Assessment of bovine mammary chemokine gene expression in response to lipopolysaccharide, lipotechoic acid + peptidoglycan, and CpG oligodeoxynucleotide 2135

Abstract

During intramammary infections pathogen associated molecular patterns (PAMPs) induce an inflammatory response, recognized clinically as mastitis. Recognition of PAMPs by mammary cells leads to the production of the pro-inflammatory cytokines, TNF-α and IL-1β. These cytokines augment the secretion of various chemokines that are responsible for directing the host cellular immune response, and consequently the outcome of infection. Previous research has shown that gram-negative and gram-positive bacteria elicit different types of innate immune responses. The purpose of this study, therefore, was to characterize the expression of various chemokine genes in bovine mammary gland explants in response to lipopolysaccharide (LPS), peptidoglycan (PTG) combined with lipotechoic acid (LTA), and CpG oligodeoxynucleotide (CpG-ODN) 2135 representing gram-negative bacteria, gram-positive bacteria, and bacterial DNA, respectively, to determine if these PAMPs induce different chemokine gene expression patterns. Explants from 3 Holstein cows were cultured with 10 μg/mL of LPS, LTA + PTG, or CpG-ODN 2135 for 6 and 24 h. Total RNA was extracted and the expression of CXCL8, MCP-1, MCP-2, MCP-3, MIP1-α, and RANTES genes was measured by real-time polymerase chain reaction (RT-PCR). Lipopolysaccharide significantly induced MCP-1, MCP-2, and MCP-3 expression, and slightly increased CXCL8 gene expression. The combined PAMPs, LTA + PTG, on the other hand, significantly induced MCP-1 gene expression, and slightly increased MCP-3 expression. No significant expression differences for any of the chemokine genes were observed in explants stimulated with CpG-ODN 2135. These results demonstrate that PAMPs associated with different mastitis-causing pathogens induce chemokine-specific gene expression patterns that may contribute to different innate immune responses to bacteria.

Introduction

Mastitis continues to be the number one cause of economic losses to the United States dairy industry, costing billions of dollars annually as a result of decreased milk yield and quality, the cost of treatment, and early culling (1,2). Escherichia coli and Staphylococcus aureus infections are major causes of mastitis in North American dairy cattle, typically causing acute and chronic mastitis, respectively (1,3,4).

Inflammation in the mammary gland is a normal host response to bacterial invasion, and it helps to direct the subsequent immune response. When inflammation is not controlled in the mammary gland, mastitis can occur leading to extensive tissue damage that affects both milk production and quality. The ability of the host to defend itself against invasion by pathogens, therefore, depends partially on its ability to elicit an efficient inflammatory response, which is characterized by pathogen recognition, the release of pro-inflammatory cytokines and chemokines, a rapid influx of effector cells, and the subsequent removal of pathogens from the mammary gland (5,6).

During intramammary infections, pathogen-associated molecular patterns (PAMPs) including gram-negative lipopolysaccharide (LPS), gram-positive lipoteichoic acid (LTA) combined with peptidoglycan (PTG), and bacterial DNA (CpG oligodeoxynucleotide) are recognized by membrane or intracellular Toll-like receptors (TLRs) expressed by macrophages (Mθ), neutrophils (PMNs), and mammary epithelial cells (MEC). Recognition of these PAMPs triggers signaling pathways that lead to the activation of nuclear transcription factor kappa B (NF-κB) and to the secretion of pro-inflammatory chemokines that differentially recruit leukocyte populations from the vasculature to the site of infection (7–9).

Host recognition of the TLR agonist LPS (TLR4) has been demonstrated to induce the secretion of pro-inflammatory chemokines such as MCP-1/CCL2, IL-8/CXCL8, MCP-5, and RANTES/CCL5 in human and murine Mθ cell lines (9), and the expression of RANTES and CXCL8 in bovine MECs and mammary tissue explants (10–12). The TLR agonist LTA (TLR1/2/6) reportedly induces the expression of the pro-inflammatory chemokines CXCL8, C5a, MCP-1, and MIP-1α in human monocytes and Mθ as well as rat bone marrow derived phagocytes (13–16), and GCP-2/CXCL6 and CXCL8 in bovine MECs (12). PTG (TLR1/2/6) have been demonstrated to induce the expression of MIP-3α and MCP-1 in murine uterine epithelial cells (17,18); whereas CpG-ODNs (TLR9) have been shown to induce the expression of human keratinocyte MCP-1, CXCL8, CCL20, and CCL27 (19), and CXCL8 in human PMNs (20). The involvement of TLR9 in PAMP recognition during intramammary infections has not been investigated; however, small changes in the expression of bovine TLR9 during mastitis suggest that it may have a limited role in PAMP recognition in the mammary gland (21).

The intent of this study was to use a bovine mammary gland explant culture system to study changes in the expression of various chemokine genes elicited by ex vivo exposure to LPS, PTG + LTA, and CpG-ODN. We hypothesize that these pathogen-specific PAMPs will elicit a unique pattern of chemokine gene expression, which may contribute to a different host immune response during E. coli and S. aureus intramammary infections.

Materials and methods

Cows

Three high-lactating Holstein cows were transported from the Elora dairy research station to the University of Guelph for slaughter over a 3-week period. These cows were culled from the herd because of poor fertility and leg conformation. On the day of slaughter, milk samples were collected aseptically from each quarter and kept on ice. These samples were sent to the Animal Health Laboratory at the University of Guelph for bacterial culturing and to measure milk somatic cell counts using a Bently 300 flow cytometry cell counter.

Tissue collection

Immediately following slaughter, 1 quarter was selected for sample collection. The site of incision was rinsed with water followed by scrubbing with 4% Germi-stat, 70% isopropyl alcohol, and Savlon. A sterile scalpel was used to make 2 incisions in the mammary gland. The two flaps of skin were pulled back and secured with clamps to expose the sterile mammary gland tissue. Using a new sterile scalpel and tweezers, mammary tissue (2 cm³) was aseptically removed from 1 quarter and placed into a 50-mL centrifuge tube (Fisher Scientific, Whitby, Ontario) containing sterile cold calcium and magnesium free Hanks Balanced Salt Solution (HBSS) (Sigma Chemical, St. Louis, Missouri, USA) supplemented with 5.6 mg/L amphotericin B (Sigma Chemical), 10 000 U/mL Penicillin/Streptomycin (Invitrogen, Burlington, Ontario) and 0.700 g/L sodium bicarbonate (Sigma Chemical). The tissue samples were immediately processed as described in the following text.

Tissue culture and treatment

Tissue samples were aseptically transferred to a petri dish containing warm supplemented HBSS. The tissue samples were blotted on sterile gauze paper and cut into explants weighing between 80 and 100 mg. The explants were then sliced to < 1-mm thick with a McIlwain Tissue Chopper (The Mickle Laboratory Engineering; Gomshall, Surrey, UK) to increase the surface area, and then placed into a 24-well plate (Fisher Scientific) containing 2 mL Dulbecco’s Modified Eagles Medium (DMEM), (Fisher Scientific) supplemented with 10% FBS, 110 mM sodium pyruvate (Invitrogen), 10 000 U/mL penicillin/streptomycin (Invitrogen), 5.6 mg/L amphotericin B (Sigma Chemical), 3.396 g/L sodium bicarbonate (Sigma Chemical), and 10 mg/mL insulin (Sigma Chemical). The explants were incubated at 37°C for 18 h in 5% CO₂ then washed in supplemented HBSS. After washing, explants from each of the 3 cows were incubated with control media or with DMEM supplemented with either 10 μg/mL E. coli LPS from serotype 0111:B4 (Sigma Chemical), 10 μg/mL LTA (Sigma Chemical) + 10 μg/mL PTG (Sigma Chemical) or 10 μg/mL of nonmethylated CpG from synthetic ODN 2135 (Sigma Chemical). Explants from the 3 PAMP treatments and the control were comprised of 6 technical replicates for each cow. After incubation for 6 and 24 h at 37°C in 5% CO₂, the 6 explants within each treatment were removed from the wells, pooled and snap frozen in cryovials (Corning, Acton, Massachusetts, USA) with liquid nitrogen.

Histology

A portion of tissue was sent to the Animal Health Laboratory at the University of Guelph for histological examination. Hemotoxylin and eosin stains as well as the immunohistochemistry stain, vimentin, were used to determine the approximate proportion of epithelial and mesenchymal cells, respectively.

Tissue homogenization

Each frozen pool of explants was removed from its respective cryovial and weighed to obtain 325 to 350 mg of tissue per treatment group. These tissues were placed in 5 mL of Quazol (Qiagen, Valencia, California, USA) in a sterile DNA/RNAase free 15 mL tube and homogenized for 90 s using a sterile Kinematica Polytron homogenizing blade (Kinematica, Lucerne, Switzerland). Each sample was allowed to sit at room temperature for 5 min as per the manufacturer’s instructions, and then stored at −80°C for total RNA extraction.

RNA extraction

Total RNA was extracted from the explants using the Qiagen RNA Lipid Midi Kit as per the manufacturer’s instructions (Qiagen, Valencia, California, USA). Initially, samples were removed from the freezer and placed in a 37°C water bath for 15 min, mixing occasionally. Chloroform (1 mL) was added to each sample tube and shaken vigorously for 15 s prior to centrifugation at 5000 × g for 15 min at 4°C. The upper aqueous phase was transferred to a new 15 mL DNA/RNAase free tube and an equal volume of 70% ethanol was added and mixed by vortexing. A volume of 4 mL of sample was transferred to a RNeasy Midi Spin Column (supplied in the Qiagen kit) and centrifuged at 4000 × g for 5 min at 24°C. The flow-through was discarded and this procedure was repeated with the remaining volume of sample. The spin column was washed with 4 mL of RW1 buffer (supplied in Qiagen kit) by centrifugation at 4000 × g for 5 min at 24°C to remove any phenol or other contaminants from the column. Subsequently, 2.5 mL of RPE buffer (supplied in the Qiagen kit) was added to the column and centrifuged at 4000 × g for 2 min at 24°C. This step was repeated again, but centrifuged for 5 min. Total RNA was eluted from the spin columns, after it was transferred to a new 15 mL DNA/RNAase free tube, by adding 250 μL of sterile water to the column and centrifuging at 4000 × g for 3 min at 24°C. The total RNA was transferred to a DNA/RNAase free 1.5 mL tube and stored at −80°C. The quality and quantity of RNA was determined using an Agilent 2100 Bioanalyzer (Agilent, Palo Alto, California, USA), a NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies, Rockland, Deleware, USA), and degradometer version 1.3 software (Microarray Unit; OSU Comprehensive Cancer Center, Columbus, Ohio, USA).

Reverse transcription

Messenger RNA was reverse-transcribed using Superscript II (Invitrogen) as per the manufacturer’s instructions. Briefly, 1 μg of total RNA was combined with 1 μL of 0.5 μg/μL Oligo(dT) (Invitrogen) and 1 μL of 10 mM dNTP mix (Invitrogen). Sterile H₂O was added to the mix in a 0.5-mL PCR tube to a final volume of 12 μL. The total RNA was heated to 65°C for 5 min and then quick-chilled on ice for 1 min. The contents of the tube was collected by a quick spin-down, and 4 μL of 5× first-time buffer followed by 2 μL 0.1 M DTT was added to the mix. Total RNA was heated to 42°C for 2 min, and then 1 μL of 200 units/μL Superscript II was added, followed by heating at 42°C for 90 min and then 70°C for 15 min.

Primer design and real-time polymerase chain reaction (RT-PCR)

Primers for target genes CXCL8, MCP-1, MCP-2, MCP-3, RANTES, and MIP1-α were designed using Primer 3 software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi). Strandberg et al (12) developed the primers for the housekeeping genes, RPLPO and GAPDH. Real-time PCR was performed using the SYBR green with Rox detection system (Invitrogen) according to the manufacturer’s instructions. A master mix was prepared in dim light using 12.5 μL SYBR Green with Rox, 1 μL of mixed forward and reverse primer (10 μM), and 6.5 μL of sterile water in a DNA/RNAase free 2.0 mL tube. A master mix volume of 20 μL was transferred to each well of a chilled 96-well PCR plate (Applied Biosystems, Foster City, California, USA) designed for use with the ABI prism 7000 thermocycler (Applied Biosystems). For each sample, a 5 μL aliquot of cDNA was added in triplicate to the 96-well plate. A negative control was included in the PCR plate by adding 5 μL of sterile water to the master mix in triplicate. The plate was then inserted into the ABI prism 7000 thermocycler, and the PCR reaction was programmed with the following protocol using the relative quantification option: 50°C for 2 min, 95°C for 2 min, followed by 45 cycles at 95°C for 15 s, the primer-specific annealing temp for 30 s (Table I) and 72°C for 30 s. Real-time PCR results were obtained using the standard curve method (User Bulletin #2, ABI Prism 7700 Sequence Detection System). A melting curve along with an electrophoresis gel was run for each gene amplified to ensure amplification of a single product with the expected product size.

Table I.

Chemokine primer sequences, annealing temperatures, and PCR product sizes used for real-time polymerase chain reaction (RT-PCR)

Gene	Gene full name	Annealing/melting temperature (°C)	NCBI accession number	Forward	Reverse	Product size
aRPLPO	Acitic ribosomal protein large PO	62.0/86	TC204704	CAACCCTGAAGTGCTTGACAT	AGGCAGATGGATCAGCCA	220 bp
aGAPDH	Glyceraldehyde-3- phosphate dehydrogenase	62.0/87	TC186924	CCTGGAGAAACCTGCCAAGT	GCCAAATTCATTGTCGTACCA	226 bp
CXCL8	Interleukin 8	62.0/77	AF232704	CATGTTCTGTGTGGGTCTGG	CAGGTGAGGGTTGCAAGATT	229 bp
MCP-1	Monocyte chemoattactant protein 1	61.5/83	M84602	CGCCTGCTGCTATACATTCA	GCTCAAGGCTTTGGAGTTTG	207 bp
MCP-2	Monocyte chemoattactant protein 2	62.0/85	AF399641	CTCGCTCAGCCAGATTCAGT	TTTGGTCCAGGAGCCTTATG	214 bp
MCP-3	Monocyte chemoattactant protein 3	61.0/86	AW655802	TCTCCTTCGCTCTGAAGCTC	CAGTAAGTGCACTGGCCTCA	236 bp
RANTES	Regulated upon activation, normal T-cell expressed and secreted	62.0/86	NM_175827	CACCCACGTCCAGGAGTATT	AGGACAAGAGCGAGAAGCAA	201 bp
MIP1-α	Macrophage inflammatory protein 1 alpha	62.0/87	AV662650	TGGTGTCATCTTCCAGACCA	GTAGCTATGGAGGCGACAGG	197 bp

^aHousekeeping genes, bp — base pair

Statistical analysis

Statistical analysis of the data was performed using SAS computer software (SAS 2002; SAS Institute, Cary, North Carolina, USA) as a split-plot design with sub-sampling in which time (6 and 24 h) was in the main plot, and treatment (control, LPS, LTA + PTG, and CpG) and the treatment bytime interaction in the sub-plot. A value of one was added to all the Ct values in order to natural log transformed them and ensure variance homogeneity. Since time by treatment interactions were not detected for any of the genes expressed, the time treatment least square means are not presented, and the main effects means were compared between challenged and non- challenged explants. A Dunnett-HSU test was performed to determine which chemokine genes in each treatment were significantly different from the control. Data was considered significant at P = 0.05, and trends are reported between P = 0.05 to 0.1.

Results

Cow characteristics, somatic cell counts, and bacteriology examination

The 3 culled cows used for this study varied dramatically in age, parity, and stage of lactation. However, all were heavy lactating cows with no evidence of mastitis as indicated by somatic cell count and bacteriological examination (Table II).

Table II.

Study cow characteristics on the day of slaughter

Cow number	Trial number	Quarter biopsy	Somatic cell count	Bacteriology	Age	Parity	Stage of lactation (days in milk)
3217	1	Left hind	74 000	Negative	4.1	1	576
3155	2	Left hind	3000	Negative	4.6	2	263
3360	3	Right hind	4000	Negative	2.4	1	23

Histology examination

Histological examination of the control explants cultured for 6 and 24 h did not reveal any differences in tissue morphology. These explants were comprised of approximately 80% epithelial cells and 20% mesenchymal cells (Figure 1A–D).

Figure 1.

Control mammary explants cultured for 6 h (A and B) and 24 h (C and D) during PAMP exposure. The explants were stained with either hematoxylin and eosin (A and C) to stain dark nuclear material and light cytoplasmic material, or vimentin (B and D) to stain cells of mesenchymal origin. These images are magnified 40 ×

Real-time polymerase chain reaction (RT-PCR) analysis of mammary explant chemokine gene expression

Time by treatment interactions were not detected for any of the expressed genes; therefore, the main effects means were compared between challenged and nonchallenged explants. Treatment with LPS significantly increased mammary explant expression of MCP-1 (P = 0.01), MCP-3 (P = 0.01), and MCP-2 (P = 0.03), and an increasing trend in CXCL8 expression (P = 0.08) was also noted (Figure 2). The expression of RANTES and MIP1-α was not significantly affected by the LPS treatment. In comparison, the LTA + PTG treatment significantly increased the expression of MCP-1 (P = 0.01), and an increasing trend in MCP-3 expression was also observed (P = 0.08) (Figure 2). The fold change in gene expression for the differentially expressed genes was: 2.99 (MCP-1), 2.88 (MCP-2), 2.17 (MCP-3), and 2.08 (CXCL8) for the LPS treatment; and 2.11 (MCP-1) and 1.62 (MCP-3) for the LTA + PTG treatment. Treatment of explants with CpG-ODN 2135 did not induce the expression of any of the chemokine genes analyzed. Expression levels of the housekeeping genes RPLPO and GAPDH were not affected by the PAMP treatments used in this study.

Figure 2.

Chemokine (CXCL8, RANTES, MIP1-α, MCP-1–3) and housekeeping gene expression (GAPDH, RPLPO) following mammary explant culture with LPS, LTA + PTG and CpG-ODN 2135. Since there was no significant treatment by time interaction, these results represent the average least squares means +/− the standard error of the 6- and 24-h sampling time points. Levels of significance of the treatment groups compared to the control are indicated by * = P < 0.10, ** = P < 0.05, and *** = P < 0.01.

Discussion

The intent of this study was to use a biologically relevant bovine mammary tissue explant culture system to compare the chemokine response elicited by ex vivo treatment with different PAMPs over time. The PAMPs, LPS, and LTA + PTG were used to model the host inflammatory response to gram-negative and gram-positive bacteria, respectively. Although it is likely that different concentrations of a given PAMP will elicit variable levels of chemokine gene expression, the concentration of PAMPs used in this study was consistent with previous in vitro experiments reported in the literature (11,12,15,22,23). The LTA + PTG were combined as 1 treatment because they both represent important membrane components of gram-positive bacteria, and studies conducted with pigs, mice, and humans have demonstrated that they act synergistically (24–26).

Previous studies have shown that gram-negative (E. coli) and gram-positive (S. aureus) bacteria elicit different innate immune responses that may account for the respective acute and chronic nature of mastitis caused by these pathogens (27,28). Bannerman et al (27) for example, measured higher milk concentrations and different temporal changes in the expression of CXCL8 and C5a in cows intramammary challenged with E. coli when compared with S. aureus. In this study, we demonstrated that various chemokine genes are also be differentially expressed in response to different PAMPs associated with E. coli and S. aureus. The expression of these chemokines may influence cell trafficking into the mammary gland during intramammary infection. The chemokines, MCP-1, MCP-2, MCP-3, and to a lesser extent CXCL8, were induced by LPS. In contrast, LTA + PTG induced MCP-1 and slightly induced MCP-3 expression. A recent study by Rabot et al (11) reported significant increases in CXCL8 expression in bovine mammary explants 6 h after E. coli LPS treatment (10 μg/mL). Although the present study showed a greater fold increase (2.08) in CXCL8 expression in response to LPS than the 1.4-fold increase reported by Rabot et al (11), our smaller sample size (n = 3) likely reduced our ability to detect this increase at a level of significance of P < 0.05.

The induction of MCP-1, MCP-2, MCP-3, and CXCL8 gene expression by LPS in mammary tissues may elicit a strong neutrophil and monocyte chemotactic response characteristic of an acute E. coli intramammary infection. Conversely, the induction of MCP-1 and MCP-3 by LTA + PTG may elicit a weaker monocyte chemotactic response that may contribute to a sustained inflammatory reaction characteristic of sub-clinical/chronic S. aureus intramammary infection. In support of this, Strandberg et al (12), reported that LPS (50 μg/mL) induced a stronger and sustained increase in bovine MEC CXCL6 and CXCL8 gene expression than did LTA (20 μg/mL). Additionally, Bannerman et al (27) reported that milk CXCL8 concentrations were increased during E. coli but not during S. aureus intramammary challenge.

Leukocyte extravasation into the mammary gland during intramammary infection is likely dependent, in part, on the milieu of chemokines released by MECs following PRR activation, since chemokines are promiscuous by nature. For example, CXCL8 can bind to CXCR1 and CXCR2; however, the CXCR2 receptor can also bind to CXCL1–7 (29). Based on this, it is expected that the induction of a greater variety of chemokines by LPS compared to LTA + PTG would elicit an overall greater chemotactic response in the mammary gland.

Interestingly, changes in RANTES expression following treatment with either LPS, or LTA + PTG were not observed in this study despite a previous report of LPS induced RANTES expression by Pareek et al (10); RANTES is chemoattractant for blood monocytes, memory T helper cells and eosinophils. In Pareek et al (10) study, MEC cells were collected at slaughter, cultured, and stimulated with 1 μg/mL of E. coli LPS; RANTES gene expression was induced 208-fold by the LPS treatment. This possible discrepancy in results may be attributed to either the different concentration of LPS, or the culture system used by this group. Culturing MEC on plastic surfaces for example, can alter MEC morphology and function leading to changes in metabolite patterns, and the synthesis and secretion of milk proteins (11). This may also be the case with PAMP recognition and chemokine gene expression. Interestingly, this group did not report any induction of MCP-1–3 or CXCL8 by LPS. It is likely, however, that MCP-2, MCP-3, and RANTES are redundant in their function since they all have the capacity to bind to CCR3 and CCR5.

The induction of MEC chemokine expression by different PAMPs appears to be influenced by different culture systems. Studies using primary MEC cultures typically report high fold changes in chemokine gene expression. A study by Strandberg et al (12) for example, showed a 100-fold increase in CXCL8 gene expression after bovine MEC, grown on a collagen matrix, were stimulated with 50 μg/mL of E. coli LPS for 24 h. As mentioned previously, MEC maintained in primary culture and stimulated with 1 μg/mL of E. coli LPS also showed a 208-fold induction in RANTES gene expression (10). In contrast, a study performed using a bovine mammary epithelial cell line (MAC-T) demonstrated a modest increase in CXCL8 gene expression (< 15-fold) following stimulation with E. coli LPS (1.5 μg/mL) and E. coli culture filtrate (30). Likewise, the present mammary explant study, and one performed by Rabot et al (11), demonstrated that LPS challenge induces small increases in chemokine gene expression; levels that we were unable to detect by microarray analysis (31). The differences in chemokine gene expression among these different model systems are unclear. It is possible that LPS may not be as readily accessible to PRRs on MEC when they are challenged as an explant culture and possibly in vivo, as opposed to in primary culture.

The present study also demonstrated that CpG-ODN 2135 did not elicit a chemokine response in mammary explants. These results support the study by Goldammer et al (21), which suggests that TLR9 contribution to PAMP recognition in the mammary gland may be limited. Although bacterial DNA has not been implicated as a cause of bovine mastitis, it has been previously used as an adjuvant for cattle (32,33). Zhu et al (34) also recently demonstrated that intramammary infusion with CpG-ODN significantly induced goat TLR9 expression, and provided protection against E. coli intramammary infection.

Synthetic ODNs approximately 20 nucleotides in length have been shown to mimic the immunomodulatory effects of bacterial DNA (35), and previous studies have reported that species-specific ODN sequence optimization is required in order to achieve a maximum immune response (23,36). CpG-ODN 2135, with a leading 5′-TCGTCGTT-3′ and 2 copies of a 5′-GTCGTT-3′ motif was found to stimulate bovine peripheral blood mononuclear cells in vitro (23), but did not stimulate mammary explants in the current study.

The present study used an ex vivo bovine mammary explant system to study the effects of bacterial PAMPs on the chemokine gene expression. The explant model proved to be an effective tool for assessing the host response to bacterial LPS and LTA + PTG. These results showed that LPS and LTA-PTG elicit unique chemokine expression patterns that support the general concept that E. coli infections lead to acute mastitis, whereas S. aureus infections lead to subclinical/chronic mastitis. CpG-ODN 2135 was shown to have no effect on mammary explant chemokine gene expression. This mammary explant model may be a more biologically relevant ex vivo alternative to existing primary MEC culture techniques and characterized MEC cell lines.