Listeria monocytogenes sequence type 1 is predominant in ruminant rhombencephalitis

Listeria (L.) monocytogenes is an opportunistic pathogen causing life-threatening infections in diverse mammalian species including humans and ruminants. As little is known on the link between strains and clinicopathological phenotypes, we studied potential strain-associated virulence and organ tropism in L. monocytogenes isolates from well-defined ruminant cases of clinical infections and the farm environment. The phylogeny of isolates and their virulence-associated genes were analyzed by multilocus sequence typing (MLST) and sequence analysis of virulence-associated genes. Additionally, a panel of representative isolates was subjected to in vitro infection assays. Our data suggest the environmental exposure of ruminants to a broad range of strains and yet the strong association of sequence type (ST) 1 from clonal complex (CC) 1 with rhombencephalitis, suggesting increased neurotropism of ST1 in ruminants, which is possibly related to its hypervirulence. This study emphasizes the importance of considering clonal background of L. monocytogenes isolates in surveillance, epidemiological investigation and disease control.

Additionally, L. monocytogenes was present in water tanks (22 %, n =7), feed bunk swabs (22 %, n =7), faeces (16 %, n =5) and commodity feed (12 %, n =4). In contrast, L. monocytogenes was detected in silage only on a single farm. The frequency of L. monocytogenes detection in farm samples was significantly higher in cattle farms (n =30) than in small ruminant farms (n =17; Fisher's test, p-value < 0.05). However, this effect was not observed when the frequency in a single environmental source was compared between cattle and small ruminant farms, possibly due to the small sample size per source. Finally, the frequency of L. monocytogenes detection was similar between control and case farms at the farm and sample level (Table. S4). In two of three farms, in which a clinical isolate was available, the ST of the clinical isolate matched the ST of the environmental and faecal isolate (ST1 and ST4). In four out of five farms where L. monocytogenes was isolated from faeces, the faecal isolate was of the same ST as the environmental isolate (ST4, ST70, ST399 and ST451).
The prevalence of L. monocytogenes in the farm environment has only been addressed by few studies [4][5][6] . We detected L. monocytogenes in the majority of ruminant farms, but in the minority of samples. The prevalence is consistent with previous studies 7, 8 challenging the view that L. monocytogenes is ubiquitous 9 . The prevalence was significantly higher in cattle farms than in small ruminant farms confirming the view that ecology of L. monocytogenes differs between cattle and small ruminant farms 8 . However, no difference in prevalence was observed between case and control farms. L. monocytogenes was most frequently detected in soil and matching STs were identified in soil and other materials suggesting that farm soil could represent a contamination source for other materials (Table. S4). In contrast, and in accordance with a previous study 10 L. monocytogenes was rarely detected in silage, which has been inculpated to be the main source of infection in farm ruminants 8,[11][12][13] . This may be related to the good quality of most of the analyzed silage in this study (data not shown). The prevalence in faecal samples was low (16% overall) but in the range of previously reported data [14][15][16] , challenging the view that ruminants act as amplification hosts and that faecal shedding is linked to silage feeding as silage was fed on most farms 17 . However, in all five farms where faecal samples were positive either feed bunk swabs, hay or water samples were also positive with the same ST, suggesting that faecal shedding is linked to contamination of feed and water.

The presence of L. monocytogenes is likely linked to farm management practices
Feeding and farm management practices were recorded in an observational cross sectional study based on a standardized questionnaire. These multiple choice questionnaires were handed out in each of the 32 visited ruminant farms and comprised a total of 100 different potential risk factors.
For identification of risk and preventive factors univariable logistic regression analysis was performed. The outcome was defined as the presence or absence of L. monocytogenes at the farm level (bivariable yes/no) and exposure was defined as the presence or absence of a given risk factor (bivariable yes/no). All risk factors presenting a p-value up to 0.20 were further analyzed in the multivariable logistic regression analysis. This was performed for each set of factors with a biological relevance, using a backward elimination procedure (removing the less significant factor) until all variables in the model were significant (p-value <0.05). Odds ratios, confidence intervals and the goodness of fit (likelihood ratio test) of each model were assessed, as well as searching for interaction terms, correlations and confounders. All statistics were produced with the R software (R Core Team (2014 OR CI = 0.01 to 0.63) and frequency of straw removal lower than once per week: coefficient = -2.0, SE = 1.0, 95% CI = -4.46 to 0.45, p-value < 0.05, OR = 0.12, 95% OR CI = 0.01to 0.63).
This analysis indicates that the presence of L. monocytogenes is likely linked to farm management practices. Deep straw housing significantly reduced the isolation of L. monocytogenes compared to straw, sand and sawdust. Similarly, a frequency of straw removal lower than once per week reduced the isolation of L. monocytogenes when compared to daily and weekly cleaning. In deep straw bedding heat is produced due to composting of the deeper layers, and the microbiome of straw-excrement mixtures is similar to the microbiome of maturing compost, wherefore growth conditions might be suboptimal for L. monocytogenes with the presence of competing microbiota 18 . Indeed, Firmicutes are much less abundant than other bacteria in deep straw litter compared to other housing systems 19 . The low frequency of cleaning may be an additional preventive factor as the composting effect in deep litter may require some time to occur. Furthermore, in these settings cleaning procedures might be more thoroughly performed.