Regulating the calcium concentration in the cell culture media produces changes in primary keratinocyte morphology and gene expression. Keratinocytes cultured in low calcium (<0.1 mM) readily proliferate and maintain a dispersed, spread morphology (Figure 1A). In nonconfluent cultures (<80%), cells only express cytokeratins 5/14 (Figure 1B and Figure 1S) consistent with a basal keratinocyte phenotype. Cells cultured in high calcium (1.5 mM) migrate over time (24–48 h) to form desmosomal structures (O’Keefe et al. 1987) and tightly packed stratified layers (Figure 1C). The size of the stratified layers grow with time (Figure 2S). Cells in the top layers appear anucleated and express cytokeratins 1/10 (Figure 1D) consistent with a differentiated keratinocyte phenotype. Under highcalcium conditions it is common (Poumay & Pittelkow 1995; Tamiji et al. 2005), however, to observe regions in the culture dish where cells attached to the plate near differentiated regions express K10/1 weakly and/or continue to express cytokeratins 5/14 (Figure 3S). We chose to conduct experiments with cells incubated in high calcium between 48 and 72 h which is the sufficient time to induce differentiation while ensuring the presence of viable cells.

We first examined the cellular response to UVB dose in the absence of QDs. Keratinocytes, cultured in either high or low calcium, were exposed to increasing doses of UVB (0 to 140 mJ/cm²). Cell viability was quantified after 24 h using the MTT assay (n = 4, different skin donors) to test for either increased cell proliferation or cell death. Cell viability for both keratinocyte phenotypes is plotted (Figure 2) normalised to the keratinocytes cultured in low calcium and 0 mJ cm⁻² UVB exposure. Results for cells cultured in low calcium indicate that UVB exposure induces a cytotoxicity onset response at 60 mJ cm⁻² (p < 0.05). Results for cells cultured in high calcium show a statistically significant lower viability (p < 0.05) at all UVB doses. The data also suggests that UVB may stimulate proliferation under high-calcium conditions, however; a statistical analysis indicates that there is no significant increase or decrease in viability at all UVB doses. When comparisons were made between cells cultured in high and low calcium at each UVB dose, a statistically significant difference (p < 0.05, * with bars) was seen at 0 mJ cm⁻², 20 mJ cm⁻² and 40 mJ cm⁻² UVB. These results are consistent with expectations that differentiated cells would exhibit lower mitochondrial activity than basal cells and that the differentiated phenotype is less sensitive to cytotoxic effects of UVB in this dose range. A 40 mJ cm⁻² dose was chosen for subsequent experiments to minimise toxic effects of UVB to the basal phenotype.

Studies of human skin report that UVB exposure has a more profound effect on basal keratinocytes (Black et al. 2008; Jans et al. 2006; Kupper et al. 1987), which our in vitro findings reported above corroborate. To explore whether the cytotoxic effects of QDs depend on keratinocyte phenotype (basal or differentiated) and if UVB preexposure modulates toxicity, we used a similar experimental strategy as described above. The MTT assay was performed on keratinocytes cultured in high and low calcium, with and without UVB preexposure, for a range of QD doses (Figure 3). In low calcium a QD exposure of 50 nM induced a statistically significant level (p < 0.01) of cytotoxicity (Figure 3A). The QD dose effect on viability was confirmed using Trypan blue exclusion assay (data not shown). The QD toxicity magnitude and the QD dose onset results are consistent with the previous reports (Ryman-Rasmussen et al. 2007; Ryman-Rasmussen et al. 2007a; Zhang et al. 2008; Zhang & Monteiro-Riviere 2009). Interestingly, we did not observe an effect of UVB (40 mJ cm⁻²) on QD-induced cytotoxicity. Similar results were observed for cells cultured in high calcium (Figure 3B).

To visualise the interaction of QDs with primary keratinocytes in culture, we imaged the cells 24 h after QD exposure (5 nM) using fluorescence microscopy (Figure 4). Prior to imaging the cultures were gently washed twice with 1X DPBS to remove any precipitated QDs. This treatment yielded an undetectable QD background on the plate for all conditions. For keratinocytes cultured in low calcium and no UVB exposure, there appeared to be QDs clustered in the perinuclear areas observed as punctate spots (Figure 4A). Exposing cells cultured in low calcium to UVB (40 mJ cm⁻²), prior to introducing QDs, appeared to increase the number punctate QD spots present in the cells (Figure 4B). Interestingly, no difference in cytotoxicity resulting from this association was observed (Figure 3A). For keratinocytes cultured in high-calcium conditions, with or without UVB preexposure, there appeared to be a substantially higher amount of cell-associated QDs (Figure 4C and 4D). Studies investigating the stability of QDs found that high-calcium media did not enhance QD precipitation onto the tissue culture plate over a 24-h period (Figure 4S). Therefore, the increased association of QDs with cells cultured in high calcium is a characteristic of the differentiated phenotype. We examined cell exposure to QD doses of 0.5 nM and 50.0 nM (data not shown) and observed similar trends between low- and high-calcium culture conditions. However, at 50 nM we saw distinct evidence of morphological changes due to cellular toxicity (Figure 3). The qualitative observation that UVB slightly increased QD uptake in basal keratinocyte without a cytotoxic consequence challenges whether the punctate QD spots are truly internalised. Confocal microscopy and transmission electron microscopy (TEM) are methods commonly used to confirm QD cellular uptake. Others have shown using these techniques that internalised QDs do appear as punctate spots (Chang et al. 2009; Geys et al. 2009; Kelf et al. 2010; Zhang & Monteiro-Riviere 2009). We have confirmed using TEM that keratinocytes uptake DLHA-QDs into vesicles (Zheng et al. 2012), but because this technique and confocal microscopy are statistically limited in the number of cells that can be analysed, we chose to use flow cytometry to further investigate the effect of keratinocyte differentiation and UVB exposure on QD cell association.

Flow cytometry was used to provide a more quantitative evaluation of the effect of QD dose, calcium concentration in the culture media and UVB preexposure on QD cell association (Figure 5). A standard gating strategy was used to define single cells from debris and multiple cell clusters (Figure 5A). Using the Calcein AM/Sytox Blue stain combination, we separated live cells from dead cells, respectively (Figure 5B). Gating on the live cell population we found a clear increase in mean fluorescence intensity signal in the QD channel with increasing QD dose, as is illustrated in the overlaid histogram distributions (Figure 5C). We found similar trends when comparing both the live and dead populations, but the live cell population consistently yielded a higher differentiation of QD intensity histogram with dose than the dead cell population. We suspect that this is due to the increased variation in cellular membrane permeability within the dead and apoptotic populations.

To explore the effect of calcium concentration and UVB exposure (40 mJ cm⁻²) on QD cell association, we quantified the relative QD fluorescence signal as a function of QD dose (Figure 6). The values reported are from the average of four experiments (n = 4, different skin donors). When the median fluorescent intensity values are plotted on a log scale, several trends become apparent (Figure 6). The most notable is a dramatic increase in fluorescence intensity with increasing QD dose. There is a steady increase of median QD fluorescent intensity with the 0.5 nM, 5.0 nM and 50 nM doses, respectively. Another obvious trend is the impact of calcium concentration on QD cell association. At each QD dose, a statistically significant higher QD fluorescence is observed for cells cultured in low calcium (basal phenotype). The effect of calcium is, however, lessened at 50 nM QD dose. We attribute this to the excessive presence of QDs producing a fluorescence intensity level that approaches the saturation limit of the flow cytometry detector. When considering the impact of UVB, the results are less clear. There is no statistically significant effect of UVB at any dose, although the data at 0.5 nM and 5.0 nM for cells cultured in low calcium suggest UVB exposure may slightly enhance uptake, consistent with qualitative findings reported in Figure 4 and expectations based on the UVB priming of basal keratinocytes to uptake melanosomes. Taken together, these results suggest that QD dose and calcium level have a strong influence on QD/keratinocyte association but that UVB preexposure is not a significant contributor, at least at the dosage and time point used in this study. The latter is consistent with the MTT assay data reported in Figure 2.

We functionalised 620 nm emitting CdSe/ZnS QD (6.55 ± 1.17 nm) purchased in toluene at a concentration of ~10 μM (NN-Labs LLC) with dihydrolipoic acid (DHLA) for use in an aqueous environment as described previously (Mortensen et al. 2009). Briefly, a 250-μL aliquot of organic QDs (1.25 mg) is removed from their solvent by addition of 1.5 mL of a 50/50 methanol/acetone mixture and centrifugation at 14,000 rpm (~23,000 g) for 5 min. The excess solvent is removed using nitrogen gas and the QDs resuspended in 250 μL tetrahydrofuran (THF). Meanwhile, a 10,000× molar excess of pure DHLA (50 μL) is added to 1 mL methanol in a different glass vial and the pH adjusted using tetramethylammonium hydroxide pentahydrate powder (Sigma-Aldrich Inc.) to pH 11. The QDs in THF are added dropwise to the methanol mixture and incubated at 60°–C with stirring for 3 h. After 3 h, the temperature is reduced to room temperature and the reaction mixture is stirred overnight. The bottom methanol layer is then removed, shaken with excess ether and centrifuged at 14,000 rpm for 5 min. The liquid is decanted, the pellet dried using nitrogen gas and the QDs resuspended in 250 μL double distilled water. To remove any excess ligand, the functionalised QDs are dialysed using a 5-kD molecular weight cut-off DispoDialyzer filter (Harvard Apparatus Inc.) and an excess volume of water for 72 h. After dialysing, the concentration is determined by measuring the absorption at the first exciton and using an extinction coefficient from the literature with Lambert–Beer’s law (Yu et al. 2003). The DLHA-coated QDs exhibited a hydrodynamic radius of 19 ± 2 nm and surface charge of −45 ± 5 mV determined by dynamic light scattering and zeta potential measurements (Malvern ZetaSizer ZS).

We obtained fresh human skin from deidentified adult donors following mammoplasty (Strong and Highland Hospitals, University of Rochester, NY), stored it at 4°C and used it within 6 h of surgery. To limit melanin content variability, experiments were limited to skin donors of Type I and Type II pigmentation. Skin samples were approved for usage by the University of Rochester Research Subjects Review Board (RSRB). To harvest keratinocytes, a modified version of the protocol outlined by Pentland and Needleman was used (Pentland & Needleman 1986). First, skin samples were rinsed with sterile 1X phosphate buffered saline (PBS), treated with 0.4 mL fungizone (Invitrogen) in 500 mL sterile 1X PBS for 10 min and rinsed again thoroughly with 1X PBS. Subcutaneous fat was removed and the dermis thinned using a sterile blade and scalpel. The skin samples were then transferred to fresh 60 mm sterile tissue culture plates with gauze and incubated overnight at room temperature in 6 mL of 0.25% Trypsin-EDTA in a sterile cell culture hood with the stratum corneum exposed to the air. The next day the epidermis was carefully separated from the dermis and mixed well with serum free keratinocyte growth media (KGM-SF) (Gibco Inc.) and 1% penicillin/streptomycin (Gibco Inc.) plus 10% foetal bovine serum (FBS). After mixing, the cells were strained using a 100-μm filter and plated in collagen-coated 12-well plates (Purecol 1:5 diluted in 1X PBS). After 24 h, the media was changed to KGM-SF without FBS, and 3 days later differential trypsin (~35–45 sec incubation with 0.25% Trypsin-EDTA) was performed to remove residual melanocytes and fibroblasts while leaving behind the keratinocytes. After 1 week, the media was changed again. When the plates reached ~80% confluence (~2 weeks), the cells were split into two groups, one with low calcium (<0.1 mM calcium chloride KGM-SF only) and one with high calcium (KGM-SF supplemented with 1.5 mM calcium chloride) to induce differentiation.

Experiments were performed 48 h after the addition of calcium to half of the harvested cells. The media was removed from both basal and differentiated cells and replaced with 300 μL 1X PBS. Cells were then exposed to UVB through a Schott WG 295 nm long-pass glass filter (BES Optics) using FS20 sunlamps (Westinghouse) as described previously (Brouxhon et al. 2007). The FS20 sunlamp emits primarily in the UVB spectrum (290–320 nm), with low amounts of UVA. Lamp output was calibrated using an IL1700 light meter (International Light) with an SED 240 probe to detect light output from 255 to 320 nm (wavelengths shorter than 295 nm removed by Schott filter). Doses from 20 to 140 mJ cm⁻² UVB were established using this technique. Sham irradiation was performed by exposing the cells to the sunlamp for equivalent amounts of time with a foil cover to block light. To determine the impact of UVB dose on cellular viability, three wells of each calcium state were exposed to a UVB dose ranging from 0 to 140 mJ cm⁻² and mitochondrial reductase activity measured 24 h later by the MTT assay (Invitrogen). Briefly, 100 μL yellow MTT reagent (1.2 mg/mL) was added to each well and incubated for 4 h. The reactive solution was then removed, replaced with 1 mL isopropanol to solubilise the resultant formazan crystals and the absorbance at 600 nm measured. UVB dose impact on viability using MTT was measured on four separate occasions with cells grown from unique donors and viability was statistically compared to unexposed control using Student’s paired t-test. Parallel experiments were performed using the same UV protocols, but the flow cytometric preparation described below.

To determine the impact of exposure to increasing doses of DHLA QD 620 nm on cell viability, basal and differentiated primary keratinocytes were incubated with QDs for 24 h with and without UVB exposure. Half of the cells were exposed to 40 mJ cm⁻² UVB, a dose that did not demonstrate statistically significant killing of the primary keratinocytes, but that has been shown by previous work to induce biological response (Brouxhon et al. 2007; Chaturvedi et al. 2004). Doses of 0 nM, 0.5 nM, 5.0 nM and 50.0 nM DHLA QD 620 nm were then introduced into the cell media and incubated for 24 h in triplicate (3 wells of a 12-well plate). The effect of QDs on mitochondrial reductase activity was measured as described above using the MTT assay on four separate occasions with cells grown from four different skin donors and viability was statistically compared to unexposed control using Student’s paired t-test. QD interaction with cells was imaged using phase contrast and fluorescence microscopy (Olympus IX70 with QImaging Retiga EXi camera) with a mercury lamp excitation source (360/30 bandpass filter) and narrow emission (620/10 bandpass filter). Images were analysed using ImageJ (NIH). Parallel experiments were performed using the same QD exposure protocols, but the flow cytometric preparation described below.

To prepare the UV-irradiated and QD-exposed cell samples for flow cytometric analysis, the media was collected from each cell sample, pooled together with each of their three repeats and the cells were trypsinised from their plates. After addition of FBS, the cells from each condition were added to their respective media and centrifuged. This procedure was enacted to preserve any cells that may have been released from the plate into the media due to cytotoxicity over the course of the experiment. The cell samples were then washed with 1X PBS, centrifuged and stained using a Calcein AM live stain (Invitrogen, Oregon, USA) and Sytox Blue dead stain (Invitrogen, Oregon, USA). Each cell sample was incubated for 10 min in the dark in a 600-μL mixture of the two stains, both of which were at 1 μM concentration. The samples were then centrifuged and the resuspended in 3% formalin (VWR International) for analysis. The appropriate single-stain controls were prepared for Sytox, Calcein and QD using HaCaT keratinocyte cells to allow fluorescence compensation. The samples were analysed using an 18-color BD LSRII flow cytometer with filters for Calcein AM (488 nm ex. 515/20 nm em.), Sytox Blue (405 nm ex. 450/50 nm em.) and QDs (405 nm ex. 660/40 nm em.), and results processed using Flow Jo (Version 7.6) software. We gated the forward scattering (FSC) and side scattering (SSC) plot to eliminate debris and multicellular events, and the Calcein/Sytox plot to allow separation of live and dead cellular fractions. The QD median intensity values were measured from four different skin donors. The results of each independent measurement were averaged and compared using the Student’s t-test.