Upregulation of Connexin 26 Between Keratinocytes of Psoriatic Lesions

Upregulation of Connexin 26 Between Keratinocytes of Psoriatic Lesions

Upregulation of Connexin 26 Between Keratinocytes of Psoriatic Lesions Marie-Pierre Labarthe, Domenico Bosco,* Jean-Hilaire Saurat, Paolo Meda,* and D...

377KB Sizes 1 Downloads 54 Views

Upregulation of Connexin 26 Between Keratinocytes of Psoriatic Lesions Marie-Pierre Labarthe, Domenico Bosco,* Jean-Hilaire Saurat, Paolo Meda,* and Denis Salomon Departments of Dermatology and *Morphology, University of Geneva Medical School, Geneva, Switzerland

To assess whether the expression of connexins (Cx) by keratinocytes is altered under conditions of abnormal epidermal differentiation, we have compared Cx26, Cx32, Cx37, Cx40, and Cx43 in the epidermis of 11 psoriatic patients who had not been treated for at least 1 mo and of seven healthy individuals. In all samples of fully mature psoriatic plaques, we have observed a massive expression of Cx26, as judged at both the transcript level (northern blot) and the protein level (immunofluorescence). This protein became consistently detected between keratinocytes of the basal and granular layers at the periphery of psoriatic plaques and in all layers of fully developed psoriatic epidermis, except in regions of parakeratosis. None or a minimal amount of Cx26 was observed in both control and nonlesional regions of psoriatic epidermis.

Psoriatic plaques also contained Cx43, the prominent gap junction protein in the interfollicular epidermis of normal human skin. The levels of this protein appeared to be slightly higher in psoriatic than in control skin, as judged at both the transcript level (northern blot) and the protein level (immunofluorescence). Three other connexins (Cx32, Cx37, and Cx40), which are not observed in control interfollicular epidermis, were not induced in either nonlesional or lesional regions of psoriatic skin. The data indicate that selective changes in the normal expression of connexins by keratinocytes are associated with the changes in the proliferation and differentiation program that these cells undergo in psoriasis. Key words: gap junctions/psoriasis. J Invest Dermatol 111:72–76, 1998

H

To assess whether connexin expression is also modified when the differentiation program of keratinocytes is altered as part of a pathologic process, we have now studied gap junction proteins in the epidermis of patients affected by psoriasis. Even though the etiology of this disease remains a matter of debate, the morphologic and biochemical alterations of interfollicular keratinocytes that take place in psoriatic epidermis are well characterized (Mc Kay and Leigh, 1995). Included in these alterations is an increase in the number of gap junctions observed by freeze-fracture electron microscopy (Caputo et al, 1978), which suggests that altered gap junctional communication somehow contributes to, or is a consequence of, the abnormal differentiation program of psoriatic keratinocytes. Here, we show that psoriatic keratinocytes massively express Cx26, a connexin that is not expressed by normal human interfollicular epidermis, in addition to native Cx43, the prominent gap junction protein of this tissue in normal human skin.

uman keratinocytes share ions (Van Heulekom et al, 1972), small metabolites (Hunter and Pitts, 1981), and low molecular weight tracers (Salomon et al, 1988) through intercellular channels, located at gap junctions. These channels are made of transmembrane proteins called connexins (Cx), which in mammals form a family of 13 members (Bruzzone et al, 1996; Kumar and Gilula, 1996). The gap junctions of human keratinocytes comprise at least two distinct connexins, referred to as Cx43 and Cx26. Cx43 has been found to be abundantly expressed throughout the interfollicular epidermis (Guo et al, 1992; Salomon et al, 1994) and in some epidermal adnexae (Salomon et al, 1994), whereas Cx26 has been found to be restricted to the latter structures (Salomon et al, 1994). Three others connexins (Cx32, Cx37, Cx40), which have not been detected in keratinocytes of normal human skin (Salomon et al, 1994), were also tested as an internal negative control. Even though the role of gap junctions in epidermal homeostasis remains to be demonstrated, it has been proposed that gap junctionmediated communication is involved in the regulation of keratinocyte growth and differentiation (Pitts et al, 1988; Salomon et al, 1993; Brissette et al, 1994; Goliger and Paul, 1995). Thus, the differential distribution of Cx43 and Cx26 within interfollicular epidermis and epidermal adnexae may be related to the specific differentiation program that keratinocytes undergo in various skin regions. Consistent with this view is the recent finding that alterations in keratinocyte differentiation induced by a chronic treatment with retinoic acid are associated with a marked increase in the expression of Cx26 within human interfollicular epidermis (Masgrau-Peya et al, 1997).

MATERIALS AND METHODS Patients Eleven psoriatic patients, who had given informed consent, were studied in accordance with the guidelines of our institutional committee for clinical investigation. All patients had chronic and stable psoriatic plaques and had not been treated by systemic drugs or phototherapy in the month preceding sampling. During this period, the skin area that was to be sampled was also not treated by any topical drug. The sex ratio of these patients was 1.8 men/women and the mean age was 38.5 y (range 21–59 y). Tissue sampling For both immunofluorescence and northern blot analysis, buttock skin was anesthetized by an intradermal injection of 1% xylocaine plus epinephrine and regions of lesional and nonlesional skin were separately sampled in five patients. In the six other patients, biopsies were performed at the border of psoriatic plaques, in order to obtain samples comprising both lesional and nonlesional areas. In the case of psoriatic plaques, samples were obtained using a dermatome set at 1.2 mm thickness, whereas for normal skin the dermatome thickness was set at 0.3–0.4 mm, so as to obtain the entire epidermis with minimal contamination of dermis. To assess whether this unavoidable

Manuscript received July 8, 1997; revised January 1, 1998; accepted for publication March 20, 1998. Reprint requests to: Dr. Denis Salomon, Clinique de Dermatologie, Hoˆpital Cantonal Universitaire, CH-1211 Gene`ve 4, Switzerland.

·

0022-202X/98/$10.50 Copyright © 1998 by The Society for Investigative Dermatology, Inc.

72

VOL. 111, NO. 1 JULY 1998

contamination influenced the connexin changes observed at the transcript level, northern blot analysis were also extended to samples of thermolysin split epidermis from control skin (Salomon et al, 1994). All samples were rapidly frozen by immersion in 2-methylbutane (Merck, ABS, Basel, Switzerland), which was cooled in liquid nitrogen and stored at –80°C until cryostat sectioning. Immunostaining Sections of 5 µm thickness were cut with a cryomicrotome (1720 digital MGW Leitz, Germany), collected on gelatin-coated slides, and fixed for 3 min in –80°C acetone. All slides were then rinsed in cold (4°C) phosphate buffered-saline containing 0.5% bovine serum albumin and processed for indirect immunofluorescence. For localization of gap junction proteins, sections were first incubated for 2 h at room temperature with one of the following antibodies: (i) affinity-purified rabbit polyclonal antibodies against liver Cx32, diluted 1:100 (Dermietzel et al, 1984); (ii) affinity-purified rabbit polyclonal antibodies against residues 108– 122 of liver Cx26, diluted 1:500 (Goliger and Paul, 1994); (iii) mouse monoclonal antibodies against a segment of the cytoplasmic loop of Cx26, diluted 1:200 (Zymed Laboratory, San Francisco, CA); (iv) affinity-purified rabbit polyclonal antibodies against the carboxy-terminal portion of mouse liver Cx26, diluted 1:100 (Traub et al, 1989); (v) affinity-purified rabbit polyclonal antibodies against residues 314–322 of heart Cx43, diluted 1:100 (Salomon et al, 1994); (vi) mouse monoclonal antibodies against 19 amino acids of the carboxy terminal portion of Cx43, diluted 1:500 (Zymed Laboratory); (vii) affinity-purified rabbit polyclonal antibodies against residues 313–330 of Cx40, diluted 1:100 (Bruzzone et al, 1993); (viii) affinity purified rabbit polyclonal antibodies against residues 266–281 of Cx37, diluted 1:200 (Haefliger et al, unpublished, Haefliger School of Medicine, University of Lausanne, Switzerland). Sections were rinsed in phosphate-buffered saline and incubated for 1 h at room temperature with fluorescein-conjugated anti-rat or anti-rabbit antibodies, whichever applicable, diluted 1:200. After further rinsing, sections were stained with a 0.03% Evans’ blue solution, covered with 0.02% paraphenylenediamine in phosphate-buffered saline-glycerol (1:2, vol:vol), and photographed with an Axioplan microscope (Zeiss, Oberkochen, West Germany) fitted with filters for fluorescein detection. In all experiments, negative controls included exposure of sections during the first incubation to either one of the following reagents: preimmune serum, normal rabbit or mouse serum, fluorescein-conjugated anti-rabbit or anti-mouse antibodies. Cryosections of rat liver, pancreas, and heart, as well as of mouse skin, were incubated in parallel with the human samples and used as positive controls for the different antibodies tested. Plasmid constructions cDNA coding for human connexins were kindly provided by Drs. G.I. Fishman (Cardiovascular Institute, New York, NY) and R. Sager (Dana Faber Cancer Institute, Boston, MA). These cDNA that contained the entire coding region of Cx43 (Fishman et al, 1990) and Cx26 (Lee et al, 1992) were subcloned in plasmid Bluescript II KS and linearized with Xho I and EcoR I (for Cx43) and Hind III and EcoR I (for Cx26) restriction site enzymes (Boehringer, Mannheim, Germany), as described in Salomon et al (1994). Connexin anti-sense probes were synthesized by in vitro transcription (Meda et al, 1993). Previous studies have shown that the probes for these two connexins have no cross-reactivity when tested on total RNA extracted from human skin (Salomon et al, 1994). Identification of connexin transcripts Skin samples were homogenized in 2.5 ml 0.1 M Tris-HCl, pH 7.4, containing 1.3 M β-mercaptoethanol and 4 M guanidium thiocyanate. After addition of solid CsCl (0.4 g per ml), the homogenate was layered on a 2 ml cushion of 5.7 M CsCl–0.1 M ethylenediamine tetraacetic acid (pH 7.4) and centrifuged for 20 h at 35,000 r.p.m. and 20°C. Pelleted RNA was resuspended in 10 mM Tris-HCl, pH 8.1, supplemented with 5 mM ethylenediamine tetraacetic acid and 0.1% sodium dodecyl sulfate, extracted twice with phenol-chloroform, precipitated in ethanol, and resuspended in water. For northern blots, total cellular RNA was denatured with 1 M glyoxal in 0.01 M phosphate buffer containing 50% dimethylsulfoxide, electrophoresed in a 1% agarose gel (5–10 µg total cellular RNA per lane), and transferred overnight onto nylon membranes (Hybond N; Amersham International). Filters were exposed for 30 s to 302 nm light and stained with methylene blue. Filters were prehybridized at 2 h and 65°C in a 0.05 M Pipesbuffered (pH 6.8) solution of 50% formamide, which was supplemented with 2 mM ethylenediamine tetraacetic acid, 0.1% sodium dodecyl sulfate, 0.1 mg salmon sperm DNA per ml, 0.8 M NaCl, and 2.53Denhardt’s solution. Filters were then hybridized for 18 h at 65°C with 23106 cpm solution 32P-labeled probe per ml, washed twice at 65°C in 33SSC and 23Denhardt’s solution, followed by three washings at 70°C in 0.23SSC, 0.1% sodium dodecyl sulfate, and 0.1% sodium pyrophosphate. Filters were then exposed to XAR-5 film (Eatsman Kodak, Rochester, NY) between intensifying screens at –80°C for 1–4 d.

CONNEXINS AND PSORIASIS

73

RESULTS Expression of Cx26 is highly upregulated in psoriatic plaques Incubations of cryosections of psoriatic plaques with two different anti-sera against Cx26 resulted in the decoration of keratinocytes by numerous immunofluorescent spots (Figs 1, 2). This immunolabeling was observed in all epidermal layers, except those showing parakeratosis, and was detected in all the patients studied (Figs 1, 2). Higher magnification views of the immunolabeled sections, revealed that virtually all the Cx26 observed in psoriatic plaques was located at discrete spots of the keratinocyte membrane, in areas of cellto-cell contact (Fig 2). In contrast, minimal levels of this protein were immunodetected in the interfollicular epidermis of control and nonlesional skin (Fig 2). These low levels escaped detection at low magnification, under conditions that clearly showed intense labeling of psoriatic skin (Fig 1). Using the same antibodies, Cx26 was also readily detected in epidermal adnexae of normal skin (not shown). Northern blot analysis showed that Cx26 mRNA was abundantly expressed in psoriatic plaques of the 11 patients studied, whereas it was not at a detectable level in samples of total control skin (Fig 3); however, minimal amounts of the Cx26 transcript were detected in preparations exclusively constituted by epidermis, which were obtained after a thermolysin splitting of normal human skin (Fig 3). Cx26 is differentially expressed throughout the epidermis of psoriatic patients Immunostaining of nonlesional skin from patients with psoriasis failed to detect Cx26 between keratinocytes of interfollicular epidermis (Fig 1); however, Cx26 was readily detected in the very same samples within hair follicles and ducts of sweat glands (not shown). Biopsies involving both nonlesional and lesional regions, revealed that Cx26 first became detectable between keratinocytes of the basal and granular layers (Fig 1). As judged by the abundance of the immunostaining, Cx26 expression appeared to progressively increase throughout the spinous layers, and was clearly detectable in all layers of viable keratinocytes within fully developed psoriatic plaques (Figs 1, 2). Northern blots of nonlesional skin of psoriatic patients also showed detectable levels of Cx26 mRNA that were much lower, however, than those found in fully developed psoriatic plaques (Fig 3). Expression of Cx43 is slightly increased in psoriatic plaques Immunostaining of cryosections using two different antibodies, revealed that Cx43 was distributed in psoriatic plaques in a similar manner to that in the interfollicular epidermis of control skin (Fig 4). Thus, in both normal and psoriatic epidermis, Cx43 was weakly immunodetected between the keratinocytes of the basal layer, was abundant in those of middle and upper spinous layers, and disappeared in those of stratum corneum, whether this layer was orthokeratotic (normal epidermis) or parakeratotic (psoriatic epidermis). No difference in Cx43 distribution was observed between nonlesional and lesional regions of the epidermis in the 11 patients examined; however, as judged by the abundance of immunofluorescent spots, Cx43 appeared more abundant in psoriatic plaques than in control and nonlesional epidermis (Fig 4). At higher magnification, it was clear that virtually all the Cx43 detected by the different antibodies used was located at discrete membrane domains (not shown). Analysis of adnexae of lesional skin failed to detect changes in Cx43 immunolabeling (not shown). Northern blot analysis showed comparable levels of Cx43 mRNA in control and psoriatic epidermis (Fig 3). Three other connexins are not upregulated in psoriatic epidermis Antibodies that identified Cx32, Cx37, and Cx40 in control human tissues (liver for Cx32, heart for Cx37 and Cx40), failed to detect these gap junction proteins in the interfollicular epidermis and epidermal adnexae of both control and psoriatic skin (not shown). These negative findings are in agreement with previous reports on both human (Salomon et al, 1994) and rodent epidermis (unpublished). DISCUSSION Human keratinocytes are connected by numerous gap junctions (Wolff and Schreiner, 1968), which allow for direct cell-to-cell exchanges of

74

LABARTHE ET AL

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Figure 1. Distribution of Cx26 in lesional and nonlesional psoriatic skin. (A) Phase-contrast view of a biopsy comprising a phenotypically normal region of skin (left side of the photograph) and a region immediately adjacent to a psoriatic plaque (right side of the photograph). (B) Immunolabeling failed to reveal Cx26 in normal interfollicular epidermis (left side of the photograph), but showed this gap junction protein in the basal, upper spinous, and granular layers of the same tissue in the region surrounding a fully developed psoriatic plaque (right side of the photograph). (C) Phase-contrast view of the fully developed psoriatic plaque that is shown in (D). (D) Within psoriatic plaques, Cx26 was detected throughout all living layers of keratinocytes, where it was expressed in much higher levels than in the nonlesional and perilesional regions of the very same patients. In contrast, no Cx26 was detected in the keratinized and parakeratotic regions of squamous layers. Scale bar: 100 µm.

Figure 2. Cx26 is expressed at the membrane of keratinocytes. High magnification analysis of the skin sections incubated with antibodies against Cx26, revealed that this protein was clearly located at the membrane of the keratinocytes that formed the bulk of psoriatic plaques (A). Comparison with the situation in normal skin (B) further revealed that the membrane spots immunostained for Cx26 were much more abundant in psoriatic plaques than in control interfollicular epidermis. Scale bar: 25 µm.

Table I. Connexins 43 and 26 are upregulated between keratinocytes of psoriatic epidermisa Psoriatic epidermis Normal epidermis

Str. Basale Str. Spinosum Str. Granulosum Str. Corneum orthokeratosis parakeratosis

Nonlesional

Border

Lesional

Cx43

Cx26

Cx43

Cx26

Cx43

Cx26

Cx43

Cx26

1 111 11

– – 1/–

1 111 11

– – 1/–

1 111 111

1 – 111

11 1111 111

1 111 111

– n.o.

– n.o.

– n.o.

– n.o.

– n.o.

– n.o.

– –

– –

a1 and –, presence and absence, respectively, of immunolabeling of connexin 43 and connexin 26; 1/–, a weak inconsistent immunolabeling of connexin 43 and connexin 26; n.o., none observed

VOL. 111, NO. 1 JULY 1998

Figure 3. Cx43 and Cx26 transcripts are differentially increased in psoriatic epidermis. Upper panel: northern blot analysis of total RNA (5 µg per lane) revealed the presence of a Cx26 mRNA in samples of both lesional (L) and nonlesional (NL) psoriatic skin, but not in samples of total skin (TS) obtained from healthy controls. In all psoriatic patients, the levels of this mRNA were much larger in plaque regions showing the full complement of psoriatic alterations (L) than in nearby, phenotypically normal regions of the skin (NL). No Cx26 mRNA was detectable in samples of total control skin (TS). By contrast, minute traces of Cx26 mRNA were barely detectable in samples of epidermis that were prepared by enzymatic splitting (SE) of normal skin. Lower panel: when the same RNA samples were tested for Cx43, a transcript encoding this connexin was detected in all samples studied. No quantitative analysis of this type of blot was performed in this study.

low molecular weight cytoplasmic molecules and ions (Pitts et al, 1988; Salomon et al, 1993; Goodenough et al, 1996). In the interfollicular epidermis of normal human skin, the predominant gap junction protein is Cx43 (Guo et al, 1992; Salomon et al, 1994), a connexin that is weakly expressed between the keratinocytes of the basal layer and abruptly increases between those of suprabasal layers (Salomon et al, 1994). In the granular layers, Cx43 is coexpressed with minute amounts of another gap junction protein, referred to as Cx26 (Kelsell et al, 1997), which is not detectable by immunolabeling elsewhere in the human interfollicular epidermis (Guo et al, 1992; Wingelbus et al, 1992; Salomon et al, 1994; Rivas et al, 1997), but is abundant within hair follicle and ducts of sweat glands (Salomon et al, 1994; Rivas et al, 1997). This differential distribution suggests that gap junctions and the direct cell-to-cell communications that these structures permit, contribute to the program of migration and acquisition of specific structural and metabolic features that lead to full keratinocyte differentiation. This contribution is further supported by the recent finding that a chronic application of all-trans retinoic acid, a treatment that is known to markedly modify the differentiation program of adult epidermis (Rosenthal et al, 1992), is associated with changes in connexins, including a modest increase in the expression of Cx43 and a much more strikingly and large induction of Cx26 (Masgrau-Peya et al, 1997). To assess whether connexin changes also take place during

CONNEXINS AND PSORIASIS

75

pathologic alterations of human epidermis, we have studied psoriasis, a skin disease characterized by major changes in the proliferation, migration, differentiation, and shedding of keratinocytes (Mc Kay and Leigh, 1995). We have found that, compared with normal interfollicular epidermis, the thickened and parakeratotic epidermis of psoriatic plaques featured a modest increase in the levels of Cx43 and a massive induction of Cx26, as judged at both the transcript and the protein levels. These findings are consistent with the recent observation that the gene encoding Cx26 is among the few that are markedly upregulated in psoriatic epidermis (Rivas et al, 1997). This study further documents that the increased expression of gap junction proteins occurred according to a specific spatial pattern. Hence, even though Cx43 and Cx26 were found increased throughout all living layers of the interfollicular epidermis, this change predominated in the suprabasal layers. At the light microscopic level, both Cx26 and Cx43 appeared to be mostly expressed at the membrane of keratinocytes, consistent with previous reports locating enlarged gap junction plaques on the membrane of psoriatic keratinocytes (Caputo et al, 1978). Interestingly, no change in either Cx26 or Cx43 was detected within epidermal adnexae, which are little, if at all, involved in psoriasis. Furthermore, the induction of Cx26 appeared less prominent at the border than in the center of psoriatic plaques, and was not observed in the noninvolved regions that displayed a histologically normal skin organization. In these regions, the interfollicular epidermis of psoriatic patients contained Cx43 and virtually no Cx26, like that of healthy controls. These findings indicate that the upregulation of the Cx26 gene and protein expression parallels the gradient of histologic and metabolic alterations that lead from a noninvolved epidermis to the hyperplastic and abnormally differentiated epidermis of fully mature psoriatic plaques. The observation of altered expression of Cx26 in areas that did not yet feature all the histologic alterations that characterize these fully mature plaques, further indicates that the connexin changes are associated with the progression of psoriatic alterations. Because many other biochemical changes also show a gradient of expression between the noninvolved skin and the fully developed plaques of psoriasis (Mc Kay and Leigh, 1995), further studies are now required to determine which of these changes is necessary, if not causal for the progression of psoriatic alterations. The modest change in Cx43 and the massive induction of Cx26 observed here, are reminiscent of the connexin changes reported after chronic application of retinoic acid on normal skin (Masgrau-Peya et al, 1997). Under these conditions the thickness of interfollicular epidermis was increased and keratinocytes were induced to express keratin K6, a marker of increased proliferation (Rosenthal et al, 1992), which is also observed in psoriasis (Stoler et al, 1988). Together with preliminary observations of analogous connexin changes in other genodermatosis (unpublished data) and skin cancers (Wingelbus et al, 1992) (unpublished data), these findings indicate that alterations in the native pattern of connexins, and most strikingly in the expression of Cx26, are consistent markers of abnormal keratinocyte differentiation. In vitro studies have shown that Cx26 and Cx43 make gap junction channels with different biophysical and regulatory properties and are not able to functionally interact with each other to form heterotypic channels (White et al, 1995). Hence, one implication of our findings is that, whereas channels made of Cx43 are sufficient to fulfill the needs of junctional communication of normal keratinocytes, some change in these needs requires an additional type of cell-to-cell channel under pathologic conditions that leads to epidermal hyperplasia. Further experiments should investigate the reasons why Cx26 is adequately tailored for these new requirements. An approach to this question is to determine whether the increased level of this connexin in psoriatic epidermis is actually paralleled by an increase in the functional coupling of keratinocytes. Future experiments will determine whether the newly formed Cx26 channels that add to those normally made by Cx43 upregulate or downregulate the native keratinocyte-to-keratinocyte communication within the interfollicular epidermis. Communication via connexins has been implicated in the regulation of cell growth (Goliger and Paul, 1995; Sawey et al, 1996; Rivas et al, 1997), migration (Pepper et al, 1989; Goliger and Paul, 1995), and differentiation (Lowenstein and Rose, 1992; Goodenough et al, 1996),

76

LABARTHE ET AL

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Figure 4. Distribution of Cx43 in lesional and nonlesional psoriatic skin. (A) Phase-contrast view of the biopsy of a psoriatic patient, comprising a phenotypically normal region of skin (left side of the photograph) and a region immediately adjacent to a fully developed psoriatic plaque (right side of the photograph). The latter region is recognizable by the characteristic lengthening of epidermal papillae. (B) Immunolabeling revealed that, in the entire biopsy, Cx43 was distributed throughout all living layers of interfollicular epidermis and predominantly in the spinous and granular layers. (C) Within fully developed psoriatic plaques, the interfollicular epidermis was markedly thickened and featured long epidermal papillae and regions of parakeratosis. (D) In these plaques, the distribution of Cx43 was qualitatively similar to that observed in the nonlesional regions of skin, i.e., was higher in the spinous and granular layers than in the basal layer of keratinocytes and was nil in the keratinized and parakeratotic regions of the squamous layers; however, the levels of Cx43 within psoriatic plaques appeared increased compared with those observed in nonlesional regions, particularly within the basal and parabasal layers of keratinocytes. Scale bar: 100 µm.

three functions whose alteration significantly contributes to the dysfunction of psoriatic keratinocytes (Mc Kay and Leigh, 1995). Elucidating which of these functions, if any, may be affected by altering connexin channels awaits the result of experiments in which in vitro reconstituted epidermis, starting from primary human keratinocytes, are combined with methods for the targeting of specific connexin genes to distinct keratinocyte subpopulations.

We thank A. Charollais, F. Cogne, J.-P. Gerber, and E. Sutter for excellent technical assistance. This work was supported by grants from the Swiss National Science Foundation (32–30211.95, 32–34086.95), the French Cancer Research Association, the Juvenile Diabetes Foundation International (197124), and the European Union (BMH4CT96–1427).

REFERENCES Brissette J, Kumar N, Gilula N, Hall J, Dotto G: Switch in gap junction protein expression is associated with selective changes in junctional permeability during keratinocyte differentiation. Proc Natl Acad Sci USA 91:6453–6457, 1994 Bruzzone R, Haefliger JA, Gimlich RL, Paul DL: Connexin 40, a component of gap junctions in vascular endothelium, is restricted with its ability to interact with other connexins. Molec Biol Cell 4:7–20, 1993 Bruzzone R, White T, Paul D: Connections with connexins: the molecular basis of direct intercellular signalling. Eur J Cell Biol 238:1–27, 1996 Caputo R, Innocenti M, Gasparini G, Peluchetti D: Plasma membrane in psoriatic cells. A freeze fracture study. J Invest Dermatol 71:245–249, 1978 Dermietzel R, Leibstein A, Frixen U, Janssen-Timmen U, Traub O, Willecke K: Gap junctions in several tissues share antigenic determinants with liver gap junctions. EMBO J 3:2261–2270, 1984 Fishman GI, Spary DC, Leinwand LA: Molecular characterization an functional expression of the human cardiac gap junction channel. J Cell Biol 111:589–598, 1990 Goliger JA, Paul DL: Expression of gap junction proteins Cx26, Cx31.1, Cx37, and Cx43 in developing and mature rat epidermis. Dev Dyn 200:1–13, 1994 Goliger JA, Paul DL: Wounding alters epidermal connexin expression and gap junctionmediated intercellular communication. Mol Biol Cell 6:1491–1501, 1995 Goodenough DA, Goliger JA, Paul DL: Connexins, connexons, and intercellular communications. Annu Rev Biochem 65:475–502, 1996 Guo H, Acevado P, Parsa FD, Bertram JS: Gap junctional protein connexin 43 is expressed in dermis and epidermis of human skin: differential modulation by retinoids. J Invest Dermatol 99:460–467, 1992 Hunter K, Pitts JD: Non-selective junctional communication between some different mammalian cell types in primary culture. J Cell Sci 49:163–175, 1981 Kelsell DP, Dunlop J, Stevens HP, Lench NJ, Liang JN, Mueller RF, Leigh IM: Connexin

26 mutations in hereditary non-syndromic sensorineural deafness. Nature 387:80– 83, 1997 Kumar NM, Gilula NB: The gap junction communication channel. Cell 84:381–388, 1996 Lee SW, Tomasetto C, Paul D, Keyomarsi K, Sager R: Transcriptional downregulation of gap-junction proteins blocks junctional communication in human mammary cell lines. J Cell Biol 118:1213–1221, 1992 Lowenstein WR, Rose B: The cell-cell channel in the control of growth. Semin Cell Biol 3:59–79, 1992 Masgrau-Peya E, Salomon D, Saurat J-H, Meda P: In vivo modulation of connexins 43 and 26 of human epidermis by topical retinoic acid treatment. J Histochem Cytochem 45:1207–1215, 1997 Mc Kay IA, Leigh IM: Altered keratinocyte growth and differentiation in psoriasis. Clin Dermatol 13:105–114, 1995 Meda P, Pepper M, Traub O, et al: Differential expression of gap junction connexins in endocrine and exocrine glands. Endocrinology 133:2371–2378, 1993 Pepper MS, Spray DC, Chanson M, Montesano R, Orci L, Meda P: Junctional communication is induced in migrating capillary endothelial cells. J Cell Biol 109:3027–3038, 1989 Pitts JD, Finbow ME, Kam E: Junctional communication and cellular differentiation. Br J Cancer 9 (Suppl.):52–57, 1988 Rivas MV, Jarvis ED, Morisaki S, Carbonaro H, Gottlieb AB, Krueger JG: Identification of aberrantly regulated genes in diseased skin using the cDNA differential display technique. J Invest Dermatol 108:188–194, 1997 Rosenthal DC, Griffiths EM, Yuspa S, Roop D, Voorhees J: Acute or chronic topical retinoic acid treatment of human skin in vivo alters the expression of epidermal transglutaminase, Loricrin, Involucrin, Filaggrin and keratins 6 and 13 but not keratins 1, 10 and 14. J Invest Dermatol 98:343–350, 1992 Salomon D, Saurat J-H, Meda P: Cell-to-cell communication within intact human skin. J Clin Invest 82:248–254, 1988 Salomon D, Masgrau E, Vischer S, Chanson M, Saurat J-H, Spray DC, Meda P: Gap junction proteins and communication in human epidermis. Prog Cell Res 3:225– 231, 1993 Salomon D, Masgrau E, Vischer S, et al: Topography of mammalian connexins in human skin. J Invest Dermatol 103:240–247, 1994 Sawey M, Goldschmidt M, Risek B, Gilula N, Lo C: Perturbation in Cx43 et Cx26 gap junction expression in mouse skin hyperplasia and neoplasia. Mol Carc 17:49–61, 1996 Stoler A, Kopan R, Duvic M, Fuchs E: Use of non-specific antisera and cRNA probes to localize the major changes in keratin expression during normal and abnormal epidermal differentiation. J Cell Biol 107:427–446, 1988 Traub O, Look J, Dermietzel R, Brummer F, Hu¨lser D, Willecke K: Comparative characterisation of the 21-kD and 26-kD gap junction proteins in murine liver and culture hepatocytes. J Cell Biol 108:1039–1051, 1989 Van Heulekom JS, Slaaf DW, Van der leun JC: Cell communications in the basal cells of human epidermis. Biophys 12:1266–1284, 1972 White T, Paul D, Goodenough DA, Bruzzone R: Functional analysis of selective interactions among rodent connexins. Mol Biol Cell 6:459–470, 1995 Wingelbus KK, Kirkpatrick CJ, Knuechel R, Willecke K, Traub O: Expression of Cx26, Cx32 and Cx43 gal junction proteins in normal and neoplastic human tissues. Inter J Cancer 51:522–529, 1992 Wolff K, Schreiner E: An electron microscopy study on the extraneous coat of keratinocytes and the intercellular space of the epidermis. J Invest Dermatol 51:418–430, 1968