Orally delivered toxin-binding protein protects against diarrhoea in a murine cholera model

Abstract

The ongoing seventh cholera pandemic, which began in 1961, poses an escalating threat to public health. There is a need for new cholera control measures, particularly ones that can be produced at low cost, for the one billion people living in cholera-endemic regions. Orally delivered V_HHs, functioning as target-binding proteins, have been proposed as a potential approach to control gastrointestinal pathogens. Here, we describe the development of an orally deliverable bivalent V_HH construct that binds to the B-pentamer of cholera toxin, showing that it inhibits toxin activity in a murine challenge model. Infant mice given the bivalent V_HH prior to V. cholerae infection exhibit a significant reduction in cholera toxin-associated intestinal fluid secretion and diarrhoea. In addition, the bivalent V_HH reduces V. cholerae colonization levels in the small intestine by a factor of 10. This cholera toxin-binding protein holds promise for protecting against severe diarrhoea associated with cholera.

Fig. 1. Blocking capacity of the V_HH constructs against the CTXB–GM1 interaction.

a The ability to block the CTXB–GM1 interaction for selected monomeric V_HHs at a molar ratio of 10:1 (V_HH:CTXB). The average blocking effect of each V_HH was calculated from a single measurement of technical duplicates normalized against a control mixture containing CTXB and a V_HH without specificity for CTXB. Error bars represent standard deviation. Source data are provided as a Source Data file. b The CTXB-blocking capacity of the monovalent BL3.1 in comparison to the bivalent (V_HH–(G₄S)₃–V_HH) BL3.2 at molar ratios of 1:1, 10:1, and 100:1 (V_HH construct:CTXB). A previously reported anti-CTX V_HH control and a negative V_HH control without specificity for CTXB were included as well. The average blocking capacity was calculated from duplicate measurements of technical triplicates. Error bars represent standard deviation. Source data are provided in a Source Data file.

Fig. 2. Thermal and simulated gastrointestinal stability profile of BL3.2.

a CTXB-binding activity of BL3.2 after incubation for up to 4 h in simulated gastric fluid (SGF; pH 1.2) or up to 5 h in simulated intestinal fluid (SIF; pH 6.8) at 37 °C. Binding activity was normalized against a BL3.2 control stored at 4 °C in phosphate-buffered saline (PBS) for 5 h. The average CTXB-binding activity is based on measurements from two experiments of one sample per condition (SGF, SIF, and PBS) analysed in technical duplicates. Error bars represent standard deviation. Source data are provided in a Source Data file. b Thermal denaturation temperature (T_m) of BL3.2 and CTX in PBS. Each sample was analysed in triplicates, with lines indicating the averages.

Fig. 3. Functional neutralization of CTX by BL3.2 in a human colon adenocarcinoma cell assay.

Reduction of intracellular cyclic adenosine monophosphate (cAMP) production in HCA-7 cells incubated with CTX (0.115 nM) pre-mixed with increasing concentrations (0.240–31.25 nM) of the bivalent V_HH construct BL3.2 (circles) or a non-specific bivalent V_HH construct control (triangles). Levels of intracellular cAMP were interpolated from a sigmoidal four parameter logistic cAMP standard curve (R² = 0.9543) based on triplicate measurements. Each data point represents an interpolated mean value from biological duplicates comprised of three technical replicates. Source data are provided in a Source Data file.

Fig. 4. Predicted interface between BL3.2 and CTX.

a The most stable conformational binding arrangement between BL3.2 (green) and CTXB pentamer (grey) based on ColabFold and molecular dynamics simulations. b The nine amino acid residues in the complementarity-determining region one (CDR1) (Asp30 and Asp31), CDR2 (Asp55 and Ser57), and CDR3 (Tyr102, Asn104, Ser105, Gln107, and Asp111) of the BL3.2 paratope predicted to be essential for CTX binding. c The 14 amino acids in the CTXB epitope identified in silico to be crucial for BL3.2 interaction, nine in the primary CTXB (His13, Asn14, Ser55, Gln56, His57, Asp59, Gln61, Trp88, and Lys91) and five from the adjacent subunit in the pentamer (Lys34, Arg35, Glu36, Ser55, and Gln56).

Fig. 5. Hydrogen–deuterium exchange mass spectrometry (HDX–MS) analysis of the CTX epitope of BL3.1.

a The level of HDX for CTXB alone and CTXB bound to BL3.1 with regards to two specific regions of the toxin: positions 29–38 and positions 50–56. The solid and dashed lines show the deuterium incorporation for CTXB and the CTX–BL3.1 complex, respectively. Each time point (1, 10, 40 and 100 min) was analysed in technical triplicates. Data are presented as mean values with standard deviation. b Surface representation of the CTXB pentamer. Shades of grey represent no HDX differences upon binding BL3.1. The level of HDX protection upon binding BL3.1 is indicated with different shades of blue. c A ribbon representation of a single subunit of CTXB with mapped HDX protection in shades of blue. A galactose molecule is shown in yellow, indicating the site of interaction with the intestinal cell receptor GM1.

Fig. 6. CTX-neutralizing capacity of BL3.2 in infant mice.

a Schematic representation of the two oral administrations of BL3.2 given to 5-day-old CD-1 mice (Charles River Laboratories, strain 022, mixed sex); 1 h before and 3 h after CTX administration. Created using BioRender. Laboratory, T. (2025) https://BioRender.com/c95v770. b Impact of the orally delivered V_HH construct BL3.2 on the severity of CTX-associated diarrhoea (weight loss) in infant mice. Mice were given two oral administrations of either BL3.2 (n = 5, 9 mg ml⁻¹) or bovine serum albumin (BSA) as a control (n = 5, 9 mg ml⁻¹); one 3 h prior to oral delivery of CTX and one 3 h after oral delivery of CTX. diarrhoeal onset (red) or diarrhoeal absence (white) was visually monitored up until the experiment was terminated, 9 h following CTX administration. Horizontal lines indicate median weight loss for BL3.2 (0.53%) and BSA (3.8%), and statistical significance (*p < 0.05) was calculated using the two-tailed Mann–Whitney U test (P = 0.0317). Source data are provided in a Source Data file. c Impact of the orally delivered bivalent V_HH construct BL3.2 on CTX-induced fluid accumulation (FA) in the small intestine of infant mice. Mice were given two oral administrations of either BL3.2 (n = 5, 9 mg ml⁻¹) or BSA as a control (n = 5, 9 mg ml⁻¹), as previously described. Statistical difference (**P < 0.01) between median FA ratio (horizontal line) for BL3.2 (5.5%) and BSA (9.8%) was calculated using the two-tailed Mann–Whitney U test (P = 0.0079). Source data are provided in a Source Data file.

Fig. 7. In vivo evaluation of BL3.2 against clinical isolate of V. cholerae.

a Schematic representation of the two oral administrations of BL3.2 given to 5-day-old CD−1 mice (Charles River Laboratories, strain 022, mixed sex); 1 h before and 5 h after orogastric challenge with a V. cholerae clinical isolate. Created using BioRender. Laboratory, T. (2025) https://BioRender.com/b75t379. b Weight loss of infant mice 22 h after orogastric inoculation with V. cholerae (2.8 × 10⁸ CFU), given two oral administrations of either BL3.2 (n = 6, 9 mg ml⁻¹) or bovine serum albumin (BSA) (n = 6, 9 mg ml⁻¹); 1 h before challenge with V. cholerae and 5 h after. Median for diarrhoea-induced weight loss (diarrhoea in red and no diarrhoea in white) determined to be 1.4% for BL3.2 and 10% for the BSA control. Statistical difference (**P < 0.01) estimated using the two-sided Mann–Whitney U test (P = 0.0022). Source data are provided in a Source Data file. c Fluid accumulation (FA) ratio (% of body weight) determined for infant mice challenged with V. cholerae (2.8 × 10⁸ CFU), given two oral administrations of either BL3.2 (n = 6, 9 mg ml⁻¹) or BSA (n = 6, 9 mg ml⁻¹); one hour before challenge with V. cholerae and five hours after. Median for FA ratio (diarrhoea in red and no diarrhoea in white) determined to 4.1% for BL3.2 and 8.7% for the BSA control. Statistical difference (**P < 0.01) estimated using the two-sided Mann–Whitney test (P = 0.0022). Source data are provided in a Source Data file. d The number of V. cholerae found in the small intestine of infant mice 22 h after orogastric inoculation with 2.8 × 10⁸ CFU (dashed line), given two oral administrations of either BL3.2 (n = 6, 9 mg ml⁻¹) or BSA (n = 6, 9 mg ml⁻¹); 1 h before challenge with V. cholerae and 5 h after. Average CFU per organ (solid line) determined to be 1.6 × 10⁸ CFU for BL3.2 and 1.8 × 10⁹ CFU for the BSA control. Statistical difference (**P < 0.01) estimated using the two-sided Mann–Whitney U test (P = 0.0022). Source data are provided as a Source Data file.