Mastitis, an inflammatory condition of the mammary gland, typically arises as a result of intramammary infections. In lactating dairy cattle, mastitis results in considerable monetary losses to the dairy industry due to depressed milk yield and quality, reduced reproductive performance, treatment, and pre-mature culling [1, 2]. Staphylococcus aureus is a Gram-positive, opportunistic, and contagious bacterium that is one of the most prevalent causes of mastitis in both humans and cattle [3]. Although infection can result in obvious clinical mastitis, S. aureus often evades the immune response resulting in a sub-clinical chronic infection that can persist for the life of the animal [4]. Classically, S. aureus is characterized as an extracellular pathogen [5]. However, an increasing number of studies elude to its ability to invade and survive within various host cell types, including mammary epithelial cells, neutrophils, and macrophages [6–8]; this may contribute to its persistence in various S. aureus-induced diseases [7]. Small colony variants (SCVs) of S. aureus may be particularly endowed for persistent survival by residing in host cells as they are characterized by slow-growth and atypical colony morphology, and are often over-looked by standard laboratory assays [9]. It has been previously demonstrated that SCVs can survive within professional and non-professional phagocytic cells [8, 10], are more persistent than typical S. aureus strains [11], and are difficult to treat with conventional antibiotics [9].

Given that different S. aureus strains appear to utilize unique infection and survival strategies, we hypothesized that the host immune response will also vary in accordance with the infecting strain. To elucidate potential differences in host responses to S. aureus infection, the present study was carried out to investigate the kinetic response of key cytokines in serum and casein-depleted milk, as well as serum haptoglobin over a 21-day period following intramammary infection with three strains of S. aureus. Additionally, a comparative proteomics platform was used to resolve the proteome of casein-depleted milk during acute and chronic phases of infection, and to identify differentially expressed proteins with potential immunomodulating effects. The S. aureus strains used for intramammary infection were: the SCV Heba3231, a naturally occurring SCV isolated from a dairy cow with history of chronic mastitis with atypical phenotypic properties [10], the 3231 parent strain displaying typical morphological and biochemical properties [10], and Newbould 305 (ATCC 29740), a teat skin strain that causes clinical bovine mastitis [12, 13] that is frequently used as a model for studying S. aureus-induced mastitis.

The current study aimed to monitor the host immune response to S. aureus intramammary infection using three unique strains, and to identify potential candidates for targeted therapies that may modulate the host defense to S. aureus challenge. Intramammary infection with different strains of S. aureus resulted in variable cytokine responses within the bovine mammary gland and circulation, supporting our hypothesis that different strains of S. aureus prompt the activation of host immune responses that are unique to the infecting strain.

Typically, an increase in milk somatic cell count (SCC) is indicative of an elevated and sustained inflammatory response caused by intramammary infection. During acute mastitis for example, neutrophil numbers in the mammary gland increase dramatically, constituting > 90% of milk somatic cells [14], whereas somatic cells in healthy glands contain < 22% neutrophils [15]. CXCL-8/IL-8 is the main tissue-derived chemoattractant for neutrophils [16, 17]; however, our observations demonstrated that the bovine IL-8 response did not always parallel the somatic cell response during the different S. aureus intramammary infections. Somatic cell scores were significantly elevated throughout the study period, and peaked during the acute phase (days 2-5 pi) in all challenge groups (reported by Atalla et al., 2009 [18]), however, we did not observe significant induction of IL-8 in the casein-depleted milk at any point during the study period, and it was only detected in the serum on days 14 and 21 pi in animals challenged with Newbould 305 and Heba 3231. The lack of an IL-8 response in the milk is consistent with two previously reported independent studies carried out by Bannerman et al. (2004) and Riollet et al. (2000b) [19, 20], suggesting that bovine IL-8 is not the only chemokine involved in trafficking neutrophils to the infected mammary gland. It is possible that other chemoattractants, such as complement cleavage factor 5a, IL-1β, TNF-α, CXCL1 and 3, and granulocyte-macrophage colony stimulating factor (GM-CSF) produced by epithelial cells, macrophages and T cells, also contribute to neutrophil trafficking to the mammary gland during intramammary infection [21–23]. For instance, neutrophil infiltration into the bovine mammary gland may be mediated by IFN-γ through the mechanism described by McLoughlin et al. (2008) [8] that involves CXC chemokines other than IL-8. Such chemokines may include CXCL1 and CXCL3, which are constitutively expressed in bovine milk [23]. In support of this, it has been demonstrated that IL-8 and GM-CSF are not active chemokines during lactation, whereas other factors such as IL-1β and TNF-α are active during any stage of lactation [22, 24].

In the present study, IFN-γ was found to be highly expressed both locally, as indicated by levels in the milk, and systemically during days 1-6 pi in animals challenged with Newbould 305, which parallels previously reported results by Bannerman et al. (2004) [19]. Serum IFN-γ levels were significantly up-regulated in all challenge groups at day 21 pi compared to day 0. This finding may be associated with the promotion of an enhanced cell-mediated immune response that includes neutrophil and macrophage activation.

During the chronic phase of infection, differences in serum TGF-β1 expression between the three challenge groups were observed with notable increases being observed for the Parent 3231 and Heba 3231 groups. Milk TGF-β1 did not significantly differ in response between any of the challenge groups; however, a trend towards altered TGF-β1 expression in animals challenged with Newbould 305 was observed, whereby the cytokine levels reached a maximum at 48 hours pi (approximately 4000 pg/ml, data not shown). This observation is in line with a previous report of milk TGF-β1 kinetics upon Newbould 305 challenge in dairy cattle [25].

The role of TGF-β in the host response to intramammary infection is uncertain because TGF-β is a pleiotropic immunoregulatory cytokine, with both pro- and anti-inflammatory functions, depending on the location and activation state of the cells that it is interacting with [25]. This is exemplified by the cytokine's dual actions on macrophages. For instance, TGF-β suppresses tissue macrophage function and inflammation by down-regulating the production of chemokines and cytokines such as IFN-γ [26], up-regulating the expression of IL-1 receptor antagonist [27], and suppressing the production of cytotoxic reactive oxygen and nitrogen intermediates [28, 29]. In contrast, TGF-β has also been shown to promote inflammation by inducing chemotaxis in human peripheral blood monocytes [30, 31], enhancing their ability for transendothelial migration to sites of infection [32], and up-regulating the transcription of pro-inflammatory cytokines such as IL-1β and TNF-α. Furthermore, TGF-β has been shown to be an important factor for the differentiation of a subset of effector Th cells known as Th17 cells, which elicit inflammatory responses during certain pathogenic infections that are not sufficiently dealt with by Th1 or Th2 immunity [33]. Lastly, regulatory cells including Tregs and M2 macrophages also produce TGF-β in humans. These cells help to prevent the deleterious effects of prolonged or excessive host immune responses [34, 35], and this process includes suppressing various components of the immune response such as antigen presentation by antigen presenting cells [36], and promoting tissue repair by increasing the deposition of extracellular matrix at the site of tissue injury, and promoting angiogenesis [37]. Since TGF-β was not observed locally in the present study, it is unlikely that it was involved in repair of the mammary gland epithelium.

There is evidence to suggest that S. aureus may be able to modulate the host immune response in part by promoting TGF-β expression; this could be a mechanism utilized by the Heba 3231 strain for example, to promote its survival within the host. In support of this, Staphylococcal superantigens (SAgs) have been demonstrated to induce the production of TGF-β by initiating the differentiation and expansion of Tregs that down-regulate the inflammatory response [38–40]. Indeed, TGF-β has been implicated in playing a role in immunosuppression by up-regulating the expansion and/or activation of anergic regulatory T cells in the presence of S. aureus SAgs, which may disrupt class II MHC antigen presentation by APCs [41], limiting cell-mediated immunity. Interestingly, this immunomodulatory mechanism is also employed by Leishmania as a means to suppress cell-mediated immunity and promote its intracellular survival [42, 43]. Lastly, it should be noted that TGF-β expression may not necessarily denote its level of bioactivity, since a study by Kehrl et al. (1986) [44] demonstrated that human B cells largely express inactive TGF-β in response S. aureus Cowan strain I.

Overall, haptoglobin concentrations increased with time in all challenge groups during the first 6 days of the study period, and declined to initial levels by days 14 and 21 pi. However, no significant differences in serum haptoglobin levels were observed between the challenge groups. This may have been due to the chosen sampling times, as no samples were taken between days 6 and 14 pi. It has been previously demonstrated that haptoglobin is up-regulated during the acute phase response, and that it may be used to discriminate between acute and chronic inflammation in cattle [45, 46]. The findings in the present study indicate that overall serum haptoglobin levels were not significantly different between any of the challenge groups, suggesting that haptoglobin is a sensitive marker of inflammation due to S. aureus but it lacks specificity to discriminate the host response to the three different strains of S. aureus.

It was previously demonstrated that two common mastitis-causing pathogens, Escherichia coli and S. aureus, elicit different immune responses and pathogenesis [19]. Here, we demonstrate that the variability in immune protein response is also strain-specific, at least in the context of S. aureus. Further studies investigating the profiles of other cytokines will add to the results of the current study. Studying IL-17 and IL-6 for instance would provide insight into the involvement of Th17 cells in S. aureus infections, as IL-17 is a Th17 cytokine and IL-6, along with TGF-β supports the differentiation of this Th subtype [47–49]. Furthermore, IL-6 is one of the major signals for the induction of the acute phase response [50], and may therefore compliment the present haptoglobin data. Monitoring chemoattractive factors, especially those under the regulation of TGF-β (i.e. IL-1β and TNF-α), would help delineate the pro- or anti-inflammatory role of TGF-β in intramammary infection with different strains of S. aureus. Although it would have been interesting to measure IL-1-β and TNF-α as well as other cytokines such as IL-4, IL-6, IL-12, and IL-10 that regulate the differentiation of these T-cell subpopulations, we had limited resources to carry out the cytokine analysis, and felt that inclusion of these cyctokies was beyond the scope of the current study as the aim was to note differences in host responses to different S. aureus strains, not to determine whether TGF-β is a pro- or anti-inflammatory cytokine.

In addition to seeing unique cytokine profiles in response to the different strains of S. aureus, the present study also found a number of milk proteins to be differentially expressed at various time-points in response to infection, including host defense proteins. Staphylococcus aureus insult to the mammary gland causes an inflammatory response that is characterized by an increase in vascular permeability, allowing exudates containing cells and proteins from the vasculature to enter the mammary gland [50]. Additionally, resident and recruited cells produce a number of proteins that target the pathogen and contribute to the restoration of homeostasis [50]. Therefore, the proteome of the milk may drastically change in response to pathogenic stimuli, and analysis of these changes may offer insight into the host response during mastitis. Indeed, the present study demonstrated changes in the milk proteome during intramammary infection with S. aureus. In order to assess protein expression in milk during the acute and chronic phases of infection, samples were evaluated on days 2 and 14 pi, respectively. Most proteins were up-regulated following Newbould 305 and Parent 3231 challenges, especially on day 2 pi during the acute phase of infection. On day 14 pi, the number of up-regulated proteins was dramatically reduced in the Newbould 305 and Parent 3231 challenge groups, as compared to day 2 pi. Results from this study indicate that differences in milk protein expression also occur as a result of intramammary infection with different strains of S. aureus. On day 2 pi for example, challenges with Newbould 305 and Parent 3231 induced 12 milk proteins each, relative to day 0. In contrast, significant changes in milk protein expression were not observed in response to challenge with Heba 3231 at this time. On day 14 pi however, 8 milk proteins were induced by challenge with the Parent 3231 strain and only 1 protein was induced by Newbould 305, whereas 8 proteins were significantly induced by the Heba 3231 challenge. These results are indicative of the acute pathogenic nature of the Parent 3231 and Newbould 305 strains and the unique phenotypic properties of SCVs, which may elicit a milder host response [18]. They suggest a highly active initial host response to challenges with the Parent 3231 and Newbould 305 strains, then perhaps resolution or establishment of chronic infection by day 14 pi. This corresponds with the somatic cell score data from these same cows presented by Atalla et al. (2009) [18], whereby the number of somatic cells in the milk during the chronic phase of infection still remained significantly up-regulated relative to day 0. Interestingly, we were not able to detect an initial host response to the SCV strain, but the induction of protein expression was demonstrated on day 14 pi. This delayed and relatively low level of host response corresponds to the ability of the Heba 3231 strain to avoid recognition and persist within host cells [10]. Future studies should include later time points in order to determine if Heba 3231 continues to up-regulate milk protein expression.

Two of the five spots corresponded to the most abundant milk proteins - β-casein precursor and β-lactoglobulin. This finding is likely a result of incomplete depletion of these highly abundant proteins. Nonetheless, the depletion protocol was found to be sufficient to yield a comprehensive view of the milk proteome. Of the five protein spots that were subjected to identification by LC-MS/MS, only one protein, CPP3 also known as lactophorin, was implicated in host defense. Component 3 of the proteose peptone is a phosphoglycoprotein uniquely expressed in the mammary gland of lactating dairy cattle and can be found in bovine milk whey. It has been previously suggested to have several functions that include antimicrobial activity, inhibition of lipolysis, mitogenic activity, and immunostimulation [51–54]. It consists of seven polypeptide components, ranging in molecular weight from 17 to 67 kDa [55, 56]. In the 2DE gels, it occurred at three different locations ranging in mass from 15-21 kDa and isoelectric point (pI) from 6-7 (Figure 3), which may correspond to its various components. In one study, the antimicrobial activity of CPP3 was investigated using a synthetic peptide called lactophoricin consisting of the 113-135 region of C-terminal of CPP3 [51]. Lactophoricin interacts with membrane phospholipids and forms voltage-dependent channels. Thus, it has been hypothesized that it may also be involved in the pore forming of natural lipid bilayers such as bacterial membranes [51, 57]. Indeed, Campagna's results demonstrated that this peptide has inhibitory-growth effects on a number of Gram-positive and Gram-negative bacteria, with a more pronounced inhibitory effect on Gram-positive bacteria, including S. aureus [51].

The present study demonstrates the complexity of the host immune protein response to various strains of S. aureus. Intramammary infection with different strains of S. aureus, elicited unique host immune protein responses, as indicated by differential expression of the cytokines IL-8, TGF-β, and IFN-γ in circulation and IFN-γ in casein-depleted milk. In general, cytokines displayed heightened responses in the serum compared to the casein-depleted milk. Interestingly, the SCV strain, Heba 3231, showed increased levels of serum IL-8, IFN-γ, and TGF-β occurring during the later stages of infection (day 21 pi). Whether this pattern of induction continues in later time points has yet to be determined. Using a proteomic approach, we have also demonstrated that intramammary infection with three different strains of S. aureus elicits differential protein expression in bovine milk during the acute and chronic phases of infection. Intramammary infection with Newbould 305 and Parent 3231 strains resulted in pronounced induction of protein expression on day 2 pi, while the Heba 3231 strain elicited no protein induction at this time. On day 14 pi, host protein activity was less pronounced for animals challenged with Newbould 305 and Parent 3231 strains, while some induction occurred in animals challenged with Heba 3231. Lastly, CPP3, a potential antimicrobial peptide that is uniquely expressed in the bovine mammary gland during lactation, appeared on our 2DE gels as three different spots, which were differentially expressed between the various S. aureus treatment groups. These results demonstrate that different strains of S. aureus affect the proteome of the milk in different ways. Of specific interest was the unique response to the SVC, Heba 3231 on day 21 pi. Therefore, investigating these changes in protein expression at later stages of infection is warranted. Overall, this study begins to shed light on the milk proteome and the changes that are elicited by intramammary infection with various strains of S. aureus and highlights an antimicrobial peptide, CPP3, which may have novel therapeutic potential.