4.3 Discussion
5.2.2 Oral Keratinocyte Stem Cell Isolation and Validation
To further understand how FOXM1 expression is regulated during keratinocyte stem cell expansion, this study aimed to directly isolate putative oral keratinocyte stem cells. This would allow keratinocyte stem cell
population expansion in vitro in a controlled environment, where FOXM1 expression could be directly assessed.
Enrichment of keratinocyte stem cells has been previously demonstrated in
vitro, based on the differential expression of cell surface markers. Immature
populations of basal epidermal keratinocytes have been identified by differential expression of β1 integrins. Integrin β1-positive fluorescence activated cell sorting (FACS) sorted fractions of epidermal keratinocytes that adhere rapidly to extracellular matrices (Collagen IV and Fibronectin), possess greater regenerative and proliferative capacity in vitro (Watt, 1998). Whereas, human oesophageal and oral mucosal keratinocyte stem cells have been previously shown to express high levels of the low affinity nerve growth factor receptor, p75NTR (Okumura et al., 2003; Nakamura et al., 2007).
In order to investigate the regulation of FOXM1 in oral keratinocyte stem cells (OKSCs), fluorescence activated cell sorting (FACS) was used to isolate specific population subsets that express either integrin β1 or p75NTRor both (Fig. 5.2A). Secondary cultures of normal human oral keratinocytes (population doublings, PD 6-9) were used for fluorescence activated cell sorting (FACS) and equal numbers from each sorted population were seeded diretly onto mitomycin C-treated 3T3 feeder layers, and in vitro clonogenic assays were performed, to examine their proliferative potential. To verify the FACS procedure integrin β1+ bright/p75NTR- (represented on Fig. 5.2A as fraction P3) and integrin β1+ bright /p75NTR+ (represented on Fig. 5.2A as fraction P4) populations were examined under fluorescence microscopy immediately after plating (Fig. 5.2B).
Figure 5.2: Oral keratinocyte stem cell isolation.
(A) FACS staining profile of primary human oral keratinocytes stained with anti- integrin β1 (CD29) and anti-p75NTR (CD271) antibodies. Populations marked as P1, P2, P3 and P4 were used for combined cell sorting analysis. The relative abundance of each population is derived from experiments performed on oral keratinocytes derived from three different donors and were carried out in triplicates. (B) Fluorescence microscopy on oral keratinocyte populations immediately after flow sorting to demonstrate successful discrimination between the indicated populations.
In line with previous observations, p75NTR+ cells possessed a superior
clonogenic capacity compared to p75NTR- (Figure 5.3A i, ii, and Figure 5.3
B) cells when examined 12 days after flow sorting. Although p75NTR- cells gave rise to some large colonies, many of them were highly irregular and displayed characteristics of terminal differentiation (Figure 5.3 Aii). The p75NTR+ cells gave rise mostly to large colonies with a smooth perimeter (Figure 5.3 Aii). The latter colony morphology signifies that such populations originate from the cells of a holoclone, which contains cells of high self- renewal and proliferative capacity (Barrandon and Green, 1987).
Figure 5.3: Oral Keratinocytes sorted for p75NTR display high clonogenic capacity
(A) Oral keratinocytes sorted for p75NTR- or p75NTR+ were allowed to grow for 12
days. Keratinocyte colonies were examined under bright field microscopy (i) and were later stained with Rhodamine B (ii) for quantification of the clonogenic potential of each separate keratinocyte population. (B) Colony growth values for each different keratinocyte population were obtained by means of area coverage (pixels) as described in materials and methods. Clonogenic assays were carried out in 6 replicate wells for each sample. RS stands for random sorted control cells. Integrin β1 positive keratinocytes were further discriminated according to p75NTR positivity. p75NTR expression determined the clonogenic potential of integrin β1 (+) keratinocytes. Values represent the fold difference from p75NTR- control samples ±SEM for each
average values. *P≤0.05, **P≤0.01, ***P≤0.001
Keratinocytes with the highest integrin β1 expression levels, were observed mainly in the p75NTR+ subset (represented as fraction P4 in Figure 5.1 A), which is consistent with the fact that β1 integrins are highly expressed in the stem cell containing basal layers of human epidermis (Watt, 1998). However,
integrin β1 positivity conferred an advantage only when combined with p75NTR positivity, since only integrin β1+ bright / p75NTR+ (Figure 5.2 A fraction P4) cells showed high levels (comparable to integrin β1+ dim / p75NTR+ (P3) or p75NTR+ alone) of clonogenic potential (Figure 5.3 B, and
Appendix 4), colony forming efficiency % (CFE), population doublings (PD)
(Figure 5.4 B) and total cell output (Figure 5.4 A). As expected, all random sorted (RS; double stained keratinocytes randomly sorted from the same pool of cells) population showed intermediate levels between the (-) and (+) populations, indicating a mixed population.
Figure 5.4: Proliferative potential of distinct keratinocyte subsets
(A) Total cell output values were calculated as mentioned in materials and methods section after 12 days of culture. RS stands for random sorted control cells. Values are representative of two independent experiments carried out in duplicates. Bars represent the average fold difference from p75NTR- control samples ±SEM.*P≤0.05,
**P≤0.005, ***P≤0.001. (B) Clonogenicity % and population doublings (PD)
measurements of all sorted oral keratinocytes after 12 days in culture. RS stands for random sorted populations. Values are representative of three independent experiments carried out in duplicate wells, ±SEM for each average value.
Next, the molecular profile of putative human OKSC (p75NTR+ cells) was investigated by qPCR. The mRNA expression levels of p75NTR were significantly elevated (~3.5 fold), in the p75NTR+ subsets proving successful FACS sorting (Figure 5.5 Ai, ii). Confluent cultures of lethally treated 3T3 feeders were also included in expression analysis to control for possible contamination during cDNA preparations. Although 3T3 feeders were routinely removed prior to experiments, there is a possibility that a small amount of 3T3 cells may interfere with expression analysis. However, qPCR
results confirmed that lethally treated 3T3 feeders did not express detectable mRNA levels of either gene (see Figures 5.5 A, 5.6 A and 5.7)
By examining the levels of a well established keratinocyte stem cell marker, ΔNp63α (Pellegrini et al., 2001), shortly after cell sorting, it was found that ΔNp63α protein was slightly more abundant in both the p75NTR+ and integrin β1+ bright / p75NTR+ ) compared to the p75NTR- population (Figure 5.5 Bi). As expected, random sorted (RS) control cells retained intermediate levels of ΔNp63α protein (Figure 5.5 Bi). The differences in ΔNp63α protein levels became much more pronounced after 12 days in culture, indicating the long term ability of OKSC derived populations to maintain the expression of stem cell associated markers (Figure 5.5 Bii). This comes in agreement with the first report identifying ΔNp63α as a keratinocyte stem cell marker, which showed that it was still present in holoclone (but not in meroclone or paraclone) derived populations of limbal and epidermal keratinocytes, after 11-12 days in culture (Pellegrini et al., 2001).
Further qPCR expression analysis of freshly isolated p75NTR+ oral keratinocytes, revealed significantly higher expression levels of two other putative stem cell genes NESTIN and NUMB, when compared to non-stem cell fractions (Figure 5.5 C). NESTIN has been shown to be neuronal stem cell marker (Lendahl et al., 1990) while NUMB was reported as a maker of asymmetric division in mouse epidermal keratinocytes in vivo (Clayton et al., 2007) and mouse satellite muscle progenitor cells in vitro (Shinin et al., 2006). The expression of cytokeratin 4 which is a marker specific for differentiating oral epithelia (Moharamzadeh et al., 2007) was also reduced in the stem cell enriched fractions of oral keratinocytes (see Appendix 5A). Collectively, these experiments validate the use of p75NTR as a stem cell marker for isolating human oral keratinocyte stem cells which are undifferentiated, highly clonogenic and proliferative.
Figure 5.5: Molecular profile of putative oral keratinocyte stem cells
(A) Absolute qPCR analysis of p75NTR expression levels in sorted oral keratinocytes three days after flow sorting. All values are representative of three independent experiments. Graphs show single (p75NTR) (i) or double-sorted (p75NTR/initegrin β1) cells (ii). Bars represent the average fold difference from p75NTR- (arbitrary
value of 1) control samples ±SEM.*P≤0.05, **P≤0.005, ***P≤0.001. (B) OKSCs retain high levels of ΔΝp63α protein after short and long term culture. Oral keratinocyte sorted populations were harvested for total protein at 72 hours (i) and 12 days (ii) and immunoblotted with anti-p63 antibody recognizing human ΔΝp63 isoforms, and β-tubulin as protein loading control. (C) Absolute qPCR analysis of p75NTR, NESTIN, and NUMB in p75NTR+ oral keratinocytes. All values are fold expression relative to p75NTR- cells (arbitrary value of 1). Each bar represents the mean ± SEM of two independent experiments performed in duplicates. .*P≤0.05,
**P≤0.005, ***P≤0.001.