Early pregnancy detection is critical to efficient livestock management, particularly in buffaloes, a significant source of dairy products in the Indian subcontinent. Accurate and timely identification of pregnancy status can optimize reproductive management strategies, improve herd productivity and enhance economic returns (Barile et al., Reference Barile, Menchetti, Casano, Brecchia, de Sousa, Zelli, Beckers and Barbato2021). Many factors contribute to declining reproductive fitness, including early embryonic mortality (Santos et al., Reference Santos, Thatcher, Chebel, Cerri and Galvão2004) during insemination to 42 days of gestation and failure to mate animals during the reproductive period (Walsh et al., Reference Walsh, Williams and Evans2011). Traditional methods for pregnancy diagnosis, such as rectal palpation (Karen et al., Reference Karen, Darwish, Ramoun, Tawfeek, Van Hanh, de Sousa, Sulon, Szenci and Beckers2007) and ultrasonography, often require considerable expertise and are limited in sensitivity during the early stages of gestation (Karen et al., Reference Karen, Darwish, Ramoun, Tawfeek, Van Hanh, de Sousa, Sulon, Szenci and Beckers2011). Other newer techniques, like the urinary metabolite-based method, have low specificity and sensitivity. The area under the curve (AUC), a statistical measure to assess the overall performance of a diagnostic test of six urinary metabolites detecting early pregnancy in Bos frontalis (Mithun), was 0.50–0.85 (Sangwan et al., Reference Sangwan, Vikram, Hooda, Choudhary, Jawla, Somagond, Balhara, Phulia, Khan, Girish, Datta, Mitra and Balhara2024).
In recent years, molecular techniques have emerged as promising alternatives for early pregnancy detection in various species. Among these, quantitative polymerase chain reaction (qPCR) has gained prominence in identifying early pregnant buffaloes using ISGs (Casano et al., Reference Casano, Barile, Menchetti, Guelfi, Brecchia, Agradi, De Matteis, Scatà, Grandoni and Barbato2022; Nag et al., Reference Nag, Arunmozhi, Sarath, Asokan and Vijayarani2018; Sharma et al., Reference Sharma, Choudhary, Ratta and Singh2023). By targeting ISG15 and RSAD2 gene expression, the accuracy of pregnancy status of cows (Cheng et al., Reference Cheng, Xiang, Hu, Yu, Xia, Tao and Zhao2019) and buffaloes (Singh and Choudhary, Reference Singh and Choudhary2024) can be increased.
Interferons are the cytokines that activate various signal transduction pathways leading to induction of hundreds of genes called interferon-stimulating genes (ISGs). As per the estimate, 10% of genes in the genome can be regulated by interferons. In addition to antiviral effects of ISGs, they may have other functions, such as ISG 15 (ISG15) and galectin-3 binding protein (LGALS3BP) are two genes that have been implicated in early pregnancy in different species. ISG15 is an interferon-induced protein with antiviral and immunomodulatory functions, while LGALS3BP is involved in various biological processes, including cell adhesion and immune response (Liu and Stowell, Reference Liu and Stowell2023). Previous studies have reported the upregulation of these genes in the maternal circulation during early pregnancy. ISGs were shown to be the best biomarkers in blood for predicting pregnancy and embryonic mortality during the peri-implantation period in buffaloes (Casano et al., Reference Casano, Menchetti, Trabalza-Marinucci, Riva, De Matteis, Brecchia, Inglesi, Rossi, Signorelli, Barile and Barbato2023). Likewise, the abundance of the transcript of LGALS3BP has been suggested as a potential novel biomarker of early pregnancy in cows (Ferraz et al., Reference Ferraz, Filho, Rocha, Neto, de Andrade Bruni, Oshiro, Baruselli, Lima and Pugliesi2021; Rocha et al., Reference Rocha, da Silva Andrade, de Melo, Motta, Coutinho, Gonella-Diaza, Binelli and Pugliesi2020), suggesting their potential as biomarkers for pregnancy detection.
Developing a duplex qPCR assay targeting ISG15 and LGALS3BP could offer a comprehensive and efficient approach for early pregnancy diagnosis in buffaloes. Previous findings indicated a significant upregulation of ISG15 and LGALS3BP transcripts at 20 days post AI in buffaloes (Sharma et al., Reference Sharma, Choudhary, Ratta and Singh2023). To enhance target gene amplification's specificity, we propose employing the TaqMan probe duplex assay for subsequent analyses, along with comprehensive validation on randomly collected blood samples from buffalo farms.
This study aims to develop and validate a TaqMan duplex qPCR assay for detecting ISG15 and LGALS3BP mRNA in buffalo blood samples and validate the results on randomly collected samples which had been inseminated before 20 ± 2 days. We evaluated the assay's diagnostic performance for early pregnant buffaloes using a receiver operating characteristic (ROC) curve and developed a formula for early pregnancy diagnosis based on ISG15 and LGALS3BP gene expression.
Material and methods
Animals and experimental design
Three animal studies (studies 1–3) were performed. In study 1, a duplex probe-based qPCR assay was developed, and in studies 2 and 3, validation of the assay was performed.
Study 1
The buffalo blood samples were collected from the Directorate of Livestock Farms, Guru Angad Dev Veterinary and Animal Sciences University, Ludhiana, Punjab, India. The experimental approaches were evaluated and permitted by the Institutional Animal Ethical Committee (GADVASU/2021/IAEC/61/05 dated 13 October 2021 and IAEC/2024/02-22 dated 10 January 2024),). In study 1, 18 animals belonging to the Murrah buffalo were selected as described earlier (Sharma et al., Reference Sharma, Choudhary, Ratta and Singh2023) to develop a duplex TaqMan probe-based assay. The animals were categorized into pregnant (n = 9) and non-pregnant (n = 9), as confirmed by ultrasonography at 40 days.
Study 2
To confirm the suitability of ISG15 and LGALS3BP as markers for early pregnancy diagnosis, blood samples of 39 randomly inseminated buffaloes at 20 ± 2 days were collected (during December 2023 to January 2024). Ultrasonography at day 40 confirmed non-pregnant (n = 23) buffaloes and pregnant (n = 17) buffaloes.
Study 3
We further validated the fluorescence probe-based qPCR duplex assay on 41 more buffalo blood samples collected (during March–April 2024) at 20 ± 2 days and analysed its diagnostic effects in detecting pregnancy.
Ultrasonography for pregnancy confirmation
The stored images were analysed using ImageJ (National Institutes of Health, Bethesda, MD, USA), an image analysis software. Each corpus luteum (CL) was delineated as a region of interest (ROI) within the captured grayscale images for pixel-based analysis, and the cross-sectional area of the CL was measured. The same ROI was utilized to quantify the pixel colour intensity in both pregnant and non-pregnant buffaloes, serving as a semi-quantitative metric for buffalo pregnancy (BP). The pixel data representing the CL and BP were then exported to calculate the percentage of the area occupied by colour pixels (Coloured area * 100/CL area).
Pregnancy diagnosis using ultrasonography
Buffaloes were assigned for pregnancy diagnosis by transrectal ultrasonography examination at 20, 25 and 40 days post-AI. During the examination, buffaloes were restrained in stanchions and the surrounding light was kept dim. Prior to the protocol, faeces were removed from the rectum by the operator, followed by the insertion of a transducer and scanning of uterine horns and ovaries. The sleeved arm of the operator, along with the transducer, was lubricated with the obstetrical lube to promote easy penetration along the anal sphincter and to provide convenient access to the base of the rectal mucosa. The transducer was placed dorsally and gently moved cranially towards the genital tract. The cervix, uterine horns and uterine structures were carefully examined. A duplex ultrasound instrument (B-mode and colour Doppler) (Sonosite M turbo; Fujifilm, HR, India) with a 5–7.5 MHz linear multi-frequency transducer was used. Images of colour Doppler were captured on days 20 and 25 to assess the luteal blood perfusions and functional CL. At 40 days, pregnancy was confirmed after observing allantochorion and a viable embryo with a heartbeat in the uterine lumen. The animals without functional CL and viable fetal mass were confirmed as non-pregnant.
Blood collection and PBMC isolation
Blood samples of 0, 20, 25 and 40 days post-inseminations were acquired from a buffalo farm in 10-ml EDTA-containing Falcon tubes (Tarsons, WB, India) immediately following collection from the external jugular vein. PBMCs were obtained following density gradient centrifugation using HiSep™LSM 1077 (1.077 g/ml; Himedia, CA, USA). In brief, 5 ml of whole blood was diluted in 1:1 with isotonic phosphate buffer saline 1× (PBS: Sigma-Aldrich, CA, USA) in 15 ml Falcon tube (Tarsons, WB, India) and gently layered over 3.5 ml of HiSep™LSM 1077. The tubes were centrifuged at 400 x g for 30 min. The phase having red blood corpuscles (RBCs) and polymorphonuclear neutrophils (PMNs) will settle at the bottom of the tube. The phase having peripheral blood mononuclear cells (PBMCs) settle over the plasma: HiSep™LSM 1077 interface. The PBMCs were collected with care and washed thrice with PBS 1×. The PBMC pellet was lysed in 1 ml of RNAiso Plus (Takara, Shiga, Japan) and stored at −80°C for downstream purposes.
RNA isolation and cDNA synthesis
As published earlier, total RNA was extracted from PBMCs using RNAiso Plus (Takara, Shiga, Japan) (Sharma et al., Reference Sharma, Choudhary, Ratta and Singh2023). The integrity of RNA was determined by the A260/A280 and A260/A230 ratios, and quantity was evaluated using NanoDrop Spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA, USA). According to the manufacturer’s protocol, 1000 ng RNA was reverse transcribed with Verso cDNA Synthesis Kit (Thermo Fisher Scientific Inc., Waltham, MA, USA). The cDNA obtained was stored at −20°C and later used as a template for TaqMan-based RT-qPCR reactions.
TaqMan-based RT-qPCR
The primer and probe sets were designed in-house against the buffalo gene sequences available in the NCBI database using PrimerQuest™ Tool from IDT (Integrated DNA Technologies, Coralville, IA, USA) software and were custom-developed (SahaGene, Hyderabad, TL, India) (Table 1). The PCR primers of these genes validated in our earlier publication (Sharma et al., Reference Sharma, Choudhary, Ratta and Singh2023). Relative quantification of candidate transcripts, namely classical ISG15, novel LGALS3BP and reference transcript PPIA (Cyclophilin A), was performed following the MIQE guidelines (Bustin et al., Reference Bustin, Benes, Garson, Hellemans, Huggett, Kubista, Mueller, Nolan, Pfaffl, Shipley, Vandesompele and Wittwer2009). All transcripts were quantified using TaqMan™ Fast Advanced Master Mix (Applied Biosystems, Waltham, MA, USA) and reactions were performed in Rotor-Gene Q thermocycler (Qiagen, Hilden, Germany). In brief, each duplex reaction well was constituted of 4 µl of cDNA, 12.5 µl of 2× TaqMan™ Fast Advanced Master Mix, a final concentration of 4 pmol of both forward and reverse primer, 2 pmol of probe for ISG15 gene, 4 pmol of forward primer and 8 pmol of reverse primer, 2 pmol of probe for LGALS3BP gene, and 5.1 µl of nuclease-free water up to a final volume of 25 µl. The assay was performed at 50°C for 2 min, then 95°C for 2 min, followed by 40 cycles at 95°C for 3 s (denaturation), and 55°C for 20 s (annealing/extension), whereas each sample was run in duplicates. The specificity of PCR amplification was assessed using agarose gel electrophoresis. Ct (threshold cycle) values were exported in Excel file format and data were assessed using the 2∆∆Ct method (Livak and Schmittgen, Reference Livak and Schmittgen2001). GraphPad (ver.9) (GraphPad Software, CA, USA) is used for the representation of graphical data and analysing multiple regression models in study 3.
Genes, their primers, probe sequences and product size of RT-qPCR amplicon used in this study
Primer sequences published in sister manuscript (Sharma et al., Reference Sharma, Choudhary, Ratta and Singh2023).
Threshold value estimation
The mean fold change values of genes expressed at days 20 and 25, normalized to day 0, were utilized to evaluate pregnancy prediction accuracy and establish a predictive threshold. The findings were as follows: pregnant buffaloes exhibited mean fold change values exceeding the threshold, corroborated by positive ultrasonographic detection (true positive [TP]). Instances where buffaloes displayed mean fold change values above the threshold but had negative ultrasonographic outcomes were categorized as false positives (FPs). True non-pregnant buffaloes presented with mean fold change values below the threshold and negative ultrasonographic results (true negative [TN]). However, non-pregnant buffaloes with lower mean fold change values but were positive on ultrasonographic detection were classified as false negatives (FNs). The positive predictive value (PPV%; TP/[TP + FP] × 100) and the negative predictive value (NPV%; TN/[TN + FN] × 100) were calculated as described earlier (Han et al., Reference Han, Austin, Rempel and Hansen2006; Yoshino et al., Reference Yoshino, Toji, Sasaki, Koshi, Yamagishi, Takahashi, Ishiguro-Oonuma, Matsuda, Yamanouchi, Hashiyada, Imai, Izaike, Kizaki and Hashizume2018a). The fold change expression of ISG15 and LGALS3BP on days 20 and 25 was used for ROC curve evaluation.
Statistical analysis
Data were normalized using the Shapiro–Wilk test and outliers beyond ±3 standard deviations were neglected. The data were investigated in SAS using the mixed linear model. Animals were considered as subjects, and days as repeated factors in study 1. The effects of fold change expression of ISG15 and LGALS3BP transcripts were assessed at four different time points (four levels: 0, 20, 25 and 40 days post-AI), the status of pregnancy (two levels: pregnant and non-pregnant) and the interrelationship between pregnancy status and time. The effect of cow was considered as a random effect. The distinction between the two groups was evaluated using a paired t-test assuming equal variances. In study 2, the effects of fold change expression of ISG15 and LGALS3BP transcripts were assessed at one time point (20 ± 2 days) using a mixed linear model, like that of study 1. Predicted probability > 0.5 was taken for pregnant buffaloes and value < 0.5 was taken as non-pregnant buffaloes. In study 3, we explored the predictive power of fold change expression of ISG15 and LGALS3BP transcripts using a full linear regression model. We wanted to test the suitability of data without random effects and to assess the robustness of our results. The full linear model Y = β0 + β1*ISG15 + β2*LGALS3BP + β3*ISG15*LGALS3BP, where Y was the predicting value, β0 as the intercept, and β1, β2 and β3 as the estimates of ISG15, LGALS3BP and ISG15*LGALS3BP, respectively. Y value ≤ 0.0.65 indicated non-pregnant and > 0.065 indicated pregnant buffaloes. The sensitivity and specificity of the method were calculated using GraphPad through ROC curves.
Results
Morphological characteristics of blood flow in CL
We evaluate changes in blood flow in CL and assess functional regression of CL. Colour Doppler was helpful in predicting non-pregnancy in cattle by 18–23 days (Gnemmi et al., Reference Gnemmi, Maraboli, Gnemmi, Saleri and De Rensis2022). The blood flow area of CL was greater in the pregnant buffalo (p < 0.001) at 20– days post-AI compared to non-pregnant ones (Fig. 1).
Ultrasonographic images of pregnant buffaloes’ corpus luteum (CL). Representative colour Doppler sonograms of CL and blood perfusion (BP) before insemination (day 0) and 20 days after insemination in non-pregnant and pregnant buffaloes. (A) A characteristic fluid-filled cavity with no appreciable CL in non-pregnant buffaloes. Images of 20-day non-pregnant (B) and pregnant (C) buffaloes. The BP area of CL, expressed in the percentage area of BP vs. total area of CL, was significantly (p < 0.0001 indicated by ****) different in 20-day pregnant buffaloes compared to 20-day non-pregnant buffaloes.
Expression of ISG15 and LGALS3BP
The ISG15 and LGALS3BP gene expression was upregulated in PBMCs of pregnant buffaloes within 20 days of insemination compared to 0 day before insemination. In non-pregnant buffaloes, however, the expression of ISG15 and LGALS3BP remained constant and at the basal level. The expression level of ISG15 (fold change or as 2−ΔΔCt) was affected by the time (p = 0.01), and interaction between time and pregnancy status (p = 0.036) (Fig. 2A). Similarly, LGALS3BP expression was affected by the time (p = 0.008), and interaction between time and pregnancy status (p = 0.033) (Fig. 2B). The marginal means of ISG15 was higher in the pregnant group than in the non-pregnant group at 20 days post-AI (vs. 0 day) (fold change; 2.60 ± 0.25 vs. 1.17 ± 0.12; p = 0.03) (Fig. 2C). Likewise, the marginal mean of LGALS3BP expression in the pregnant group was significantly higher at 20 days (vs. 0 day) (fold change; 2.21 ± 0.13 vs. 1.07 ± 0.06; p = 0.011) (Fig. 2D). In the non-pregnant group, the difference in the expression of both genes did not differ (p > 0.05) at 20 and 40 days post-AI (vs. 0 day).
Differential gene expression of two markers of early pregnancy, ISG15 and LGALS3BP, in test data. A linear mixed model analysis was conducted to examine the effects of ISG15 (A) and LGALS3BP (B) expression, and time of insemination on pregnancy status. The results revealed a significant interaction between time of insemination and gene expression (p < 0.05). In early pregnant buffaloes, both ISG15 (p = 0.03) and LGALS3BP (p = 0.011) exhibited significantly higher expression levels at 20 days post-insemination compared to 0 day. No significant (ns) changes were observed in non-pregnant groups (p > 0.05).
Assay validation
Average expression of ISG15 and LGALS3BP genes in 39 buffalo samples was validated in the study 2 dataset. Retrospectively, of 39 buffaloes, 20 were pregnant and 19 were non-pregnant buffaloes determined by ultrasonography at 35 days post-AI and then by rectal palpation. The ROC curves were established for the expression of ISG15 and LGALS3BP. Duplex gene amplification plot of ISG15 (green channel, solid lines) and LGALS3B gene (yellow channel, lines with circles) showed similar Ct value (Fig. 3A). This suggests that the interaction between ISG15 and LGALS3BP significantly contributes to the model's predictive power beyond the individual effects of genes. There was a very high positive correlation (r = 0.78; p < 0.0001) in the expression of ISG15 and LGALS3BP in pregnant buffaloes (Fig. 3B). The effectiveness of pregnancy was represented by the area under the ROC curve (AUC). The AUC for the model with ISG15 expression was 0.80. The AUC for the model with LGALS3BP was 0.60. The interaction model with the interaction of expression of two genes alone (ISG15*LGALS3BP) demonstrated an AUC of 0.86. When the whole model was considered taking expression of both the genes with the interaction, the AUC was 0.957 (Fig. 3C). Furthermore, 17/40 buffaloes were found TP, and 3 were found FP (Fig. 3D) in study 2. The negative predictive power (NPV) and positive predictive power (PPV) of the assay were 95% and 85%, respectively (Table 2).
Validation of duplex qPCR analysis of ISG15 and LGALS3BP for early pregnancy detection in buffaloes. (A) Simultaneous amplification curves for ISG15 (green) and LGALS3BP (yellow) genes. The similar Ct values indicate comparable expression levels. (B) Scatter plot demonstrating a strong positive correlation between ISG15 and LGALS3BP expression. (C) ROC curves for the model (AUC = 0.95), ISG15 (AUC = 0.80), LGALS3BP (AUC = 0.60) and the interaction of ISG15 and LGALS3BP (AUC = 0.86). (D) Cut-off probability values for early pregnant and non-pregnant buffaloes. The duplex assay exhibits a specificity of 95% and a sensitivity of 85%.
Of the 39 animals (validation data from study 2) from the validation data set 2, the table shows the negative predictive value (NPV) and positive predictive value (PPV) of the probe-based duplex assay of RT-qPCR
We further validated the robustness of the duplex qPCR assay in study 3 using 41 randomly selected animals. Similar to the result of study 2, ISG15 expression was significantly upregulated in pregnant buffaloes at 20 ± 2 days post-AI (13.0 ± 2.43 vs. 3.0 ± 0.79; p < 0.0001) (Fig. 4A). LGALS3BP exhibited a marked tendency towards increased levels in pregnant buffaloes (2.28 ± 0.54 vs. 1.44 ± 0.21; p = 0.08) (Fig. 4B). The predictive value (Y) of pregnancy was significantly higher in pregnant compared to non-pregnant animals (0.24 ± 0.07 vs. −0.06 ± 0.03; p < 0.0001) (Fig. 4C). The AUC for Y was 0.89 (p = 0.0002). In this, 9 out of 41 animals were identified as TP, while 3 animals were detected FP (Fig. 4D). Additionally, 26 out of 41 buffaloes were TN, and 3 buffaloes were found FN. The NPV and the PPV in this dataset were 90% and 75%, respectively (Table 3).
Revalidation of duplex qPCR analysis of ISG15 and LGALS3BP for early pregnancy detection in buffaloes. (A, B) Relative expression levels of ISG15 (A) and LGALS3BP (B) in peripheral blood mononuclear cells (PBMCs) of early pregnant and non-pregnant buffaloes, normalized to basal expression in non-pregnant animals. (C) Predictive threshold value (Y) for early pregnancy diagnosis, determined by receiver operating characteristic (ROC) curve analysis. (D) Cut-off threshold values (Y > 0.065) for differentiating early pregnant from non-pregnant buffaloes. The duplex assay exhibited a specificity of 90% and a sensitivity of 75% for early pregnancy detection.
Of the 41 animals (validation data from study 3) from the validation data set 3, the table shows the negative predictive value (NPV) and positive predictive value (PPV) of the probe-based duplex assay of RT-qPCR
Predicting power of pregnancy was based on full linear model with interaction.
Discussion
Early and accurate identification of pregnant buffaloes is crucial for effective herd management. Traditionally, different classical and metabolite-based methods, like rectal palpations, ultrasound and hormonal assays, have been used. However, all these techniques incur some kind of limitation in terms of precision, timing and economic efficiency. Alternatively, a molecular approach targeting gene expression, specifically interferon-stimulated genes (ISGs), provides high sensitivity and effective methodology for early pregnancy diagnosis. The expression of ISGs in PBMC, in bovines, during early gestation, has been suggested as a reliable indicator of pregnancy (Casano et al., Reference Casano, Menchetti, Trabalza-Marinucci, Riva, De Matteis, Brecchia, Inglesi, Rossi, Signorelli, Barile and Barbato2023; Nag et al., Reference Nag, Arunmozhi, Sarath, Asokan and Vijayarani2018; Yoshino et al., Reference Yoshino, Toji, Sasaki, Koshi, Yamagishi, Takahashi, Ishiguro-Oonuma, Matsuda, Yamanouchi, Hashiyada, Imai, Izaike, Kizaki and Hashizume2018). This study examines the simple indicator of early pregnancy in buffaloes utilizing classical marker ISG15 and new marker LGALS3BP for improving diagnostic methods of bubaline gestation. ISG15 expression levels in pregnant buffaloes were significantly higher at 20 days than those in non-pregnant buffaloes consistently in all our studies (studies 1–3). A simple threshold value was then determined to diagnose pregnancy status. The average ISG expression values of buffaloes during the oestrus cycle were significantly lower than those in pregnant buffaloes. We selected expression levels of two candidate genes: ISG15 and LGALS3BP, in PBMC and then measured the fold change value of non-pregnant buffaloes to determine early pregnancy.
During pregnancy, the physiological changes in the buffalo’s reproductive system include significant increases in blood flow to the uterus and placenta to support the developing foetus. Blood flow area is the best predictor of pregnancy in cattle on days 17–20 (Kanazawa et al., Reference Kanazawa, Seki and Iga2022). Recognition of the embryonic vesicle and embryo proper with a heartbeat is used as the criterion for positive diagnosis of pregnant buffaloes (Ali and Fahmy, Reference Ali and Fahmy2008; Barile et al., Reference Barile, Menchetti, Casano, Brecchia, de Sousa, Zelli, Beckers and Barbato2021; Pawshe et al., Reference Pawshe, Appa Rao and Totey1994).
Doppler ultrasonography is used for early pregnancy detection in cattle herds, typically between days 20 and 22, with an accuracy exceeding 90% in several studies (Pugliesi et al., Reference Pugliesi, Miagawa, Paiva, França, Silva and Binelli2014). This method boasts high sensitivity, nearing 100%, and produces very few FNs; however, it has a 15–20% likelihood of false-positive diagnoses (Fontes and Oosthuizen, Reference Fontes and Oosthuizen2022). An additional method for early pregnancy detection involves assessing CL blood perfusion, where decreased blood flow in the CL can signal pregnancy loss or absence, as it indicates a potential pregnancy (Fontes and Oosthuizen, Reference Fontes and Oosthuizen2022). Our findings reveal a striking difference in the maximum vascularization observed in the CL. Specifically, the percentage area of BP relative to the total area of CL was found to be highly significant (p < 0.0001) in pregnant buffaloes at the 20-day stage compared to non-pregnant buffaloes at the same stage after insemination. These results shed light on the dynamic vascular changes occurring in the CL during early pregnancy in buffaloes.
We measured the expression levels of ISG15 and LGALS3BP genes in PBMCs of buffaloes at different days: 0 (before insemination), 20, 25 and 40 days post-insemination. At 20 days post-AI, increased ISG15 (2.6-fold vs. 1.17-fold; pregnant vs. non-pregnant) and LGALS3BP (2.21 = fold vs. 1.07-fold; pregnant vs. non-pregnant) levels were observed, similar to the previous findings in cows and buffaloes (Ferraz et al., Reference Ferraz, Filho, Rocha, Neto, de Andrade Bruni, Oshiro, Baruselli, Lima and Pugliesi2021; Sharma et al., Reference Sharma, Choudhary, Ratta and Singh2023). No upregulation of ISG15 was observed in non-pregnant buffaloes, which signifies the absence of a viable conceptus. Low expression level of ISG15 in blood was the indicator of non-pregnant cows (Han et al., Reference Han, Austin, Rempel and Hansen2006). The elevated expression of ISG15, along with increased blood flow in CL during ultrasonography, will increase the accuracy of early pregnancy detection in buffaloes.
The LGALS3BP gene belongs to the galectin family and has a significant contribution to the interaction of cell–cell and cell–matrix. The elevated expression of LGALS3BP transcripts in uteri of early pregnant buffaloes (Baba et al., Reference Baba, Panigrahi, Verma, Sadam, Sulabh, Chhotaray, Parida, Krishnaswamy and Bhushan2019), especially in the luminal epithelium of uterine wall (Bauersachs et al., Reference Bauersachs, Ulbirch, Gross, Schmidt, Meyer, Wenigerkind, Vermehren, Sinowatz, Blum and Wolf2006), signifies its contribution to the attachment of conceptus to the uterine wall. The elevated expression of LGALS3BP transcripts at 20 days post-AI in pregnant buffaloes when compared with the same animals at day 0 non-pregnant buffaloes signifies galectin 3-binding protein contribution in establishing pregnancy. The elevated expression of LGALS3BP was consistent with the expression in PBMC of pregnant heifers at day 18 (Ferraz et al., Reference Ferraz, Filho, Rocha, Neto, de Andrade Bruni, Oshiro, Baruselli, Lima and Pugliesi2021; Rocha et al., Reference Rocha, da Silva Andrade, de Melo, Motta, Coutinho, Gonella-Diaza, Binelli and Pugliesi2020).
ROC analysis established the strong diagnostic efficacy of this molecular approach, emphasizing its advantage in accurately differentiating pregnant from non-pregnant animals during the early gestation period (Kanazawa et al., Reference Kanazawa, Seki and Iga2022; Yoshino et al., Reference Yoshino, Toji, Sasaki, Koshi, Yamagishi, Takahashi, Ishiguro-Oonuma, Matsuda, Yamanouchi, Hashiyada, Imai, Izaike, Kizaki and Hashizume2018). The AUC of accurate pregnancy prediction using a simple scoring system in pregnant women with foetal heartbeat was 0.901 (Bottomley et al., Reference Bottomley, Van Belle, Kirk, Van Huffel, Timmerman and Bourne2013) and a lower AUC of 0.724 for predicting an ongoing pregnancy (Liu et al., Reference Liu, Wei and Wang2024). The AUCs obtained in our case were 0.95 and 0.90 in study 2 and study 3 datasets, respectively. The AUC of ISG15 as a pregnancy predictor in cow PBMC was 0.92 (Ferraz et al., Reference Ferraz, Filho, Rocha, Neto, de Andrade Bruni, Oshiro, Baruselli, Lima and Pugliesi2021). It is worth mentioning here that 100% success for the diagnosis of pregnancy status of buffalo cannot be practically obtained due to many reasons. Of those, early embryonic mortality is a key reason. In buffalo, being the seasonal breeder, the incidence of early embryonic mortality varies from 7% to 40%. Early embryonic (before 21 days of gestation) mortality and pregnancy failure rates rise in buffaloes mated during the transition from winter to spring, when daylight hours are lengthening (Campanile et al., Reference Campanile, Neglia and D'Occhio2016). Buffaloes conceived during daylight lengthening experienced higher embryo loss (20–40%), while those conceived during decreasing daylight had lower loss (7%) (Barile et al., Reference Barile, Terzano, Pacelli, Todini, Malfatti and Barbato2015; Campanile et al., Reference Campanile, Neglia and D'Occhio2016). The false outcomes are attributed to animal biological differences, disease conditions like brucellosis, mastitis and other reproductive diseases that may influence the expression of ISGs. When buffaloes are not pregnant, attempts should be made to redo insemination in the next oestrous cycle. In field conditions and many buffalo, hormonal treatments such as prostaglandins or gonadotropin-releasing hormones are injected to induce oestrus or to synchronize oestrus. Such hormonal treatment runs the risk of affecting the viability of embryo, altering gene expression of PBMC, and thus can provide FN results.
Additionally, our results show a high correlation between ISG15 and LGALS3BP expression (r = 0.78; p < 0.0001), suggesting a functional relationship that might play a crucial role in early pregnancy events in buffaloes. This is in line with findings in cattle, where ISG15 has been shown to be an important marker during the maternal recognition of pregnancy. LGALS3BP, a protein involved in cell adhesion and immune modulation (Margadant et al., Reference Margadant, Van Den Bout, Van Boxtel, Thijssen and Sonnenberg2012), might work synergistically with ISG15 to support the establishment of pregnancy. This gene interaction model is novel in the context of pregnancy detection, providing insights that may improve the accuracy of diagnostics compared to traditional methods such as ultrasonography or progesterone assays.
When considering the practical application of this method, the NPV and PPV of 95% and 85%, respectively, highlight the assay's reliability. Compared to other pregnancy detection studies that report NPV values between 80% and 90%, our assay appears to provide higher diagnostic confidence within 20 ± 2 days post-insemination. This is a critical period for pregnancy detection in livestock management, as early detection allows for timely decisions regarding reproductive interventions or re-breeding.
While the assay is highly promising, the study also showed a critical limitation: elevated levels of ISGs during disease conditions such as viral disease (Lang et al., Reference Lang, Li, Luo, Liu, Zhang, Guo, Xu, Bao, Dong and Yu2022), which leads to confusion between pregnant and diseased buffaloes. This overlap indicates the need for further refinement to ensure its reliability and specificity.
In conclusion, this study demonstrates the enhanced diagnostic accuracy of combining ISG15 and LGALS3BP gene expression for early pregnancy detection in buffaloes. TaqMan-based duplex assay offers a better alternative to classical methods of pregnancy diagnosis, with promising potential for improving the reproductive efficiency. A combination of ultrasonography with qPCR assay would significantly improve predictive power, accuracy and early detection periods. Further research should explore the application of this model in larger buffalo populations and across different breeds to validate the robustness of these findings.
