Introduction

All-trans retinoic acid (ATRA) and its derivatives, commonly called retinoids, are important regulators of several biological processes, such as embryogenesis, reproduction, differentiation, proliferation, and apoptosis (Ahuja et al., 2003; Wilson et al., 2003). Retinoids exert potent keratolytic effects in skin and are frequently prescribed in severe forms of psoriasis (Smit et al., 2004). Retinoids inhibit sebocyte proliferation and differentiation and are of therapeutic value for the treatment of acne vulgaris (Gollnick, 2003). Retinoids are also potent stimulators of keratinocyte proliferation (Fisher and Voorhees, 1996) through the release of heparin-binding epidermal growth factor produced by suprabasal keratinocytes (Chapellier et al., 2002). Retinoids also inhibit the expression of key genes involved in keratinocyte differentiation such as caspase 14 or differentiation-specific keratins (van Rossum et al., 2000; Rendl et al., 2002). By regulating keratinocyte proliferation and differentiation, retinoids increase stratum granulosum thickness and are widely used in cosmetics for the treatment of skin aging (Varani et al., 2000; Roeder et al., 2004).

Water channel aquaporin 3 (AQP3) is a member of a subgroup of the aquaporin family, called aquaglyceroporins, which transports both water and small neutral solutes, such as glycerol. AQP3 was first cloned from rat kidney (Sasaki et al., 1994) and subsequently found in red blood cells (Roudier et al., 1998), chondrocytes (Mobasheri et al., 2004), and in epithelial cells from the urinary, digestive, and respiratory systems (Frigeri et al., 1995; Nielsen et al., 1997; Koyama et al., 1999; Matsuzaki et al., 1999). In skin, AQP3 is constitutively expressed by epidermal keratinocytes (Sougrat et al., 2002). AQP3-deficient mice suffer from reduced water and glycerol permeabilities and decreased water holding capacity of the stratum corneum, demonstrating a pivotal role of this channel in the maintenance of skin hydration (Ma et al., 2002). AQP3-deficient mice also show delayed barrier recovery following tape-stripping disruption together with delayed wound healing, suggesting a possible involvement of AQP3 in the regulation of keratinocyte differentiation and proliferation (Hara et al., 2002). In support of this hypothesis, it has been demonstrated that AQP3 is downregulated by differentiating agents such as high calcium concentrations and 1,25-dihydroxyvitamin D3 (Zheng and Bollinger Bollag, 2003).

In this study, we investigated whether ATRA, a potent modulator of keratinocyte proliferation and differentiation, can regulate AQP3 expression, tissue distribution, and biological activity.

Results

AQP3 is the most abundant AQP in the skin

The subtypes of AQPs expressed in normal human epidermal keratinocytes (NHEK) and in human epidermis were identified using reverse transcription-PCR with AQP1-9-specific primers (S1). NHEK expressed mainly AQP3 transcripts (Figure 1a). A faint band corresponding to AQP9 was also detected in NHEK (Figure 1a). In contrast, in human epidermis, a single band corresponding to AQP3 was visualized (Figure 1b).

Figure 1.

AQP3 is the most abundant AQP in the skin. PCR analysis of AQP expression patterns was performed in (a) cultured NHEK and (b) in human epidermis (b). One microgram RNA was reverse transcribed into first-strand cDNA and amplified by PCR using AQP (A1–9)-specific primers. As a control, PCR was performed using -actin (act)-specific primers.

Full figure and legend (51K)

ATRA is increasing AQP3 expression in NHEK through RAR

NHEK were treated with 1 M ATRA. Cells were collected after 0.5, 1, 2, 3, 24, and 48 hours of incubation and the effect of ATRA on gene expression was determined by means of real-time quantitative PCR (rtQPCR). Application of 1 M ATRA significantly increased AQP3 gene expression (+214%) after 2 hours (Figure 2). A significant accumulation of AQP3 transcripts was also noted after 3 hours (+112%) and 24 hours (+148%) of incubation (Figure 2). After 48 hours of incubation, the expression of AQP3 returned to control levels.

Figure 2.

ATRA is time-dependently increasing AQP3 gene expression in cultured NHEK. The time course effect of ATRA on AQP3 expression in NHEK was analyzed by rtQPCR analysis. The cells were seeded in 25-cm² flasks and cultured until they reached 80% confluence. They were then treated with 1 M ATRA for incubation times ranging from 0.5 to 48 hours. At the indicated times, the cells were pelleted and RNA was extracted. One microgram RNA was reverse transcribed into first-strand cDNA and rtQPCR was performed using AQP3-specific primers. The expression levels of AQP3 mRNA were normalized using the actin expression to correct for RNA quantity and integrity. Data are expressed as percentage of the corresponding control for each time point and are presented as meanSEM (n=5). ^*P<0.05; ^**P<0.01 vs corresponding control. CTL, control.

Full figure and legend (12K)

NHEK were exposed to several concentrations of ATRA ranging from 0.1 to 10 M to test the dose dependency for the AQP3 expression. Application of increasing concentrations of ATRA produced a significant increase in the level of AQP3 transcripts (+241% for 5 M and +494% for 10 M, Figure 3a). To identify which retinoic acid receptor (RAR) is involved in the effect of ATRA on NHEK, the cells were exposed to several concentrations of the specific RAR agonist, Am580 and the specific RAR agonist, CD437. Application of CD437 enhanced AQP3 mRNA expression in a dose-dependent manner (+445% for 5 M, Figure 3b). In contrast, Am580 at concentrations ranging from 0.1 to 5 M produced a limited and non significant increase in AQP3 gene expression (Figure 3c).

Figure 3.

The effect of ATRA on AQP3 gene expression mainly involves RAR receptors. The effects of increasing concentrations of (a) ATRA, (b) RAR agonist CD437, and (c) RAR agonist Am580 on AQP3 expression were investigated by rtQPCR. Cells were seeded in 25-cm² flasks and cultured until they reached 80% confluence at passage 1. They were then treated with ATRA, CD437, or Am580 at concentrations ranging from 0.1 to 10 M for 24 hours. At the end of incubation, the cells were pelleted and RNA was extracted. One microgram RNA was reverse transcribed into first-strand cDNA and rtQPCR was performed using AQP3-specific primers. The expression levels of AQP3 mRNA were normalized using the actin expression to correct for RNA quantity and integrity. Data are expressed as percentage of untreated control and are presented as meanSEM (n=3). ^*P<0.05, ^**P<0.01 vs untreated control.

Full figure and legend (26K)

ATRA is increasing AQP3 protein expression in NHEK

To analyze the ATRA effect on the AQP3 protein, cultured NHEK were treated with ATRA at three concentrations (0.1, 1, and 10 M) for 3, 24, and 48 hours. Proteins were extracted from pelleted cells under non-reducing conditions and subjected to analysis by SDS-PAGE followed by Western blot. AQP3 antiserum revealed the presence of two bands exhibiting apparent molecular weights of 28 kDa (Figure 4a) and 40 kDa (not shown). These data are in good agreement with previous studies reporting that antisera directed against the C-terminal end of the protein recognize both unglycosylated (30 kDa) and glycosylated (40 kDa) forms of AQP3 (Ishibashi et al., 1994). Our results also show that 3-hour application of ATRA at concentrations of 1 and 10 M provoked an accumulation of AQP3-immunoreactive material, whereas treatment with ATRA at the lowest concentration did not show any effect (Figure 4a). Application of ATRA for 24 and 48 hours did not increase the amount of 28-kDa immunoreactive material (Figure 4a). These results were confirmed by densitometric analysis of the signals generated after two independent experiments, which revealed that AQP3 immunoreactive protein significantly accumulated after treatment by 1 and 10 M ATRA for 3 hours (+101 and +131% respectively, Figure 4b).

Figure 4.

ATRA is stimulating AQP3 protein accumulation in cultured NHEK. The effects of ATRA on AQP3 protein were analyzed by Western blot. The cells were treated with ATRA (0.1–10 M) for 3, 24, and 48 hours. At the end of incubation, the cells were pelleted and proteins were extracted. The protein extracts were subjected to electrophoresis, transferred to polyvinyldifluoride membranes, and incubated with anti-AQP3 or anti--actin antibodies. (a) The antibody against AQP3 revealed a major band migrating at 30 kDa. (b) The immunostaining levels of the 30-kDa band were compared in each condition by densitometric quantification and were expressed as a ratio of -actin immunoreactive signal. Data are expressed as percentage of untreated control and are presented as meanSEM (n=2). ^*P<0.05, ^**P<0.01 vs untreated control. CTL, control.

Full figure and legend (72K)

ATRA is stimulating glycerol uptake in NHEK

The noted increase in AQP3 gene and protein expression induced by ATRA may also modulate cellular transport processes. It has been shown that AQP3 is a glycerol-permeable channel. We have therefore investigated the effects of ATRA on glycerol uptake. NHEK were incubated in culture medium containing 1 M ATRA for 3, 24, 48, and 72 hours before glycerol transport was measured. When NHEK were treated for 3 hours with ATRA, the [³H]glycerol amount detected in treated cells was not significantly different from that of untreated cells (Figure 5). In contrast when longer incubations of NHEK in ATRA were performed, glycerol uptake was significantly stimulated (+30, +50, and +48%, 24-, 48-, and 72-hour incubations respectively, Figure 5).

Figure 5.

ATRA is enhancing glycerol uptake by cultured NHEK. Cultured NHEK were treated with ATRA (10 M) for 3, 24, 48, and 72 hours. At the indicated times, the medium was replaced by freshly prepared buffer containing 1 Ci [³H]glycerol. After a 5-minute incubation, the radioactivity of cell lysates was counted. Data are presented as meanSEM (n=3). ^*P<0.05; ^**P<0.01, ^***P<0.001 vs corresponding control. CTL, control.

Full figure and legend (12K)

ATRA is increasing AQP3 expression in human skin explants

An oil-in-water carbopol-based emulsion containing 0.05% ATRA was applied topically on skin explants for 3 and 24 hours. The epidermis was collected and further processed for rtQPCR. Application of the placebo emulsion for 3 hours did not modify AQP3 gene expression in the epidermis (Figure 6a). Formulated ATRA induced a limited and not significant increase in mRNA levels at 3 hours (+67%, Figure 6a). In contrast, a marked increase in AQP3 mRNA was observed when skin explants were treated for 24 hours (+95%, Figure 6b).

Figure 6.

ATRA is increasing AQP3 gene expression in human skin explants. The effects of ATRA on AQP3 expression in human epidermis were investigated by rtQPCR. Explants were punched from skin biopsies and cultured in maintenance medium overnight. Five microliters of ATRA-containing emulsion were applied on skin explants for (a) 3 hours and (b) 24 hours. At the indicated times RNA was extracted from epidermises. One microgram RNA was reverse transcribed into first-strand cDNA and rtQPCR was performed using AQP3-specific primers. The expression levels of AQP3 mRNA were normalized using the actin expression to correct for RNA quantity and integrity. Data are expressed as percentage of control (control 3 hours=100%) and are presented as meanSEM (n=4). ^***P<0.001 vs control 3 hours. CTL, control.

Full figure and legend (23K)

The effect of ATRA-containing emulsion on AQP3 protein expression was also investigated in skin explants by immunohistochemistry. As confirmed by Masson Trichrome histological staining, incubation of skin explants for 48 hours in maintenance medium under a 5% CO₂ humidified atmosphere did not affect tissue integrity (Figure 7a–c). Under control conditions, AQP3 antiserum stained exclusively the viable cells of the epidermis (except from the stratum granulosum) and the labeling was typically concentrated on the plasma membrane (Figure 7d, f, and h). It is noteworthy that AQP3 immunoreactivity is more abundant in the basal layers than in the upper layers of the epidermis (Figure 7d, f, and h). No variation in staining intensity was observed during the course of the treatment, indicating that incubation of skin explants for 48 hours did not alter AQP3 immunoreactivity (Figure 7d, f, and h).

Figure 7.

Topical application of ATRA-containing emulsion is inducing an accumulation of AQP3 immunoreactivity in human skin explants. To visualize the tissue integrity, untreated skin sections were stained using Masson Trichrome (a–c). Skin explants were treated with 5 l of placebo (d, f, h) or ATRA-containing emulsion (e, g, i) for 3 hours (d, e), 24 hours (f, g), and 48 hours (h, i). Skin sections were immunostained by rabbit antiserum directed against AQP3 (d–i). Bar=50 m.

Full figure and legend (271K)

Application of 0.05 % ATRA for 3 hours did not provoke any visible variation in staining intensity and did not modify AQP3 localization in the epidermis (Figure 7e). Similarly, when ATRA was applied for 24 hours, no modification in AQP3 immunoreactivity could be observed (Figure 7g). In contrast, when explants were treated by 0.05% ATRA for 48 hours, an increase in staining intensity was visualized within all the epidermal layers. In addition, basal layers of the epidermis exhibited a remarkable intracellular labeling, suggesting that newly synthesized AQP3 may accumulate in the cytoplasm of the basal cells (Figure 7i).

Discussion

AQPs are water channels present in many organs and tissues. AQP3 is the most abundant AQP in human epidermis (Sougrat et al., 2002). In this study, we have investigated the effect of ATRA, a known regulator of keratinocyte proliferation and differentiation, on AQP3 expression and function. Our results show that application of micromolar concentrations of ATRA to NHEK for up to 24 hours induced a strong increase in AQP3 gene expression presumably through activation of RAR receptor. Concomitantly, ATRA produced a rapid accumulation of 28-kDa immunoreactive form of AQP3. When applied topically on human skin, ATRA induced an overexpression of AQP3 both at the gene and protein levels. Using the tritiated glycerol uptake assay, we finally demonstrated that the increase in AQP3 expression was accompanied by a stimulation of glycerol uptake, suggesting that the increase in AQP3 expression was translated in functional enhancement of cellular transport. Whereas ATRA provoked a long-lasting stimulation of AQP3 gene expression for up to 24 hours, we could not detect by Western blot any accumulation of AQP3 immunoreactivity in NHEK after 24 hours of incubation, although a strong increase in 28-kDa immunoreactive staining was observed after 3 hours of treatment. Molecular analysis of AQPs has shown that these proteins are based on tetrameric arrangements of four channels that stick together within the cell membrane and that appear to be resistant to detergent-based purification (Smith and Agre, 1991; Stroud et al., 2003). AQP3 antibodies directed against the C-terminal end of AQP3 monomer can detect preferentially both glycosylated and unglycosylated monomeric forms of the protein (Ishibashi et al., 1994). The fact that we did not observe any effect of ATRA on 28-kDa AQP3 protein after 24 hours of incubation might suggest that newly synthesized AQP3 monomers do not accumulate into the cytoplasm and are rapidly transformed in tetramers that were not fully extracted in our Western blot procedure. In support of this hypothesis, Ecelbarger et al. (1995) have demonstrated by immunoelectron microscopy that AQP3 is mainly located at the membrane compartment with little or no staining in intracellular vesicles, suggesting that this protein rapidly transits from Golgi apparatus to the cell membrane. Alternatively, we could not rule out the possibility that acute post-transcriptional regulation of AQP3 may occur in NHEK and that the increase in AQP3 transcripts was not translated into proteins. On the other hand, our functional assay demonstrated that ATRA significantly enhanced tritiated glycerol uptake after 24 hours of incubation and that the stimulation of glycerol transport was still noted after 48 and 72 hours of incubation. These data strongly suggest that the AQP3-specific glycerol transport was actually stimulated, most likely through increased AQP3 expression.

In skin explants, we have shown that ATRA induced AQP3 mRNA expression 24 hours after topical application and we have demonstrated by immunohistochemistry that AQP3 is visibly overexpressed 48 hours after the application of ATRA-containing gel. Retinoids and other active molecules may take long time to exert an effect in skin, compared with the rapid onset of the same effect in cell cultures. For instance, although heparin-binding epidermal growth factor expression is stimulated by ATRA in NHEK as early as 3 hours after application (personal observation, Yoshimura et al., 2003), it has been shown that in human skin maximal induction of heparin-binding epidermal growth factor was obtained 24 hours following application of 0.1% ATRA (Rittie et al., 2006).

Representing about 90% of the RARs expressed in the epidermis (Fisher and Voorhees, 1996), RAR is mediating most of the effects of ATRA, such as keratinocyte growth in normal skin or antihyperproliferation in non-melanoma skin cancers (Fisher and Voorhees, 1996; Papoutsaki et al., 2004). By using the RAR- and RAR-specific agonists, Am580 and CD437, we have determined that predominantly CD437 could significantly increase AQP3 transcripts, whereas Am580 did not show any effect. This may designate RAR as the main receptor involved in the effects of ATRA on AQP3 expression. However, it has been shown that CD437 can induce apoptosis in human squamous cell carcinoma cells and that this effect is not abolished by the RAR antagonist, indicating that CD437 is able to induce RAR-independent effects (Sun et al., 2000). Further studies are therefore needed to determine whether CD437 is acting directly via RAR.

AQP3-deficient mice show delayed barrier recovery after tape-stripping disruption and delayed wound healing (Hara et al., 2002), suggesting a possible involvement of AQP3 in the regulation of keratinocyte differentiation and proliferation. In fact, AQP3 expression is downregulated by differentiating agents, such as high Ca²⁺ concentrations and 1,25 dihydroxyvitamin D3 (Zheng and Bollinger Bollag, 2003). Here, we demonstrated that AQP3 expression and biological activity were stimulated by ATRA, a known inducer of keratinocyte proliferation. These data suggest that AQP3 is finely regulated during epidermal differentiation. However, the physiological role of this channel in epidermal maturation remains unclear. Previous studies have reported that AQP3 colocates with phospholipase D2 in caveolin-rich membrane microdomains and it has been assumed that AQP3 may provide glycerol to phospholipase D2 to generate phosphatidylglycerol, which in turn might initiate early differentiation (Zheng and Bollinger Bollag, 2003). The finding that retinoids increase AQP3 expression and stimulate glycerol transport further confirms that, beyond its humectant properties, glycerol may actually play a biological role in epidermal maturation.

In conclusion, we have demonstrated that ATRA stimulated AQP3 gene and protein expression in NHEK as well as in skin explants and increased glycerol transport capacity, indicating that stimulation of AQP3 expression was accompanied by an enhancement of biological activity. Overexpression of functional AQP3 may increase skin glycerol content, which in turn may be a key messenger of keratinocyte proliferation and early differentiation processes.

Materials and Methods

Preparation of skin explants and keratinocyte primary cultures

Abdominal skin samples were obtained from normal human adults undergoing plastic surgery. For the ex vivo experiments, 0.93-cm² biopsies were punched in skin samples under sterile conditions and placed in maintenance medium under a 5% CO₂ humidified atmosphere overnight until treatment. Skin biopsies taken from three to five individuals were used in each experiment. For the in vivo experiments, NHEK were dissociated by enzymatic and mechanical dispersion and seeded at a density of 40 000 cells/cm² in KGM-2 (Cambrex Bio Science Paris, Emerainville, France) containing an antifungal solution.

This study was conducted in accordance with the Declaration of Helsinki Principles and was approved by Johnson & Johnson consumer R&D Board. Informed consent was obtained from each patient.

Chemicals and treatments

ATRA, the RAR analogue Am580, and the RAR analogue CD437 were obtained from Sigma (St Louis, MO).

For the in vitro agonist studies stock solutions of ATRA, Am580, and CD437 (10 mM) were prepared in DMSO and subsequent dilutions were made in culture medium, so that the final concentration of DMSO was always 0.1%. Concentrations of ATRA, CD437, and Am580 ranging from 0.1 to 10 M were applied on NHEK for 24 hours. For the study of kinetics, 1 M ATRA was applied on NHEK for 0.5, 1, 2, 3, 24, and 48 hours. At the end of the treatment, the cells were rinsed in phosphate-buffered saline, pelleted, and stored at -80°C until further assayed.

For the ex vivo treatments, skin biopsies were treated by topical application of ATRA formulated at a concentration of 0.05% in an oil-in-water emulsion containing carbopol gel. Five microliters were applied onto skin explants and application was repeated after 24 hours of incubation. A placebo emulsion was prepared as a control and applied onto skin following the same procedure. Skin explants were collected after 3, 24, and 48 hours of incubation. For the rtQPCR analysis of AQP3 gene expression, the epidermis was separated from the dermis by heat. The samples were further homogenized in TriReagent (Sigma, Saint-Quentin Fallavier, France) and kept at -80°C until RNA extraction. For the analysis of protein expression and localization, skin explants were collected, frozen on dry ice, and kept at -80°C until immunohistochemistry.

RNA extraction

The total RNA from NHEK pellets was extracted using the SV total RNA Isolation System (Promega, Charbonnières, France) following the manufacturer's instructions. RNA from skin explants was extracted by the acid guanidium–thiocyanate–phenol–chloroform method using TriReagent. The concentration of total RNA was determined by measuring the optical density at 260 nm.

Reverse transcription and PCR

One microgram of total RNA was converted into first-strand cDNA using the ImProm-II Reverse Transcription system with either Oligo(dT)₁₅ primers (semiquantitative PCR) or random primers (rtQPCR), as suggested by the manufacturer's instructions (Promega). Parallel reactions for each RNA sample were run in the absence of reverse transcriptase to assess any genomic DNA contamination of the RNA.

For the semiquantitative PCR experiment, the reverse transcription product was amplified using specific primers (Proligo, Paris, France) designed from human AQPs 1-9 cDNAs and the -actin cDNA. The amplified PCR products were subjected to electrophoresis on 1.5% agarose gel and visualized on a UV table.

The rtQPCR experiments were carried out in an M 3000 p detection system (Stratagene, Amsterdam, The Netherlands) using a SYBR Green PCR Mastermix (Stratagene). Each sample was analyzed in duplicate along with standard and no-template controls. The reaction contained 30 ng cDNA in 1 Mastermix, including pre-set concentrations of deoxyribonucleotide triphosphates, MgCl₂, and buffers, along with 300 nM forward and reverse primers and the SYBR Green reporter dye. The PCR parameters were 95°C for 10 minutes, 40 cycles at 95°C for 15 seconds, 60°C for 1 minute, and 72°C for 30 seconds. RNA concentrations were determined by comparing cDNA-generated signals in samples with those generated from known amounts of cDNA. RNA levels were corrected with the -actin cDNA signal for variations in amounts of input RNA. Product purity was confirmed using a dissociation standard curve and agarose gel electrophoresis. Primers were designed from the human AQP3 cDNA and -actin cDNA using the Beacon Designer 3.01 software (Premier Biosoft International, Palo Alto, CA): AQP3 forward primer, 5'-ACCTTTGCCATGTGCTTCCT-3' AQP3 reverse primer, 5'-GCGTCTGTGCCAGGGTGTA-3' -actin forward primer, 5'-CTGGCACCCAGCACAATG-3'; -actin reverse primer, 5'-GCCGATCCACACGGAGTACT-3' (Proligo).

Immunohistochemistry

Skin explant sections (7-m thick) were cut in a cryomicrotome (2800 Frigocut; Leica, Heidelberg, Germany) and collected on 3-aminopropyltriethoxysilane-coated glass slides. After fixation in acetone, the sections were incubated for 1 hour in phosphate-buffered saline containing 0.3% BSA and rabbit anti-AQP3 antiserum diluted 1:500 (Chemicon International, Paris, France). Skin sections were rinsed in phosphate-buffered saline and incubated for 1 hour in the presence of biotinylated goat anti-rabbit Igs diluted 1:2500 (Zymed Laboratories, CliniSciences, Montrouge, France). Reaction products were detected by FITC-conjugated streptavidin diluted 1:1250 (Caltag Laboratories, Burlingame, CA). Finally, sections were counterstained with propidium iodide, mounted in Fluoprep (Biomérieux, Marcy l'Etoile, France), covered, and stored in the dark at 4°C. The integrity of the explants was assessed by Masson Trichrome histological staining.

SDS-PAGE and Western blot analysis

Proteins were analyzed by PAGE under non-reducing conditions and electroblotted onto polyvinyldifluoride membranes (Millipore, Molsheim, France). The membranes were incubated for 2 hours with AQP3 antibodies diluted 1:200 (Calbiochem, Nottingham, UK) or actin antibodies diluted 1:1000 (Calbiochem) in 0.1 M Tris-buffered saline containing 5% non-fat dry milk and 0.05% Tween 20. The membranes were then rinsed in Tris-buffered saline /non-fat dry milk/Tween 20 and incubated for 1 hour with horseradish peroxidase-conjugated goat anti-rabbit Igs diluted 1:2000 (Interchim, Montluçon, France). Finally, the membranes were rinsed in Tris-buffered saline and the reaction product was detected by using a chemiluminescence detection kit (Amersham Biosciences, Freiburg, Germany). The resulting signals were analyzed by densitometry and the results were expressed as the ratio of the optical density of the AQP3-corresponding band to that of actin.

Tritiated glycerol uptake

NHEK were grown in 12-well plates and treated by 10 M ATRA for 3, 24, 48, and 72 hours as indicated above. At the end of the treatment, the cells were incubated for 5 minutes in KGM-2 containing 0.1 M N-2-hydroxyethylpiperazine-N'-2-ethanesulphonic acid and 1 Ci [³H]glycerol (Amersham Biosciences). The medium was then removed and the cells were rinsed three times in cold phosphate-buffered saline and lysed in 0.3 M NaOH. The radioactivity of the cell lysates was counted by a -counter (Wallac Tricarb, Packard, Schwadorf, Austria).

Statistical analysis

Results are expressed as meansSEM. For comparisons, the Student's t-test was performed and statistical significance was considered for P<0.05. Statistical analysis was performed using GraphPad software.

Conflict of Interest

The authors state no conflict of interest.

References

1. Ahuja HS, Szanto A, Nagy L, Davies PJ (2003) The retinoid X receptor and its ligands: versatile regulators of metabolic function, cell differentiation and cell death. J Biol Regul Homeost Agents 17:29–45 | PubMed | ISI | ChemPort |
2. Chapellier B, Mark M, Messaddeq N, Calleja C, Warot X, Brocard J et al. (2002) Physiological and retinoid-induced proliferations of epidermis basal keratinocytes are differently controlled. EMBO J 21:3402–3413 | Article | PubMed | ISI | ChemPort |
3. Ecelbarger CA, Terris J, Frindt G, Echevarria M, Marples D, Nielsen S et al. (1995) Aquaporin-3 water channel localization and regulation in rat kidney. Am J Physiol 269:F663–F672 | PubMed | ISI | ChemPort |
4. Fisher GJ, Voorhees JJ (1996) Molecular mechanisms of retinoid actions in skin. FASEB J 10:1002–1013 | PubMed | ISI | ChemPort |
5. Frigeri A, Gropper MA, Umenishi F, Kawashima M, Brown D, Verkman A (1995) Localization of MIWC and GLIP water channel homologs in neuromuscular, epithelial and glandular tissues. J Cell Sci 108:2993–3002 | PubMed | ISI | ChemPort |
6. Gollnick H (2003) Current concepts of the pathogenesis of acne: implications for drug treatment. Drugs 63:1579–1596 | Article | PubMed | ISI | ChemPort |
7. Hara M, Ma T, Verkman AS (2002) Selectively reduced glycerol in skin of aquaporin-3-deficient mice may account for impaired skin hydration, elasticity, and barrier recovery. J Biol Chem 277:46616–46621 | Article | PubMed | ISI | ChemPort |
8. Ishibashi K, Sasaki S, Fushimi K, Uchida S, Kuwahara M, Saito H et al. (1994) Molecular cloning and expression of a member of the aquaporin family with permeability to glycerol and urea in addition to water expressed at the basolateral membrane of kidney collecting duct cells. Proc Natl Acad Sci USA 91:6269–6273 | Article | PubMed | ChemPort |
9. Koyama Y, Yamamoto T, Tani T, Nihei K, Kondo D, Funaki H et al. (1999) Expression and localization of aquaporins in rat gastrointestinal tract. Am J Physiol 276:C621–C627 | PubMed | ISI | ChemPort |
10. Ma T, Hara M, Sougrat R, Verbavatz JM, Verkman AS (2002) Impaired stratum corneum hydration in mice lacking epidermal water channel aquaporin-3. J Biol Chem 277:17147–17153 | Article | PubMed | ISI | ChemPort |
11. Matsuzaki T, Suzuki T, Koyama H, Tanaka S, Takata K (1999) Water channel protein AQP3 is present in epithelia exposed to the environment of possible water loss. J Histochem Cytochem 47:1275–1286 | PubMed | ISI | ChemPort |
12. Mobasheri A, Trujillo E, Bell S, Carter SD, Clegg PD, Martin-Vasallo P et al. (2004) Aquaporin water channels AQP1 and AQP3, are expressed in equine articular chondrocytes. Vet J 168:143–150 | Article | PubMed | ISI | ChemPort |
13. Nielsen S, King LS, Christensen BM, Agre P (1997) Aquaporins in complex tissues. II. Subcellular distribution in respiratory and glandular tissues of rat. Am J Physiol 273:1549–1561
14. Papoutsaki M, Lanza M, Marinari B, Nistico S, Moretti F, Levrero M et al. (2004) The p73 gene is an anti-tumoral target of the RARbeta/gamma-selective retinoid tazarotene. J Invest Dermatol 123:1162–1168 | Article | PubMed | ISI | ChemPort |
15. Rendl M, Ban J, Mrass P, Mayer C, Lengauer B, Eckhart L et al. (2002) Caspase-14 expression by epidermal keratinocytes is regulated by retinoids in a differentiation-associated manner. J Invest Dermatol 119:1150–1155 | Article | PubMed | ISI | ChemPort |
16. Rittie L, Varani J, Kang S, Voorhees JJ, Fisher GJ (2006) Retinoid-induced epidermal hyperplasia is mediated by epidermal growth factor receptor activation via specific induction of its ligands heparin-binding EGF and amphiregulin in human skin in vivo. J Invest Dermatol 126:732–739 | Article | PubMed | ISI | ChemPort |
17. Roeder A, Schaller M, Schafer-Korting M, Korting HC (2004) Tazarotene: therapeutic strategies in the treatment of psoriasis, acne and photoaging. Skin Pharmacol Physiol 17:111–118 | Article | PubMed | ISI | ChemPort |
18. Roudier N, Verbavatz JM, Maurel C, Ripoche P, Tacnet F (1998) Evidence for the presence of aquaporin-3 in human red blood cells. J Biol Chem 273:8407–8412 | Article | PubMed | ISI | ChemPort |
19. Sasaki S, Fushimi K, Saito H, Saito F, Uchida S, Ishibashi K et al. (1994) Cloning, characterization, and chromosomal mapping of human aquaporin of collecting duct. J Clin Invest 93:1250–1256 | PubMed | ISI | ChemPort |
20. Smit JV, de Jong EM, van Hooijdonk CA, Otero ME, Boezeman JB, van de Kerkhof PC (2004) Systemic treatment of psoriatic patients with bexarotene decreases epidermal proliferation and parameters for inflammation, and improves differentiation in lesional skin. J Am Acad Dermatol 51:257–264 | Article | PubMed | ISI |
21. Smith BL, Agre P (1991) Erythrocyte Mr 28, 000 transmembrane protein exists as a multisubunit oligomer similar to channel proteins. J Biol Chem 266:6407–6415 | PubMed | ISI | ChemPort |
22. Sougrat R, Morand M, Gondran C, Barre P, Gobin R, Bonte F et al. (2002) Functional expression of AQP3 in human skin epidermis and reconstructed epidermis. J Invest Dermatol 118:678–685 | Article | PubMed | ISI | ChemPort |
23. Stroud RM, Savage D, Miercke LJ, Lee JK, Khademi S, Harries W (2003) Selectivity and conductance among the glycerol and water conducting aquaporin family of channels. FEBS Lett 555:79–84 | Article | PubMed | ISI | ChemPort |
24. Sun SY, Yue P, Chandraratna RA, Tesfaigzi Y, Hong WK, Lotan R (2000) Dual mechanisms of action of the retinoid CD437: nuclear retinoic acid receptor-mediated suppression of squamous differentiation and receptor-independent induction of apoptosis in UMSCC22B human head and neck squamous cell carcinoma cells. Mol Pharmacol 58:508–514 | PubMed | ISI | ChemPort |
25. van Rossum MM, Mommers JM, van de Kerkhof PC, van Erp PE (2000) Coexpression of keratins 13 and 16 in human keratinocytes indicates association between hyperproliferation-associated and retinoid-induced differentiation. Arch Dermatol Res 292:16–20 | Article | PubMed | ISI | ChemPort |
26. Varani J, Warner RL, Gharaee-Kermani M, Phan SH, Kang S, Chung JH et al. (2000) Vitamin A antagonizes decreased cell growth and elevated collagen-degrading matrix metalloproteinases and stimulates collagen accumulation in naturally aged human skin. J Invest Dermatol 114:480–486 | Article | PubMed | ISI | ChemPort |
27. Wilson L, Gale E, Maden M (2003) The role of retinoic acid in the morphogenesis of the neural tube. J Anat 203:357–368 | Article | PubMed | ISI | ChemPort |
28. Yoshimura K, Uchida G, Okazaki M, Kitano Y, Harii K (2003) Differential expression of heparin-binding EGF-like growth factor (HB-EGF) mRNA in normal human keratinocytes induced by a variety of natural and synthetic retinoids. Exp Dermatol 12(Suppl 2):2–34 | Article |
29. Zheng X, Bollinger Bollag W (2003) Aquaporin 3 colocates with phospholipase d2 in caveolin-rich membrane microdomains and is downregulated upon keratinocyte differentiation. J Invest Dermatol 121:1487–1495 | Article | PubMed | ISI | ChemPort |

Acknowledgments

We thank Dr Georgios Stamatas for carefully reviewing this paper. We also thank Jose Serrano for expert technical assistance.

SUPPLEMENTARY MATERIAL

Table S1. Specific primer pairs used for semiquantitative PCR analysis of aquaporin expression in cultured NHEK and human epidermis.