Conventionally, diagnosis of ALA is confirmed by finding the E. histolytica trophozoites in liver pus aspirate obtained via ultrasound guided percutaneous aspiration biopsy, but the parasites are often absent as most of them are located at the margin on the peripheral of the abscess [20, 21]. Therefore, serodiagnosis is widely adopted for diagnosis of ALA, detecting either amoebic antigens or antibodies from serum samples. Although numerous commercial antigen detection kits have been developed i.e. TechLab E. histolytica II ELISA (TechLab, Blacksburg, VA), Entamoeba CELISA-PATH (Cellabs Pty Ltd., Brookvale, Australia), Optimum S Entamoeba histolytica antigen ELISA (Merlin Diagnostika, Berheim-Hersel, Germany), Triage parasite panel BIOSATE (diagnostic, San Diego, CA), and ProSpecT Entamoeba histolytica microplate assay (REMEL Inc., Lenexa, KS), only TechLab E. histolytica II ELISA was evaluated for the detection of circulating antigen in ALA patient serum samples . Haque et al. reported that the kit detected Gal/GalNAc lectin in the ALA patient serum samples with a sensitivity of 96%. However, reports by Parija and Khairnar  and Zeehaida et al. revealed lower sensitivities at 50% and 8.6%, respectively.
Instead of antigen detection, diagnosis of ALA can be performed by detecting circulating antibodies because hepatic amoebiasis raises a strong humoral response, especially immunoglobulin G . Although a variety of serological assays was reported in diagnosis of ALA based on detecting circulating antibodies i.e. IHA, latex agglutination, indirect immunofluorescence, counter-immunoelectrophoresis, gel diffusion, complement fixation and ELISA; IHA and ELISA are still the preferred choices . ELISA is commonly used as the routine diagnostic assay in diagnosis of ALA because it can be developed for in-house use based on different amoebic antigen preparations such as CSA, excretory-secretory antigens, plasma membrane antigens, purified antigenic proteins or recombinant proteins.
The CSA-based ELISA technique has either been used in routine diagnosis of ALA or field screening of amoebiasis [8, 10, 27, 28]. In the diagnosis of ALA, CSA-based ELISA was reported to be 100% sensitive and > 90% specific [8, 29]. Excretory-secretory antigens have been reported to be useful in diagnosis of ALA, which showed sensitivity of ~80% in Western blot analysis [7, 30]. Another ELISA using amoebic plasma membrane antigen for diagnosis of ALA was reported to be 95% sensitive and 91% specific against anti-amoebic IgG; while detection of anti-amoebic IgM was 91% sensitive and 95% specific .
Nonetheless, crude preparations of amoebic antigens produced variable diagnostic efficacies, which may lead to false positive in the diagnosis of ALA. These antigen preparations require maintenance of E. histolytica cultures, which is costly and tedious. Moreover, the proteins are not well-defined; hence the interpretation of results may vary from batch to batch. In this study, analysis of the CSA protein profile showed that most of the antigenic proteins recognized by the human ALA serum antibodies ranged from 25 kDa to 170 kDa. Lower percentage (9%) polyacrylamide gel revealed that the ~170 kDa, ~100 kDa and ~70 kDa protein bands were well recognized by human ALA serum samples. Here, the ~70 kDa protein with 70.83% sensitivity and 96.83% specificity was selected for cloning and production of recombinant protein. The ~170 kDa protein also revealed almost similar sensitivity and specificity, but was not selected because it was most probably similar to the previously described lectin surface antigen of E. histolytica.
Prior to elucidating the identity of the ~70 kDa protein, it was further separated using 2-DE electrophoresis, a widely used technique in biomarker discoveries . Proteins of CSA were fractionated via OFFGEL approach based on pI separation. This approach is advantageous when serum sample is limited. Multiple individual serum samples can be applied on the selected pI protein fraction containing the protein of interest in NC strip blot format, instead of using the whole gel NC membrane format in the conventional 2D-Western blot.
The ~70 kDa protein in this study was identified as E. histolytica PGM by mass-spectrometry analysis. This protein is reported to be involved in the glycolytic pathway of E. histolytica, which catalyses the inter-conversion of glucose 6-phosphate and glucose 1-phosphate. Previously, PGM was used as the gold standard for differentiation between E. histolytica and E. dispar via isoenzyme electrophoresis together with hexokinase . Interestingly, there is no report on the value of this protein in diagnosis of invasive amoebiasis.
Well-defined antigens such as purified antigenic proteins and recombinant proteins of E. histolytica can overcome some of the setbacks posed by utilization of crude soluble preparations. For instance, purified lectin antigen and purified 29 kDa cysteine-rich surface protein in an ELISA format were reported to be useful for diagnosis of ALA [34, 35]. According to Flores et al. (1993), the purified native 29 kDa cysteine-rich surface protein showed 79% sensitivity and 98% specificity in the diagnosis of ALA . Several recombinant proteins have been produced such as SREHP, Gal/GalNAc-specific lectin, 29 kDa cysteine-rich surface protein and 40 kDa NADP+-dependent alcohol dehydrogenase (EhADH1) [13, 36]. Recombinant proteins of SREHP, Gal/GalNAc-specific lectin and 29 kDa cysteine-rich surface protein have been evaluated for diagnosis of ALA. SREHP/MBP fusion protein, which was among the first recombinant proteins used in serodiagnosis of ALA was reported to have a sensitivity and specificity of 74% and 55%, respectively. Interestingly, its diagnostic validity increased to 79% and 87%, respectively, after the removal of its MBP component . The validity is almost similar with rPGM protein in the current study, which is 79.17% sensitive and 86.67% specific. Recombinant full-length 170 kDa Gal/GalNAc-specific lectin protein has been produced and its validity in diagnosis of ALA was about 91% sensitivity . Besides, several recombinant fragments of the Gal/GalNAc-specific lectin protein have also been investigated for the diagnosis of ALA. For example, recombinant protein of a truncated immunodominant domain of the Gal/GalNAc-specific lectin protein with 1.85 kbp cDNA insert showed more than 90% for both sensitivity and specificity . Another study using recombinant Gal/GalNAc-specific lectin-derived protein (LC3) showed 100% sensitivity in diagnosis of ALA . A study carried out by Tachibana et al. reported that the 150 kDa recombinant intermediate subunit (IgI) of the Gal/GalNAc-specific lectin protein used in diagnosis of ALA was only 8.6% sensitive. A recombinant 29 kDa cysteine-rich surface protein was also developed and evaluated as its purified form showed moderate validity in diagnosis of ALA. The evaluation of this recombinant protein in the diagnosis of ALA showed increment of sensitivity (90%) compared to its purified native form protein (79%) .
Although several recombinant proteins have been evaluated, they were not specific for diagnosis of ALA because several studies have reported that those recombinant proteins were immunoreactive with serum samples for amoebic dysentery and asymptomatic cases [15, 16, 29, 38, 39]. In the current study, the novel rPGM protein showed diagnostic validity of ~80% sensitivity and ~90% specificity in diagnosis of ALA. The efficacy of rPGM-ELISA was almost comparable to previously reported recombinant proteins such as SREHP, Gal/GalNAc-specific lectin and 29 kDa cysteine-rich surface protein, which showed both sensitivity and specificity ranging from 80% to 100% [16, 34, 39].