Combination of Low Calcium with Y-27632 Rock Inhibitor Increases the Proliferative Capacity, Expansion Potential and Lifespan of Primary Human Keratinocytes while Retaining Their Capacity to Differentiate into Stratified Epidermis in a 3D Skin Model

Human keratinocytes are difficult to isolate and have a limited lifespan. Traditionally, immortalised keratinocyte cell lines are used in vitro due to their ability to bypass senescence and survive indefinitely. However these cells do not fully retain their ability to differentiate in vitro and they are unable to form a normal stratum corneum in organotypic culture. Here we aimed to generate a pool of phenotypically similar keratinocytes from human donors that could be used in monolayer culture, without a fibroblast feeder layer, and in 3D human skin equivalent models. Primary human neonatal epidermal keratinocytes (HEKn) were cultured in low calcium, (0.07mM) media, +/-10μM Y-27632 ROCK inhibitor (HEKn-CaY). mRNA and protein was extracted and expression of differentiation markers Keratin 14 (K14), Keratin 10 (K10) and Involucrin (Inv) assessed by qRT-PCR and Western blotting. The differentiation potential of the HEKn-CaY cultures was assessed by increasing calcium levels and removing the Y-27632 for 72hrs prior to assessment of K14, K10 and Inv. The ability of the HEKn-CaY, to form a stratified epithelium was assessed using a human skin equivalent (HSE) model in the absence of Y-27632. Increased proliferative capacity, expansion potential and lifespan of HEKn was observed with the combination of low calcium and 10μM ROCK inhibitor Y-27632. The removal of Y-27632 and the addition of high calcium to induce differentiation allowed the cells to behave as primary keratinocytes even after extended serial passaging. Prolonged lifespan HEK-CaYs were capable of forming an organised stratified epidermis in 3D HSE cultures, demonstrating their ability to fully stratify and retain their original, primary characteristics. In conclusion, the use of 0.07mM Calcium and 10μM Y-27632 in HEKn monocultures provides the opportunity to culture primary human keratinocytes without a cell feeder layer for extended periods of culture whilst retaining their ability to differentiate and form a stratified epithelium.


Introduction
Human skin cells, especially keratinocytes, are difficult to isolate and have a limited lifespan in which they may be used [1]. This often means that repeated isolations of cells are required from multiple donors which not only can be difficult to source and time consuming, but can give rise to experimental variation due to genetic differences, location of skin, patient age and sex. Keratinocytes rapidly differentiate in culture, leading to both a reduction in cell numbers due to the loss of the proliferative, basal cells and also phenotypic variance of cells in culture, over time [2,3]. These confounding factors greatly affect the reproducibility and accuracy of in vitro keratinocyte experiments. In addition it is well known that keratinocytes in vivo undergo a process of differentiation, forming a multi-layered epidermis and stratum corneum to form a functional skin barrier (Fig 1). In vitro human skin explant models have been developed which allow investigations concerning keratinocyte differentiation and skin barrier formation to be assessed however these models require the effective maintenance of a basal pool of keratinocytes with full differentiation capabilities.
Traditionally, immortalised keratinocyte cell lines have been used in vitro due to their ability to bypass senescence and survive indefinitely in culture. The most commonly used human keratinocyte cell line, HaCaT's, was spontaneously immortalised from adult human keratinocytes [4]. The HaCaT cell line, cannot however, be considered analogous to the primary phenotype of keratinocytes, as it has acquired a p53 UV-linked genetic mutation [5]. In fact, the cell line has been described as representative of epithelial cells in the early stages of carcinogensis [6][7][8] and investigations of the HaCaT cell line during long term culture have revealed some phenotypic variations which can differ greatly from the original phenotype, and also between laboratories and over time [9]. Contrary to original belief, these cells do not fully retain the ability to differentiate in vitro and they are unable to form a normal stratum corneum in organotypic culture [10]. It has often therefore been necessary to return to primary cell culture with all its limitation and difficulties in order to investigate how keratinocytes respond under different conditions.
The rho kinase (ROCK) inhibitor Y-27632 has been shown to inhibit terminal differentiation and increase proliferation in keratinocytes in culture [11]. Primary keratinocytes cocultured with irradiated fibroblast feeder cells, in keratinocyte growth media supplemented with 10μM Y-27632, maintain proliferation for up to 190 population doublings (PD) and as such it has been suggested that they be considered immortal [12,13]. Unlike HaCaT cells, the  karyotype of Y-27632 immortalized cells remains normal and the morphology of the cells continue to resemble that of basal keratinocytes. Where the HaCaT cell line exhibits a p53 UVlinked mutation, Y-27632 treated keratinocytes continue to express p53 and exhibit a normal DNA damage response. Moreover, the removal of Y-27632 from the culture media allows these cells to maintain a primary phenotype, forming a stratified epithelium in a 3D model of keratinocyte differentiation [12]. However these keratinocytes could not be sustained in the absence of a feeder layer under these conditions, with cultures undergoing senescence after around 35 population doublings [12]. Changes in the concentration of calcium have been shown to affect keratinocyte differentiation. High levels of calcium stimulate differentiation of keratinocytes while low calcium has been used to prolong the lifespan of basal primary keratinocytes in culture by preventing loss of basal cells due to keratinocyte differentiation [14]. Keratinocyte growth media has been formulated commercially with less than 0.1mM calcium to promote the maintenance of a basal proliferative keratinocyte population.
The purpose of this study was to generate a pool of phenotypically similar keratinocytes from human donors that could be used in both monolayer culture and 3D human skin equivalent models without a cell feeder layer. Using freshly isolated human keratinocytes we investigated using very low calcium in combination with ROCK inhibition to prolong the lifespan of keratinocytes in culture whilst maintaining the primary keratinocyte phenotype. Assessments of proliferative capacity, expansion potential and lifespan of primary human keratinocytes and the retention of the capacity to differentiate into stratified epidermis in a 3D skin model were performed.

Human Keratinocyte Culture
Primary human epidermal keratinocytes (HEKn) were isolated from donor neonatal foreskin, collected under human research ethics approval from the Calvary Health Care Adelaide Human Research Ethics Committee (11-CHREC-F007), in accordance with the Declaration of Helsinski principles. Written informed consent was obtained from the next of kin, caretakers, or guardians on behalf of the children enrolled in the study. The written consent form was approved by the human research ethics committee which was signed prior to the procedure occurring.
The epidermis was separated from the dermis via overnight incubation with a 10mg/ml Dispase (Sigma, St Louis, Missouri) solution containing 100 IU/ml penicillin, 100μg/ml streptomycin and 0.25μg/ml amphotericin B (Sigma, St Louis, Missouri), followed by TrypLE (Invitrogen, USA) dissociation of the keratinocytes within the epidermal sheet. Cells were then centrifuged at low speed (160 x g for 5 min) and seeded for culture in CELLnTEC (Bern, Switzerland) low calcium, (0.07mM) 'CnT-07' media containing 100 IU/ml penicillin, 100μg/ml streptomycin and 0.25μg/ml amphotericin B (Sigma St Louis, Missouri) at a density of 5 x 10 4 cell/cm 2 in standard tissue culture vessels.
The cells were grown in low calcium media with or without supplemental 10μM Y-27632 Rock Inhibitor (Sigma, St Louis, Missouri) which was added at the time of initial seeding. These cells are abbreviated as HEKn-CaY. Following the first 2 passages, performed when the cells reached 90% confluence, antibiotics and antimycotics were removed from the culture media and population doubling (PD) rates assessed during serial culture. Cell number and viability were calculated using trypan blue exclusion during passage at 90% confluence. Population doubling was calculated using the equation: n = 3.32 (log UCY-log I) + X, where n = the final PD number at the end of a given subculture, UCY = the cell yield at that point, I = the cell number used as inoculum to begin that subculture, and X = the doubling level of the inoculum used to initiate the given subculture [15].

Human Skin Equivalent Model
To investigate the stratification ability of the cultured keratinocytes, HSEs were prepared using 3x10 5 HaCaT, passage 10 HEKn-CaY or passage 3 HEKn and 1x10 5 passage 6 human foreskin fibroblasts on acellular de-epidermised dermis (DED) in Green's media (1.725μM calcium) in the absence of Y-27632 according to protocol described by McNeil et al [16]. 21 days after raising to air-liquid interface, the HSEs were fixed in 10% buffered formalin and embedded in paraffin wax for histological analysis. 4μm histological sections were stained with haematoxylin and eosin (H&E) and differentiation markers were assessed by immunofluorescence following antigen retrieval according to manufacturer's protocols (DAKO Glostrup, Denmark) with trypsin digestion. Following blocking in 3% normal goat serum, primary antibodies against K10, K14 and Inv (1:200, ThermoScientific, Cheshire, UK) were applied, followed by detection with species-specific, Alexa-488 conjugated secondary antibodies (1:200 Molecular Probes, Life Technologies, Waltham, USA). For verification of staining non-specific binding was determined by omitting primary or secondary antibodies. All control sections had negligible immunofluorescence.

Statistical analysis
Statistical differences were determined using the Student's t-test or an ANOVA. For data not following a normal distribution, the Mann-Whitney U-test was performed. A P-value of less than 0.05 was considered significant.

Results
Primary keratinocytes can be maintained in a proliferative basal state in very low calcium and with ROCK inhibition without a cell feeder layer for greater than 40 population doublings Human neonatal epidermal keratinocytes (HEKn) cultured in low, 0.07mM calcium basal media (CnT-07) reached senescence, defined as less than 0.2 PD/day within 4 weeks and after 35PD which is in line with expected culture rates described by the manufacturer (CELLnTEC). However, cell viability was low, with trypan blue exclusion assay showing cell viability dropping from 87.5% at the third passage to around 50% after 10 passages and 0% at passage 13 (Fig 2A). This is indicative of the cells undergoing terminal differentiation and the loss of basal keratinocytes from culture [17]. In contrast, HEKn cultured in low, 0.07mM calcium basal media (CnT-07) supplemented with 10μM Y-27632 ROCK inhibitor (HEKn-CaY) maintained a high (95%) viability and a steady growth rate for greater than 40 population doublings ( Fig  2B). The addition of 10μM Y-27632 increased proliferation rates with a growth rate of 0.5PD per day being maintained in late passages (beyond 10 passages). At passage 13, around 48 PD, cell viability remained high at an average of 87.5% and the growth rate continued at 0.44PD/ day (Fig 2A). Phase contrast microscope images confirm the macroscopic phenotype was similar between the passage 10 HEKn-CaY and that of the primary and passage 3 HEKn usually used in in vitro experiments (Fig 2C). In addition to the inability of passage 10 HEKn cells to reach confluence due to the decline in proliferation and low viability of these cells, the cells exhibited marked changes in morphology including increased size and more rounded shape. As such, these cells were not included in further investigations.
HEKn-CaY basal phenotype is confirmed with expression of basal keratinocyte markers only when cultured with 10μM Y-27632 Rock Inhibitor in 0.07mM Ca 2+ RNA and protein were extracted from monolayer cultures of HaCaT, passage 1 HEKn, passage 3 HEKn, passage 3 HEKn-CaY and passage 10 HEKn-CaY. qRT-PCR and Western blot were performed to assess the expression of differentiation markers K14, K10 and Inv with HaCaT lysate used as controls (Fig 3). K14 is a marker of basal keratinocytes. K10 is an early marker of differentiation and is expressed by keratinocytes in the suprabasal layers of the epidermis. Inv is a terminal differentiation marker and is expressed by cells within the cornified envelope of the cornified layer of the epidermis. No mRNA expression of the differentiation markers K10 (A) or Inv (B) was observed in HEKn or HEKn-CaY cells indicating they neither of these cell populations had undergone differentiation. Expression of K14 mRNA was observed in all cell populations with no significant difference in levels, confirming their basal phenotype (C).

HEKn-CaY cells express early differentiation marker in 2D differentiation model
The potential of the passage 10 HEKn-CaY cells to differentiate was assessed using a 2D monolayer culture and using high calcium exposure to induced differentiation. Expression of differentiation markers was investigated using immunocytochemistry (Fig 4). K14 expression was observed in both the low level 0.07mM calcium treated keratinocytes (D) and in high 1.2mM calcium induced cells indicating the presence of basal cells (G). K10 expression was absent in the low calcium cultures (E) however it was induced in the high calcium treated cultures (H) indicating that the HEKn-CaY cells had differentiated into different types of keratinocytes. Inv was absent in low calcium treated cells (F) but was also observed in the high calcium treated cells suggesting that some of these cell had undergone terminal differentiation (I). DAPI nuclear counterstain confirms the presence of cells (A-C, J-L) even where no K10 or Inv staining was observed in 0.07mM calcium treated keratinocytes (B,C). HEKn-CaY but not HaCaT's are capable of forming fully stratified epidermis in 3D human skin equivalent culture Passage 10 HEKn-CaY cells and HaCaT's were stratified in a 3D human skin equivalent (HSE) culture and gross morphology compared to normal human skin and a 3D skin equivalent created with passage 3 HEKn as controls (Fig 5). H&E sections (A) reveal that HaCaT cells form a disorganised epithelium with no observable cornified layer. The gross morphology of both passage 3 HEKn and passage 10 HEKn-CaY appear more similar to normal epithelium complete  with evidence of full stratification and cornification. Immunofluorescent images of stratification markers K10 (B), K14 (C) and Inv (D) in green demonstrates that HEKn-CaY but not HaCaT's are capable of forming fully stratified epidermis in a manner similar to low passage HEKn cells. HSE prepared with HaCaT's express K10 but not K14 nor Inv suggesting a more suprabasal keratinocyte phenotype that are not capable of stratification to form the cornified epithelium. HEKn and HEKn-CaY cells however both express K14 in the basal cells, K10 in the suprabasal layers and exhibit Inv expression in the cornified layers. Whilst K14 expression appears stronger in the suprabasal cells of the in vitro HSE than in normal epithelium and some difference in the layers of epithelium is apparent between HEKn-CaY and HEKn, the cells do exhibit the ability to stratify fully to the cornified layer and express all three markers of differentiation in a similar manner. It is important to note that the HSE have thicker cornified layers than normal human skin as it is protected from sloughing which removes much of the cornified layer of normal human skin.

Discussion
Primary keratinocytes are known to have a limited lifespan in culture. Typically, keratinocytes spontaneously differentiate and quickly reach senescence, becoming unviable for further use in in vitro culture experiments [1]. As such, it is generally suggested that keratinocytes used in investigations of skin biology and wound healing including 3D HSE models be limited to passage 1 to 3 cells [16]. The requirement of repeated isolations of primary keratinocytes can lead to increased variation in experiments and has often lead researchers to rely upon immortalised cell lines, such as the HaCaTs, however these are now recognised to vary greatly in phenotype from primary keratinocytes [9]. Indeed, as we have seen, HaCaTs are unable to fully stratify and form the cornified layer of an epidermis making them unsuitable for more sophisticated in vitro models of skin biology such as the 3D HSE model.
Early work showed the use of feeder layers to support the culture of keratinocytes and that the addition of exogenous agents, such as EGF could greatly increase the lifespan of keratinocytes in culture [18,19]. There is however a push towards culture of keratinocytes for clinical use without the use of a feeder to reduce the dependence upon animal products and also the recognition that keratinocytes in monolayer are more suited to biological investigations than those grown in the presence of a fibroblast feeder layer [20,21]. The Rho kinase inhibitor Y-27632 inhibits terminal differentiation and increases proliferation in keratinocytes in culture [11,13]. The addition of Y-27632 to normal keratinocyte growth media has been shown to maintain proliferation indefinitely provided the keratinocytes are supported by an irradiated fibroblast feeder layer [12]. Whilst Chapman et al indicated the requirement of a fibroblast feeder layer to maintain keratinocytes beyond PD35, we have shown that with the combination of Y-27632 and very low calcium, it is possible to maintain a healthy, responsive population of keratinocytes in monolayer culture beyond this point. Whilst we saw senescence in keratinocytes cultured in very low calcium media alone by PD35, the viability of those we cultured in very low calcium supplemented with Y-27632 remained high beyond 40PD with a continued steady growth rate. Thus, we demonstrate that Y-27632 supplementation in combination with low calcium increases the proliferative capacity, expansion potential and lifespan of primary human keratinocytes without the need for a feeder layer.
We have demonstrated that HEKn-CaY in monolayer culture maintain a basal phenotype with no expression of differentiation makers at the protein or RNA level. Critically, the removal of the Y-27632 from the culture media allows the cells to behave as early passage primary keratinocytes, even after extended serial culture. Identification of agents which stimulate differentiation of keratinocytes in a 2D, monolayer setting is also possible using the immunohistochemical markers of early keratinocyte differentiation K10 compared with basal marker K14 as we have shown induction of differentiation only occurs with the addition of high calcium to the media, not simply immediately after removal of Y-27632. This may be important in screening agents to accelerate the reformation of skin barrier function post injury or in various dermatological conditions. Moreover, the same pool of HEKn-CaY may also be used to set up 3D models of human skin which can then be used in sophisticated in vitro assays. Using passage 10 Y-27632 treated keratinocytes and passage 3 keratinocytes cultured in normal low calcium media as a control from the same donor source, we have successfully generated 3D HSE's, demonstrating the ability of these prolonged lifespan cells to fully stratify and retain their original, primary characteristics. Therefore, many more experiments may be set up from one donor sample, increasing the reproducibility and accuracy of research. Our studies suggest that a combination of very low calcium and the addition of ROCK inhibitor can prolong the lifespan of keratinocytes in the absence of a feeder layer enabling the rapid expansion of basal keratinocytes which retain a normal primary phenotype whilst being capable of full differentiation and can used in a range of in vitro assays of skin biology including monolayer screening and 3D HSE models.