Aquaporin-3 re-expression induces differentiation in a phospholipase D2-dependent manner in aquaporin-3-knockout mouse keratinocytes

Abstract

Aquaporin-3 (AQP3) is a water and glycerol channel expressed in epidermal keratinocytes. Despite many studies, controversy remains about the role of AQP3 in keratinocyte differentiation. Previously, our laboratory has shown co-localization of AQP3 and phospholipase D2 (PLD2) in caveolin-rich membrane microdomains. We hypothesized that AQP3 transports glycerol and "funnels" this primary alcohol to PLD2 to form a pro-differentiative signal, such that the action of AQP3 to induce differentiation should require PLD2. To test this idea, we re-expressed AQP3 in mouse keratinocytes derived from AQP3-knockout mice. The re-expression of AQP3, which increased [3H]glycerol uptake, also induced mRNA and protein expression of epidermal differentiation markers such as keratin 1, keratin 10, and loricrin, with or without the induction of differentiation by an elevated extracellular calcium concentration. Re-expression of AQP3 had no effect on the expression of the proliferation markers keratin 5 and cyclin D1. Furthermore, a selective inhibitor of PLD2, CAY10594, and a lipase-dead (LD) PLD2 mutant, but not a LD PLD1 mutant, significantly inhibited AQP3 re-expression-induced differentiation marker expression with calcium elevation, suggesting a role for PLD2 in this process. Thus, our results indicate that AQP3 has a pro-differentiative role in epidermal keratinocytes and that PLD2 activity is necessary for this effect.

Figure 1.. Validation of expression and activity of re-expressed AQP3 in AQP3-knockout keratinocytes

Primary cultures of AQP3-knockout keratinocytes were allowed to reach approximately 70-80% confluence and then infected with adenoviruses expressing either wild-type AQP3 or vector alone using a MOI of 25 for 24 hours. Fresh control medium (50μM calcium) or medium containing 125μM calcium was added for another 24 hours. (a) Total cell lysates were analyzed by Western blotting using antibodies against AQP3 and the GAPDH loading control. AQP3 is seen as a non-glycosylated, 28 kDa band and a diffuse band at 35-40 kDa, representing the glycosylated form. The figure shown is representative of three independent experiments. (b) [³H]glycerol uptake by the cells is shown as cpm/μg protein. The data represent the means ± SEM from three independent experiments. *p<0.05 versus vector-infected keratinocytes.

Figure 2.. AQP3 re-expression increases expression of differentiation markers in AQP3-knockout keratinocytes

Keratinocytes from AQP3-knockout neonatal mice were treated as described in Figure 1. Cells were harvested for quantitative RT-PCR (qRT-PCR) analysis of the expression of (a) keratin 1 (K1), (b) keratin 10 (K10) and (c) loricrin (LOR). Expression was analyzed with the ΔΔCt method using an average value of GAPDH and RPLP0 as endogenous controls, and values are shown as the fold over the vector-infected group. (d and e) Cells were harvested for Western analysis using antibodies against (d) keratin 10 (K10) and (e) loricrin (LOR) with GAPDH as the loading control. Quantitation of the values from three independent experiments is shown, with values expressed as the percent (%) maximal response. *p<0.05, **p<0.01 compared to the vector-infected group and the other indicated groups (in panel d); ^† or ^#p<0.05 versus the indicated groups.

Figure 3.. AQP3 re-expression had no effect on the expression of markers of keratinocyte proliferation

Keratinocytes from AQP3-knockout neonatal mice were treated as described in Figure 1. Cells were harvested for qRT-PCR analysis, and the mRNA expression of (a) keratin 5 (K5) and (b) cyclin D1 was analyzed using the ΔΔCt method with the average of GAPDH and RPLP0 expression as the endogenous control. These results are shown as the fold over the vector-infected group and were derived from the analysis of at least three independent experiments; ns: non-significant.

Figure 4.. The PLD2 inhibitor, CAY10594, inhibited the AQP3 re-expression-induced increase in mRNA levels of markers of keratinocyte differentiation

Primary AQP3-knockout mouse keratinocytes were allowed to reach approximately 70-80% confluence and then infected with adenoviruses expressing either wild-type AQP3 or vector alone using an MOI of 25 for 24 hours. Medium containing the PLD2-selective inhibitor, CAY10594 (1 μM) alone or in combination with an elevated calcium concentration (125μM) was added to the keratinocytes for an additional 24 hours. (a) The cells were harvested for qRT-PCR analysis, and mRNA expression of (a) keratin 1 (K1), (b) keratin 10 (K10) and (c) loricrin (LOR) was determined using the ΔΔCt method with GAPDH as the endogenous control. The results are shown as the fold over the vector-infected group and were derived from at least three independent experiments. *** p<0.001 compared to all other groups; ns, non-significant.

Figure 5.. The PLD2 specific inhibitor, CAY10594, inhibited the AQP3 re-expression-induced increase in protein levels of markers of keratinocyte differentiation

Experiments were performed as described in Figure 4 and cells were harvested for Western analysis using antibodies against (a) keratin 10 (K10) and (b) loricrin (LOR). β-actin served as the loading control. Quantitation of the values from three independent experiments is shown, with values expressed as the percent (%) maximal response. **p<0.01, ***p<0.001 versus the vector-infected control and the other indicated groups; ^† or ^#p<0.05, ^##p<0.01 versus the indicated groups.

Figure 6.. The lipase-dead PLD2 mutant, but not lipase-dead PLD1, inhibited the AQP3 re-expression-induced increase in markers of keratinocyte differentiation

(a,b,c) AQP3 knockout primary keratinocytes at approximately 70-80% confluence were infected with adenovirus expressing either wild-type AQP3 or vector using a MOI of 25 in two groups. One group received additional adenovirus (25 MOI) expressing the lipase-dead PLD2 mutant (PLD2 LD) while the other group received adenovirus expressing vector. After 24 hours fresh medium with basal or an elevated calcium concentration (125μM) was added to the keratinocytes for an additional 24 hours. The cells were harvested for qRT-PCR analysis and mRNA expression of (a) keratin 1 (K1), (b) keratin 10 (K10) and (c) loricrin (LOR) was determined using the ΔΔCt method with GAPDH as the endogenous control. These results are shown as the fold over the vector/vector-infected group and were derived from at least three independent experiments. *** p<0.001 versus all other groups unless indicated otherwise; ^###p<0.001 versus vector/vector; ^†p<0.05, ^††p<0.01 as indicated. (d,e,f) AQP3 knockout primary keratinocytes were infected as above except that an MOI of 12.5 was used both for adenovirus expressing AQP3 or vector and the additional adenovirus expressing lipase-dead PLD1 mutant (PLD1 LD) or vector. mRNA expression of (d) keratin 1 (K1), (e) keratin 10 (K10) and (f) loricrin (LOR) was determined as above. ***p<0.001 versus all other groups unless otherwise indicated; **p<0.01 and *p<0.05 versus vector/vector; ^†p<0.05, ^††p<0.01 as indicated. ns: non-significant. Similar results were obtained in another experiment (repeated once for n=2) using an MOI of 25 for both viruses.