Dermatology Branch, National Cancer Institute, Bethesda, Maryland, USA

Correspondence: Jonathan C. Vogel, MD, Dermatology Branch National Cancer Institute, NIH Bldg. 10/Room 12N260, 10 Center Drive MSC 1908, Bethesda, MD 20892-1908, USA. Email: jonvogel@mail.nih.gov

Received 5 May 2004; Revised 11 June 2004; Accepted 30 June 2004; Published online 7 October 2004.

To the Editor:

Reliable in vivo assays to assess the repopulation ability of stem cell candidates are essential for progress in stem cell research. We have designed an in vivo competitive repopulation assay that recapitulates human epidermis on immunocompromised mice, and uses MHC-Class I markers to distinguish and evaluate the relative abundance of long-term repopulating keratinocytes and their progeny in the epidermis. The accuracy and durability of this assay over time is demonstrated.

The remarkable progress in hematopoietic stem cell research has been driven by the development of sophisticated in vivo repopulation assays (^{Mazurier et al, 2003}). Research on stem cells in other tissues, including epidermis of the skin, has been hampered by the lack of reliable in vivo assay systems. Improved in vivo repopulation assays that are accurate, durable, and utilize bioengineered skin grafts in the correct anatomical location are needed to characterize keratinocyte stem cells (KSC;^{Schneider et al, 2003;Li et al, 2004}).

To develop an improved competitive in vivo repopulation assay that allows us to evaluate the relative abundance of human KSC, we used grafts of bioengineered human skin equivalents (HSE;^{Kolodka et al, 1998;Kuhn et al, 2002;Pfutzner et al, 2002}) composed of competing keratinocyte populations that can be distinguished by the presence or absence of an intrinsic genetic marker, MHC-class I HLA-A2. Since HLA-A2 is positive in 30%–50% of the US population (^{Natali et al, 1989}), competing keratinocyte populations, either HLA-A2-positive or -negative, can easily be obtained. HSE composed of known numbers of HLA-A2-positive and -negative cells are grafted onto immunocompromised mice to establish stable HSE, and the ratios between A2-positive and -negative cells in the HSE are monitored by an HLA-A2-specific monoclonal antibody and flow cytometry (FACS). Since each keratinocyte stem cell is believed to produce equivalent numbers of differentiated progeny, extending upward from the basal layer toward the skin surface (^{Potten and Allen, 1975;Kolodka et al, 1998}), the relative percentages of A2-positive and -negative keratinocytes in HSE after many epidermal turnovers accurately reflect the numbers of KSC in the initial A2-positive and -negative populations. Initial populations that are enriched for KSC will yield a greater percentage of keratinocytes within the HSE in long-term studies. To demonstrate, however, that this in vivo assay is accurate and suitable for competitive repopulation, it has to first be proved that A2-positive and -negative cells repopulate the grafted HSE equally well and are able to maintain their original ratio following many epidermal turnovers.

The HSE used in our study were produced by a raft culture-grafting model (^{Pfutzner et al, 2002}); human keratinocytes were seeded onto a dermal equivalent that was made of bovine type I collagen (Organogenesis, Canton, Massachusetts) and contained primary human dermal fibroblasts (Figure 1a and b). HSE grafted on the backs of recipient animals were stably maintained without loss or shrinkage for more than 12 mo (Figure 1c). Immunohistochemistry with anti-human involucrin showed that the keratinocytes within the HSE were exclusively of human origin and had the histology of normal human epidermis (data not shown). At the time of analysis, the whole HSE was excised with the surrounding mouse skin, and processed with dispase and trypsin to prepare epidermal cell suspension. For FACS analysis, living human cells (a blue box in Figure 1d) were identified with anti-HLA-A, B, C, and 7-AAD (Figure 1d), and anti-HLA-A2, clone BB7.2, was used to evaluate the percentages of HLA-A2-positive and -negative cells (^{Brodsky et al, 1979}; Figure 1e).

Figure 1.

The in vivo assay system to compare the relative abundance of repopulating cells in two populations of epidermal cells. Human epidermal cells were prepared from six pieces each of HLA-A2-positive and -negative neonatal foreskins by enzymatic treatments (^{Pfutzner et al, 1999}). The equivalent numbers of A2-positive and -negative cells were mixed, and 600,000 cells were seeded onto a dermal equivalent prepared by established methods to make human skin equivalents (HSE;^{Garlick and Taichman, 1994}; a). HSE showed histology of close resemblance to the normal human skin (b; E: epidermis; D: dermis). HSE were grafted on the backs of immunocompromised mice 3 d after seeding epidermal cells. The grafted HSE (pigmented patch on panel c) were chased for 5–26 wk and harvested by wide excision containing the surrounding mouse skin. An epidermal cell suspension was prepared by the enzymatic treatments and stained with anti-HLA-A, B, C (FITC-conjugate; BD Biosciences Pharmingen, San Diego, California), anti-HLA-A2 (clone BB7.2; BD Biosciences Pharmingen) and 7-AAD (BD Biosciences Pharmingen) to identify living cells. Goat anti-mouse IgG_2b conjugated with AlexaFluor 647 (Molecular Probes, Eugene, Oregon) was used as a secondary antibody for BB7.2. Cells were analyzed using FACSCalibur and CellQuest software (BD Biosciences Immunocytometry Systems, San Jose, CA). Living human cells gated with a blue box (d) were evaluated for the percentages of HLA-A2-positive and -negative cells (e).

Full figure and legend (109K)

To test our hypothesis that A2-positive and -negative keratinocytes repopulate the HSE equally well and can stably maintain the original A2-positive/negative ratio in HSE, we grafted a mixture with equivalent numbers of A2-positive and -negative cells, which presumably contain equivalent numbers of KSC, onto three strains of immunocompromised mice (Swiss nude, NOD-scid and scid-beige), and assessed the percentages of A2-positive cells by FACS at 26 wk, corresponding to more than six complete epidermal turnovers (Figure 2). In all mouse strains tested, no significant change in the percentages of A2-positive cells occurred during the time course; p values were 0.59, 0.73, and 0.75 for Swiss nude, NOD-scid, and scid-beige, respectively. There was also no significant difference between three strains of mice at 5, 10, and 26 wk; p values were 0.58, 0.40, and 0.29, respectively. In the real experimental setting to test the long-term repopulation potential of stem cell candidates, the test population (stem cell candidate) with an A2-positive phenotype would be mixed with the A2-negative control population (or vice versa). Prior to grafting the HSE, the accurate plating efficiency of the A2-positive and -negative populations needs be confirmed (e.g., by FACS) to ensure that the in vivo assay accurately measures the long-term repopulating ability of the initial A2-positive and -negative populations. If the test population is truly enriched for stem cells, whereas the control is not, the percentage of A2-positive cells should increase over time to predominate the HSE.

Figure 2.

The system stably maintained the 1:1 ratio between two populations of cells that should contain the equivalent numbers of repopulating cells. The percentages of HLA-A2-positive cells in the HSE, started with 47%, remained constant on the backs of immunocompromised mice (Swiss nude and scid-beige: Taconic, Germantown, NY; NOD-scid: The Jackson Laboratory, Bar Harbor, Maine) for 5–26 wk (n=2–3 at each time point in each strain). Averages (SEM) in one of two independent sets of experiments with similar results are shown.

Full figure and legend (23K)

The use of an intrinsic HLA marker in our system has advantages over cell-marking methods using viral vectors. In contrast to variable percentages of keratinocytes expressing transduced viral markers, pure keratinocyte populations that are completely A2-positive or -negative can easily be obtained. Also, intrinsic markers avoid the risk of introducing a growth advantage or disadvantage caused by extrinsic molecules, and minimize the in vitro manipulation that could harm the primitive nature of stem cells. Additionally, HLA-A2 differences can be detected on a genomic level even if gene expression is lost. This system also permits serial transplantation experiments, which have been critical for demonstrating the self-renewal capacity of hematopoietic stem cell candidates. The value of this assay is that the repopulation ability of different KSC candidates can be accurately compared in an anatomically correct in vivo setting. The animal studies described in this manuscript were approved by the National Cancer Institute (NCI) Animal Care and Use Committee (ACUC).