Antibody response to enterotoxigenic Bacteroides fragilis of Filipino colorectal cancer patients

Ethical clearances were obtained from the Research Ethics Review Committees (RERCs) of the hospital study sites – the Mariano Marcos Memorial Hospital and Medical Center in Ilocos Norte (MMMH-RERC-17-001) and the University of Santo Tomas Hospital (IRB-2016-11-191-IS-A1/CR2) in Manila. Written informed consent was obtained from all the participants.

In this case-control study, patients born to Filipino parents who were histologically confirmed CRC patients seen at University of Santo Tomas Hospital (USTH) and Mariano Marcos Memorial Hospital and Medical Center (MMMH-MC) between April 2018 and March 2019 were recruited and enrolled as cases. There were initially 72 potential CRC participants but only 39 were enrolled as cases since they were the only ones who have not yet undergone any form of treatment for their cancer during recruitment, their primary cancer is CRC, and they had no history of bowel disorder such as IBD, polyposis syndromes, and Lynch syndrome. Each of the cases was age- (±2 years) and sex-matched with physician-assessed clinically healthy controls from the same locality (Figure 1). The controls were physically assessed to be free from any form of malignancies. Clinical data of the cases were retrieved from histopathology reports and medical records. This prospective study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) case-control reporting guidelines [12].

Blood collected in Ethylenediaminetetraacetic acid (EDTA) tubes from the enrolled participants was immediately centrifuged at 2,500 rpm for 15 min. Separated plasma samples were stored at −20°C until serologic analysis. All plasma samples were coded to avoid bias in the analysis and to maintain the confidentiality of the identity of the study participants.

The enterotoxigenic B. fragilis (ETBF 86-5443-2-2) culture used in this study was generously given by Dr. Cynthia Sears of John Hopkins University. Following the collection of 48 h ETBF culture grown anaerobically in brain heart infusion (BHI) agar, total protein was extracted using ReadyPrep^TM Protein Extraction Kit (Bio-Rad Laboratories) following the manufacturer's protocol with slight modification. The super-natant was collected, and protein concentration was determined using bicinchoninic acid (BCA) assay (QuantiPro^TM BCA Assay Kit, Sigma-Aldrich). For the culture broth, BHI culture broth was obtained from 48 h culture, which was centrifuged to separate the bacteria from the liquid media.

The chessboard titration method was initially deployed to determine the optimum working conditions. ETBF lysate and culture broth were diluted with bicarbonate/carbonate coating buffer (pH 9.6) to concentrations of 80 μg/mL, 50 μg/mL, 25 μg/mL, 10 μg/mL, and 5 μg/mL. Pooled plasma samples were diluted to 1:50, 1:100, 1:500, and 1:1,000 and the horseradish peroxidase (HRP)-conjugated anti-IgG and anti-IgA with a concentration of 1.2 mg/mL (Abcam) were diluted to 1:500 and 1:1,000. ELISA of the pooled samples from CRC patients and healthy controls was performed using the ETBF lysate coating concentration (micrograms per milliliter), plasma sample dilution, and HRP-conjugated secondary antibody dilution, which were simultaneously tested. The optimum working conditions were decided based on the combined lowest concentrations of the different components that gave the strongest signal versus low background.

Optimized indirect ELISA conditions identified after deployment of the chessboard titration method were used to determine the IgG and IgA antibody levels of CRC patients and their case-matched clinically healthy controls against ETBF. The 96-well plates were coated with 50 μL of 5 μg/mL ETBF lysate, incubated overnight, and blocked with bovine serum albumin (BSA) for 1 h. This was followed by addition of 50 μL of 1:100 diluted plasma of CRC patients and their case-matched controls. Following 1 h of incubation, HRP-conjugated goat anti-human IgG or IgA antibodies (50 μL; 1:1,000) (Abcam) were dispensed in each well and incubated for 1 h. Visualization was performed by adding 50 μL of tetramethyl benzidine (TMB) (TMB Substrate Set, BioLegend), followed by reading the absorbance at 450 nm (OD_{450 nm}) (SPECTROstar Nano, BMG Labtech). The same procedure was performed using the ETBF culture broth.

The COVs of anti-ETBF IgG or IgA antibody levels of the control samples were used in determining positive reactions among the CRC cases. The COVs were computed based on the formula: x¯+SDt1+1/n {\rm{\bar x}}\; + \;{\rm{SD}}\;{\rm{t}}\sqrt {1 + \left( {1/n} \right)} [13], where SD is standard deviation.

Samples with absorbance values above the respective COV were considered reactive or positive while all values below COV were considered non-reactive or to have negative reactivity with the ETBF lysate. The plasma reactivity of cases and controls in terms of anti-ETBF IgG and IgA antibody titers was also compared using an independent t test. The association of antibody levels with tumor grade and tumor stage was ascertained using an independent sample t test, in which cases with missing data were not included. Stata 14 was used to analyze the data and P < 0.05 was considered significant.

A total of 39 histologically confirmed preoperative/pretreatment CRC cases and 39 age- and sex-matched clinically assessed healthy controls seen at the University of Santo Tomas Hospital (Manila) and Mariano Marcos Memorial Hospital and Medical Center (Ilocos Norte, Philippines) between April 2018 and March 2019 were enrolled in this study.

The majority of the CRC cases were males (n = 22, 56%) who were aged more than 50 years with moderately and poorly differentiated (n = 17, 44%) tumors, and were in an advanced stage (III/IV) of the disease (n = 21, 54%) (Table 1).

The optimum working conditions were based on the lowest concentrations of the different components that gave the strongest signal versus low background. In this study, the optimum conditions were 5 μg/mL for the coating protein, 1:100 plasma dilution, and 1:1,000 detection antibody dilution (Figure 2E).

COVs for ETBF lysate were 0.074 and 0.077 for IgG and IgA antibodies, respectively. Using culture broths, COVs for IgG and IgA antibodies were ascertained at 0.079 and 0.074, respectively.

Using ETBF lysate as coating antigen, it was ascertained that 38 (38/39, 97%) of the CRC cases had anti-ETBF IgG antibodies above the cut-off, while all (39/39, 100%) of the cases had anti-ETBF IgA antibodies above the cut-off. Likewise, all (39/39, 100%) cases tested positive for anti-ETBF IgG and IgA antibodies to ETBF culture broths as coating antigen. For the clinically matched controls, 36/39 or 92% had anti-ETBF IgG and IgA antibodies above the cut-off when bacterial lysates were used. All the cases and controls (39/39, 100%) were positive for anti-ETBF IgA antibodies when ETBF culture broth was used as coating antigen. Statistical analysis showed no significant difference on the number of CRC cases who tested positive for anti-ETBF IgG (lysate: P = 0.615; broth: P = 0.494) and IgA (P = 0.240) antibodies compared to their matched controls in both bacterial lysates and culture broths (Figure 3).

Figure 2.

Figure 3.

CRC cases with well-differentiated tumors (11/28, 39%) and moderately to poorly differentiated tumors (17/28, 61%) had higher IgA than IgG levels against ETBF. Similarly, cases at the early (8/29, 28%) and advanced tumor stage (21/29, 72%) had greater anti-ETBF IgA than IgG levels. When categories within tumor grade were analyzed, the result showed that the mean IgG and IgA levels of those with moderately and poorly differentiated tumor were not significantly higher than those of the cases with well-differentiated tumor (IgG: 0.3876 ± 0.1632 vs. 0.3042 ± 0.1556, P = 0.180; IgA: 0.4139 ± 0.1719 vs. 0.3887 ± 0.1410, P = 0.678). There were no significant differences in the antibody levels observed in CRC cases at the early and advanced stages of malignancy (IgG: 0.4374 ± 0.1246 vs. 0.3244 ± 0.1667, P = 0.095; IgA: 0.4538 ± 0.1665 vs. 0.4313 ± 0.1576, P = 0.738) (Figure 4).

Figure 4.

Figure 5.

The mean IgA antibody levels of cases, whether having well, moderately, or poorly differentiated tumor, at the early or advanced tumor stage, were higher than their IgG levels against ETBF. T test analysis showed that the mean anti-ETBF IgG and IgA levels of CRC cases with well-differentiated tumors were not significantly different from the mean antibody levels of the cases with moderately and poorly differentiated tumors (IgG: 0.2932 ± 0.1191 vs. 0.2671 ± 0.1200, P = 0.571; IgA: 0.6238 ± 0.1907 vs. 0.7448 ± 0.1938, P = 0.109). Likewise, the mean anti-ETBF IgG and IgA levels of early-stage CRC were not significantly higher than those of advanced stage CRC (IgG: 0.3320 ± 0.0760 vs. 0.3034 ± 0.1191, P = 0.618; IgA: 0.7448 ± 0.1930 vs. 0.7044 ± 0.1965, P = 0.624) (Figure 5).