Stem Cell Properties of Normal Human Keratinocytes Determine Transformation Responses to Human Papillomavirus 16 DNA

ABSTRACT

Human papillomavirus (HPV) infection of the genital tract is common; however, only about 10 to 15% of infections persist, and approximately 10 to 15% of these persistent infections result in cancer. Basal epidermal stem cells are the presumed target cells for HPV infection, providing a reservoir of latently infected cells that persist over time and initiate lesions. However, it is not known whether stem cell density has any influence on transformation of human keratinocytes by HPV. We explored the relationship between stem cell properties of normal human keratinocytes and their susceptibility to transformation by HPV16 DNA. Normal human keratinocyte isolates (NHKc) derived from different donors were cultured in three-dimensional anchorage-free suspension to assess their spheroid-forming ability. NHKc spheroids were then plated back into plastic monolayer culture and transfected with full-length HPV16 DNA, which we have previously shown to integrate into the host cell genome upon transfection. Spheroid-derived NHKc (SD-NHKc) and fluorescence-activated cell sorting-purified populations of basal stem-like keratinocytes, expressing low levels of epidermal growth factor receptor and high levels of integrin alpha 6 (EGFR^lo/ITGα6^hi), responded to transfection with HPV16 DNA with more vigorous proliferation, greater immortalization efficiency, and faster progression to differentiation resistance than autologous mass-cultured cells. Conversely, cells committed to terminal differentiation (EGFR^hi/ITGα6^lo) grew slowly after transfection with HPV16 and failed to generate immortalized or DR clones. HPV16 DNA induced stem cell properties in mass-cultured NHKc. We conclude that HPV16 preferentially immortalizes basal keratinocytes with stem cell properties and that these cells readily achieve a differentiation-resistant phenotype upon immortalization by HPV16.

IMPORTANCE This paper explores the relationship between the stem cell properties of normal human epidermal cells in culture and these cells' susceptibility to transformation by HPV16 DNA, the HPV type present in about 50% of cervical cancers. We report variable susceptibilities to HPV16-mediated transformation among different keratinocyte isolates derived from neonatal foreskin. Our findings provide strong experimental evidence that HPV16 preferentially transforms basal keratinocytes with stem cell properties. Insights gained from these studies increase our understanding of the host cell-specific factors influencing individual susceptibility to HPV-driven transformation and the contributing factors leading to preneoplastic and neoplastic progression of HPV-positive lesions.

INTRODUCTION

Human papillomavirus (HPV) infections are thought to arise in epithelial regions rich in stem/progenitor-like cells, such as the squamocolumnar junction of the cervix, the lingual and palatine tonsils in the oropharynx, and the transitional zone in the anal-rectal canal (1–3). The target cell of HPV infection remains unclear, but infection requires virus particles to gain access to the epidermal basal layer, where they can infect basal keratinocytes and induce aberrant cellular proliferation (4–6). Therefore, keratinocyte stem cells (KSC) have long been postulated as the target cells of HPV infection (4, 7–9). KSC are characterized by their spheroid-forming abilities and intense expression of basal epidermal markers, such as low levels of epidermal growth factor receptor (EGFR) and elevated levels of integrin alpha 6 (INTα6) (10–13), a transmembrane glycoprotein expressed in the basal compartment of the epidermis that is postulated to be the main receptor for HPV16 (14–16).

We have previously reported that normal human foreskin keratinocytes (NHKc) isolated from different donors vary in their basal levels of EGFR and in their sensitivities to HPV16-mediated immortalization (17): lower EGFR levels in the basal compartment of the epidermis in foreskin tissue corresponded with increased keratinocyte proliferation after infecting individual NHKc isolates with a retrovirus encoding the HPV16 E6 and E7 oncogenes, while NHKc isolates expressing higher basal EGFR levels responded to infection with sluggish proliferation (17). Others demonstrated that HPV E6- and E7-expressing keratinocytes display stem-like properties and greater sphere-forming abilities (18), indicating that a basal stem-like phenotype is also selected for during immortalization by HPV. Moreover, recent findings have demonstrated that HPV oncogene expression in basal KSC can contribute to neoplasia (3, 8). However, a nonbasal subpopulation of cells found in the transformation zone (TZ) has also been reported to be responsible for cancer initiation in HPV-driven cancers (19). Thus, the precise role of epidermal stem cells in HPV-driven neoplasia continues to be a matter of debate. The question remains as to the relationship between the density of basal stem/progenitor cells in an epithelium and susceptibility of that epithelium to HPV-mediated transformation.

In our model system for HPV16-mediated human cell transformation, monolayer cultures of NHKc are stably transfected with full-length HPV16 DNA to give rise to HPV16-immortalized lines (HKc/HPV16) (20–22). Selection of HKc/HPV16 lines in medium supplemented with 5% serum and 1 mM CaCl₂ yields differentiation-resistant, preneoplastic cell lines (HKc/DR) which exhibit gene expression patterns similar to those found in cervical cancer (22, 23). Upon exogenous overexpression of the homeobox gene SIX1, HKc/DR acquire tumorigenic potential in nude mice (24), while SIX1 does not induce tumorigenicity in early-passage, immortalized HKc/HPV16 (25). However, upon transfection with exogenous SIX1, HKc/HPV16 lines acquire the HKc/DR phenotype (24). Thus, SIX1 is a key mediator of the HKc/HPV16 transition to HKc/DR, the hallmark of preneoplastic progression in HPV16-immortalized keratinocytes (20, 25–27).

In this study, we sought to determine the primary cellular phenotypes associated with HPV16-mediated immortalization and in vitro progression of HKc/HPV16 toward an HKc/DR phenotype. Using our in vitro model system, we explored in detail the relationship between basal stem/progenitor-like keratinocyte density in primary epidermal NHKc cultures and the susceptibility of these cultures to HPV-mediated immortalization and transition to HKc/DR. We hypothesized that cultures rich in epidermal stem cells (EpSCs) would be considerably more sensitive to HPV16-mediated immortalization and may also be more efficient at undergoing transition to HKc/DR upon immortalization with HPV16 DNA than mass-cultured cells. To this aim, we transfected progenitor/stem-like NHKc cultures, and autologous NHKc mass cultures, from several different individuals with the full-length HPV16 DNA and assessed growth responses and immortalization efficiencies in vitro. We found that NHKc cultures enriched in stem-like cells responded to transfection with greater proliferation and more readily achieved HKc/DR status when subcultivated in the presence of 5% serum and 1 mM CaCl₂. Moreover, HPV16 DNA conferred epidermal stem/progenitor-like cell characteristics to most, but not all, successfully transfected NHKc lines and stimulated the expression of several genes involved in basal progenitor cell proliferation. Conversely, autologous NHKc cultures enriched for cells committed to terminal differentiation (TD) were considerably more resistant to immortalization by HPV16 DNA and failed to establish proliferating clones in serum and high-calcium-containing media. Our findings reveal various transformation susceptibilities among different keratinocyte populations and show that basal stem-like cells play a vital role in HPV16-mediated preneoplastic initiation and progression. These data strongly suggest that an increase in stem/progenitor-like cells within keratinocyte cultures leads to increased susceptibility to HPV16-mediated transformation.

RESULTS

NHKc spheroids produce functionally viable colonies with increased proliferative potential when replated in monolayer culture.

To investigate the ability of NHKc derived from skin explants of different donors to escape anoikis when cultured in suspension, cells were plated at a density of 2 × 10⁴/well in keratinocyte serum-free medium-stem cell medium (KSFM-scm) in 96-well round-bottomed plates coated with a semisolid polymerized mixture of agarose and KSFM-scm (Fig. 1A). Under these conditions, cells produce self-aggregating nonadherent three-dimensional spheroids of reproducible sizes (Fig. 1B). We found that some individual NHKc isolates aggregated into multicellular spheroids in suspension culture (spheroid-forming HKc, or SF-HKc), while others consistently failed to form spheroids under the same culture conditions (non-spheroid-forming HKc, or NF-HKc). To assess cell viability, we used trypan blue dye exclusion staining to measure the ratio of viable cells in NF-NHKc and SF-NHKc suspension cultures. We found that suspension cultures from SF-NHKc maintained significantly greater cell viabilities than NF-NHKc (Fig. 1C). We then tested the proliferation capacity of suspension cultures from SF-NHKc and NF-NHKc isolates by plating them back into polystyrene cell culture dishes (Fig. 1D). We observed that spheroids from SF-NHKc readily attached to the cell culture dish and gave rise to proliferating cells that formed numerous holoclonal (28) colonies, while NF-NHKc from suspension culture produced abortive colonies (Fig. 1E). We next examined the colony-forming efficiency (CFE) of replated cell suspensions by quantifying the ratio of Giemsa-stained colonies to the number of cells seeded. We found that cells derived from SF-NHKc spheroid suspensions (SD-NHKc) were 74.5-fold more efficient at generating colonies than suspensions from NF-NHKc (Fig. 1F). This indicates that spheroid-derived NHKc populations that demonstrated resistance to anoikis have increased proliferative potential in monolayer culture.

FIG 1.

NHKc spheroids produce functionally viable colonies with increased proliferation potential. (A) Schematic overview of the spheroid formation assay. NHKc are isolated from fresh foreskin explants (a) and cultivated in 100-mm dishes (b) and then passaged into 96-well round-bottomed plates coated with a polymerized mixture of agarose and KSFM-scm (c). Cells then generate suspension cultures of reproducibly sized spheroids in each individual well. (B) Cells (2 × 10⁴) cultured in KSFM-scm are floating onto the polymerized soft agarose coating individual wells of a 96-well plate. Cell suspensions from non-spheroid-forming NHKc (NF-NHKc) (left) and spheroid-forming NHKc (SF-NHKc) (right) isolates are shown. Images were taken on a Nikon TMS phase microscope using Infinity Analyze software. (C) Proportion of viable cells in NF-NHKc and SF-NHKc cell suspensions after 24 h. (D) Schematic illustrating a suspension spheroid (a) being plated back into an empty uncoated 2D polystyrene cell culture dish (b). (c) After 15 to 20 days, the adherent culture gives rise to a monolayer of spheroid-derived cells (SD-NHKc) which are cultivated and maintained in 2D culture. (E) Giemsa staining of suspension cultures from NF-NHKc and SF-NHKc after replating in 2D culture. (F) Colony-forming efficiencies (C.F.E) of 2D-plated cell suspensions from non-spheroid-forming (NF) and spheroid-forming (SF) NHKc isolates. Bars indicate standard deviations, and *, **, and *** indicate statistically significant differences, with P values of ≤0.05, <0.01, and <0.001, respectively.

Spheroid-derived NHKc are enriched in P63/K14 double-positive cells that maintain subapoptotic (low) EGFR levels in culture.

To further assess the growth potential of SD-NHKc in adherent culture, we performed extensive clonal analysis using SD-NHKc generated after spheroids were transferred to two-dimensional (2D) monolayer culture. We observed that small cells migrated out of spheroids plated in plastic dishes to form a continuous monolayer of cells (Fig. 2A and A1). After a few rounds of subcultivation in adherent culture, SD-NHKc progenies maintained the cobblestone appearance typical of actively proliferating NHKc (Fig. 2B), whereas clones generated from mass cultures acquired the morphology of senescent keratinocytes after 15 population doublings (PD) (Fig. 2C). To determine the basal epidermal status of SD-NHKc progenies, we assessed their nuclear expression of P63 and cytoplasmic expression of basal cytokeratin 14 (K14) by immunofluorescence (Fig. 2D). We found that over 60% of SD-NHKc clones expressed nuclear P63 or basal K14, whereas less than 20% of clones generated from corresponding mass-cultured cells expressed K14 and only 10% expressed nuclear P63 (Fig. 2E). SD-NHKc cultures also contained 26 times more K14/P63-coexpressing cells than their mass-cultured counterpart, suggesting a marked enrichment of stem/progenitor-like keratinocytes in the spheroid-derived cultures (Fig. 2E). We next measured levels of mRNAs encoding pan-P63, cytokeratin 14, and EGFR and found a 4.6-fold increase in P63 mRNA levels and a 2.1-fold increase in K14 mRNA levels in SD-NHKc compared to those of their corresponding mass cultures. EGFR mRNA levels in SD-NHKc were not significantly different from those of corresponding mass cultures (Fig. 2F). To further examine EGFR expression in SD-NHKc, we measured their cell surface levels of EGFR as well as those of corresponding monolayer cultures using fluorescence-activated cell sorting (FACS) analysis. We found that cell surface EGFR expression increased over 100-fold in mass cultures after 2 rounds of subcloning in cells maintained in monolayer culture but remained comparatively stable in SD-NHKc from the same NHKc isolate (Fig. 2G). Higher cell surface EGFR levels in mass cultures also corresponded to a loss of spheroid formation ability, elongated cell morphologies, and cell senescence. In contrast, secondary SD-NHKc cultures retained spheroid-forming abilities (Fig. 2H), accumulated more PD, and consisted of small-sized cells that could be subcultivated for more than 10 weeks as monolayers. These observations mirror previous reports describing low cell surface levels of EGFR in NHKc as a feature of stem-like keratinocytes and intense upregulation of cell surface EGFR as a precursor of terminal differentiation (10, 11). Taken together, these data indicate that SD-NHKc are a proliferating population of basal keratinocytes with characteristics of stem/progenitor epidermal cells.

FIG 2.

Spheroid-derived cells display several attributes of stem/progenitor-like keratinocytes. (A) A single multicellular spheroid was transplanted back into a plain uncoated 2D polystyrene cell culture dish. Arrows depict direction of cell outflow from the spheroid onto the dish surface. Scale bar, 100 μm. (A1) Magnified image of SD-NHKc actively migrating and proliferating out of the spheroid. Scale bar, 20 μm. Adherent SD-NHKc (B) and their corresponding mass-cultured (Mass cul.) cells (C) after 15 cumulative population doublings in monolayer culture at ×20 magnification. Scale bar, 50 μm. (D) SD-NHKc and their corresponding mass cultures immunostained with antibodies against basal cytokeratin 14 (K14; green) and pan-tumor protein 63 targeting all TP63 isoforms (P63; red) at ×40 magnification. Scale bar, 25 μm. (E) The ratio of P63, K14, and P63/K14 coexpressing cells was quantified by measuring the numbers of stained cells per visual field. (F) The expression of mRNA encoding pan-TP63 (all P63 isoforms), cytokeratin 14, and EGFR in SD-NHKc (passage 1) relative to corresponding mass cultures (passage 0) as determined by RT-PCR. The segmented line indicates mRNA expression levels in controls (monolayer mass cultures) to which a value of 1 was assigned. Data were normalized to GAPDH expression and reported in relation to their controls. Each bar represents the means ± standard deviations from triplicate experiments using three different keratinocyte isolates. (G) Cell surface expression of EGFR measured by fluorescent-activated cell sorting (FACS) in SD-NHKc and their corresponding mass cultures after the initial passage (P1), passage 2 (P2), and passage 3 (P3) in monolayer culture. (H) Spheroid-forming assay conducted on SD-NHKc and their corresponding SF-NHKc mass cultures after 3 passages in monolayer culture. (I) Representative flow cytometry data illustrating the use of the epidermal cell markers integrin alpha 6 (ITGα6) and EGFR to purify desired cell populations within SD-NHKc and autologous mass cultures. (J) Fraction of EGFR^lo/ITGα6^hi, EGFR^hi/ITGα6^hi, and EGFR^hi/ITGα6^lo cells in mass cultures of individual NHKc isolates and their corresponding SD cultures. (K) Cumulative population doublings (P.D) observed in EGFR^lo/ITGα6^hi, EGFR^hi/ITGα6^hi, and EGFR^hi/ITGα6^lo cells over the course of 50 days in monolayer culture.

Given that SD-NHKc intensely expressed basal epidermal markers, we suspected that their prolonged growth potential was primarily sustained by a distinct fraction of KSC. To explore this possibility, we used FACS to identify cell populations expressing various levels of EGFR and ITGα6 within SD-NHKc cultures and autologous monolayer cultures. We found that nearly 20% of cells in SD-NHKc cultures retained low levels of EGFR and high levels of ITGα6 (EGFR^lo/ITGα6^hi) (Fig. 2I and J). In contrast, EGFR^lo/ITGα6^hi populations only made up 3% of cells within autologous mass cultures (Fig. 2I and J). To explore the growth potential of the cell populations represented in spheroid-forming cultures, we sorted and collected the EGFR^lo/ITGα6^hi, EGFR^hi/ITGα6^hi, and EGFR^hi/ITGα6^lo fractions from SF-NHKc mass cultures and cultivated them as monolayers. We found that EGFR^hi/ITGα6^lo cells accumulated the fewest number of PD in culture (∼3) and generated dividing colonies for approximately 18 days in culture (Fig. 2K). EGFR^hi/ITGα6^hi cells accumulated a total of 11.3 PD in culture and continued to generate dividing colonies for nearly 30 days (Fig. 2K). Interestingly, EGFR^lo/ITGα6^hi cells were slow growing during the first 15 days in culture and then proliferated at a rate exceeding that of the other isolated cell fractions (Fig. 2K). Furthermore, EGFR^lo/ITGα6^hi cells displayed the greatest growth potential of all FACS-isolated cells and generated dividing colonies for at least 50 days in culture (Fig. 2K). We further characterized the EGFR^lo/ITGα6^hi and EGFR^hi/ITGα6^lo fractions to compare their abilities to form holoclonal colonies, their expression of keratinocyte stem cells markers, and their spheroid-forming ability (Fig. 3A to I) and determined that the EGFR^lo/ITGα6^hi cells formed holoclonal colonies, expressed keratinocyte stem cell markers, and formed spheroids when cultured in suspension (Fig. 3A to I). Taken together, these results clearly indicate that suspension culture conditions select for EGFR^lo/ITGα6^hi cells, which have characteristics of keratinocyte stem cells. In addition, we asked how long FACS-isolated EGFR^lo/ITGα6^hi cells would retain their spheroid-forming ability and EGFR^lo/ITGα6^hi phenotype after successive cycles of spheroid formation in suspension and replating in mass culture (Fig. 3J to M). We found that spheroid-forming ability was fully retained for at least two cycles of spheroid formation and replating in monolayer cultures but was lost by cycle 4 (Fig. 3J and K). In parallel, cells with the EGFR^lo/ITGα6^hi phenotype were depleted by the fourth cycle (Fig. 3L and M). Keratin 14 expression also progressively decreased, in parallel with the loss of stem-like cells in the population, during the spheroid-forming and replating cycles (Fig. 3K).

FIG 3.

FACS-sorted EGFR^lo/ITGα6^hi cells express stem cell markers, form spheroids in suspension culture, and retain their ability to form spheroids for at least two cycles of spheroid formation and replating in mass culture. (A) FACS-sorted EGFR^lo/ITGα6^hi cells were plated into a 6-well plate at a density of 10,000 cells/well and cultured until approximately 80% confluence. Colony density was assessed by Giemsa staining. (B) Phase-contrast image of proliferating EGFR^lo/ITGα6^hi cell progenies in adherent culture. (C) EGFR^lo/ITGα6^hi cell progenies immunostained with antibodies against pan-tumor protein 63 targeting all TP63 isoforms (P63, red) and basal cytokeratin 14 (K14, green). Scale bar, 100 μM. (C1 to C3) Cells stained with the nuclear stain DAPI (C1) and for nuclear P63 (C2) and cytoplasmic keratin 14 (C3). (C4) Magnified (×40) image of a single EGFR^lo/ITGα6^hi cell progeny immunostained for nuclear P63 (red) and cytoplasmic K14 (green). Scale bar, 20 μM. (D) EGFR^hi/ITGα6^lo colonies stained with Giemsa. (E) Phase-contrast images of EGFR^hi/ITGα6^lo cell progenies in adherent culture. (F) EGFR^hi/ITGα6^lo cells immunostained with antibodies against pan-tumor protein 63 targeting all TP63 isoforms (P63; red) and basal cytokeratin 14 (K14; green). Scale bar, 100 μM. (F1 to F3) Cells stained with the nuclear stain DAPI (F1) and for nuclear P63 (F2) and cytokeratin 14 (F3). (F4) Magnified (×40) image of a single EGFR^hi/ITGα6^lo cell progeny immunostained for nuclear P63 (red) and cytoplasmic K14 (green). Scale bar, 20 μM. (G) The expression of mRNAs encoding TP63 (all isoforms) and cytokeratin 14 in EGFR^hi/ITGα6^lo and EGFR^lo/ITGα6^hi cells relative to unsorted populations as determined by RT-PCR. Data were normalized to GAPDH expression and are reported as means ± standard deviations. (H and I) Phase-contrast images of suspension cultures from EGFR^hi/ITGα6^lo (H) and EGFR^lo/ITGα6^hi (I) cells isolated from an SF-NHKc strain. Scale bar, 100 μM. (J) Schematic depicting the spheroid passaging assay. (K) Phase-contrast images of NHKc spheroid cultures after 4 cycles of spheroid passaging. Scale bar, 100 μM. (L) Representative flow cytometry data illustrating gated fractions of EGFR^lo/ITGα6^hi, EGFR^hi/ITGα6^hi, and EGFR^hi/ITGα6^lo cells detected in SD-NHKc after 1 and 5 passage cycles in culture. (M) Quantification of sorted cell fractions in SD-NHKc after 1 and 5 passage cycles.

Spheroid-derived keratinocytes display vigorous proliferation after transfection with HPV16 DNA.

To gain insight into SD-NHKc responsiveness to transfection by HPV16 DNA, each individual multicellular spheroid was plated onto an uncoated 60-mm dish and maintained until confluence (Fig. 4A). SD-NHKc (1 × 10⁵ cells/well) and corresponding mass-cultured cells were then passaged into a 6-well plate and transfected with the plasmid pMHPV16d, which contains two copies of the HPV16 genome in a head-to-tail configuration within the pdMMTneo vector, or with the pdMMTneo vector only and cotransfected with a plasmid transiently expressing enhanced green fluorescent protein (eGFP) to monitor transfection efficiency (Fig. 4B and C). To establish successfully transfected clones, HKc/pMHPV16d and HKc/pdMMTneo were serially passaged at low density (1:500 dilutions) in 100-mm dishes until their HKc/pdMMTneo control cells senesced (Fig. 4D and E). SD-NHKc responded to transfection with robust proliferation and underwent nearly 20 PD within 20 days in culture, compared to the ∼10 PD acquired by their mass-cultured HKc/HPV16 counterparts in the same time span (Fig. 4F). Furthermore, SD/HPV16 reached complete colony confluence within 10 days in culture, while HKc/HPV16 established from mass cultures failed to achieve complete confluence in the same time span (Fig. 4G and H). However, all HPV16-transfected cells ultimately produced immortalized lines, while all vector-transfected controls eventually senesced. These findings indicate that enriched populations of stem/progenitor-like keratinocytes transfected with HPV16 DNA have higher proliferative potential than their mass-cultured counterparts.

FIG 4.

SD-NHKc isolates respond to HPV16 transfection with vigorous cell proliferation. (A) Schematic depicting a single NHKc spheroid in suspension (a) transplanted back into an uncoated 60-mm dish, and then 1 × 10⁵ spheroid-derived cells (SD-NHKc) (b) were passaged into each well of a 6-well plate for transfection with the full-length HPV16 genome (pMHPV16d) or vector only (pdMMTneo) (c). (B) NHKc transfected with 1 μg of pMHPV16d or pdMMTneo vector control and cotransfected with GFP (3:1 wt/wt) to monitor transfection efficiency. Scale bar, 20 μm. (C) GFP expression was analyzed 48 h posttransfection, and fluorescent cells were quantified by ImageJ. n.s, not significant. Transfected cells were selected with G418 (100 mg/ml) for 48 h (D) and repeatedly subcultured until vector-transfected cells senesced (E). Scale bars, 50 μm (D) and 1 cm (E). (F) Cumulative population doublings of SD-HKc/HPV16 isolates derived from four different individuals, along with the corresponding mass-cultured HKc/HPV16 and respective HKc/pdMMTneo controls. Bars indicate standard deviations, and an asterisk indicates statistically significant P values of ≤0.01. (G) Cells (5 × 10³) were plated in a 100-mm dish after successful transfection, and colony patterns were examined by Giemsa staining at day 10. (a and b) SD-HKc/HPV16 isolate (a) and corresponding mass-cultured HKc/HPV16 isolate after 10 days in culture (b). (c and d) Phase-contrast images of SD-HKc/HPV16 colonies (c) and corresponding HKc/HPV16 mass-cultured colonies (d). Scale bars, 1 cm and 50 μm, respectively. (H) Fold increase in total colony area was determine by quantifying the percentage of dish surface area occupied by SD-HKc/HPV16 colonies after 10 days in culture relative to the corresponding mass-cultured HKc/HPV16 isolate.

Full-length HPV16 DNA abrogates suspension-induced cell death and triggers a basal stem/progenitor-like program in NHKc.

Given that the stem/progenitor-like SD-NHKc cultures were markedly more responsive to transfection by HPV16 DNA, we investigated the influence of HPV16 DNA on keratinocyte stem cell characteristics. We observed that while NHKc isolates transfected with pdMMTneo senesced after 5 passages in culture (Fig. 5A, a and b), all 15 HKc/HPV16 lines established from mass-cultured keratinocytes grew continuously and maintained small-sized cell morphology throughout their time in culture (Fig. 5B, c and d). Moreover, while protein levels for P63 and K14 decreased considerably in HKc/neo after 5 passages (Fig. 5C, e), HKc/HPV16 lines established from mass cultures maintained intense expression of these two markers through multiple rounds of subcultivation (Fig. 5C, f). We have previously demonstrated that the HPV16 E6 and E7 oncogenes increase EGFR mRNA levels and overcome mechanisms by which excessive EGFR signaling shortens NHKc life span (17, 29). Therefore, we investigated cell surface EGFR expression in HKc/HPV16 lines established from monolayer cultures of NHKc isolates of known spheroid-forming potential by FACS. We detected an ∼10-fold increase in cell surface EGFR in all HKc/HPV16 lines, i.e., lines established from SF-NHKc isolates and NF-NHKc isolates alike, relative to HKc/neo (Fig. 5D). These findings reveal that regardless of innate spheroid-forming ability, transfection by HPV16 DNA induces increased EGFR expression. Given that all HKc/HPV16 lines intensely expressed EGFR yet continually maintained a phenotype typical of basal keratinocytes, we decided to measure their mRNA levels of genes responsible for epidermal basal stem/progenitor cell maintenance and proliferation. We found a 17.3-fold induction of P63, a 9.6-fold induction of EGFR, an 11.3-fold induction of Ki-67, and an 11.0-fold induction of ALDH1 in HKc/HPV16 relative to HKc/neo controls (Fig. 5E). This indicates that genes responsible for basal stem/progenitor cell proliferation and maintenance are activated in HPV16-transfected cells, while mechanisms of senescence and differentiation are inhibited. We determined the effects of HPV16 DNA on NHKc spheroid formation: we found that SF-NHKc retained their spheroid-forming abilities after transfection with HPV16 DNA (Fig. 5F). Strikingly, most NF-NHKc isolates acquired spheroid-forming abilities after transfection with HPV16 DNA (NF/HPV16^Δpos) (Fig. 5F, c and d), with a few isolates resisting spheroid formation despite being transfected with HPV16 DNA (NF/HPV16^Δneg) (Fig. 5F, e and f). In total, of the 15 established HKc/HPV16 lines, 100% of those derived from SF-NHKc isolates (5/5) retained spheroid-forming abilities after transfection with HPV16 DNA, and 80% of NF-NHKc isolates (8/10) acquired spheroid-forming abilities (Table 1). These findings suggest that HPV16 DNA strongly promotes epidermal stemness in NHKc.

FIG 5.

HPV16 DNA induces basal stem/progenitor cell characteristics in mass-cultured NHKc. (A) Mass-cultured keratinocytes after two (a) and five (b) passages following transfection with pdMMTneo vector. (B) Corresponding HKc/HPV16 cell line after 2 (c) and 25 (d) passages following transfection with HPV16 DNA. (C) HKc/neo colony (e) and a corresponding HKc/HPV16 colony (f) immunostained with antibodies against basal cytokeratin 14 (K14; green) and pan-tumor protein 63 targeting all TP63 isoforms (P63; red) after 5 and 25 passages in culture, respectively. DAPI (blue) depicts nuclear staining. Scale bar, 100 μm. (D) Flow cytometry analysis of cell surface EGFR levels in mass-cultured HKc/HPV16 and HKc/neo lines established from spheroid-forming (SF) and non-spheroid-forming (NF) NHKc isolates. (E) Expression levels of mRNAs encoding pan-TP63, EGFR, Ki-67, and ALDH1 in HKc/HPV16 relative to HKc/neo cells, both at passage 5, as determined by RT-PCR. Data were normalized to GAPDH expression and reported as means ± standard deviations. Bars indicate standard deviations, and *, **, and *** indicate statistically significant P values of ≤0.05, <0.01, and <0.001, respectively. (F) NHKc suspension cultures from SF-NHKc (a) and NF-NHKc (c and e) isolates before transfection and after transfection with full-length HPV16 DNA (b, d, and f). (d) Spheroid-forming ability acquired (^Δpos) by a keratinocyte isolate that was initially NF. (f) Spheroid-forming inability maintained (^Δneg) by an NF keratinocyte isolate despite being transfected with full-length HPV16 DNA. Scale bar, 100 μm.

TABLE 1.

Distribution of keratinocyte isolates with spheroid-forming ability before and after transfection with HPV16 DNA

Isolate	Spheroid-forming isolates (n = 5) [no. positive/total no. (%)]	Non-spheroid-forming isolates (n = 10) [no. positive/total no. (%)]
		NFΔpos (n = 8)	NFΔneg (n = 2)
NHKc	Yes, 5/5 (100)	No, 0/8 (0)	No, 0/2 (0)
HKc/HPV16	Yes, 5/5 (100)	Yes, 8/8 (80)	No, 0/2 (20)

Stem/progenitor-like keratinocytes are preferentially immortalized by HPV16 DNA.

Given that HPV16 DNA stimulated robust proliferation in SD-NHKc cultures and conferred basal stem-progenitor-like properties in NHKc mass cultures, we decided to assess the immortalization response to HPV16 of various FACS-purified basal and nonbasal keratinocyte populations (Fig. 6A and B). We found that ITGα6^high/EGFR^low cells displayed the most rapid growth response after transfection with HPV16 DNA, followed by ITGα6^high/EGFR^high cell populations and ITGα6^high cells. Conversely, ITGα6^low/EGFR^high and ITGα6^low cells consistently failed to proliferate beyond 25 days in culture and usually senesced after accumulating 5 PD, despite successful transfection with HPV16 DNA (Fig. 6C and D). To challenge these findings, we made critical dilutions (1 to 5 cells/well) of various HKc/HPV16 lines established from FACS-purified populations and seeded them into individual uncoated wells of 96-well plates. Resultant colonies were repeatedly subcultured and assessed for colony renewal (Fig. 6E). We found that ITGα6^low/EGFR^high-HPV16 and ITGα6^low-HPV16 populations had difficulty sustaining repeated subcloning at critical dilutions (Fig. 6F). Conversely, ITGα6^high/EGFR^low-HPV16 cells gave rise to many dividing clones even after six rounds of subcloning at critical dilutions. This sustained clonal growth for several PD after G418 selection is indicative of stable viral DNA integration and immortalization, as previously described by our laboratory (20, 21). Notably, ITGα6^high/EGFR^low-HPV16 cells produced the greatest number of stably immortalized clones that sustained the subcultivation protocol with greater fitness. We next assessed the immortalization responses of various HKc/HPV16 lines established from mass cultures (i.e., NF^Δneg/HPV16, NF^Δpos/HPV16, and SF/HPV16). We found that theses cultures sustained clonal passaging with difficulty, while HKc/HPV16 lines established from spheroid-derived cultures (SD/HPV16) gave rise to a significantly greater number of stably immortalized clones (Fig. 6F). When quantifying the ratio of successfully proliferating colonies generated from each line, we found that lines derived from mass-cultured NHKc isolates consistently achieved less than 25% immortalization efficiency, with no significant difference seen between NF^Δneg/HPV16 and NF^Δpos/HPV16 isolates (Fig. 6G). Strikingly, SD/HPV16 cultures achieved over 2-fold greater immortalization efficiencies (57%) than corresponding mass cultured SF/HPV16 (23%) (Fig. 6G). SD/HPV16 also generated greater numbers of confluent colonies that were more densely stained by Giemsa than colonies from mass-cultured HKc/HPV16 lines (Fig. 6F), suggesting that immortalized stem/progenitor-like populations are more efficient at self-renewal and maintain greater proliferative potential throughout their time in culture. Our data indicate that differentiating cells are considerably more resistant to immortalization by HPV16 DNA, while keratinocyte stem-like cells are significantly more susceptible to immortalization by the full-length HPV16 genome. Furthermore, our findings reveal that the innate spheroid-forming phenotype present in certain primary NHKc isolates may give them a growth advantage in culture and slightly increase immortalization efficiency by HPV16 DNA. However, purification of stem/progenitor-like cells from primary keratinocyte cultures, whether spheroid-forming or not, dramatically increases immortalization efficiency by HPV16 DNA.

FIG 6.

Stem/progenitor-like keratinocytes are preferentially immortalized by HPV16 DNA. (A) Fluorescence-activated cell sorting (FACS) of NHKc populations isolated based on cell surface levels of epidermal cell markers integrin alpha 6 (ITGα6) and EGFR. (B) Gated fractions of ITGα6^hi- and ITGα6^lo-expressing cells purified by FACS. (C) Cumulative population doublings observed in EGFR^lo/ITGα6^hi, EGFR^hi/ITGα6^lo, EGFR^hi/ITGα6^hi, ITGα6^hi, and ITGα6^lo cells after transfection with full-length HPV16 DNA. (D) FACS-isolated HPV16-tansfected cells were plated into a 100-mm dish at a density of 5,000 cells/dish until cells reached approximately 80% confluence. Cells were then serially passaged 1:500 for several passages after HKc/neo cells senesced, and cells were counted at each passage to derive PD number. (E) Schematic depicting the limiting dilution assay used to assess immortalization efficiency. HPV16-transfected cells are seeded at a critical density of approximately 1 to 5 cells per well in individual wells of an uncoated 96-well plate. Upon confluence, initial colonies are then serially transposed at a 1:5 ratio into new wells of a 96-well plate until clonogenecity potential is exhausted or stably immortalized clones are formed. (F) Immortalization efficiencies of HKc/HPV16 lines established from various FACS-purified populations, as well as NF, SF, and SD isolates. Col., colony. (G and H) Immortalization efficiencies of 8 different HKc/HV16 lines established from monolayer and spheroid-derived NHKc (G) and FACS-purified NHKc populations (H).

HKc/HPV16 lines established from stem/progenitor-like keratinocytes readily achieve differentiation resistance.

Given that resistance to differentiation stimuli, such as elevated calcium and serum levels, is a pivotal hallmark in the preneoplastic progression of HKc/HPV16 (20, 25–27), we chose to assess whether stably immortalized HKc/HPV16 clones established from our various cell lines could directly transition from growth in basal serum-free medium to growth in differentiation-resistant (DR) medium and whether they could sustain continual subcultivation at limited dilutions under this condition (Fig. 7A). We found that SD/HPV16 and ITGα6^high/EGFR^low-HPV16 cells gave rise to significantly more DR clones than ITGα6^high/EGFR^high-HPV16 and HKc/HPV16 lines established from mass cultures (Fig. 7B and C), suggesting that stem-like NHKc and KSC transition to differentiation resistance more readily than other keratinocyte cell fractions (Fig. 7D). We next assessed cell surface levels of EGFR in immortalized HKc/HPV16 and their corresponding HKc/DR lines relative to HKc/neo. Similar to levels seen in freshly established HKc/HPV16 (Fig. 5D), we detected a nearly 10-fold induction in cell surface EGFR intensity in established HKc/HPV16 relative to HKc/neo lines and a near 100-fold increase when HKc/HPV16 transitioned into HKc/DR relative to HKc/neo (Fig. 7G). When grown onto 100-mm monolayer dishes, we found that while HKc/HPV16 lines maintained a cobblestone-like morphology typical of basal epidermal cells (Fig. 7D, a), HKc/DR displayed a morphology reminiscent of fibroblastic cells (Fig. 7D, b). When assessing levels of gene products associated with epithelial-to-mesenchymal transition (EMT) in HKc/DR and corresponding HKc/HPV16 lines, we found that HKc/HPV16 expressed low cytoplasmic levels of fibronectin (FN1) and intense levels of E-cadherin on their plasma and nuclear membranes (Fig. 7E). Conversely, HKc/DR expressed intense cytoplasmic levels of fibronectin and markedly reduced levels of E-cadherin (Fig. 7E). Furthermore, HKc/DR acquired a spindle-shaped morphology similar to that of fibroblasts, while HKc/HPV16 conserved their round cobblestone-like shapes (Fig. 7E). We also observed that levels of E-cadherin mRNA were higher in HKc/HPV16 than HKc/DR, FN1 mRNA was expressed at lower levels in HKc/HPV16 than HKc/DR, and Snail and Twist mRNA were greatly increased in HKc/DR compared to the level in HKc/HPV16 (Fig. 7F). When assessing spheroid-forming abilities in HKc/DR from various keratinocyte isolates, we found that all HKc/DR lines formed spheroids in suspension (not shown). Strikingly, HKc/DR lines established from NF^Δneg/HPV16 also displayed spheroid-forming abilities, suggesting that spheroid formation is a phenotype selected for during in vitro progression of HPV16-immortalized cells (Fig. 7H). Histological examination of H&E-stained HKc/DR spheroid aggregates determined that they appeared remarkably similar to a preneoplastic epithelial lesion, with several cells exhibiting hyperchromatic features (Fig. 7I). Spheroid-derived HKc/DR formed larger spheroids than their corresponding spheroid-derived HKc/HPV16 lines (Fig. 8A). Moreover, wound-scratch assays determined that spheroid-derived HKc/DR could restore over 80% of the wounded area within 24 h, while the corresponding HKc/HPV16 restored only 25% of the wound over the same period (Fig. 8B and C). We next investigated the invasiveness of spheroid-derived HKc/DR lines by determining their migration in Matrigel. We found that spheroid-derived HKc/DR minimally invaded through Matrigel compared to the tumorigenic HKc/DR cell line HKc/DR-Six1, which we used here as a positive control (24) (Fig. 8D and E). These data indicate that spheroid-derived HKc/DR have undergone EMT but still lack invasive potential and that additional events are required for these HKc/DR lines to achieve invasiveness in culture and, ultimately, tumorigenicity in nude mice (24, 25).

FIG 7.

HKc/HPV16 established from stem cell-enriched cultures readily acquire differentiation resistance. (A) Schematic depicting the differentiation resistance assay: stably immortalized HKc/HPV16 colonies (a) are subcultivated 1:5 in differentiation resistance (DR) medium (KSFM; 1 mM calcium chloride and 5% FBS) (b) until selected clones arise that can continually sustain growth in DR media (c). (B and C) Proportion of differentiation-resistant clones observed in various stably immortalized HKc/HPV16 lines after the differentiation resistance assay. (D) Phase-contrast image of a stably immortalized HKc/HPV16 line established from SF-NHKc (a) and its corresponding HKc/DR line (b). Scale bar, 50 μm. (E) Intensity of cell membrane E-cadherin and cytosolic fibronectin evaluated by a Zeiss LSM 710 confocal microscope. (a to c) Magnified image of HKc/HPV16 cells (a) and corresponding HKc/DR cells (b and c). Scale bar, 20 μm. (F) Relative mRNA expression of genes involved in HKc/HPV16 transition to HKc/DR cells, as determined by RT-qPCR, relative to NHKc. Data were normalized to GAPDH expression and are reported as means ± standard deviations. (G) Flow cytometry analysis of cell surface EGFR levels in NHKc, stably immortalized HKc/HPV16, and HKc/DR. (H) Magnification (×200) of a spheroid aggregate growing from HKc/DR after 5 days in suspension culture. Scale bar, 100 μm. (I) Hematoxylin and eosin (H&E) staining of an HKc/DR spheroid aggregate at day 5 (×400 magnification). Preneoplastic DR cells are seen throughout the thickness of the spheroid section. Arrows point to cells displaying hyperchromatic and/or atypical nuclei. Scale bar, 1,000 μm.

FIG 8.

Immortalized and differentiation-resistant keratinocytes derived from spheroids are migratory noninvasive cells. (A) Spheroids formed by homologous, SD HKc/HPV16, and HKc/DR cell lines. (B) Scratch-wound assay depicting the migratory behaviors of established, SD HKc/HPV16, and HKc/DR lines after 24 h in culture. Black lines mark the original wound boundaries, and yellow lines mark the front of migrating cells. (C) Percentage of wound closure achieved 24 h after scratching SD HKc/HPV16 and HKc/DR monolayers. (D) Matrigel cell invasion by SD HKc/DR and tumorigenic HKc/DR-SIX1 cells in transwell tissue culture inserts. Invading cells were quantified after crystal violet staining. Images are shown at ×100. (E) The number of migrating cells was directly counted. Each column represents the means from five different fields. Bars indicate standard deviations, and * indicates statistically significant P values of ≤0.05.

DISCUSSION

Basal epithelial stem cells have widely been speculated to be targets of cancer initiation in a variety of tissues, including the cervix, skin, ovaries, breast, and prostate (30–34). In HPV-driven cancers, stem cells of the epidermis are presumed to be preferentially targeted and transformed by HPV (4, 7–9). These cells can be distinguished by intense expression of ITGα6 and reduced expression of EGFR (10, 13). We have previously shown that low basal levels of EGFR in NHKc are associated with an increase in proliferation after infection with retroviruses expressing the HPV16 E6 and E7 oncoproteins (17). Conversely, elevated basal levels of EGFR, a characteristic of committed progenitor cells (CP), are associated with less robust proliferative response to E6/E7; we later demonstrated that HPV16 oncoproteins cooperate to increase EGFR by overcoming mechanisms by which excessive EGFR signaling stimulates terminal differentiation by decreasing p53 levels, which in turn upregulates YY1, a positive regulator of EGFR expression (17, 29).

In this study, we describe the relationship between the stem cell properties of NHKc and their susceptibility to transformation by HPV16 DNA. We found that progenitor/stem-like populations respond to HPV16 transfection with robust proliferation. Moreover, transfection of NHKc with the full-length HPV16 genome induces basal stem-like characteristics in most keratinocyte isolates and generally confers spheroid-forming abilities and prolonged cell survival under anchorage-independent conditions. We observed that P63/K14 double-positive keratinocytes are strongly selected for after each round of subcloning of HKc/HPV16, and these cultures appear increasingly homogeneous with passaging, displaying uniformly small-sized cells, tightly packed in cobblestone-like patterns. Conversely, HKc/neo controls mostly contained loosely packed flattened cells with a low proliferative potential and considerably reduced expression of P63 and K14. We therefore speculated whether EpSCs present within keratinocyte cultures naturally outcompeted neighboring CP and TD cells during clonal selection or whether EpSCs were more efficiently immortalized by HPV16 DNA, which in turn enabled them to outcompete their CP and TD counterparts. We expected that if EpSCs naturally outcompeted neighboring cells, then HKc/neo would also increasingly be enriched for stem-like basal cells during the subcloning process, but we observed the opposite: P63/K14-expressing cells dramatically decreased in HKc/neo lines after each round of passaging, indicating that their EpSC reserves were continually depleted, as previously reported in NHKc cultures maintained on plastic (35, 36). Thus, the enrichment of basal stem/progenitor-like cells in HKc/HPV16 lines despite sustaining multiple rounds of passaging is indicative of better immortalization efficiencies in EpSCs, enabling them to successfully self-renew and outcompete neighboring CP and TD cells. Stem/progenitor-like enriched cultures are likely at a clonal advantage over mass-cultured cells, enabling them to exhibit better responses when transfected with HPV16 DNA. The successful establishment of stable HKc/HPV16 lines in culture is also likely a result of a few immortalized epithelial stem cells that are continually selected for throughout the clonal pressure of low-density passaging. This is supported by our repeated limited dilution experiments, which demonstrated that the long-term clonal growth of HKc/HPV16 was better sustained in lines established from SD/NHKc and EGFR^lo/ITGα6^hi populations than in autologous lines established from mass-cultured cells. Furthermore, our clonal dilution experiments revealed that CP and TD cells were unable to achieve stable immortalization irrespective of the spheroid-forming status of the NHKc isolate from which they were derived. This could be due to CP and TD cells having exited the cell cycle, as HPV16 requires cells to be actively dividing to carry forth immortalization events (37). Although previous studies from our laboratory have demonstrated that HPV16 DNA integration is consistent among HKc/HPV16 lines established from different individuals (20, 21, 24), differences in immortalization efficiencies observed in stem-like cultures relative to mass cultures could be indicative of greater viral integration events occurring in basal stem-like cells relative to CP and TD populations. Further studies will need to be carried out to carefully investigate differences in HPV16 DNA copy number and integration state between keratinocyte lines immortalized from stem/progenitor cells and those derived from mass cultures.

We and others have shown that when HKc/HPV16 are cultured in medium containing 1 mM CaCl₂ and 5% fetal bovine serum, the cells differentiate and cease to proliferate (21, 22, 26, 38, 39). Only a very few cells that can proliferate under these culture conditions emerge to give rise to HKc/DR lines. However, immortalized HKc/HPV16 lines established from populations of stem/progenitor-like keratinocytes readily achieved a DR status without necessitating a growth factor-independent (GFI) phase. Conversely, HKc/HPV16 lines established from autologous mass cultures were significantly less efficient at achieving DR status. Interestingly, the spheroid-forming abilities of these cells were not merely conserved but were considerably improved as the cells transitioned toward the DR phenotype. This is particularly striking in NF/HPV16^Δneg, whose HKc/DR clones displayed a robust spheroid-forming ability in suspension. Spheroid-forming cells are likely present in all HKc/HPV16 lines, including NF/HPV16^Δneg, but are continuously selected for during clonal passaging, thereby enabling them to achieve greater homogeneity and yielding a spheroid-forming phenotype as cells undergo in vitro progression. Given that SD-NHKc are actively proliferating populations of keratinocytes, their rapid clonal divisions may uniquely predispose them to HPV-mediated immortalization, as suggested by recent findings (40, 41). Moreover, the presence of hyperchromatic cells within HKc/DR aggregates indicates that they are further along the preneoplastic progression process than their HKc/HPV16 counterparts. Nonetheless, HKc/DR cells' noninvasiveness in Matrigel suggests that additional genetic events, in addition to a stem-like status, are required to achieve tumorigenic potential (20, 24, 25). These observations may help explain recent reports correlating successful tumor grafting and cancer progression in mice (42): it is possible that epithelial tissue dense in stem cells is more prone to transformation when exposed to an oncogenic insult, thus yielding clones with greater grafting potential in xenograft hosts.

Taken together, our findings reveal that, at least in culture, epidermal stem cells are not only readily immortalized but also acutely undergo DR transition when transfected with HPV16 DNA. The proliferative responses seen in SD/HPV16 established from several different donors exclude the possibility that variations in the host cell genetic background play a major role in producing these characteristics, indicating that such a response is a result of alterations at the transcriptome and epigenetic level. This supports the notion that HPV16-immortalized stem/progenitor-like cells, rather than cells committed to terminal differentiation, are poised to repopulate in culture, consistent with the cell competition model proposed in several reports (18, 43). Our data confirm epidermal stemness as a foremost phenotypic determinant of susceptibility to HPV-mediated immortalization and progression to differentiation resistance in primary human keratinocyte cultures, supporting recent observations by Hufbauer et al. (18). This sheds light on the contributing factors leading to preneoplastic and neoplastic lesion progression in HPV-infected individuals and suggests that an increased density in tissue stem/progenitor-like cells among certain individuals predisposes them to HPV-driven neoplasia. The assessment of tissue stem cell density in primary tissue could be explored as a predictive biomarker of susceptibility to HPV-mediated carcinogenesis, as this may help identify, among the many people who present with persistent HPV infections, the relatively few who are truly at risk for developing cancer. Further studies are warranted to define more clearly the role of epidermal tissue stem cell density and HPV-mediated transformation in vivo.

MATERIALS AND METHODS

Cell culture and cell lines.

NHKc were isolated from neonatal foreskin as previously described (17, 20, 21, 29) and cultured in modified keratinocyte serum-free medium (KSFM) supplemented with 20 ng/ml EGF, 10 ng/ml basic fibroblast growth factor, 0.4% bovine serum albumin (BSA), and 4 μg/ml insulin, adapted from the spheroid-forming (44) and sphere-propagating (45) media. This medium is referred to as KSFM-stem cell medium (KSFM-scm). NHKc were passaged into individual wells of a 6-well plate (∼200,000 cells/well) and transfected with 1 μg of a plasmid containing two copies of the full-length HPV16 genome (pMHPV16d) or with the vector control (pdMMTneo) as previously described (20, 21). All cells were cotransfected with the pMSCV-IRES-EGFP plasmid (1:3, wt/wt, HPV or control plasmid) to monitor transfection efficiency. HPV16 immortalized cell lines (HKc/HPV16) were established from NHKc isolates derived from 15 different human donors. Ten of these NHKc isolates were unable to form spheroids in suspension culture (NF-NHKc), while five were able to form spheroids in suspension (SF-NHKc). Proliferating spheroid-derived cells (SD-NHKc) were trypsinized, passaged into 6-well plates, and transfected to establish HPV16-immortalized spheroid-derived lines (SD/HPV16). Another group of HPV16-immortalized lines was established from fluorescence-activated cell sorting (FACS)-purified keratinocyte populations from three individual donors (see below for FACS conditions). Cells (2 × 10⁴) from respective FACS-sorted populations were directly seeded into individual wells of a 6-well plate and allowed to grow until ∼75% confluence. Confluent FACS-sorted cells were then transfected by following the protocol described above. Moreover, differentiation-resistant cells (HKc/DR) were obtained from each stably immortalized line (HKc/HPV16) by subcultivating in KSFM supplemented with 1 mM calcium chloride and 5% fetal bovine serum (FBS) (21). All cell lines were routinely subcultivated at 1:500 when confluent, and medium was changed 24 h after passaging and every 48 h thereafter.

Clonal growth assay.

Cells were plated in duplicate 100-mm dishes at low density (5,000 cells/dish) and fed 20 ml of KSFM-scm every 10 days. Cultures were then serially passaged at low density in 100-mm dishes to determine cumulative population doubling (PD) over time. Cell numbers and viability were obtained by utilizing a Countess automated cell counter (Invitrogen, CA). Cumulative population doublings were calculated according to the formula PD = (log N/N₀)/log₂, where N represents the total cell number obtained at each passage and N₀ represents the number of cells plated at the start of the experiment (46). Matching dishes were fixed in methanol at room temperature and stained with Giemsa (10%) to assess colony size and morphology.

Immortalization and transformation assay.

HKc/HPV16 lines were serially diluted to a final concentration of ∼1 to 5 cells per 200 μl and seeded into individual wells of duplicate uncoated 96-well plates. Colonies were serially plated every 20 days into new wells of two 96-well plates at a 1:5 ratio until their proliferation potential was exhausted or stably immortalized clones were established. Matching duplicate plates were fixed in methanol and stained with Giemsa (10%). The ratio between the number of clones remaining at the end and the initial number of clones observed at the first passage was taken as a measure of immortalization efficiency. Such clones were considered stably immortalized (HKc/HPV16). A similar in vitro limiting dilution assay was performed to assess the efficiency of HKc/HPV progression to HKc/DR, but this time experiments began with HKc/HPV16 lines and cells were continually maintained in KSFM supplemented with 1 mM calcium chloride and 5% FBS (DR medium) (21).

Spheroid formation and spheroid aggregation assay.

Cultured NHKc (2 × 10⁴) were seeded into a 96-well round-bottom plate coated with a polymerized mixture of agarose and KSFM-scm to a final concentration of agarose of 1% (wt/vol) in medium. The spheroids were maintained for at least 24 h in suspension and then used for further experimentations. Cell viability was assessed by trypan blue staining, while spheroid size and morphology were determined by image analysis with Lumenera Infinity1 software (Lumenera Corporation, Ottawa, ON, Canada). For bulk spheroid aggregate formation, 2 × 10⁶ cells from individual clonally transformed HKc/DR lines were plated onto a 100-mm dish coated with 3.0 ml of a polymerized mixture of agarose (0.5% final concentration) and DR medium. Cells were allowed to aggregate overnight, and spheroid size was monitored for 10 days in suspension culture. Size and morphology were determined by imaging and Lumenera Infinity1 analysis (Lumenera Corporation). At day 5 and day 10, single spheroid aggregates were isolated, fixed in 2% paraformaldehyde (PFA), sectioned, and counterstained with hematoxylin and eosin (H&E). Sections were imaged on a Nikon Eclipse E600 microscope.

Real-time PCR.

Total RNA was isolated from cells using the All Prep DNA/RNA minikit (Qiagen, CA) according to the manufacturer's protocol. Reverse transcription was carried out with 1 μg of total RNA using the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA). Real-time PCR was performed using iQ SYBR green supermix (Bio-Rad) by following the manufacturer's instructions. Amplicon products were validated by gel electrophoresis on a 2% agarose gel. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control. All samples were assayed in triplicate. The sequences of the primers used for real-time PCR are shown in Table 2.

TABLE 2.

Primers used for real-time PCR

Gene name	Primer sequence (5′-3′)
	Forward	Reverse
ALDH1	GCACGCCAGACTTACCTGTC	CCTCCTCAGTTGCAGGATTAAAG
BIRC5	AGGACCACCGCATCTCTACAT	AAGTCTGGCTCGTTCTCAGTG
CDH1	CGAGAGCTACACGTTCACGG	GGGTGTCGAGGGAAAAATAGG
EGFR	AGGCACGAGTAACAAGCTCAC	ATGAGGACATAACCAGCCACC
FN1	CGGTGGCTGTCAGTCAAAG	AAACCTCGGCTTCCTCCATAA
GAPDH	GGAGCGAGATCCCTCCAAAAT	GGCTGTTGTCATACTTCTCATGG
K14	TGAGCCGCATTCTGAACGAG	GATGACTGCGATCCAGAGGA
KI-67	ACGCCTGGTTACTATCAAAAGG	CAGACCCATTTACTTGTGTTGGA
KLF4	CCCACATGAAGCGACTTCCC	CAGGTCCAGGAGATCGTTGAA
SNAIL	TCGGAAGCCTAACTACAGCGA	AGATGAGCATTGGCAGCGAG
SIX1	ATTCTCACCTCCCCAAAGTC	ACTTAGGACCCCAAGTCCAC
TP63	GGACCAGCAGATTCAGAACGG	AGGACACGTCGAAACTGTGC
TWIST	GTCCGCAGTCTTACGAGGAG	GCTTGAGGGTCTGAATCTTGCT
β-Actin	CATGTACGTTGCTATCCAGGC	CTCCTTAATGTCACGCACGAT
ΔN TP63	ATGTTGTACCTGGAAAACAATGCC	CAGGCATGGCACGGATAAC

Immunohistochemistry.

Cells were grown on coverslips coated with polylysine until 75% confluent. The cells were then washed twice in ice-cold phosphate-buffered saline (PBS), fixed with 4% paraformaldehyde for 20 min at room temperature, permeabilized with 0.5% Triton X-100 in 1% glycine, and then blocked using 0.5% bovine serum albumin (BSA) and 5% goat serum for 30 min at room temperature. Samples were next incubated with antibodies against P63 (1:200 dilution; Thermo Scientific), cytokeratin 14 (1:200 dilution; Santa Cruz Biotechnology), fibronectin (1:200 dilution; Santa Cruz Biotechnology), or E-cadherin (1:200 dilution; Santa Cruz Biotechnology) in blocking solution overnight at 4°C. Samples next were washed three times with PBS–0.05% Tween 20 (PBST) followed by incubation with fluorescein isothiocyanate (FITC)- and Alexa 568-conjugated secondary antibodies (at 1:1,000 dilution; Invitrogen). Nuclei were stained with a 1:5,000 dilution of 4′,6-diamidino-2-phenylindole (DAPI) (Invitrogen) before cells were mounted. Samples were observed using a Nikon Eclipse E600 or Zeiss confocal laser scanning microscope.

FACS.

Cells (2 × 10⁶ to 4 × 10⁶ cells/ml) were stained with FITC-conjugated anti-integrin alpha 6 (ITGα6) (Abcam, Cambridge, MA) and phycoerythrin (PE)-conjugated anti-EGFR (BD Pharmingen San Jose, CA). FACS-sorted populations were collected by centrifugation at 112 × g for 5 min, the supernatant was removed, the cell pellets were resuspended in fresh KSFM-scm, and then they were seeded into individual wells of a 6-well plate for subsequent experiments. FACS analysis was performed using a BD FACSAria II flow cytometer (BD Biosciences, San Jose, CA) and analyzed with Flowing Software (V2.5.1; Turku, Finland).

Cell migration and invasion assay.

HKc/HPV16 and their respective HKc/DR lines were grown to ∼80% confluence on 60-mm dishes in a humidified 5% CO₂ incubator at 37°C. To generate wounding, a single linear scratch was created using a 200-μl micropipette tip. Repair of the wound area was monitored every 6 h, and images were processed by Lumenera Infinity1 software analysis (Lumenera Corporation, Ottawa, ON, Canada). The invasive ability of HKc/DR and HKc/DR-Six1 was measured as previously described (24). Briefly, 2 × 10⁴ cells were seeded onto the upper surface of a Matrigel matrix in a transwell dish; cells that invaded through the matrix and into the lower chamber were fixed and stained with 1% crystal violet and counted in five randomly chosen areas. Each experiment was performed in triplicate wells and repeated three times.

Statistical analysis.

Data are expressed as the means ± standard deviations (SD). Differences between mean values were analyzed using Student's t test. P values of <0.05, <0.01, or <0.001 were considered statistically significant and are indicated in the figures by one, two, or three asterisks, respectively.