Supply Chain Disruptions During COVID-19 Pandemic Uncover Differences in Keratinocyte Culture Media.

Various culture media are used to propagate keratinocytes (KCs) in vitro. The COVID-19 pandemic resulted in supply chain shortages necessitating substitutions to standard laboratory protocols, which resulted in many laboratories having to use culture media different from those they typically use. We screened available media on the KC line N/TERT2G and found that biological responses varied considerably across three culture media: KC serum-free media, KC growth medium 2, and defined media. We observed qualitative and quantitative differences in proliferation; KCs cultured in defined media had significantly lower proliferative capacity. KC differentiation was assessed by western blot for CLDN1, occludin, cytokeratin-10, and loricrin. Elevated expression of differentiation markers was observed in cells cultured in either KC growth medium 2 or defined media compared with those in cells cultured in KC serum-free media. KC barrier function was measured by transepithelial electrical resistance. KCs cultured in KC growth medium 2 and defined media developed significantly higher transepithelial electrical resistance than those cultured in KC serum-free media, and when treated with IL-4 and IL-13 or IL-17A, we observed variable responses. H&E staining on day 5 -post-differentiation showed greater epithelial thickness in KCs cultured in defined media and KC growth medium 2 than in those cultured in KC serum-free media. These findings show that the choice of culture media impacts the biological response of KCs in a manner that persists through differentiation in the same media.

INTRODUCTION:

Within the field of epidermal biology, various culture media are used to propagate keratinocytes (KC). While within a lab the same media is commonly used, supply chain disruptions that occurred throughout the COVID-19 pandemic required our lab and labs around the globe to identify substitutions for many standard protocols. We found the use of alternative KC culture media gave us highly variable results in the immortalized KC line, N/TERT2G

(Dickson et al., 2000, Moran et al., 2021).

Differences in cell behavior based on culture media have been observed by other groups. Broadbent et al. found the type of media used for differentiation and maintenance of airway epithelial cells impacts experimental results, including morphology, epithelial integrity, and response to viral infection (Broadbent et al., 2020). Zorn-Kruppa et al. also commented on KC characteristics altered by the propagation media (Zorn-Kruppa et al., 2016). Key differences to note in these studies include: 1) primary KC were used, 2) comparison was with different propagation media than we studied (e.g., KGM2, DermaLife, and EpiLife; some of these media and supplements were backordered so we were unable to include them in our study), and 3) they used a protocol for KC differentiation (high calcium added to the culture media rather than differentiation media) which is not commonly used.

We sought to assess relevant biological responses (i.e., proliferation, protein expression, barrier function, etc.) of KC cultured in different media. While researchers may be aware that cell behavior may vary with media composition, we hoped to share the significant differences that can be observed among culture media and further highlight that media is not interchangeable.

In this study we assessed the question of whether the choice of culture media affects critical aspects of KC biology. We measured several characteristics of the KC cell line, N/TERT2G, which were propagated in three routinely used culture media: Keratinocyte Serum Free Media (KSFM; Gibco, #17005042), Keratinocyte Growth Medium 2 (KGM2; PromoCell, #C-20211) and Defined (Gibco, #10744019) (Table 1 ). We focused on four KC assays: proliferation, differentiation marker expression, barrier function and susceptibility to viral infection (Figure 1 ).

Table 1

Media Composition

Media Type	Vendor	Catalog Number	Known Supplements Added by Vendor (Concentration)	Supplements Added by Our Lab (Concentration)	Calcium Concentration
Keratinocyte Serum-Free Medium (KSFM)	Gibco	17005042	oBPE (0.2% v/v)oEGF (5 ng/mL)	oPenicillin (10units/mL)oStreptomycin (10μg/mL)oAmphotericin B (500ng/mL)	0.09mM
Keratinocyte Growth Medium 2 (KGM2)	PromoCell	C20111	oBPE (0.4% v/v)oEGF (0.125ng/mL)oInsulin (5μg/mL)oHydrocortisone (0.33μg/mL)oEpinephrine (0.39μg/mL)oTransferrin (10μg/mL)oCaCl2 (0.06mM)	oPenicillin (10units/mL)oStreptomycin (10μg/mL)oAmphotericin B (500ng/mL)	0.06mM
Defined	Gibco	10744019	No information provided on supplements	oPenicillin (10units/mL)oStreptomycin (10μg/mL)oAmphotericin B (500ng/mL)	0.09mM
EpiLife	Cascade Biologics	MEPI500CA	oBPE (0.2% v/v)oEGF (0.2ng/mL)oInsulin (5μg/mL)oHydrocortisone (0.18 μg/mL)oTransferrin (5μg/mL)oCaCl2 (0.06mM)	oPenicillin (10units/mL)oStreptomycin (10μg/mL)oAmphotericin B (500ng/mL)	0.06mM
DermaLife	LifeLine Cell Technology		oBPE (0.4%)oInsulin (5μg/mL)oHydrocortisone (0.08μg/mL)oEpinephrine (0.33μg/mL)oTransferrin (5μg/mL)oCaCl2 (0.05mM)oTGF-alpha (0.5ng/mL)oL-glutamine (6mM)oPenicillin/Streptomycin (100μg/mL)

Figure 1

Schematic of cell culturing and experimental methods. N/TERT2G were cultured in different media including KSFM, KGM2, and Defined. Once 30% confluent, cells were plated at equal numbers in 96-well plates for proliferation assays, 6.5mm transwell inserts in 24-well plates transwells for TEER assays, or 24-well plates for Western and infection assays. In all assays cells reached confluency 48 hours after plating (24 well pates/transwells) and were switched to DMEM media containing calcium (1.8 mM) to initiate differentiation. Schematic created using BioRender.com

RESULTS & DISCUSSION:

When propagating N/TERT2G in KSFM, KGM2, and Defined media, we observed similar cell morphology but differences in how quickly culture flasks reached confluency (Figure 2 a). To quantify changes in proliferation across different media, we utilized the Click-iT™ EdU Proliferation Assay. N/TERT2G were plated in 96-well plates (28,000 cells/well) and after 24 hours, 10 μM EdU was added to each well and incubated for 6 to 24 hours (Figure 1). Prior to adding the EdU, we observed decreased confluency of cells cultured in Defined, whereas confluency was comparable for cells cultured in KSFM and KGM2 (data not shown). This finding was confirmed by the EdU fluorescent readout, which demonstrated a significant decrease in the proliferative capacity of N/TERT2G grown in Defined compared to KSFM (p < 0.05) (Figure 2b). In all media we observed increased proliferation from 6 to 24 hours, however the degree of this was lower in cells cultured in Defined (Figure 2b).

Figure 2

Propagation of KC in KGM2 and Defined media leads to greater barrier integrity and epithelial thickness than KSFM. Cells were cultured in KSFM, KGM2 or Defined media. Cell morphology was similar across all media, scale bar 200 μm (a). Proliferation was measured in undifferentiated cells cultured in KSFM, KGM2 or Defined media using the Click-iT EdU assay at 6 and 24 hours post addition of EdU (n = 3 experiments), Friedman test with Dunn’s post-hoc test (pairwise) (b). Cells were cultured in KSFM, KGM2 or Defined media, grown to confluency, and switched to DMEM media containing calcium (1.8 mM) to initiate differentiation. Transepithelial electrical resistance (TEER) was measured for 8 days post-differentiation (n = 5 experiments), Friedman test with Dunn’s post-hoc test (pairwise) comparing KSFM vs KGM2 (*) or KSFM vs Defined (#) (c). Membranes were removed from transwells on D5, formalin fixed, processed, embedded, sectioned, and stained for hematoxylin and eosin (H&E), scale bar 50 μm (d). Epithelial thickness was measured across 5 representative images per media type, Kruskal-Wallis test with Dunn’s post-hoc test (unpaired) (e). Significance: *, #, p<0.05, ** p<0.01, *** p<0.001. Data presented as median (Q1, Q3).

As KC differentiate, their proliferative capacity decreases and they form tight junctions, which are critical for establishing epithelial barrier. Transepithelial electrical resistance (TEER) is a widely accepted and quantitative measure of tight junction function, where higher TEER values indicate greater barrier function. N/TERT2G (75,000 cells/well) were plated in transwells (6.5 mm insert, 0.4 μm polyester membrane, Costar) and grown to confluency (Figure 1). TEER was measured prior to inducing differentiation (D0) and every day following (D1-D8) with media replacements every two days. Notably, continued growth in KSFM, KGM2 and Defined media (without switching to DMEM) did not result in increased TEER (data not shown). After inducing differentiation with DMEM media, we observed that KC cultured in KGM2 (p < 0.01) or Defined (p < 0.05) rapidly developed significantly higher TEER than cells initially cultured in KSFM (Figure 2c). This indicates that the initial propagation of KC in Defined or KGM2 promotes greater barrier function of differentiated cells. Hematoxylin and eosin staining demonstrated greater stratification and epithelial thickness (p < 0.001) in cells that had been cultured in Defined or KGM2 compared to KSFM media, which also may explain the increased level of TEER observed with Defined and KGM2 media (Figure 2d & e). We could not delineate from the specimens whether the thicker epidermis was a consequence of taller cells or simply more layers of cells.

To further understand the characteristics of N/TERT2G grown in these different media, we analyzed expression of differentiation markers by Western blot analysis (Goleva et al., 2019). Upon reaching confluency, KC were switched to high calcium DMEM media (1.8 mM) to induce differentiation (Bikle et al., 2012). In undifferentiated KC, we observed surprising differences in the expression of the barrier proteins claudin-1 (CLDN1) and occludin (OCLN) and a commonly used marker of differentiation, cytokeratin-10 (CK10). These proteins were below the limit of detection in nearly all the samples cultured in KSFM, whereas expression was detected in samples cultured in either KGM2 or Defined (Figure 3 a). Undifferentiated cells are reminiscent of cells within the stratum basale of the epidermis, where OCLN and CK10 are not typically detected at the protein level; therefore, it is surprising that we could detect these proteins in undifferentiated KC (Fuchs and Green, 1980, Kirschner et al., 2010). This suggests that KC cultured in KSFM media have similar protein expression to in vivo observations, while culturing in KGM2 or Defined media may prime cells for differentiation. Upon differentiation, we observed earlier and/or greater expression of the differentiation markers and barrier proteins CLDN1, OCLN, and loricrin (LOR) in cells that had been cultured in KGM2 or Defined media compared to KC cultured in KSFM (Figure 3b-e). Expression of these proteins while significantly lower in undifferentiated KC (KSFM, p < 0.05), became comparable in differentiated cells. These findings reveal that culturing KC in different media fundamentally alters the cells in ways that impact protein expression in both undifferentiated and differentiated KC, even after switching to the same media.

Figure 3

Markers of KC differentiation are significantly impacted by propagation media. Cells were cultured in KSFM, KGM2 or Defined media, grown to confluency, and switched to DMEM media containing calcium (1.8 mM) to initiate differentiation. Cell lysates were collected from undifferentiated KC 24 hours after reaching confluency (a) and collected from differentiated KC at Day (D) 1, 2 and 3 post differentiation (b-e). Differentiation markers cytokeratin 10 (CK10), loricrin (LOR), and tight junction proteins claudin-1 (CLDN1) and occludin (OCLN) were detected by Western blot analysis. A representative blot for each protein is provided (f). Expression was quantified by densitometry with normalization to β-actin expression. n = 3 experiments, graphs with less than 3 points indicates protein was not detected (ND) in experiments. Solid bars indicate undifferentiated KC in respective media, open bars indicate KC that have been switched from respective media to DMEM. Friedman test with Dunn’s post-hoc test (pairwise) of KSFM vs KGM2 or KSFM vs Defined. Significance: *p<0.05. Data presented as median (Q1, Q3).

The significantly increased barrier function observed in KC propagated in Defined or KGM2 (Figure 2b) may be explained by our Western blot findings in which two important tight junction proteins, CLDN1 and OCLN, were already detected in undifferentiated KC propagated in Defined or KGM2 (Figure 3a). Furthermore, the relative intensity of CLDN1 and OCLN strongly correlated with TEER values of matched samples (Table 2 ).

Table 2

Barrier function correlates with expression of tight junction proteins. Transepithelial electrical resistance (TEER) and relative intensity of tight junction proteins claudin-1 (CLDN1) and occludin (OCLN), as measured by Western blot quantification, were analyzed for correlation using the Pearson correlation test over the course of three days of differentiation. n=2 matched sample experiments.

Media	Tight Junction Protein	Pearson r	95% confidence interval	P value
KSFM	CLDN1	0.921	0.435 – 0.992	0.009
KSFM	OCLN	0.951	0.613 – 0.995	0.004
KGM2	CLDN1	0.948	0.591 – 0.994	0.004
KGM2	OCLN	0.984	0.855 – 0.998	0.0004
Defined	CLDN1	0.636	-0.363 – 0.955	0.175
Defined	OCLN	0.985	0.865 – 0.998	0.0003

Our lab studies atopic dermatitis and frequently employs the in vitro model of KC conditioned with type 2 cytokines IL-4 and IL-13 (IL-4/13) to study epidermal biology; with a focus on barrier development (Figure 4 a). We observed a range of responses to IL-4/13 treatment, with KSFM propagated KC demonstrating an initial drop in TEER (D2-D3), followed by a significant increase in TEER at later timepoints (D5-D8, p < 0.05). Cells cultured in Defined media demonstrated no significant TEER changes after IL-4/13 treatment. Lastly, we observed a modest but significant decrease in TEER upon IL-4/13 treatment in cells that had been cultured in KGM2 (D6, p < 0.05). To test if this difference in response to cytokine stimulation was observed with other cytokines, we treated KC with the type 3 cytokine commonly observed in psoriasis, IL-17A. Similar to IL-4/13 stimulation, we saw a range of responses depending on which media KC had been propagated in (Figure 4b). KC propagated in KSFM demonstrated a transient enhancement in TEER (D2-3) followed by a significant decrease in TEER (D4-5, p < 0.05). KC propagated in KGM2 also resulted in a significant decrease in TEER from D4-D8 (p < 0.01 - 0.05). Similar to treatment with IL-4/13, treatment with IL-17A had no significant impact on TEER in KC propagated in Defined media, although a modest decrease was observed. This highlights that the culture media used may change the KC inflammatory response. It is possible that culturing KC in these different media results in differential expression of cytokine receptors which may contribute to the variable effects of IL-4/13 and IL-17A on KC barrier function.

Figure 4

KC propagation media significantly impact barrier responses to cytokine stimulation. Cells were cultured in KSFM, KGM2 or Defined media, grown to confluency, and switched to DMEM media containing calcium (1.8 mM) to initiate differentiation. Cells were treated with 50 ng/mL of IL-4/13 (n = 3 experiments) (a) or 50 ng/mL of IL-17A (n = 4 experiments) (b) at the time of differentiation and TEER was measured for 8 days post differentiation. Paired t test comparing KSFM vs KSFM + cytokines (&) or KGM2 vs KGM2 + cytokines ($), black points indicate matched, untreated media controls. Significance: $, & p<0.05, $$ p<0.01, $$$ p<0.001. Data presented as median (Q1, Q3).

We have previously shown the state of KC differentiation impacts susceptibility to infection with vaccinia virus (VV) (Moran et al., 2021). Since Western blot and TEER findings suggest differences in the kinetics of differentiation in cells propagated in Defined or KGM2 compared to KSFM media, we evaluated whether susceptibility to viral infection differed. N/TERT2G were grown to confluency and cells were infected with a low MOI of VV one (D1) or two (D2) days post-differentiation. Undifferentiated (Undiff) KC were similarly infected. We observed comparable infection kinetics over the course of differentiation across the three media. Undifferentiated KC being were most resistant to infection, D1 differentiated KC were more susceptible, and D2 differentiated KC became less susceptible to infection. Infection was significantly greater in undifferentiated cells (KGM2) compared to KSFM (p < 0.05), which may be partially due to the more “differentiated” nature of KC cultured in KGM2, which is supported by the Western blot analysis showing increased levels of differentiation markers in KC grown in this media (Figure 3a). There were no significant differences in infection among the D1 or D2 infected cells across media, indicating the overall kinetics of susceptibility is similar among the different media types (Figure 5 ).

Figure 5

Viral infection kinetics are similar across KSFM, KGM2 and Defined culture media. Cells were cultured in KSFM, KGM2 or Defined media, grown to confluency, and switched to DMEM media containing calcium (1.8 mM) to initiate differentiation. Cells were infected with a low MOI (0.0001) of vaccinia virus while undifferentiated (Undiff), at day 1 (D1) post differentiation, or day 2 (D2) post differentiation. Plates were stained with crystal violet 48 hours post-infection to visualize plaque formation. ImageJ was used to calculate the percentage of the monolayer within each well that was cleared by plaques (n = 4-6 experiments). Wilcoxon matched-pairs signed rank test for undifferentiated KSFM vs KGM2. Significance: *p<0.05. Data presented as median (Q1, Q3).

The use of different culture media fundamentally alters KC in a manner that persists even after switching cells to the same media (DMEM) (Table 3 ). These studies demonstrate that culture conditions are able to induce lasting changes in KC that impact numerous biological functions. While understanding which components of each media are responsible for such changes is critically important, the formulations of these media are in part proprietary and therefore we are unable to test individual components. Our results highlight specific differences between commonly used and commercially available media so that labs may be more aware of what results to expect depending on their media of choice. These findings may also support why specific experimental results may not be repeatable across labs if different culture media is being used. This observation should motivate the field to be precise in our experimental methodology reporting, particularly regarding which media and supplements are used and from which vendor.

Table 3

Summary of Key Findings.

Assay	KSFM	KGM2	Defined	Figure #
Proliferation	=	=	↓	2b
TJ Barrier Function (TEER)	↑	↑↑↑	↑↑	2c
Epithelial Thickness	↑	↑↑	↑↑	2d-e
Undifferentiated Condition (CLDN1, OCLN, CK10)	↓	↑↑	↑↑	3a
Differentiated Condition – (D1) (CLDN1)	↓	↑	↑↑	3e-f
Differentiated Condition – (D2, D3)	=	=	=	3e-f
Response to Type 2 Cytokines	↑	=	=	4a
Response to Type 3 Cytokines	↓↓	↓↓	↓	4b
Pattern of Viral Infection	=	=	=	4

MATERIALS & METHODS:

N/TERT2G cells were provided by Ellen H. van den Bogaard and grown to 30% confluency as previously described (Dickson et al., 2000, Moran et al., 2021, Smits et al., 2017). Cells were cultured in one of three media; KSFM (Gibco, #17005042), KGM2 (PromoCell, #C-20211), and Defined (Gibco, #10744019). N/TERT2G were switched to DMEM media supplemented with 4 mM glutamine and 1.8 mM Ca2+ to induce differentiation. Days post differentiation (D) refers to the number of days since exposure to DMEM. Representative images were taken at 140x magnification using the EVOSTM XL Core Imaging System (ThermoFisher Scientific).

: N/TERT2G were plated in 96-well plates at 28,000 cells/well. After 24 hours, 10 μM EdU was added to each well and incubated for 6 or 24 hours. Proliferative capacity was assessed by the Click-iT™ EdU Proliferation Assay for Microplates (Invitrogen C10499) and the protocol was followed as written in the product information sheet.

: N/TERT2G were plated in 24-well plates at 75,000 cells/well. Cell lysates (RIPA buffer containing protease and phosphatase inhibitors with 0.2% SDS) were collected from undifferentiated cells and cells at days 1, 2, and 3 post-differentiation. Samples were run on Invitrogen NuPAGETM 4-12% Bis-Tris gels and transferred to a PVDF membrane (BioRad). Membranes were probed with the antibodies provided in the table below. Antibodies were detected using SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Scientific). Relative protein expression was determined by densitometry calculated using ImageJ software. Samples were normalized to β-actin expression and protein content (Pierce® BCA Protein Assay Kit).

Western Blot Antibodies:

TEER measurements were done as previously published (De Benedetto et al., 2011). Measurements of TEER were taken for up to 8 days following the initiation of differentiation and exposure to type 2 cytokines IL-4 and IL-13 at 50 ng/mL of each (Biolegend #574004, #571104) or type 3 cytokine IL-17A (R&D Systems 317ILB). To measure epithelial thickness, membranes were cut from transwells on day 5 post differentiation and fixed in formalin. URMC Pathology processed, embedded, sectioned and stained the membranes for hematoxylin and eosin (H&E). Representative images were taken using the EVOSTM XL Core Imaging System (ThermoFisher Scientific) at 560x. Epithelial thickness was quantified at six equally distributed locations across five representative images per media type using ImageJ software. The straight-line tool was used to span the thickness of the layers and then measured using the Analyze -> Measure function.

: N/TERT2G were plated at a density of 150,000 cells/well in a 24 well plate. Cells were infected at a low multiplicity of infection with the Western Reserve strain of vaccinia virus (MOI 0.0001) while undifferentiated, one day post differentiation (D1), or two days post-differentiation (D2). Crystal violet was added to the cells 48 hours after infection. ImageJ software was used to calculate the percentage of the monolayer within each well that was cleared by plaques. To do this, each well was selected with the region of interest tool (circle), and the image was duplicated (right click, duplicate). Total area of the circle was determined by using Analyze -> Measure. Next, the outside of the circle was cleared using the Edit -> Clear Outside command then the Threshold function was applied to the image so that the cleared monolayer (plaques) were white: Image -> Adjust -> Threshold -> Apply. Finally, all area considered to be plaques was selected using the Edit -> Selection -> Create Selection and the selection was inverted using the Make Inverse function. The area covered by plaques was measured with the Analyze -> Measure and the area covered in plaques was divided by the total area to get % monolayer cleared.

Statistics were run using GraphPad Prism. The Friedman test with Dunn’s post-hoc test for multiple pairwise comparisons was used to compare proliferation among the media (KSFM vs KGM2, KSFM vs Defined, KGM2 vs Defined) at both 6 hours and 24 hours. The Friedman test with Dunn’s post-hoc test (pairwise) using KSFM as the control group (KSFM vs KGM2, KSFM vs Defined) was run at each day of differentiation (D0-D8) in which TEER was measured. The Kruskal-Wallis test with Dunn’s post-hoc test for multiple comparisons (unpaired) was used to compare epithelial thickness among the three media (KSFM vs KGM2, KSFM vs Defined, KGM2 vs Defined). For each protein and at each day of differentiation (undifferentiated, D1, D2, D3) the Friedman test with Dunn’s post-hoc test (pairwise) using KSFM as the control group (KSFM vs KGM2, KSFM vs Defined) was used to test for differences in protein expression. Correlation between TEER and tight junction protein expression (CLDN1 & OCLN) across three days of differentiation was assessed using the Pearson correlation test. At each day of differentiation (D0-D8), a paired t test was run comparing media alone (KSFM, KGM2 or Defined) to media with cytokines (either IL-4/13 or IL-17) to test if treatment with cytokines significantly changed TEER within each media. The Wilcoxon matched-pairs signed rank test was used to compare infection in undifferentiated KC cultured in KSFM or KGM2. No statistics were run on Defined undifferentiated KC as there was only an n of 2. Kruskal-Wallis tests with Dunn’s post-hoc test (unpaired) comparing KSFM vs KGM2, KSFM vs Defined, and KGM2 vs Defined were run at D1 and D2 of differentiation and showed no significant differences in infection.

DATA AVAILABILITY STATEMENT: No large datasets were generated or analyzed during the current study.

Anti-CLDN1 (519000)	Invitrogen	1:1000
Anti-β-actin (C4) HRP	Santa Cruz	1:5000
Anti-Keratin 10 (Poly 19054)	Biolegend	1:1000
Anti-Loricrin (Poly 19051)	Biolegend	1:1000
Anti-Occludin (OC3F10)	Invitrogen	1:500
Anti-Mouse IgG HRP (NA931V)	Sigma-Aldrich	1:5000
Anti-Rabbit IgG HRP (NA934V)	Sigma-Aldrich	1:5000