Protein A/G-based enzyme-linked immunosorbent assay for detection of anti-Pythium insidiosum antibodies in human and animal subjects

Objectives Pythiosis is a deadly infectious disease caused by Pythium insidiosum. Reports of both human and animal pythiosis are on the rise worldwide. Prognosis of the pythiosis patients relies on early diagnosis and prompt treatment. There are needs for an immunodiagnostic test that can detect the disease in both humans and animals. This study aims at reporting an optimized protocol for the development of a protein A/G-based enzyme-linked immunosorbent assay (ELISA) for the detection of anti-P. insidiosum antibody in multiple host species. Results A total of 25 pythiosis and 50 control sera, obtained from humans, horses, dogs, cats, and cows, were recruited for the assay development. With a proper ELISA cutoff point, all pythiosis sera can ultimately be distinguished from the control sera. The successfully-developed protein A/G-based ELISA can detect the anti-P. insidiosum antibodies in serum samples of both humans and animals. It is a versatile, feasible-to-develop, and functional immunodiagnostic assay for pythiosis.

Immunodiagnostic tests for pythiosis have gained popularity due to simplicity and short turnaround time. Various immunodiagnostic methods, such as immunodiffusion (ID), enzyme-linked immunosorbent assay (ELISA), hemagglutination (HA), immunochromatographic test (ICT), have been developed to facilitate the diagnosis of pythiosis [10][11][12][13][14][15][16][17][18][19][20]. Each immunodiagnostic test has some advantages/disadvantages over the others. For example, ID and HA detect the antibodies in both humans and animals but possess limited sensitivity. ELISA is a multi-step assay but exhibits relatively-high diagnostic performance. ICT is a rapid and user-friendly assay, but the development of this test is complicated.
Because reports of both human and animal pythiosis are on the rise worldwide, there are needs for an immunodiagnostic test that can detect the disease in a broad range of host species. Due to the lack of a versatile, feasibleto-develop, and functional immunodiagnostic assay, the current study aims at reporting an optimized protocol to develop an ELISA for the detection of anti-P. insidiosum antibodies in humans and other animals. The assay relied on the use of protein A/G, which is a molecule that can bind the immunoglobulins from various animal species [16].
A total of 25 serum samples were derived from 20 Thai patients, 4 horses, and a dog with the clinical features of pythiosis [1][2][3], together with either successful detection of P. insidiosum by culture and molecular analysis [6,[24][25][26] or serum anti-P. insidiosum antibodies by ICT [16]. Fifty control (negative) sera were obtained from 10 each of individuals, horses, dogs, cats, and cows with no clinical sign of pythiosis in the absence of serum anti-P. insidiosum antibodies by ICT. Positive (from a known pythiosis patient) and negative (from a healthy blood donor) control sera were analyzed in parallel. All serum samples were leftover specimens obtained from the Faculty of Medicine, Ramathibodi Hospital, Mahidol University, and the Faculty of Veterinary Medicine, Kasetsart University. Each serum sample was tested in duplicate by the protein A/G-based ELISA. ELISA value (EV) was an OD value of a serum sample divided by an OD value of the negative control serum. ELISA cutoff value was derived from the mean EV of the control sera plus 2, 3, 4, or 5 standard deviations (SD). Diagnostic performance (i.e., specificity, sensitivity, and accuracy) were calculated as described previously [15] ( Table 1). Distribution of EVs (representing a detectable level of anti-P. insidiosum antibodies) of the pythiosis (n = 25) and control (n = 50) sera from both humans and animals was shown in Fig. 1a. Mean EV of the pythiosis sera was 41.0 (range 23.1-46.9), whereas that of the control sera was 3.0 (range 0.4-12.2). The "mean EV + 5SDs" cutoff point (EV, 13.4) provided the best sensitivity, specificity, and accuracy for the detection of the anti-P. insidiosum antibodies (Table 1). Furthermore, protein A/G-based ELISA was performed using serial twofold dilutions (from 1:1600 to 1:102,400) of a serum sample from a human, a horse, and a dog with pythiosis, and a healthy individual (control). The serum dilutions and the obtained EVs of each animal species showed some degree of a linear relationship (i.e., the higher dilution, the lower EV) (Fig. 2). The result suggests that the protein A/G-based ELISA can quantitate an anti-P. insidiosum antibody level for monitoring the clinical course of pythiosis.
All serum samples were also tested in duplicate by the established human-specific ELISA employing goat anti-human IgG antibody conjugated with peroxidase (Bio-Rad Laboratories, California, USA) [15,23]. The anti-human IgG antibody-based ELISA with an optimal cutoff point (mean EV + 4SDs, 15.4; Table 1; Fig. 1b) did not detect the anti-P. insidiosum antibodies in any control sera tested (EV range 0.0-9.3). This assay can detect the anti-P. insidiosum antibodies only in the sera from humans (EV range 19.8-72.0) and a dog (at a relativelylow EV, 17.9), but not horses (EV range 6.1-6.6), with pythiosis. The result indicates that the goat anti-human IgG antibody could strongly bind human antibodies and slightly cross-react with dog antibodies. Based on the same serum samples from humans and animals, the antihuman IgG antibody-based ELISA showed equivalent detection specificity (100%), but markedly-lower detection sensitivity (84% vs. 100%), when compared with the protein A/G-based ELISA (Table 1).
With a proper cutoff point, all pythiosis sera can ultimately be distinguished from the control sera by the protein A/G-based ELISA (Table 1; Fig. 1a). Unlike the species-specific ELISAs established here (Table 1; Fig. 1b) and by other investigators [10][11][12][13][14][15], the protein A/G-based ELISA is capable of detecting not only the anti-P. insidiosum antibodies in a single animal species, but also in humans and various animals (i.e., horses and dogs) those are the most affected hosts of pythiosis. A key feature of our ELISA is the use of protein A/G (rather than a species-specific IgG) as a conjugate, which can bind the anti-P. insidiosum immunoglobulins from humans and other animal species. Recently, Intaramat et al. have developed the protein A/G-based ICT, as a rapid and user-friendly assay for diagnosis of pythiosis in humans and animals [16]. They performed a side-by-side performance comparison of their ICT [16] and the species-specific ELISA [11,13,15] and showed that ICT has equivalent specificity but relatively-lower  Table 1) sensitivity. One major obstacle for clinical use of ICT is the assay production, which is very complicated and requires a skilled developer and some special reagents that are not available in general clinical laboratories. As opposed to ICT, although ELISA is a multi-step assay and has a longer turnaround time, the production of ELISA is more feasible and requires widely-available reagents. In conclusion, protein A/G-based ELISA has been successfully developed for immunodiagnosis of both human and animal pythiosis. It is a versatile, feasible-to-develop, and functional immunodiagnostic assay.

Limitations
The newly-developed protein A/G-based ELISA overcame the limitation of the previously-established ELISAs [10][11][12][13][14][15], as it can detect the serum anti-P. insidiosum antibodies in multiple host species, including the most affected hosts: humans, horses, and dogs. Other investigators could adopt our protocol to develop their own in-house ELISA. However, there are several limitations in association with the development and performance evaluation of the assay that need to be addressed: 1. The P. insidiosum strain Pi-S was used to prepare the coating antigen. Although this strain has been studied extensively [21,22,27], it is not available in the ATCC and CBS-KNAW culture collection centers. Alternatively, any other P. insidiosum strains could be used to prepare the coating antigen as well. 2. In the evaluation of our protein A/G-based ELISA, only pythiosis serum samples from 4 horses and a dog were tested, due to the low prevalence of the disease in animals in Thailand. 3. The current study focuses on reporting an optimized protocol for the development of protein A/G-based ELISA. Thus, the diagnostic performance (i.e., detection sensitivity and specificity) of the assay was not extensively investigated, using a variety of serum samples from humans and animals with pythiosis and other diseases. 4. Like the previously-established ELISAs, one extra step to perform the assay is to make a serum dilution (our assay requires 1:1600 dilution). This step is necessary for reducing the background ELISA signals from the bindings of non-specific antibodies and P. insidiosum's crude protein extract. In the future, if a purified P. insidiosum-specific protein is available for coating the ELISA plate, then an undiluted serum could be directly tested. 5. Pythiosis patients usually present with a chronic infection, as the mean duration of symptoms before receiving medical care is several months [3]. Direct detection of P. insidiosum is an indicator of an active infection. Isolation of P. insidiosum from a clinical specimen is difficult and challenging. Antibody detection is an alternative method, providing indirect diagnostic evidence for the P. insidiosum infection. Like the other established immunodiagnostic tests for pythiosis, the protein A/G-based ELISA cannot selectively detect an acute-phase infectious marker, such as IgM. Therefore, it cannot directly differentiate between acute, chronic, and previous P. insidiosum infections. However, in a patient with the typical features of pythiosis, an elevated or high level of the serum anti-P. insidiosum antibodies detected by ELISA suggests an active infection.