docosanoids semi-quantification

The following methods were used, as described in a previous study (27): Method A from 0.0–9.0 min for PGF2 349.0 -> 192.7, collision energy 22 eV; PD1 and PD1 isomers, like PDX, 359.0 -> 153.3, collision energy 17 eV; TXB2 369.0 -> 195.0, collision energy 13 eV; PGE3 349.0 -> 233.0, collision energy 17 eV; LXA4/LXB4 351.0 -> 115.0, collision energy 12 eV; 8iPGF3 351.0 -> 193.0, collision energy 22 eV; 20-COOH-LTB4 365.0 -> 195.0, collision energy 13 eV; RvD1, RvD2 375.0 -> 141.3, collision energy 13 eV; RvE1 375.1 -> 141.3, collision energy 13 eV. From 9.0–12.5 min for 13-HODE 294.7 -> 170.7, collision energy 16 eV; 9-HODE 294.7 -> 194.7, collision energy 16 eV; 5-HEPE 317.0 -> 115.0, collision energy 17 eV; 12-HEPE 317.0 -> 179.0, collision energy 17 eV; 15-HEPE 317.0 -> 219.0, collision energy 17 eV, LTC4 623.9 -> 272.0, collision energy 14 eV. From 12.5–16.0 min for LA 279.3 -> 279.0, collision energy 10 eV; 8-HEPE 317.0 -> 255.0, collision energy 17 eV; 18-HEPE 317.0 -> 259.0, collision energy 17 eV; 5-oxoETE, 12-oxoETE and 15-oxoETE 317.0 -> 273.0, collision energy 17 eV; 20-HETE 319.0 -> 245.0, collision energy 10 eV; LTC4 623.9 -> 272.0, collision energy 14 eV; LTE4 438.0 -> 333.0, collision energy 13 eV. From 12.5–16.0 min for LA 279.3 -> 59.2, collision energy 25 eV; EPA 301.0 -> 203.2, collision energy 12 eV; AA 303.0 -> 259.3, collision energy 14 eV; DHA 327.1 -> 29.3, collision energy 14 eV.

Method B (27) from 0.0–9.8 min for PGE2, d15d12PGD2, PGD2, d15d12PGJ2 and PGJ2 315.0 -> 271.3, collision energy 13 eV; LTB5 333.0 -> 195.0, collision energy 13 eV; LTB4 335.0 -> 195.0, collision energy 13 eV; RvE1 349.1 -> 195.3, collision energy 13 eV; HXA3, HXB3 and 20-COOH-AA 335.0 -> 273.3, collision energy 13 eV; LXA5 349.0 -> 115.0, collision energy 12 eV; 20-OH-LTB4 351.0 -> 195.0, collision energy 13 eV; MaR 359.0 -> 250.0, collision energy 13 eV. From 9.8–12.5 min for 5-HETE 318.7 -> 115.0, collision energy 14 eV; 8-HETE 319.0 -> 155.0, collision energy 14 eV; 11-HETE 319.0 -> 167.0, collision energy 14 eV; 12-HETE 319.0 -> 179.0, collision energy 14 eV; 15-HETE 319.0 -> 218.9, collision energy 11 eV; 4-HDHA 343.0 -> 101.0, collision energy 10 eV; 10-HDHA 343.0 -> 181.0, collision energy 10 eV; 14-HDHA 343.0 -> 205.0, collision energy 10 eV; 17-HDHA 343.0 -> 245.0, collision energy 14 eV, 20-HDHA 343.0 -> 285.0, collision energy 10 eV; 13-oxoODE 293.0 -> 249.0, collision energy 17 eV. And from 12.5–16.0 min for LA 279.3 -> 59.2, collision energy 25 eV; EPA 301.0 -> 203.2, collision energy 12 eV; AA 303.0 -> 259.3, collision energy 11 eV; and DHA 327.1 -> 229.3, collision energy 14 eV.

Standard solutions

Stock solutions of the PUFAs, eicosanoids and docosanoids were prepared by dissolving the solutions in methanol to yield a final concentration of 10 μg/ml. The solutions were obtained from Cayman-Chemicals (Tallinn, Estonia), BioMol International (Kastel-Med KFT, Budapest, Hungary), Sigma-Aldrich (Hungary), Larodan Lipids (Malmö, Sweden) and Dr. Charles Serhan (Harvard, MA, USA). All stock solutions were stored in the dark, at −80°C until further use. The reference PUFAs, eicosanoids and docosanoids (BioMol International/Enzo Life Sciences, Farmingdale, NY, USA) were used for assay validation.

Quantification

Individual eicosanoids and docosanoids were quantified based on the determination of the area under the curve (AUC), which was compared with the AUC of the standard compounds. To ensure optimal extraction, isotope-labelled standard compounds were used. This analytical procedure was previously established for liquids and tissue analysis (28).

mRNA preparation, reverse transcription and quantitative polymerase chain reaction (qPCR)

Total RNA extraction, reverse transcription, and a subsequent qPCR, were performed as described previously (7). The primers used are listed in Table I.

Table I Oligonucleotides used in quantitative polymerase chain reactions.

Table I

Oligonucleotides used in quantitative polymerase chain reactions.

Gene	Sequence
Mouse 36B4	5′-AGA TTC GGG ATA TGC TGT TGG-3′ 5′-AAA GCC TGG AAG AAG GAG GTC-3′
Mouse GAPDH	5′-CAA GGT CAT CCA TGA CAA CTT TG-3′ 5′-GGC CAT CCA CAG TCT TCT GG-3′
Mouse cyclophyllin	5′-TGG AGA GCA CCA AGA CAG ACA-3′ 5′-TGC CGG AGT CGA CAA TGA T-3′
Mouse FABP7	5′-GAG TAC ATG AAA GCT CTG GGC-3′ 5′-AAC CGA ACC ACA GAC TTA CAGT T-3′
Mouse COX2	5′-GTG AAG GGA AAT AAG GAG CTT CC-3′ 5′-GTG ATT TAA GTC CAC TCC ATG GC-3′
Human 36B4	5′-CCA TTG AAA TCC TGA GTG ATG TG-3′ 5′-GTC GAA CAC CTG CTG GAT GAC-3′
Human actin	5′-GAC CCA GAT CAT GTT TGA GAC C-3′ 5′-CAT CAC GAT GCC AGT GGT AC-3′
Human cyclophyllin	5′-GTC TCC TTT GAG CTG TTT GCA GAC-3′ 5′-CTT GCC ACC AGT GCC ATT ATG-3′
Human PARP-1	5′-CAC TGG TAC CAC TTC TCC TGC TTC-3′ 5′-CTT TGC CTG TCA CTC CTC CAG-3′

Establishment of a PARP-1 knockdown HEK293T cell line and luciferase reporter assays

HEK293T human embryonic kidney cells (American Type Culture Collection, Manassas, VA, USA) were maintained in Dulbecco’s modified Eagle’s medium, supplemented with 4.5 g/l glucose, 10% fetal calf serum, 2.5 μg/ml puromycin. These cells stable expressing a small hairpin (sh)RNA specifically targeting PARP-1 (pLKO.1-PARP1). pLKO.1 was used in the control cells (Sigma-Aldrich, USA). PARP-1 expression was subsequently suppressed, as confirmed by qPCR.

pLightSwitch_PromFABP7 (Switchgear Genomics, Carlsbad, CA, USA) and pCMV-βgal were transfected into the above cell line using jetPEI (PolyPlus Transfections SA, Illkirch, France). Following a 24 h incubation, the luciferase and β-galactosidase activities were determined. Luciferase activity was normalized to β-galactosidase activity.

Statistical analysis

The results of the cell assays are expressed as the means ± standard deviation, and statistical significance between the groups was determined by Student’s t-test. The results of the in vivo studies are expressed as the means, and statistical significance was determined using the Kruskal-Wallis and Mann-Whitney tests. A P<0.05 was considered to indicate a statistically significant difference.

Results

Neutrophil infiltration and edema at the site of the CHS reaction was observed in the histological examinations; however, the levels of these were reduced in the PARP-1−/− mice (Fig. 1A and B). This finding indicates that PARP-1 may contribute to inflammation in CHS, therefore PARP-1−/− mice may be protected against CHS. These results are concordant with our previous observations (7,26,30).

Figure 1 Deletion of poly(ADP-ribose) polymerase-1 (PARP-1) suppresses neutrophil migration. Formalin-fixed, paraffin-embedded tissue sections were stained with (A) hematoxylin and eosin and (B) DAPI counterstain (blue) and an anti-neutrophil elastase antibody (green) to detect neutrophil elastase. The arrows indicate the neutrophil elastase-positive cells. The neutrophil elastase staining was observed to be cytoplasmic. Scale bars: A, 30 μm; B, 10 μm.

The qPCR detected higher mRNA expression levels of fatty acid binding protein (FABP)7 in the PARP-1−/− mice (Fig. 2). The depletion of PARP-1 expression (Fig. 3A) induced the activity of the FABP7 promoter (Fig. 3B), suggesting that PARP-1 may have direct effects on the FABP7 promoter.

Figure 2 Eicosanoid and docosanoid signaling alterations in the skin of poly(ADP-ribose) polymerase-1 (PARP-1)−/− mice, due to enhanced fatty acid binding protein (FABP)7 expression. The FABP7 mRNA expression levels, and eicosanoid and docosanoid concentrations, were determined in the skin of control (PARP-1+/+ NS), OXA-sensitized PARP-1+/+ (PARP-1+/+ SENS) and respective PARP-1−/− mice (n=5/5/6/5). AA, arachidonic acid; DHGLA, dihomo-gamma-linoleic acid; EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid; MET, metabolites; COX, cyclooxygenase; COX2, cyclooxygenase 2; Pg, prostaglandin.

Figure 3 Reduction in the expression levels of poly(ADP-ribose) polymerase-1 (PARP-1) in response to specific small hairpin (sh)RNA treatment of HEK293T human embryonic kidney cells. PARP-shRNA also induced the activity of the fatty acid binding protein (FABP)7 promoter. (A) PARP-1 expression levels were reduced in HEK293T cells using a construct coding an shRNA against PARP-1 (shPARP-1), as determined by quantitative polymerase chain reaction (n=3/3). (B) Promoter activity of FABP7 was determined in luciferase reporter assays in control (scPARP-1 HEK2935T) and shPARP-1 HEK293T cells (n=3/3). *P<0.05, PARP-1+/+ mice, compared with PARP-1−/− mice; ***P<0.001, scPARP-1, compared with shPARP-1 HEK293T cells.

Using a lipidomic approach, the composition of eicosanoids and docosanoids in the skin was determined. There was an increase in four of the major PUFAs, which are precursors of eicosanoids and docosanoids: Arachidonic acid (AA; 20:4 n-6), dihomo-gamma-linoleic acid (DHGLA; 20:3 n-6), docosahexaenoic acid (DHA; 22:5 n-3) and eicosapentaenoic acid (EPA; 20:5 n-3). These free fatty acids had increased levels in the skin of the control and OXA-challenged PARP-1−/− mice, as compared with their respective PARP-1+/+ cohorts (Fig. 2).

The levels of AA-metabolites (AA MET), as compared with the mother compound AA (AA MET/AA), as well as EPA MET/EPA and DHA MET/DHA, had lower ratios in the non-sensitized animals, and the n3-PUFAs ratios were increased in the sensitized animals (Fig. 2). The sum of the cyclooxygenase (COX) metabolites (COX SUM), and the levels of COX metabolites prostaglandin E2 (PgE2) and PgD2, as well as the PgE2/AA and PgD2/AA ratios, were all decreased showing reduced synthesis of pro-inflammatory lipid mediators. The levels of anti-inflammatory or pro-resolving n3-PUFA metabolites were mainly increased, resulting from EPA and DHA-metabolism (31) in the PARP-1−/− animals (Fig. 2). These data are concordant with the decreased expression levels of COX-2 in the PARP-1−/− mice(Fig. 2).

Discussion

In the present study, it was determined that ablation of PARP-1 expression had characteristic effects on skin eicosanoid and docosanoid signaling, and metabolism. The two n-3 PUFAs (DHA and EPA), DHA- and EPA-metabolites, and the ratios of DHA- and EPA-metabolites vs. AA-metabolites were increased, indicating a pro-resolving, anti-inflammatory environment (31) in the PARP-1−/− mice, especially following OXA-sensitization. The formation of the pro-inflammatory COX-metabolites was lower in the skin of the PARP-1−/− mice, which is concordant with a previously determined anti-inflammatory environment (6). PARP-1 has previously been shown to regulate COX-2 expression (32) which may explain the observed decrease in COX-2 metabolites.

It may be suggested that increased FABP7-levels are responsible for the observed alterations of lipid mediator signaling. FABP7 has preference towards n-3, as compared with n-6, PUFAs (33), this may be a plausible explanation for the accumulation of EPA-derived metabolites, and increased ratios of PUFA-metabolites (DHA/AA MET, EPA/AA MET) in the PARP-1+/+ mice.

The results of the present study widen the spectrum of altered lipid metabolism upon PARP-1 expression ablation (34,35), by implicating skin eicosanoid and docosanoid metabolism and signaling. A reduction in PARP-1 expression results in a reduced pro-inflammatory and increased pro-resolving or anti-inflammatory environment in the skin. Higher levels of the DHA and EPA metabolites, and reduced COX-pathway metabolites were previously implicated in a reduction of DNFB-induced CHS (36), peritonitis (31), atopy (37) or allergy (38).

These data are concordant with our previous findings, which showed that genetic or pharmacological inhibition of PARP-1 is capable of protecting against CHS (7,26,39). Notably, differences in eicosanoid and docosanoid metabolism and signaling are also present in unchallenged mice and it may be speculated that these minor changes may render PARP-1−/− mice less susceptible to CHS. Furthermore, the changes to gene expression levels and lipid mediator composition are associated with inflammatory pathologies that are mediated not only by T-helper (Th) 1 or Th2 cells, but also the newly defined Th17 cells, which have previously been associated with the induction of FABPs in psoriasis (40). In conclusion, alteration in skin eicosanoid and docosanoid metabolism and signaling may modify the barrier function of the skin, wherein the role of PARPs remains relatively unexplored.

Acknowledgements

The present study was supported by grants from the National Innovation Office (Baross program Seahorse grant; no. TÁMOP-4.2.2. A-11/1/KONV-2012-0025), OTKA (nos. PD83473, K105872, K108308, CNK80709, K104720 and K108308), and the Medical and Health Science Center (Mecenatura no. Mec-8/2011). PB and MS are recipients of the Bolyai fellowship from the Hungarian Academy of Sciences. The authors of the present study would like to acknowledge the technical assistance of Mrs Erzsébet Herbály.

Abbreviations:

References