A sublingual nanofiber vaccine to prevent urinary tract infections

Abstract

Urinary tract infections (UTIs) are a major public health problem affecting millions of individuals each year. Recurrent UTIs are managed by long-term antibiotic use, making the alarming rise of antibiotic resistance a substantial threat to future UTI treatment. Extended antibiotic regimens may also have adverse effects on the microbiome. Here, we report the use of a supramolecular vaccine to provide long-term protection against uropathogenic Escherichia coli, which cause 80% of uncomplicated UTIs. We designed mucus-penetrating peptide-polymer nanofibers to enable sublingual (under the tongue) vaccine delivery and elicit antibody responses systemically and in the urogenital tract. In a mouse model of UTI, we demonstrate equivalent efficacy to high-dose oral antibiotics but with significantly less perturbation of the gut microbiome. We also formulate our vaccine as a rapid-dissolving sublingual tablet that raises response in mice and rabbits. Our approach represents a promising alternative to antibiotics for the treatment and prevention of UTIs.

Fig. 1.. Sublingual nanofiber vaccine raises antibody responses against three B cell epitopes with broad expression across UPEC strains.

(A) Percentage of clinical UPEC isolates that contain the gene encoding the parent proteins of the pIroN, pIutA, and pIreA epitopes (2). Transmission electron microscopy images of nanofibers composed of (B) PEG-Q11pIreA, (C) PEG-Q11pIutA, (D) PEG-Q11pIroN, and (E) VAC-Q11, or (F) a coassembly of these four together. (G) Zeta potential measurements of nanofiber assemblies. (H) Mice were immunized sublingually with coassembled PEG-Q11 nanofibers containing the PADRE T helper epitope and either pIreA, pIroN, or pIutA B cell epitope, plus cholera toxin B (CTB) adjuvant. Mice were boosted at weeks 1 and 3, and serum immunoglobulin G (IgG) titer against the immunizing epitope was measured [two-way repeated-measures analysis of variance (RM-ANOVA), Tukey’s test, n = 4 per group]. (I) Effects of titrating T cell epitope content with sublingual nanofiber vaccines. Mice were immunized sublingually with PEG-Q11(pIreA/PADRE) nanofibers containing variable concentrations of PADRE plus CTB and were boosted at weeks 1, 3, and 6 (n = 4 per group). (J) Coassembly with the VAC epitope, but not PADRE, elicits antibody responses against pIutA and pIroN. Mice immunized with the pIutA or pIroN epitope and VAC are compared to responses with PADRE shown in (F). All mice were boosted at weeks 1 and 3, and formulations contained CTB adjuvant (two-way RM-ANOVA, Tukey’s test, n = 4 per group).

Fig. 2.. Fully coassembled nanofibers elicit polyvalent systemic and urinary antibody responses that specifically target UPEC.

Mice were immunized sublingually with either a mixture of three separately assembled nanofibers [PEG-Q11(pIreA/PADRE) + PEG-Q11(pIutA/VAC) + PEG-Q11(pIroN/VAC)] or a single fully coassembled nanofiber [PEG-Q11(pIreA/pIutA/pIroN)] and either CTB or cyclic di-AMP (c-di-AMP) adjuvant and were boosted at weeks 1, 3, 6, and 8. (A) Schematic depicting the epitope composition of the single- and three-nanofiber (NF) formulations. (B to D) IgG levels were measured individually against each of the three peptide epitopes. (E) To compare the overall response, a titer sum was calculated by arithmetic addition of the titers against each of the three epitopes (two-way RM-ANOVA, Tukey’s test, n = 5 per group). (F) Serum antibody isotype and IgG subclasses were measured by enzyme-linked immunosorbent assay (ELISA) against a 1:1:1 mixture of pIreA, pIutA, and pIroN. (G) Urinary antibody levels were determined by ELISA on undiluted urine samples (two-way ANOVA, n = 1 to 4 per group). (H) Vaccine-induced serum antibodies bound specifically to UPEC. Week 13 serum IgG titers were measured by ELISA against a UPEC strain (CFT073) or a nonpathogenic laboratory strain (BL21) (multiple t tests, Holm-Šídák correction, n = 4 to 5 per group).

Fig. 3.. Formulation with STING and TLR9 agonists promotes strong serum and urinary antibodies against UPEC without accompanying gut responses.

(A) Mice were immunized with PEG-Q11(pIreA/pIutA/pIroN/VAC) nanofibers and indicated adjuvants and boosted at weeks 1, 3, 6, and 10 (two-way RM-ANOVA, n = 5 per group). (B) Heatmap showing pIreA-, pIutA-, and pIroN-specific serum antibody responses at week 36. (C) IgG subclasses measured on serum diluted 1:1000 against 1:1:1 mixture of pIreA, pIutA, and pIroN. (D) Week 11 serum IgG measured by ELISA against UPEC strain (CFT073) or nonpathogenic laboratory strain (BL21) (multiple t tests, Holm-Šídák correction). (E and F) Urinary IgA and IgG at week 11 measured on undiluted urine against 1:1:1 mixture of pIreA, pIutA, and pIroN (one-way ANOVA, Dunnett’s multiple comparisons test against c-di-AMP only). (G) No antigen-specific fecal antibodies were observed after sublingual immunization. (H) Mice immunized in Fig. 3 (green arrows signify immunizations) were challenged intraperitoneally (I.P.) 26 weeks after final boost with 1 × 10⁸ CFU CFT073 cultured in Luria broth (pink arrow). (I) Lowest temperature recorded for each mouse (one-way ANOVA, Dunnett’s test versus unimmunized, n = 4 to 5 per group). (J) Body weight (wt.) over time. Bold lines represent mean of each group, and shaded boundaries indicate SEM. Weight and temperature curves for individual mice are in fig. S3. (K) Immunized groups were combined to show vaccine’s amelioration of weight loss versus unimmunized mice [two-way RM-ANOVA, n = 4 (unimmunized) or n = 17 (combined immunization groups)]. (L) Survival curve for unimmunized versus combined immunization groups (log-rank test).

Fig. 4.. A highly accessible tablet delivery vehicle enables sublingual immunization against UPEC epitopes in mice and rabbits.

(A) Schematic depicting tablet-making process. (B) Camera image of tablet vaccine (diameter, 5 mm). Mice were immunized with tablets containing PEG-Q11(pIreA/pIutA/pIroN/VAC) or PEG-Q11(pIreA/pIutA/pIroN; no VAC) and indicated adjuvants and boosted at weeks 1, 6, and 8. (C) Serum antibody responses against each UPEC epitope. (D) Serum antibody responses over time (two-way RM-ANOVA, Dunnett’s test against No VAC, n = 5 per group). (E and F) Urinary IgA and IgG in undiluted urine against 1:1:1 mixture of pIreA, pIutA, and pIroN (one-way ANOVA, Dunnett’s test against No VAC). (G) Experimental timeline. Immunizations are signified by green arrows; infection is signified by pink arrow. (H) Bold lines represent mean temperature, and shaded boundaries indicate SEM (two-way RM-ANOVA, Dunnett’s test against unimmunized, n = 5 per group). (I) Mice were sacrificed at humane endpoints or at 72 hours. Weight and temperature curves for individual mice are in fig. S4 (J) Contingency table depicting outcomes for unimmunized (n = 5) or immunized (n = 15) groups (Fisher’s exact test). (K) Camera image showing placement of tablet in rabbits. (L) Rabbits were immunized with vaccine tablets containing PEG-Q11(PIreA/PIutA/PIroN/VAC) or PEG-Q11(PIreA/IutA/IroN; No VAC), plus c-di-AMP and CpG adjuvants, and boosted at weeks 2, 4, 7, and 10 (n = 3 to 4 per group). (M) Urinary antibody responses in rabbits using undiluted urine against 1:1:1 mixture of pIreA, pIutA, and pIroN (two-way ANOVA, Šidák’s test). (N) Week 11 serum IgG against nonpathogenic BL21 E. coli or UPEC strain CFT073 cultured under normal or iron-limited conditions (two-way ANOVA, Tukey’s test).

Fig. 5.. Sublingual anti-UPEC vaccine is as effective as high-dose antibiotics at preventing transurethral infection, but without accompanying disruption of the microbiome.

(A) Mice (n = 10 per group) were immunized sublingually with PEG-Q11(pIreA/pIroN/pIutA) plus c-di-AMP and CpG and boosted at weeks 2, 4, 9, and 12 (green arrows). Control groups were unimmunized or immunized against irrelevant T and B cell epitope (OVA). After transurethral infection with 5 × 10⁷ CFU CFT073 (pink arrow), mice were sacrificed at 48 hours or humane endpoints. Unimmunized mice were given oral antibiotics (fosfomycin) at 1000 mg/kg daily for 3 days before and during infection (blue arrows). (B) Serum antibody responses in anti-UPEC vaccine group. Bold line shows mean titer, and faded lines represent individual mice. Titers against pIreA, pIroN, and pIutA are in fig. S6. (C) Serum IgG against OVA in mice given control vaccine. (D) Unimmunized mice raised no antibody responses. (E) Survival curve. (F) Mouse temperatures over time (two-way RM-ANOVA, Dunnett’s test against OVA vaccine). Bold lines show mean, and shaded areas represent SEM. (G) Body weight over time (two-way RM-ANOVA against OVA vaccine). Weight and temperature for individual mice are in fig. S7. (H to M) Feces collected before treatment (week 0) and week 13 for all groups plus week 13.5 for unimmunized groups (after 3 days of antibiotics) were used to compare effects of vaccination and antibiotics on microbiome. Richness sampling in fig. S8. (H) Estimation of microbiome’s OTU richness. (I) α-Diversity (one-way ANOVA, Tukey’s test). (J) Unweighted UniFrac principal coordinates analysis plot using β-diversity. (K) Family-level diversity. Rows are individual mice at week 13 (after vaccination) or week 13.5 (after antibiotic treatment). Full heatmap in fig. S9. (L) Overrepresented and (M) underrepresented species in antibiotic-treated mouse’s microbiome at the species level.