Main

It is a well-known fact that breast-feeding reduces both morbidity and mortality of infants (1). Breastmilk contains a wide variety of anti-infectious factors such as immunoglobulins, cellular components, and cytokines, which give breastmilk bactericidal, antiviral, and antifungal properties (2, 3). These factors are believed to contribute to a lower incidence of necrotizing enterocolitis, respiratory tract infections, and other gastrointestinal illnesses in breastfed infants (4–7).

Defensins are a group of naturally occurring peptides that display antibiotic and cytotoxic properties. Their cytotoxic activity is nonspecific and their antimicrobial spectrum includes both Gram-positive and Gram-negative bacteria, mycobacteria, fungi, and some enveloped viruses (8). Defensins are small peptides containing 29–45 amino acid residues, weighing only 3–4 kD. Two structurally distinct defensin peptide families exist in humans: α-defensins, found in phagocytic cells and Paneth cells of the small intestine, and β-defensins, expressed in epithelial tissues. To date, two β-defensin peptides have been identified in humans: human β-defensin-1 (hBD-1) and human β-defensin-2 (hBD-2). hBD-1 is expressed in a wide range of epithelial tissues such as kidney, pancreas, salivary gland, airway epithelia, female urogenital system, and placenta (9–11), whereas hBD-2 is expressed in tissues including skin, lung, and the urogenital system (12–15). While hBD-1 expression is constitutive, hBD-2 is induced by Gram-positive and Gram-negative bacteria, Candida albicans, and tumor necrosis factor-α (12).

Because of the epithelial origin and secretory nature of the β-defensins, we hypothesized that human mammary gland secretes defensins onto its surface. During lactation, secretion of defensins into milk may protect the tissue surface from bacterial colonization and contribute to the feeding-mediated passive transfer of innate immunity from mothers to their breast-fed infants. In this study, we report that human mammary gland epithelia express hBD-1.

METHODS

Cell culture.

The MCF-12A cell line (a human mammary gland epithelial cell line) was obtained from the American Type Culture Collection (Rockville, MD, U.S.A.). Cells were maintained in a 1:1 mixture of Ham's F12 medium and Dulbecco's modified Eagle's medium with 5% horse serum, supplemented with 0.1 μg/mL cholera enterotoxin, 10 μg/mL insulin, 0.5 μg/mL hydrocortisone, and 20 ng/mL epidermal growth factor. Routinely, cells were grown to 80–90% confluency and subcultured by trypsinization.

Mammary tissue processing.

Following approval of our protocol by Hospital for Sick Children and Sunnybrook and Women's College Health Sciences Centre Research Ethics Review Boards, discarded mammary gland tissue was obtained from nine individuals undergoing elective bilateral breast reduction operations. Their age ranged from 18 to 54 y. Mammary tissue was immediately dissected following surgery. The parenchyma (the ductular-lobular-alveolar structures), identified as tissue at the end of the whitish strands (lactiferous ducts), were subsequently snap frozen in liquid nitrogen and stored at –80°C until used for RNA preparation. Mammary tissue used for immunohistochemistry was fixed with 10% formalin in PBS immediately after surgery and subsequently embedded in paraffin and sectioned to about 5 μm thickness.

Milk cells.

Under the same ethics approval (see above), milk samples were obtained from six lactating women whose postpartum age ranged from 5 d to 7 mo. The cellular fraction was isolated by centrifugation at 1500 rpm for 10 min at 4°C, and processed for RNA purification.

Blood leukocytes.

Heparinized venous blood was obtained from a healthy female volunteer and leukocytes isolated by density centrifugation. Whole blood was centrifuged at 2000 rpm for 10 min at 18°C. The white blood cell layer (the layer at the interface of plasma and erythrocytes) was removed and treated with sterile water for 30 s to lyse contaminating red blood cells. Following resuspension in isotonic PBS, the mixture was centrifuged again at 1200 rpm for 10 min at 18°C. The lysis and wash steps were repeated until the cell pellet was free of contaminating erythrocyte debris. The white cell pellet was immediately processed for RNA purification.

RNA purification and reverse transcription polymerase chainreaction (RT-PCR).

Total RNA was isolated from cultured cells, mammary tissue, milk cells and leukocytes using the TRIZOL® Reagent (GIBCO BRL, Burlington, ON, Canada). Then 20 μL cDNA was synthesized by reverse transcription from 1 μg total RNA, all of which was used in a 50 μL PCR solution containing 2.5 mM MgCl2, and 25 μM each of forward and reverse primers (Dalton Chemical Labs, Toronto, ON, Canada). Primers used for amplification were: hBD-1: forward (5′-TTGTCTGAGATGGCCTCAGGTGGTAAC-3′), reverse (5′-TATTTCCTTTTAACGAAAACTTCATA-3′); hBD-2: forward (5′-CCAGCCATCAGCCATGAGGGT-3′), reverse (5′-GGAGCCCTTTCTGAATCCGCA-3′). The expected amplification product using these primers is 253 bp for hBD-1 (Fig. 1) (9) and 254 bp for hBD-2 (12).

Figure 1
figure 1

hBD-1 cDNA and amino acid sequences. The amino acid sequence is shown as a single-letter code below the nucleotide sequence. Nucleotide numbering is indicated on the left. The hBD-1 primers used for PCR are shown as arrows. Amplification by these primers results in an hBD-1 fragment of 253 bp. Start (atg) and stop (tga) codons are indicated in bold.

Full size image

PCR was performed with a Perkin Elmer automated thermal cycler (Foster City, CA, U.S.A.). Initial denaturation was carried out at 95°C for 5 min, followed by 30 cycles of amplification with denaturation at 94°C for 10 s, primer annealing at 62°C for 10 s, and elongation at 72°C for 40 s. A negative control (no cDNA) was also included to test for contamination. Relative levels of expression were normalized, using the above PCR conditions, with 450 bp β-actin products (forward primer: 5′-CTACAATGAGCTGCGTGTGG-3′; reverse primer: 5′-TAGCTCTTCTCCAGGGAGGA-3′). All PCR products were visualized after electrophoresis with ethidium bromide using a 1% agarose gel.

Subcloning, restriction mapping, and sequencing of hBD-1product.

The hBD-1 PCR product (253 bp) was excised from the agarose gel and purified using the QIAEX II Gel Extraction Kit (QIAGEN Inc., Chatsworth, CA, U.S.A.). Purified hBD-1 PCR product was mapped by digesting with Alu I and Bgl II (Pharmacia Biotech Inc., Baie d'Urfe, PQ, Canada). Digested fragments were visualized using the VisiGel™ Separation Matrix (Stratagene, La Jolla, CA, U.S.A.), stained with ethidium bromide.

Purified hBD-1 DNA was subcloned into the pCR®2.1 vector (Invitrogen, Carlsbad, CA, U.S.A.) and then used to transform INVα F' One Shot™ cells (Invitrogen). Subcloned hBD-1 PCR products were also mapped by digestion with Bgl II, Xba I, and Hin d III (Pharmacia Biotech Inc.). Products were visualized on 1% agarose gels stained with ethidium bromide.

Finally, DNA sequencing was performed using the T7Sequencing Kit™ (Pharmacia Biotech Inc.). Sequence data were analyzed using the MacMolly (Soft Gene GmbH, Berlin, Germany) DNA analysis software.

Immunohistochemical detection of human β-defensin peptides.

Using rabbit polyclonal antiserum to the hBD-1 (11) and hBD-2 (15) peptides previously developed, we immunostained human mammary gland tissue. Briefly, deparaffinized human tissue sections were treated to inactivate endogenous peroxidase. Slides were washed and incubated for 18 h at room temperature with a 1:800 dilution of rabbit polyclonal serum in antibody solution containing 1% gelatin, Tris-buffered saline (TBS: 500 mM NaCl, 20 mM Tris, pH 7.5), 0.05% Tween 20 (Sigma Chemical Co., St. Louis, MO, U.S.A.), and 0.01% thimerosal. After washing in TBS with 0.05% Tween 20 (TTBS), the slides were incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG (Bio-Rad Laboratories) diluted 1:2000 in antibody solution for 18 h, washed in TTBS, and developed for 1–3 min in diaminobenzidine solution and hydrogen peroxide. After developing the color, the slides were counterstained with Harris Hematoxylin stain (Fisher Scientific, Fair Lawn, NJ, U.S.A.). Preimmune rabbit serum was used as a control.

RESULTS

Detection of Human β-Defensin-1 mRNA

MCF-12A cells.

The MCF-12A cell line is a nontumorigenic epithelial cell line derived from long-term culture of normal human mammary tissue. Fig. 2A displays the results of an RT-PCR assay using both β-actin and hBD-1 primers. The MCF-12A cells resulted in a PCR product band compatible with hBD-1 in the expected region of 253 bp. To verify that this product band was indeed an hBD-1 PCR product and not a PCR artifact, we purified and mapped this fragment with Alu I and Bgl II (Fig. 2B). To further verify the findings, we subcloned the PCR product into the pCR®2.1 vector for sequencing.

Figure 2
figure 2

hBD-1 mRNA expression in MCF-12A cells. (A) hBD-1 mRNA expression was detected after reverse transcriptase PCR by electrophoresis as a 253 bp fragment on a 1% agarose gel stained with ethidium bromide. Also shown is β-actin expression, detected as a 450 bp fragment. (B) Restriction digest analysis of the hBD-1 PCR fragment. The 253 bp product was digested with Alu I and Bgl II and visualized on a VisiGel™ separation matrix. Uncut product (no enzyme) is shown in the last lane as a control. These results, in addition to sequence analysis (not shown) confirmed the PCR fragment as an hBD-1 product.

Full size image

Restriction digests with Bgl II, Xba I, and Hin d III as well as unidirectional sequencing confirmed that the amplified and cloned PCR fragment was a human β-defensin-1 product (data not shown). Therefore, the human mammary gland epithelial cell line, MCF-12A, expresses the mRNA transcript for the human β-defensin-1 gene.

Human mammary gland tissue.

The mammary gland tissue samples taken from nine women were subjected to the same RT-PCR assay as described for the MCF-12A samples and then mapped accordingly. The results of the PCR amplification of hBD-1 from breast tissue samples 1 through 9 are shown in Figure 3. Again, a product band in the expected region (253 bp) resulted from RT-PCR using the designated hBD-1 primers. Restriction mapping of the amplified products, using the enzymes Alu I and Bgl II, verified the identity of the fragments as hBD-1 products (digest results not shown).

Figure 3
figure 3

RT-PCR analysis of hBD-1 mRNA expression in human mammary tissue. RT-PCR was performed on each of nine human mammary tissue samples. Products were visualized on an ethidium bromide-stained agarose gel. Shown in lanes 1 to 9 are signals for β-actin (left side) and hBD-1 (right side) for each sample.

Full size image

Human milk cells.

One of the ultimate goals of our study was to determine whether hBD-1 is present in mammary gland epithelia at the lactation stage. Since lactating tissue is difficult to obtain, we used the cellular fraction of milk as a surrogate for lactating mammary tissue. This fraction of cells contains sloughed-off epithelial cells from the lactating breast, enabling us to indirectly assay lactating epithelium. We proceeded to isolate the cellular fraction of human milk samples obtained from six lactating women, and performed our RT-PCR assay for hBD-1. The results of this analysis revealed that all milk cell samples express the hBD-1 mRNA transcript (Fig. 4).

Figure 4
figure 4

Analysis of hBD-1 mRNA expression in milk cells and blood leukocytes. Milk cells, harvested from six human milk samples, and blood leukocytes from one healthy female volunteer were subjected to RT-PCR. All milk samples express hBD-1 mRNA, as detected by a 253 bp signal. Blood leukocytes (WBC) do not express hBD-1 mRNA, as indicated by the absence of a 253 bp signal.

Full size image

A substantial proportion of cells within milk are blood-derived cells, including lymphocytes and granulocytes. Therefore, to further characterize the results of our RT-PCR analysis on milk cells, we performed our established RT-PCR assay on human blood leukocytes. The results indicate that blood leukocytes do not express the hBD-1 transcript, suggesting that the PCR signal we obtained from milk cells is most likely derived from the sloughed-off epithelial cells in milk (Fig. 4). We performed immunohistochemistry next to extend our RT-PCR results to the peptide level and to localize the expression of the hBD-1 peptide in the mammary gland.

Immunohistochemical Detection of Human β-Defensin-1 Peptide

To determine cellular localization of hBD-1 peptide expression, paraffin-embedded sections from normal human mammary tissue were stained with hBD-1 polyclonal antiserum. In accordance with the data obtained by RT-PCR, immunohistochemistry showed staining for the hBD-1 peptide in human mammary gland epithelia (Fig. 5a-d). Specificity of immunostaining was established by observing a lack of staining with the same concentrations of preimmune serum or anti-hBD-2 serum in the same tissue. These results indicate that hBD-1 peptide expression is mainly localized in the epithelia lining the mammary ducts.

Figure 5
figure 5

Immunohistochemical staining of hBD-1 in human mammary gland tissue sections. (a) Control preimmune staining (original magnification 250X). (b) hBD-1 staining with anti-hBD-1 serum (original magnification 250X). Staining is evident as a brownish-orange color. Arrowheads in (a) and (b) indicate blood vessels. Panels (c) and (d) represent insets of (a) and (b), respectively (original magnification 1000X). The arrowheads in panel (d) indicate areas of staining within the epithelium (brownish-orange color). E, epithelium; L, duct lumen; C, connective tissue. Scale bars: (a) and (b), 100 μm; (c) and (d), 20 μm.

Full size image

Detection of Human β-Defensin-2 (hBD-2) mRNA and Peptide

We failed to detect hBD-2 mRNA in the MCF-12A cell line and in human mammary gland tissue by RT-PCR (data not shown). Similarly, immunohistochemistry for hBD-2 was negative (data not shown).

DISCUSSION

In 1996, Zhao et al. reported the constitutive expression of hBD-1 in various human-derived cell lines including mammary cells (9) of myoepithelial cell origin (16). The human mammary gland consists of two main cell types that make up its pseudostratified epithelium: epithelial cells and myoepithelial cells. Myoepithelial cells are contractile cells that form basket-like structures in the basal layer of the mammary gland epithelium, such that they do not make contact with the lumen of the mammary duct and alveoli. Their main function is to contract upon stimulation, causing milk to be released (16). In contrast, the epithelial layer faces the lumen and anything secreted by these cells directly enters the lumen of the mammary duct. While low-level expression was reported in myoepithelial cells (9), whether the milk-producing mammary gland epithelial cells express hBD-1 was unknown. Our study provides the first evidence of hBD-1 expression in mammary gland epithelia, suggesting that hBD-1 is secreted onto the epithelial surface and possibly into human milk.

Lactational mastitis is an infection of the breast tissue that afflicts between 4–33% of breast-feeding women (17–22). Although the majority of women infected with mastitis continue breast-feeding, symptoms can become significantly debilitating to cause some women to discontinue breast-feeding. This is of concern, given the well-established advantages of breast-feeding for neonatal immunity.

The anti-infective activity of mother's milk has been a long-time subject of research efforts and investigation. Although the antibacterial activity of human milk has been extensively investigated, the entirety of milk's anti-infective properties remains to be elucidated. For instance, milk possesses an unidentified heat-stable resistance factor that provides protection against staphylococcal infection (1). Also, outbreak of necrotizing enterocolitis in premature infants following withdrawal of boiled milk (23) suggests the presence of an anti-infective factor in human milk that, unlike other known milk proteins such as IgA and lysozyme, is heat stable. It is quite possible that this undefined factor could be hBD-1 because defensins are heat stable, retaining their antibacterial activity after boiling (8, 24).

hBD-1 was first identified in 1995 by Bensch et al. in human hemofiltrate from patients with end-stage renal disease (25). It was found to be a short peptide of only 36 amino acids, with a molecular mass of around 3 kD. Since then, investigations into the expression of β-defensins have revealed the peptides to be expressed in tissues of epithelial nature (9, 10), emphasizing their importance in mucosal defenses. It is thought that β-defensins are secreted by the epithelium onto its surface, thus providing a protective “antimicrobial” barrier for the tissue. Because relatively low NaCl concentrations are needed for optimal antibacterial activity of β-defensins, it is speculated that the high salt in airway fluids of patients with cystic fibrosis compromises hBD-1 function, enhancing the patients' susceptibility to infection from Pseudomonas and Staphylococcus aureus (24, 26). Importantly, this inactivation is not expected in human milk because the NaCl concentrations are as low as 10–20 mM (1).

Using RT-PCR, we have shown that hBD-1 mRNA is expressed in cultured mammary gland epithelial cells as well as in mammary glandular tissue of nonlactating women. We were also able to show that only sloughed-off milk epithelial cells, and not milk leukocytes, express the hBD-1 transcript. In their study on expression of defensins in human airway epithelia, McCray and Bentley have made brief mention that human macrophages express hBD-1 mRNA (10). Zhao et al. also detected hBD-1 mRNA expression in peripheral blood leukocytes, albeit at very low levels (9). We were unable to confirm these results by RT-PCR.

Immunostaining of mammary tissue sections for hBD-1 confirmed our RT-PCR results and localized the expression of hBD-1 to the mammary epithelium.

In contrast with the previous findings that showed hBD-2 mRNA expression in the mammary gland by in situ hybridization (14), we failed to detect hBD-2 mRNA and peptide. It is unclear if this discrepancy is due to the difference in the method of detection.

While expression of hBD-2 is inducible by immunologic and bacterial factors, hBD-1 is expressed constitutively. Coupled with our findings, this suggests that defensins may play a role in preventing infection of the breast both at the lactating and nonlactating stages. Since mastitis is most commonly caused by Staphylococcus aureus and defensins have shown antibacterial activity against this organism, it is possible that women who inherently express less hBD-1 are at an increased risk for bacterial colonization. Investigations into the interindividual variation of hBD-1 expression among nonlactating women is necessary to provide insight on differences in susceptibility to infection. Given the secretory nature of β-defensins, it is also possible that hBD-1 may be secreted into breast milk providing immunity to the child. Detection and quantitation of hBD-1 in milk is crucial to provide data on its role in the human mammary gland as a significant component of passive immunity for neonates.

In summary, we have found that human mammary tissue and a representative mammary epithelial cell line express hBD-1 mRNA, and that the protein is present in the luminal epithelia of the mammary gland. Further studies are needed to determine whether it functions only as a local anti-infective barrier or whether it contributes in furthering feeding-mediated passive immunity for the neonate.