Experimental ETEC infection study

The present study is based on data and specimens collected during an already completed human volunteer experimental infection study performed in the Division for Infectious Diseases ward at the Department of Medicine at Haukeland University Hospital, Bergen, Norway, between 2014 and 2018 [27]. In that study, 21 healthy adult volunteers aged between 19 and 28 years were experimentally infected with ETEC strain TW10722 (O115:H5; GenBank BioProject: PRJNA59745) [30], by receiving doses between 1 × 106 to 1 × 1010 colony forming units (CFU) and followed for up to 10 days with collection of symptoms data and specimens. The infection was cleared 5 days after dose ingestion by treatment with ciprofloxacin, or earlier if a volunteer experienced severe diarrhea or moderate diarrhea lasting for ≥ 24 h. ETEC strain TW10722 encodes ETEC colonization factors Coli Surface antigen 5 (CS5) and CS6, the human variant of the heat-stable enterotoxin (STh) [30], as well as YghJ.

ALS preparation

ALS sampling was done a few hours prior to dose ingestion (Day 0) and 7 days after, at which time the recirculating blood plasmablast concentrations are usually peaking. ALS was derived from cultures of peripheral blood mononuclear cells (PBMCs) isolated from venous blood collected in BD Vacutainer CPT Mononuclear Cell Preparation Tubes (BD Biosciences, Franklin Lakes, NJ). After centrifuging at 1800 × g for 20 min at room temp, the PBMC-containing fraction was extracted by pipetting and added to 15 mL Phosphate Buffered Saline ([PBS] 10 mM Na2HPO4, 1.8 mM KH2PO4, 2.7 mM KCl, 137 mM NaCl, pH 7.4), centrifuged at 300 × g for 15 min, and washed once with 15 mL PBS before re-centrifugation and resuspending the resulting pellet in 10 mL complete RPMI 1640 medium (Thermo Fisher Scientific, Waltham, MA).

The lymphocyte concentration in this cell suspension was determined by incubating 10 µL of the cell suspension with CD45-FITC (BD Biosciences) to stain all leukocytes, and CD14-PC5 (Beckman Coulter, Brea, CA) to stain all monocytes. After adding AccuCheck Counting Beads (BD Biosciences), the concentration of lymphocyte cells, identified CD45-FITC-positive and CD14-PC5-negative cells was determined by analysis on a C6 Accuri flow cytometer (BD Biosciences). The cell suspension was subsequently re-centrifuged and the resulting pellet was resuspended to a concentration of 1 × 107 lymphocytes/mL in complete RPMI 1640 medium containing 10% heat-inactivated fetal bovine serum, 2 mM L-glutamine, and 50 µg/mL gentamicin sulfate. One mL was incubated for 72 h in a 24-well microplate without antigenic stimulation in an atmosphere of 5% CO2, humidified at 37°C. The resulting supernatant, representing ALS, was stored frozen at – 80 °C until analyses.

Saliva sample collection and preparation

Saliva specimens were collected on the day of dose ingestion and 10 days after. Whole saliva was collected for the first 9 volunteers, while sublingual saliva was collected for the last 12 because a new study suggested that IgA antibodies in sublingual saliva provide a better reflection of the IgA secreted in the gut in response to an ETEC infection [14]. To collect the whole saliva, volunteers placed Salimetrics Oral Swabs (Salimetrics LLC, State College, PA) under the tongue for up to 15 min without salivary stimulation. Collection of sublingual saliva was done the same way, except that the parotid ducts were first blocked by using dental cotton rolls. The saturated swabs were then placed in Salimetrics Swab Storage Tubes (Salimetrics) and centrifuged at 3000 × g at 4 °C for 10 min. The supernatants were frozen at -80 °C after total IgA concentration determination, which was done by using the Human IgA Flex Set Kit (BD Biosciences) following the manufacturer’s protocol. The total IgA concentration estimates were later used to normalize ETEC-specific saliva anti-YghJ IgA level estimates to enable comparison of responses between volunteers and between timepoints [14, 31].

Before analyses, thawed specimens were centrifuged at 16,000 × g for 3 min, and to minimize background assay signals, any remaining mucoid aggregates were removed from the resulting supernatant by performing size-exclusion chromatography. For this procedure, Sephadex® G-25 Superfine gel filtration media (Cytiva, GmbH, Germany) was allowed to swell overnight in PBS, and 300 µL of the resulting media slurry was added to each well of a Nunc™ 278011 fritted 96-well DeepWell Filter Plates (Thermo Fisher Scientific). The plate was centrifuged at 300 × g for 1 min before washing by adding 80 µL PBS to the top of each column and re-centrifuging. Finally, 80 µL saliva supernatant was added to the top of the column, and after centrifuging the plate, the flow-through was diluted 1:8 in PBS, and immediately used in the bead-based immunological assays described below.

Intestinal lavage and serum collection and preparation

The collection and treatment of serum and intestinal lavage specimens has been described previously [28]. Briefly, venous blood was collected on the day before and 10 days after dose ingestion, and serum was obtained by centrifugation. The intestinal lavage specimens were collected around 2 weeks prior to dose ingestion and 10 days after. For these specimens, the volunteers drank a polyethylene glycol (PEG)-based laxative until they excreted clear, watery stools. Protease inhibitors were added to the stools before they were centrifuged, sterile filtered, and stored at – 80 °C.

YghJ production

The current study is based on the same glycosylated and non-glycosylated YghJ (gYghJ and nYghJ, respectively) preparations that were also used in a previous study [28]. In brief, to produce the glycosylated version of the YghJ, a 3xFLAG reporter gene was added to the YghJ gene in wild-type TW10722 to facilitate purification. For the non-glycosylated version, the 3xFLAG tagged YghJ gene was cloned into an expression vector and expressed in E. coli strain MG1655ΔhldE, which lacks the ability to synthesize the heptose glycans that are needed for E. coli protein glycosylation [32]. Both proteins were purified by using affinity chromatography and dialysis. The quality of the YghJ was tested by SDS-PAGE and quantified by using the Bicinchoninic Acid (BCA) assay [33]. The glycosylation patterns of gYghJ were characterized by using the beta-elimination of O-linked glycans followed by Michael-addition of a phosphonic acid derivative (BEMAP) analysis [28].

Anti-YghJ IgA assay

Anti-YghJ IgA levels were estimated by using a bead-based flow-cytometry assay as previously described [28]. Purified gYghJ and nYghJ were independently covalently coupled to 4 µm Cyto-Plex (Thermo Fisher Scientific) or MagPlex carboxylated beads (Luminex Corp, Austin, TX) as described earlier [28]. Cyto-Plex beads were used to analyze the ALS specimens, while the more expensive MagPlex beads were used for saliva analyses because initial tests indicated that these beads gave lower background signals when analyzing saliva. Specimens from different timepoints belonging to the same volunteer were analyzed on the same microplate. Around 5000 each of gYghJ and nYghJ-coupled beads were used for all assays.

Prior to analyses, ALS specimens were thawed and diluted 1:5 in Assay buffer (PBS containing 1% bovine serum albumin [BSA], and 0.05% Tween-20). Wells of MultiScreen HTS HV filter plates (Merck KGaA, Darmstadt, Germany) were wetted with 50 µL Assay buffer, followed centrifugation at 300 × g for 45 sec immediately before adding 50 µL Assay buffer containing nYghJ and gYghJ-coupled beads and 50 µL diluted ALS specimen. After incubating the plates for 30 min at room temperature on a microplate shaker at 600 rpm at room temperature, the filter plates were centrifuged at 50 × g for 45 sec. Later, the beads were washed twice by adding 200 µL Assay buffer followed by re-centrifugation. After a final 30 min incubation with 50 µL Assay buffer containing 1:400 diluted Alexa Fluor 488 AffiniPure Goat Anti-Human Serum IgA antibody (Jackson ImmunoResearch, West Grove, PA), the beads were washed, as described above, twice with 200 µL Assay buffer. Beads were resuspended in 200 µL Assay buffer and analysed by flow cytometry on a LSR Fortessa flow cytometer (BD Biosciences).

As a measure of anti-YghJ IgA levels, we calculated the median fluorescence intensity (MFI) of the gYghJ-coupled beads by using FlowJo, version 10.4.2 (BD Biosciences). Each flow-cytometry analysis included a tenfold dilution series (in duplicate) of a specimen that had high anti-YghJ IgA levels, and the mean of resulting MFI standard curves generated from those dilution series were used for normalizing the estimated MFI, with the maximum set to 10,000 arbitrary units (AUs). The nYghJ-coupled beads were included in the assay merely for quality control purposes, and data from these beads were not included in any of the data analyses. To estimate the fold change in anti-YghJ antibody levels between before and after dose ingestion, we divided the estimated post-infection IgA levels with the corresponding pre-infection levels. Volunteers who had a ≥ 2.0-fold increase were considered to be responders.

Saliva specimens were analysed and responders were defined as explained for ALS, above, except that the MagPlex beads were first suspended in 20 µL PBS containing 5% BSA and 0.25% Tween-20 in a 96-well Bio-Plex Pro™ Flat Bottom Plate (Bio-Rad, Hercules, CA), followed by adding 80 µL of the diluted saliva, and that the washes were done by pipet-aspiration of liquid after immobilizing the beads on a Luminex Magnetic Plate Separator (Luminex). After calculating the AU, the estimates were further normalized by dividing with the total IgA concentration for the given specimen, and the resulting normalized AU (nAU) estimates were used in data analyses.

Glycosylation-specific proportion assay

A glycosylation-specific proportion (GSP) assay was performed to evaluate the degree to which anti-YghJ antibody responses specifically targeted glycosylated epitopes [28]. In this competition-based assay, an excess of unbound gYghJ or nYghJ is added to the specimens prior to performing the above-described anti-YghJ IgA assay in order to selectively hinder different types of anti-YghJ IgA in the specimens to bind to the bead-coupled YghJ. Unbound gYghJ exposes both glycosylated and non-glycosylated epitopes, while nYghJ only exposes non-glycosylated epitopes [28], so treating with an excess of gYghJ in effect removes all YghJ-specific antibodies from the specimen, while treating with nYghJ only removes antibodies that bind to non-glycosylated YghJ epitopes, leaving antibodies that bind to glycosylated YghJ epitopes free to react with the bead-bound YghJ. Subtracting the IgA levels after gYghJ treatment (background level) from those after nYghJ-treatment (levels of IgA targeting glycosylated YghJ epitopes) and untreated specimens (levels of IgA targeting any YghJ epitopes) and subsequently dividing the latter two thus gives the estimated proportion of all anti-YghJ antibodies that target glycosylated epitopes (i.e. the GSP estimate). When estimating GSP for one isotype (e.g. IgA), other isotypes (e.g. IgG and IgM) may interfere with analyses and may need to be removed before analyses.

The GSP assays were performed as detailed earlier [28], except that when analyzing ALS, we added MW 89,000–98,000 polyvinyl alcohol (Sigma-Aldrich, St. Louis, MO) to the Assay buffer to a 2.0% final concentration to help increase recovery of IgA when removing IgG and IgM.

Statistical analyses

Prism, version 10, (GraphPad Software, San Diego, CA) was used for statistical analyses and for preparing graphs. Wilcoxon signed rank test were used to test for differences in anti-YghJ IgA antibody levels, GSP, and total IgA content between specimens collected from the same volunteers but at different timepoints. Mann–Whitney U test was used to test for differences in anti-YghJ IgA antibody levels between specimens collected from volunteers who developed diarrhea and those who did not, between volunteers who contributed whole saliva to those who contributed sublingual saliva, and in GSP differences between saliva and lavage specimens. To evaluate whether the fold change increase in anti-YghJ IgA correlated between different types of specimens, we calculated Pearson correlation coefficients. Lastly, IBM SPSS (IBM SPSS Statistics, Chicago, IL), version 29, was used to determine agreement between GSPs of different specimens using Lin’s concordance correlation coefficients.

Ethical approval

The study (NCT02870751 at ClinicalTrials.gov) was approved by the Regional Committee for Medical and Health Research Ethics, Health Region West (REC-West; case number 2014/826). All volunteers signed written informed consent to participate in the study.