Identification of an Antagonistic Probiotic Combination Protecting Ornate Spiny Lobster (Panulirus ornatus) Larvae against Vibrio owensii Infection

Evan F. Goulden, Michael R. Hall, Lily L. Pereg, and Lone Høj PLoS One. 2012; 7(7): e39667.

Diseases of marine animals, particularly farmed shrimp, shellfish and fish, are an increasing priority from a somewhat off-beat veterinary health perspective. From my perspective, this is a nice system for probiotic development because it is so obvious that yogurt isn't going to be the probiotic of choice. This opens up the choice of organisms more broadly; and in this case, marine isolate collections are the obvious source of probiotic strains to prevent infection with a pathogen.

In the formulation of a probiotic, resource competition and active antagonism are two major mechanisms by which the probiotic may prevent the pathogen from infecting a sensitive host (immunostimulation, anti-virulence, and direct predation are other strategies which have been investigated; in some sense, immunostimulation is actually a live vaccine so I wouldn't put it in the same category). The advantage of resource competition as a strategy is relative robustness. If the probiotic (single strain or multi-strain) fills a niche required for pre-infection colonization, it is likely difficult for the pathogen to evolve around that requirement or outcompete the probiotic. Antagonism, on the other hand, tends to be brittle; evolution to resistance is straight-forward in a number of well-characterized cases (including several bacteriocins). The selection to resistance is strong and the path typically involves loss or truncation of a receptor.

The attempt to formulate antagonistic Escherichia coli probiotics in enteric diseases has hit dead ends historically as a result of resistance. However, attempts in Streptococcus mutans (for caries prevention) has been successful enough to lead to a marketed product (Oragenics Inc). In this current paper, a large number (~500) of marine isolates were screened for antagonism against a type strain of the pathogen. 500 isolates may not seem like much to high-throughput screeners for small molecules; but the handling and management of that many environmental isolates and the execution of the assay on that scale could be significant. These isolates were screened against a single strain of the pathogen.

This is a point of criticism. The routine underestimate of microbial, particularly pathogen, diversity, is a very serious blind spot. Many researchers falsely presume that a single model strain can reasonably represent the pathogen population for many assays. Even worse, researchers may presume that some set of traits makes a model 'relevant' or creates a 'bad bug' - a priority organism - and that diversity away from the model is irrelevant. In this case, strain V. owensii DY05 (http://www.ncbi.nlm.nih.gov/pubmed/22307306, http://onlinelibrary.wiley.com/doi/10.1111/j.1574-6968.2009.01850.x/full) is a recognized lobster pathogen which induces high mortality. However, despite there being some other known strains which are pathogenic in prawns, the apparent known diversity of lobster pathogens appears minimal. The prawn pathogens which are related to the lobster pathogen are not particularly pathogenic in the lobsters and mention of a second lobster pathogen strain is not made in papers addressing the phylogeny of these Vibrio (http://eprints.jcu.edu.au/23845/; http://www.ncbi.nlm.nih.gov/pubmed/22055753). As a result, a therapy which addresses this single pathogenic strain, DY05, may do so for reasons which are unrelated to the general model for lobster pathogenesis. This may create a brittle therapy. The appropriate solution is to find a range of lobster pathogenic strains and do a screen against all of them.

The antagonism/co-culture assays were conducted both in isolation (agar well method) and in biofilms; a sound experimental practice; the subsequent screen eliminated potential pathogens of the lobsters. The biofilm assay was crude - crystal violet microwell assay - but traditional. Given that there is only one pathogen strain included, it seems an appropriate assay.

Members of 7 genera antagonized DY05; drawn from new marine isolates as well as an established marine strain collection. From the 91 isolates of these 7 genera, ease of culture and relationship to putative human pathogens restricted the total list to 16. It seems that this screen may be overly aggressive at this stage of probiotic development; but understandably convenient. From the remaining 16, the multispecies biofilm study eliminated 6 strains as potentially enhancing pathogen biofilm formation, through unknown mechanisms. Exclusion, competition and displacement studies were each done; antagonistic strains performed these tasks differently.

Four strains were advanced into live protection/treatment trials. One thing that was skipped: testing whether the strains antagonized each other... so it isn't clear exactly why the four together are less efficacious either of the two best, or a pair comprised of those two, but it may be that one of the four antagonizes the others. Regardless, they found a pair which reduced the mortality impact of pathogen exposure to statistically undetectable. This is an unquestionably impressive result; but it also depended on a relatively dense inoculum of the protective organisms. The authors suggest that this observation is normal in the literature; the claim requires more investigation (that is, more reading to do...).

This paper is part of a growing field for multi-species probiotics, exemplified by the Clostridium difficile work at the Sanger Institute (http://www.plospathogens.org/article/info%3Adoi%2F10.1371%2Fjournal.ppat.1002995). This paper did not engage with the microbiome/metagenomic sequencing field directly, as the C. difficile research did, but could probably benefit from engagement with those methods. Perhaps that is yet to come, given the development in this area, led by the Høj group.