An optimization protocol for Swiss 3T3 feeder cell growth-arrest by Mitomycin C dose-to-volume derivation strategy

Abstract

Feeder cell functionality following growth-arrest with the cost-effective Mitomycin C vis-à-vis irradiation is controversial due to several methodological variables reported. Earlier, we demonstrated variability in growth arrested Swiss 3T3 feeder cell life-span following titration of feeder cell densities with Mitomycin C concentrations which led to the derivation of doses per cell. Alternatively, to counter the unexpected feeder regrowth at high exposure cell density, we proposed titration of a fixed density with arithmetically derived volumes of Mitomycin C solution that corresponded to permutations of specific concentrations and doses per cell. We now describe an experimental procedure of inducing differential feeder cell growth-arrest by titrating with such volumes and validating the best feeder batch through target cell growth assessment. A safe cell density of Swiss 3T3 tested for the exclusion of Mitomycin C resistant variants was titrated with a range of volumes of a Mitomycin C solution. The differentially growth-arrested feeder batches generated were tested for short-term and long-term viability and human epidermal keratinocyte growth supporting ability. The feeder cell extinction rate was directly proportional to the volume of Mitomycin C solution within a given concentration per se. The keratinocyte colony forming efficiency and the overall growth in mass cultures were maximal with a median extinction rate produced by an intermediate volume, while the faster and slower extinction rates by high and low volumes, respectively, were suboptimal. The described method could counter the inadequacies of growth-arrest with Mitomycin C.

Electronic supplementary material

The online version of this article (doi:10.1007/s10616-017-0064-9) contains supplementary material, which is available to authorized users.

Introduction

Growth arrested feeder cells serve as a supportive surface for several cells which are otherwise difficult to grow (Puck and Marcus 1995). Since then, a variety of feeder cells were used for adult, embryonic and induced pluripotent stem cells (Lee et al. 2004; Aasen et al. 2008; Barrier et al. 2012; Higuchi et al. 2015). More importantly, growth-arrested murine Swiss 3T3 cells were used to produce graftable cultured epidermal sheets for clinical application in full thickness burns for about 3 decades (Rheinwald and Green 1975; O’Connor et al. 1981; Green 2008). The growth-arrest in feeder cells is cost-effectively achieved by treatment with Mitomycin C (MC), but the comparative effectiveness of MC vis-à-vis irradiation is a controversy (Schrader 1999; Roy et al. 2001; Fleischmann et al. 2009; Llames et al. 2015; Jiang et al. 2015). The issue is more complicated by the large reported inconsistencies in MC concentrations, duration of exposure, feeder cell type and exposure cell density employed (Roy et al. 2001; Ponchio et al. 2000; Gragnani et al. 2003; Nieto et al. 2007; Omoto et al. 2009; Chugh et al. 2016). The feeder-dependent culture systems therefore, remained less defined resulting in experimental variability. Considering the deficiencies, in addition to the xenogeneic concern with murine feeders, the feeder-free stem cell culture methods are encouraged (Amit and Itskovitz-Eldor 2006). However, they are not fully satisfactory and the feeder-dependent systems are still the indispensable gold standard with unequivocal superiority (Green 2008; Atkinson et al. 2013).

Earlier, we showed that the combined factors of exposure cell density of Swiss 3T3 cells and the concentration of MC for inducing growth-arrest determined the derivation of specific MC dose per cell which appeared critical for the ultimate feeder cell functionality (Chugh et al. 2016). Alternatively, to counter the unexpected resumption of feeder proliferation, we suggested fixing the exposure cell density to a safe constant and substituting the demonstrated combined factors with permutations of concentration and dose per cell. In practical terms, such permutations represent a range of volumes of treating MC solutions which are directly and indirectly proportional to the derived dose per cell and concentration per se, respectively. Because, the 3T3 cultures were known to gradually accumulate spontaneous variants with altered characteristics (Rubin and Xu 1989; Matthews 1993), the exposure cell number has to be set to such a constant level that the cell population does not accumulate MC-resistant variants through successive passaging (Chugh et al. 2015a). Accordingly, the titration of a safe density of feeder cells with varied volumes of a MC solution of a less toxic but effective concentration would then enable identification of a specific concentration–dose permutation that produced ideal feeder cell extinction which in turn influenced optimal stem cell proliferation. This approach would also be valuable in addressing the reported subdued stimulation of epidermal keratinocyte proliferation by the MC feeders as compared to irradiated ones in producing the graftable cultured epithelia (Chugh et al. 2015b).

In this study, we describe a strategy of testing a range of doses per cell by combining volumes of Mitomycin C solutions with concentrations in the order of 3–10 μg/ml for inducing differential feeder cell extinctions in Swiss 3T3 cells. We employed human epidermal keratinocytes as target stem cells to identify the best feeder processing protocol.

Materials and methods

Swiss 3T3 cell culture

The SWISS 3T3 fibroblast cells supplied at 115th passage (CCL-92; www.atcc.org) were cultured and expanded for a total of 6 passages before cryopreservation by adopting a two-tiered banking protocol. The frozen working bank vials were quickly thawed and grown in 3T3 medium containing Dulbecco’s modified eagle medium (Gibco-Invitrogen, Carlsbad, CA, USA) with 10% donor calf serum (Hyclone, Logan, UT, USA) and 1.5 g of sodium-bi-carbonate under standard cell culture conditions of 37 °C in humidified atmosphere with 5% CO₂. The experimental cells were established in either T25 or T75 flasks by sub-culturing the working bank cells once by following a previously reported protocol and the 3rd day cultures, tested for the absence of Mitomycin C (MC) resistance and anchorage independent growth were used in the experiment (Chugh et al. 2015a). The cultures were tested for Mycoplasma contamination (Kumar et al. 2008).

Growth arrest with concentration–dose permutations of MC

The experimental cells were titrated with different volumes of MC that corresponded to specific permutations of concentrations per se and doses per cell as shown schematically in Fig. S1. Each volume of treating solution, υ in ml, was calculated by dividing the product of pre-exposure cell number, Σ in millions, and the chosen dose, Δ expressed as pg/cell, with the concentration, C in μg/ml, as per the following formula.

$$\mathit{\upsilon }=\frac{\mathit{\Sigma }\mathrm{\Delta }}{\text{C}}$$

The pre-exposure cell number was the average of cell counts in randomly picked parallel flasks in triplicate. A broader range of doses than the one demonstrated earlier for concentrations of 4 and 5 μg/ml through exposure cell density titrations (Chugh et al. 2016) was chosen for volume titrations to increase the chance of producing an obvious differential in post-exposure cell extinction. Thus, each concentration was sub-divided into doses of 15, 75, 150, and 450 pg/cell whereas 3 and 10 μg/ml that revealed no significant extinction differential in preliminary dose titrations were combined with doses of 10 and 30 pg/cell, respectively, and served as low and high concentration controls for comparison. Each permutation is depicted with the concentration and the dose shown on the left and right side of a hyphen, respectively.

The cells in T25 or T75 flasks were treated with a 2 h-pulse of MC (Sigma-Aldrich, St. Louis, MO, USA; Cat. No. M4287), which was prepared by proportionately diluting MC stock solution of 200 μg/ml of HEPES Buffered Earl’s Salt (HBES) with 3T3 culture medium to yield the desired dose permutation. A median volume of culture medium containing HBES was used as the vehicle-control. At the end of MC exposure, cells were trypsinized using a solution of 0.25% trypsin and 0.03% EDTA in PBS and viable cells were counted by trypan blue exclusion in Neubauer chamber. The cells exposed to MC in T75 flasks were replated in triplicate wells of 24 well-plates at a density of 7000 cells/cm² and incubated until they were trypsinized for viable cell counting at intervals of 3 days until 12 days. The viable cell counts performed immediately after a 2-h exposure in T25 flasks represented the short-term influence of MC. On the other hand, the periodical cell counts of the replated 3T3 cells after MC exposure in T75 flasks represented the long-term outcome on cell extinctions.

Keratinocyte—feeder co-culture

Primary keratinocytes were obtained from Genlantis (Cat No. PH10205A, www.genlantis.com) which were cryo-preserved at the end of first culture of epidermal cells isolated from healthy adult human skin biopsy after growing in a feeder-free and serum-free culture system. The cells were passaged twice to expand sufficiently enough for undertaking the co-culture experiments with the differentially growth arrested feeders. The basic Rheinwald and Green (1975) technique was adopted for the co-culture (Chugh et al. 2015a). In brief, the keratinocyte medium comprised of Dulbecco’s modified eagle medium and Ham’s F-12 at 3:1 ratio, 10% (V/V) Fetal Calf Serum with each ml containing 10 μg ciprofloxacin, 5 μg insulin, additional 110 μg l-glutamine, 1 μg dexamethasone, 24.32 μg adenine, 20 μg l-serine, 0.4 μg hydrocortisone, 10 ng Cholera toxin, 1.346 ng tri-iodo-thyronine and 5 μg transferrin, 10 ng epidermal growth factor (added in the medium on day 2). The keratinocytes were isolated by initially treating with 0.02% EDTA in PBS to remove the feeder cells followed by detachment of keratinocytes using 0.08% trypsin together with 0.01% EDTA and 0.025% glucose in PBS and viable cells were counted in Neubauer chamber. An initial co-culture was conducted to identify the optimal keratinocyte-feeder seeding ratios. Keratinocytes at initial seeding densities of 5000, 7500 and 10,000/cm² were co-cultured with feeder cells of 4-15, 4-150 and 4-450 permutations at a constant density of 7500/cm², which corresponded to keratinocyte-feeder ratios of 1:0.75, 1:1 and 1:1.5, respectively. These were compared with co-culture of 7500 keratinocytes/cm² with 15,000 feeders/cm² cells which, corresponding to a ratio of 1:2, yielded optimal keratinocyte growth. Therefore, this ratio was adopted uniformly in the final experiments employing feeders of doses of 15, 150 and 450 pg/cell combined with concentrations of 4 and 5 μg/ml and 3-10 and 10-30 were included as controls for comparison. Additionally, the short-listed feeders were tested for differential extinctions at the raised seeding density of 15,000/cm².

Feeder performance on epidermal keratinocytes at clonal density

Colony forming efficiency

A set of triplicate wells plated with 15,000 cells/cm² for each feeder group were inoculated with 250 human epidermal keratinocytes per well and incubated in keratinocyte medium at standard culture conditions for ten days with medium change on alternate days. The colonies were fixed in 4% para-formaldehyde prepared in phosphate buffered (pH 7.2) saline for 45 min, stained with 1% Rhodamine B in distilled water for 30 min and color differentiated in tap water. The preparations were air-dried after the areas of keratinocytes retained red color. The discrete keratinocyte colonies with 8 or more cells were counted. Small colonies of highly irregular shape containing broad, flattened and terminally differentiated cells were considered as aborted and the rest constituted the proliferative colonies.

Digital image analysis

The total growth area of keratinocytes in colony forming assay plates was additionally performed by an image analysis method described previously (Chugh et al. 2015b). Briefly, the stained plates were photographed under uniform light and distance and the digital images were subjected to image analysis using Adobe Photoshop version 7. The keratinocytes area in red color was selected using the Magic Wand tool by manually controlling the precision of selection with the Magic Wand palette and Select options in menu bar. The pixels in the selected area were measured using the Histogram command in the Image menu and isolated to create a separate image and the remaining image area of the plate represented the feeder cells. Similarly, the pixels in the areas representing the feeders and the total growth surface of the well were quantified. The keratinocyte growth area was calculated using the following formula:

$$\text{Keratinocytes}\text{growth}\text{area}\left(\text{in}{\text{cm}}^2\right)=\left(\text{R}\times 9.6\right)/\text{T}$$

R is the number of red pixels; T refers to the total number pixels of the total growth surface and 9.6 is the known growth surface area of the well in cm².

Statistics

The immediate influence of 2-h exposure to MC on cell viability was visualized by column graphs. The permutations were compared with 3-10 and 10-30 by Student’s T test and the vehicle control was compared with pre-exposure cell number. Additionally, dose dependent fall in short-term viability across the permutations was tested by regression. The periodical cell extinctions caused by each permutation of a given MC concentration was depicted in line diagrams with viable 3T3 cell number on y-axis against post-treatment time points on x-axis. The cell viability at each time point was separately tested for variance in cell extinction across the dose permutations by Kruskal–Wallis test. Additionally, linear trend lines were constructed using cell viability on y axis against doses on x axis by least squares fit to calculate R² values by regression. Furthermore, the feeder extinctions were analyzed by constructing column graphs depicting viable cell count of each time point as a column and grouping the columns of each permutation as a single cluster. The inter-cluster comparisons were performed by Student’s T test by pairing the average values of corresponding time points in order to view the overall differences in extinctions.

For evaluating the influence of various seeding densities, the line diagrams were constructed and the significance of variance at each time point was tested by Kruskal–Wallis test. The differential epidermal keratinocyte growth by permutations at high seeding density was represented in column graphs and the significance of variance between permutations was tested by Kruskal–Wallis test. The colony forming efficiency and growth area assessment were tested by Student’s T test. The significance was considered if P < 0.05. Experiments were performed in triplicate and repeated at least twice.

Results

Logistics of volumetric titration and dosing

The minimum volume of treating solution to sufficiently submerge the entire cell monolayer was the lower limit, while, the maximum capacity of the chosen culture flask was the upper limit. The permutation of dose and concentration along with exposure cell number determined the total amount of MC in the final treating volume as shown in Table 1. The permutations led to the derivation of volumes ranging from a minimum of 1.47 ml as in 5-15 and 10-30 to a maximum of 55.125 ml as in 4-450 for T25 flask of 65 ml capacity, whereas the working volumes for the T75 flask of 265 ml capacity were spread across a minimum of 5.281 ml to a maximum of 198.028 ml for the same permutations, respectively. The average pre-exposure cell densities per flask in T25 and T75 flasks were 490,000 ± 7775 (n = 6) cells and 1760,247 ± 42,266 (n = 22), respectively. The vehicle control cells were not considered for comparisons for long-term influence as they grew with a normal sigmoid curve and reached confluence in a week’s time.

Table 1.

Quantitative details of Mitomycin C employed for treating Swiss 3T3 feeder cells with a 2-h pulse

Concentration (µg/ml)	Dose (pg/cell)	T75 flaska		T25 flaska
		Total amount (µg)b	Derived volumesc	Total amount (µg)b	Derived volumesc
3	10	17.603	5.868	4.900	1.633
4	15	26.404	6.601	7.350	1.838
4	75	132.019	33.005	36.750	9.188
4	150	264.038	66.009	73.500	18.375
4	450	792.113	198.028	220.500	55.125
5	15	26.404	5.281	7.350	1.470
5	75	132.019	26.404	36.750	7.350
5	150	264.038	52.808	73.500	14.700
5	450	792.113	158.423	220.500	44.100
10	30	52.808	5.281	14.700	1.470

^aThe average exposure cell number in T75 and T25 flasks were 1760,247 ± 42,266 (n = 22) and 490,000 ± 7775 (n = 6), respectively

^bTotal amount = Exposure cell number (millions) × dose per cell (pg/cell)

^cVolume = (Exposure cell number × dose per cell)/concentration

Short-term viability of dose-titrated feeders

There was no influence of the vehicle on cell viability as reflected by the absence of significant difference in cell counts before and after 2-h exposure (Fig. 1). But the exposure to MC resulted in a significant (P < 0.05) reduction of cell viability in all feeder groups as compared to the control. Interestingly, the cell viability in 4-15 and 5-450 were comparable to the lower 3-10 and the higher 10-30, respectively. The regression analysis revealed a significant dose dependent fall in viability among all the permutations of 4 μg/ml (R² = 0.926, P < 0.01) and 5 μg/ml (R² = 0.898, P < 0.02).

Fig. 1.

Swiss 3T3 cell viability after 2-h pulsed exposure to Mitomycin C solution. The concentrations of 4 µg/ml (a) and 5 µg/ml (b), each of which was combined with doses of 15, 75, 150 or 450 pg/cell. Each permutation was compared with concentrations of 3-10 and 10-30 by Student’s T test and indicated as significant at P < 0.05 (*) or insignificant (NS). Dose dependent fall in viability among all the permutations of 4 µg/ml (R² = 0.926, P < 0.01) or 5 µg/ml (R² = 0.898, P < 0.02) was tested by regression. Pre-exposure cell number (ECN) was determined by cell counts from three random flasks before treatment. Control represents viability in flasks sham exposed to only Mitomycin C vehicle solution. Each column depicted average with standard deviation from triplicates

Extinction of dose-titrated feeders at low density

The extinction of feeder cells replated at a density of 7000/cm² revealed significant variance among the tested doses after post-treatment days of 6 (P < 0.02), 9 (P < 0.03) and 12 (P < 0.03) in the 4 μg/ml group (Fig. 2a), whereas in the 5 μg/ml group, the dose dependent variation was significant on day 6 (P < 0.02), and day 12 (P < 0.04) (Fig. 2b). The regression analysis after dose titration with 4μg/ml (Fig. 2c) revealed significant dose dependent intensification of cell extinction on day 3 (R² = 0.985, P < 0.01), day 6 (R² = 0.969, P < 0.01), day 9 (R² = 0.9, P < 0.02) and day 12 (R² = 0.888, P < 0.02). The dose titration with 5 μg/ml (Fig. 2d) produced significant dose-dependent extinction on day 6 (R² = 0.987, P < 0.01), day 9 (R² = 0.914, P < 0.02) and day 12 (R² = 0.899, P < 0.02), while it was insignificant for day 3 (R² = 0.6). The day matched comparisons of feeder extinctions following 4 or 5 μg/ml revealed that the dose of 15 pg/cell was essentially comparable to 3-10. Similarly, the feeder extinction caused by 10-30 was not significantly different in comparison to either 150 or 450 pg/cell (Fig. 2e, f), although the cell death in 4-150 in particular on day 6 was significantly (P < 0.04) slower. Additional inter-dose comparisons showed that 4-75 produced significantly different cell death as compared to 4-15 (P < 0.001) or 4-150 (P < 0.01), but no significance was noted between 5-75 and 5-150. The dose of 75 pg/cell was therefore, omitted in subsequent experiments.

Fig. 2.

Differential cell extinctions of Mitomycin C treated Swiss 3T3 cells at low density of 7000/cm². Feeder cells were pulse-exposed to Mitomycin C concentrations of 4 µg/ml (a, c, e) and 5 µg/ml (b, d, f) by combining with doses of 15, 150 and 450 pg/cell and viable cells counted on days 3, 6, 9 and 12 (a, b), results analysed by Kruskal–Wallis test and the P value less than 0.05 was indicated. The dose dependent variation in cell extinctions after 4 µg/ml (c) and 5 µg/ml (d) was represented by linear trend lines. Regression analysis was undertaken for doses of 4 μg/ml on day 3 (R² = 0.985, P < 0.01), day 6 (R² = 0.969, P < 0.01), day 9 (R² = 0.9, P < 0.02), and day 12 (R² = 0.888, P < 0.02). The regression with 5 μg/ml was similarly calculated on day 3 (R² = 0.6, NS), day 6 (R² = 0.888, P < 0.02), day 9 (R² = 0.914, P < 0.02), and day 12 (R² = 0.899, P < 0.02). The periodic cell extinctions were represented by clustered column diagrams (e, f). Each cluster represented viable cell number from a single permutation on 3, 6, 9 and 12 post-treatment days. The clusters of 3-10 and 10-30 served as controls for comparison by paired ‘T' test and indicated as significant at P < 0.05 (*) or insignificant (NS)

The differential extent of cellularity at comparable time points was also noted microscopically among the doses. The vacuolated cells denoting cellular disintegration were rare in feeders of 4-15, less frequent in 4-150, but were numerous in 4-450 after 6 days post-MC exposure (Fig. 3a–c). A correspondingly differential loss of cellularity was observed after 12 days which was fair, moderate and intense in 4-15, 4-150 and 4-450, respectively (Fig. 3d–f). Conspicuously, the cells gradually assumed broader aspect as their numbers depleted.

Fig. 3.

Differential loss of cellularity. The variation in periodic loss of Mitomycin C exposed 3T3 cells on days 6 and 12 after they were re-plated into 24-well plates following a pulsed exposure to Mitomycin C in T25 flasks. The concentration of 4 µg/ml, formulated in a way to give dose permutations of 15 (a, d), 150 (b, e) or 450 (c, f) pg/cell. Vacuolated cells (arrows) after 6 days were rare in feeders of 4-15 (a), few in 4-150 (b) and numerous in 4-450 (c). Overall cellularity on culture surface after 12 days post-treatment was fair in 4-15 (d), moderate in 4-150 (e) and poor in 4-450 (f). Marker length is 10 µM

The optimal keratinocyte-feeder ratio

The statistical comparisons of the primary screening of seeding density of feeders of 4 μg/ml revealed production of significant (P < 0.05) variance in keratinocyte stimulation by the short-listed feeders of 4-15, 4-150 and 4-450 on all days in case of keratinocyte: feeder ratio of 1: 0.75 (Fig. S2a) but it was limited to days 6, 9, 12 in case of ratios of 1:1 and 1:1.5 (Fig. S2b, c). A maximal keratinocyte proliferation as early as day 9 was observed with a ratio of 1:2 in which the seeding densities of feeders and keratinocytes were 15,000/cm² and 7500/cm², respectively, although no significant variance was observed (Fig. S2d). As this ratio was chosen for further evaluation, the same feeder groups were additionally tested for differential extinctions at the raised seeding density of 15,000/cm².

Extinction of dose-titrated feeders at high density

Unlike the significant variance in the feeder extinction of 7000 cells/cm² observed discretely on certain post-plating days following treatment with MC doses within 4 or 5 μg/ml, the short-listed feeders plated at 15,000/cm² revealed significance only on the 12th day with 5 μg/ml (Fig. 4a, b). On the other hand, the regression analysis of dose titrations revealed an overall comparable response. The dose titrations with 4 μg/ml (Fig. 4c) showed significant regression on day 3 (R² = 0.964, P < 0.02), day 6 (R² = 0.964, P < 0.02), and day 9 (R² = 0.993, P < 0. 01) barring insignificance on day 12 (R² = 0.75) in contrast to significance on all time-points with low density feeders. The regression with 5 μg/ml (Fig. 4d) was significant on day 3 (R² = 0.942, P < 0.05), day 6 (R² = 0.993, P < 0.01), day 9 (R² = 0.964, P < 0.02), and day 12 (R² = 0.964, P < 0.02), contrary to insignificance on day 3 with low density feeders. Similarly, the comparisons of the overall feeder extinctions performed across all the day-matched data unlike the discrete single time point analyses in line graphs revealed a relative persistence of significant dose-dependent differential extinctions as also observed with low density seeded feeders. The minor differences compared with low density feeders were that the feeders of 4-150 and 3-10 exhibited significantly (P < 0.02) slower extinctions than 10-30 and 4-15, respectively, (Fig. 4e), while the outcome with 5 μg/ml revealed significantly (P < 0.05) slower extinction in 5-150 than in 10-30 (Fig. 4f).

Fig. 4.

Differential cell extinctions of Mitomycin C treated Swiss 3T3 cells at high density of 15,000/cm². Feeder cells were pulse-exposed to Mitomycin C concentrations of 4 µg/ml (a, c, e) and 5 µg/ml (b, d, f) by combining with doses of 15, 150 and 450 pg/cell and viable cells counted on days 3, 6, 9, and 12 (a, b), results analysed by Kruskal–Wallis and the P value less than 0.05 was indicated. The dose dependent variation in cell extinctions after 4 µg/ml (c) and 5 µg/ml (d) was represented by linear trend lines. Regression analysis was undertaken for doses of 4 μg/ml on day 3 (R² = 0.964, P < 0.02), day 6 (R² = 0.964, P < 0.02), day 9 (R² = 0.993, P < 0. 01), and day 12 (R² = 0.75, NS). The regression with 5 μg/ml was similarly calculated on day 3 (R² = 0.942, P < 0.05), day 6 (R² = 0.993, P < 0.01), day 9 (R² = 0.964, P < 0.01) and day 12 (R² = 0.964, P < 0.02). The periodic cell extinctions were represented by clustered column diagrams (e, f). Each cluster represented viable cell number from a single permutation on 3, 6, 9 and 12 post-treatment days. The clusters of 3-10 and 10-30 served as controls for comparison by paired ‘T’ test and indicated as significant at P < 0.05 (*) or insignificant (NS)

Effectiveness of dose-titrated feeders on keratinocyte growth

The differentially growth-arrested feeder cells produced by dose titrations revealed significant (P < 0.04) keratinocyte growth stimulation exclusively by 4-150 as compared to 10-30 on day 6 and it was again significant (P < 0.05) in comparison to both 3-10 and 4-450 on day 9 (Fig. 5a), while no such difference was observed with the doses of 5 μg/ml (Fig. 5b). At the same time, it is important to note that the keratinocyte growth produced by either 4-15 exhibiting slower extinction or 4-450 with faster extinction was not statistically different when compared to the control feeder groups of 3-10 and 10-30, respectively. Further comparisons between the feeder groups of 4 and 5 μg/ml revealed significantly (P < 0.05) higher keratinocyte output in 4-150 than in any of the feeders of 5 μg/ml group on day 6 and continued to be higher than 5-15 and 5-450 until day 9.

Fig. 5.

Growth patterns of human epidermal keratinocytes. Column diagram showing the periodical growth of human epidermal keratinocytes grown in presence of Mitomycin C feeders of 15, 150 and 450 pg/cell under concentrations of 4 (a) and 5 (b) µg/ml and compared with those of 3-10 and 10-30 feeders. The statistical comparisons between two independent feeder cell groups for each time point were performed by Kruskal–Wallis test indicated as significant at P < 0.05 (*) or insignificant (NS)

Dose-titrated feeders and keratinocyte clonal growth

The colony forming efficiency of keratinocytes plated over the various dose-titrated feeder cells revealed significantly (P < 0.01) high number of proliferative colonies in 4-150 as compared to other feeders within 4 μg/ml (Fig. 6a), rather, it was highly significant (P < 0.001) in comparison to any feeders of 5 μg/ml dose-group as well. Interestingly, 5-150 also produced significantly (P < 0.05) more colonies than other feeders within the group, except 5-15 (Fig. 6b).

Fig. 6.

Colony forming efficiency of keratinocytes and growth area measurement. Keratinocytes co-cultured with various feeder groups of 4 µg (a, c) and 5 µg (b, d), each of which was combined with doses of 15, 150 or 450 pg/cell. Feeders of 3-10 and 10-30 were included as controls for comparison. The keratinocytes and feeders were seeded at densities of 250 and 144,000 per well, respectively. Colonies were counted after staining with Rhodamine B (a, b). Keratinocyte growth area was digitally assessed after isolating the Rhodamine B stained colonies (c, d). Each untreated image, ‘U’ from triplicate cultures was used to produce independent images representing keratinocyte colonies, ‘K’ and feeders, ‘F’. These images were superimposed to produce a corresponding merged image, ‘M’ (e). The inter-group comparisons were made by Student’s T test and indicated as significant at P < 0.05 (*) or insignificant (NS)

The superior functionality of 4-150 over other feeder cells was further evident by the significantly (P < 0.02) larger total growth area of keratinocytes (Fig. 6c) while such a convincing outcome was not achieved by feeders of 5 μg/ml dose-group (Fig. 6d). The growth area measurement proved to be a sensitive tool in differentiating the influence of variedly growth arrested feeders as demonstrated by the distinct color isolation of keratinocyte colonies from the feeder cell area (Fig. 6e).

Discussion

Exposure cell density versus volume titration

Earlier we proposed regulation of feeder to target cell ratio through employment of permutations of exposure cell density and concentration of MC during a pulsed treatment protocol for growth-arresting the Swiss 3T3 cells (Chugh et al. 2016). The permutations were shown to result in corresponding doses per cell represented by Δ (pg/cell) depending upon the exposure cell number (Σ) as per the following formula:

$$\mathrm{\Delta }=\frac{\text{C}\mathit{\upsilon }}{\mathit{\Sigma }}$$

When the volume (υ in ml) of treating solution of MC is kept constant, the permutation of exposure cell number (Σ in millions) and concentration of MC (C in µg/ml) determine the value of dose per cell. However, considering the growth-arrest failure in that study following a permutation of low MC concentration and a higher confluent exposure cell density that was previously shown by us to accumulate MC-resistant variants (Chugh et al. 2015a), we proposed substitution of titrations of exposure cell densities with the volumes of MC solutions. In order to obtain the analogous doses per cell, it then required derivation of volumes (υ) of treating solutions as per the following modification:

$$\mathit{\upsilon }=\frac{\mathit{\Sigma }\mathrm{\Delta }}{\text{C}}$$

In view of the proposal, we conducted the experiments in which Σ was kept to a safe constant by adopting a safe and validated subculture scheme to exclude the MC-resistant variants (Chugh et al. 2015a) and the concentration was varied to pinpoint the right permutation. The differential cell extinctions produced thereby with volume titrations using a median concentration of 4 µg/ml, albeit showing minor deviations with change in seeding density, not only mimicked what was shown earlier in cell density titrations (Chugh et al. 2016), but also translated into differential keratinocyte stimulation. The volume titration strategy thus identified an optimized keratinocyte stem cell culture system.

Volume regulation of operational dose per cell

Conventionally, in vitro treatment protocols for effectiveness of a given drug prescribe the strength of such agents in terms of its concentration per se in treating solutions (Barlogie and Drewinko 1980; Ponchio et al. 2000; Connor 2000). Logically, the cells exposed to a toxic agent in a culture flask are presumed to share equally the whole amount present in the total volume of treating solution and upon gradual increase of exposure cell number, a range of doses per cell proportional to the raise would eventually become operational (Yerneni and Jayaraman 2003; Chugh et al. 2016). We observed this phenomenon with only the moderately acting concentrations of 4 and 5 μg/ml, while it was not manifested by either the weakly acting 3 μg/ml or the stronger 10 μg/ml which perhaps was due to experimental limitation of usable volumes. Interestingly, the extinction profile following the weaker dose of 15 pg/cell and intense 450 pg/cell was similar to 3 and 10 μg/ml, respectively, demonstrating the impact of volume titrations in regulating the effect of concentration.

From the published reports it is difficult to deduce the accurate working doses in 3T3 or other feeder cells due to lack of details about precise exposure cell density and volume of treating solution (Schrader 1999; Fleischmann et al. 2009; Ponchio et al. 2000; Nieto et al. 2007; Chugh et al. 2015b). The standard working volumes of 10-15 ml for a T75 flask with constant exposure cell density would correspond to doses of 23-34 pg/cell for the concentration of 4 μg/ml and 28-43 pg/cell for 5 μg/ml which are much lower to the optimally performing dose of 150 pg/cell. Considering the exposure densities in previous studies as confluent or sub-confluent, the dose/cell would still remain much lower leaving the higher doses unexplored.

Implications for feeder-dependent culture systems

We have for the first time demonstrated the concept of employing higher doses of MC per cell through volume derivations for eventual titrations without raising the concentration per se and produced significant alterations in the overall post-exposure life span of feeders thereof with maneuvering capability on target cell stimulation. It is important to recognize that the cell density variation strategy provided primary estimates of effective concentration–dose permutations for volume titrations. The control of net life span of volume/dose-titrated feeders is perhaps the chief controlling factor in optimizing such cultures, since an ideal ratio is a well identified key growth regulating factor in vitro (Sun et al. 2008; Zhou et al. 2009; Jubin et al. 2011). Accordingly, the feeders of 4-15 and 4-450 exhibiting slower and faster extinctions, respectively, performed sub-optimally, while the feeders of 4-150 showing a median extinction brought about maximal keratinocyte stimulation.

We earlier recognized that the feeder regrowth by MC-resistant variants is a sporadic and very late occurrence resulting in a largely overlooked contamination of the target cells with proliferative feeders and recommended specific sub-culture procedures to avoid the same (Chugh et al. 2015a). Therefore, identification of the right sub-culture protocol for banking of feeder cells and long-term validation of the growth-arrested feeders for the absence of regrowth are crucial in allowing the usage of a low MC concentration. The lower MC concentrations would also counter the toxic traces of MC which produced genetic modifications in target stem cells with 10 μg/ml (Zhou et al. 2014). Additionally, we believe that the most commonly used cultures of primary mouse embryonic fibroblast (MEF) and human dermal fibroblasts (HDF) with finite proliferative potential may also harbor and distribute variants with distinct MC sensitivities in a sub-culture dependent manner as there were inconsistencies on concentration and duration of exposure to MC (Nieto et al. 2007; Fleischmann et al. 2009; Lu et al. 2012; Jiang et al. 2015). We also observed highly varied life-spans of MC treated HDF generated by different passage protocols (data not shown).

Cost of cultured epidermal autograft production

Earlier, while working with a concentration of 4 μg/ml, we reported the superior functionality of irradiated feeders over the MC feeders while evaluating the choice of feeder cell growth arrest for the production of cultured epidermis (Chugh et al. 2015b). However, the effective working dose in that study must have been close to the underperforming 4-15 of the present study and there is a probability that the superior 4-150 could then functionally equate with irradiation. Since, the irradiation equipment and the associated facilities are reportedly expensive and time consuming (Llames et al. 2015), the described dose optimization approach could serve as an economical alternative for preparing cultured epidermal sheets for burns patients, particularly in the low and middle income countries where 90% of world’s burns incidences occur (Atiyeh et al. 2009).

Implications for in vitro toxicological assays

In routine toxicological evaluation studies, concentration is the term of reference for in vitro studies and the in vitro dose response curve truly represents concentration dependent evaluation, while dose is specific for in vivo studies (Eisenbranda et al. 2002). In our method, both concentration and dose per given cell population were taken into account as discrete variables while making endpoint observation on cell death. Therefore, we propose that if volume variation strategy should be adopted following cell density titrations in an in vitro toxicological and/or pharmacological study design, it could perhaps form the basis of extrapolating a compound’s operational in vivo dose from the most active permutation of concentration and dose studied on a fixed population of cells in vitro. The volume titrations could further help in estimating accurate effective dosing of anti-cancer agents in a pre-clinical efficacy evaluation setting, particularly while testing moderately toxic concentrations in an attempt to simulate containment of unwanted side-effects in vivo.

Conclusion

The described strategy is projected to streamline the epithelial cell culture system to the extent of maximizing the feeder efficacy, countering the reported inconsistent comparisons on the efficacy of MC versus irradiation and ascertain experimental reproducibility. Additional investigations with other type of feeder cells like the MEF and HDF for hESC and iPSC lines while comparing the irradiation method would further signify the application spectrum of volume titrations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

MC

Mitomycin C

ECN

Exposure cell number

HEPES

4-(2-Hydroxy ethyl)-1-piperazineethanesulfonic acid

Compliance with ethical standards

Conflict of interest

The authors have no conflict of interests to declare.