Recombinant cystatin-like protein-based competition ELISA for Trichinella spiralis antibody test in multihost sera

Objectives Trichinella spiralis is a zoonotic parasite with a complex parasitic life cycle and exposed to animals or humans by infectious meat. To control transmissions of T. spiralis through the food chain to humans, sensitive and selective multihost sera-diagnosis is urgent needed for monitoring T. spiralis exposure. Methods A competition enzyme-linked immunosorbent assay (cELISA) for T. spiralis infection diagnosis in multihost sera was developed based on recombinant cystatin-like protein (rCLP-cELISA) as well as monoclonal antibodies. The sensitivity and accuracy of the rCLP-cELISA were quantified using swine (n = 1316), mice (n = 189) and human (n = 157) serum samples. T. spiralis-antibody targeting test ability of the rCLP-cELISA in swine (n = 22) and human (n = 36), instead of other parasites or viruses antibodies, was evaluated. Results The rCLP-cELISA showed high agreement with commercial ELISA kits in field swine sera assessed by Cohen’s kappa value (κ = 0.7963). And it showed 100% specificity in human trichinellosis detection with sensitivity of 96.49%, no cross-reaction with other parasite or virus infections, and high positive detection rate of 87.5% in low-dose infected swine. Besides, the rCLP-cELISA exhibited potential in the detection of T. spiralis, T. nelsoni and Trichinella T8 infections. Conclusions The rCLP-cELISA can be used for T. spiralis-associated antibody test in multihost sera.


Introduction
Trichinella spiralis is an important zoonotic parasite and commonly associated with domestic pigs with highly infective for humans [1][2][3]. T. spiralis is accounts for serious economic losses in the pig industry and hundreds of human infections worldwide annually [2][3][4]. Domestic pork is the main source of human trichinellosis, which results from consumer ingestion of infectious meat [5,6]. In China, the southwestern, central and northeastern parts are the main endemic areas of T. spiralis infections in herds, which is consistent with the highest prevalence of human trichinellosis reported in Yunnan, Hubei and Henan provinces [7,8]. In these endemic provinces, the habit of eating raw/undercooked pork meat is common, which accidentally introduces humans into the exposure. The potentially high rate of infection in domestic herds is responsible for the threat of human infection by T. spiralis, therefore, monitoring the exposure of herds, vectors and populations is beneficial for T. spiralis control and public health.
Parasites of the genus Trichinella are transmitted among hosts by the ingestion of infected muscles revealing two epidemiological cycles: domestic and sylvatic [9,10]. The domestic cycle occurs particularly between domestic swine and rodents [9][10][11]. With the presence of illegal feeding by left-over food containing Trichinella larvae, an important way of introducing the parasite to swine is transmission by vectors, such as rodents, in pig farms without proper control and presenting low levels of sanitary conditions in rural areas [9,11]. Considering the growing scale of meat market of rodents [12,13] as the vector of T. spiralis transmission, as well as high prevalence of the infection in domestic swine [3,7,8], it is important to focus on epidemiological investigations of hosts in the food chain for risk management and control of this food-born parasite.
Serological assays such as enzyme-linked immunosorbent assay (ELISA) have been recommended by the International Commission on Trichinellosis (ICT) for surveillance of T. spiralis infections in herds to ensure food safety [14,15]. ELISA is also the first recommendation for screening of human trichinellosis, since clinical symptoms are nonspecific and resemble those of influenza or other disorders [15,16]. A number of ELISAs have been established for monitoring T. spiralis infections in animal hosts [17,18] or human host [19], respectively. Most of the assays are indirect ELISAs employing excretory-secretary (ES) products (ES-iELISA), a mixture-type antigens from muscle larvae collected strictly by living animals, which possess defects in reproducibility, quality control, high cost and cross-reactivity [17][18][19]. To control transmissions of T. spiralis through the food chain to humans, sensitive and selective multihost sera-diagnosis is urgent needed based on definitive antigenic components of ES products [18,20]. To date, only one multihost universal test has been reported [21] for the detection of T. spiralis infection, in which the mixed antigens were employed.
In our previous study [22,23], an antigenic cystatin-like protein (CLP) was proved to be expressed in various developmental stages of T. spiralis, and monoclonal antibodies (MAbs) against CLP were prepared [23]. In this study, a recombinant CLP (rCLP)-based competition ELISA (rCLP-cELISA) employing MAb for the detection of T. spiralis antibodies was proposed and evaluated using multiple host serum samples. Compared with commercial ELISA kits (ES-iELISAs, swine and human), the proposed rCLP-cELISA showed similar sensitivity and no cross-reactions with other parasite or virus infections in swine and human sera.

Ethics statement
Human and animal serum samples, as well as different species/genotypes of Trichinella larvae, were collected and conserved by the OIE Collaborating Centre for Food-Borne Parasites from the Asian-Pacific Region, Jilin University. All experiments in this study were approved by the Ethical Committee of Jilin University, China (permit number: 20170318 for animal serum samples; 2019-H-K13 for human sera). All the human serum samples were collected from adults, and the written informed consent was acquired from the adults before samples were used. The authors confirm that the ethical policies of this study, as noted on the journal's author guidelines page, have been adhered to ethical approvals as this report solely contains diagnostic samples taken in the field.

Parasite and sera
A total of 1720 serum samples used in this study were shown in Table 1. Negative swine sera were reconfirmed with a commercial ELISA kit (Qiagen, Cat. 273501), the corresponding diaphragm tissues of which were tested by artificial digestion [24]. Positive sera from swine and mice infected with T. spiralis (iss 534) were collected and stored at -80˚C. Positive human sera were collected from two family infection cases in which pork was privately raised and slaughtered in Yunnan province. Serum samples from 100 healthy donors from the nearby villages of the two cases were reconfirmed with a commercial ELISA kit (Abcam, ab108780). Sera from patients with Clonorchis sinensis were collected in Jilin province. Sera from patients with enterovirus 71 (EV71), coxsackievirus A16 (CA16) and other parasites infections were provided by National Institute of Parasitic Diseases of Chinese Centre for Disease Control and Prevention.

Immunoassay procedure of the rCLP-cELISA and cut-off values
The MAb 1H9 binding to native epitope 39 HEALFSSDLKQESGV 53 of CLP (GenBank: ABY60755.1) was selected using competitive ELISA (S1 Text) and was labelled with biotin by Sangon Biotech (Shanghai) Co., Ltd. Briefly, 96-well ELISA plates were coated with 1.25 μg/mL rCLP antigens in 100 μL per well. The plates were blocked with PBST containing 1% BSA at 37˚C for 1 h. MAb 1H9 with biotin conjugation (0.3 μg/mL) was prepared in 0.9% NaCL solution (w/v). Equal volumes of the serum (50 μL) and the MAb solution (50 μL) were dispensed into wells and incubated at 37˚C for 1 h. Unbound serum and/or MAb were removed by washing steps. Then, HRP-labelled avidin (Invitrogen, LOT 2197902) was diluted 1:500 in PBST containing 1% BSA, added to wells and incubated at 37˚C for 30 min. After three washing steps, colour was developed for 8 min using TMB substrate solution (Tiangen The result for each sample was expressed as the competitive inhibition (PI) ratio [21]. The PI ratio was defined as follows: PI = (1 -OD sample /OD MAb ) × 100%. The cut-off value in rCLP-cELISA was calculated based on the mean of negative sera plus two standard deviations (SD) (swine, n = 270; mice, n = 30) or the receiver operator characteristics curves (ROC) analysis (human, n = 157) [21,25,26].
In ES-iELISAs, the signal-to-positive (S/P) ratio, cut-off values in swine sera, the OD 450nm values of human sera as well as controls (negative, cut-off and positive controls) were all calculated according to the manufacturer's instructions.

Evaluation of the rCLP-cELISA performance
2.3.1 Analysis of antibody kinetics using T. spiralis infected swine sera. Antibody kinetics of T. spiralis infection in swine at various stages were determined using rCLP-cELISA and ES-iELISA (Qiagen). Serum samples were obtained from three different infection doses (200, 400 and 600 larvae) at different days post infection (dpi). Seroconversion was defined as the time point when the concentration of T. spiralis antibody caused a PI ratio or S/P ratio that exceeded the cut-off value.

Accuracy test using T. spiralis infected swine sera.
The accuracy of the rCLP-cELISA was evaluated by analyzing sera from swine infected with three different doses of T. spiralis including 50, 100 and 200 larvae. The results were compared with those obtained by ES-iELISA (Qiagen) using the same serum samples. According to the recommendation of the ICT [14,24], diaphragm tissue of individual pig was detected with artificial digestion to calculate the mean larvae per gram (LPG) of muscle in order to confirm the larvae burden. Serum samples (n = 920) collected from backyard pigs from Hubei province were used to assess the sensitivity and specificity [27][28][29] of the rCLP-cELISA in the field, corresponding diaphragm tissues (100 g per pig) were detected to confirm the infection status of each pig.

2.3.3
The rCLP-cELISA validation in mice sera. Antibody kinetics of T. spiralis infection in mice at various stages were determined using rCLP-cELISA. Serum samples were collected from four different infection doses (10, 20, 200 and 400 larvae) on different infection days.
Serum samples of mice infected with 17 isolates belonging to 12 species/genotypes of Trichinella were collected. The PI ratio of each serum sample was compared with the threshold to evaluate the potential suitability of the rCLP-cELISA used in sera infected with different species of Trichinella larvae. These samples were reconfirmed by crude worm extract-western blot (CWE-WB) using corresponding larvae and MAb 1H9 according to previous studies [23].

The rCLP-cELISA validation in human sera.
The developed rCLP-cELISA was validated by analyzing 100 healthy donors sera and 57 sera from trichinellosis patients using ROC analysis. Six of these positive sera with different OD 450nm values as well as different PI ratios were analyzed by western blot employing ES products (ES-WB) of T. spiralis (iss 534). ES products were produced as described in a previous study [22], and the procedure for ES-WB was performed as published protocols [30,31]. Two serological negative sera coming from the nearby villages of trichinellosis patients were used as controls.

Serological cross-reactivity validation of the rCLP-cELISA.
The cross-reactivity of the rCLP-cELISA was evaluated using sera of other parasite infections and virus immunizations/infections in swine and human hosts, which compared with that of ELISA kit. According to the recommendation of the ICT [14,15], serological positive samples by ELISAs were reconfirmed with ES-WB, as well as the western blot employing rCLP (rCLP-WB). The procedure for rCLP-WB was performed as our previous study [23].

Statistical analysis
The results were expressed as the mean ± SD of the antibody kinetics of T. spiralis infection in swine and mice. Other results were expressed as the mean. Statistical analysis was performed by GraphPad Prism 6 software for Windows.

Analytical performance of the rCLP-cELISA in swine sera
The MAb 1H9, with highest negative-to-positive ratio among three MAbs, was suitable for the development of rCLP-cELISA (S1 Text). The cut-off value of the rCLP-cELISA was determined to be 52% (S1 Fig). Antibody dynamics were determined in three different dose groups with rCLP-cELISA and ES-iELISA (Fig 1). The continuous serological positivity detected by rCLP-cELISA occurred later than that of ES-iELISA in the 600 and 400 larvae dose groups, in which seroconversion for continuous serological positivity delayed one sampling time point of 5 (600 larvae, Fig 1A and 1D) or 15 (400 larvae, Fig 1B and 1E) days. This delay time in seroconversion using rCLP-cELISA cannot be determined accurately due to the setting interval of sampling time points. When the infective dose was reduced from 600 to 200 larvae per pig, seroconversion for continuous serological positivity was delayed from 35 to 90 dpi in rCLP-cELISA and from 30 to 90 dpi in ES-iELISA (Fig 1A, 1C, 1D and 1F). The PI ratios exceeded the threshold at 17 dpi and 19 dpi (Fig 1D and 1E). A similar pattern of antibody dynamics was detected in the 200 larvae group, although PI ratios did not exceed the cut-off value (Fig 1F). The accuracy of rCLP-cELISA was evaluated through a recovery test (Fig 2), analyzing the matrix (sera) spiked samples of the low-dose groups including 50, 100 and 200 larvae. The results of rCLP-cELISA were compared with those detected by ES-iELISA. All pigs were examined by artificial digestion to calculate the mean larvae burdens (S1 Table). In 200 larvae group, the mean larvae burden of the diaphragm was less than 0.50 LPG, and the 100% positive rate of rCLP-cELISA was the same as that of ES-iELISA. When the mean larvae burden reduced to 0.025 LPG, 5/6 pigs were serologically positive using rCLP-cELISA. Since there were 4 positive pigs using ES-iELISA in the 50 larvae group, 3/4 pigs tested positive with rCLP-cELISA when the larvae burden of the diaphragm was as low as 0.005 LPG.
The agreement between rCLP-cELISA and ES-iELISA in field swine sera (n = 920) was assessed by Cohen's κ analysis ( Table 2). According to κ values reported previously [27], a substantial agreement (κ = 0.7963) was observed between the two methods. In Table 3, field samples were classified as true positive or negative according to their larvae burden in diaphragm. The results showed that the proposed rCLP-cELISA are high sensitivity (se: 1.0000) with specificity of 0.9956.

Evaluation of the rCLP-cELISA using infected mice sera
The cut-off value of the rCLP-cELISA in mice sera was 39% (S2 Fig). In the different infection dose groups of mice (Fig 3), PI ratios exceeded the threshold at 14 dpi followed by a decrease, and again exceeded the cut-off value at 35 dpi (10 larvae), 42 dpi (20 larvae), 30 dpi (200 larvae)

PLOS NEGLECTED TROPICAL DISEASES
and 21 dpi (400 larvae). The pattern of antibody dynamics detected in mice was similar to that in swine, regardless of whether the dose of infection was high (200 and 400 larvae) or low (10 and 20 larvae). Serum samples of mice infected with different species/genotypes of Trichinella were tested with rCLP-cELISA at 60 dpi. As shown in Fig 4A, PI ratios of sera from 5 isolates were significantly above the threshold, including T. spiralis (T1, iss 534, iss 4 and iss 533), T. nelsoni (T7, iss 37) and Trichinella T8 (iss 124), which reflected the diagnostic potential of rCLP-cELISA used in these isolates of Trichinella infections. The CWE-WB was performed to confirm the presence of the epitope identified by MAb 1H9 (Fig 4B). Consistent with the results of rCLP-cELISA, MAb 1H9 can identify different patterns of bands specifically with the crude antigens of corresponding isolates. For T. spiralis (T1) and T. nelsoni (T7), characteristic bands were found at about 45 kDa and 2 or 3 bands at 25-35 kDa. The band at about 45 kDa of Trichinella T8 (iss 124) was higher than that of the two species with no band at 25-35 kDa. Although similar bands were found in the other 12 isolates, their corresponding sera were not serologically

Preliminary validation of the rCLP-cELISA in humans
The developed rCLP-cELISA distinguished well between 57 sera of trichinellosis patients and 100 sera of healthy donors (Fig 5A). According to the ROC analysis (Fig 5B), the threshold was set as 40.72% when the specificity reached 100% with sensitivity of 96.49%. The area under the curve (AUC) of PI was 0.9981, showing that the rCLP-cELISA are highly accurate (0.9 < AUC < 1.0) [25,26]. This result indicated that the defined cut-off value was suitable for evaluating rCLP-cELISA result to determine the presence/absence of Trichinella antibodies in human serum. In Fig 5C, each serum sample of trichinellosis patient was provided two values, one recorded with ES-iELISA and the other one recorded with rCLP-cELISA. The correlation

PLOS NEGLECTED TROPICAL DISEASES
between the two ELISAs was intense (R 2 = 0.7887) [29], suggesting a direct relationship between OD scores in ES-iELISA and PI ratios in rCLP-cELISA in human sera. A three-band pattern at 45-65 kDa [16,30,32] was determined in the ES-WB test using positive sera (Fig 5D  lanes 2-6, black box). With PI ratios as well as OD values of positive sera decreased, the threeband gradually weakened, revealing the high consistency among the three methods. However, the characteristic pattern of sera with weak antibody positivity was not significantly distinguishable from those of antibody-negative samples (Fig 5D lanes 7-9). The results showed that the rCLP-cELISA exhibited higher sensitivity than traditional ES-WB in screening human trichinellosis [14][15][16].

Cross-reactivity of the rCLP-cELISA
All serum samples from swine infected with other parasites or immunized with virus vaccines were serologically negative by rCLP-cELISA and ES-iELISA (Fig 6A and 6B). Compared with the rCLP-cELISA, cross-reaction occurred in human sera in two patients with Clonorchis sinensis using ES-iELISA (Figs 6C and 6D, and S3), although OD 450nm values of other six samples in Clonorchis sinensis group did not exceed the threshold (S3 Fig). The two samples showing cross-reaction as well as one sample that slightly lowered the threshold were detected by ES-WB compared with rCLP-WB. Since the molecular weight of rCLP was about 47 kDa [23], no specific bands were tested using corresponding sera (Fig 6E). But the bands of the two samples at 45-65 kDa were tested in ES-WB (Fig 6F lanes 3-4), which were partly consistent with the three-band pattern of the positive reference sera (Fig 5D lanes 2-7). In addition, the reactivity of antibody-negative samples tested by ES-iELISA in the two regions (Yunnan and Jilin provinces) were completely different (Figs 5D lanes 8-9 and 6F lane 2). The results indicated that rCLP-cELISA exhibited higher selectivity than ES-iELISA in patients with Clonorchis sinensis.

Discussion
ES products, as a mixture of multiple proteins (S4 Fig), require appropriate culture conditions and suitable purification methods to ensure the specificity of the detection results [17][18][19]. In the case of that every component of ES products is unknown, inter-batch and intra-batch quality controls of ES production are difficult. In addition to the differences in species/genotypes of Trichinella and in host species, the seroconversion time point of swine with the same infection dose varied greatly (S2 Table), which was largely related to the quality differences in ES products among different laboratories.
Our previous studies have shown that indirect ELISA employing rCLP can detect specific antibodies in swine sera collected from pigs infected with high dose (� 1000 larvae per pig) of T. spiralis [23]. MAb 1H9 against epitope 39 HEALFSSDLKQESGV 53 of CLP was evaluated and found to be useful for establishing a competition ELISA. Detection methods based on specific antigens and MAbs can be beneficial to improve sera-diagnosis of T. spiralis infection, taking advantage of its controllable quality, low cost, good stability and reproducibility. The limit of detection (LOD) of artificial digestion was 1.0 LPG, the result of which was not stable when sampled near the LOD due to the limited sample size of tissue and personnel operation factors [24,33]. According to previous studies (S2 and S3 Tables), we focused on the sensitivity of rCLP-cELISA at a low infective dose (< 800 larvae per pig). The rCLP-cELISA showed a 100% positive detection rate when the larvae burden was lower than 1.0 LPG as well as 66.7% when the larvae burden was even lower than 0.03 LPG. And it had a good agreement rate with ES-iELISA in field samples (κ = 0.7963). Results of field samples testing showed that rCLP-cELISA had considerable sensitivity (se = 1.0000) with good application prospects for monitoring swine infection status to ensure food safety.
Previous studies have shown that there is a delay in seroconversion in swine at low infective doses [15]. Our results also confirmed that when the infective dose was reduced from 600 to 200 larvae per pig, the load of larvae in diaphragm tissues reduced 10-fold (S3 Table), and continuous serological positivity was delayed by 60 days using ES-iELISA and by 55 days using rCLP-cELISA. It is worth noting that the antibody dynamic curve of CLP was completely different from that of ES at low infective doses. In rCLP-cELISA, the abundance of specific antibodies against CLP shifted below the threshold at 19 dpi up to 35 dpi (600 larvae) or to 60 dpi (400 larvae). Considering the delay time in antibody production for hosts, the deficiency of specific antibodies might be caused by the lack of CLP in newborn larvae with inadequate CLP expression in adult worms [22]. Moreover, the absence of persistent increases in CLP-associated antibodies resulted from parasitic immune status of the host [34], such as immunosuppression [35][36][37] or immune tolerance [38,39] in T. spiralis intestinal infection stage, to restrict the production of specific antibodies. In addition, a similar pattern of the antibody dynamic curve of CLP was demonstrated in mice in different dose groups. In the 400 larvae group, the level of specific antibodies against CLP remained above the threshold, which was consistent with our previous findings in swine at 50000 and 1000 larvae infections [23]. As detected in swine samples, CLP antibody levels in 10 and 20 larvae groups in mice also tended to change from above the threshold to below and then to consistently positive.
Although scientists have debated whether rodents are accidental hosts or stable reservoirs in livestock circulation (S4 Table), it is undeniable that rodents play an important role in the transmission of Trichinella and are also an important route for swine infection [9][10][11][12]. T. spiralis (T1) originating from Eastern Asia is now distributed world-wide due to its highly infective for domestic and sylvatic swine, mice and rat [1]. The risk of Trichinella infection in pig herds is presented mainly by T. spiralis (T1), representing a significant public health risk especially in Asia [1,7,8]. The proposed rCLP-cELISA was able to test specific antibody in mice infected with 3 isolates of T. spiralis (T1) and performed well in sera-diagnosis of T. nelsoni (T7) and Trichinella T8 infection. The results confirmed that the proposed rCLP-cELISA had the potential for cross-species diagnosis of Trichinella infection. For the twelve species/genotypes, the inconsistencies in the results of CWE-WB and rCLP-cELISA revealed differences in the antigenicity of the same epitope recognized by MAb 1H9 between these isolates. In addition, two different isolates of the same species, T. murrelli (T5), occasionally showed the presence or absence of the epitope (Fig 4B lanes 9-10 and S2 Text). To date, three proteins belonging to the cystatin superfamily have been reported, including CLP, muti-cystatin-like domain protein 1 (MCD-1) [40] and T. spiralis novel cystatin (TsCstN) [41], with no report on serological diagnosis-related applications about other two proteins. Epitopes recognized by MAb 1H9 were confirmed in 11 isolates, but the size and abundance of these bands varied from each other, suggesting the possibility of the presence of novel proteins that may or may not belonged to the cystatin superfamily. And in terms of inhibition rate, there are great differences in antigenicity between them.
Most of the data in the specificity study of ES-iELISA are from swine [14,15], ES-iELISA as well as rCLP-cELISA has good selectivity in the sera-diagnosis of swine infection. In contrast to rCLP-cELISA, ES-iELISA had cross-reactions in the serological detection of Clonorchis sinensis patients (Figs 5D and 6E and 6F, and S3). But rCLP-cELISA showed 100% specificity in the detection of patients with enterovirus and other parasites infections. The two samples that cross-reacted with ES-iELISA were tested by ES-WB, and bands of similar size to trichinellosis positive sera were found (Figs 5D and 6F) with no bands found in rCLP-WB for corresponding samples (Fig 6E). Besides, all of the 8 samples from patients with Clonorchis sinensis showed higher OD values than other 28 samples (S3 Fig), although values of only 2/8 samples in Clonorchis sinensis group exceeded the threshold. Moreover, the two groups of samples were from two provinces with relatively far geographical locations, local residents had different dietary habits, and the negative serum had different band reactions in ES-WB (Figs 5D lanes 8-9 and 6F lane 2). Thus, it was believed that partly similar band patterns can be caused by the reaction between ES and Clonorchis sinensis infected human sera. However, rCLP-cELISA showed better selectivity than ES-iELISA with PI ratios lower than 40% in these samples. Besides, the ROC analysis of 100 negative and 57 positive serum samples showed that rCLP-cELISA can distinguish between the presence and absence of Trichinella antibodies with considerable specificity (100%) and sensitivity (96.49%). Therefore, the proposed rCLP-cELISA has great potential in sera-diagnosis of human trichinellosis.
In conclusion, a competition ELISA employing the CLP and MAb was proposed and evaluated using positive sera from multiple T. spiralis infected hosts. Compared to ES-iELISA, the proposed rCLP-cELISA showed a similar positive detection rate in swine at low infection doses, with the seroconversion slightly delayed. In addition, the rCLP-cELISA exhibited good versatility in mice infected with different species/genotypes of Trichinella, and can distinguish the status of trichinellosis patients from healthy donors with considerable sensitivity and specificity. Taking advantage of its controllable quality stability, the rCLP-cELISA had potential applications for T. spiralis-related serological investigation and screening for multihost serum samples. With the development of MAb modification strategies [42,43] and the discovery of antigenic proteins from ES products [20,44], the proposed rCLP-cELISA provides useful reference for the improvement of serological assay for monitoring the exposure of T. spiralis.