Corynebacterium and Dolosigranulum: future probiotic candidates for upper respiratory tract infections

The presence of the bacterial genera Corynebacterium and Dolosigranulum has consistently been associated with a healthy upper respiratory tract (URT). Commonly occurring together in the nasopharynx of healthy children, the role of these commensal organisms in nasopharyngeal health is unknown, as few studies have sought to determine whether they actively contribute to maintaining a healthy state. We recently identified Corynebacterium pseudodiphtheriticum and Dolosigranulum pigrum as the major nasopharyngeal species associated with resistance to recurrent ear infections, via 16S rRNA gene sequencing and metagenomics. Using in vitro bacterial interference assays, we observed a reduction in the growth of Moraxella catarrhalis – one of the three major otopathogens – in the presence of C. pseudodiphtheriticum. Further in vitro and in vivo studies of the interactions between commensal C. pseudodiphtheriticum and D. pigrum strains, URT pathogens, and the human host will help to clarify their role in nasopharyngeal health. If they play a protective role, these organisms are promising candidates for the development of a probiotic therapy for the treatment or prevention of URT diseases in children.

Human microbiome research has traditionally focused on differences in the composition of the microbial community between healthy and disease states. This research has naturally progressed into ‘mining the microbiota for therapeutics’; the process of identifying members of the microbiota that contribute to health and developing them into therapeutic agents, either in the form of probiotics (live, beneficial microorganisms) or isolated antimicrobial compounds. Several examples mirroring this process have already been demonstrated, where defined bacterial mixtures have successfully treated¹ or prevented² infection, or the production of an antimicrobial substance by a commensal organism inhibits a pathogenic close relative^3,4.

In the microbiome of the upper respiratory tract (URT), the genera Corynebacterium and Dolosigranulum have been consistently associated with health, particularly in children (Table 1). In these studies, they are frequently reported to co-occur, their presence corresponds with breastfeeding and a lower risk of respiratory infection, and they appear to be negatively impacted by antibiotic use. This suggests that these genera are important characteristics of a healthy nasopharynx. It may be hypothesised that the absence or reduction (by antibiotic use) of Corynebacterium and Dolosigranulum from the URT microbiota results in respiratory pathogens preferentially colonising the nasopharynx and an increased susceptibility to respiratory disease.

Table 1.  Studies observing the presence of Corynebacterium and Dolosigranulum in the URT microbiota and their association with a healthy state.

While it is clear that Corynebacterium and Dolosigranulum are strongly associated with health, little work has been done in vitro to characterise how they contribute to health. Do they compete with respiratory pathogens for nutrients and space, or prevent infection via bacterial antagonism? No previous studies appear to have directly tested D. pigrum for such activity. Of those that have investigated Corynebacterium, one early study reported C. pseudodiphtheriticum binding to pharyngeal cells, competitively excluding M. catarrhalis; though no zone of inhibition was observed on blood agar¹⁸. A more recent study reported that inoculating C. pseudodiphtheriticum strain 090104 into mice improved resistance to infection by respiratory syncytial virus and secondary pneumococcal infection, where non-viable 090104 had a weaker effect¹⁹. Recently, the commensal also demonstrated inhibition of S. aureus by exploiting its own virulence components²⁰. Other Corynebacterium species have shown similar promising antimicrobial activity; C. propinquum uses siderophores to restrict the availability of iron to coagulase-negative staphylococci²¹, and C. accolens inhibits S. pneumoniae via antimicrobial free fatty acids, which C. accolens requires for growth¹⁰.

Recently, we used 16S rRNA gene sequencing to characterise the nasopharyngeal microbiomes of children with recurrent ear infections (recurrent acute otitis media; rAOM), and healthy rAOM-resistant children¹⁷. The relative abundance of Corynebacterium and Dolosigranulum was significantly higher in the rAOM-resistant children, concordant with previous studies comparing the nasopharyngeal microbiota of children with and without OM^6,7,13. Metagenomics further revealed that C. pseudodiphtheriticum and D. pigrum were the dominant species in the healthy rAOM-resistant nasopharynx with C. propinquum and C. accolens present to a lesser extent²². To investigate whether these commensal organisms have a protective role in the nasopharynx interfering with the growth of pathogens, we conducted bacterial interference assays²². C. pseudodiphtheriticum and D. pigrum were assessed for inhibitory activity against the major otopathogen species (non-typeable Haemophilus influenzae – NTHi, Streptococcus pneumoniae and Moraxella catarrhalis); two strains of each commensal species and seven strains of each otopathogen were tested, including mixtures of both commensals together. Growing the pathogens adjacent to the commensals on agar revealed that neither commensal affected the growth of S. pneumoniae or NTHi, but all seven M. catarrhalis strains were inhibited by C. pseudodiphtheriticum (Figure 1). This effect was not influenced by the presence of D. pigrum.

Figure 1.  C. pseudodiphtheriticum inhibits M. catarrhalis on blood agar. The growth of M. catarrhalis (outer drops on all images) is inhibited by (a) C. pseudodiphtheriticum (inner drops) but not by (b) D. pigrum or (c) broth alone (negative control). The inhibitory effect is the same with (d) C. pseudodiphtheriticum and D. pigrum together.

To attempt to identify the production of an antimicrobial compound by C. pseudodiphtheriticum as a potential mechanism, we inoculated filter paper discs with cell-free C. pseudodiphtheriticum broth culture and cell lysates for a disc diffusion assay on M. catarrhalis. No inhibition was observed (Figure 2)²². It is possible that the inhibitory effect of C. pseudodiphtheriticum is not mediated by an antimicrobial substance, for example by competitive acquisition of nutrients. However, it is also likely that such a substance is produced, but was not successfully extracted from C. pseudodiphtheriticum culture in a functional state or at sufficient concentrations to produce a zone of inhibition, or its production requires a trigger (like the presence of M. catarrhalis). This preliminary work demonstrates that C. pseudodiphtheriticum inhibits the growth of M. catarrhalis; while this is the least common of the major otopathogens, it may be an important one to target given its apparent synergistic interactions with NTHi and S. pneumoniae^23,24. However, the mechanism of interference, the role of D. pigrum and whether Corynebacterium species can inhibit URT pathogens at a clinically useful degree is yet unknown.

Figure 2.  Cell-free extracts from C. pseudodiphtheriticum do not inhibit M. catarrhalis. The central disc contains 30 μg kanamycin (positive control). There are no zones of inhibition around the four other discs containing cell-free C. pseudodiphtheriticum extracts from either strain of the commensal (broth supernatants on discs 1 and 2; cell lysates on discs 4 and 5) or the two negative controls (negative broth on disc 3, negative lysate on disc 6).

The consistent correlation between Corynebacterium, Dolosigranulum and upper respiratory health presents a very promising avenue of research towards development of probiotic therapies for the URT. The recent studies demonstrating antimicrobial activity by C. accolens against S. pneumoniae¹⁰, C. pseudodiphtheriticum against both S. aureus²⁰ and M. catarrhalis²², and C. propinquum against coagulase-negative staphylococci²¹ suggest it is possible that multiple Corynebacterium species are beneficial to the URT, acting via different mechanisms against different pathogens. The role of D. pigrum in these interactions remains unknown, however it has been hypothesised that the production of lactic acid by D. pigrum lowers the pH of the local environment, selecting for the growth of Corynebacterium species²⁵. This is highly plausible, as several species of Lactobacillus inhibit M. catarrhalis via lactic acid²⁶, which also exhibits antimicrobial and anti-inflammatory properties in the human vagina²⁷. A niche-specific probiotic therapy containing multiple species with complementary effects may be an effective approach for the URT, especially for restoration after antibiotic use, which has been demonstrated to reduce Corynebacterium and Dolosigranulum populations (Table 1). Previous clinical trials using nasal spray probiotics for children with rAOM have suggested that pre-treatment with antibiotics allows re-establishment of the healthy microbiota, preventing the re-emergence of pathogens^28,29.

To fully investigate the possibility that these commensal organisms work synergistically to provide resistance to respiratory pathogens, and the potential for this to be translated into a therapy, there is an array of possibilities for future research. First, it remains likely that C. pseudodiphtheriticum produces an antimicrobial substance that inhibits M. catarrhalis. To isolate it, a range of commensal nasopharyngeal strains could be screened for optimal antimicrobial activity, and optimisation of culture conditions or enrichment of proteins may be required, similar to the approach taken by Latham et al. (2017) for the isolation of an antimicrobial substance from Haemophilus haemolyticus⁴. The commensal Corynebacterium genomes are not well characterised; a very recent sequencing effort by Stubbendieck et al. (2019) has improved the available genomic information for C. pseudodiphtheriticum and C. propinquum²¹, but these draft genome assemblies are still incomplete as of August 2019 (NCBI Genome Assembly and Annotation reports). There is now one strain of D. pigrum with a complete assembly (83VPs-KB5) available. Mining the genomes of multiple commensal strains for putative antimicrobial genes may provide further evidence for protective factors. Additionally, epithelial cell models have previously been successful in understanding the competition between H. haemolyticus and NTHi in the URT³⁰ and would also be useful to obtain a deeper insight into the competitive mechanisms and host responses, to see if colonisation with these organisms influences susceptibility to later infection with S. pneumoniae, M. catarrhalis or S. aureus.

Further research into the potential of Corynebacterium and Dolosigranulum to compete with respiratory pathogens, alone or in synergy, will provide a strong direction for the use of animal models and clinical trials in the development of a probiotic therapy, which has the potential to treat or prevent a range of upper respiratory illnesses in children.

The authors declare no conflicts of interest.