Semisynthetic Glycoconjugate Vaccine Lead against Klebsiella pneumoniae Serotype O2afg Induces Functional Antibodies and Reduces the Burden of Acute Pneumonia

Abstract

Carbapenem-resistant Klebsiella pneumoniae (CR-Kp) bacteria are a serious global health concern due to their drug-resistance to nearly all available antibiotics, fast spread, and high mortality rate. O2afg is a major CR-Kp serotype in the sequence type 258 group (KPST258) that is weakly immunogenic in humans. Here, we describe the creation and evaluation of semisynthetic O2afg glycoconjugate vaccine leads containing one and two repeating units of the polysaccharide epitope that covers the surface of the bacteria conjugated to the carrier protein CRM₁₉₇. The semisynthetic glycoconjugate containing two repeating units induced functional IgG antibodies in rabbits with opsonophagocytic killing activity and enhanced complement activation and complement-mediated killing of CR-Kp. Passive immunization reduced the burden of acute pneumonia in mice and may represent an alternative to antimicrobial therapy. The semisynthetic glycoconjugate vaccine lead against CR-Kp expressing O2afg antigen is awaiting preclinical development.

Figure 1

Glycoconjugate vaccine design and retro-synthetic analysis to obtain the epitopes. (A) Trisaccharide repeating unit of Gal-III from KPST258. (B) Synthetic trisaccharide repeating unit 1 and hexasaccharide 2 containing two repeating units, each connected to an amine linker. (C) Retrosynthetic analysis of trisaccharide 1 and hexasaccharide 2.

Scheme 1. Synthesis of Trisaccharide 1 and Hexasaccharide 2: (A) Assembly of Protected Trisaccharide 3: (B) Global Deprotection of 3

Galactofuranose partially cleaved during Pd catalyzed hydrogenation. With one-pot Birch reduction, trisaccharide 1 was obtained without cleavage of the Galf residue. (C) Synthesis of the modified galactofuranose building block 17. (D) Assembly of trisaccharide fragments 9 and 10 for the assembly of hexasaccharide target 2.

Figure 2

Determination of human antibody binding to synthetic O-antigens derived from O2afg serotype. (A) Synthetic glycans 1 and 2 were immobilized on a glass slide and sera from CRK-Kp infected patients, and healthy human were incubated with the antigens. Antibody binding was determined by using a glycan microarray assay, and the fluorescence intensity indicates the number of bound IgG antibodies. (B) Graphs showed the titer of bound human IgG antibodies to either O2afg trisaccharide 1 (on the left) or the hexasaccharide 2 (on the right). Both antigens showed positive antibody binders, but only the hexasaccharide showed a significant increase in fluorescence intensity, when compared to the negative control group. A trisaccharide containing the 6-deoxy hexose l-rhamnose (Rha) was used as human antibody positive control. The analysis was performed using the unpaired t-test of individual serum mean values from triplicate printed glycan spots. CRKP—carbapenem-resistant K. pneumoniae, MFI—mean fluorescence intensity in human sera evaluated by glycan array (left panel), and the mean fluorescence intensity of bound antibodies to O2afg synthetic hexasaccharide from different groups was compared (graph on the right). The sera at time point 0 h was collected as control of preinfection. Two-way ANOVA was used in the analysis. The data represent the mean of MFI values from triplicated printed spots of three individuals per group ±95% CI. **p < 0.01.

Figure 3

Rabbits immunized with CRM₁₉₇-2 produce antibodies that recognize both LPS epitopes 1 and 2. (A) Schematic schedule for rabbit immunization. Five individuals were immunized on day 0 with CRM₁₉₇-2 (1 μg antigen per dose) formulated in alum and boosted on days 14 and 28 with the same formulation. The long-term immune response was measured 11 days after a boost injection on day 133. Negative control group containing three individuals received PBS with alum. (B) Levels of IgG specific to the hexasaccharide 2. (C) Levels of IgG specific against the trisaccharide 1. (D) Levels of IgA specific to the hexasaccharide 2. (E) Levels of IgA specific antibody against the trisaccharide 1. The data represent the 95% CI distribution of five animals in the CRM₁₉₇-2 group and three animals in the PBS group. Two-way ANOVA was used for statistical analysis. MFI—mean fluorescence intensity measured by ELISA. *p < 0.05 and **p < 0.01.

Figure 4

Rabbit anti-O2afg IgG binding assay measured by flow cytometry and opsonophagocytic killing activity. (A,B) K. pneumoniae O2afg or O1 were incubated with sera from day 35 (diluted 1:100) from rabbits immunized with either glycoconjugate (CRM₁₉₇-2 group, red) or PBS (PBS group, blue). A negative control containing bacteria with fluorescent secondary antibody was used as reference to establish the threshold for positive binding (Figure S4). The positive bind quantification shown in the right panel (MFI) was based on the product of the percent of gated events that passed the fluorescence threshold and the median fluorescence of those events that passed the threshold. MFI—mean fluorescence intensity, K. pn—K. pneumoniae. The error bars represent the SD of three independent experiments. One-way ANOVA was used for statistical analysis. **p < 0.01. (C–H) Killing activity of antibodies present in the sera of rabbits was evaluated with K. pneumoniae expressing a capsule (C,D) and a mutant strain without capsule (F,G). The opsonic index represents the sera dilution for 50% of bacteria killing (E,H) and was based on a four-parameter linear regression of sera dilution curves from two independent assays performed in duplicate. The percentage bacterial killing data (D,G) were mean ± SD of colony-forming units (CFUs) reduction relative to negative control (sample lacking sera with complement and effector cells) of two independent assays performed in duplicate. Unpaired t-test was used in the analysis. *p < 0.05, **p < 0.01.

Figure 5

Passive immunization in a murine pneumonia model. (A) Mice were intranasally infected with 1 × 10⁸ of O2afg-CR-Kp, and 2 h later passively immunized with polyclonal sera of rabbits vaccinated with either synthetic anti-O2afg vaccine (CRM₁₉₇-2) or PBS vaccine (PBS). Lung, BALF, and blood were collected 48 h postinfection. Mice infected with PBS (PBSinf) but receiving CRM₁₉₇-2 polyclonal rabbit sera were used as negative control. (B) Lung (left), BALF (middle), and blood (right) were collected 48 h postinfection and analyzed for bacterial burden by quantifying CFU counting. (C) Myeloperoxidase (MPO) activity and lung permeability were measured in BALF of mice infected with CR-Kp 48 h postinfection. Levels of MPO was measured by its enzymatic activity against TMB substrate. Lung permeability was measured by the amount of total protein present in BALF. The data represent the mean ± SE of 12 mice per group. Mann–Whitney was used for the statistical analysis. **p < 0.01 and ****p < 0.0001. One mouse was excluded from the Kpn-infected PBS group due to death after narcosis, and one mouse outlier was excluded from the CR-Kp-infected CRM₁₉₇-2 group. (D) Lung sections stained with hematoxylin and eosin (H&E) from mice noninfected (PBSInf), mice infected and treated with polyclonal sera of rabbits immunized with O2afg vaccine (Kpn + CRM₁₉₇-2), and mice infected and treated with polyclonal of rabbits vaccinated with PBS (Kpn + PBS), 48 h postinfection. (E) Histopathology scores of the lungs of mice. From left to right: total cells in lungs; inflammation score of lungs; alveolar edema of lungs; and perivascular edema of lungs. The data represent the mean ± SE of five mice per group. Kruskal–Wallis was used for the statistical analysis. *p < 0.05 and **p < 0.01.