Background
High risk papillomavirus (HPV16 and 18) have been widely studied due to its association with cervical cancer and tumors of other anogenital sites [
1]. These are small un-envolved virus with a circularized double stranded DNA genome, containing eight open reading frames that are expressed either early (E1, E2, E4, E5, E6 and E7) or late (L1 and L2) during the infection, and a noncoding region known as long control region (LCR) which contains the viral early promoter, the enhancer and the origin of replication [
2,
3].
The HPV life cycle is intimately linked to epithelial differentiation [
1]. HPV normally infects dividing cells at the basal layer of the epithelium but completes its life cycle by amplifying progeny DNA genomes in the spinous layer and carrying out viral encapsidation in the uppermost epithelial layer named the granular layer [
4].
One of the proteins that plays a crucial role in the papillomaviruses life cycle is the E2 protein. It possesses a DNA-binding domain and a transactivation domain that are linked by a serine-arginine-rich hinge region [
5]. E2 as a homodimer binds the cognate sequences ACCGN4CGGT (E2-binding sites [E2BSs]) in the viral LCR, and different experimental systems have shown that it regulates the viral E6 and E7 oncogenes transcription, acting as an activator or as a repressor depending on the protein levels and the specific sites contacted [
6,
7]. However, the evidence points to a repressive function of E2 in controlling the viral early promoter [
8]. On the other hand, another important role of E2 is to participate as an auxiliary replication factor, since at the origin of DNA replication, E2 interacts with and loads the HPV E1 replication helicase, recruiting then the cellular DNA replication machinery [
8‐
11]. E2 also has an crucial role in segregation of viral episomal genomes during the division of infected cells by interacting with chromatin adapter proteins that tether them to host mitotic chromosomes [
8].
In addition, possibly by an indirect mechanism involving its interaction with several cellular proteins, the expression of E2 also has an impact on the cellular transcription profile, affecting the expression of genes involved in key processes such as proliferation, differentiation, apoptosis and senescence [
12‐
20].
The biological effects of E2 on mammalian cells have been generally studied using transiently transfected cancer cell lines. Early studies in HPV-positive cervical carcinoma derived cell lines [
21], indicated that transient overexpression of E2 induces apoptosis in these cells due to repression of endogenous E6/E7 expression [
22‐
27]. However, also in HPV negative cells, E2 overexpression induces apoptosis by both transactivation independent and mediated through activation of caspase 8 [
28,
29]. It has been also observed that sustained expression of E2 in HPV positive cells, induces a prolonged growth arrest and induces irreversible senescence [
30‐
33], probably as a default pathway for cells which exit the cell cycle, since most of these cells are incapable of terminal differentiation.
Then, the observed effects of E2 on the processes of apoptosis and senescence could be related to its role in differentiation, since it is well known that a critical step to start this process is an arrest in the cell cycle followed by the expression of genes involved in early differentiation. If these genes are not correctly expressed, cells could keep in an irreversible senescence status and progress later to apoptosis or immortalization programs [
34,
35].
Considering the technical difficulty to establish a convenient cellular stratification system to follow step by step the progress of the infection, most of the studies on the effects of HPV gene expression on epithelial differentiation process have been performed on carcinoma derived cell lines [
36‐
38], or immortalized keratinocytes from several tissues [
39‐
41].
In fact, the induction of cell differentiation has been demonstrated in HPV16-E2 stably transfected HaCaT cells, showing a typical differentiated morphology, cell elongation and multilayer colonies growth, besides the expression of the classical keratinocytes differentiation markers involucrin, filagrin and cytokeratins 1 and 10 [
34].
However, since E2 is one of the first expressed genes in the epithelial basal layer during infection [
42,
43], the understanding of its effects on early differentiation must be studied on cellular models that mimic this layer in the epithelium, considering the cellular subpopulations that constitute it, including epithelial stem cells.
Different strategies have been used to identify these cells from epithelial derived transformed cell lines and keratinocytes primary cultures, such as the expression of the membrane molecules desmoglein-3 and β1-integrin, the nuclear presence of p63, or metabolic characteristics of these cells, such as the over-expression of aldehyde-dehydrogenase enzyme (ALDH) or the ATP-binding cassette subfamily G (ABCG) transporter pump [
44‐
47]. Nevertheless, the simultaneous detection of α6-integrin (CD49f) and the transferrin receptor CD71 levels has demonstrated in a very confident form a hierarchical organization in keratinocytes primary cultures with the presence of three cellular subpopulations [
48].
In the present work, we have taken advantage of the α6-integrin and CD71 detection and cell separation described by Schluter et al. [
48], and demonstrated that HaCaT cell line derived from spontaneously immortalized keratinocytes, where HPV genomes are absent and have been extensively used to simulate epithelial tissues, possess cellular subpopulations with similar characteristics to those from the stratified epithelial basal layer, corresponding to: cells in an early differentiation process α6-integrin
dim (8.16 ± 0.52%), transitory amplifying cells α6-integrin
bri/CD71
bri (87.27 ± 1.21%) and progenitor cells α6-integrin
bri/CD71
dim (1.16 ± 0.08%) . This latter subpopulation expressed high mRNA levels of
SOX2,
NANOG and
OCT4 factors, a high self-renewal activity and a high proportion of holoclones formation in clonogenic assays, all of them characteristics of epithelial stem cells.
Besides, we demonstrated that HPV16-E2 expression modifies the relative abundance of these subpopulations, favoring the enrichment of the early differentiated subpopulation in a comparable way than the differentiation processes produced by the induction with retinoic acid (RA) or calcium chloride (CaCl2) in these cells.
Methods
Cell cultures
HEK293-FT cells from ATCC and HaCaT cells (a generous gift from Dr. Norbert Fusenig) were grown in culture dishes in Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen, CA, USA) supplemented with 10% fetal bovine serum (FBS, Gibco, NY, USA), L-glutamine (2 mM), sodium pyruvate (1 mM), penicillin (50 U/ml), and streptomycin (50 μg/ml). Both cell lines were incubated in a humidified atmosphere with 5% CO2 at 37 °C and maintained in exponential growth phase.
Lentiviral generation
A lentiviral system containing a cassette for puromycin selection and the transgene expression controlled by the promoter for the elongation factor 1-α (EF1-α), was used in this work. The E2 gene from HPV16 was amplified by PCR with the forward (Fw) primer 5′ ATTCCGAATTCATGGAGACTCT 3′ and the reverse (Rev) primer 5′ TTCGGGATCCTCATATAGACAT 3′, using as a template the plasmid pcDNA3-E2. The corresponding amplicon was cloned in the pSin-EF2-Pur plasmid (Addgene, MA, USA) using the EcoRI and BamHI restriction sites, generating the vector pSin-EF2-E216-Pur. A pSin-EF2-Vac-Pur vector was also built, incorporating the EcoRI-BamHI fragment from the pSin-EF2-Pur plasmid. This vector pSin-EF2-Vac-Pur allowed us to generate a lentivirus that does not contain expression cassette, denominated Lenti-Vac. Lentivirus were generated by co-transfection of the corresponding pSin-EF2-X-Pur with pMD2.G and psPAX2 plasmids into packing HEK293-FT cells using Lipofectamine Transfection Reagent (Invitrogen, CA, USA) during 24 h. After 48 h transfection, the supernatant from the cell cultures were ultracentrifugated (25,000 rpm in SW41 Ti rotor) for 2 h at 4 °C, to purify the lentiviral particles. The pellets were suspended in cold phosphate buffer saline (PBS) containing 0.01% bovine serum albumin (BSA) and stored at -70 °C.
Lentiviral transduction
2.5 × 105 HaCaT cells were seeded in DMEM with 10% SFB 24 h before the infection. The cell cultures were then incubated with 1 MOI (multiplicity of infection) of either HPV16-E2 lentivirus or Lenti-Vac for 24 h in DMEM with 10% SFB and polybrene (8 μg/ml), in order to allow virus adsorption. The viral stock was then removed away and 48 h post-infection the puromycin (Sigma-Aldrich, MO, USA) selection (0.45 μg/ml) was started.
RNA extraction and gene expression analysis
Total RNA was extracted from cells using the TRIzol method, treated with RQ1 DNase (Promega, WI, USA) for 2 h at 37 oC and 2 μg of RNA were reverse transcribed into cDNA using the enzyme M-MLV RT at 42 °C and Oligo-dT15 (Promega, WI, USA).
To determinate the transduction and the transgene expression, we amplify by PCR a 250 bp fragment of the HPV16-E2 gene, using primers Fw: 5′ TTGGGGATCCGTGTTTAGCAGCAACGAAGTAT 3′ and Rev: 5′ ATCCGAATTCTCAGTTAATCCGTCCTTTGTGTGAGCT 3′. HPV16-E2 expression in transduced cells was monitored daily.
To evaluate the mRNA expression of the stem cells markers we performed Real-Time PCR (qPCR) using the ABsolute qPCR SYBR Green Mix (Thermo Scientific, PA, USA) and an ABI StepOnePlus Real-Time PCR System, using the following primers:
SOX2 Fw: 5′ TCAGGAGTTGTCAAGGCAGAG 3′, Rev: 5′ AGAGGCAAACTGGAATCAGGA 3′;
NANOG Fw: 5′ GCAATGGTGTGACGCAGAAG 3′, Rev: 5′ ATTGGAAGGTTCCCAGTCGG 3′;
OCT4 Fw: 5′ CTTCGCAAGCCCTCATTTCACC 3′, Rev: 5′ GGTCCGAGGATCAACCCAG 3′. As a control for endogenous constitutively expressed gene, we used β-actin Fw: 5′ GCGGGAAATCGTGCGTGACATT 3′, Rev: 5′ GATGGAGTTGAAGGTAGTTTCGTG 3′. Relative quantification values and the cycle thresholds (Cts) for the target amplicon and the endogenous control (β-actin) were determined for each sample (Total, α6-integrin
bri/CD71
dim and Non-α6-integrin
bri/CD71
dim). The value for the target, normalized to the endogenous control for the samples, relative to the value in the total cells was then calculated with the formula
ΔΔCt [
49].
Immunoblotting analysis of proteins
Total cell lysates were prepared from HaCaT wild type (HaCaTwt) and HaCaT-HPV16-E2 using Bolen-modified RIPA buffer (20 mM MOPS-NaOH pH 7.0, 150 mM NaCl, 1% sodium deoxycholate, 1% Nonidet P-40, 1 mM EDTA, 0.1% SDS) and protease inhibitor cocktail (Roche Diagnostics, Basel, Switzerland). Total proteins (50 μg) were separated through SDS 10% polyacrilamide gels (SDS-PAGE), transferred to nitrocellulose membranes, and immunoblotted using either specific anti-HPV16-E2, anti-cytokeratin 14 (Santa Cruz Biotecnology Inc., CA, USA), anti-cytokeratin 10 (Abcam PLC, MA, USA) or anti-β-actin (a generous gift of Dr. Manuel Hernández) monoclonal antibodies. Secondary antibodies peroxidase conjugated AffiniPure goat anti-mouse IgG (H + L) or peroxidase conjugated AffiniPure goat anti-rabbit IgG (H + L) (Jackson ImmunoResearch Laboratories, PA, USA) were used. Immunoreaction was developed using the SuperSignal West Pico Chemiluminiscent Substrate (Thermo Scientific, IL, USA) and the proteins were quantified with Image J software (National Institutes of Health, MD, USA).
Flow - cytometric analysis and cell sorting
HaCaT cell cultures (lentivirus transduced or wt) were washed twice in PBS, trypsinized, and suspended in ice-cold PBS containing 2% FBS/2% BSA and kept for 15 min at 4 °C. Cells were pelleted by centrifugation and suspended at 1 × 106/100 μl in ice-cold PBS containing 1% BSA and then processed for single or double staining with PE-Cy5 mouse anti-human CD71 and PE rat anti-human α6 integrin (BD Biosciences, NJ, USA) during 45 min at 4 °C, using the appropriate negative controls to establish the compensation settings on the FACS. Cells were washed twice with cold PBS, suspended in PBS at 2–3 × 106/ml, and kept at 4 °C until their flow-cytometric analysis. Cell sorting was performed using MoFlo XDP flow cytometer (Beckman Coulter, CA, USA).
Self-renewal assay
α6-integrinbri/CD71dim and Non-α6-integrinbri/CD71dim HaCaTwt cells sorted after α6-integrin-CD71 staining, were seeded in six-wells plates at 4.5 × 103 cells/well and cultured under standard conditions. After 10 days culture, cells were analyzed for cell surface phenotype (α6-integrinbri/CD71dim and Non-α6-integrinbri/CD71dim) and sorted at subsequent passage for another cycle of 10 days to evaluate the self-renewal capacity.
Colony heterogeneity assay
α6-integrinbri/CD71dim and Non-α6-integrinbri/CD71dim HaCaTwt cells sorted after α6-integrin-CD71 staining, were seeded in six-wells plates at 4.5 × 103 cells/well and cultured under standard conditions. After 7 days culture, the morphology of the clones (classified as holoclones and paraclones) was observed and counted under an inverted light microscope evaluating the percentage of each type of clone.
Differentiation analysis
HaCaTwt cells were seeded and cultured under standard conditions during 24 h. The cell cultures were then washed twice with PBS and incubated with complete media containing CaCl2 5 mM or RA 1 μM (Sigma-Aldrich, MO, USA) during additional 5 days. Then cells were harvested and lysates prepared for the analysis of epithelial differentiation markers.
HaCaTwt and HaCaT-HPV16-E2 were treated with CaCl2 or RA as above indicated. After that, the cells were harvested and analyzed by flow cytometry for the α6-integrin-CD71 cell surface phenotype, in order to compare the differentiation status in HaCaTwt and HaCaT-HPV16-E2.
Statistical analysis
Data were analyzed using GraphPad Prism® v6.0 (GraphPad Software). Results represent the mean of at least three independent experiments; bar ± standard deviation (SD). An ANOVA test was used to determine the statistical significance of differences in values between two groups. Statistical significance was defined as p value <0.05.
Discussion
HPV infection in the basal epithelial layer may occur in any cellular type that integrates this layer and then the expression of the early genes, E2 one of the most important in this virus, could have particular effects on each one of the constituent cell types. A limited number of cells in this epithelial layer, possess cell progenitor characteristics and this small subpopulation has been identified and isolated from primary cultures and transformed cell lines by using different experimental approaches such as detection of desmoglein-3, α6- and β1-integrins, CD71, the expression and nuclear presence of p63, or the metabolic activity of ALDH or ABCG transporters in these cells [
44‐
47].
HaCaT cell line that derives from spontaneously immortalized keratinocytes has been widely used to simulate epithelial tissues [
54,
55]. Using the method described by Schluter and cols. [
48], in this work we demonstrated for the first time that cultures of the HaCaT cell line, similarly to the reported in keratinocytes primary cultures, contain cellular subpopulations, based on the identification and simultaneous detection of the membrane receptors α6-integrin and CD71, corresponding to: early differentiated cells (α6-integrin
dim), transitory amplifying cells (α6-integrin
bri/CD71
bri) and progenitor cells (α6-integrin
bri/CD71
dim) (Fig.
1, panel
a).
Self-renewal assays, clonogenicity and expression of the transcription factors
NANOG,
SOX2 and
OCT4 have been currently used to demonstrate the stemness of putative progenitor cell populations, no matter the method used for its isolation, either based on metabolic characteristics or expression level of specific membrane proteins [
45,
56‐
58].
The self-renewal capability observed in cells from the α6-integrin
bri/CD71
dim subpopulation in HaCaT cells reported in this work, was similar to that reported in several works in the identification of both adult normal and cancer stem cells, where the isolated progenitor subpopulation was also able to reestablish the initial phenotype of the total population, increasing also the number of putative “progenitor” cells after each cycle of separation [
46,
59,
60]. However, in a clear contrast with cancer stem cells, putative progenitor cell population in HaCaT was considerably lower and its enrichment after reseeding was not exponential.
The ability of cells from the α6-integrin
bri/CD71
dim subpopulation to form in clonogenic assays mostly holoclones (Fig.
1, panel
c), in a similar way than the reported in keratinocytes primary cultures [
59], suggest also that cells from this subpopulation have properties of progenitor epithelial cells.
The presence of mRNA from transcription factors
SOX2,
NANOG and
OCT4 (Fig.
1, panels
d and
e), whose expression has been reported for progenitor cells in normal keratinocytes [
56,
61], only in the α6-integrin
bri/CD71
dim subpopulation of HaCaT cells, largely explains its self-renewal capability and the formation mainly of holoclones. In agreement with our results, a recent work reported the immunodetection of NANOG and OCT4 proteins in HaCaT cell line growing in monolayer conditions, despite in those particular culture conditions and unlike than the cancer stem cells evaluated in that work, HaCaT cells were not capable to form spheroids [
45]. Taken together these results confirm that α6-integrin
bri/CD71
dim subpopulation that constitutes approximately 1% of the total population in HaCaT cell line, is enriched in progenitor cells.
Although HaCaT is an immortalized cell line, it possesses particular characteristics that make it very attractive to be used as biological system, such as the absence of viral sequences in its genome, and its ability to respond to a variety of differentiation stimuli such as CaCl
2, RA and cell-cell contact [
62‐
64]. Our demonstration that this cell line possesses, in a low but constant proportion, a cellular subpopulation that express stemness markers, with self-renewal capability similar to the observed in normal epithelial stem cells, constitutes an additional property that must be exploited in this cell line. Then, HaCaT cell line would seem to be a very appropriate model to study the effects of viral gene expression on the three main subpopulations that constitute the epithelial basal layer and that can be target for the HPV infection.
We observed that expression of HPV16-E2 in HaCaT cells generated a significant change in the proportion of the three cellular subpopulations, possibly favoring the early differentiation, since the relative abundance of progenitor (α6-integrin
bri/CD71
dim) cells decreased notably, while cell subpopulation committed to differentiation (α6-integrin
bri) increased (Fig.
2, panel
c). These results confirm and complement the observed by Burns and cols., whom reported the expression of early differentiation markers in the total population of HaCaT cells expressing HPV16-E2 [
34]. In the same way, previous reports from our research group in C-33A cells, indicated that the presence of HPV16-E2 modifies the gene expression profile affecting importantly several signaling pathways, such as integrins, WNT/β-catenin, RhoA and Notch [
20], all of them relevant for proliferation and cell differentiation responses.
Although the stem like properties of the HaCaT-HPV16-E2 cells might not change drastically, some aspects of stemness, such as
SOX2,
NANOG and
OCT4 expression are altered. Then, changes in the relative abundance of the subpopulations could be just another effect of the expression of HPV16-E2, considering that our interpretation is based exclusively on the expression of α6-integrin and CD71. Interestingly the effect on the expression of the stemness factors is differential, since while for
SOX2 and
NANOG the expression is higher in these cells, the opposite effect is observed for
OCT4 (Fig.
3), suggesting that characterization of the E2 cells may not be so straightforward.
The modification in the expression level of these stemness factors could be related to the ability of HPV16-E2 protein to physically interact with several transcription factors such as CBP/p300, SP1, TAF1 and BRD4 [
33,
65‐
68], all of them involved in the regulation of crucial genes for proliferation and differentiation pathways, and also in the regulation of
SOX2,
NANOG and
OCT4 expression [
56,
61]. As an example, it has been demonstrated that BRD4 is a very important piece in the regulation of
NANOG gene expression [
68], and the well known capability of HPV16-E2 to stabilize the binding of BRD4 to cellular promoters [
69] could explain the overexpression observed in cells where HPV16-E2 is present. On the other side, it has been reported that SP1 factor binds directly to the
OCT4 promoter favoring its transcription [
70]. In this way, the physical interaction of E2 protein with SP1 [
71,
72] could prevent its union to several gene promoters,
OCT4 among them, explaining the low expression level of this factor observed in our assays.
The changes generated in the expression pattern of the stemness genes in the progenitor cells in HaCaT-HPV16-E2, suggest that the viral protein has a similar effect than the previously described for differentiation inducer agents [
73‐
76]. Conditions previously reported for stimulation of differentiation in HaCaT cells with CaCl
2 or RA [
77,
78], caused a similar effect on the subpopulations profile to that observed in HaCaT cells expressing HPV16-E2 (Fig.
4, panels c, d and f), confirming that this viral protein by itself is capable to induce a differentiation process. Until our knowledge, the effect of these differentiation inducers agents on the cellular subpopulations constituents of this cell line, has not been reported, then the present work is pioneering in this respect.
RA treatment of HaCaT-HPV16-E2 cells generated the apparent acceleration in the differentiation process, since the majority of cells in the total population corresponded to those committed to early differentiation (Fig.
4, panel
g). However, CaCl
2 treatment generated a profile where almost 70% of the cells corresponded to the transitory amplifying subpopulation. This means that in cells expressing HPV16-E2, the stimulus induces initially an accelerated proliferation process, and then cell differentiation (Fig.
4, panel
e). This behavior could be explained by an increase in the activity of the cip/kip family member p21, since it has been demonstrated that this gene is transcriptionally activated by HPV16-E2 protein [
33,
71,
79], and also by the cascade S100C/A11 directly induced by CaCl
2, leading the cells to hyperproliferation [
77]. The same observed behavior could be also related to the expression or the absence of different regulators of the epidermal differentiation, such as the insulin growth factor binding protein 3 (IGFBP-3), which is characteristically expressed in cells from the epithelial basal layer and the repression of this factor by CaCl
2 stimulus, leads cells to proliferation [
80]. Interestingly, the effect of either CaCl
2 or RA treatment on the relative abundance of the progenitor subpopulation in HaCaT-HPV16-E2 cells was radically different than the observed in HaCaTwt cells, inducing an increase and reaching 0.98% for RA and 1.38% for CaCl
2 induction (Fig.
4, panels
d,
e,
f and
g). This comportment generated by E2 expression could be due to both, the direct transcriptional regulation of genes involved in particular pathways, as well as to indirect mechanisms at epigenetic level, in genic regions that regulate cell fate decision to stemness or to differentiation.
Taken together these results suggest that the synergy between the differentiation stimulus that represents the expression of HPV16-E2 accompanied of a second differentiation stimulus such as CaCl2 or RA (HPV16-E2/CaCl2 and HPV16-E2/AR) can promote the exit of some of the “progenitor cells” to the next immediate higher differentiation level, while the remaining percentage of this subpopulation is apparently less receptive to the stimulus, giving to these cells the opportunity to stays in a stemness status. These effects could guarantee the continuation of the replicative cycle that depends of the epithelial differentiation, besides the preservation of a small amount of progenitor cells ensuring the viral persistence.