Validation of Buffalo PAG Antisera for Early Pregnancy Diagnosis in Bovine 

Early pregnancy diagnosis is crucial for effective reduction of the massive economical and biological losses associated with fertility problems in animal husbandry practice (Rao Krishna and Veena, 2009). Although the diagnosis of pregnancy depends on many factors, the sensitivity and specificity of the method are the most important factors that determine the effectiveness of the diagnostic method (Abdullah et al., 2017). The ability to diagnose pregnancy using pre-determined concentrations of PAG based on these specific antibody combinations have been identified (Al-Samawi et al., 2015). At present, blood tests for early detection of pregnancy in cattle based on PAGs are commercially available (Northrop et al., 2019). However, the heterologous RIA systems were developed for evaluating accurate measurements of PAG concentration in cows.

Bovine PAG with a molecular mass of 67 kDa was isolated from fetal cotyledons and characterized by Zoli et al., (1991). These authors reported later that this PAG can be detected by RIA (RIA-1, AS#497) in maternal circulation in all cows from 30 to 35 Days of pregnancy (Zoli et al., 1992). Some years later, PAG molecules were isolated from caprine placental extracts (Garbayo et al., 1998). A semi purified preparation containing caPAG55+62kDa was used for the production of a polyclonal antiserum (RIA 2, AS#706), which has been shown to be able to measure increasing concentrations of PAG in pregnant cows as early as days 21-25 after conception (Perenyi et al., 2002). Attempts have been made to isolate and purify pregnancy associated glycoproteins in buffalo but only partial purification could be achieved (Barbato et al., 2003).  Protein sequence with the name of PAG75_BUBBU (P85048) has been submitted by Barbato et al., (2006). The apparent molecular masses of the immunoreactive bands determined by western blotting with anti-PAG sera were ranged from 59.5 to 75.8 kDa in early pregnancy placentas and from 57.8 to 73.3 kDa in the mid-pregnancy and late-pregnancy placentas. After the results of microsequencing of the N-terminal extremity, buffalo PAG appeared as new members of the PAG family (Barbato et al., 2008). More recently, the isolation and purification of PAGs from buffalo placenta allowed for the development of a specific RIA system for buffalo (Barbato et al., 2017).

Despite the fact that a number of studies have been published on early pregnancy diagnosis in bovine by using different antibodies types to determination of PAG concentration (Perenyi et al., 2002). The use of antibody against buffalo PAG to detect PAG in bovine samples was not yet reported. In this study, we evaluated the availibity of determination of blood PAG concentration and pregnancy diagnosis in bovine by RIA system using antibody against buffalo PAG in comparison with systems using antisera against bovine and caprine PAG.

Animals and sample collection
 
The study was performed on 437 commercial samples from cows in Belgium. All animals were reared within the herds. The samples were collected after day 21 since artificial insemination or natural mating. The heparinized blood samples were withdrawn from each animal from the tail vein. These blood samples were centrifuged (15 min at 1500 x g) within 30 min of collection and plasma was stored at -20^oC until assayed.
 
Antibody
 
Three radioimmunoassays (RIA 1, RIA 2 and RIA 3) differing in antiserum, were used in this experiment to measure the PAG concentrations in samples. In RIA 1 (AS#497) anti-bovine PAG 1 (67 kDa, rabbit polyclonal antiserum raised against the major bovine PAG-1) (Zoli et al., 1992), in RIA 2 (AS#706) anti-caprine PAG (rabbit polyclonal antiserum raised against mixing caprine PAG 55 kDa and PAG 62 kDa (accession numbers P80935 and P80933) (Perenyi et al., 2002) and in RIA3 (AS#859) anti-buffalo PAG (rabbit polyclonal antiserum raised against mixing buffalo PAG molecular mass ranging from 61 to 73 kDa (Accession numbers P86373 and P86374) were used (Barbato et al., 2017). These antisera were produced from rabbits after intradermally immunizing them with the appropriate antigen.
 
Radioiodination of PAG tracers
 
Iodination (NaI125, PerkinElmer, Boston, USA) was carried out according to the chloramine T method (Zoli et al., 1992). The antigen was dissolved in phosphate buffer (0.2 M, pH 7.5) to obtain 1 mg/ml concentration. The radioiodination mixture was prepared by adding 10 ml of chloramine T solution (5 mg/ml dissolved in water) and 10 ml of  NaI125 (1 mCi, approximately 3.7 ´ 107 disintegrations per second) to 10 ml of antigen solution. After one minute of stirring 10 ml of sodium metabisulfite solution (30 mg/ml dissolved in water) were added in order to terminate the reaction. This mixture was loaded onto a G-75 column, which was previously calibrated with Tris BSA buffer (0.025 M, pH 7.5) for the separation of free I125 and labeled antigen. Eluted fractions of 1 ml were collected. The fractions were tested before binding by Geiger mini counter. The fractions giving more than 1000 count per second were selected to use.
 
Radioimmunoassay procedure
 
A PAG radioimmunoassay was performed using the methods of Perenyi et al. (2002). Briefly, 0.1 ml of PAG standard (bovine PAG 67 kDa purified according to the protocol of Zoli et al., (1991) at dilutions ranging from 0.8 to 100 ng/mL) and 0.05 ml of different fluid samples were diluted in 0.3 ml of Tris-BSA buffer. After adding an appropriate dilution of antisera (0.1 ml), 0.1 ml of radiolabelled 125I-PAG was poured in. Both the serum samples and standard tubes were incubated overnight at room temperature. The next day, the addition of 1 ml second antibody polyethylene glycol (PEG) solution (0.17% (vol/vol) normal rabbit serum, 0.83% (vol/vol) sheep anti-rabbit IgG, 0.3% (wt/vol) BSA, 4% (wt/vol) polyethylene glycol 6000 in Tris buffer and incubated for 30 min at room temperature. Finally, 2 ml of Tris-BSA buffer was added to the tubes and centrifuged (20 min at 1,500 x g). After centrifugation, the tubes were aspirated and counted using a gamma counter. Details concerning RIA-1 and RIA-2 validations were previously reported by different authors (Perenyi et al., 2002).
 
Data analysis
 
Results are expressed as mean±S.E.M. The minimum detection limit was determined as the mean concentration minus twice the standard deviation of 20 replicates of the zero standards. Parallelism was assessed by serially diluting PAG standard with PAG-free serum collected from non-pregnant bovine.

The correlation between the different RIA was calculated by comparing the systems two by two (RIA 1 versus RIA 2, RIA 1 versus RIA 3 and RIA 2 versus RIA 3) using the ANOVA single factor in Excel office program. Regression and the variance analysis were done by Excel. The regression between measurements based on two different RIA systems was expressed as:

Y= aX + b

Where,
X and Y are PAG concentrations measured by the different RIA.

The serial dilutions of different samples showed dose response curves parallel to the standard curves at all the three RIA systems (Fig 1).



The results of analysis of a total of 437 bovine samples is presented in Table 1. The proportions of PAG values in the ‘negative’ column were significantly higher (P<0.05) for the RIA-3 system (AS#859) compared to the RIA-2 (AS#706) and RIA-1 (AS#497) systems. The rate of doubtful was lowest in RIA-3 system. 



Table 2 showed ED 80, ED 50 and ED 20 in three RIA system using different antisera. ED 80 was higher in RIA-2 and lower in RIA-1 and RIA-3. The value for the PAG concentration was only considered in samples with a PAG concentration between ED-80 and ED-20. It was observed that the average concentrations were not significantly different in RIA-1 and RIA-3 (3.07±2.00 and 3.16±2.77 ng/ml, respectively) and lower than that in RIA-2 (5.40±7.59 ng/ml). The regressions between the three systems were showed in Fig 2. Overall, the correlation between the three systems was high. However, the correlation between RIA-1 and RIA-3 was the highest (R2=0.9364).

​

​

The CaPAG12 contained a complete of 1140 nucleotides corresponding to an inferred polypeptide length of  380 amino acids, it was most related to boPAG15, sharing 90% nucleotide and 83% inferred acid amine sequence identities (Garbayo et al., 2008). Whereas the boPAG-2 is a polypeptide of 372 amino acids long, structurally related to boPAG-11, -12 and -13 (62%, 83% and 96% amino acid sequence identity, respectively) (Xie et al., 1994; Barbato et al., 2008). However, the percentages of identity with boPAG67 kDa varied with other boPAGs from 80% to 84.62% (Xie et al., 1995). The immunoassay methods for PAG by testing new antisera  can be explained by different phenomena such as the temporal expression of different PAG molecules during early pregnancy, the high N-terminal amino acid identities and distinct glycosylation patterns (and probably half-life) of PAG molecules purified from ovine, caprine or bovine placenta (Barbato et al., 2008). These parameters can explain the specific ability of different antisera to detect PAG during early pregnancy in cattle (Ayad et al., 2007; Chavatte-Palmer et al., 2006; Lopez-Gatius et al., 2007). The relationship of PAG gene in buffalo and bovine throughout the pregnancy was reported (Jerome 2012). The sequence nomenclatures were showed the maximum similarity between sequences of Bos Taurus and Bubalus bubalis href="#lotfan_2018">(Lotfan et al., 2018).

The measurement of PAG concentrations in peripheral blood in bovine was used not only for pregnancy diagnosis and following up on ongoing pregnancy physiology but was also used for pathological diagnosis, such as the embryonic status and fetal mortalities (Sousa et al., 2006). An original approach was developed by Lopez-Gatius et al., (2007), who determined plasma levels of boPAG (measured by both RIA-1 and RIA-2) and progesterone in Holstein Friesian dairy cows (Barbato et al., 2008). Venipuncture was performed on cows on Days 35, 42, 49, 56 and 63. The sensitivity of the PAG-RIA test was 11.1% at days 19-24 and reached 100% from day 31 after breeding. The results showed an interaction between the milk production and the PAG levels measured by both RIA systems (Lopez-Gatius et al., 2007). Although the PAG concentrations obtained by RIA-2 were higher than those obtained by RIA-1 (Perenyi et al., 2002; Ayad et al., 2007; Lopez-Gatius et al., 2007). In spite of this, the RIA-1 system was used more widely for pregnancy diagnosis in bovine species (Ayad et al., 2007). It was also demonstrated that false pregnancy diagnosis based on PAG concentrations was reduced in high-producing dairy cows when measured by RIA-2 than by the RIA-1 system (Lopez-Gatius et al., 2007).

With the antisera raised against PAG purified from different species including bovine, caprine and buffalo placenta with a single bovine PAG67 kDa preparation as standard and tracer, the results described in the present study showed that all RIA systems proved to be sensitive, repeatable and accurate for measuring PAG concentrations (Van Hanh et al., 2012). In this research, plasma concentrations of PAGs were determined by 3 heterologous RIA systems with a bovine standard and tracer. Similar to the results reported by Perenyi et al., (2002), the measured PAG concentrations in RIA-2 were always higher than this in RIA1. In our results, the PAG concentrations are also higher than in RIA-3 (P<0.05), meanwhile the PAG concentrations in RIA-1 and RIA-3 showed no significant difference (P<0.05). The correlation coefficient between the RIA-1 and RIA-2 system ranged from 0.87 to 0.96 (Perenyi et al., 2002). However in our study, the correlation between RIA-1 and RIA-3 with RIA-2 was 0.67 and 0.73 (respectively), with the correlation between RIA-1 and RIA-3 the best (r=0.936). Correlations of RIA in these cases are similar to that reported using different antisera to detect buffalo PAG (Van Hanh et al., 2012).

In conclusion, our results clearly show that a new antiserum (As#859) is available for early to diagnose pregnancy by determination of PAG in bovine. Whether, the accuracy of system using this antiserum (As#859) shown ability to recognize fewer doubtful results remains to be confirmed.

We thank Dr J. Sulon for help and advice with the RIA procedures and Mrs. R. Fares-Noucairi for her editorial assistance.