Functional characterization of highly adherent CD34+ keratinocytes isolated from human skin


Ander Izeta, PhD, Fundación Inbiomed, Paseo Mikeletegi 61, 20009 San Sebastián, Spain, Tel.: +34-94-3309064, Fax +34-94-3308222, e-mail:


Please cite this paper as: Functional characterization of highly adherent CD34+ keratinocytes isolated from human skin. Experimental Dermatology 2010; 19: 685–688.

Abstract:  Compared to murine models, data on cells responsible for the homeostasis of human epidermis are scarce and often contradictory. Given the conflicting results and the availability of clinical grade protocols to purify CD34 cells from a given tissue, we pursued to phenotypically characterize human epidermal CD34+ population. After magnetic separation of whole skin CD34+ and CD34− cell fractions and selection for cells highly adherent to extracellular matrix, both CD34± fractions retained the ability to form a stratified epidermis in organotypic cultures and presented similar in vitro migratory phenotypes. However CD34− cells showed higher clonogenic potential and in vitro proliferative capacity. These results indicated that CD34− cell fraction contains stem/early progenitor cells, while CD34+ cells might be a transit-amplifying precursor for hair follicle (HF) sheath cells. The ability to isolate living cells using differential cell adhesion and surface markers provides an opportunity to study cells from different morphological regions of the HF.


extracellular matrix


hair follicle


hair follicle stem cell


outer root sheath


Decades after successful establishment of human keratinocyte cultures (1), isolation of pure epidermal and/or hair follicle stem cell (HFSC) populations in sufficient quantities for clinical use is pending because of inconsistent cell isolation and non-comparable results obtained in animal models (2). Available clinical evidence supports maintenance of interfollicular epidermis stem cells within standard keratinocyte culture. Permanent coverage of large burn wounds by keratinocyte cultures (3) has now been confirmed by lineage tracing of gene-corrected stem cells (4). Besides, naturally occurring genetic reversions of congenital mutations produce large healthy skin patches that remain stable during patient life (5). Hair follicle (HF) transplants also show life-long self-renewal capacity (6), both the upper and lower portions of bisected HFs surviving when implanted (7,8). Localization of cells with the highest in vitro proliferative potential in the human HF, a landmark of epidermal stem cells, is controversial (9–11). Isolation of stem cells in serial culture (12) or after transplantation to nude mice (13) established assays to functionally test stemness in human epidermal cells. Other evidence included differential adhesion of stem cells to extracellular matrix (ECM) (14), marker expression (2,14–18), organotypic culture (19,20) and cell migration assays (21). Expression of CD200 and negative selection for an antibody cocktail was used to successfully isolate HFSCs, although with limited functional evidence (22) and not from interfollicular epidermis (23).

Questions addressed

Mouse HFSCs are well characterized both in terms of gene expression (24–27) and membrane marker phenotype (28,29). Localization of human HFSCs appears to be similar but bulge cells present a different membrane marker profile (2,14–18). As CD34 is used for clinical grade cell isolation, we set to phenotypically characterize CD34+ cells from human skin and their stem/precursor cell status.

Experimental design

We have processed human skin biopsies (n = 7, Table S1) and comparatively analysed phenotype of highly adherent CD34± cell fractions regarding in vitro proliferation, cell migration and differentiation in organotypic culture (Appendix S1).


CD34+ cells might in principle be present both in HF and interfollicular epidermis. In our quest to isolate CD34+ cells from human skin, we first performed immunofluorescence staining on wholemount epidermis (30). As expected, CD34+ cells were clearly detected in the follicular central area and in colocalization with α6 integrin (CD49f) positive cells of the outer HF layer. Under these experimental conditions, no CD34+ cell was detected in the interfollicular epidermis. In contrast, anti-CD49f antibody stained basal keratinocytes as expected (data not shown).

Functional characterization of CD34± cell fractions highly adherent to ECM

Basal keratinocytes and epidermal stem cells express high levels of α6:β4 integrin, which mediates cell adhesion to Laminin-332 (13,14). To enrich for α6:β4hi cells, magnetically isolated CD34+ and CD34− cells from whole skin biopsies were seeded on Laminin 332-rich ECM obtained from 804G cells (31). Cells that adhered to ECM within 10 min were termed ‘adhesion high’ (Ahi). Cells that did not adhere were put into another ECM-covered plate for 1 h. Attached cells were termed ‘adhesion low’ (Alo). Cells that did not adhere were put into standard culture plates and termed ‘adhesion negative’ (A−) (Fig. 1a). Keratinocyte colonies were obtained from all fractions with the exception of A− CD34+ cells (Fig. 1b, b3). However, important phenotypic differences were observed. CD34− fraction had a better keratinocyte yield per seeded cell than CD34+, and Ahi fractions better than Alo (Fig. S1). To evaluate their in vitro long-term self-renewal capacity, cell fractions were subjected to clonogenic and serial passage assays (Fig. 1c). Plates were stained with rhodamine B to assess colony formation. Ahi CD34− keratinocytes had a higher clonogenic potential than Ahi CD34+, giving rise to more colonies and of bigger size (Fig. 1d, d1–d2). Five keratinocyte colonies of each Ahi fraction were cloned and further passaged until no more colonies were observed. CD34+ keratinocytes showed limited proliferative capacity (Fig. 1c). Accordingly, A− CD34+ keratinocytes did not grow in primary culture, while Alo CD34+ keratinocytes grew for one passage. In passage 3, only 3/5 clones of Ahi CD34+ keratinocytes were able to grow, with no single culture passaged further than P4. In contrast, Ahi CD34− keratinocytes were grown for eight passages (Fig. 1c and d, d3, d4). These results indicated that epidermal stem cells cofraction with Ahi CD34− keratinocytes in our experimental set-up.

Figure 1.

 Differential adhesion to ECM of the CD34± cell fractions permits analysis of long-term proliferative capacity in vitro. (a) Strategy for the ECM adhesion assay. Culture dishes of 60 mm were covered with 804G cell line-derived ECM and magnetically separated CD34± cell fractions were let to adhere for 10 min [adhesion high (Ahi) fraction]. Cells in suspension were transferred to new ECM-covered dishes and let to adhere for 1 h [adhesion low (Alo) fraction]. Cells that were still in suspension were transferred to new dishes without ECM [adhesion negative (A−) fraction]. (b) Primary keratinocyte culture of the Ahi, Alo and A− cell fractions. All fractions (b1–b2 and b4–b6) but A− CD34+ cells (b3) gave rise to keratinocyte colonies. Scale bars: 100 μm (c) In vitro proliferative potential of the Ahi, Alo and A− cell fractions. Starting from a secondary cell culture (P1), a clonogenic assay was performed to determine cell propagation properties in a clonal manner. Each clone was seeded in passage P3 and used for serial passaging until proliferative capacity (represented by grey bars) was exhausted (maximum P8). Ahi fractions showed a higher proliferative capacity than Alo and A− fractions. The Ahi CD34− fraction showed a higher proliferative capacity than the Ahi CD34+ fraction. (d) Rhodamine B staining of the clonogenic assay at passage P2 (d1–d2) and microscopic aspect of typical clones obtained at passage 3 (d3–d4). The Ahi CD34− fraction (d2) showed a higher clonogenic capacity than the Ahi CD34+ cells (d1). CD34+ fraction had colonies that were small and stratified (d3, arrows) when compared to the big colonies (composed of small undifferentiated cells) of the Ahi CD34− fraction (d4). A representative experiment with cell fractions obtained from one biopsy (B111, Table S1) is shown. Scale bars: 12 mm (black), 100 μm (white).

CD34± fractions show no clear phenotypic difference in short-term assays

HFSCs reside in a niche of which their progeny migrates to regenerate adjacent tissue. Accordingly, cell migration has been used as a functional assay to distinguish HFSC and precursor cells (21). To determine in vitro migration capacity, Ahi keratinocytes were incubated for 8 h and their migration recorded for the following 12 h (Fig. 2). Keratinocytes of the Ahi CD34+ (Video S1), Ahi CD34− (Video S2) and Ahi unsorted (Video S3) fractions migrated similar distances (Fig. 2a). Per cent migrating cells was higher in the Ahi CD34− fraction (Fig. 2b), although overall the difference was deemed not significant.

Figure 2.

 Functional characterization of the highly adherent (Ahi) CD34± cell fractions. (a) Migration assay. Cells were plated into ECM-covered dishes and maintained in the incubator for 8 h. Photographs of one field were taken for each condition every 2 min, during 15 h, using a confocal microscope coupled to a cell incubation system that maintained atmosphere at 37°C and 5% CO2. Distance covered by each cell between every measurement was registered, and the average of the distance covered between two measurements was calculated. Results for two independent biopsies are shown (B150, B161; Table S1; a1–a2). There were no statistically significant differences between the migration rate of the three cell fractions analysed (Ahi unsorted, Ahi CD34+ and Ahi CD34−). (b) Per cent migrating cells was measured for each cell fraction. Results for two independent biopsies are shown (b1–b2). The percentage of migrating cells was lower in the Ahi CD34+ fraction when compared to the Ahi CD34− and Ahi unsorted fractions. (c) Epidermal regeneration capacity of the CD34+ and CD34− fractions in organotypic cultures. (c1) When CD34+ total cell fraction was seeded as an epidermal component, no epidermis was formed. (c2) A completely stratified epidermis was formed when the total CD34− fraction was seeded as an epidermal component. (c3–c4) When Ahi CD34+ (c3) and Ahi CD34− (c4) cell fractions were used as epidermal component, both Ahi keratinocyte fractions were able to form a fully stratified epidermis, suggesting that both epidermal cell fractions are equally capable of short-term epidermal regeneration. The enrichment cells of dermal origin in the total CD34 +  fraction could explain the inability of these cells to give rise to stratified epidermis unless the Ahi subfraction is specifically selected. (d) Terminal differentiation as demonstrated by immunofluorescent staining with anti-Loricrin antibody. Loricrin is a well-characterized component of the cornified cell envelope in the uppermost layers of the epidermis. Organotypic cultures of the AhiCD34− (d1), AhiCD34+ (d2) and Ahi Unsorted (d3) cell fractions were plated on transwell supports on top of an artificial, fibrin-based dermis with embedded fibroblasts. Cells were allowed to proliferate for 3–5 days in submerged culture and to stratify for 15–17 days in the air–liquid interface. Organotypic skin sections were then stained with anti-Loricrin antibody (green) and cell nuclei counterstained with Hoescht (blue). Dermoepidermal junction is represented by a dashed line. A representative experiment with cell fractions obtained from one biopsy (B161, Table S1) is shown. Scale bars: 100 μm.

Organotypic cultures have been used to maintain human epidermal stem cells long-term in vivo (20,32) and in vitro (19). To characterize CD34± fractions in organotypic culture, whole skin CD34+ and CD34− fractions were magnetically purified and seeded on a fibrin dermis with embedded fibroblasts (33). Cells from the total CD34+ fraction did not generate an epidermis (Fig. 2c, panel c1). In contrast, cells from the total CD34− fraction generated an stratified epidermis, with a well-organized basal stratum and differentiated cell layers stratifying over it, as did the unsorted cells (Fig. 2c, c2; and data not shown). These results suggested the CD34− fraction contained keratinocytes able to generate full-thickness epidermis. Because of the fact that CD34+ cells included a high percentage of dermal cells, Ahi fractions were then tested. Interestingly, both CD34+ and CD34− Ahi fractions generated fully stratified epidermis (Fig. 2c – c3, c4 and Fig. 2d). These data suggested that CD34+ and CD34− keratinocyte fractions are functionally equivalent in cell migration and short-term tissue reconstitution assays.


Classical stemness assays such as clonogenic ability and in vitro proliferation capacity show a marked difference that suggests human CD34− cells are stem/early progenitor cells, while CD34+ cells might be a transit-amplifying precursor for HF sheath cells. On the other hand, CD34+ cells of human HF, which reside in suprabulbar outer root sheath (ORS), are functionally equivalent to CD34− cells residing in HF bulge and interfollicular epidermis, at least with regard to their migratory properties and ability to generate an epidermis in short-term organotypic cultures. Long-term follow-up of Ahi CD34± skin equivalents transplanted in vivo should further clarify the observed phenotypic similarities and differences.


We thank Alvaro Meana, Fernando Larcher and Marcela del Río for helpful experimental suggestions, Clara I. Rodriguez for help with cell migration experiments, and Gabriel Márquez, Angel G. Martín and Rosario Sánchez-Pernaute for critical reading of the manuscript. This work was supported by grants from Diputación Foral de Gipuzkoa (OF94/2008 and OF808/2004), the Spanish Ministry of Science and Innovation (PSE-090100-2006-3 and PSE-010000-2008-7), and Instituto de Salud Carlos III (FIS PI052614) from Spain. A.I. was supported by a “Miguel Servet” contract from the Spanish Ministry of Science and Innovation (CP03/00056). A.G-R. was supported by a Fellowship from the Fundación Ramón Areces (Spain).

Note Added in Proof

While this manuscript was in revision, Inoue et al. (Lab Invest 89:844) reported that FACS-purified human CD200+ CD34− bulge cells had much higher colony-forming abilities than the CD34+ population. These findings further support a HF stem cell-like role for the bulge-residing CD200+ population when compared to CD34+ cells in the sub-bulge and lower ORS regions.