Multidrug Tolerance and the Stringent Response in Diverse Clinical Isolates of Haemophilus influenzae

Background: Persistent bacterial infections constitute a significant burden on human health. While genetically encoded resistances are on the rise, a second less-studied but pervasive phenotype is common: the formation of multidrug-tolerant persister cells. Microbes embedded within the dense communities of biofilms become dormant, so they do not express the growth-promoting pathways typically targeted by antibiotics and thereby exhibit "multidrug tolerance" (MDT). Dormant cells exhibiting MDT are known as 'persisters'. A widely conserved pathway that signals dormancy is the "stringent response", which signals the downregulation of growth-promoting pathways and activates stress responses via production and breakdown of the second messenger guanosine (penta-)tetra-phosphate, or (p)ppGpp, via the RSH (RelA/SpoT homolog) family of enzymes. Haemophilus influenzae are responsible for high morbidity due to their role in respiratory infections, and their biofilms have been implicated in chronic and recurrent diseases of the middle ear and lungs. Clinical isolates of H. influenzae show widely varying phenotypes, but variation in biofilm-induced MDT and the contribution of the stringent response remain unknown. We hypothesized that diverse strains would present diverse biofilm and MDT phenotypes. We also hypothesized that the stringent response would contribute to MDT across diverse strains. Methods: To test for variation in MDT and other biofilm-related phenotypes, experiments used a 'diversity panel' of 12 strains selected to maximize genomic diversity but minimize the number of strains, as well as isogenic mutants of the lab strain Rd with ΔrelA and ΔrelA ΔspoT knockout mutations. Growth and viability assays were conducted for planktonic and biofilm cultures. Minimum inhibitory concentrations (MICs) were performed for 3 distinct antibiotics during planktonic growth, followed by biofilm-induced MDT assays of the 3 antibiotics at excess MIC. Experiments using TEM-1 positive ampicillin-resistant strains used treatment with clavulanate to inhibit β-lactamase activity. MDT assays were also conducted with or without the addition of novel small molecule HEJ14 (RelA inhibitor), designed to competitively inhibit the RelA active site and thereby block the stringent response. Results: Verifying the importance of the stringent response on biofilm-induced MDT, results using isogenic mutant derivatives of the laboratory strain Rd showed that deletion of relA or both relA and spoT resulted in 10 to 100-fold reduced biofilm-induced MDT compared to wild-type, though had little impact on planktonic or biofilm growth. Strains on the diversity panel showed ~100-fold variation in the viable cell densities within 24-hour static culture biofilms, high variability in MDT, and distinct tolerance to antibiotic classes. Treatment of biofilm cultures with the putative RelA inhibitor HEJ14 generally reduced MDT across the diversity panel, depending on antibiotic class, although some strains showed high potentiation by HEJ14 and others almost none. Surprisingly, the HEJ14 compound reduced MDT for wild-type Rd and relA and relA spoT mutants. Conclusion: This work demonstrates that the stringent response is at least partially responsible for starvation-induced MDT in Haemophilus influenzae, suggesting that stringent response inhibitors could be potent adjunct therapies to help clear persistent infections. However, evidence also indicates that other pathways contribute to MDT, and these contributions vary across strains. The surprising finding that HEJ14 reduces MDT even relA spoT mutants suggests that the molecule may instead inhibit a downstream effector of the stringent response that increases MDT. We propose a model for how HEJ14 acts and experiments for follow-up investigations.