Local host response following an intramammary challenge with Staphylococcus fleurettii and different strains of Staphylococcus chromogenes in dairy heifers

Coagulase-negative staphylococci (CNS) are a common cause of subclinical mastitis in dairy cattle. The CNS inhabit various ecological habitats, ranging between the environment and the host. In order to obtain a better insight into the host response, an experimental infection was carried out in eight healthy heifers in mid-lactation with three different CNS strains: a Staphylococcus fleurettii strain originating from sawdust bedding, an intramammary Staphylococcus chromogenes strain originating from a persistent intramammary infection (S. chromogenes IM) and a S. chromogenes strain isolated from a heifer’s teat apex (S. chromogenes TA). Each heifer was inoculated in the mammary gland with 1.0 × 106 colony forming units of each bacterial strain (one strain per udder quarter), whereas the remaining quarter was infused with phosphate-buffered saline. Overall, the CNS evoked a mild local host response. The somatic cell count increased in all S. fleurettii-inoculated quarters, although the strain was eliminated within 12 h. The two S. chromogenes strains were shed in larger numbers for a longer period. Bacterial and somatic cell counts, as well as neutrophil responses, were higher after inoculation with S. chromogenes IM than with S. chromogenes TA. In conclusion, these results suggest that S. chromogenes might be better adapted to the mammary gland than S. fleurettii. Furthermore, not all S. chromogenes strains induce the same local host response.


Introduction
Coagulase-negative staphylococci (CNS) are the principal cause of subclinical mastitis in dairy cattle [1], especially in primiparous cows [2]. The impact of CNS on the udder health of dairy cattle has gained more attention in the past decade. The CNS were initially reported as one large, uniform group of bacteria [3,4]. However, thanks to recent advances in molecular identification techniques, individual CNS species have become easier to identify and study [5,6]. Over 20 different CNS species have been isolated from bovine milk [7]. Nonetheless, the bovineassociated CNS cover a wide range of ecological habitats, varying from essentially environmental species to hostadapted species [8]. Some CNS species rarely occur in bovine milk, but rather thrive in the environment of the cow and the barn (e.g. air, sawdust, bedding, and floors) [9]. These so-called environmental CNS species include, among others, Staphylococcus equorum and Staphylococcus fleurettii [8]. On the other end of the spectrum are the so-called host-adapted CNS species, specialized in survival in the udder and on the cow. Staphylococcus chromogenes is considered such a species [10], since it is the predominant CNS species found in milk [8,11]. Furthermore, S. chromogenes is also present on the teat apex [12,13], streak canal, and other extra-mammary body sites [14]. Staphylococcus chromogenes is one of the main CNS species involved in intramammary infections (IMI) [15,16]. In general, S. chromogenes causes a minor to moderate increase in the milk somatic cell count (SCC) [17], although one study noted a rise in SCC comparable to that of S. aureus infections [18]. On the other hand, teat apex colonization with S. chromogenes has also been associated with a lower quarter milk SCC in early lactating heifers [12], whereas some strains can even inhibit the growth of other mastitis pathogens in vitro [19].
Epidemiological data suggest that not all CNS species exhibit the same degree of pathogenicity [18], but little is known about the different host responses caused by an IMI with representatives of the supposed environmental or host-adapted species (or strains). Therefore, the first objective of the research was to examine the host response and bacterial shedding following an experimental intramammary inoculation in heifers with one distinctive hostadapted CNS species and another typical environmental one (S. chromogenes versus S. fleurettii). The second objective was to evaluate whether the elicited host response and bacterial shedding differs between strains belonging to the same species, in this case S. chromogenes.

Materials and methods
The study is in compliance with the European Directive 2010/63/EU, and was approved by the Ethics Committee of the Faculty of Veterinary Medicine, Ghent University (EC2012/73).

Animals
The experiment was performed at the research dairy farm of Ghent University (Biocentrum Agri-Vet, Melle, Belgium). Eight clinically healthy Holstein-Friesian heifers in mid-lactation (78-278 days in milk) were selected. Heifers with a previous history of clinical mastitis or persistent high SCC (>150 000 cells/mL) on Dairy Herd Improvement records were not included. To increase the likelihood that all quarters were free from IMI, the animals received 15 days before the start of the experiment 3 daily intramuscular injections of 10 g penethamate hydroiodide (Mamyzin, Boehringer Ingelheim GmbH) combined with an intramammary treatment of 200 mg cephalexin and 100 000 I.U. kanamycin (Ubrolexin, Boehringer Ingelheim GmbH) in each quarter for 2 days. The heifers were moved to a separate tie-stall barn 48 h prior to inoculation, and kept there until the end of the experiment (i.e. 78 h after inoculation). The heifers were milked twice a day, at 08:00 and 20:00 h. After milking, the teats were dipped with an iodine-based barrier dip (Io-Shield, Ecolab, Northwich, UK) and before sampling, the teats were cleansed with a lactic acid based foam product (Oxy-Foam D, Ecolab, Northwich, UK).

Study design
All heifers were challenged following a split-udder design [20,21]. Three quarters of each heifer were simultaneously inoculated with 1.0 × 10 6 colony forming units (CFU) of the bacterial strains (one strain per udder quarter) in 5 mL sterile phosphate-buffered saline (PBS) using a sterile polyvinyl chloride catheter of 18 cm. The remaining quarter was inoculated in the same manner with 5 mL sterile PBS (Thermo Scientific, Waltham, USA) and served as a control. All inocula were directly infused into the gland cistern. To ensure a balanced distribution between the quarter positions, the inocula were allocated to the quarters using restricted randomization. The heifers were examined clinically and their rectal temperature was registered at each sampling.

Inocula
Three different CNS strains were used in this experimental challenge (Table 1). Two field strains of S. chromogenes were used; one isolated from the teat apex of a heifer with no signs of mastitis (S. chromogenes TA) [19] and the other from the left hind quarter of a multiparous cow with a persistent IMI lasting over 300 days (S. chromogenes IM) [9]. Although no genotypic strain typing was performed, there were notable phenotypic differences between both S. chromogenes isolates (Table 1). For instance, the bacterial inhibitory capacities of the present strains were tested against a field isolate of S. aureus in a previously described modified cross-streaking culture [19]. Staphylococcus chromogenes TA is able to inhibit the growth of S. aureus, whereas S. chromogenes IM is not. Both strains elicit a different immune response in mice [22] and interact differently with mammary epithelial cells [23]. The third CNS strain used in this study was S. fleurettii [9]. The strains were initially stored at −80 °C (Microbank, Pro-Lab Diagnostics, UK). A growth curve was set up for each strain by incubating one colony in brain-heart infusion broth at 37 °C. The bacteria were collected during the late logarithmic growth phase. The bacteria were washed three times in sterile PBS by centrifugation at 4000 × g for 10 min. The pellet was resuspended in PBS with 15% (v/v) glycerol and stored at −80 °C. To confirm the viable bacterial count in the stock solution, serial dilutions were plated on tryptic soy agar (TSA, Oxoid, Basingstoke, UK). An infection dose of 1.0 × 10 6 CFU was selected based on the results of a preliminary challenge trial to induce subclinical mastitis rather than clinical mastitis [21].

Microbiology and somatic cell counts
The SCC was determined by a DeLaval Cell counter (DeLaval, Tumba, Sweden). The milk samples (10 µL) were plated in duplicate on aesculin-blood and MacConkey agar (Oxoid, Basingstoke, UK) according to the guidelines of the National Mastitis Council [24]. The plates were incubated at 37 °C in aerobic conditions, and examined after 24 and 48 h. Additionally, serial dilutions of the milk were plated in duplicate on TSA for counting the CFU. Every morphologically dissimilar colony type on TSA was collected, and subjected to Gram staining, together with catalase, DNAse and tube coagulase testing (i.e. the routine identification methods for CNS). One colony of each morphologically dissimilar Gram-positive, catalasepositive and coagulase-negative isolate was stored at −80 °C (Microbank, Pro-Lab Diagnostics, Richmond Hill, Canada) and subjected to transfer RNA-intergenic spacer PCR (tDNA-PCR) for further identification at the species level [5]. Only if the isolates could not be identified using tDNA-PCR, rpoB sequencing was carried out [25].

Apoptosis and necrosis of milk PMN
The apoptosis and necrosis of the milk PMN, considered an indirect indicator of their impaired functionality [26], were determined 24 h bi and at 0, 6, 12, 18, 24, 48 and 72 h pi by means of flow cytometry [27]. This dual staining technique with annexin V-fluorescein isothiocyanate (FITC) and propidium iodide (PI) differentiates the (early) apoptotic (FITC+/PI−) and necrotic (FITC+/ PI+) PMN from the intact, viable cells (FITC−/PI−). The raw data were acquired and analyzed with FACSDiva Software (BD Biosciences, San Jose, USA).

Statistical analysis
The entire data of one quarter were not included in the final analysis due to an elevated SCC at the moment of inoculation. The data of another quarter were also omitted from the analysis due to a naturally occurring IMI with Staphylococcus epidermidis, but only 12 h after the inoculation.
Linear mixed regression analysis was used to model the relationship between the inoculum (categorical variable: control, S. fleurettii, S. chromogenes IM and TA), the time of sampling (continuous variable: from 4 h until 78 h pi), the quadratic term of time of sampling (continuous variable) and the different outcome variables (bacterial count, SCC, % apoptotic milk PMN, and % necrotic milk PMN) (PROC MIXED, SAS 9.4, SAS Institute Inc.). The bacterial count was log 10 -transformed, whereas the % apoptotic PMN, % necrotic PMN, and SCC/µL underwent a natural logarithmic transformation to obtain a normalized distribution. Heifer and quarter were incorporated as random effects in every model to account for the correlated nature of the data. Compound symmetry was selected as a covariance pattern to correct for the clustering of multiple samplings per quarter. To emphasize the response after challenge, measurements prior to inoculation were not included in the analysis. The interaction between inoculum and time of sampling was tested each time, and kept in the model when significant. The significance level was set at P ≤ 0.05. For pairwise comparisons between the different bacterial strains and control, a Bonferroni adjustment was used.

Clinical parameters
None of the inoculated quarters showed physical signs of clinical mastitis during the trial. A short period of fever was observed in three heifers though (>39.5 and <40.6 °C) between 9 and 12 h pi.

Somatic cell count
The milk SCC was significantly higher in the quarters inoculated with any of the isolates compared to the control quarters (adjusted P < 0.01; Table 2). No significant difference was found between the SCC response in quarters challenged with S. fleurettii, and those challenged with either S. chromogenes strain (S. chromogenes IM and TA: adjusted P = 0.43 and P = 1.00). However, the SCC tended to be more pronounced in the quarters challenged with S. chromogenes IM than with S. chromogenes TA (adjusted P = 0.06). The evolution over time of the quarter milk SCC after challenge differed significantly between quarters (interaction inoculum x time of sampling: P < 0.01; Table 2). After 78 h, the SCC continued to decline in the challenged quarters ( Figure 1).

Bacterial shedding
The control quarters remained culture-negative throughout the trial. The bacterial shedding was significantly higher in the S. chromogenes TA-and S. chromogenes IM-challenged quarters than in the control quarters (adjusted P values < 0.01). It was significantly more pronounced in S. chromogenes IM than in S. chromogenes TA (adjusted P ≤0.01). The bacterial shedding was so low in the S. fleurettii-inoculated quarters, that there was no significant difference with the control quarters (adjusted P value = 0.25). The evolution over time of bacterial shedding differed significantly between quarters (interaction inoculum x time of sampling: P < 0.01; Table 2). In each cow, S. fleurettii was eliminated from the mammary gland within 12 h, whereas S. chromogenes IM and TA remained present for a longer period of time ( Figure 2). No CNS were found in the inoculated quarters after 144 h pi though.

Milk PMN apoptosis and necrosis
The highest proportion of apoptotic and necrotic PMN was found in the control quarters. Compared to the control quarters, only the S. fleurettii-and S. chromogenes IM-challenged quarters yielded significantly less apoptotic milk PMN (adjusted P value <0.01 and 0.02 respectively; Table 2). No significant difference was found in the number of apoptotic PMN between the control quarters and S. chromogenes TA (adjusted P value = 0.13).
Staphylococcus fleurettii was the only strain that resulted in significantly less necrotic PMN than in the control quarters (pairwise comparison: P = 0.03; Figure 3). No significant differences were found when comparing the different CNS isolates with each other in terms of PMN apoptosis or necrosis. The evolution over time of the proportion of apoptotic and necrotic PMN differed significantly between quarters though (interaction   Table 2). It should also be noted that the proportion of apoptotic and necrotic PMN was reduced in all quarters immediately before the inoculation, compared to 24 h in advance (Figures 3 and 4).

Milk IL-1β, IL-8 and TNF-α
The IL-8 and IL-1β ELISA had an intra-and inter-assay coefficient of variation (CV) of <6 and <15%, respectively. The TNF-α ELISA showed lower precision (intra-and inter CV: 13 and 28%, respectively). Since the inoculation  of the CNS strains generally elicited a minimal, or nondetectable cytokine response, the IL-1β, IL-8 and TNF-α levels were not statistically analyzed. Instead, descriptive statistics were generated. In the S. fleurettii-and S. chromogenes IM-challenged quarters, a low, transient IL-8 response (<30 pg/mL) was observed within 28 h after inoculation ( Figure 5). These quarters also showed limited, erratic TNF-α peaks (<10 ng/mL; Figure 6). Staphylococcus chromogenes IM caused a higher, long-lasting IL-1β response compared to the TNF-α response, starting at 12 h pi (Figure 7). Staphylococcus chromogenes TA did not evoke a detectable IL-8 response, but the strain caused a major increase in IL-1β (peaking at 300 pg/mL 60 h pi) in one heifer. This particular heifer was also the only animal that demonstrated a TNF-α reaction to S. chromogenes TA peaking at 60 h pi. Notably, the data of the S. fleurettii quarter of the same heifer were omitted from the analysis 12 h pi due to the occurrence of a spontaneous, natural infection in that quarter. Another heifer showed no discernable cytokine response in any challenged quarter. No cytokines were found in the control quarters.

Discussion
First, we wanted to compare the host responses in dairy cattle with subclinical mastitis caused by representative isolates of either an "environmental" or a "hostassociated" CNS species. For this purpose, S. fleurettii was selected as the environmental species, and S. chromogenes as the host-adapted species. The second objective was to study the host's reaction to different strains of the same species, in this case S. chromogenes. Due to the split-udder design, which partially circumvents variation between heifers, the host response could be studied using only a limited number of experimental animals. In contrast to other experimental CNS trials [28,29], all CNS species found in the milk samples were identified with tDNA-PCR, in addition to conventional culturing techniques. Through the molecular identification of the CNS species, a natural infection with S. epidermidis was detected. The data of this particular quarter were subsequently discarded from the analysis. Based on the changes in SCC and bacterial shedding, all quarters were successfully challenged with the different CNS strains, whereas the control quarters remained culture-negative throughout the entire trial. Staphylococcus fleurettii was fairly quickly (i.e. within 12 h) eliminated from the mammary gland, and the bacterial shedding was lowest for this species. The latter finding was also observed in an experimental trial in mice using the same CNS isolates [22]. Both S. chromogenes strains, on the other hand, persisted in the mammary gland for at least 3 days, accentuating the host-adapted nature of this species. Altogether, this is still a short period, considering that S. chromogenes IM was originally isolated from a cow suffering from a persistent IMI lasting over 300 days [9]. The bacterial count of S. chromogenes TA was significantly lower than S. chromogenes IM though. In fact, S. chromogenes IM seemed to be the only strain able to multiply in the mammary gland in the first 6 h after inoculation. The S. chromogenes IM strain tended to elicit a larger increase in SCC than S. chromogenes TA.
The lower cellular response in the S. chromogenes TAinfected quarters could possibly explain the slower bacterial clearance. These results might indicate a difference in pathogenicity and in vivo growth capacity between the  different strains. Interestingly, in the murine experimental trial, no differences in bacteriological shedding nor in neutrophil influx were observed between S. chromogenes IM and S. chromogenes TA [22].
The intramammary challenge had a significant effect on the apoptosis of PMN in milk, which was likely the result of an influx of young, activated PMN from the blood stream to the site of infection along with a potentially  delayed PMN apoptosis. Aging PMN eventually undergo apoptosis (or programmed cell death), impairing the functionality of the cells [30,31]. Delayed apoptosis might therefore contribute to a faster clearing of the bacterial infection [26]. However, PMN apoptosis and subsequent phagocytosis by macrophages is a physiological necessity to curb unbridled inflammatory responses that result in tissue damage [30]. Since the lactation stage and parity may affect the survival of PMN [26,32], only midlactation heifers were used in this experiment, allowing us to study the effect of the infection stage [33]. Staphylococcus fleurettii evoked the greatest decrease in PMN apoptosis during the first 48 pi, at least partly explaining the higher (albeit insignificant) SCC increase compared to S. chromogenes TA. Still, the PMN apoptosis was already decreased in all quarters right before the inoculation. The pre-inoculation drop in apoptosis might have been caused by stress [34] associated with the beginning of the experiment, although this could not be confirmed. A longer adaptation period (>48 h) in the tie-stall facility prior to the inoculation might have mitigated these findings. Pro-inflammatory cytokines, such as TNF-α and IL-1β, are involved in a plethora of immune functions on a local and systemic level (e.g. the endothelial adhesion of immune cells, induction of fever or the production of other cytokines) [35,36]. The further recruitment of PMN to the infection site is mediated by IL-8 and other chemokines [37]. In contrast to Escherichia coli infections, S. aureus mastitis does not evoke a significant IL-8 or TNF-α response, which might partially explain the chronic nature of S. aureus IMI [38,39]. As demonstrated in previous research [28], other CNS species (S. epidermidis and Staphylococcus simulans) appear to induce a clear pro-inflammatory response with TNF-α, IL-8 and IL-1β nonetheless.
In this study, S. chromogenes IM induced the largest overall pro-inflammatory cytokine response, starting with an increase in IL-8 at 9 h pi. The production of IL-1β occurred later at 12 h pi, but lasted longer (>78 h pi). The IL-1β response was less pronounced in the S. fleurettii-challenged quarters than in the S. chromogenes IM-quarters. The IL-8 response induced by these strains seemed to be smaller than the response described for S. epidermidis or S. simulans [28]. However, a higher infection dose was used in the latter study. Furthermore, the authors transformed the IL-8 data by multiplying it by 100, since they observed that the human IL-8 ELISA measures bovine IL-8 100-times less efficiently. That might explain why their IL-8 results fall in the ng/mL range, whereas our results are found in the pg/mL range. Also, we did not find a distinction between early (peaking at 12 h) or late (peaking at 30 h) IL-8 responders as seen in [28]. In fact, none of our animals displayed any IL-8 response after 30 h pi.
Staphylococcus chromogenes TA did not elicit any detectable cytokine response, except in one particular heifer. In a similar murine infection trial, some mice displayed a disproportionate reaction to S. chromogenes TA as well, resulting in a high IL-1β response and an intense clinical reaction [22]. Another heifer showed no cytokine response at all to any CNS strain. Everything considered, the pro-inflammatory cytokine response appeared to vary greatly between individual animals in our research.
As discussed in other research [40], the split udder design used in this experiment is founded on withincow comparisons. On the one hand, this study design partly circumvents high variation between individual animals, and reduces the needed number of research animals substantially. On the other hand, it assumes that all mammary quarters are completely separate anatomical entities within the udder. This is not necessarily the case, as demonstrated by previous research that illustrated that the immune response of the neighboring (uninfected) quarters is affected by an experimental challenge with major mastitis pathogens [41]. This should be kept in mind when interpreting the results of this study. Nonetheless, we were still able to demonstrate different host responses to each CNS strain while using the split-udder design.
None of the heifers in our study showed any local signs of clinical mastitis (such as clots in the milk or swelling of the quarters), although 3 animals experienced a brief bout of fever in the first 12 h after inoculation. In a similar challenge study with S. chromogenes, the heifers only developed very mild clinical signs of inflammation, even though the inoculation dose was more than double ours (2.1 × 10 6 versus 1.0 × 10 6 CFU) [29]. In the preliminary challenge trial conducted prior to this study [21], higher doses of the same S. chromogenes strain evoked a more pronounced immune response as opposed to lower doses with the appearance of mild clinical signs at a dose of 2 × 10 6 CFU. Observational research has demonstrated that approximately half of all cows with an intramammary S. chromogenes infection showed clinical signs of mastitis [11,42]. Notwithstanding a different infection dose or host-immune status, this could suggest that not all S. chromogenes strains are equally pathogenic. This is in accordance with S. aureus, where certain strains have also been linked to a more severe clinical outcome, with a reduced persistence in the mammary gland [43]. When extrapolating the current results to practice, it should be noted that natural infection doses might vary from doses used here, and that the experimental strains were directly infused into the teat cistern (as opposed to natural infections, where CNS have to overcome the teat barrier.) The results of our experimental trial with dairy heifers should be interpreted with caution, since the outcome of an IMI may vary with parity, lactation stage and other cow factors.
Even in case of large inoculation doses, bovine-associated S. chromogenes and environmental S. fleurettii strains trigger a similar, relatively mild local response. The environmental CNS species, S. fleurettii, evokes a profound cellular response in dairy heifers nonetheless, akin to the host-adapted species S. chromogenes. However, S. fleurettii is eliminated more rapidly from the mammary gland than S. chromogenes. This might indicate that certain bovine-associated CNS species, like S. chromogenes, are better able to withstand and thrive in the mammary gland. The present in vivo study also suggests that not all S. chromogenes strains exhibit the same degree of pathogenicity. The clinical outcome of a natural S. chromogenes mastitis might therefore not only depend on the infection pressure and the resistance of the cow, but also on the pathogenicity of the particular strain.