Abstract
Enterotoxigenic Escherichia coli (ETEC) is a major cause of diarrheal illness in the developing world. Enterotoxigenic E coli vaccinology has been challenged by genetic diversity and heterogeneity of canonical antigens. Examination of the antigenic breadth of immune responses associated with protective immunity could afford new avenues for vaccine development.
Antibody lymphocyte supernatants (ALS) and sera from 20 naive human volunteers challenged with ETEC strain H10407 and from 10 volunteers rechallenged 4–6 weeks later with the same strain (9 of whom were completely protected on rechallenge) were tested against ETEC proteome microarrays containing 957 antigens.
Enterotoxigenic E coli challenge stimulated robust serum and mucosal (ALS) responses to canonical vaccine antigens (CFA/I, and the B subunit of LT) as well as a small number of antigens not presently targeted in ETEC vaccines. These included pathovar-specific secreted proteins (EtpA, EatA) as well as highly conserved E coli antigens including YghJ, flagellin, and pertactin-like autotransporter proteins, all of which have previously afforded protection against ETEC infection in preclinical studies.
Taken together, studies reported here suggest that immune responses after ETEC infection involve traditional vaccine targets as well as a select number of more recently identified protein antigens that could offer additional avenues for vaccine development for these pathogens.
Enterotoxigenic Escherichia coli (ETEC) is an exceedingly common cause of diarrheal illness with hundreds of millions of infections estimated annually. Much of the burden and mortality from these common infections occurs disproportionately among young children in resource-poor regions. Multiple studies have recently reaffirmed the importance of ETEC as an important cause of moderate-to-severe diarrhea and indicated that afflicted children are more likely to have poor health outcomes [1]. Likewise, ETEC remains an important cause of severe diarrheal illness and death in older individuals [2] and the most common cause of diarrhea in travelers to endemic regions. Although the death rate from diarrheal diseases has declined, due in part to the availability of oral rehydration therapy, it has become increasingly apparent that ETEC and other enteric pathogens are linked to postdiarrheal sequelae in young children including malnutrition, growth stunting, and impaired cognitive development [3–5], greatly compounding the impact of these infections.
In the classic paradigm of ETEC pathogenesis plasmid-encoded colonization factor (CF) or coli surface, antigens mediate colonization of the small intestine [6]. Here, intimate association with small intestinal enterocytes facilitates effective delivery of heat-stable enterotoxins (ST) and/or heat-labile enterotoxins (LT) to their respective epithelial receptors. These toxins alter salt and water transport resulting in net fluid losses into the intestinal lumen and ensuing watery diarrheal illness.
At present, there is no licensed vaccine for ETEC based on this classic paradigm. One feature of ETEC that has confounded development of a broadly protective vaccine based on the canonical antigens, namely CFs and LT, is the inherent genetic plasticity of E coli, and with the advent of large-scale bacterial genome sequencing, it has become increasingly apparent that ETEC comprises a highly variable E coli pathovar [7, 8].
To date, ETEC molecular pathogenesis, immunology [9], and vaccinology efforts have focused on a relatively narrow subset of classic antigens. However, emerging data suggest that both the microbial pathogenesis of these organisms and the immune response to ETEC [10] may be significantly more complex than had been appreciated, thereby affording additional antigens to target in future efforts to develop a more comprehensively protective vaccine [11]. The present studies incorporate whole ETEC genomes in the design of ETEC proteome microarrays to encompass both canonical virulence factors as well as novel antigens in an attempt to provide a more holistic examination of immune responses associated with protection.
A controlled human infection model (CHIM), in which volunteers are challenged with virulent wild-type strains of ETEC, has been used for decades to investigate pathogenesis [12], to decipher immune responses after infection [9], and to assess candidate antigens and vaccines [13]. The ETEC H10407 strain, originally isolated from a case of severe cholera-like watery diarrhea in Bangladesh [14], is the most commonly used ETEC challenge strain. Prior CHIM studies with H10407 have shown that homologous rechallenge with this strain typically results in robust protection against symptomatic infection [15]; however, the precise mechanism of the protection afforded by initial ETEC exposure is unknown. The present studies were undertaken to comprehensively assess the adaptive immune response to ETEC infection that could provide immunologic benchmarks of protection that inform future vaccinology efforts.
METHODS
Comparative Genomics of Enterotoxigenic Escherichia coli Isolates
Genes were selected from the sequenced genomes of 3 parental ETEC isolates (WS_1858B, WS_2773E, and WS3504D) that were used in the construction of a live-attenuated ETEC vaccine candidate, ACE527 [16], as well as the genome of E coli H10407 [7], a prototype ETEC strain that has been examined in several human clinical trials (Supplementary Table 1). The genome content of the 4 isolates was compared using the Large-Scale BLAST Score Ratio Analysis [17], and encoded products that were common in all 4 isolates, as well as having a signal for potential secretion to the surface, were identified. Three algorithms (PSORT [18], TMHMM [19], and SignalP [20]) that were used to identify potential surface molecules identified 800 antigens in H10407. An additional 157 antigens present in 1 or more of the 3 isolates of the ACE527 vaccine candidate were lacking in the H10407 genome. Gene identifiers, deoxyribonucleic acid sequences, and the isolates used as the template for isolation are included in Supplementary Data s1. Included among the informatically selected features were a number of known ETEC antigens. These included the heat-labile toxin (subunits A and B), CF antigens for CFA/I and CS3, the serine protease EatA [21], the EtpA adhesin [22, 23], and the metalloprotease YghJ [24], in addition to conserved and serotype-specific regions of flagellin (FliC) molecules represented in the different strains.
Protein Microarrays
Microarray Construction
Genes encoding potential surface proteins were amplified by polymerase chain reaction (PCR) and cloned by high-throughput PCR recombination cloning into a linearized complementary pXI T7 expression plasmid vector as previously described [25]. Proteins were expressed in a cell-free Rapid Translation System kit (BiotechRabbit, Hennigsdorf, Germany) by in vitro transcription—translation (IVTT). Each expressed protein includes a 5’ polyhistidine (HIS) epitope and 3’ hemagglutinin (HA) epitope. Recombinant proteins including EtpA [26], the passenger domain of EatA [21], YghJ [24], antigen 43 [27], EaeH [28], and FliC [29] subunits were produced at Washington University (St. Louis, MO) as HIS-tagged proteins and purified by metal affinity chromatography as previously described. Both IVTT and recombinant proteins were printed onto nitrocellulose-coated glass AVID slides (Grace Bio-Labs, Inc., Bend, OR) using an Omni Grid Accent robotic microarray printer (Digilabs, Inc., Marlborough, MA). Each slide contained 3 nitrocellulose “pads” for which the full array was printed in replicate, allowing 3 samples to be probed per slide. Microarray chip printing and IVTT protein expression were quality checked by probing random slides with anti-HIS and anti-HA monoclonal antibodies with fluorescent labeling.
Microarray Antigen Content
The array comprises IVTT-expressed proteins (n = 957) and purified recombinant proteins representing known ETEC antigens (n = 9). Also included on the array are IVTT control spots (n = 28), positive control spots for immunoglobulin (Ig)G secondary antibody (n = 16), positive controls for human IgG (n = 16), positive control spots for IgA secondary antibody (n = 16), and positive controls for human IgA (n = 16) (Supplementary Data s1).
Validation of Array Results by Enzyme-Linked Immunosorbent Assay
Recombinant proteins were used to coat 384-well microtiter plates (product no. 3540; Corning) as previously described. Serum samples were diluted 1:1000 in phosphate-buffered saline (PBS) containing 0.05% Tween-20 and 1% bovine serum albumin (PBST-B), whereas antibody lymphocyte supernatant (ALS) samples were added without dilution to microtiter wells and incubated for 1 hour at room temperature. After washing with PBS, secondary antibodies (goat antihuman IgG [H+L] horseradish peroxidase [HRP] conjugate [catalog no. 31410; Thermo Fisher Scientific]), diluted 1:5000 in PBST-B, or goat antihuman IgA/heavy chain HRP conjugate (ThermoFisher 31417), diluted 1:2000 in PBST-B, were added to the wells and incubated for 1 hour at room temperature, washed, then developed with 3,3’,5,5’ tetramethybenzidine substrate, read kinetically at 650 nm (Eon; BioTek, Winooski, VT), and the data were reported as Vmax (milliunits/minute).
Enzyme-Linked Immunosorbent Assay Controlled Human Infection Studies
Samples used in the present studies were derived from an earlier ETEC human volunteer challenge study at Johns Hopkins University [15]. The challenge cohort analyzed in the present study included 20 adult volunteers. After fasting overnight, 10 of the volunteers were ETEC-naive and received a single dose of 2 × 107 colony-forming units (CFU) of ETEC strain H10407 suspended in bicarbonate buffer, whereas the remaining 10 volunteers were challenged in an identical fashion but then rechallenged 2 to 3 months later with 2 × 107 CFU of strain H10407 with bicarbonate, again after an overnight fast.
Enzyme-Linked Immunosorbent Assay Strain Validation
The ETEC H10407 challenge strain used in these studies belongs to serotype O78:H11, expresses the CFA/I CF, both the STh and STp heat-stable toxins, heat-labile toxin, the EtpA adhesin [30], and the EatA mucinase [31] in addition to YghJ, a metalloprotease secreted by the type 2 secretion system (T2SS) [24, 32] that also secretes LT [33]. The clinical challenge strain used in these studies was derived from cGMP Batch Production Record (no. 285-000, Lot 0519, date of manufacture February 26, 1998) at the WRAIR Pilot Bioproduction Facility. A lyophilized aliquot of this lot was shipped to Washington University School of Medicine (St. Louis, MO), where frozen glycerol stocks were prepared and maintained at −80°C for use in subsequent studies. After overnight culture of the strain in Lysogeny broth, at 37°C, motility was confirmed in soft agar (1% tryptone, 0.7% NaCl, 0.35% agar), as previously described [22]. Retention of genes encoding the plasmid encoded antigens including STp, STh, LT, EtpA, and EatA were verified by PCR using the primers listed in Supplementary Table 2. Immunoblotting was then performed as previously described [34] to verify production of the secreted proteins EtpA, EatA, and Ygh by the challenge strain, and the production of LT was confirmed by GM1 ganglioside-binding enzyme-linked immunosorbent assay (ELISA) [35].
Production of Subcellular Fractions of Enterotoxigenic Escherichia coli
To further characterize the proteome of the H10407 challenge strain, we produced subcellular fractions that included outer membrane proteins (OMPs), outer membrane vesicles (OMV), and concentrated culture supernatants, which were then analyzed by tandem mass spectrometry (MS/MS).
Construction of a Type 2 Secretion System Mutant
To specifically examine the relative abundance of proteins secreted by the T2SS encoded on the chromosome of H10407, we constructed an isogenic strain bearing a mutation in gspG, which encodes the major pseudopilin subunit required for secretion through the T2SS [33] (Supplementary Table 1). The resulting gspG mutant (jf1121) was processed in parallel for MS/MS with the H10407 strain.
Proteomic Analysis of Enterotoxigenic Escherichia coli Subcellular Fractions
Subcellular fractions were examined by MS/MS at the Washington University Proteomics Shared Resource. The previously sequenced H10407 genome [7] was used to provide predicted peptide sequences against which to query mass spectrometry data. Additional detailed methods can be found in the Supplementary Data.
RESULTS
Validation of the H10407 Challenge Strain Proteome
Because recent studies have suggested that ETEC may present a range of antigens that extend beyond classic vaccine targets [10], we first examined subcellular fractions of the challenge strain to identify potential antigens that might be recognized by the host during infection. Using the same growth conditions and the same strain lot of H10407 that were used to prepare the challenge inoculum that was administered to human volunteers, we prepared subcellular fractions consisting of those most likely to contain surface-expressed antigens, including OMPs, OMVs, and secreted proteins, that were then subjected to MS/MS. In this study, we identified the classic target antigens for ETEC vaccines including CFA/I from which we were able to detect the outer membrane usher protein CfaC, and the major pilin subunit CfaB, as well as CfaE, the CFA/I tip adhesin protein in OMV (Supplementary Figure 1) and/or OMP fractions (Figure 1, Supplementary Data File s2, Supplementary Data File s3). Likewise, we detected both subunits of and heat-labile toxin (LT-A and LT-B) in all of the fractions tested. In addition, we also detected a number of antigens that are not presently targeted in vaccine approaches, but which have been associated with protection in animal models including EtpA, EatA, antigen 43, and FliC. To further validate our approach, we compared the wild-type H10407 strain to a strain with a mutation in gspG required for protein export through T2SS, which is known to be responsible for secretion of both the LT [33] and the highly conserved chromosomally encoded YghJ metalloprotease [24, 36]. As predicted, we found that these proteins were either greatly diminished or altogether undetectable in supernatants from the T2SS mutant (Figure 1). As predicted from earlier immunoproteome studies [10], the data presented here suggest that ETEC present a large repertoire of surface-exposed antigens that may provide relevant targets for immune neutralization of effective pathogen-host interaction.
Antigen production by challenge strain H10407. Shown are antigens detected after growth of H10407 under conditions that were used to prepare the challenge strain. Values in columns 1, 2, 4, and 5 reflect protein probability of correct tandem mass spectrometry (MS/MS) identification. Relative peptide abundance (*) values are shown in columns 3 and 6. A subset of the antigens detected is listed in the column at left. Columns 1 and 2 represent proteins identified from wild-type (wt) H10407 outer membrane vesicles (omv) or purified outer membrane proteins (omp) fractions, respectively, whereas column 3 compares the relative peptide abundance of each protein in omv and omp preparations. Columns 4 and column 5 depict the identification of supernatant proteins from the wt and the general secretion pathway (gspG) mutant, respectively, whereas column 6 compares the relative abundance of the peptides in the gsp mutant and wt. The top 5 row features correspond to canonical vaccine antigen targets CFA/I with individual protein subunits CfaE (CFA/I tip adhesin), CfaB (CFA/I major fimbrial subunit), and CfaC (CFA/I outer membrane usher protein), and heat-labile toxin subunits (LT-A, and LT-B). Heat-labile enterotoxin and the other type 2 secretion system (T2SS) effector protein, the YghJ metalloprotease, are grouped together. These are followed by noncanonical putative virulence antigens.
Enterotoxigenic Escherichia coli Challenge Elicits Systemic Humoral and Mucosal Antibody Responses to Classic Vaccine Antigens
Because ETEC vaccine development has primarily centered on CFs and LT, we focused our initial analysis on immune responses to these antigens. After H10407 challenge, 11 of 20 volunteers mounted ≥2-fold increases in serum IgA antibody titer to the B subunit of heat-labile toxin (LT-B) with 7 of 20 demonstrating increased IgG titers in endpoint ELISA assays. Likewise, 12 of 20 and 7 of 20 volunteers exhibited ≥2-fold increases in serum IgA and IgG titers, respectively, to CFA/I (Figure 2, Supplementary Data File s4).
Systemic immunoglobulin (Ig)A and IgG responses to canonical enterotoxigenic Escherichia coli (ETEC) antigens. Serum IgA and IgG endpoint enzyme-linked immunosorbent assay responses to the B subunit of heat-labile enterotoxin (LT-B) and CFA/I fimbriae are shown on day 0 before challenge with ETEC H10407 and on days 7, 10, 28, and 84 after infection. Numbers in the top right corner of each graph indicate the fraction of volunteers with a ≥2-fold increase in titer compared with the day 0 sample. * indicates the peak titer days for volunteer 314 who was the only individual not protected from diarrheal illness on subsequent rechallenge.
After intestinal infection, lymphocytes captured from the peripheral circulation can be cultured in vitro without stimulation, yielding antibodies in the culture supernatant (ALS) that serve as excellent surrogate markers of the intestinal immune response [37]. Therefore, we also interrogated the ALS response to classic antigens on the array. The majority of volunteers (16 of 20) demonstrated at least a 2-fold increase in ALS IgA response to CFA/I after challenge, and 8 of 10 volunteers mounted further increases in their response to this antigen on rechallenge (Figure 3A). Some volunteers (6 of 20) mounted >2-fold increases to LT-B on initial challenge, and 3 additional volunteers who failed to respond initially mounted robust ALS IgA anamnestic responses ranging from 5- to 33-fold of their baseline response on rechallenge (Figure 3B).
![Mucosal immune response to canonical vaccine antigens. Antibody lymphocyte supernatant (immunoglobulin A, antibody lymphocyte supernatants [ALS]) responses of individual subjects (n = 20) to CFA/I (A) and the heat-labile toxin B subunit (LT-B [B]). Subject responses are represented by colored lines (identified by subject number in the key at right) with dashed lines representing those (n = 10) that were challenged once and solid lines representing those (n = 10) who were rechallenged. The x-axis of each graph depicts the days before (day −1) or after (days 7, 10, 28) each challenge, and the y-axis of each graph is the base-2 log of the microarray raw signal intensity. Subject 314 (*) was the only subject not protected on rechallenge.](https://cdn.statically.io/img/oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/jid/218/9/10.1093_infdis_jiy312/1/m_jiy31203.jpeg?Expires=1724520293&Signature=PYiaH1lMdNbhiIhn1mw8OKzFRLugS3sQCSpTu7gXWIJ0laSIfLFsATlZrLeuwto4OpRk6hAPr47pEAZ4kfB7Zc4aeP0-eRrVeTRs0QmfbH60xn3CKbca4mXdJZjn4Wx2CwBpt8aVQcJagYlrBlJgtU8dH9EpzBFQ6HG6HER4Px2y017mXRQqNOeYsDlELMGuW9vQ4OFUxSSD909n49tk7goXNLXvLU6qncFFEShIedRGHIKX-syIVwr0LZv4tOJXh7qo-s2cV7F9LuyK5mhUgNAN3khouhZrggJIu~3Z0pBmbxkqsZwf0Gvx1hSCeVFqmHbFGk6DL~Xa2Gop0As9kg__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Mucosal immune response to canonical vaccine antigens. Antibody lymphocyte supernatant (immunoglobulin A, antibody lymphocyte supernatants [ALS]) responses of individual subjects (n = 20) to CFA/I (A) and the heat-labile toxin B subunit (LT-B [B]). Subject responses are represented by colored lines (identified by subject number in the key at right) with dashed lines representing those (n = 10) that were challenged once and solid lines representing those (n = 10) who were rechallenged. The x-axis of each graph depicts the days before (day −1) or after (days 7, 10, 28) each challenge, and the y-axis of each graph is the base-2 log of the microarray raw signal intensity. Subject 314 (*) was the only subject not protected on rechallenge.
Immune Responses to Noncanonical Novel Antigens
At present, there is no established mechanistic correlate of protection for ETEC infections. A potential advantage of the microarray approach is that it permits an unbiased and holistic assessment of immune responses associated with protection. We found that in addition to the classic vaccine antigens, the majority of volunteers mounted robust ALS responses to a relatively small number of novel antigens that are not currently targeted in ETEC vaccines (Supplementary Figure 2). Although there were more than 900 antigen features included on the arrays, we found that only a small number of these antigens were differentially recognized after infection with H10407. Using stringent criteria for immunoreactivity, only 23 antigens were reactive in more than 10% of subjects. Among the antigens identified in these studies were 3 secreted high molecular weight proteins YghJ, the passenger domain of the EatA autotransporter protein, and the EtpA two-partner secretion exoprotein (Figure 4, Tables 1 and 2).
(A) Antibody lymphocyte supernatant (ALS) responses to recombinant purified proteins. Heatmap depicts results for individual subjects on day –1 before and day 7 after challenge. Signals shown are base-2 logarithm of raw signal. Subject numbers are shown on the x-axis, and the ordering is the same for both sample days. Proteins are ordered with the pathovar-specific noncanonical secreted antigens EtpA and EatA at top, followed by conserved proteins YghJ and EaeH, and the classic vaccine targets LT, and CFA/I in the last 3 rows. (B) The ALS responses to in vitro transcription—translation proteins. Values correspond to normalized array signals, and antigens are grouped by function with secreted antigens EtpA, EatA, and YghJ; followed by flagellar proteins, FliC (full-length H11 flagellin flagellar structural subunit, H111-487, H11 serotype-specific region of, H11174-399), and FlgE the flagellar hook protein; and membrane-associated proteins, ompW, and DedD; and 3 pertactin-like adhesin-autotransporter (AT) proteins including aantigen 43 (ETEC_4662), bETEC_2119, and cETEC_2366. Is the heat-stable precursor protein identified in this screening?
ETEC Microarray ALS IgA (Day –1 to Day 7) Responses to Recombinant Proteins
| Averageb | ||||||
|---|---|---|---|---|---|---|
| Protein | [mg/mL]a | Day –1 | Day 7 | Increase | cP = | Frequency |
| YghJ | 1 × 10–1 | 7.98 | 10.92 | 2.93 | 3.2 × 10–08 | 0.90 |
| CFA/I | 1 × 10–1 | 6.14 | 7.88 | 1.74 | 3.6 × 10–03 | 0.80 |
| EaeH | 1 × 10–1 | 7.89 | 8.75 | 0.86 | 4.3 × 10–07 | 0.70 |
| EtpA | 1 × 10–4 | 10.12 | 12.1 | 1.97 | 3.9 × 10–03 | 0.65 |
| EatA | 1 × 10–1 | 7.95 | 9.12 | 1.17 | 1.8 × 10–03 | 0.50 |
| LT-B | 1 × 10–1 | 7.92 | 8.75 | 0.83 | 8.0 × 10–03 | 0.40 |
| LT-A | 1 × 10–1 | 7.44 | 7.64 | 0.2 | 2.9 × 10–01 | 0.30 |
| Averageb | ||||||
|---|---|---|---|---|---|---|
| Protein | [mg/mL]a | Day –1 | Day 7 | Increase | cP = | Frequency |
| YghJ | 1 × 10–1 | 7.98 | 10.92 | 2.93 | 3.2 × 10–08 | 0.90 |
| CFA/I | 1 × 10–1 | 6.14 | 7.88 | 1.74 | 3.6 × 10–03 | 0.80 |
| EaeH | 1 × 10–1 | 7.89 | 8.75 | 0.86 | 4.3 × 10–07 | 0.70 |
| EtpA | 1 × 10–4 | 10.12 | 12.1 | 1.97 | 3.9 × 10–03 | 0.65 |
| EatA | 1 × 10–1 | 7.95 | 9.12 | 1.17 | 1.8 × 10–03 | 0.50 |
| LT-B | 1 × 10–1 | 7.92 | 8.75 | 0.83 | 8.0 × 10–03 | 0.40 |
| LT-A | 1 × 10–1 | 7.44 | 7.64 | 0.2 | 2.9 × 10–01 | 0.30 |
Abbreviations: ALS, antibody lymphocyte supernatants; ETEC, enterotoxigenic Escherichia coli; IgA, immunoglobulin A.
aProtein printing concentration.
bSignals are in base-2 logarithm of raw signal.
cPaired t test.
NOTE: Frequency = proportion of 20 subjects that had at least a 50% increase in average signal intensity from day −1 to day 7 after initial challenge. Proteins are rank-ordered by the increase in frequency followed by the average increase from day –1 to day 7.
ETEC Microarray ALS IgA (Day –1 to Day 7) Responses to Recombinant Proteins
| Averageb | ||||||
|---|---|---|---|---|---|---|
| Protein | [mg/mL]a | Day –1 | Day 7 | Increase | cP = | Frequency |
| YghJ | 1 × 10–1 | 7.98 | 10.92 | 2.93 | 3.2 × 10–08 | 0.90 |
| CFA/I | 1 × 10–1 | 6.14 | 7.88 | 1.74 | 3.6 × 10–03 | 0.80 |
| EaeH | 1 × 10–1 | 7.89 | 8.75 | 0.86 | 4.3 × 10–07 | 0.70 |
| EtpA | 1 × 10–4 | 10.12 | 12.1 | 1.97 | 3.9 × 10–03 | 0.65 |
| EatA | 1 × 10–1 | 7.95 | 9.12 | 1.17 | 1.8 × 10–03 | 0.50 |
| LT-B | 1 × 10–1 | 7.92 | 8.75 | 0.83 | 8.0 × 10–03 | 0.40 |
| LT-A | 1 × 10–1 | 7.44 | 7.64 | 0.2 | 2.9 × 10–01 | 0.30 |
| Averageb | ||||||
|---|---|---|---|---|---|---|
| Protein | [mg/mL]a | Day –1 | Day 7 | Increase | cP = | Frequency |
| YghJ | 1 × 10–1 | 7.98 | 10.92 | 2.93 | 3.2 × 10–08 | 0.90 |
| CFA/I | 1 × 10–1 | 6.14 | 7.88 | 1.74 | 3.6 × 10–03 | 0.80 |
| EaeH | 1 × 10–1 | 7.89 | 8.75 | 0.86 | 4.3 × 10–07 | 0.70 |
| EtpA | 1 × 10–4 | 10.12 | 12.1 | 1.97 | 3.9 × 10–03 | 0.65 |
| EatA | 1 × 10–1 | 7.95 | 9.12 | 1.17 | 1.8 × 10–03 | 0.50 |
| LT-B | 1 × 10–1 | 7.92 | 8.75 | 0.83 | 8.0 × 10–03 | 0.40 |
| LT-A | 1 × 10–1 | 7.44 | 7.64 | 0.2 | 2.9 × 10–01 | 0.30 |
Abbreviations: ALS, antibody lymphocyte supernatants; ETEC, enterotoxigenic Escherichia coli; IgA, immunoglobulin A.
aProtein printing concentration.
bSignals are in base-2 logarithm of raw signal.
cPaired t test.
NOTE: Frequency = proportion of 20 subjects that had at least a 50% increase in average signal intensity from day −1 to day 7 after initial challenge. Proteins are rank-ordered by the increase in frequency followed by the average increase from day –1 to day 7.
In addition to these antigens, which have been the subject of earlier antigen discovery and characterization efforts, we also found that the majority of volunteers mounted striking responses to FliC, the structural subunit of the flagellar shaft, with very strong responses to the serotype-specific region (amino acids 174–399) of the FliC H11 molecule but also to the highly conserved amino-terminal region of FliC (amino acids 1–174). Likewise, on initial challenge with H10407, the majority of volunteers mounted robust IgA responses to FlgE, the protein comprising the flagellar hook [38] (Figure 4B, Table 2).
Differentially Reactive Antigens, ALS IgA, Day-1 to Day 7 After Initial Challenge
| Averagea | ||||||
|---|---|---|---|---|---|---|
| Protein ID | Description | Day –1 | Day 7 | Increase | bP = | Frequency |
| ETEC_2032 | flagellin (FliC, serotype H11) | 0.367 | 5.79 | 5.42 | 8.0 × 10–14 | 1.00 |
| ETEC_2032174-399 | FliC, H11 serotype-specific region | –0.005 | 3.746 | 3.75 | 4.5 × 10–11 | 0.95 |
| ETEC_3241 | YghJ | –0.110 | 2.01 | 2.12 | 5.4 × 10–07 | 0.85 |
| ETEC_1141 | flagellar hook protein FlgE | –0.073 | 1.84 | 1.92 | 3.3 × 10–05 | 0.70 |
| ETEC_p948_0020 | EatA passenger domain | –0.178 | 1.28 | 1.46 | 7.3 × 10–04 | 0.55 |
| ETEC_4462 | antigen 43 (fluffing protein) | 0.205 | 1.64 | 1.44 | 2.3 × 10–03 | 0.55 |
| ETEC_1358 | outer membrane protein W | 0.463 | 1.07 | 0.60 | 5.2 × 10–04 | 0.50 |
| ETEC_2450 | peptidoglycan-binding protein | 0.483 | 1.00 | 0.52 | 3.0 × 10–04 | 0.45 |
| WS_3504D_00539 | flagellin (FliC, serotype H16c) | 0.28 | 0.91 | 0.63 | 3.7 × 10–04 | 0.40 |
| ETEC_p948_0110 | EtpA | –0.171 | 0.65 | 0.82 | 4.1 × 10–03 | 0.35 |
| WS_2773_FliC1-174 | FliC, conserved region, AA 1-174 | 0.127 | .486 | 0.36 | 1.4 × 10–02 | 0.30 |
| WS_1858B_FliC | flagellin (FliC, serotype H45d) | –0.0160 | 0.397 | 0.41 | 8.3 × 10–03 | 0.25 |
| ETEC_2119 | putative adhesin autotransporter | 0.255 | 0.63 | 0.37 | 4.0 × 10–02 | 0.25 |
| ETEC_p666_0810 | 2-component sensor protein | 0.482 | 0.860 | .38 | 3.2 × 10–04 | 0.25 |
| ETEC_2366 | putative adhesin autotransporter | 0.109 | 0.495 | 0.39 | 1.2 × 10–02 | 0.20 |
| ETEC_4444 | transcriptional activator | –0.063 | 0.217 | 0.28 | 9.7 × 10–03 | 0.20 |
| Averagea | ||||||
|---|---|---|---|---|---|---|
| Protein ID | Description | Day –1 | Day 7 | Increase | bP = | Frequency |
| ETEC_2032 | flagellin (FliC, serotype H11) | 0.367 | 5.79 | 5.42 | 8.0 × 10–14 | 1.00 |
| ETEC_2032174-399 | FliC, H11 serotype-specific region | –0.005 | 3.746 | 3.75 | 4.5 × 10–11 | 0.95 |
| ETEC_3241 | YghJ | –0.110 | 2.01 | 2.12 | 5.4 × 10–07 | 0.85 |
| ETEC_1141 | flagellar hook protein FlgE | –0.073 | 1.84 | 1.92 | 3.3 × 10–05 | 0.70 |
| ETEC_p948_0020 | EatA passenger domain | –0.178 | 1.28 | 1.46 | 7.3 × 10–04 | 0.55 |
| ETEC_4462 | antigen 43 (fluffing protein) | 0.205 | 1.64 | 1.44 | 2.3 × 10–03 | 0.55 |
| ETEC_1358 | outer membrane protein W | 0.463 | 1.07 | 0.60 | 5.2 × 10–04 | 0.50 |
| ETEC_2450 | peptidoglycan-binding protein | 0.483 | 1.00 | 0.52 | 3.0 × 10–04 | 0.45 |
| WS_3504D_00539 | flagellin (FliC, serotype H16c) | 0.28 | 0.91 | 0.63 | 3.7 × 10–04 | 0.40 |
| ETEC_p948_0110 | EtpA | –0.171 | 0.65 | 0.82 | 4.1 × 10–03 | 0.35 |
| WS_2773_FliC1-174 | FliC, conserved region, AA 1-174 | 0.127 | .486 | 0.36 | 1.4 × 10–02 | 0.30 |
| WS_1858B_FliC | flagellin (FliC, serotype H45d) | –0.0160 | 0.397 | 0.41 | 8.3 × 10–03 | 0.25 |
| ETEC_2119 | putative adhesin autotransporter | 0.255 | 0.63 | 0.37 | 4.0 × 10–02 | 0.25 |
| ETEC_p666_0810 | 2-component sensor protein | 0.482 | 0.860 | .38 | 3.2 × 10–04 | 0.25 |
| ETEC_2366 | putative adhesin autotransporter | 0.109 | 0.495 | 0.39 | 1.2 × 10–02 | 0.20 |
| ETEC_4444 | transcriptional activator | –0.063 | 0.217 | 0.28 | 9.7 × 10–03 | 0.20 |
Abbreviations: ALS, antibody lymphocyte supernatants; DNA, deoxyribonucleic acid; ETEC, enterotoxigenic Escherichia coli; IgA, immunoglobulin A; IVTT, in vitro transcription—translation.
aSignals are normalized IVTT protein data.
bPaired t test based on IVTT normalized data. Proteins are rank-ordered by the increase in frequency.
cPredicted by DNA sequence using Serotype Finder 1.1 https://cge.cbs.dtu.dk/services/SerotypeFinder/. Original serotype H12.
dPredicted by DNA sequence, original serotype H−.
Differentially Reactive Antigens, ALS IgA, Day-1 to Day 7 After Initial Challenge
| Averagea | ||||||
|---|---|---|---|---|---|---|
| Protein ID | Description | Day –1 | Day 7 | Increase | bP = | Frequency |
| ETEC_2032 | flagellin (FliC, serotype H11) | 0.367 | 5.79 | 5.42 | 8.0 × 10–14 | 1.00 |
| ETEC_2032174-399 | FliC, H11 serotype-specific region | –0.005 | 3.746 | 3.75 | 4.5 × 10–11 | 0.95 |
| ETEC_3241 | YghJ | –0.110 | 2.01 | 2.12 | 5.4 × 10–07 | 0.85 |
| ETEC_1141 | flagellar hook protein FlgE | –0.073 | 1.84 | 1.92 | 3.3 × 10–05 | 0.70 |
| ETEC_p948_0020 | EatA passenger domain | –0.178 | 1.28 | 1.46 | 7.3 × 10–04 | 0.55 |
| ETEC_4462 | antigen 43 (fluffing protein) | 0.205 | 1.64 | 1.44 | 2.3 × 10–03 | 0.55 |
| ETEC_1358 | outer membrane protein W | 0.463 | 1.07 | 0.60 | 5.2 × 10–04 | 0.50 |
| ETEC_2450 | peptidoglycan-binding protein | 0.483 | 1.00 | 0.52 | 3.0 × 10–04 | 0.45 |
| WS_3504D_00539 | flagellin (FliC, serotype H16c) | 0.28 | 0.91 | 0.63 | 3.7 × 10–04 | 0.40 |
| ETEC_p948_0110 | EtpA | –0.171 | 0.65 | 0.82 | 4.1 × 10–03 | 0.35 |
| WS_2773_FliC1-174 | FliC, conserved region, AA 1-174 | 0.127 | .486 | 0.36 | 1.4 × 10–02 | 0.30 |
| WS_1858B_FliC | flagellin (FliC, serotype H45d) | –0.0160 | 0.397 | 0.41 | 8.3 × 10–03 | 0.25 |
| ETEC_2119 | putative adhesin autotransporter | 0.255 | 0.63 | 0.37 | 4.0 × 10–02 | 0.25 |
| ETEC_p666_0810 | 2-component sensor protein | 0.482 | 0.860 | .38 | 3.2 × 10–04 | 0.25 |
| ETEC_2366 | putative adhesin autotransporter | 0.109 | 0.495 | 0.39 | 1.2 × 10–02 | 0.20 |
| ETEC_4444 | transcriptional activator | –0.063 | 0.217 | 0.28 | 9.7 × 10–03 | 0.20 |
| Averagea | ||||||
|---|---|---|---|---|---|---|
| Protein ID | Description | Day –1 | Day 7 | Increase | bP = | Frequency |
| ETEC_2032 | flagellin (FliC, serotype H11) | 0.367 | 5.79 | 5.42 | 8.0 × 10–14 | 1.00 |
| ETEC_2032174-399 | FliC, H11 serotype-specific region | –0.005 | 3.746 | 3.75 | 4.5 × 10–11 | 0.95 |
| ETEC_3241 | YghJ | –0.110 | 2.01 | 2.12 | 5.4 × 10–07 | 0.85 |
| ETEC_1141 | flagellar hook protein FlgE | –0.073 | 1.84 | 1.92 | 3.3 × 10–05 | 0.70 |
| ETEC_p948_0020 | EatA passenger domain | –0.178 | 1.28 | 1.46 | 7.3 × 10–04 | 0.55 |
| ETEC_4462 | antigen 43 (fluffing protein) | 0.205 | 1.64 | 1.44 | 2.3 × 10–03 | 0.55 |
| ETEC_1358 | outer membrane protein W | 0.463 | 1.07 | 0.60 | 5.2 × 10–04 | 0.50 |
| ETEC_2450 | peptidoglycan-binding protein | 0.483 | 1.00 | 0.52 | 3.0 × 10–04 | 0.45 |
| WS_3504D_00539 | flagellin (FliC, serotype H16c) | 0.28 | 0.91 | 0.63 | 3.7 × 10–04 | 0.40 |
| ETEC_p948_0110 | EtpA | –0.171 | 0.65 | 0.82 | 4.1 × 10–03 | 0.35 |
| WS_2773_FliC1-174 | FliC, conserved region, AA 1-174 | 0.127 | .486 | 0.36 | 1.4 × 10–02 | 0.30 |
| WS_1858B_FliC | flagellin (FliC, serotype H45d) | –0.0160 | 0.397 | 0.41 | 8.3 × 10–03 | 0.25 |
| ETEC_2119 | putative adhesin autotransporter | 0.255 | 0.63 | 0.37 | 4.0 × 10–02 | 0.25 |
| ETEC_p666_0810 | 2-component sensor protein | 0.482 | 0.860 | .38 | 3.2 × 10–04 | 0.25 |
| ETEC_2366 | putative adhesin autotransporter | 0.109 | 0.495 | 0.39 | 1.2 × 10–02 | 0.20 |
| ETEC_4444 | transcriptional activator | –0.063 | 0.217 | 0.28 | 9.7 × 10–03 | 0.20 |
Abbreviations: ALS, antibody lymphocyte supernatants; DNA, deoxyribonucleic acid; ETEC, enterotoxigenic Escherichia coli; IgA, immunoglobulin A; IVTT, in vitro transcription—translation.
aSignals are normalized IVTT protein data.
bPaired t test based on IVTT normalized data. Proteins are rank-ordered by the increase in frequency.
cPredicted by DNA sequence using Serotype Finder 1.1 https://cge.cbs.dtu.dk/services/SerotypeFinder/. Original serotype H12.
dPredicted by DNA sequence, original serotype H−.
Several of the IVTT proteins highlighted by ALS responses after initial infection have recently been shown to be modulated by a single transcriptional regulator CfaD, which regulates production of the CFA/I cfaABCE operon [39]. Included in these CfaD-regulated proteins are 2 adhesin-autotransporter proteins antigen 43 and flu as well as the etpBAC locus responsible for secreting the EtpA adhesin. Similar to EtpA, we observed significant ALS responses to 3 similar adhesin-like chromosomally encoded autotransporter proteins antigen 43 (ETEC_4462), ETEC_2119, and ETEC_2366, each of which contain pertactin-like domains, with antigen 43 (ETEC_4662) and ETEC_2119 sharing a very high degree of homology (Supplementary Figure 3).
For each of the secreted high molecular weight proteins (YghJ, EatA, and EtpA), we observed substantial increases in ALS responses postchallenge, typically peaking at 7 days after infection, with similar responses in the rechallenged volunteers (Figure 5A). Likewise, direct analysis of ALS samples by ELISA demonstrated significant increases in the IgA responses to recombinant versions of EtpA, the exposed passenger domain of EatA, and antigen 43 (Figure 5B).
H10407 challenge elicits robust mucosal responses to noncanonical antigens. (A) Antibody lymphocyte supernatant (IgA, ALS) responses of individual subjects (n = 20) are represented by colored lines (identified by subject number in the key at right) with dashed lines representing those (n = 10) that were challenged once, and solid lines representing those (n = 10) who were rechallenged. The x-axis of each graph depicts the days before (day –1) or after (days 7, 10, 28) each challenge, and the y-axis of each graph is the base-2 log of the microarray raw signal intensity. Graphs from left to right demonstrate ALS responses, the YghJ metalloprotease, the EatA mucinase, and the EtpA adhesin. Respectively, subject 314 (*) was the only subject not protected against diarrhea on rechallenge. (B) Kinetic enzyme-linked immunosorbent assay data demonstrating ALS (IgA) responses to noncanonical proteins after initial challenge with H10407. P values reflect 2-tailed Mann-Whitney nonparametric comparisons. Dashed lines in each group represent geometric means.
Taken together, these data suggest that although the immune response to ETEC extends beyond canonical vaccine targets, a relatively small number of antigens are recognized during the course of infection. Because a number of these antigens, including FliC, EtpA, EatA, and the antigen-43 like autotransporters, have been associated with protection in preclinical models, the present studies may inform future attempts to engineer broadly protective ETEC vaccines.
DISCUSSION
Enterotoxigenic E coli remains a leading cause of diarrheal morbidity and mortality in countries where lack of access to clean water and effective sanitation continue to provide a permissive environment for the dissemination of these pathogens. Although deaths from infectious diarrhea generally are estimated to have decreased significantly within the past decade [40], granular data on causes of death and the morbidity attributable to diarrhea are often the most difficult to ascertain in regions where the impact of these illnesses is likely to be highest [41], and, conservatively, it is estimated that 1 child still dies each minute from infectious diarrhea [42].
An effective ETEC vaccine could prevent deaths from these pathogens and reduce the substantial morbidity associated with diarrheal illnesses as well as their pernicious sequelae [43]. All ETEC vaccine candidates that have advanced to clinical trials to date have exclusively targeted the plasmid-encoded CFs and/or LT [44], and seminal studies in humans clearly support the importance of CFs [45] or their associated subunits [13] as protective antigens. Nevertheless, development of a broadly protective ETEC vaccine has been confounded by the remarkable underlying genetic plasticity of E coli [8] and a still-developing understanding of mechanistic correlates of protection [46].
Emerging data from preclinical antigen discovery and vaccinology studies suggest that more recently discovered antigens could provide molecular targets that complement classic approaches to ETEC vaccines, but there are no prior studies that have comprehensively examined the immune response to ETEC infections. The present studies combine genomic, proteomic, and protein microarray data to provide a more holistic view of target antigen expression and the humoral immune responses to ETEC that are associated with protection.
An advantage of the present studies is that they involve infection with a single well defined ETEC strain, H10407, for which the complete genome was available [7]. This made it possible to interrogate the surface proteome of the challenge strain to determine precisely what peptides would be presented by the challenge inoculum and, subsequently, to assess the breadth of the immunologic response to these peptides. The data presented here suggest that although the immune response to ETEC is complex, infected hosts generally respond to classic vaccine antigens and to a finite number of noncanonical protein antigens during the course of infection.
A number of the immunodominant proteins highlighted in the microarray studies have been recently investigated as potential protective antigens based on their surface expression and/or similarity to subunit components of vaccines for other pathogens [47]. EtpA, a two-partner secretion exoprotein adhesin, is similar to filamentous HA, a component of acellular pertussis (aPV) vaccines, whereas antigen 43 and the adhesin-autotransporter proteins belong to the same molecular family as pertactin, another aPV subunit [48]. Vaccination with EtpA, alone [23] or in combination with LT [49], the adhesin autotransporter proteins [27], as well as EatA [21] and FliC [29], all afford some degree of protection in a murine model of infection. In addition, the highly conserved YghJ metalloprotease, also referred to as SslE, has been shown to provide broad protection against a variety of E coli pathovars [50]. Data presented here demonstrate that each of these antigens are recognized during the course of ETEC infection, raising the possibility that they could serve as viable vaccine components.
Heterogeneity of classic ETEC antigens has significantly impeded progress toward a protective vaccine. Therefore, molecular conservation of the noncanonical antigens will be an important parameter to consider with regard to their potential utility as vaccine targets. Data emerging from ongoing molecular epidemiology studies on a diverse collection of hundreds of isolates suggest that some of the more recently discovered antigens including EatA, YghJ, and EtpA are relatively conserved across a broad range of geographically and phylogenetically disparate stains, and that these secreted molecules exhibit surprisingly little antigenic variation.
The responses to FliC, the major subunit of flagellar filaments responsible for ETEC motility, are intriguing. We observed responses to the serotype (H11)-specific region of FliC present in the H10407 challenge strain, to FliCs from heterologous serotypes, and specifically to highly conserved regions common to all E coli FliC molecules. Prior studies demonstrated that these conserved regions FliC are exposed, bind to the EtpA adhesin [22], and are protective [29]. However, additional study is needed to determine whether these highly immunogenic proteins offer protection against ETEC and other important Gram-negative infections.
In the canonical view of ETEC pathogenesis that has formed the basis for vaccine development, these organisms adhere to the small intestine via CFs and deliver LT and/or ST. However, it appears likely that ETEC coordinate the deployment of both classic vaccine antigens and novel antigenic targets in an elegantly orchestrated engagement of the host. Indeed, the recent discovery that CFA/I, EtpA, and the pertactin-like adhesin autotransporter proteins are all governed by a single transcriptional regulator [39] might also suggest that successful infection with these pathogens represents a coordinated, serial engagement of the host that culminates in the recognition of multiple antigens. Therefore, an expanded paradigm for ETEC pathogenesis could open additional avenues to interdict effective pathogen host interactions and novel rational approaches to vaccines.
CONCLUSIONS
Although the current studies provide a broadened perspective of host responses to ETEC infection, significant additional effort will be needed to precisely define distinct immunologic benchmarks associated with protection, test candidate antigens, and to integrate these into formulation of a broadly protective vaccine. Nevertheless, these recent data provide a strong incentive as well as a technology platform for further investigation.
Supplementary Data
Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Acknowledgments. We thank Edwin Oaks (Walter Reed Army Institute of Research), Stephen Savarino, and Stephen Poole (Naval Medical Research Center) for providing select purified antigens that were used in production of the arrays.
Disclaimer. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health (NIH), National Institute of Allergy and Infectious Diseases (NIAID), National Center for Advancing Translational Sciences (NCATS), the Department of Veterans Affairs, or PATH.
Finanical support. Research reported in this publication was supported by PATH; funding from NIAID of the NIH under award numbers R01AI089894and R01AI126887 (to J. M. F.), the Washington University Institute of Clinical and Translational Sciences grant UL1 TR000448 from the NCATS of the NIH, and the Department of Veterans Affairs (5I01BX001469; to J. M. F.).
Potential conflicts of interest. J. M. F. is listed as an inventor on patent 8323668 related to the EtpA protein. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
