Effect of Tumor Necrosis Factor-α on in vitro Prostaglandin Production in Buffalo Corpus Luteum

Premature luteal regression or inadequate luteal function is a major cause of early embryonic mortality in domestic ruminants including buffaloes. Early embryonic mortality is a major cause of reproductive failure in buffaloes during pre-implantation period (Mondal and Prakash 2002; Mondal et al., 2009; Mondal et al., 2010) and accounts for 20% loss in buffaloes living closer to equator (Vale et al., 1989). Prostaglandins namely prostaglandin E2 (PGE_α) and prostaglandin F2a (PGF_2α) play important role during maternal recognition of pregnancy, implantation and establishment of pregnancy in livestock including buffalo (Nandi et al., 2012; Mondal, 2013). The life span of CL is controlled by complex interactions among various luteotrophic and luteolytic factors including prostaglandins. Prostaglandin E₂ protects the CL from spontaneous regression and play important role in maintenance of pregnancy (Akinlosotu et al., 1986; Christenson et al., 1994) whereas PGF_2α primarily secreted from the surface epithelium of the uterus cause regression of CL (Davis and Rueda, 2002). In spite of PGF_2α produced from endometrium, its secretion is also reported from CL in cows (Townson and Pate, 1994, 1996) and pigs (Diaz et al., 2000). The small amounts of uterine PGF_2α stimulate the production of PGF_2α in the CL which indicates the presence of a positive feedback pathway in the ruminant CL (Diaz et al., 2002).  PGF_2α can autoregulate its synthesis in CL and it has been proposed that PGF_2α produced locally in CL acts via a paracrine and/or autocrine mechanism to induce luteolysis (Wiltbank and Ottobre, 2003). Progesterone, oxytocin and various cytokines have been shown to affect luteal PGF_2α production in different species (Pate, 1988; Wiltbank and Ottobre, 2003, Tripathi et al. 2018, Grazul et al., 1989).

Tumour Necrosis Factor-alpha (TNF_α), a cytokine possessing wide biological actions, is mainly produced by ovarian macrophages and increased during regression of the CL (Penny et al. 1999). TNF_α participates in the regression of the corpus luteum in bovine (Pate, 1995). Highest expression of its specific receptors (TNF-RI) is present in the bovine corpus luteum during luteolysis (Friedman et al. 2000; Korzekwa et al. 2008). The luteolytic action of TNF_α may be mediated by the intraluteal production of PGF2a as TNFa dramatically stimulated PGF_2α production by cultured luteal cells of cow (Benyo and Pate, 1992).  There is an integrated role of luteal PGF_2α and PGE₂ in auto-regulation of CL function (Arosh et al., 2004). Intrauterine administration of PGE₂ had been found to protect the CL from spontaneous and/or induced luteolysis in ruminants (Pratt et al., 1977). The production of endometrial prostaglandins is governed by the enzymes prostaglandin E and F synthase (PGES and PGFS) which convert PGH₂ to PGE₂ and PGF_2α, respectively. There is paucity of information regarding the role of TNFa in regulation of PGF2a and PGE2 secretion and expression of PGES and PGFS mRNA in buffalo CL. Keeping this in view the present work was undertaken to investigate the effect of TNFa on in vitro prostaglandin production and expression of PGES and PGFS mRNA in buffalo CL.

Collection of corpus luteum and identification of stage of estrous cycle
 
Buffalo ovaries with CL were collected from a local abattoir within 1 hr after exsanguinations. The ovaries were then submerged in ice-cold physiological saline (0.9% NaCl; pH 7.4) containing antibiotics [penicillin (10 IU/ml), streptomycin (100 g/ml), amphotericin (2 g/ml) and L-glutamine (100 μg/ml)] and transported to the laboratory. The stages of the estrous cycle were determined according to previously described methods (Mondal et al., 2004; Ghosh and Mondal, 2006, Chouhan et al., 2013) and were categorized to Stage I (days 3 to 5 of the cycle), Stage II (days 6 to 15 of the cycle) and Stage III (days 16 to 21 of the cycle). The Stage I is the developing stage, Stage II is mature one while Stage III is in the regressed stage of corpus luteum. The mid-luteal phase of estrous cycle was determined by macroscopic examination of the CL by observing the change of colour from dark red to orange, firm in consistency on palpation and demarcated from rest of the ovary (Mondal et al., 2013).
 
Luteal cell Isolation
 
Mid stage corpus luteum were immediately separated from the surrounding tissues of the ovaries, chopped, rinsed a few times with Hanks Balanced Salt Solution (HBSS, H9394 Sigma) medium supplemented with gentamycin (50 μg/μl, pH 7.4, TC026, Hiimedia) and 0.1% bovine serum albumin (BSA, RM3158, Himedia) and incubated at 37°C for 1 hr in HBSS containing 0.1% collagenase (Type II collagenase from Clostridium histolyticum, pH 7.4, C6885, Sigma). The cell suspension obtained from the digestion was filtered through plastic strainer (70 μM) to remove undissociated tissue fragments. There after the filtrate was washed three times by centrifugation at 600 x g for 10 min with HBBS supplemented with gentamycin and 0.1% BSA. The number of viable cells that excluded Trypan blue was counted using a haemocytometer.   
 
Culture of luteal cells and treatment with TNFa
 
After cell counting and viability determination the luteal cells were seeded at the rate of 1×105 viable cells in RPMI (Roswell Park Memorial Institute) 1640 medium supplemented with 100µg glutamine, 1µl gentamicin (50 µg/µl; TC026-1G, Himedia) and 10% FBS (Fetal Bovine Serum; F2442, Sigma) and were treated with increasing doses of TNFa (0.1, 0.5 and 1.0 nM) followed by incubation at 38.5°C and 5% CO2 level for 24 hours. Cells along with spent media were harvested and centrifuged at 8000 rpm for 6 minutes to isolate cell pellet. The spent media was collected for quantification of prostaglandins. The cell pellet was kept in 100 µl RLT-lysis buffer (Qiagen, Germany) containing β-mercaptoethanol and stored at -80°C for expression of PGEs and PGFS mRNA.
 
Quantification of prostaglandins
 
PGF2a concentrations were estimated in 50 ml aliquots of culture medium after 10fold dilution with extraction buffer using ELISA kit (404710, Neogen, USA). The sensitivity of the assay was 0.002 ng/ml. The cross reactivity of the antisera against 6-keto prostaglandin F1a, 13, 14 dihydro-15 keto-prostaglandin F2a, prostaglandin D2 and prostaglandin E2 were 3.05%, 0.05%, 0.05% and <0.01%, respectively. The intra- and inter-assay coefficients of variation were less than 15%. Concentrations of PGE2 were estimated in 50 ml aliquots of culture medium after 5fold dilution with extraction buffer. The sensitivity of the assay was 0.002 ng/ml. The cross reactivity of the antisera against 6-keto prostaglandin F1, 13,14 dihydro-15 keto-prostaglandin F2a, prostaglandin D2 and prostaglandin F2a were >0.01% for all. Intra and inter-assay coefficients of variation were less than 13%.
 
RNA isolation
 
Cultured cell pellets of both the treatment and control groups were used for the extraction of total RNA using RNeasy Minikit (M/s Qiagen, Germany) as per manufacturer’s  protocol and quantified by nanodrop spectrophotometer (Eppendorf, Germany) at 260 nm and 280 nm. RNA integrity was checked by formaldehyde-agarose gel (1%) electrophoresis and visualized under UV light after staining with ethidium bromide.
 
Reverse transcription polymerase chain reaction (RT-PCR)
 
cDNA synthesis kit (Biorad iScriptTM cDNA synthesis kit, 1708890, Biorad, USA) was used to synthesize cDNA following the protocol - I script reverse transcriptase (1μl), 5X iScript reaction mix (4µl), Nuclease free water (14 µl) and (100 ng) RNA (1 µl) to make total volume 20 µl. The reactants were  incubated at 25°C for 5 min, 42°C for 30 min followed by 85°C for 5 min. Following first strand synthesis, PCR amplification of PGES and PGFS were carried out using gene specific primers designed on the basis of aligned nucleotide sequences available in GenBank (Table 1). PCR amplification was carried out in 25 µl total volume. Reaction mixture contained of 2 µl first strand of cDNA, 10X Buffer (2.5 µl), 50 mM dNTP Mix (0.5 µl), 25 mM MgCl2 (2.5µl), forward primer (1 µl), reverse primer (1 µl), Mili Q water (15.3 µl) and 0.2 µl Taq DNA Polymerase (5U/µl). Following initial denaturation at 95°C for 5 min, amplification (30 cycles) was carried out under the conditions mentioned in Table 1, followed by final extension at 72°C for 10 min. The generated cDNA fragments were resolved by agarose gel (1%) electrophoresis.


 
Real time PCR
 
PCR analyses were performed with three replicates per sample of each gene using the STEP ONE PLUS real time PCR system (Applied Bio systems, USA), Fast SYBR Green Master Mix (KK4604, Applied Bio systems, USA) and gene specific primers for both housekeeping and target genes (Table 1). Primer 3 program (Rozen and Skaletsky, 2000) was used for designing the primers with an annealing temperature of 59°C and amplification size of less than 250 bp. Primer efficiencies were checked by 10fold serial dilution of cDNAs (Svec et al., 2015) and ranged between 95 and 105%. β-actin was used as endogenous control. Thermal cycling was carried out as per manufacturer’s protocol (95°C, 20 sec followed by 40 cycles of 95°C, 3 sec and 60°C, 25 sec and a melt curve of 95°C (15 sec), 60°C (1 min) and 95°C (15 sec). The specificity of each PCR product was determined by melt curve analysis and amplicon size determination was carried out by agarose gel electrophoresis (2%). Negative control consisted of all the components of qRT-PCR mix except cDNA was used for all primers. The relative quantification of gene expression changes were recorded after normalizing for β-actin expression computed by using 2DDCT method (Svec et al., 2015) in which CT value from controls serves as calibrator.
 
Statistical analysis
 
The results are shown as the Mean±S.E.M. of values obtained from separate experiments, each performed in triplicate. The differences in concentrations of PGF2a, PGE2 and transcript abundance were analyzed by ANOVA followed by Tukey’s multiple comparison tests using the statistical package of Graph Pad Prism 5, San Diego, USA.  A value of p<0.05 was considered statistically significant.

The effect of various doses of TNFa (0.1 nM, 0.5 nM, 1.0 nM) on in vitro PGF2a  production in buffalo luteal cells is presented in Fig 1(a). The concentrations of PGF2a. were 110.2±11.53, 122.6±19.04, 124.6±22.68, 130.4±15.96 pg/ml in control, 0.1 nM, 0.5 nM and 1.0 nM TNFα treatment group, respectively. The effect of various doses of TNFα (0.1 nM, 0.5 nM and 1.0 nM) on in vitro PGE₂ production in buffalo luteal cells is presented in Fig 1 (b). The levels of PGE2 were 2.27±0.36, 2.84±0.37, 2.99±0.42 and 3.17±0.57 pg/ml in control, 0.1 nM, 0.5 nM and 1.0 nM TNFa, respectively.



The effect of different doses (0.1 nM, 0.5 nM, 1.0 nM) of TNFa on PGES and PGFS mRNA expression is presented in Fig 2(a) and 2(b). TNFa causes dose dependent increase in PGFS mRNA expression. However, there was no significant difference in PGFS expression among different groups. There was non-significant (P>0.05) decrease in PGES mRNA expression at 0.1 nM and 0.5 nM but increase at 1.0 nM concentrations of TNFa in comparison to control group. However, there was no significant differences (P>0.05) in PGES and PGFS mRNA expression among various groups.

​

To the best of our knowledge and belief, this is the first study to report the effect of TNF_α on in vitro prostaglandin production and expression of PGFS and PGES mRNA in buffalo Corpus luteum. The present study showed that TNF_α. stimulated PGF_2α and PGE₂ production by cultured buffalo luteal cells. There was dose dependent increase in the concentrations of both PGF_2α and PGE₂ with the increasing doses of TNF_α. However, no significant difference in PGF_2α and PGE₂ concentrations were found among the different treatment groups. Since early corpus luteum is relatively resistant to PGF_2α, mid stage CL was chosen in the present study to see the effect of TNF_α on luteal cell PGF_2α production. Our results are in alignment with the earlier report of Benyo and Pate (1992) wherein TNFa stimulated PGF_2α secretion by the cells in a dose-dependent fashion in mid-stage bovine CL. Similarly, Sakumoto et al., (2000a) reported that TNFa significantly stimulated both PGF_2α and PGE₂ secretion in the bovine luteal cells in a dose-dependent fashion. In porcine species also TNFa had been shown to stimulate both PGF_2α and PGE₂ secretion in a dose-dependent manner in cultured mid luteal cells (Miyamoto et al., 2002). The production of both PGE2 and PGF2a was dose-dependently increased by recombinant human TNF_α (rTNF_α) in cultured human granulosa-lutein cells (Wang et al. 1992). However, our results are not in agreement with the findings of Skarzynski et al., (2003) and Korzekwa et al., (2008) who reported that infusion of lower concentration of TNFa caused the increase in plasma PGF2a level and luteolysis whereas its higher concentration stimulated PGE₂ level and prolongs the estrous cycle  in cattle. (Skarzynski et al., 2003, Korzekwa et al., 2008) which may be due to species difference.  The stimulatory effect of TNF_α on PGF2_α/PGE₂ had also been shown in various tissues of different species including bovine endometrium (Miyamoto et al., 2000; Okuda et al., 2002; Skarzynski et al., 2003), equine endometrium (Szóstek et al. 2014), porcine endometrium (Blitek and Ziecik, 2006; Waclawik et al., 2010), porcine maternal placenta (Jana et al., 2008), human luteal phase endometrial cells (Chen et al., 1995) and human fetal membranes (Kent et al., 1993).

Most prostaglandins are produced from arachidonic acid (AA) which is released from the membrane phospholipids via the action of Cytosolic phospholipase A₂ (cPLA₂) (Wiltbank and Ottobre, 2003). Cyclooxygenases (COX-1 and COX-2) convert AA into PGH₂, the common metabolite for various PGs including PGE₂ and PGF_2α. PGE synthase (PGES) and PGF synthase (PGFS), catalyze the formation of PGE₂ and PGF_2α respectively from PGH₂ (Smith and Dewitt, 1996). The bovine CL possesses all of the PGE₂ and PGF_2α machineries necessary for autoregulation of its function and PGE₂ and PGF_2α machineries are selectively and specifically expressed in CL (Arosh et al., 2004). The present study showed that TNFa increased PGFS mRNA expression in dose dependent manner. Our findings gain support from the finding of Korzekwa et al., (2011) who reported that TNF_α increased mRNA expression of PGFS in immortalized bovine luteal endothelial cell line. In our results the stimulatory effect of TNF_α on both PGF_2α and PGE₂ production indicated that the regulation of synthesis of both PGs by TNFa may be at some common sites like substrate availability, phospholipase A₂ or cyclooxygenase enzyme activity as well at PGFS level in case of PGF_2α synthesis but it may not be at PGES level in case of PGE₂ synthesis as Parent et al., (2002) reported that TNFa stimulated PGE₂ output but not COX-2 expression. Therefore, it is possible that TNF_α may stimulate the production of PGE₂ by increasing the availability of arachidonic acid which is the common substrate for PGs synthesis. Another source of PGF_2α is PGE₂, which could be converted to PGF_2α by PGE₂ 9-keto reductase (Duffy, 2015).  TNFa stimulated PGF_2α secretion through stimulation of phospholipase A₂ (PLA₂) and MAPK pathways in bovine luteal cells (Sakumoto et al., 2000b). TNFa is potent stimulators of luteal prostaglandin production and appears to act primarily by stimulating phospholipase A₂, (Pate, 1995). TNF_α stimulated PGF_2α production by luteal cells is dependent upon the stimulation of phospholipase A₂ through mechanisms requiring synthesis of RNA and protein (Townson and Pate 1996). The mechanism of action of TNFa on PGE₂ is not well documented in luteal cell. However, Waclawik et al., (2010) reported that TNFa significantly stimulated PGE₂ synthesis and release through up-regulation of prostaglandin-endoperoxide synthase 2 mRNA, as well as PGE₂ synthase (mPGES1) mRNA and protein expression in endometrial luminal epithelial cells of porcine collected from days 11-12 of the estrous cycle and pregnancy. Similarly, in another experiment in equine endometrium, Szóstek et al. (2014) reported that TNF_α stimulated PGE₂ production to a greater extent and PGF_2α secretion to a lower level through up-regulation of PG synthases mRNA transcription. 

The overall result indicates that TNF_α may serve as an important modulator of prostaglandins production in the developing stage of buffalo corpus luteum; however, the possible physiological role of TNF_α for simultaneous increase of intra-luteal PGF_2α and PGE₂ production seems counterintuitive and needs further detailed studies to clarify its precise role.

We thank Director, NIANP for providing the necessary facilities for conducting the research work. Thanks are due to Director, CIRB for granting permission to conduct research work at NIANP Bangalore, India.