A class I Haemophilus ducreyi strain containing a class II hgbA allele is partially attenuated in humans: implications for HgbA vaccine efficacy trials

Haemophilus ducreyi causes chancroid and is a major cause of cutaneous ulcers in children. Due to environmental reservoirs, both class I and class II H. ducreyi strains persist in cutaneous ulcer-endemic regions following mass drug administration of azithromycin, suggesting the need for a vaccine. The hemoglobin receptor (HgbA) is a leading vaccine candidate, but its efficacy in animal models is class-specific. Controlled human infection models can be used to evaluate vaccines, but only a class I strain (35000HP) has been characterized in this model. As a prelude to evaluating HgbA vaccines in the human model, here we tested whether a derivative of 35000HP containing a class II hgbA allele (FX548) is as virulent as 35000HP in humans. In 8 volunteers infected at 3 sites with each strain, the papule formation rate was 95.8% for 35000HP vs. 62.5% for FX548 ( P = 0.021). Excluding doses of FX548 that were ≥ 2-fold higher than those of 35000HP, the pustule formation rate was 25% for 35000HP vs. 11.7% for FX548 (P = 0.0053). By Western blot analysis, FX548 and 35000HP expressed equivalent amounts of HgbA in whole cell lysates and outer membranes. The growth of FX548 and 35000HP was similar in media containing hemoglobin or hemin. By whole genome sequencing and single nucleotide polymorphism analysis, FX548 contained no mutations in open reading frames other than hgbA . We conclude that by an unknown mechanism, FX548 is partially attenuated in humans and is not a suitable strain for HgbA vaccine efficacy trials in the model.