Reduced serum methods for contact-based coculture of human dermal fibroblasts and epidermal keratinocytes

In the proliferative phase of wound healing, the wound bed is progressively filled with granulation tissue [1], followed by re-epithelialization to restore surface integrity [2]. Granulation tissue consists of, among other components, fibroblasts from the dermis, which proliferate and migrate into the wound bed [3]. Fibroblasts secrete signaling factors that recruit adjacent epidermal keratinocytes from the wound edge and epithelial appendages (Figure 1A) [3,4]. Migration of keratinocytes and fibroblasts results in regions of direct contact between the two cell types [5,6], observed as both cells coexisting in the wound bed or as keratinocytes lying on top of fibroblasts, as seen at the wound edges. Cell-to-cell interactions between these two cell types critically impact cellular proliferation, differentiation and migration and thereby affect wound bed structure and functions [6–9].

Contact-based coculture of human dermal fibroblasts (fibroblasts) and epidermal keratinocytes (keratinocytes) has been a challenging issue in the field of wound and skin studies (Table 1) [10–19]. Fibroblasts and keratinocytes require a host of growth factors and supplements to support their in vitro proliferation [20–23], the concentrations and compositions of which vary between the two cell types. Animal serum (fetal bovine serum [FBS] or fetal calf serum) is an undefined mixture of various growth supplements widely used in cell culture media [10,12–14,17,24]; however, given the scientific, technical and ethical limitations associated with its use, there is a global push to reduce the usage of animal serum in cell culture studies. Further, serum poses a special concern in the context of fibroblast and keratinocyte coculture. The presence of serum is required to support the proliferation of fibroblasts, in contrast to primary keratinocytes, which typically grow poorly or fail to attach and grow in the presence of serum and are therefore cultivated in serum-free media [16,25–30]. Several wound and skin studies have overcome this problem by using immortalized human keratinocyte cell lines, such as HaCaT cells [29,31], which are amenable to long-term passage and display better serum tolerance, as a model for keratinocyte functions [30,32,33]. While HaCaT cells are more tolerant to serum than primary human keratinocytes, they have been shown to demonstrate impaired proliferation in coculture with fibroblasts at high concentrations of serum, such as 10% FBS, which is the concentration usually employed for fibroblast growth [12].

Taken together, existing approaches for contact-based coculture of fibroblasts and keratinocytes are not only technically demanding and labor intensive, but are also limited by the differential effects of serum on the two cell types, the need for complex cell feeder platforms, and material-intensive media formulations (Table 1) [24,34–38].

Materials & methods

Cell culture & maintenance

Adult primary human dermal fibroblasts (HDFa) were obtained from PromoCell (Germany) and maintained in all-in-one, ready-to-use Fibroblast Growth Media (FGM) (Cell Applications 116-500) containing 2% FBS. Immortalized adult human keratinocytes (HaCaT) were a gift from Madhur Motwani (Linq Labs, Pune, India) and were maintained in an all-in-one, ready-to-use Keratinocyte Serum-free Growth Medium (KGM) (Cell Applications 131-500A). Cells were maintained at 37°C in a 5% CO₂ humidified incubator. Coculture was done in T25 flasks (Tarsons) and six-well tissue culture-treated plates (Thermo Scientific). Cell counting was done with a hemocytometer. Imaging was performed at magnification 100× with a Magnus INVI microscope equipped with a Magcam DC5 camera. For all experiments in minimal essential medium (MEM), penicillin and streptomycin (Gibco) was added to media in a final concentration of 1×.

Reduced serum media formulations for coculture of HDFa & HaCaT cells

Reduced serum formulations based on all-in-one fibroblast & keratinocyte growth media

Equal parts of FGM (containing 2% serum) and KGM were mixed, resulting in a composite medium with an effective concentration of 1% serum (referred to as FGM+KGM in this study). FGM and KGM are commercially available formulations, serving as an ‘all-in-one’ composition of growth factors, supplements and antibiotics. Notably, FGM contains 2% FBS, whereas KGM is serum-free. Media formulations were freshly prepared prior to each use.

Reduced serum formulations based on MEM with cell-specific growth supplements

MEM containing Earle's salts, 2 mM L-glutamine, 1 mM sodium pyruvate, nonessential amino acids and 1.5 g/l sodium bicarbonate (Himedia, AL047S), was supplemented with Low Serum Growth Supplement for Fibroblasts (Gibco, S00310, referred to as Fibroblast Growth Supplement [FGS] in this study) and Human Keratinocyte Growth Supplement (Gibco, S0015, referred to as Keratinocyte Growth Supplement [KGS] in this study) at concentrations of 1× each; the resultant medium is referred to here as MEM+FGS+KGS. FGS contains fetal bovine serum, hydrocortisone, human EGF, basic FGF and heparin. When FGS is added to the medium at a concentration of 1×, the resultant medium contains 2% FBS (information from manufacturer). KGS contains bovine pituitary extract, recombinant human IGF-1, hydrocortisone, bovine transferrin and human EGF, and is serum-free. Media formulations were freshly prepared prior to each use.

Coculture of HDFa & HaCaT cells with layering & co-seeding approaches

For the layering approach, HDFa cells were counted (∼10⁵ cells) and seeded in tissue culture-treated T25 flasks in FGM or MEM supplemented with 1× FGS (both containing 2% serum), incubated at 37°C in a 5% CO₂ incubator and grown to 90% confluency. When the cells were confluent, the medium was removed and replaced with FGM+KGM (in a 1:1 ratio; 1% serum) or MEM containing both FGS and KGS (MEM+FGS+KGS, 1× each, 2% serum). HaCaT cells were counted (∼10⁵ cells) and seeded on the confluent layer of fibroblasts. Flasks were imaged with a 10× objective (total magnification 100×) every 24 h for 5 days.

For the co-seeding approach, HDFa and HaCaT cells were counted and ∼10⁵ cells of each type were mixed (resulting in a 1:1 ratio) in FGM+KGM (1% serum) or MEM+FGS+KGS (2% serum), and co-seeded into T25 flasks. Co-seeded cells were incubated at 37°C in a 5% CO₂ incubator. The T25 flasks were imaged with a 10× objective (total magnification 100×) every 24 h for 5 days.

Wound scratch assay on cocultured platforms

HDFa and HaCaT were seeded in six-well tissue culture-treated plates under reduced serum conditions using the co-seeding or layering approaches. When the cocultured cells reached 90% confluency, the medium was replaced with fresh reduced serum medium (FGM+KGM or MEM+FGS+KGS), and a scratch was made across each well using a sterile 200-μl pipette tip. Following this, the wells were imaged with a 10× objective (total magnification 100×) at 0, 4, 24 and 48 h (three images were captured per scratch at different regions). The images were analyzed in ImageJ [39] by drawing a freehand outline of the scratch and using ImageJ tools to calculate the area within the outlined region. This was done for three images per well for three replicates for each condition, and an average wound area was calculated. Percentage of wound closure percentage was calculated as

where W₀ = wound area at 0 h and W_x = wound area at ‘x’ h (calculated at 0, 4, 24 and 48 h).

Results & discussion

Reduced serum conditions developed using commercial media components

To coculture HDFa and HaCaT cells, two reduced serum media were formulated using commercially available media components (Figure 1B). These formulations result in effective concentrations of 1% FBS in FGM+KGM and 2% FBS in MEM+FGS+KGS, representing a five- to tenfold reduction in serum concentration as compared with previous studies (that use 5–10% serum) [10–12,14–16,19,40]. In addition to reduced serum concentrations, our formulations also contain combinations of cell-specific growth factors and supplements, recommended for fibroblast and keratinocyte culture. These reduced serum formulations are simple to develop and use widely available media components. Notably, we have leveraged commercial ‘all-in-one’ formulations (of media and growth supplements) to develop reduced serum concentrations and incorporate cell-specific growth supplements. This is in contrast to previous approaches for fibroblast and keratinocyte coculture, which add serum (most often 10%), growth factors and supplements individually to the cell culture media; this approach is significantly more resource, time, and labor intensive [10,11,15,16,19,24]. Although the KGS is serum-free, it does contain bovine pituitary extract (0.2% v/v, an undefined animal product); however, the resultant formulation (MEM+FGS+KGS) contains very low concentrations of this product.

Reduced serum conditions support the coculture of HDFa & HaCaT cells using the layering technique

For the layering approach, HDFa cells were seeded and grown to 90% confluency in tissue culture-treated T25 flasks, following which HaCaT cells were counted and seeded on layers of confluent HDFa. HDFa cells were initially cultured in FGM (containing 2% FBS) or MEM+1×FGS (final concentration 2% FBS), where they were observed to attach and proliferate, and formed confluent layers in 2–3 days under reduced serum conditions (Figure 2A). This is notable, given that previous protocols have employed media containing 10% FBS to grow these HDFa layers [10,11,15,16,19] and HDFa cells show poor proliferation under serum-free conditions [12]. After HDFa layers grew to confluency, HaCaT cells were introduced in FGM+KGM (1% FBS) and MEM+FGS+KGS (2% FBS). Starting from day 1, HaCaT cells were observed to attach onto HDFa layers under both reduced serum conditions. As seen in Figure 2A, HaCaT cells formed distinct clusters on and between sheaths of HDFa cells from day 1 (MEM+FGS+KGS) or day 2 (FGM+KGM), with no observable difference in the coculture morphology between the two medium formulations. Our formulations enable layered coculture of HDFa and HaCaT cells using reduced serum concentrations at all stages of the protocol, and thereby represent an advancement to previous protocols that employ high-serum (10% FBS or fetal calf serum) media and different media compositions (10% FBS and subsequently serum-free) for the two cell types [10,11,15,16,19].

Reduced serum conditions support the coculture of HDFa & HaCaT cells using the co-seeding technique

Because the simultaneous attachment and proliferation of the two cell types is important to explore the range of cellular interactions in the wound bed [4,11,12,17], we proceeded to employ a co-seeding approach in which both cell types were seeded simultaneously. HDFa and HaCaT were seeded in a 1:1 ratio in tissue culture-treated T25 flasks. Co-seeding was done in FGM+KGM (1% serum) and in MEM+FGS+KGS (2% serum). After 24 h, HDFa and HaCaT were seen to attach and proliferate, and showed characteristic cell morphology which was similar across both reduced serum conditions (Figure 2B). Starting from day 2, the HaCaT cells were observed to cluster in colonies which became progressively dense and tightly packed, surrounded by sheaths of HDFa cells. The coculture platform became fully confluent by day 5 (Figure 2B).

In another study [10], human dermal fibroblasts and epidermal keratinocytes were cocultured in a layered fashion. In this approach, fibroblasts were cultured on six-well Falcon plates or cell culture inserts for 48 h, after which keratinocytes were seeded on top of them. For the cultivation of fibroblasts, Roswell Park Memorial Institute 1640 medium containing 10% FBS was used; when keratinocytes were added, the medium was switched to serum-free conditions (KGM with growth supplements). Our study represents a significant advance to this work, as we have successfully managed to coculture fibroblasts and keratinocytes with the layering approach using 1–2% FBS at all stages of the protocol, including in the initial step of fibroblast culture. Further, our reduced serum approaches successfully cocultured the two cell types simultaneously using the co-seeding approach, which cannot be achieved with protocols that require two different media formulations.

A previous study reported attempts at layering and co-seeding approaches for HDFa and HaCaT coculture under serum-free conditions, using Dulbecco's modified Eagle medium with glutamine and antibiotics [12]. To achieve coculture, HaCaT cells were added soon after the introduction of HDFa cells (∼40 min) or across 5 days. However, HDFa cells (seeded at a density of 5 × 10³ cells/cm²) showed significantly impaired proliferation under serum-free conditions in monoculture and coculture, even after 5 days of culture; even when the seeding density was increased, HDFa cells in coculture displayed impaired proliferation and failed to achieve confluency. For example, at the seeding density of 5 × 10³ cells/cm², HDFa cells achieved a maximum confluency of 1.35% (vs 60% confluency under 10% FBS conditions). Even with increased seeding density (80,000 cells/ml), the maximum confluence achieved by HDFa cells was 44.87%, and this density also supported HaCaT proliferation poorly in coculture. Furthermore, HaCaT cells also did not show the typical morphology and proliferation characteristics under these serum-free conditions. When quantified, at a seeding density of 5 × 10³ cells/cm², the maximum confluence achieved by HaCaT cells was 0.13%; while proliferation of HaCaT cells improved at higher densities, the proliferation and confluence of HDFa remained suboptimal.

Our reduced serum approaches used seeding densities of ∼10⁵ cells, in a 1:1 ratio of the two cell types, and demonstrated successful coculture of HDFa and HaCaT cells. In the layering approach, HDFa cells grew to confluence in the reduced serum conditions in 2–3 days, and HaCaT cells added subsequently proceeded to form distinct clusters on and between sheaths of HDFa cells (Figure 2). With the co-seeding approaches, both cell types were seen to attach and proliferate, and the cocultured platform reached confluence in 5 days (Figure 2).

Reduced serum conditions support characteristic HDFa & HaCaT morphologies under coculture conditions

In coculture under reduced serum conditions, HaCaT cells displayed typical polygonal cell clusters (colonies) with mosaic, cobblestone cell morphology and were surrounded by characteristic sheaths of spindle-shaped HDFa cells (Figure 2 & Supplementary Figure 1A). This characteristic coculture arrangement was observed with both layering and co-seeding approaches and is similar to that seen in previous coculture studies with 10% FBS or serum-free conditions with a range of growth supplements [12,17,24]. Serial images from days 1–5 (Figure 2) show both cell types displaying robust attachment and subsequent proliferation (reaching confluency in the co-seeding approach) in coculture under reduced serum conditions. This is seen as progressively enlarging clusters of HaCaT cells and the formation of HDFa sheaths around HaCaT colonies.

It is important to note that the reduced serum conditions in this study include cell-specific growth supplements (either as components of the ‘all-in-one’ specialized cell media or as a mixture of growth supplements added to minimal media) which, along with optimum serum concentrations, were observed to be critical to establish coculture. As seen in Supplementary Figure 1B, when co-seeded in the absence of serum and growth factor supplements, HDFa cells showed poor proliferation even after 5 days, and HaCaT cells failed to attach and proliferate; in the presence of 10% FBS alone (no growth supplements), HaCaT showed minimal proliferation, and HDFa failed to form typical confluent sheaths around HaCaT colonies (day 3 onwards; compare with Figure 2). This underscores the need for both optimum serum concentrations and cell-specific growth supplements for successful coculture of HDFa and HaCaT; both of these factors are included in the reduced serum formulations in this study. Previous serum-free approaches have added a range of growth supplements individually, which is both resource- and time-intensive [34,35].

Co-cultured platforms, under reduced serum conditions, enable the study of cell migration & wound closure using scratch assays

To enable the study of wound bed functions using the scratch assay, HDFa and HaCaT cells were cocultured in tissue culture-treated six-well plates under reduced serum conditions, FGM+KGM (1% serum) and in MEM+FGS+KGS (2% serum). As seen in Supplementary Figure 2A, using both layering and co-seeding approaches, HDFa and HaCaT cells showed characteristic coculture morphology in tissue culture-treated six-well plates, similar to the results in tissue culture-treated T25 flasks (Figure 2). After the scratch was made, the wound area was imaged at 0, 4, 24 and 48 h. From images for each time point, the wound area was measured (using ImageJ) and the percentage wound closure was calculated (Supplementary Figure 2B).

In the layering approach under both reduced serum conditions, HDFa and HaCaT cells started filling the wound gap by 4 h, with near-complete closure of the gap by 24 h (Figure 3A & B). In the co-seeding approach, under both reduced serum conditions, HDFa and HaCaT proliferation and migration started filling the wound gap by 4 h. The gap continued to be filled between 24 and 48 h, with the wound area showing complete closure (100%) with MEM+FGS+KGS and 80% closure for FGM+KGM after 48 h (Figure 3C & D). In the first 24 h, the ∼20% closure of the wound gap is likely due to a combination of proliferation and early migration in the periphery of the scratch (wound). Subsequently, the gap was seen to be filled predominantly by migrating HDFa cells, while HaCaT clusters were seen to proliferate (seen as increasing size of HaCaT clusters) and fill the wound area from the periphery.

Conclusion

The reduced serum approaches in this study can be adopted for coculture of these two cell types for wound and skin studies, tissue bioengineering and cell therapy applications. Until serum alternatives (human plasma, platelet lysate) or synthetic (serum-free) cell culture media are in routine use for fibroblast and keratinocyte culture, and are optimized to support coculture of these cell types, these reduced serum approaches will serve to reduce the scientific, technical and ethical limitations associated with the use of animal serum in wound and skin studies [41]. For example, serum-containing media contain high levels of calcium [41]. Although calcium supports cell attachment, signaling, movement and architecture in cell culture studies, in the context of keratinocytes it also facilitates differentiation and cornification [42]. However, in studies to recapitulate the structure and functions of the wound bed (as opposed to skin), media containing high calcium levels is not physiologically representative, nor is further cornification of HaCaT cells relevant to the experiments. Therefore in employing reduced serum concentrations, our approaches provide an additional advantage of lowered calcium levels and a more physiological cocultured wound bed milieu.

Future perspective

The proposed approaches will also lend well to further modifications and applications. While HaCaT cells serve as a good model platform for keratinocyte studies [30,32,33], a viable next step would be to optimize the coculture of HDFa with primary human epidermal keratinocytes (HEKa). Given the low serum tolerance of HEKa, this could require further reducing the concentration of FBS in the formulations, while retaining that of the growth supplements, to achieve a ‘sweet spot’ between HDFa serum requirements and HEKa serum tolerance. Another important application of the reduced serum approaches is related to coculturing the recapitulated wound bed with microbes (bacteria, fungi) to mimic an infected wound state. Given that serum is known to be microbicidal [43,44], the presence of reduced concentrations of serum will be advantageous toward optimizing host cell–microbe culture. Finally, while we have developed our approaches in standard cell culture flask and six-well systems, our approaches can be easily taken to 3D coculture and organoid systems, with the addition of matrix and other relevant components.