Salivary cell-free HSD17B1 and HSPA1A transcripts as potential biomarkers for estrus identification in buffaloes (Bubalus bubalis)

Abstract Estrus detection is a major problem in buffaloes because of the poor expression of estrus signs leading to low reproductive efficiency. Salivary transcripts analysis is a promising tool to identify biomarkers; therefore, the present study was carried out to evaluate their potential as estrus biomarkers. The levels of HSD17B1, INHBA, HSPA1A, TES transcripts were compared in saliva during estrous cycle stages [early proestrus (day −2, EP), late proestrus (day-1, LP), estrus (E), metestrus (ME) and diestrus (DE)] of cyclic heifers (n = 8) and pluriparous (n = 8) buffaloes by employing quantitative real-time polymerase chain reaction (qRT-PCR). The levels of HSD17B1 (EP/DE 1.46–2.43 fold, LP/DE 2.49–3.06 fold; E/DE 7.21–11.9-fold p < 0.01; ME/D 1.0–1.16 fold) and HSPA1A (EP/DE 0.93–2.39 fold, LP/DE 2.68–3.23 fold; E/DE 8.52–15.18 fold p < 0.01; ME/D 0.86–1.01 fold) were significantly altered during the estrus than other estrous cycle stages in both cyclic heifers and pluriparous buffaloes. Receiver operating characteristic curve analysis revealed the ability of salivary HSD17B1 (AUC 0.96; p < 0.001) and HSPA1A (AUC 0.99; p < 0.01) to differentiate E from other stages of the estrous cycle. Significantly higher levels of HSD17B1 and HSPA1A transcripts in saliva during the estrus phase suggest their biomarkers potential for estrus detection in buffaloes.


Introduction
Identifying the estrus phase accurately and efficiently is crucial to successful artificial insemination (AI) and conception in farm animals.Nevertheless, the buffalo is considered a shy breeder and the expression of estrus signs is not very prominent, especially during the summer months which leads to a higher incidence of silent estrus (29%) 1 and beget estrus detection difficult. 2In addition, the varied detection tools used for estrus detection in cattle aren't very effective in buffaloes since their sensitivity and specificity vary widely. 3Therefore, finding an easy, reliable, and accurate method for estrus detection in buffaloes is of utmost importance.
Saliva is preferred to other biological fluids for the discovery of biomarkers as it can be collected in a noninvasive, stress-free and un-stimulated condition. 4,5Saliva contains specific biomolecules like mRNA, miRNA, DNA, or protein that reflect the individual's physiology or disease condition. 6For instance, a lower level of polypeptide 1 (CYP27A1) and sialic acid acetyltransferase (SIAE) transcript in the saliva was found to be associated with oral squamous cell carcinoma (OSCC) in humans (Area Under Curve (AUC) 0.84). 7Similarly in humans, pancreatic disease has been shown to increase the levels of specific transcripts (MBD3L2, KRAS, STIM2, ACRV1, DMD and CABLES1) in saliva. 8In contrast to humans, salivary transcriptome analysis in farm animals has not been explored.For the first time, our group has reported direct salivary transcript analysis (DSTA) as a novel noninvasive method for estrus detection in buffaloes. 9The study found that salivary HSP70 and TLR4 levels were higher during estrus than at diestrus during the buffalo estrous cycle.Additionally, DSTA against target miRNAs identified miR-16, miR-191, and miR-223 as indicators of dominant ovarian follicles in buffaloes. 10In a recent study, Hebbar et al. 11 identified urinary cell-free miR-99a-5p as a potential candidate for estrus detection in buffaloes.Their study demonstrated a significantly lower level of miR-99a-5p during estrus and receiveroperating characteristic (ROC) analysis revealed its ability (AUC of 0.64, p < 0.08) to differentiate estrus from the diestrus stage of the estrous cycle.Transcriptomic analysis of goat ovarian tissues showed altered mRNA expression between estrus and diestrus. 12Since the ovary has a rich blood supply, it seems rational to hypothesize that mRNA may be excreted differentially into the peripheral circulation depending on the stage of the estrous cycle and some of these systemic transcripts may be in saliva as a transudate, which can mirror ovarian transcription and act as estrus biomarkers.Therefore, the present study examined salivary transcript levels of estrus indicator proteins during the buffalo estrous cycle.In this study, we selected 4 transcripts viz., 17beta-hydroxysteroid dehydrogenase type 1 (HSD17B1), inhibin Subunit Beta A (INHBA), heat shock 70 KDa Protein 1 A (HSPA1A), and testin (TES) based on their function associated with estrous physiology as per the available literature 5 and our earlier findings. 9,13We used quantitative real-time polymerase chain reaction (qRT-PCR) to detect levels of selected transcripts in buffalo saliva and compared those levels within different stages of the estrous cycle.ROC was performed to detect transcripts' ability to discriminate estrus from other non-estrus stages.Finally, in silico analysis was performed to identify the involvement of transcripts in various signaling pathways.

Selection of animals
The study was conducted in buffaloes (Bubalus bubalis) maintained at Livestock Research Center, ICAR-National Dairy Research Institute, Karnal, Haryana.For the study, heifers (n ¼ 8) and pluriparous (n ¼ 8) buffaloes of 2nd-5th parity maintained under isomanagerial conditions were considered for the experimentation.The animals' nutrient requirement was mainly fulfilled with ad-libitum green fodder and measured amount of concentrate as per the National Research Council (NRC) requirement.The Murrah buffaloes were examined for the presence of dominant follicle (DF) or corpus luteum (CL) on the ovarian surface using per-rectal examination and trans-rectal ultrasonography (USG) (Aloka, MODEL UST-5820-5).Buffaloes with cyclic CL were administered with PGF 2 a for estrus induction (Vetmate; 500 mg I/M).Samples were collected from the next spontaneous estrous cycle.The study was approved by the Institute Animal Ethics Committee (42-IAEC-18-9).

Detection of estrus in buffaloes
The onset of estrus in buffaloes was detected by physical observation of estrus signs, and confirmed by reproductive tract examination (uterine horn tonicity, cervical relaxation, tumefaction of vulva, hyperemia or reddening of vulvar mucous membrane), biochemical parameters (cervicovaginal mucus crystallization, fluidity, spinnbarkeit value) and progesterone hormone estimation in blood serum.Buffalo bull parading was carried out twice a day, morning (6-7 AM), evening (5-6 PM) and each time at least for 30 min for the exhibition of estrus signs by female buffaloes.Estrus signs observed were: standing to be mounted, sniffing/ licking the vulva, chin resting, flehmen's reaction, and mounting on or by other buffaloes, restlessness, bellowing, and frequent micturition.Further, trans-rectal USG was conducted to confirm the presence of preovulatory follicle (POF) on the day of estrus, the onset of ovulation after the end of estrus, and the presence of CL during the diestrus stage.Samples were collected from buffaloes that have shown standing estrus along with other signs of estrus.

Collection and processing of samples
Saliva samples were collected in the morning before feeding, on the day of estrus (E, day 0), during proestrus fearly proestrus (EP, day À2), late proestrus (LP, day À1)g, metestrus (ME, day þ3) and diestrus (day DE þ10) stage of the estrous cycle.Saliva ($5 mL) was collected from all selected buffaloes using a 20 ml syringe without the needle in 15 mL centrifuge tubes in unstimulated conditions by aspiration directly from the lower jaw of buffaloes.Samples were transported to the laboratory on ice and centrifuged at 3000 g for 10 min at 4 C. Supernatant or cell-free saliva was transferred to another microcentrifuge tube and an equal amount of TRIzol reagent (Sigma Aldrich, St. Louis, USA) was added and further processed for RNA isolation and gene expression study.

Collection and analysis of cervico-vaginal mucus parameters
The cervico-vaginal mucus (CVM) was collected from buffaloes during the peri-estrus period (proestrus, estrus, and metestrus) by aspiration method using a sterile blue sheath (IMV Technologies, France) fitted with a universal AI gun through the recto-vaginal method.Immediately after collection, samples were transported to the laboratory on ice.Initial parameters of CVM such as quantity, consistency, and crystallization/fern pattern were recorded.For crystallization/ fern pattern, a small quantity of mucus was smeared onto a glass slide and the smear was allowed to dry at room temperature.Then the slide was observed under the microscope (Nikon Eclipse Ti, Japan) to confirm the estrus stage.

Blood collection and its processing for estimation of progesterone hormone
Blood samples were collected during different estrous cycle stages using 9 mL serum collection tubes (BD Vacutainer).Blood samples were kept at room temperature for 1 hr in a slanting position and centrifuged at 3000 g for 10 min.Serum was separated and stored in cryovials at À80 C for further progesterone estimation using an ELISA kit (Cayman Chemical, USA) according to the manufacturer's protocol.The progesterone concentration was measured in Nano Quant Infinite M200 PRO (TECAN, Seestrasse 103, and Switzerland) at a wavelength of 410 nm.

Total RNA extraction
Total RNA was extracted using TRIzol reagent as per the manufacturer's instructions with slight modifications.In brief, TRIzol reagent was added to each microcentrifuge tube containing saliva and vortexed for 15 sec, followed by the addition of chloroform (200 lL) and vortexed for 15 sec for mixing of contents properly.Samples were kept on ice for 10 min and centrifuged at 11,363 g for 10 min at 4 C; the uppermost aqueous phase (containing RNA) was collected in a new 2 mL microcentrifuge tube without disturbing the interface.An equal quantity of isopropanol was added to each tube and kept at À80 C overnight.The total content of the tube was loaded to the respective RNeasy mini spin columns (RNeasy mini kit, Qiagen) placed in a 1.5 mL collection tube, and the lid was closed properly and centrifuged for 1.5 min at 11,363 g.The flow-through was discarded and 600 lL of chilled ethanol (75%) was added to each column and centrifuged for 1.5 min at 11,363 g twice.The flow-through was discarded.The column was centrifuged again at 11,363 g for 2 min to remove the traces of ethanol.The RNeasy Mini spin column was placed in a new 1.5 mL collection tube and elution of RNA was done by placing 20 lL of RNase free water on the spin column membrane and kept for 5 minutes on ice followed by centrifugation for 2.0 min at 11,363 g.The step was repeated twice to get the maximum yield of RNA.Concentration and purity (260/280) of total RNA extracted was measured in Nano Quant Infinite M200 PRO (TECAN, Seestrasse 103, and Switzerland).The samples with a 260/280 ratio between 1.8 and 2.0 were used for cDNA synthesis and downstream processing.

cDNA synthesis and optimization of PCR
Total RNA (250 ng) was reverse transcribed into cDNA in a 20 lL reaction mixture using RevertAid First-strand cDNA synthesis kit (Thermo Scientific, USA) as per the manufacturer's protocol.A reverse transcription reaction was carried out at 65 C for 5 min, 42 C for 60 min, and 70 C for 5 min in a thermal cycler (Bio-Rad, USA).In brief, 12 mL of reaction mixture containing 250 ng of the total RNA, 0.5 mL of oligo-dT primer, 0.5 mL of random primer, nucleasefree water to adjust the volume, was incubated at 65 C for 5 min.On completion of incubation following components (f5X Reaction Buffer, 4 mL; Ribolock RNase Inhibitor, 1 mL (20 U/ll); 10 mM dNTP Mix, 2 mL; RevertAid M-MuLV RT, 1 mL (200 U/ll)g, nuclease-free water to make the volume up to 20 mL) were added to the above reaction mixture and contents were appropriately mixed, incubated at 42 C for 60 min in a thermal cycler.The reaction was finally terminated by heating at 70 C for 5 min.The reverse transcription reaction product (cDNA) was stored at À20 C until further use.The sequences of the selected genes were retrieved from the NCBI database and primers were designed using Gene tool software (Table 1).Cyclic conditions of PCR were optimized for each target gene and a reaction mixture of 10 mL was prepared containing 1 lL of cDNA, 5.0 lL Dream Taq PCR Master Mix (Thermo Scientific, USA), and 0.5 lM of each forward and reverse primer and nuclease-free water to adjust the volume and PCR reaction was set for initial denaturation at 95 C for 3 min followed by 35 cycles at 95 C for 30 sec, annealing temperature for the gene for 30 sec, 72 C for 30 sec and final extension at 72 C for 1.2 min in a Gradient Thermal Cycler (C1000 Touch Thermal Cycle, Biorad).After optimization, the presence of a single amplicon was confirmed by agarose gel (1.5%) electrophoresis.

Real-time PCR
Salivary levels of mRNAs were determined by qRT-PCR in a CFX96 Touch Real-Time PCR Detection System (Biorad, USA) using PowerUp TM SYBR TM Green Master Mix.In brief, 10 mL PCR reaction mixture containing 1 mL of cDNA, 0.5 mM of each of forward and reverse primer, 5.0 mL of PowerUp TM SYBR TM Green master mix, and the rest of the volume was adjusted with nuclease-free water.All the samples were run in duplicate along with nontemplate control.The real-time PCR was performed under thermocycling conditions: 95 C for 10 min, 40 cycles of 95 C for 30 sec, annealing temperature (Table 1) for 10 sec, and final extension at 72 C for 15 min.The amplification plot and melting curves for all the reactions were analyzed for the specific amplification in all samples.Expression of GAPDH was taken as endogenous control.The relative quantification of the target gene was done using the 2 ÀDDCT method. 14he expression of genes in different samples was estimated as: Relative gene expression, i.e., Fold change Where DCt ¼ Ct of target gene -Ct of reference gene (GAPDH)

Bioinformatics analysis and statistical analysis
Search Tool for the Retrieval of Interacting Genes (STRING) (https://www.string-db.org/),an online tool, was used to understand the involvement of selected genes in different signaling pathways.Statistical analysis was performed by one-way analysis of variance (ANOVA) and post hoc Tukey test or paired t-test for group-wise comparison using SPSS software (version 16).Results are represented as mean ± SEM.
Receiver operating characteristics (ROC) curves were performed using DCt values of the HSD17B1 and HSPA1A in the EP, LP, E, ME, and DE stage samples in order to determine their diagnostic biomarker potential.ROC curves plotted sensitivity against 1-specificity over the whole range of cutoff values.When the AUC was near 1 and statistical significance (p < 0.01) was observed, the transcript's ability to differentiate estrus from other non-estrus stages was considered.The optimal cutoff value of transcripts was established with the highest sensitivity (Se) and specificity (Sp) for differentiating the estrus from other non-estrus stages.

Identification and confirmation of estrus
Estrus was determined by physical signs of estrus, confirmed by reproductive tract examination (uterine horn tonicity, tumefaction of vulva, hyperemia of vulval mucous membrane, cervical relaxation), biochemical parameters (CVM characteristics), and progesterone hormone in blood serum.Behavioral signs of estrus are shown in Table 2.Among behavioral signs, standing to be mounted, flehmen's reaction, and licking/sniffing the vulva by teaser bull were mostly observed in female buffaloes during estrus.Other estrus signs were categorized as mild, moderate and intense based on their intensity of expression in individual animals, as shown in Table 3. Tonicity of the uterus was found to be most intense (89.5%) during estrus and in a few buffaloes, moderate tonocity (10.5%) was also recorded.Similarly, CVM discharge and hyperemia/reddening of the vulval mucous membrane were most intense during estrus and in only 10.5% buffaloes, it was moderate.CVM crystallization/fern pattern was found to be typical with score 3 and 4 in 79.3% and atypical with score 2 in 21.1% buffaloes (Fig. S1).Transrectal-USG revealed the presence of dominant follicle (DF) during the follicular phase, and the size of DF was 11.5 mm, 12.5 mm, and 13.3 mm during EP, LP, and E, respectively (Fig. 1).The absence of preovulatory follicle confirmed ovulation during the ME and CL presence during the DE.Serum progesterone levels was estimated and it was found lowest at E (0.32 ± 0.06 ng/mL, p < 0.05), LP (0.51 ± 0.12 ng/mL, p < 0.05), EP (0.93 ± 0.03 ng/mL, p < 0.05) as compared to ME (1.58 ± 0.24 ng/mL) and DE (2.84 ± 0.26 ng/mL) stages (Table 4).
Target mRNA levels in cell-free saliva We selected 4 candidate genes based on our earlier findings and available literature (     2C).However, in pluriparous buffaloes, INHBA levels increased at E (4.94-fold) as compared to ME (0.9fold) but not from EP (2.79-fold) and LP (4.28-fold) stages (Fig. 3C).The salivary level of the TES gene revealed its higher level at E (3.65-fold, p < 0.05) as compared to EP (1.13-fold) and DE stage but not from LP (2.52-fold) and ME (1.44-fold) stage of cyclic buffalo heifers (Fig. 2D).Similarly, its level did not significantly change at E (2.8-fold) as compared to EP (1.7-fold), LP (2.57-fold) and ME stage (1.3-fold) of pluriparous buffaloes (Fig. 3D).

Receiver operating characteristics (ROC) curve analysis to predict the diagnostic ability of salivary transcripts
ROC curve analysis was performed to determine the discriminatory ability of HSPA1A and HSD17B1 for differentiating E from other stages of the estrous cycle.ROC curve analysis showed that at the 0.26 cutoff value, the sensitivity and specificity of HSPA1A in differentiating the E stage from EP, LP, ME, and DE stages were 100 and 95.65%, respectively; the AUC was 0.99 at a 95% confidence interval (p < 0.0001) as shown in Fig. 4. Similarly, the cutoff value for HSD17B1 was 0.095 and the sensitivity and specificity in differentiating the E from EP, LP, ME, and DE stages were 90 and 94.74%, respectively; the AUC was 0.96 at 95% confidence interval (p < 0.0001) (Fig. 5).In silico analysis for identification of signaling pathways In silico analysis was performed for HSD17B1 and HSPA1A genes to identify their involvement in different signaling pathways using the online STRING tool.It revealed a significant association of HSD17B1 with other genes such as CYP19A1, CYP1A1, CYP2A13, CYP1B1, HSD17B2, HSD17B3, HSD17B7, HSD17B and involved in steroid hormone biosynthesis/ovarian steroidogenesis and aromatase signaling pathways (P ¼ 8.10E-15), Fig. S2.Similarly, HSPA1A in association with other genes (BAG1-3, BAG5, DNAJA1-2, DNAJB1, DNAJB1, DNAJC1), regulates the cellular response to the stress pathway (P ¼ 7.42e-06), Fig. S3.

Discussion
Accurate and efficient estrus detection is important for successful conception and efficient reproduction management in farm animals.Because buffalo don't show obvious estrus signs, identifying estrus biomarkers for this species is of paramount importance.In recent years, several studies performed on saliva reported biomarker potential of circulating cell-free RNAs including mRNA and miRNA for the diagnosis of several diseases 4,8,15 and physiological alterations. 9,10In our previous study, we have identified 62 proteins specific to the estrus stage using a global proteome analysis of buffalo saliva and among them, HSPA1A, HSD17B1, INBHA, and TES proteins are found to be most predominant during estrus. 13Using direct salivary transcript analysis, we also discovered that HSP70 is more abundant in buffalo saliva during the estrus than diestrus stage. 9Therefore, the present study examined the estrus biomarker potential of the HSD17B1, HSPA1A, INHBA, and TES genes in the saliva of buffaloes.Among the 4 transcripts evaluated in the present study, we found elevated levels of HSD17B1 and HSPA1A during estrus compared to other estrous cycle stages in both cyclic heifers and pluriparous buffaloes.The highest level of HSD17B1 in saliva coincided with the presence of a large dominant follicle and its level decreased after ovulation and during the luteal phase, suggesting its crucial role in granulosa cells function, ovarian steroidogenesis, and functional status of the dominant follicle.Shashikumar et al. 13 also identified HSD17B1 protein in the saliva of buffaloes, especially during the estrus stage and not at any other stages of the estrous cycle.HSD17B1 is a short-chain alcohol dehydrogenase reductase family member that catalyzes the conversion of low active 17-ketosteroids to the highly active 17b hydroxysteroids and vice versa. 16Although several HSD17B enzymes play a role in steroid metabolism, the HSD17B1 enzyme preferably catalyzes the estrone to estradiol 17 and plays a central role in granulosa cell estradiol synthesis and ovarian function. 16Other studies demonstrated down-regulation of this enzyme in luteinizing granulosa cells and not present in CL, further supporting our findings on the lower level of HSD17B1 during the metestrus and diestrus stage. 18,19imilarly, the transcriptome profile of granulosa cells from bovine ovarian follicles showed down regulation of HSD17B1 gene in small atretic follicles than in healthy follicles 20 demonstrating its important role in follicular development.All these studies demonstrated cyclic variation in the expression of HSD17B1 in granulosa cells.Since HSD17B1 is one of the major enzymes regulating ovarian estradiol production, there is a possibility of excretion of this enzyme into circulation and then into the saliva.Another possibility is that higher estradiol levels during late proestrus and estrus may induce the expression of HSD17B1 in salivary glands, leading to its increased level in saliva, which needs to be explored.Interestingly, we observed elevated levels of HSPA1A toward late proestrus and a very high level during estrus than in other stages of the estrous cycle.Muthukumar et al. 5 also identified HSP70 protein in the CVM and saliva of buffaloes, particularly during the estrus stage and not at the diestrus stage of the estrous cycle.An earlier study by Onteru et al. 9 using direct saliva transcript analysis also demonstrated a higher level of HSP70 transcript in saliva during estrus than diestrus stage, suggesting it as a good indicator of estrus in buffaloes.Previously, we also identified HSP70 protein in the saliva of buffaloes during the estrus stage and not at any other stages, suggesting stage-specific regulation of HSPA1A during the estrous cycle. 13HSPA1A is a molecular chaperone and it protects cells from stress by promoting the folding of nascent polypeptides and correcting the misfolding of denatured proteins. 21A previous study also reported its involvement in the estrogen signaling pathway.HSP70 regulates the responsiveness of ERa to its ligands (17b-estradiol) and their binding to estrogenresponse elements (ERE) and induces transcription of target genes. 22Estradiol hormone induces the estrus signs in animals and estradiol level increased to a higher level just before the onset of estrus (22.4-35 pg/ mL) 23,24 and higher expression of HSPA1A during estrus may be a protective mechanism to protect cells from estradiol-induced stress.Previously it was reported that estradiol treatment-induced heat shock proteins (HSPs) expressions in rat brain vasculature 25,26 demonstrate HSPs induction as an important protective mechanism for estrogen induced stress or injury.Suggestive increased expression of HSPA1A level during estrus could be a protection from estrogen-induced stress to granulosa cells and its possible role in follicular growth and steroidogenesis process.Hatzirodos et al. 20 also reported down-regulation of HSPA1A expression in small atretic follicles than healthy follicles, demonstrating its essential role in follicular development.
In addition, we also observed high levels of INHBA during estrus than in the ME and DE stages but not from EP and LP stages.In contrast, Shashikumar et al. 13 reported INHBA protein in buffalo saliva during estrus and not at any other stages.In concordance with our observation, an increased level of plasma inhibin A concentration was reported during the follicular phase i.e., from $50 pg/mL before luteolysis to a peak level of $125 pg/mL during preovulatory estradiol/LH surge and then decreased sharply to a concentration of 55 pg/mL after commencement of ovulation. 27Inhibin belongs to the transforming growth factor (TGFb) superfamily.Several tissues and reproductive organs such as ovarian granulosa cells, fetus, placenta, express INHBA.It is mainly synthesized and secreted by the follicular granulosa cells and regulates proliferation, apoptosis of granulosa cells, steroidogenesis, and follicular development during the estrous cycle. 28It also plays a vital role in maintaining a species-specific number of ovulations.Hatzirodos et al. 20 also reported down-regulation of INHBA expression in small atretic follicles than healthy follicles, demonstrating its important role in follicular development.Our findings are the first report on salivary INHBA transcript level in general and with the buffalo estrous cycle.A higher level of INHBA during proestrus and estrus can be correlated with its involvement in the selection and development of the dominant follicle.
Further, we also observed an increased level of TES gene on the estrus day compared to DE stage but not from EP, LP and ME stages.In contrast, Shashikumar et al. 13 reported TES protein in buffaloes' saliva exclusively during the estrus stage and not at any other estrous cycle stages.Immuno-histochemical study in the rat has identified localization of TES protein in the granulosa cells and its expression increased during follicular development but is down-regulated drastically when the follicle is undergoing atresia. 29Testin is a 47 kDa protein composed of 421 amino acids, and it has 3-C terminal LIM domain.It interacts with the Vangl2 gene during female reproductive tract development in mice. 30Suggestive increased expression of the TES gene in saliva during the estrus stage can be correlated with its involvement in folliculogenesis and follicular development.
To further understand the functions of HSPA1A and HSD17B1, we performed in silico analysis to predict possible roles in different signaling pathways using the STRING online tool (v.11.5).HSD17B1 plays an important role in ovarian function, involving steroid hormone biosynthesis/ovarian steroidogenesis and aromatase signaling pathways.Similarly, HSPA1A regulates the cellular response to the stress pathway and maybe a protective mechanism to protect granulosa cells from estradiol-induced stress during estrus.In addition, ROC curve analysis also strongly indicated the ability of HSPA1A and HSD17B1 to differentiate estrus from other stages of the estrous cycle.Finally, the dynamic presence of the above two transcripts in saliva during the estrous cycle indicates their biomarker potential for estrus detection in buffaloes.However, this data needs to be validated in a large cohort of buffaloes, including different breeds and seasons, to further qualify them as estrus biomarkers.

Conclusion
Using cell-free salivary transcript analysis, we identified significantly higher levels of HSD17B1 and HSPA1A transcripts in buffalo saliva during the estrus stage and their potential as biomarkers for distinguishing estrus from other non-estrus stages in both cyclic heifers and pluriparous buffaloes.These findings may be helpful to develop a salivary transcript-based-estrus detection tool in the near future.

Table 2 .
Expression of estrus signs in buffaloes.

Table 1 .
Details of primers used in the study.

Table 1
) for qRT-PCR analysis.Salivary levels of HSD17B1 and was found to be significantly (p < 0.01) higher at E (15.18-fold) followed by LP (3.23-fold), EP (0.93fold), and ME (0.86-fold) as compared to the DE stage (Fig.2B).A similar trend was also observed in pluriparous buffaloes wherein the highest level (p < 0.05)

Table 3 .
Intensity of expression of other estrus signs in buffaloes.