Department of Regenerative Medicine, Research Institute, International Medical Center of Japan, Tokyo
Department of Regenerative Medicine, Research Institute, International Medical Center of Japan, 1-21-1 Toyama, Shinjuku-ku, Tokyo 162-8655, Japan. Telephone: 81-3-3202-7181, ext. 2825; Fax: 81-3-3202-7192
Recently, the detection of side population (SP) cells, which have the ability to strongly efflux Hoechst 33342 fluorescence dye, has attracted attention as a method of stem cell isolation. We identified SP cells from mouse skin using the same method as from bone marrow. This population almost completely disappeared after treatment with the calcium channel blocker verapamil. SP cells were mainly localized in the epidermis, with a few in the dermis. The ratio of SP cells decreased as the mouse became older. Surface marker analysis revealed that the sorted SP cells expressed α6-integrin, β1-integrin, Sca-1, keratin 14, and keratin 19, which are proliferating and progenitor cell markers, at levels higher than in non-SP cells, while they expressed E-cadherin, CD34, and CD71 at lower levels. The expression of breast cancer resistance protein 1 (BCRP1), which participates in dye efflux, was expressed at high levels at both the protein and mRNA level in sorted SP cells. Immunohistochemical analysis showed that BCRP1 was expressed in the basal layers and hair bulge regions of mouse skin. BCRP1 mRNA was found in basal layers and hair follicles of newborn skin by in situ hybridization. These results indicate that the localization of BCRP1-positive cells is compatible with that of keratinocyte stem cells. Based on the close relationship between BCRP1 and the SP cell phenotype, we conclude that keratinocyte stem cells are closely related to the SP- or BCRP1-positive cells.
Many studies have investigated the localization, character, and function of stem cells in the skin. Multipotent skin stem cells have been thought to be localized in the basal layers or the arrector pili muscle–attaching area in the follicle, called the bulge area, where they have been detected by labeling skin with titrated thymidine or bromodeoxyurindine [1, 2] or colony-forming culture . Several specific molecular markers can distinguish stem cells from other skin cells; α6-integrin, β1-integrin, CD34, and keratin 19–positive and CD71-negative skin cells are thought to be keratinocyte stem cells [4–8]. Although many methods have been developed for the isolation and analysis of specific cell types, these cannot be used for medical experimentation on living material because of the damaging techniques, such as isotope radiation and cell fixation.
Recently, it was discovered that hematopoietic stem cells with the characteristics of immature stem cells could be isolated as a specific cell population that had the capability to strongly efflux Hoechst 33342 DNA-binding fluorescent dye . The method relies on incubating the target cells with Hoechst 33342 and performing subsequent fluorescence-activated cell sorter analysis of dual-wavelength Hoechst fluorescence with gating on a specific side population (SP) displaying low red and low blue fluorescence. This low-staining population is called the SP.
These interesting phenomena are explained by the mechanism of a novel stem cell half-transporter. Breast cancer resistance protein (BCRP1), which is one of the multidrug resistance proteins (MDRPs) on the cell membrane and an ATP-binding cassette transporter, predominantly effluxes Hoechst 33342 . MDRPs are associated with resistance to some carcinostatics and are overexpressed in several cancer cell lines [11, 12]. In hematopoietic stem cells, SP cells express relatively high levels of this BCRP1 . Zhou et al. [10, 14] determined that bone marrow SP cells require surface expression of BCRP1 for the efflux of Hoechst dye. Therefore, BCRP1 expression may serve as a new marker for stem cells, not only in hematopoietic cells but also in other types of cells.
As for skin SP cells, only a few studies have been previously performed [15–18]. It has not been clarified where SP cells are localized and whether skin SP cells possess stem cell characteristics. Accordingly, in the present study, we analyzed skin cells from newborn and adult mice for the presence of SP cells. We further examined the characteristics of the isolated skin SP cells and localized the SPs that were BCRP1-positive in mouse skin.
Materials and Methods
Isolation of SP Cells from Mouse Skin
C57BL/6J mice (Clea Japan, Tokyo, http://www.clea-japan.com) were used. Animal studies were conducted according to protocols approved by the International Medical Center of Japan. For the isolation of skin cells, mice were shaved and the dorsal and ventral skin was removed. The skin was minced in cold phosphate-buffered saline (PBS) and incubated for 60 minutes at 37°C in 0.25% trypsin-EDTA (Gibco, Grand Island, NY, http://www.invitrogen.com). Newborn skin was first incubated in 2000 U/ml dispase (Godo Shusei, Tokyo, http://www.godo.jp/english) overnight at 4°C and then separated into epidermis and dermis. Each tissue was incubated for 30 minutes at 37°C in 0.25% trypsin-EDTA. Cells were washed, resuspended at 1 × 106 cells/ml in PBS with 3% fetal bovine serum, and incubated in 5 μg/ml Hoechst 33342 dye (Sigma, St. Louis, http://www.sigmaaldrich.com) for 60 minutes at 37°C. An aliquot was stained with Hoechst in the presence of 100 μM verapamil (Wako Chemical, Osaka, Japan, http://www.wako-chem.co.jp/english), which blocks the action of the transporter responsible for Hoechst exclusion. After terminating the staining, the cells were washed and then analyzed and sorted by EPICS ALTRA flow cytometer (Beckman Coulter, Fullerton, CA, http://www.beckman.com). They were first excited with 50 mW of UV (351–364 nm), and then the emission was detected through a 450/20-nm (Hoechst blue) band-pass filter and a 675-nm (Hoechst red) long-pass filter. All of the parameters were collected using linear amplification in list mode and displayed in a Hoechst blue versus Hoechst red dotplot to visualize SP.
After sorting, aliquots of the cells were incubated for 30 minutes at 4°C with anti-mouse α6-integrin antibody (fluorescein isothiocyanate [FITC]–conjugated, 1:100), anti-mouse β1-integrin antibody (FITC-conjugated, 1:100), anti-mouse stem cell antigen-1 (Sca-1) antibody (FITC-conjugated, 1:100), anti-mouse CD34 antibody (FITC-conjugated, 1:100), anti-mouse CD71 antibody (phycoerythrin [PE]-conjugated, 1:100), anti-mouse E-cadherin antibody (FITC-conjugated, 1:100), anti-mouse CD3 antibody (FITC-conjugated, 1:100), anti-mouse B220 antibody (FITC-conjugated, 1:1000,), or anti-mouse CD45RO antibody (PE-conjugated, 1:100) (all from BD Biosciences, San Jose, CA, http://www.bdbiosciences.com). For keratin 14, keratin 19, and BCRP1 staining, we used an IntraPrep Permeabilization Kit (Immunotech, Marseille, France, http://www.bc-cytometry.com). Anti-mouse keratin 14 and 19 monoclonal antibodies (1:100) were kind gifts from Professor E.B. Lane, University of Dundee School of Life Sciences, Dundee, Scotland, U.K. Anti-mouse BCRP1 antibody (BXP-9, 1:50) came from Kamiya Biomedical (Seattle). FITC-conjugated anti-mouse immunoglobulin G (IgG) antibody (1:1000, Molecular Probes, Eugene, OR, http://probes.invitrogen.com) and PE-conjugated anti-rat IgG antibody (1:200, BD Biosciences) were used as the secondary antibodies.
Western Blotting for BCRP1 Expression on SP and Non-SP Cells
Each fraction of sorted and unsorted cells from epidermal and dermal tissues was collected by centrifugation at 4°C. The pellets were homogenized in RIPA buffer (1% Nonidet P-40, 0.5% sodium deoxycholate, 150 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, and 1 μg/ml leupeptin) at 4°C. The lysates were analyzed by SDS-PAGE. Equal amounts of protein were electrophoresed in a 7.5% SDS-polyacrylamide gel and then transferred onto a nitrocellulose membrane. The membrane was blocked with 2% bovine serum albumin in tris-buffered saline (TBS) for 2 hours. Primary antibody staining with anti-mouse BCRP1 monoclonal antibody (BXP-53, 1:100, Kamiya Biomedical) was performed in TBS with 0.1% Tween 20 (TBST) overnight at 4°C. Secondary antibody staining with anti-rat IgG (HRP-conjugated, 1:1000, Santa Cruz Biotechnology, Inc., Santa Cruz, CA, http://www.scbt.com) was performed for 1 hour at room temperature. Last, chemiluminescence was performed with the SuperSignal Chemiluminescent Substrate (Pierce, Rockford, IL, http://www.piercenet.com).
Reverse Transcription–Polymerase Chain Reaction for BCRP1 Gene
Total RNA was extracted from the pellets of sorted cells using ISOGEN (Nippon Gene, Tokyo, http://www.nippongene.com) according to the manufacturer's protocol. Reverse transcription–polymerase chain reaction (RT-PCR) was then performed using an OneStep RT-PCR Kit (Qiagen, Hiden, Germany, http://www.qiagen.com). The primers used for PCR amplification were as follows: BCRP1: 5′-CCA TAG CCA CAG GCC AAA GT-3′ and 5′-GGG CCA CAT GAT TCT TCC AC-3′, ACTIN: 5′-GCT CGT TGC CAA TAG TGA TG-3′ and 5′-AAG AGA GGT ATC CTG ACC CT-3′. Conditions for BCRP1 were 94°C for 1 minute, 56°C for 1 minute, 72°C for 1 minute, 38 cycles. Conditions for ACTIN were 94°C for 1 minute, 55°C for 1 minute, 72°C for 1 minute, 33 cycles. Acquired DNA samples were loaded onto a 2% agarose gel and analyzed.
Immunohistochemical Staining of Mouse Skin with Anti-BCRP1 Antibody
Five micrometers of frozen mouse skin fragments were dried and fixed in 100% ethanol for 10 minutes at 4°C. After washing and blocking, they were incubated with primary anti-mouse BCRP1 monoclonal antibody (BXP-9, 1:100, Kamiya Biomedical), anti–Ki-67 monoclonal antibody (SP6, Lab Vision, Fremont, CA, http://www.labvision.com), or anti-CD31 antibody PE-conjugated (1:100, BD Biosciences) for 1 hour at room temperature. Anti-rat IgG (1:100, HRP-conjugated, Amersham Biosciences, Piscataway, NJ, http://www.amersham.com) enhanced with TSA system Alexa 488–conjugated (Molecular Probes) and anti-rabbit IgG (1:300, Alexa 594–conjugated, Molecular Probes) were used as secondary antibodies. The samples were mounted and observed under a confocal microscope LSM510 (Carl Zeiss, Thornwood, NY, http://www.zeiss.com).
In Situ Hybridization
These methods have been previously described . Briefly, for the synthesis of the probe, a 386-bp BCRP1 cDNA sequence was subcloned into pGEM-T (Promega, Madison, WI, http://www.promega.com). The linearized plasmid was incubated with T7 or SP6 RNA polymerase, transcription buffer, digoxigenein-labeled UTP, and nucleotides at 37°C for 2 hours. For in situ hybridization, rehydrated paraffin-embedded tissue sections were fixed in 4% paraformaldehyde in PBS for 15 minutes and then treated with 7.5 μg/ml proteinase K in PBS at 37°C for 1 hour. They were refixed with 4% paraformaldehyde in PBS and placed in 0.2 M HCl for 10 minutes. After washing, they were acetylated by incubation in 0.1 M triethanolamine-HCl, pH 8.0, for 1 minute and further in 0.1 M triethanolamine-HCl, 0.25% acetic anhydride for 10 minutes. After washing, they were dehydrated through an ethanol series. Hybridization was performed with probes at concentrations of 200–500 ng/ml in a hybridization solution (50% formamide, 5 × standard saline citrate [SSC], 1% SDS, 50 μg/ml tRNA, and 50 μg/ml heparin) at 55°C for 16 hours. After hybridization, they were washed in 5 × SSC at 55°C for 15 minutes and then in 50% formamide, 2 × SSC at 55°C for 15 minutes, followed by RNase treatment in 50 μg/ml RNase A in 10 mM Tris-HCl, pH 8.0, 1 M NaCl, and 1 mM EDTA. Then they were washed twice with 2 × SSC at 50°C for 15 minutes, twice with 0.2 × SSC at 50°C for 15 minutes, and once with TBST (0.1% Tween 20 in TBS) for 5 minutes. After treatment with 0.5% blocking reagent (Roche, Mannheim, Germany, http://www.roche.com) in TBST for 1 hour, they were incubated with anti-DIG AP conjugate (Roche) diluted 1:2000 with TBST for 1 hour, washed twice with TBST containing 2 mM levamisole, and then incubated in 100 mM NaCl, 50 mM MgCl2, 0.1% Tween 20, 100 mM Tris-HCl, pH 9.5, and 2 mM levamisole. Coloring reactions were performed with BM purple substrate (Roche) overnight. The samples were dehydrated and mounted on a slide glass.
Aging analyses were performed in at least three independent experiments that yielded highly comparable results. Data are presented as mean values ± SD, as indicated in the Figure 1. Differences between mean values were analyzed with Student's t-test, and p < .05 was considered statistically significant.
SP Cells Were Detected in Mouse Skin
Flow cytometric analysis of skin cell suspensions revealed that approximately 1% of total cells were found distinctly and reproducibly in the tail of the curve (Fig. 2A), representing a side population analogous to bone marrow SP cells (Fig. 2C). This SP was eliminated by verapamil treatment (Figs. 2B, 2D). Interestingly, the frequency of SP cells was relatively higher in mouse skin (approximately 1.4%) than in bone marrow (approximately 0.1%). Cells isolated from the ear skin suspension gave similar results (data not shown). Next, we also measured the SP cells from the epidermis and the dermis separately. SP cells were present at high levels in the epidermis (approximately 5.1%, Fig.2E); however, they were rarely found in the dermis (approximately 0.1%, Fig. 2F), suggesting that SP cells are located mainly in the epidermis in vivo. We then compared the ratios of SP cells among various ages of mice. Epidermal cells from newborn mice had a very high ratio of SP cells. The ratio of SP cells significantly tended to decrease in proportion with aging (Fig. 1).
Characterization of the Sorted SP Cells by Several Markers
To investigate whether the SP cells expressed different surface markers than non-SP cells did, epidermal cells were separately sorted as shown in Figure 3A. We then examined putative stem cell markers. We found that SP cells were more strongly stained with α6-integrin, β1-integrin, Sca-1, and keratin 14 than non-SP cells (Figs. 3B–3E). Keratin 19 was weakly stained in SP cells but not in non-SP cells (Fig. 3F). On the other hand, we found that SP cells were very weakly stained by CD34, CD71, and E-cadherin compared with non-SP cells (Figs. 3G–3I). Both SP cells and non-SP cells other than epidermal Langerhans cells were negative for CD3, B220 (Figs. 3J, 3K), and CD45RO (Fig. 3L), suggesting that the preparations were not contaminated by hematopoietic cells.
BCRP1 Was Strongly Expressed in SP Cells at Both the Protein and mRNA Levels
Next we examined BCRP1 expression on skin SP cells, because BCRP1 is a key molecule for the SP cell phenotype [10, 13, 14]. After the sorting of both SP and non-SP cells, we analyzed the cell-surface BCRP1 protein. Flow cytometric analysis revealed that BCRP1 was expressed on the cell membrane of skin SP cells more strongly than on non-SP cells (Fig. 4A). Using an anti- BCRP1 monoclonal antibody as a probe during Western blot analysis of sorted cells and tissue lysates, we detected a single protein band with a molecular weight of approximately 72 kDa, the expected size of BCRP1 (Fig. 4B). We found that the SP cells and the epidermis expressed BCRP1; however, non-SP cells and those in the dermis expressed little if any BCRP1. These results suggest that skin SP cells express BCRP1 as do other types of SP cells and that skin SP cells are located mainly in the epidermis, which is consistent with our flow cytometric analysis of epidermal and dermal cell suspensions.
To further analyze BCRP1 mRNA, RT-PCR was performed. We found that SP cells expressed BCRP1 strongly but non-SP cells in either adult mouse (Fig. 4C) or newborn mouse (Fig. 4D) showed a weak expression. These results suggest that SP cells strongly express both BCRP1 protein on the cell membrane and BCRP1 mRNA.
Localization of BCRP1 in Mouse Skin
Based on the correlation between SP cell and BCRP1 expression, we determined where the BCRP1-positive cells existed. Because BCRP1 is also expressed in blood vessels, we performed CD31 double staining to clarify the location of BCRP1 in the skin. The immunohistochemical staining of newborn mouse skin with anti-mouse BCRP1 and CD31 antibodies combined with the usual staining with hematoxylin and eosin revealed that positive staining was found on the cell membrane in almost all of the basal layers and bulge regions of hair follicles in both newborn (Figs. 5A, 5B) and 1-week-old (Figs. 5C, 5D) mice, consistent with another new report of human skin staining . The skin of adult mouse (Figs. 5E, 5F) stained positively for BCRP1 only in the basal layers and hair bulge regions. The staining intensity of adult skin was therefore relatively weaker than that of newborn skin. The dermal papillae and surrounding lower hair follicle cells did not positively stain in either newborn or adult mice.
We further performed in situ hybridization using BCRP1 mRNA probes. Scattered BCRP1 mRNA expression was detected along the basal layers of the epidermis and hair follicles in newborn mouse skin (Figs. 6A, 6B). No accumulation of positive cells in the bulge regions or dermal papillae was observed. The lung epithelial cells in the newborn mouse were strongly positive, as previously reported  (Fig. 6C). In adult mouse skin, BCRP1 mRNA expression was scarcely detected (not shown).
Next, we compared the Ki-67 and BCRP1 expression in mouse skin to elucidate the proliferation or self-renewal capacity of the skin SP cells. We found that over half of the BCRP1-positive cells in the basal layer were also positive for Ki-67 (Figs. 7A–7C).
SP cells were first discovered in the hematopoietic system by Goodell et al. ; they are now considered primitive stem cells. Since then, SP cells have been found in several tissues and experimental cell lines, including bone marrow, muscle, lung, mammary epithelium, embryonic stem cells, neurosphere cells, and the A549 cancer cell line [9, 10, 21–25]. SP cells are also present in epidermal skin suspensions [15–18]. The SP phenotype identifies a subset of stem cells. SP cells from bone marrow have been shown to be able to rescue lethally irradiated mice  and to differentiate into both lymphoid and myeloid lineages . SP cells from bone marrow and skeletal muscles express Sca-1 and c-kit [22, 26], and SP cells from mammary epithelium express α6-integrin .
We found that the percentage of skin SP cells was relatively higher than that from other organs such as bone marrow, which was reported to be approximately 0.05%–0.1% (Figs. 2A–2D). The probable reason for the high ratio is the fact that the epidermis is a functional barrier organ that can strongly efflux foreign bodies, toxic substances, and dyes, including Hoechst dye. Previous studies have reported that the ratio of skin SP cells is approximately 1.1% in the mouse (varying by the concentration of Hoechst) [17, 18] and 0.01%–5.4% in the human . We are the first to show that newborn skin has a very high ratio of SP cells and that the percentage of skin SP cells decreases in proportion with aging (Fig. 1). Other reports have found no relationship between SP and aging [15, 17]. These discrepancies can be explained by the differences of observation periods; ours extended from newborn to 24-month-old mice, which is much longer than others. Considering the fact that embryo and newborn skin cells can reconstitute hair on nude mice but adult cells cannot , it is not surprising that newborn epidermal cells are highly plastic. Our data also clarified that the dermis is composed of only 0.1% SP cells, which is similar to bone marrow (Figs. 2C, 2F). Therefore, epidermal SP cells have a stronger dye-effluxing potential than dermal and bone marrow cells.
Many markers for epidermal stem cells have been investigated. We demonstrated that α6-integrin , β1-integrin [4, 5], Sca-1 , keratin 14 , and keratin 19  were highly expressed in SP cells and that CD34 , CD71 , and E-cadherin  were more weakly expressed in SP cells than in non-SP cells (Fig. 3). These results indicate that SP cells have several stem cell markers and are not differentiated cells. Our data for Sca-1 were compatible with previous reports [15,17,18]. However, CD34 expression on epidermal stem cells is controversial. CD34 is reported by several authors to be highly expressed in the bulge region [8, 31, 32], but another study reported that epider-mallabel-retaining cells do not express CD34 . Our results showed that SP cells rarely expressed CD34 but that non-SP cells did express CD34, in agreement with the report that BCRP1 did not appear to be enriched in CD34-high bulge stem cells . However, Montanaro et al.  claimed that CD34 is expressed in SP cells (91%) more frequently than in non-SP cells (73%). It has been suggested that these discrepancies could be due to the fact that SP- or BCRP1-positive cells are located not only in the bulge region but also in the basal layer, but it is certain that further examinations are needed.
BCRP1 is a novel stem cell transporter. BCRP1 mRNA is expressed at high levels in primitive hematopoietic stem cells and is sharply downregulated with differentiation . Our flow cytometric analysis, Western blotting, and RT-PCR analysis revealed that there is strong expression of BCRP1 in SP cells and epidermis but not in non-SP cells and only weakly in the dermis (Fig. 4B). We initially pinpointed the location of the BCRP1-positive cells in the mouse skin. Immunohistochemical staining and in situ hybridization of BCRP1 revealed that BCRP1-positive cells are located mainly in the epidermal basal layers and hair bulge regions of both newborn and adult mice (Fig. 5). However, BCRP1 expression in the dermal papilla and lower hair follicle was not detected in either newborn skin or in adult skin. The frequency of BCRP1-positive cells in the skin was similar to that of SP cells in skin suspension. These results confirm that the presence of BCRP1 is closely correlated with that of SP cells and that BCRP1 can therefore be used as a marker for primitive quiescent stem cells.
Although it has been suggested that SP cells protect the integrity of stem cells and can contribute to all cell lineages, not all SP cells function as primitive stem cells. To use hematopoietic stem cells, both SP and c-kit+, Sca-1+, CD34− cells are used for the regeneration of ischemic cardiac muscle . SP cells that have a subset of c-kit+, Thy-1+, Sca-1+, and linage− represent long-term repopulating cells . There are controversies about the stem cell properties of skin SP cells. Although it has been claimed that adult skin SP cells can engraft in dystrophic skeletal muscle , another report showed that overexpression of BCRP1 compromises the growth potential of the cells  and that SP cells are not label-retaining cells . It has also been reported that SP cells express Sca-1; however, the bulge region does not express Sca-1 , and BCRP1 does not seem to be enriched in CD34-high bulge stem cells . Considering these facts, being SP- or BCRP1-positive is not considered completely indispensable criteria for stem cells, but they are among the characteristics of the stem cells. Therefore, it is reasonable to combine SP cell screening with that for other surface markers to isolate pure epidermal stem cells. We demonstrated the proliferative character and self-renewal capacity of SP cells using Ki-67 and BCRP1 costaining (Fig. 7). To prove their abilities, we need to develop a suitable in vitro culture system for epidermal stem cells in the near future.
In conclusion, we have demonstrated that SP cells can be isolated from mouse skin and that they display several epidermal stem cell markers. These SP cells are abundant in newborn skin, have strong dye-effluxing potential, and significantly decrease in proportion with aging. The location of these SP cells in the skin was elucidated in both newborn and adult mice. Our observations open doors for future research in stem cell biology, skin carcinogenesis, wound healing, and hair and skin regeneration. Using the plasticity of skin SP cells as a source of stem cells for other organ regeneration is both promising and attractive. We are next going to transplant skin SP cells into lethally irradiated mice and determine whether the mice are rescued and the SP cells can differentiate into other cell types.
This work was supported by a Health Science Research Grant from the Japanese Ministry of Health, Labor, and Welfare. The authors thank Dr. Miho Mizukami, Yasuhiko Nagasaka, and Eri Watanabe for technical advice.