Genetic Fusions of a CFA/I/II/IV MEFA (Multiepitope Fusion Antigen) and a Toxoid Fusion of Heat-Stable Toxin (STa) and Heat-Labile Toxin (LT) of Enterotoxigenic Escherichia coli (ETEC) Retain Broad Anti-CFA and Antitoxin Antigenicity

Immunological heterogeneity has long been the major challenge in developing broadly effective vaccines to protect humans and animals against bacterial and viral infections. Enterotoxigenic Escherichia coli (ETEC) strains, the leading bacterial cause of diarrhea in humans, express at least 23 immunologically different colonization factor antigens (CFAs) and two distinct enterotoxins [heat-labile toxin (LT) and heat-stable toxin type Ib (STa or hSTa)]. ETEC strains expressing any one or two CFAs and either toxin cause diarrhea, therefore vaccines inducing broad immunity against a majority of CFAs, if not all, and both toxins are expected to be effective against ETEC. In this study, we applied the multiepitope fusion antigen (MEFA) strategy to construct ETEC antigens and examined antigens for broad anti-CFA and antitoxin immunogenicity. CFA MEFA CFA/I/II/IV [CVI 2014, 21(2):243-9], which carried epitopes of seven CFAs [CFA/I, CFA/II (CS1, CS2, CS3), CFA/IV (CS4, CS5, CS6)] expressed by the most prevalent and virulent ETEC strains, was genetically fused to LT-STa toxoid fusion monomer 3xSTaA14Q-dmLT or 3xSTaN12S-dmLT [IAI 2014, 82(5):1823-32] for CFA/I/II/IV-STaA14Q-dmLT and CFA/I/II/IV-STaN12S-dmLT MEFAs. Mice intraperitoneally immunized with either CFA/I/II/IV-STa-toxoid-dmLT MEFA developed antibodies specific to seven CFAs and both toxins, at levels equivalent or comparable to those induced from co-administration of the CFA/I/II/IV MEFA and toxoid fusion 3xSTaN12S-dmLT. Moreover, induced antibodies showed in vitro adherence inhibition activities against ETEC or E. coli strains expressing these seven CFAs and neutralization activities against both toxins. These results indicated CFA/I/II/IV-STa-toxoid-dmLT MEFA or CFA/I/II/IV MEFA combined with 3xSTaN12S-dmLT induced broadly protective anti-CFA and antitoxin immunity, and suggested their potential application in broadly effective ETEC vaccine development. This MEFA strategy may be generally used in multivalent vaccine development.


Introduction
Virulence heterogeneity among bacterial and viral strains or isolates has long been one major challenge in vaccine development. Like many other infectious pathogens, enterotoxigenic Escherichia coli (ETEC) strains (i.e., E. coli producing heat-labile and heat-stable enterotoxins) are immunologically heterogeneous. ETEC strains are the most common bacterial cause of diarrhea which continues to be the second leading cause of death in children younger than 5 years who live in developing countries and remains a major threat to global health [1,2]. These ETEC strains express immunologically different colonization factor antigen (CFA) adhesins and enterotoxins. CFA adhesins mediate bacterial attachment to host receptors and facilitate colonization in host small intestines. Enterotoxins disrupt fluid and electrolyte homeostasis in host small intestinal epithelial cells that leads to fluid hyper-secretion and diarrhea [3]. Therefore, CFA adhesins and enterotoxins are recognized the major virulence determinants. There are at least 23 CFA or CS (coli surface antigen) adhesins and two very distinct enterotoxins characterized among ETEC strains isolated from humans with diarrhea [3][4][5][6]. Enterotoxins produced by ETEC strains associated with human diarrhea are heat-labile toxin (LT) and heatstable toxin type Ib (STa, human-type STa, STh or hSTa). Although heat-stable toxin type Ia (porcine-type STa, STp or pSTa) and heat-stable toxin type II (STb), together with LT and/or STa, are occasionally detected in ETEC strains isolated from human diarrheal patients, pSTa and STb toxins cause diarrhea only in animals [3]. Since ETEC strains expressing any one or two CFA or CS adhesins and LT or STa enterotoxin cause diarrhea, developing broadly protective vaccines against ETEC diarrhea continues to be very challenging [7,8].
Anti-CFA antibodies specific to an individual CFA or antitoxin antibodies to LT were reported to provide protection to human volunteers against homologous challenge [9][10][11][12], but not against ETEC strains expressing immunologically heterogeneous CFAs or the STa toxin. Early experimental vaccine studies showed that killed ETEC prototype strain H10407 (O78: H11, LT + STa + CFA/I + ) induced anti-CFA/I and anti-LT antibodies that protected volunteers against challenge of strain H10407 or a homologous strain [13,14]. These observations led to a conceptual ETEC vaccine, that is, an ETEC vaccine that should induce antibodies protecting against multiple CFA adhesins and toxin LT [15]. Consequently, products that were the result of mixing together a few strains that expressed several CFA adhesins and recombinant LT-B subunit protein or the homologous cholera toxin B subunit (CT-B) were examined for protection against ETEC diarrhea. Experimental vaccines currently under development include two oral whole-cell ETEC vaccine candidates, rCTB-CF and ACE527. The rCTB-CF consists of five killed strains expressing six CFA adhesins plus recombinant CT-B subunit protein [16,17], and the ACE527 is composed of three live attenuated E. coli strains that express five CFA adhesins, one CFA subunit, and LT-B subunit [18,19]. Recombinant CT-B subunit was included in the rCTB-CF product because anti-CT-B antibodies were shown to provide short-term protection against LT-producing ETEC strains [20]. Field studies showed that the rCTB-CF vaccine induced antibody responses and protected adults traveling from developed countries to ETEC endemic countries against the risk of disease by 60% to 70% [16,17] or against moderate to severe diarrhea [21]. This product, however, provided no protection to children, especially very young children living in endemic areas against ETEC diarrhea, and caused some adverse effects in very young children when an adult dose was given orally [22,23]. In addition, it did not reduce the overall diarrhea rate among US adults traveling to Guatemala and Mexico [21]. The live ACE527 product was found associated with some adverse effects in volunteer studies, but the adverse effects were reduced or eliminated when a lower dose was given [24]. This ACE527 induced antibody responses to LT-B, CFA/I, CS3, and CS6 among adult volunteers, but protected against only the severity of diarrhea outcome from homologous challenge [18,25].
Although efforts have been undertaken to continuously improve current vaccine candidates [24,26,27], these products might still be unable to provide truly broad protection against ETEC diarrhea because they carry only LT-B or CT-B subunit antigen but no STa antigens to induce antibodies against STa toxin. Anti-LT antibodies were found protective against ETEC strains expressing LT but not against ETEC strains expressing the heat-stable toxin [28,29]. Furthermore, cocktail products require a relatively high oral dose to stimulate host immune responses against individual CFA adhesins and LT. A high oral dose delivers many somatic antigens, particularly lipopolysaccharide (LPS) which is thought to be the cause of the gastrointestinal side effects associated with these vaccine candidates, especially in young children [30], and may also mask stimulation of host anti-CFA and anti-LT (or anti-CT) specific antibody responses [8,31].
Ideally, a vaccine should induce antibodies against all CFA adhesins and both toxins, particularly STa toxin. In fact ETEC strains producing STa as the only toxin are the frequent cause of ETEC diarrhea in developing countries [32]. ETEC strains that produce STa with or without LT are associated with over two-thirds of human ETEC diarrhea cases [5], and tend to cause moderate to severe diarrhea more frequently [21,33]. Therefore, ETEC vaccines may also need to induce protective anti-STa antibodies in order to provide broad protection against ETEC diarrhea.
Although STa is poorly immunogenic because of its small size (19 amino acids), it is a very potent enterotoxin. Thus, STa presents a problem for use as a potential immunogen because it does not induce immunity following exposure and also causes disease unless inactivated [34,35]. However, recent studies demonstrated that STa, when had a single amino acid substituted, was less or non-toxic; moreover, derived STa toxoids became immunogenic after being genetically fused to a strongly immunogenic LT toxoid monomer (a single peptide consisting of a mutated LT-A subunit and one LT-B subunit) [36,37]. LT-STa toxoid fusions that carried different STa toxoids were found to induce neutralizing antibodies against both LT and STa toxins [38].
To induce anti-CFA antibodies effectively protecting against heterogeneous CFA adhesins is another challenge in ETEC vaccine development. A recent study showed that a single CFA/I/ II/IV multiepitope fusion antigen (MEFA) was constructed and this CFA/I/II/IV MEFA induced antibodies cross protective against seven ETEC CFA adhesins, CFA/I, CFA/II (CS1, CS2, CS3) and CFA/IV (CS4, CS5, CS6) [39]. However, although these seven CFA adhesins are expressed by ETEC strains causing 70-80% of ETEC diarrhea cases [23], their prevalence varies greatly at different geographic locations [40]. Thus, antibodies against these seven CFA adhesins may still not be sufficient to effectively protect against ETEC diarrhea in some regions.
In contrast to CFA adhesins with variable prevalence geographically, LT and STa, alone or together, are expressed by all ETEC strains that cause diarrhea in humans. But anti-toxin antibodies without assistance from antibodies against CFA adhesins may be less effective against ETEC diarrhea [41]. If the CFA/I/II/IV MEFA is combined with or further fused to an optimal LT-STa toxoid fusion that induces protective antibodies against both toxins, the resultant product should be able to induce immunity to provide even broader protection against ETEC diarrhea, since the antitoxin antibodies provide supplementary protection against ETEC strains expressing these seven CFA adhesins and independent protection against the ETEC strains expressing the other CFA adhesins.
To create a single antigen inducing antibodies broadly protecting against seven CFA adhesins and both toxins and to explore potential application of such an antigen in ETEC vaccine development, in this study we first genetically fused the CFA/I/II/IV MEFA gene [39] to a LT-STa toxoid fusion gene [38] for a CFA/I/II/IV-STa -toxoid -dmLT MEFA and examined anti-CFA and antitoxin immunogenicity in a murine model. We then evaluated the induced antibodies for neutralization activities against adherence from these seven CFA adhesins and enterotoxicity of both toxins. In addition, we immunized mice with the CFA/I/II/IV MEFA combined with toxoid fusion 3xSTa N12S -dmLT, and compared induced antigen-specific antibody responses with those induced by the CFA/I/II/IV-STa N12S -dmLT MEFA to assess whether genetic fusion affected antigenic property of the fused CFA adhesin and toxin antigen components, thus enabling application of MEFA strategy for multivalent vaccine development.
Expression, extraction and detection of CFA/I/II/IV-STa -toxoid -dmLT MEFA proteins E. coli strain BL21 was used to host plasmid carrying CFA/I/II/IV-STa A14Q -dmLT or CFA/I/II/ IV-STa N12S -dmLT chimeric gene to express two CFA/I/II/IV-STa -toxoid -dmLT MEFA proteins. Each recombinant strain (from a single colony) was cultured in 5 ml Luria Bertani (LB) broth supplemented with kanamycin (30μg/ml) at 37°C on a shaker (150 rpm). Overnight grown culture was added to 500 ml 2x YT (2x Yeast Extract Tryptone) medium broth for continued incubation. Bacteria were induced with isopropyl-1-thio-β-D-galactoside (IPTG; 0.5 mM) after culture optical density reached 0.5 at 600 nm (OD 600 ), and were incubated for 4 more hours. Bacterial culture was centrifuged at 5,000 X g for 20 min, and pellets were used to extract total insoluble proteins (inclusion body fraction) using bacterial protein extraction reagent (B-PER, in phosphate buffer; Pierce, Rockford, IL) by following the manufacturer's protocol.

Mouse immunization with CFA/I/II/IV-STa -toxoid -dmLT MEFA proteins
Mouse immunization studies complied with the Animal Welfare Act by following the 1996 National Research Council guidelines [45], and were approved and supervised by a state veterinarian and the Kansas State University Institutional Animal Care and Use Committee. Data generated from this study are publicly available. Before being used in mouse immunization, CFA/I/II/IV-STa -toxoid -dmLT MEFA proteins were verified non-toxic using T-84 cells and EIA cAMP and cGMP kits (Assay Design, MI) as previously described [34,37,38].
Four groups of 6-to 8-week-old female BALB/c mice (Charles River Laboratories International, Inc., Wilmington, MA) were included in the immunization study. The first group of 8 mice was each injected intraperitoneally (i. Blood samples were collected from each mouse before the primary immunization and 10 to 12 days after each immunization to prepare serum samples. Serum samples were stored at -80°C until use. On day 37 post primary immunization, mice were anesthetized with CO 2 and exsanguinated.

Anti-CFA antibody adherence inhibition assay
Mouse serum samples pooled from each group were examined for in vitro antibody adherence inhibition activity against ETEC strains expressing CFA/I, CS3, CS4/CS6, CS5/CS6, or CS6 and E. coli recombinant strains expressing CS1 or CS2, using Caco-2 cells (ATCC, #HTB-37) as previously described [39]. ETEC or E. coli bacteria expressing CFA/I, CFA/II or CFA/IV adhere to Caco-2 cells, whereas neutralizing anti-CFA antibodies in mouse serum samples block adherence of these ETEC or E. coli bacteria to Caco-2 cells. Therefore, by mixing the mouse serum and bacteria of each strain, incubating the mixture with Caco-2 cells, and counting the bacteria adhered to the incubated cells, we were able to measure serum antibody adherence inhibition activities against ETEC or E. coli expressing these seven CFA adhesins. Briefly, Caco-2 cells were seeded and grown in 75 ml flask (Corning). After growth of confluence, Caco-2 cells were transferred to each well of a 12-well tissue culture plate containing Dulbecco's modified Eagle's medium (DMEM)-20% fetal bovine serum (FBS) (Fisher Thermo Scientific, Pittsburg, PA), and grown to a confluent monolayer (7x10 5 per well). ETEC and E. coli bacteria, after overnight growth on sheep blood agar plates at 37°C, were scraped off with cotton swabs and were gently suspended in sterile PBS. One hundred microliters of each bacterial suspension (3.5x10 6 bacteria; with a multiplicity-of-infection ratio set at five bacteria to one Caco-2 cell) were incubated with 20 μl serum sample pooled from the mice in each group on a shaker (50 rpm) for 1 h at room temperature. The bacteria and serum mixture was brought to 300 μl with PBS, and was added to each well containing the Caco-2 cells (in 700 μl cell culture medium). After incubation in a CO 2 incubator (5% CO 2 ) for 1 h at 37°C, wells were gently washed with PBS to remove non-adherent ETEC or E. coli bacteria. Washed Caco-2 cells were dislodged through incubation with 0.25% trypsin (200 μl per well) in a CO 2 incubator for 30 min at 37°C. Dislodged Caco-2 cells (with adherent ETEC or E. coli bacteria) were collected by centrifugation (15,000 g for 10 min) and then suspended in 1 ml PBS. Suspensions were gently but well mixed, serially diluted, and plated on LB plates. After overnight growth at 37°C, ETEC or E. coli bacteria colony forming units (CFUs) were counted.

Anti-LT and anti-STa antibody neutralization assays
Serum samples pooled from mice in each group were also examined for in vitro antibody neutralization activities against STa and CT using EIA cAMP and cGMP kits (Assay Design) and T-84 cells. STa stimulates an increase of intracellular cyclic GMP levels and CT elevates intracellular cAMP levels in T-84 cells. Neutralizing antitoxin antibodies neutralize enterotoxicity thus prevent STa and CT from stimulating intracellular cGMP or cAMP. Therefore, by incubating serum with the toxin, adding the mixture to T-84 cells, and measuring cGMP or cAMP levels in the cells, we were able to evaluate neutralization activities of mouse serum IgG antibodies against STa or CT. As described previously [34,[36][37][38], 30 μl pooled serum sample of each immunization group or the control group was incubated with 2 ng STa toxin or 10 ng CT for 30 min at room temperature, and the serum/toxin mixture was added to T-84 cells. After incubation in a CO 2 incubator for 1 h (for STa to measure cGMP) or 3 h (for CT to measure cAMP), T-84 cells were measured for intracellular cGMP or cAMP levels (pmol/ml) with EIA cGMP or cAMP kit by following the manufacturer's protocol (Assay Design). STa or CT alone (without serum) was used as the control to show enterotoxicity in stimulation of cGMP or cAMP in T-84 cells, and culture medium only (without toxin or serum) was used to show a baseline of intracellular cAMP or cGMP level in T-84 cells.

Statistical analysis
Data were analyzed using SAS for Windows, version 8 (SAS Institute, Cary, NC). Results were presented as means ± the standard deviations. A Student's t-test was used to compare different treatment groups. Calculated p values of less than 0.05 were considered as significant when treatment groups were compared using two-tailed distribution and two-sample unequal variance. In addition, the non-parametric Mood's Median Test was carried out to assess differences of antibody neutralization activities (cAMP and cGMP, pmol/ml) from serum samples of the immunized group and the control group.

Results and Discussion
Constructed CFA/I/II/IV-STa -toxoid -dmLT MEFAs consisted of epitopes of the seven CFA adhesins, two copies of the STa toxoid, and a part of the dmLT monomer The overlap of CFA/I/II/IV MEFA and the 3xSTa A14Q -dmLT or 3xSTa N12S -dmLT PCR amplified products yielded two CFA/I/II/IV-STa -toxoid -dmLT chimeric genes (Fig. 1). Two recombinant strains, 9208 and 9401, were constructed to express two CFA/I/II/IV-STa -toxoid -dmLT MEFA proteins. These two MEFA proteins differed only at the STa toxoid, with STa A14Q in 9208 and STa N12S in 9401 (Table 1). DNA sequencing showed each chimeric gene was a single open reading frame coding a single 6xHis-tagged CFA/I/II/IV-STa -toxoid -dmLT protein. Each fusion protein consisted of 20 amino acids from the pET28α vector including the 6xHis tag (six histidines), the CFA/I/II/IV MEFA (150 amino acids) carrying epitopes of CFA/I and CS1-CS6 major subunits (CfaB, CooA, CotA, CstH, CsaB, CsfA and CssA), two copies of STa toxoid STa A14Q or STa N12S , 109 amino acids of dmLT-A subunit (132-240, at the C-terminus; with R192G and L211A mutations) and one copy of the LT-B subunit (100 amino acids), and four intra-peptide linkers (Fig. 1C). The first copy of the STa toxoid (without the stop codon) with a 'GPVD' linker was located after the mutated 192th amino acid residue (Arg!Gly) of the LT-A subunit, and the second STa toxoid (with the stop codon) with an L-linker was at the C-terminus of the fusion protein.

CFA/I/II/IV-STa
There were no IgG antibodies specific to CFA/I, CFA/II, CFA/IV, STa or LT detected in the serum samples of the control mice or serum samples collected before the primary immunization.  Table 1 (500 ng per well of a 2HB plate), STa-ovalbumin (10 ng per well of a Costar plate), or LT (List Biological Laboratories, Inc.; 100 ng per well of a 2HB plate) and HRP-conjugated goat-antimouse IgG (1:3300; the secondary antibodies) were used to titrate IgG antibodies specific to CFA/I, CS1, CS2, CS3, CS4/6, CS5/6 and to STa and LT toxins, respectively. The antibody titer was calculated from the highest dilution of a serum sample that produced an ELISA optical density of greater than 0.3 (above the background) and presented in a log 10 scale. Each dot represented a mouse IgG titer, and the bars indicated the mean titer of the group. Serum samples of the immunized mice were shown to inhibit adherence of ETEC strains expressing CFA/I, CS3, CS4/CS6, CS5/CS6 or CS6 and E. coli strains expressing CS1 or CS2 to Caco-2 cells Serum samples pooled from the mice co-immunized with CFA/I/II/IV and 3xSTa N12S -dmLT exhibited significant inhibition activities against adherence of H10407 (CFA/I + LT + STa + ), E116 (CS3 + LT + STa + ), E106 (CS4 + CS6 + LT + STa + ), UM75688 (CS5 + CS6 + LT + STa + ) and 2423/ ETP98066 (CS6 + LT + STa + ), and E. coli recombinant strains expressing CS1 adhesin or CS2 adhesins (Table 3). Serum sample pooled from the mice immunized with CFA/I/II/ IV-STa N12S -dmLT showed significant adherence inhibition activities against all examined ETEC and E. coli strains except the recombinant E. coli strain expressing CS2.

Antibodies in serum samples of the immunized mice showed neutralizing activities against STa toxin and CT in vitro
Pooled serum samples from mice immunized with CFA/I/II/IV-STa A14Q -dmLT or CFA/I/II/ IV-STa N12S -dmLT, or co-immunized with CFA/I/II/IV MEFA and 3xSTa N12S -dmLT showed neutralization activity against both STa and CT. Intracellular cGMP levels in T-84 cells incubated with 2 ng STa toxin and the serum of the mice immunized with CFA/I/II/IV-STa A14Q -dmLT, CFA/I/II/IV-STa N12S -dmLT, or CFA/I/II/IV combined with 3xSTa N12S -dmLT were 36.7 ±20.7, 23.2±7.0, 0.31±0.42 (pmol/ml), respectively (Fig. 3). These cGMP levels were significantly Table 2. Anti-CFA/I, -CS1, -CS2, -CS3, -CS4/CS5, and anti-CS5/CS6, and anti-STa and anti-LT IgG antibody titers (in log 10 ; mean ± standard deviation) detected in the serum of mice immunized with the CFA/I/II/IV-STa N12S -dmLT or co-immunized with the CFA/I/II/IV MEFA and the 3xSTa N12S -dmLT toxoid fusion.   (Fig. 4). These cAMP levels were significantly lower than those in cells incubated with the toxin alone (53.7±1.3; p<0.01 in Student's t-test and Mood's Median Test) or with the toxin and the serum sample of the control mice (40.1±6.5; p<0.01 in Student's t-test and Mood's Median Test).
Results from this study showed that CFA/I/II/IV-STa A14Q -dmLT and CFA/I/II/IV-STa N12S -dmLT MEFAs retained anti-CFA and antitoxin antigenicity as both induced antibody responses to seven heterogeneous ETEC CFA adhesins and LT and STa toxins. That suggested potential application of CFA/I/II/IV-STa -toxoid -dmLT MEFAs in development of broadly protective ETEC vaccines. Vaccines inducing anti-CFA antibodies to broadly prevent ETEC adherence and colonization and also antitoxin antibodies to neutralize both LT and STa toxins are expected to be broadly effective against ETEC diarrhea. To develop a vaccine carrying representative antigens of all or even a majority of the 23 characterized CFA adhesins and also two toxins is an unreachable goal currently. But it appears feasible to develop a vaccine protecting against ETEC strains that cause a majority of clinical cases including moderate to severe cases. Vaccine candidates currently under development carry antigens of up to six CFA adhesins and CT-B or LT-B subunit, but they do not induce anti-STa immunity and also require a high oral delivery dose with excessive amount of somatic antigen. These vaccine candidates would have to include additional strains to deliver additional CFAs and STa antigens in order to induce Table 3. Results of in vitro antibody adherence inhibition assays a , using serum samples of mice immunized with CFA/I/II/IV-STa N12S -dmLT, co-administrated with CFA/I/II/IV MEFA and toxoid fusion 3xSTa N12S -dmLT, or the negative control mice. The number of ETEC or E. coli bacteria adhered to Caco-2 cells was used to indicate activity of anti-CFA antibodies against bacteria adherence. broader anti-CFA and antitoxin antibody responses. But inclusion of additional strains would require an even higher oral dose leading to potentially greater side effects. A vaccine with fewer strains, or ideally a single strain, to express multiple CFA adhesins and LT and STa toxins (or antigenic elements of these CFA adhesins and toxins) could overcome the difficulties encountered by the cocktail vaccine candidates. Recently, we demonstrated that a single CFA/I/II/IV MEFA carrying epitopes of seven CFA adhesins [CFA/I, CFA/II (CS1, CS2, CS3) and CFA/IV (CS4, CS5, CS6)] induced antibodies broadly inhibiting adherence of ETEC and E. coli strains expressing these seven CFA adhesins [39]. Protection against these seven CFA adhesins can potentially protect against ETEC strains associated with 70% to 80% of ETEC diarrhea cases [23,40] and also the moderate to severe cases [46]. Unlike cocktail vaccine candidates, this CFA/I/II/IV MEFA was expressed as a single protein by a single E. coli strain, and for the first time induced antibody responses in mice protecting against all seven 3) pooled from each group of mice immunized with CFA/I/II/IV-STa A14Q -dmLT or CFA/I/II/IV-STa N12S -dmLT, co-immunized with CFA/I/II/ IV and 3xSTa N12S -dmLT, the control group, or the serum sample collected prior to immunization was incubated with STa toxin (2 ng, in 150 μl cell culture medium) for 30 min at room temperature, and the serumtoxin mixture was added to T-84 cells (1 ml of final volume with cell culture medium). Intracellular cGMP concentration in T-84 cells was measured after 1 hour incubation at a CO 2 incubator, with the mean cGMP and standard deviation (from four to six replicates) of each group indicated as columns and bars. The cGMP levels in T-84 cells cultured with cell culture medium alone (without STa toxin or serum; no STa toxicity) or with STa toxin in culture medium (without serum; STa toxicity) were used as controls.
doi:10.1371/journal.pone.0121623.g003 CFA adhesins. Products developed from this CFA/I/II/IV MEFA can be formulated and administrated at a regular or lower dose, thus will eliminate the risk of adverse effects associated with the cocktail vaccine candidates. In addition, using a peptide or epitopes instead of an entire adhesin as the antigen may have advantage. A peptide from an adhesin subunit can be more effective in inducing adhesin-specific antibody responses. Experimental vaccines carrying intact F18 fimbriae, physically purified (as a subunit vaccine) or expressed by a live E. coli strain (as a live vaccine), were unable to induce effective antibody responses to protect pigs against F18-fimbrial ETEC challenge [47,48]; however, a small peptide from the minor subunit of the F18 fimbrial adhesin was shown to induce strongly protective antibodies against this F18 adhesin [44,49]. Data from this study showed that mice immunized with the CFA/I/II/IV MEFA 3) pooled from each group of mice immunized with CFA/I/II/IV-STa A14Q -dmLT or CFA/I/II/IV-STa N12S -dmLT, co-immunized with CFA/I/II/IV and 3xSTa N12S -dmLT, or the control group was incubated with CT toxin (10 ng, in 150 μl cell culture medium) for 30 min at room temperature, and the serum-toxin mixture was added to T-84 cells (1 ml of final volume with cell culture medium). Intracellular cAMP concentration (pmole/ml) in T-84 cells was measured after 3 hour incubation at a CO 2 incubator, with the mean cAMP and standard deviation (from four to six replicates) of each group indicated as columns and bars. The cAMP levels in T-84 cells incubated with cell culture medium alone (without CT or serum; no CT toxicity) or with CT in culture medium (without serum; CT toxicity) were also measured as references. (co-administrated with the toxoid fusion) or the CFA/I/II/IV-STa -toxoid -dmLT that carried a part of the CFA/I major subunit peptide developed a similar anti-CFA/I IgG titer as mice immunized with CFA/I fimbriae.
We reported recently that a STa toxoid, such as STa A14Q or STa N12S , can be genetically fused to a monomeric LT toxoid peptide and the resultant LT-STa toxoid fusions elicited antibodies neutralizing STa and LT toxins [37,38]. Results from the toxoid fusion and the CFA/I/ II/IV MEFA studies led us to believe that a single CFA/I/II/IV-STa -toxoid -dmLT MEFA could induce broad anti-CFA and antitoxin antibody responses and can be potentially used for development of a broadly protective ETEC vaccine. Data from the present study showed mice immunized with CFA/I/II/IV-STa A14Q -dmLT or CFA/I/II/IV-STa N12S -dmLT developed antibody responses to all seven adhesins and both toxins. That indicated that the CFA/I/II/IV-STa -toxoid -dmLT MEFA retained antigenicity of individual CFA adhesins and toxins, and suggested potential application of CFA/I/II/IV-STa -toxoid -dmLT MEFA in ETEC vaccine development.
Data from this study showed that mice immunized with the CFA/I/II/IV-STa N12S -dmLT developed greater titer of anti-STa IgG antibodies (2.50±0.52; in log 10 ) than mice immunized with CFA/I/II/IV-STa A14Q -dmLT (2.40±0.75); also, serum antibodies induced by CFA/I/II/ IV-STa N12S -dmLT exhibited greater neutralizing activity against STa enterotoxicity (Fig. 3). That confirmed STa N12S is the preferred toxoid for fusions to induce anti-STa antibody response. However, it was noted that IgG antibody titers to all CFA adhesins and LT in serum of the mice immunized with the CFA/I/II/IV-STa N12S -dmLT were found systematically lower than those in the serum of mice immunized with CFA/I/II/IV-STa A14Q -dmLT. Given the fact that the only difference between these two MEFA antigens was the different STa toxoid, these two MEFA antigens should have induced similar titers of antibody responses to the CFAs and LT. Future studies will be needed to explore whether different STa toxoids alter antigenic structure of CFA/I/II/IV-STa -toxoid -dmLT MEFAs.
Since CFA/I/II/IV-STa N12S -dmLT MEFA induced antibodies with greater neutralizing activity against STa (compared to CFA/I/II/IV-STa A14Q -dmLT MEFA) and also significant adherence inhibition against ETEC or E. coli strains expressing six of the seven adhesins, this MEFA was evaluated for potential application in multivalent vaccine development. By comparatively examining antigen-specific antibody responses in mice immunized with the single CFA/I/II/IV-STa N12S -dmLT MEFA and in mice immunized with a combination of CFA/I/II/ IV MEFA and toxoid fusion 3xSTa N12S -dmLT, we found that mice in the two immunization groups developed similar levels of anti-CS1, -CS3, -CS4/6 and anti-CS5/6, and anti-LT IgG antibody responses ( Table 2). That suggested that the fusion process did not compromise antigenicity of a majority of the individual antigenic components. It was observed, however, that mice co-immunized with CFA/I/II/IV MEFA and 3xSTa N12S -dmLT fusion developed moderately greater titers of anti-CFA/I (2.69±0.05 vs 2.63±0.9; p = 0.04), anti-CS2 (2.53±0.08 vs 2.44 ±0.10; p = 0.01) and anti-STa (2.88±0.08 vs 2.50±0.52; p = 0.01) serum IgG antibodies. The negative effect on anti-CFA/I and anti-CS2 immunogenicity could be resulted from alteration of the CFA/I and CS2 epitope antigenic topology in the CFA/I/II/IV-STa N12S -dmLT MEFA, since the CFA/I epitope ( 159 SGVVSLVMT 167 ) and the CS2 epitope were located at the C-terminus of the CFA/I/II/IV MEFA to which the toxoid fusion was fused. Future protein structure studies may help to reveal if any structural alteration occurred. Since the anti-adhesin antibodies induced by the CFA/I/II/IV-STa N12S -dmLT MEFA were able to significantly inhibit adherence of ETEC H10407 strain but not against CS2 strain (Table 3), we may only need to optimize the construction to enhance anti-CS2 immunogenicity. To relocate the CS2 epitope toward the N-terminus or to extend the linker between CFA/I/II/IV MEFA and the toxoid fusion may improve CS2 epitope presentation and anti-CS2 immunogenicity.
Data from the study showed a lower anti-STa IgG antibody response in mice immunized with CFA/I/II/IV-STa N12S -dmLT MEFA compared to the mice co-administrated with the CFA/I/II/IV and 3xSTa N12S -dmLT (Table 2). That was not surprising because only two copies of STa N12S were carried by CFA/I/II/IV-STa N12S -dmLT MEFA (versus three STa N12S toxoids were carried by the 3xSTa N12S -dmLT used for co-administration). Additional copies of a STa toxoid in a LT-STa toxoid fusion were shown to enhance anti-STa antigenicity [37]. That also explained that only serum sample from the co-administrated mice was able to completely neutralized 2 ng STa (Fig. 3). Had CFA/I/II/IV-STa N12S -dmLT MEFA carried three copies of STa N12S , we would have observed a similar level of anti-STa antibody response and antibody neutralization activity against STa from the serum of the immunized mice. Future studies to modify the CFA/I/II/IV-STa N12S -dmLT MEFA to carry three copies of STa N12S , and to examine whether it induces the same level of anti-STa antibody response and more importantly antibodies with similar neutralizing activity against STa as the combination of 3xSTa N12S -tmLT and CFA/I/II/IV will be informative.
Data from the study also showed that serum of mice co-immunized with CFA/I/II/IV MEFA and 3xSTa N12S -dmLT showed greater neutralization activity against CT (Fig. 4). In contrast to CFA/I/II/IV-STa A14Q -dmLT or CFA/I/II/IV-STa N12S -dmLT MEFA which contained the last 109 amino acids of LT-A1, the co-administrated toxoid fusion 3xSTa N12S -dmLT carried the entire LT-A1 peptide. That suggested the antibody response to the LT-A1 peptide could play an important role in neutralizing against LT or CT toxin.
Although results from the current study suggested that a fusion of the CFA/I/II/IV MEFA and the toxoid fusion did not significantly alter antigenicity of the carried adhesin and toxin antigenic components, the combination of the CFA/I/II/IV MEFA and toxoid fusion 3xSTa N12S -dmLT seemed the better antigens for future ETEC vaccine development as these two antigens induced greater titers of anti-STa, anti-CFA/I and anti-CS2 antibodies than the single CFA/I/II/IV-STa N12S -dmLT MEFA. But a vaccine product derived from a single antigen or a single strain decreases complexity and cost in product manufacture. If the CFA/I/II/ IV-STa N12S -dmLT MEFA, with modification to carry an additional copy of STa N12S and to have anti-CS2 antigenicity enhanced, is safe and induces equivalent antibody responses to all seven adhesins and both toxins as the combination of CFA/I/II/IV and 3xSTa N12S -dmLT, this modified CFA/I/II/IV-STa N12S -dmLT MEFA should be the preferred antigen for ETEC vaccine development.
It needs to point out that only intraperitoneal immunization route was used to examine immunogenicity of the CFA/I/II/IV-STa -toxoid -dmLT MEFAs in this study. Future studies using intradermal, intramuscular, subcutaneous or even oral route, and perhaps with different adjuvants such as dmLT which is also an antigen to further induce neutralizing anti-LT antibodies, will help us to characterize better the immunogenicity of a modified CFA/I/II/ IV-STa N12S -dmLT MEFA. In addition, fusion protein structure, protein integrity and stability, and more importantly safety of this CFA/I/II/IV-STa N12S -dmLT MEFA, as well as the CFA/I/ II/IV MEFA and the 3xSTa N12S -dmLT (for co-administration), would have to be thoroughly characterized before it can be assessed for ETEC vaccine development. Furthermore, animal models including a piglet challenge model and a rabbit colonization model, and perhaps the modified RITARD (removable intestine tie adult rabbit diarrhea) model, will be needed in preclinical studies to evaluate protective efficacy of the induced anti-CFA and antitoxin immunity against ETEC diarrhea prior to human volunteer studies or field trials. Additionally, despite of results from a recent study indicated that anti-STa monoclonal and polyclonal antibodies had no reactivity with guanylin and a very low reactivity with uroguanylin [50], anti-STa antibodies induced by the toxoid fusion or CFA-toxoid MEFAs would need to be evaluated for cross reactivity with guanylin and uroguanylin. This study, nevertheless, demonstrated that epitopes from multiple CFA adhesins and antigens from both toxins can be fused together as a single MEFA protein to induce broadly protective antibody responses, and suggested potential application of MEFA for effective ETEC vaccine development. In addition, as plasticity in MEFA construction allows inclusion of additional antigens for even broader protection, this MEFA approach perhaps can be generally used in effective vaccine development against diseases caused by other immunologically heterogeneous pathogenic strains or isolates.

Conclusions
Results from this study indicated CFA/I/II/IV-STa -toxoid -dmLT MEFA or CFA/I/II/IV MEFA combined with 3xSTa N12S -dmLT induced broadly protective anti-CFA and antitoxin immunity, and suggested their potential application in broadly effective ETEC vaccine development. This MEFA strategy may be generally used in multivalent vaccine development.