Inflammatory cell infiltration as an indicator of

Veterinary Immunology and Immunopathology 104 (2005) 163–169 www.elsevier.com/locate/vetimm

Inflammatory cell infiltration as an indicator of Staphylococcus aureus infection and therapeutic efficacy in experimental mouse mastitis Eric Brouillette, Gilles Grondin, Brian G. Talbot, Franc¸ois Malouin* Centre d’E´tude et de Valorisation de la Diversite´ Microbienne (CEVDM), De´partement de Biologie, Faculte´ des Sciences, Universite´ de Sherbrooke, 2500 Boul, Universite´, Sherbrooke, Que., Canada, J1K 2R1 Received 18 May 2004; received in revised form 5 October 2004; accepted 5 November 2004

Abstract Staphylococcus aureus intramammary colonization of the mouse mammary gland induces migration of polymorphonuclear neutrophils (PMNs) similar to that observed during bovine mastitis. In the present study, a method combining acridine orange staining, fluorescence microscopy and computer-assisted image analysis has been developed to quantitate PMN infiltration in a mouse model of mastitis. This was carried out using paraffin embedded sections, and using this method, we showed that the presence of PMNs increased with the number of bacteria present in tissues. Nearly 400 and 1100 times more PMNs were counted in the mammary gland tissue after 12 and 24 h of infection, respectively, compared to mice infected for 6 h. Treatment with the antibiotic cephapirin at 10 or 25 mg/kg reduced PMN infiltration by 71 and 85%, respectively. In conclusion, this method can be used to quantitate PMN infiltration as a marker of inflammation and bacterial burden in infected tissue sections. # 2004 Elsevier B.V. All rights reserved. Keywords: Neutrophil; PMN; Mastitis; Staphylococcus aureus; Acridine orange

1. Introduction Microbial infection of bovine mammary gland induces inflammation and causes mastitis. This inflammation reduces the quantity and quality of the milk produced by lactating cows. In the dairy industry, the Abbreviations: PMN, polymorphonuclear neutrophil; CEP, cephapirin; IMI, intramammary infection * Corresponding author. Tel.: +1 819 821 8000x1202; fax: +1 819 821 8049. E-mail address: [email protected] (F. Malouin).

level of inflammation in the bovine mammary gland is evaluated by counting the number of cells that are present in milk. These cells are macrophages, lymphocytes, epithelial cells, and during acute mastitis, almost exclusively polymorphonuclear neutrophils (PMNs) (Paape et al., 2000). The cell counts, referred to as the somatic cell counts (SSC), are used to fix the commercial value of the milk. The mouse model of Staphylococcus aureus mastitis has been used to study bacterial pathogenesis and its treatment (Reid et al., 1976; Anderson and Holmberg, 1977; Anderson, 1979; Craven and

0165-2427/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.vetimm.2004.11.006

164

E. Brouillette et al. / Veterinary Immunology and Immunopathology 104 (2005) 163–169

Anderson, 1983; Jonsson et al., 1985; Bramley et al., 1989; Mamo et al., 1991; Sanchez et al., 1994; Kerr et al., 2001 and Brouillette et al., 2003). S. aureus intramammary infection (IMI) in the mouse is similar to that in the cow in that it induces inflammation and PMN infiltration of mammary tissue. However, in contrast with cows, milk collection from mice for somatic cell count analysis is a difficult and timeconsuming procedure. In order to use mouse mammary gland sections for the evaluation of tissue alterations and inflammation during mastitis, an efficient method for quantification of PMN infiltration is necessary. Until now, it has only been possible to carry out semi-quantitative observations to evaluate PMN infiltration in tissue sections (Owens et al., 1992; Guhad et al., 2000). In the present study, the staining of paraffin embedded tissue sections with acridine orange permitted differentiation between infiltrated inflammatory cells and resident epithelial mammary cells. Acridine orange is a stain that can be visualized with a fluorescence microscope equipped with appropriate filters. By the use of computer-assisted image analysis software, huge numbers of inflammatory cells were counted over time after bacterial infection and following antibacterial treatment. The method demonstrated that the presence of PMNs varied with the number of bacteria present in tissue.

2. Materials and methods 2.1. The mouse model of S. aureus mastitis Experimental conditions used here for the mastitis model were previously optimized for S. aureus Newbould and antibiotic treatment (Brouillette et al., 2004). In brief, 1 h following removal of 12– 14 day-old offspring, lactating CD-1 mice (Charles River, St-Constant, Que., Canada) were anesthetized and mammary ducts were exposed by a small cut at the ends of the teats. Under a binocular microscope, a 100 ml-bacterial suspension containing 1.4 to 3 102 CFU of S. aureus Newbould 305 (ATCC 29740), in endotoxin-free PBS (Sigma), was injected through the teat canal using a 33-gauge blunt needle. For each animal, two glands (fourth on the right and fourth on the left) were inoculated and aseptically

harvested at the indicated times. Control animals were inoculated with PBS instead of bacteria. For bacterial CFU counts, 4–5 mice (i.e. 8–10 mammary glands) were used for each condition tested, whereas two glands from two mice were used for microscopic analysis. The tissues used for CFU counts were homogenized in 2 ml of PBS and the bacterial content was evaluated by serial logarithmic dilutions on agar. The detection limit was 100 CFU/g of gland. In experiments implicating antibiotic treatment, cephapirin (CEP), which is widely used to treat S. aureus mastitis in cattle, was administered at t = 0 and t = 10 h by i.m. injection at 10 or 25 mg/kg in PBS (controls were treated with PBS). Infection was allowed to proceed for 24 h. The guidelines of the Canadian Council on Animal Care (Olfert et al., 1993) were respected during all the procedures. 2.2. Tissue preparation The mammary glands used for microscopy were fixed in 4% formaldehyde in PBS for 24 h at room temperature and embedded in paraffin wax. Fivemicrometers thick sections were stained with acridine orange as previously described (Humanson, 1979) with minor modifications. Briefly, tissue sections on slides were deparaffined and gradually hydrated with five dips in each of 80%, 70% and 50% ethanol followed by four dips in 1% acetic acid. The slides were then rinsed in water for 2 min and transferred to McIlvaine buffer (76 mM Na2HPO4, 65 mM citric acid monohydrate and 0.13% sodium azide) for 3 min. Tissues were stained for 3 min in a solution composed of one part of acridine orange stock solution (0.1% acridine orange) and nine parts of McIlvaine buffer. Slides were then rinsed for 4 min in McIlvaine buffer followed by distilled water. Mounting media and cover glasses were then applied to the slides. The mounting medium, which contained 5% gelatin, 27% glycerol and 0.1% sodium azide, was melted before use in a water bath at 60 8C. The sections were mounted from warm distilled water. 2.3. Microscopic observations and analysis of mammary tissue images Stained tissue sections were examined with a standard Carl Zeiss epifluorescence microscope. The


images used for quantification were taken at a final magnification of 136 with a fluorescence emission of 450–490 nm. Pictures of 1700 mm2 were taken with a Kodak MDS 120 color camera. Ten images were analyzed for each individual sample. One mammary gland per mouse and two mice were used for each experimental condition for a total of 20 images per condition. Based on the differential staining exhibited by cells in the mouse mammary tissue, epithelial cells and PMNs were counted with image analysis software developed by the company DLW Information Inc. (Sherbrooke, Que., Canada; website: www.dlw-information.com). An independent operator from DLW Information Inc., unaware of any specific information about the samples, analyzed the digital images. On the images, cells are present in the form of local color peaks. Mathematical morphology techniques (dilation, erosion, opening and closing, etc.) were applied to analyze these peaks in terms of shape in order to specifically identify cells while eliminating the background. The color surrounding the cells were used by the software to differentiate attached epithelial cells, free epithelial cells and PMNs. Before each experiment and determination of cellular counts, three images were used to set the level of sensitivity and specificity of detection, i.e. the ability of the software to specifically identify and count cells present on the images was carefully evaluated and adjusted so that cell counts generated by the software accurately represented visual observation of the cells and their respective type. All the images were then processed with these settings. 2.4. Statistical analysis PMN and bacterial log10 CFU counts obtained for each experimental group were compared for statistical significance by using the non-parametric Kruskall– Wallis test followed by the Dunn’s multiple comparisons test. P < 0.05 was considered significant.

3. Results and discussion In preliminary studies with Giemsa staining, it was determined that PMNs accounted for more than 95% of the cells that infiltrated mouse mammary tissue during S. aureus IMI (data not shown). Here,

165

fluorescence microscopy was used to observe PMN infiltration in tissue after staining with acridine orange. It allowed visualization of epithelial cells as orange cells containing a yellow nucleus. Free epithelial cells detaching from the gland were visible as round isolated forms, the attached ones formed multiple lines of cells looking like a mesh arrangement (Fig. 1A). In the case of PMNs, they were visible as yellowish light-green spots with the characteristic lobulation of the neutrophil nucleus (Fig. 1B). At 6 h post-infection, bacteria were present at 105 CFU/g of gland, as determined by a CFU plate count of tissue homogenates, and PMNs were present in a relatively small number (Fig. 1C). However, approximately 630 and 2400 times more bacteria were present at 12 and 24 h of infection, respectively, and proportionally, a massive infiltration of PMNs was visible in tissues (Fig. 1D and E). As hoped, antibiotic treatment with cephapirin diminished in a dose-dependent manner the number of PMNs at 24 h post-infection (Fig. 1G and H) as compared to that observed in PBS-treated mice (Fig. 1F). Epithelial cells and PMNs were counted with image analysis software (see Section 2). Fig. 1I shows how the software discriminated between epithelial cells (red crosses) and PMNs (blue crosses) based on the differential staining exhibited by these cells and the surrounding environment. The free epithelial cells never accounted for more than 0.3–1.5% of total epithelial cells (data not shown) and henceforth, ‘‘epithelial cells’’ will refer exclusively to attached epithelial cells. Fig. 2 shows the absolute numbers of PMNs in relation with time and the bacterial load of the mammary glands. The detection of PMNs by the image analysis software can be influenced by the sensitivity settings that are adjusted at the beginning of the process. This is carried out using representative control images (three images were used in the present case). To be sure that independent experiments would be comparable and not influenced by the initial settings, the PMN counts for uninfected tissues (inoculated with PBS) were subtracted from the PMN counts determined for infected tissues at each experimental time point (Fig. 3A, solid bars). The PMN counts obtained for uninfected tissues were thus considered here as background values. At 6 h of infection, the PMN count was found to be slightly negative, after background subtraction, indicating that

166


Fig. 1. PMN tissue infiltration during S. aureus IMI in the mouse. Staining of mammary tissue sections with acridine orange reveals, by the use of fluorescence microscopy, the presence of milk (green), epithelial cells (orange) and PMNs (yellow–light-green). Inocula of 1.4 to 3 102 CFU/gland were used for the IMI. (A) Epithelial cells were visible as being attached (big arrows) or free (small arrows) (12 h, non-infected). (B) Yellow–light-green cells that are PMNs (arrows), with characteristic lobulation of nuclei, infiltrated the tissues after S. aureus inoculation, thus indicating inflammation of the mammary gland (24 h, infected). (C, D, and E) Infected mammary tissues at 6, 12 and 24 h, respectively. A small number of PMNs can be seen at 6 h whereas infiltration is clearly visible at 12 and 24 h showing an impressive number of PMN cells. (F, G, and H) mammary tissues of mice infected and treated with i.m. injections of PBS, 10 and 25 mg/kg of CEP, respectively (24 h). (I) Computer-assisted identification of cells. The red and blue crosses on the image respectively represent the epithelial and PMN cells automatically detected by the image analysis software. The picture was originally taken at a magnification of 136 and then used for quantitation without any modification. It is here magnified to allow observation of the quality of the detection. Approximate magnification on the figure: (A, B) 600, (C–H) 240, (I) 1200.


Fig. 2. Computer-assisted quantitation of PMNs in tissue during S. aureus IMI. The absolute numbers of PMN cells are indicated on the graph. Also shown are the CFU counts obtained for S. aureus through time. Values are given as medians with the corresponding interquartile (Q1–Q3). The numbers of PMNs are statistically different (P < 0.05) from each other at each time point whereas, for bacteria, only CFU counts between 12 and 24 h do not differ significantly.

almost no additional PMNs, compared to uninfected tissues, were detected by the software. However, the PMN counts increased approximately 1100 folds between 6 and 24 h (Fig. 3A, solid bars). For epithelial cells, similar counts were obtained at each time point for either infected or non-infected tissues (not shown). However, between 6 and 12 h after removal of the offspring, a diminution of approximately 40% in the number of epithelial cells was observed while this number remained relatively constant after 12 h of infection (Fig. 3A, empty bars).

167

This result confirmed, as we and others have previously observed (Anderson, 1981; Brouillette et al., 2003), that milk accumulates in mammary glands when suckling is stopped and that the gland tissue consequently expands to accommodate this increased quantity of milk. To account for this expansion, the number of infiltrated PMNs thus needs to be quantitated in relation to the number of epithelial cells. With this in mind, a ratio of PMNs to epithelial cells, denoted as the ‘‘inflammation index’’, was calculated at each time point and represented 0.01, 0.15 and 0.39 at 6, 12 and 24 h, respectively (Fig. 3B). Representative images with such inflammation indices are shown in Fig. 1C–E. However, it must be kept in mind that 20 images each one ten-times larger than those of Fig. 1C–E were used for the calculation of the inflammation index for each given condition. This adds up to a total of 200 times the surface of a picture presented in Fig. 1C–E. The practical applicability of this computerassisted method for measuring PMN infiltration following a given pharmacological treatment was evaluated by the administration of the antibiotic CEP to animals. Two CEP treatments were administered to infected mice (10 or 25 mg/kg of body weight) and compared to PBS-treated infected mice. The calculated inflammation index values (ratios of PMNs to epithelial cells) for PBS-treated control mice and CEP-treated mice, respectively, were 0.65 and 0.19 for the 10 mg/kg treatment and, 0.33 and 0.05 for

Fig. 3. Determination of the ‘‘inflammation index’’. (A) The number of epithelial cells (empty bars) and PMNs (solid bars) in relation with time is shown for mice that received an intramammary injection of S. aureus. The number of PMNs for non-infected mice was subtracted from the number of PMNs determined for infected mice, at each experimental time point. Resulting values are given as medians with the corresponding interquartile (Q1–Q3). The numbers of PMNs are statistically different (P < 0.05) from each others at each time point whereas, for epithelial cells, only the counts between 12 and 24 h do not differ significantly. (B) Variation with time of the ‘‘inflammation index’’ (ratio of PMNs to epithelial cells).

168


the 25 mg/kg treatment. This indicates that PMN infiltration in the mammary tissue was reduced by 71% with the 10 mg/kg dose whereas the 25 mg/kg dose reduced the number of PMNs by 85%. The calculated inflammation indices for such treatments corresponded well to the cell infiltration observed in glands from treated (Fig. 1G–H) and non-treated mice (Fig. 1F). Bacterial enumeration of tissue homogenates also showed that 58% and 92% of mammary glands were sterilized at 24 h (CFU below the detection limit of 100 CFU/g of gland) when mice were treated with CEP at 10 and 25 mg/kg of weight, respectively. Histological examination of glands from non-infected control mice that received CEP showed no effect of the antibiotic on the tissue (not shown). These results indicate that, in a dose-dependent manner, antibiotic treatment ultimately reduced the inflammation in terms of PMN infiltration by diminishing the bacterial burden in mammary glands. The degree of PMN infiltration thus appears to be a good indicator of the level of infection and of the effect of a therapeutic treatment. The image analysis software permitted the efficient counting of several hundred thousand cells. On each image analyzed, 1400 to 8000 cells were counted. For each experimental condition, 20 images from two specimens were analyzed i.e., 34,000 mm2, leading to the detection of 54,000–96,000 cells. Throughout this study, more than 740,000 cells were counted. Interestingly, the method can be applied to archive samples embedded in paraffin wax and could probably be also used with other types of tissues, depending on the staining exhibited by local resident cells. An important consequence of bovine mastitis is the damage caused by inflammation to the secretory epithelial cells that ultimately leads to destruction of the tissue. The mouse mastitis model of infection and the method of PMN quantification presented here thus constitute interesting tools that can be used to investigate the efficacy of any potential antibacterial agents or anti-inflammatory molecules. In this report, it was shown that there was a good association between the bacterial burden and PMN infiltration in mammary gland tissue sections prepared from a mouse model of S. aureus mastitis. This outcome was maintained during therapy with an antibiotic that indirectly reduced PMN infiltration by diminishing the bacterial load in tissues.

Acknowledgements This study was supported in part by grant no. MOP57701 to FM from the Canadian Institutes for Health Research and by grant no. 2003-NO-87490 to BGT from the program FQRNT-Novalait Actions Concerte´ es. References Anderson, J.C., Holmberg, O., 1977. The invasiveness of Staphylococcus aureus and Staphylococcus epidermidis for the mammary gland of the mouse. Acta Vet. Scand. 188, 129. Anderson, J.C., 1979. Experimental staphylococcal mastitis in the mouse: the persistence of chronic infection from one lactation to the next. Res. Vet. Sci. 26, 213. Anderson, J.C., 1981. The effect of suckling on the course of experimental staphylococcal mastitis in the mouse. Br. Vet. J. 137, 489. Bramley, A.J., Patel, A.H., O’Reilly, M., Foster, R., Foster, T.J., 1989. Roles of alpha-toxin and beta-toxin in virulence of Staphylococcus aureus for the mouse mammary gland. Infect. Immun. 57, 2489. Brouillette, E., Talbot, B.G., Malouin, F., 2003. The fibronectinbinding proteins of Staphylococcus aureus may promote mammary gland colonization in a lactating mouse model of mastitis. Infect. Immun. 71, 2292. Brouillette, E., Grondin, G., Lefebvre, C., Talbot, B.G., Malouin, F., 2004. Mouse mastitis model of infection for antimicrobial compound efficacy studies against intracellular and extracellular forms of Staphylococcus aureus. Vet. Microbiol. 101, 263. Craven, N., Anderson, J.C., 1983. Antibiotic activity against intraleukocytic Staphylococcus aureus in vitro and in experimental mastitis in mice. Am. J. Vet. Res. 44, 709. Guhad, F.A., Jensen, H.E., Hau, J., 2000. Complement activation in SCID and nude mice is related to severity of tissue inflammation in the Candida mastitis model. FEMS Microbiol. Lett. 192, 231. Humanson, G.L., 1979. Animal Tissue Techniques, fourth ed. W.H. Freemen and Company, San Francisco, p. 468.. Jonsson, P., Lindberg, M., Haraldsson, I., Wadstrom, T., 1985. Virulence of Staphylococcus aureus in a mouse mastitis model: studies of alpha hemolysin, coagulase, and protein A as possible virulence determinants with protoplast fusion and gene cloning. Infect. Immun. 49, 765. Kerr, D.E., Plaut, K., Bramley, A.J., Williamson, C.M., Lax, A.J., Moore, K., Wells, K.D., Wall, R.J., 2001. Lysostaphin expression in mammary glands confers protection against staphylococcal infection in transgenic mice. Nat. Biotechnol. 19, 66. Mamo, W., Lindahl, M., Jonsson, P., 1991. Enhanced virulence of Staphylococcus aureus from bovine mastitis induced by growth in milk whey. Vet. Microbiol. 27, 371. Olfert et al., 1993. second ed. Olfert, E.D., Cross, B.M., McWilliam, A. (Eds.), 1993. Guide to the care and use of experimental animals, vol. 1. Canadian Council on Animal Care, Ontario, Canada.

E. Brouillette et al. / Veterinary Immunology and Immunopathology 104 (2005) 163–169 Owens, W.E., Nickerson, S.C., Washburn, P.J., 1992. Effect of a milk-derived factor on the inflammatory response to Staphylococcus aureus intramammary infection. Vet. Immunol. Immunopathol. 30, 233. Paape, M.J., Shafer-Weaver, K., Capuco, V.A., Van Oostveldt, K., Burvenich, C., 2000. Immune surveillance of mammary tissue by phagocytic cells. Adv. Exp. Med. Biol. 480, 259.

169

Reid, I.M., Harrison, R.D., Anderson, J.C., 1976. Experimental staphylococcal mastitis in the mouse. A morphometric study of early changes in mammary gland structure. J. Comp. Pathol. 86, 329. Sanchez, M.S., Ford, C.W., Yancey Jr., R.J., 1994. Efficacy of tumor necrosis factor-alpha and antibiotics in therapy of experimental murine staphylococcal mastitis. J. Dairy Sci. 77, 1259.