In vivo long-term effects of retinoic acid exposure in utero on induced hyperplastic epidermal foci in murine skin

© 2007 The Authors. Journal compilation © 2007 ESVD and ACVD. 18; 287–293 287

Blackwell Publishing Ltd

In vivo long-term effects of retinoic acid exposure

in utero on induced hyperplastic epidermal foci in

murine skin

Rosa A. García-Fernández*, Claudia

Pérez-Martínez†, Javier Espinosa-Álvarez†

and Maria J. García-Iglesias†

*Histology and Pathological Anatomy Section, Department of Animal Medicine and Surgery, Faculty of Veterinary Science, University of Madrid (UCM), Madrid, Spain†Histology and Pathological Anatomy Section, Department of Animal Health, Faculty of Veterinary Science, University of León, León, SpainCorrespondence: Maria J. García-Iglesias, Histology and Pathological Anatomy Section, Department of Animal Health, Faculty of Veterinary Science, University of León, León, Spain. E-mail: [email protected]

What is known about the topic of this paper

• Retinoic acid (RA) plays an important role in embryogenesisand is necessary to maintain the differentiation pathwayof the epithelial tissues (skin, and other epithelia).

• RA has been shown to inhibit the induction of cancer invarious organs including the skin.

• Retinoids are used to treat skin diseases such as psoriasis,actinic keratosis, squamous and basal cell carcinomas.

What this paper adds to the field of veterinary

dermatology

• RA administered to pregnant mice shows an in vivo long-term action on development of hyperplastic lesions inadult skin of their offspring.

• This retinoid inhibits the cell differentiation and stimulatesthe cell proliferation in those hyperplastic lesions.

• The chemopreventive effect of RA is questioned usingour experimental mouse model.

Abstract

Adult Naval Medical Research Institute (NMRI) mice,

after prenatal exposure to retinoic acid (RA), were

treated with a standard two-stage skin carcinogenesis

regime to characterize hyperplastic epidermal foci

that precede the appearance of cutaneous papillomas,

and to investigate the in vivo long-term action of RA

on adult mouse skin treated with DMBA (7,12 dimethyl

benz[a]anthracene) and TPA (12-O-tetradecanoylphorbol

13-acetate). The results demonstrate that RA admin-

istered to pregnant mice had a long-term inhibitory

action on the cell differentiation and development of

hyperplastic lesions occurring prior to cancer on the

adult skin of their offspring as well as a stimulatory

effect on cell proliferation of these hyperplastic lesions.

Accepted 8 June 2007

Introduction

Retinoic acid (RA), the most biologically active naturalmetabolite of circulating vitamin A,1 plays an importantrole in embryogenesis1 and is fundamental to maintenanceof the differentiation pathway of epithelial tissues.2 Previousstudies have demonstrated that prenatal RA exposureinduces earlier epidermal differentiation3 and a stimulatoryeffect on cell proliferation of the developing epidermis3

and hair follicles.4 RA and its analogues (retinoids) areused in the therapy of dermatological5,6 and neoplastic dis-eases,7 although their clinical usefulness is limited by theirproven teratogenic potential.8–10 Retinoids have been shownto be effective in the prevention of skin cancer7 in differentanimal models, following either topical application11 or oraladministration.12,13 However, there appears to be nothingknown about the in vivo long-term effects on skin carcino-genesis after prenatal RA exposure in mice.

The effects of retinoids on skin carcinogenesis havebeen studied using a two-stage chemical model involvingthe application of 7,12 dimethyl benz[a]anthracene (DMBA)as an initiator and 12-O-tetradecanoylphorbol 13-acetate(TPA) as a promoter.7 Polymerase chain reaction analysisspecific for the Ha-ras gene has provided molecular evidencethat squamous cell hyperplastic foci are precursors ofcutaneous papillomas.14 In this way, it is possible to usesquamous cell hyperplastic foci as the earliest indicator ofpotential mouse skin tumorigenicity.14

Morphological modifications of the epidermis in bothmultistage mouse skin carcinogenesis15,16 and topical RAtreatment17 have been associated with changes in keratinexpression. As keratins are considered to be ‘permanent’markers for keratinocytes18 as well as useful tools forinvestigation of skin morphogenesis and differentiation,19

the aim of this study was to investigate the in vivo long-termeffects of prenatal RA exposure on hyperplastic epidermalfoci (HEFs) induced by DMBA/TPA in NMRI adult mouseskin by quantification of the number of HEFs and analysisof their keratin expression.

Materials and methods

Animals

Two-week-old albino NMRI mice (n = 56) were maintained in a cycleof 12 h light and 12 h dark, at a controlled temperature (21 ± 1 °C)and relative humidity (55 ± 10%) with food and water ad libitum. Allexperiments were performed following the guidelines of the EuropeanLaw on Laboratory Animal Protection.20

Twenty-eight of these mice had previous exposure to RA in uteroand were obtained from nine pregnant females treated by oral admin-istration with 30 mg kg–1 body weight of all-trans-retinoic acid (RA)(Sigma Chemical Company, St. Louis, MO, USA) in corn oil on day

288 © 2007 The Authors. Journal compilation © 2007 ESVD and ACVD.

García-Fernández et al.

11.5 of gestation according to a previously described schedule.3,4 Theother 28 were obtained from nine untreated pregnant females.

Two-stage skin carcinogenesis protocol

The animals were assigned to four treatment groups:

1 RA-carcinogen group: 20 mice (12 females and 8 males) withprevious exposure to RA in utero were treated with a two-stageskin carcinogenesis protocol. This involved shaving the dorsal skin3 days before topical application of a single dose of DMBA (20 µgin 100 µL of acetone) (Sigma Chemical Company) followed aweek later and thereafter twice weekly by application of thetumour promoter TPA (2 µL in 100 µL of acetone) (Sigma ChemicalCompany) for 20 weeks.

2 Carcinogen control group: 19 mice (8 females and 11 males) withno previous exposure to RA in utero were treated with DMBAand TPA following the two-stage skin carcinogenesis protocoldescribed above and served as carcinogen treatment controls.

3 RA group: eight mice (4 females and 4 males) with previousexposure to RA in utero received only topical application ofacetone (100 µL per mouse) on the shaved area of skin from thesecond week of life.

4 Vehicle control group: nine mice (4 females and 5 males) with noprior exposure to RA in utero also received only topical applicationof acetone (100 µL per mouse) on the shaved area of skin fromthe second week of life.

The mice were sacrificed by cervical dislocation 7 days after thelast treatment, and the dorsal skin was removed and fixed in Bouin’ssolution.

Histology

The whole fixed dorsal skin from each animal was divided into2-mm-thick samples that were routinely processed and paraffin-embedded. Sections (3 µm) were stained with haematoxylin andeosin and Masson’s trichrome stains.

Immunohistochemistry

Sections from all four groups of animals, mounted on poly-L-lysine-coatedslides, were stained for different keratin types using the avidin-biotin-peroxidase complex (ABC) method (Peroxidase Standard,Vectastain, ABC kit, Vector Laboratories, Burlingame, CA, USA). Theworking dilutions and keratin specificity of the primary antibodiesused are shown in Table 1. The immunoreactivity of masked antigenswas restored by heat treatment21 and diaminobenzidine (DAB, substratekit for peroxidase, Vector Laboratories) was used as the chromogen;the sections were counterstained with Harris haematoxylin. Sectionsfrom mouse dorsal epidermis were included as positive controls witheach assay. As a negative control, the primary antibody was replacedby its diluent, TBS (Tris-buffered solution).

Classification of hyperplastic epidermal foci

HEFs were characterized by the focal increase of non-dysplasticepidermal layers. Their growth pattern was clearly not exophytic, amajor morphological difference from papilloma where the epithelialand fibrovascular components rise well above the surface of theadjacent skin.

HEFs were classified into three stages (1, 2 and 3) on the basis oftwo major criteria: (i) the number of keratinocyte layers in the livingepidermis above the stratum basale and (ii) K1/K10 and K6 immuno-histochemical expression in these layers. Classification was estab-lished from both histological and immunohistochemical evaluation ofcarcinogen-induced HEFs.

Statistical analysis

Student’s t-test22 was used to test whether the mean values of HEFsdiffered between the carcinogen control and the RA-carcinogengroups. Data were expressed as mean ± standard deviation (SD). Tocompare the results of keratin expression in HEFs between thecarcinogen control and RA-carcinogen groups, a chi-squared test was

used. Data on keratin expression were presented as percentages.Differences were considered statistically significant at P < 0.05.Analysis was carried out using SPSS 13.0® (SPSS Inc., Chicago, IL,USA) for Windows.

Results

Vehicle control group

In mice with no previous exposure to RA in utero and nocarcinogen treatment, histological evaluation revealed anormal epidermis and an absence of HEFs (Table 2). Thethin epidermis of these control animals consisted of asingle layered stratum basale, one to two suprabasal celllayers (strata spinosum and granulosum) and a stratumcorneum (Fig. 1a). The stratum basale was composed ofa single cell layer that expressed K5 (Fig. 1b) and K14(Table 3). The one to two suprabasal cell layers reactedpositively for both K1 (Fig. 1c) and K10 keratins, whereasthe stratum corneum was completely negative for keratins(Table 3). The entire epidermis was negative for K6 keratin(Table 3), which was, however, expressed in hair follicles(Fig. 1d).

RA group

No hyperplastic epidermal foci were observed in RA-exposedmice receiving only topical application of acetone (vehicle)(Table 2). The main structural and immunohistochemicalcharacteristics of their epidermis were similar to thosedescribed above in vehicle control group (Fig. 1) (Table 3).

Carcinogen control group: classification of HEFs

Histological evaluation showed foci of increased epidermalthickness, which were referred to as hyperplastic epidermal

Table 1. Primary antibodies used for immunohistochemical labelling

CloneCharacter/species Specificity Dilution* Source

AF 109 po/Rb K1 1 in 2 × 105 BabCO, California, USAAF 138 po/Rb K5 1 in 105 BabCO, California, USAMK6 po/Rb K6 1 in 105 BabCO, California, USAMK10 po/Rb K10 1 in 4 × 105 BabCO, California, USAAF 64 po/Rb K14 1 in 105 BabCO, California, USA

po, polyclonal; K, keratin; Rb, rabbit. *µL in µL (microlitre in microlitre).

Table 2. Mean and standard deviation of DMBA-TPA-induced hyperplastic epidermal foci (HEFs): statistical comparison between carcinogen control and retinoic acid (RA)-carcinogen groups

Stage HEFs

Treatment groups

Vehicle control RA

Carcinogen control RA-carcinogen

Stage 1 0 0 10.89 ± 4.89 11.10 ± 7.02Stage 2 0 0 10.00 ± 5.68* 6.55 ± 8.70*Stage 3 0 0 1.84 ± 1.86 1.35 ± 1.95Total 0 0 22.74 ± 9.98 19.00 ± 13.75

Stage 1, HEFs with three to five suprabasal cell layers. Stage 2, HEFs with six to eight suprabasal cell layers. Stage 3, HEFs with > 8 suprabasal cell layers. Total, mean number of HEFs. *Significant differences between carcinogen control and RA-carcinogen groups (P < 0.01) in stage 2 HEFs.

© 2007 The Authors. Journal compilation © 2007 ESVD and ACVD. 289

Long-term RA action on chemically induced epidermal hyperplasia

foci (HEFs). Differences in the increase of epidermallayers and immunostaining of these HEFs allowed theirclassification into three stages (1, 2 and 3) that appearedto be consistent with temporal development of hyperplasticchanges.

Stage 1 HEFs exhibited a slightly thicker epidermis(cf. Fig. 2a,b) that consisted of three to five suprabasal celllayers expressing K6 in 14.3% of those examined. Immu-nostaining for K5 and K14 keratins was always positive inthe basal cell layer. The HEFs also exhibited a reaction forboth K5 (64.7%) (Fig. 3) and K14 (100%) in their suprabasalcell layers (Table 4).

Stage 2 HEFs (Fig. 2c) were composed of six to eightsuprabasal cell layers. Their keratin expression patternwas qualitatively similar to that of stage 1. However, anincreased percentage of HEFs expressing K6 (63.6%) wasfound in comparison to stage 1 HEFs (14.3%), althoughthis difference was not statistically significant (Table 4).This immunohistochemical finding was one of the criteriaused to distinguish the two stages.

Stage 3 HEFs (Fig. 2d) were characterized by thepresence of more than eight suprabasal cell layers and areduced expression for both K1 (11.1%) (Fig. 4) and K10(100%) in the suprabasal cell layers of those examined

(Table 5). This immunohistochemical finding had not beenobserved in prior stages 1 and 2 (Fig. 5). In stage 3 HEFs,the expression of K5, K14 and K6 was qualitatively similarto previous stages (Table 4).

In the epidermal sections examined, different admixedHEFs stages were often seen. In this way, it was notuncommon to find stage 2 or 3 HEFs surrounded by anepidermis showing features of stage 1 or 2 HEFs, respec-tively, suggesting a possible evolution from one stage toother.

Quantification of chemically induced HEFs:

differences among groups

As statistical analysis did not show any sex influence inthe number (mean ± SD) of carcinogen-induced HEFswithin both RA-exposed (24.27 ± 10.03 in males vs.20.63 ± 9.99 in females) and not exposed to RA(26.57 ± 13.41 in males vs. 16.17 ± 12.44 in females)groups; later statistical analysis was carried out independentof sex. Table 2 shows the mean number of HEFs in thetreatment groups taking into account their stages.

An important difference was that only carcinogen-treatedanimals (exposed and non-exposed to RA) developedHEFs in contrast to non-carcinogen-treated mice thatnever showed epidermal hyperplasia (Table 2).

In carcinogen-treated animals, a significant decreasein the mean number of stage 2 HEFs was shown in RA-exposed mice compared with the group not exposed toRA. Likewise, a decrease in the average number of HEFsregardless of their stage, was also observed, but it wasnot statistically significant (Table 2).

Histological and immunohistochemical differences

among groups

Carcinogen treatment-induced immunohistochemical changesin the strata spinosum and granulosum of HEFs comparedwith the normal keratin expression were shown in the

Figure 1. Mouse dorsal skin from vehiclecontrol group. (a) Structure of the epidermis:stratum basale (SB), suprabasal cell layers (SCL)(including strata spinosum and granulosum),stratum corneum (SC). Masson’s trichromestain. (b) Epidermis showing K5 positivity in thestratum basale. (c) Epidermis showing K1 positivityin the suprabasal cell layers. (d) Epidermisshowing negativity for K6 expression butpositivity in the hair follicle (arrow). Avidin–biotin–peroxidase complex; Harris haematoxylin.

Table 3. Characteristic keratin expression in basal and suprabasal cell layers of mouse epidermis in noncarcinogen-treated animals

Vehicle control and RA groups

Keratin expression pattern in mouse epidermal cells

Epidermal layers K14 K5 K1 K10 K6

Basal cell layer + + – – –Suprabasal cell layers – – + + –Stratum corneum – – – – –

RA, retinoic acid. +, presence of keratin expression in all epidermal cells. –, absence of keratin expression in all epidermal cells.

290 © 2007 The Authors. Journal compilation © 2007 ESVD and ACVD.

García-Fernández et al.

epidermis from mice without carcinogen treatment (vehiclecontrol and RA groups). Therefore, HEFs from both exposedand non-exposed mice to RA receiving carcinogen treatmentoften showed an aberrant expression for K14, K5 and K6in suprabasal epidermal cells as well as a reduced expres-sion for suprabasal keratins K1 and K10. No differences forboth K5 and K14 expression in the stratum basale of theepidermis were found among the four groups examined(Tables 3 and 4).

When both groups of carcinogen-treated mice (animalsexposed and non-exposed to RA) were compared, thehistological features of HEFs did not differ betweenmice whether exposed or not to RA, whereas RA-inducedchanges in keratin expression in their strata spinosum andgranulosum were seen when compared with that observed

in mice not exposed to this retinoid. Thus, prenatal RAexposure induced a statistically significant increase in thepercentage of stage 1 HEFs expressing K6 (47.1%) incomparison with that (14.3%) found in non-exposed miceto RA (Table 4). The suprabasal expression of K5 wasobserved in a higher percentage of HEFs from RA-exposedmice, although this increase with respect to mice notexposed to RA was only statistically demonstrated instages 1 and 2 HEFs (Table 4). No K14 expression differences

Figure 2. Mouse dorsal skin. (a) Vehicle controlgroup. Normal structure. (b) Carcinogen controlgroup. Stage 1 hyperplastic epidermal foci (HEFs)(arrowhead). (c) Carcinogen control group. Stage 2HEFs (arrowhead). (d) Carcinogen control group.Stage 3 HEFs (arrowhead). Masson’s trichromestain.

Figure 3. Mouse dorsal skin from the carcinogen control group.Stage 1 hyperplastic epidermal foci. Note the K5 expression in boththe basal and the suprabasal cell layers of the epidermis and in the hairfollicles. Avidin–biotin–peroxidase complex; Harris haematoxylin.

Table 4. Percentage of mouse hyperplastic epidermal foci (HEFs) with aberrant expression of K14, K5 and K6 in the suprabasal cell layers

Treatment groupsPercentage of HEFs with positive keratin expression

Epidermal layers K14 K5 K6

Carcinogen controlBasal cell layer

Stages 1, 2 and 3 HEFs 100 100 0Suprabasal cell layers

Stage 1 HEFs 100.0 64.7* 14.3**Stage 2 HEFs 100.0 72.73* 63.6Stage 3 HEFs 100.0 75.0 100.0

RA-carcinogenBasal cell layer

Stages 1, 2 and 3 HEFs 100 100 0Suprabasal cell layers

Stage 1 HEFs 100.0 100.0* 47.1**Stage 2 HEFs 100.0 100.0* 72.7Stage 3 HEFs 100.0 100.0 100.0

Stage 1, HEFs with three to five suprabasal cell layers. Stage 2, HEFs with six to eight suprabasal cell layers. Stage 3, HEFs with > 8 suprabasal cell layers.*Significant differences for K5 expression between carcinogen control and RA-carcinogen groups (P < 0.05) in stages 1 and 2 HEFs.**Significant differences for K6 expression between carcinogen control and RA-carcinogen groups (P < 0.05) in stage 1 HEFs.

© 2007 The Authors. Journal compilation © 2007 ESVD and ACVD. 291

Long-term RA action on chemically induced epidermal hyperplasia

were found between carcinogen control and RA-carcinogengroups. Another RA-induced immunohistochemical changewas a reduced expression for K1 and K10 in the stage 2HEFs (cf. Figs 5, 6) in comparison to non-exposed RAmice, whereas this loss of expression was observed forthe first time in stage 3 HEFs from non-exposed mice toRA (Fig. 4) (Table 5).

Discussion

The results show that prenatal exposure of mice to RA onday 11.5 of gestation decreased the number of hyperplasticepidermal foci and modified the keratin expression patternin skin lesions occurring during induced multistage carcino-genesis in adult mice. These findings support a conclusion

that in vivo long-term activity of RA can influence chemicalinduction of HEFs.

The development of epidermal hyperplasia in mouseback skin and the aberrant expression of K14/K5 (keratinslinked to cellular proliferative capacity)23 and K6 (commonlyexpressed in hyperproliferative epidermis)18 suggest astimulatory effect on cell proliferation, which has beenpreviously associated with multistage mouse skin carcino-genesis and therefore with a carcinogen action.14,16,24 Astriking finding was that RA exposure in utero increasedsignificantly this proliferative activity in stages 1 and 2HEFs compared with the carcinogen effect, as was shownby K5 and K6 immunostaining. This RA hyperproliferativeeffect has also been described in vitro,25 in mice treatedtopically,17,26 and in newborn mice exposed to RA in utero.3

Similarly, the loss of cell differentiation shown by areduced K1 and K10 expression at the advanced stage 3of epidermal hyperplasia was also associated with thecarcinogenic action. Interestingly, an earlier loss of celldifferentiation related to in utero RA exposure was shown

Figure 4. Mouse dorsal skin from the carcinogen control group.Stage 3 hyperplastic epidermal foci (HEFs) showing reduced K1expression in the suprabasal cell layers. Some groups of cells did notexpress K1 (arrowheads). Avidin–biotin–peroxidase complex; Harrishaematoxylin.

Table 5. Percentage of mouse hyperplastic epidermal foci (HEFs) with reduced K1/K10 expression in the suprabasal cell layers

Treatment groupsPercentage of HEFs with reduced K1/K10 expression

Epidermal cell layers of HEFs Stages K1 K10

Carcinogen controlBasal cell layer

Stages 1, 2 and 3 HEFs 0 0Suprabasal cell layers

Stage 1 HEFs 0 0Stage 2 HEFs 0 0Stage 3 HEFs 11.1 100.0

RA-carcinogenBasal cell layer

Stages 1, 2 and 3 HEFs 0 0Suprabasal cell layers

Stage 1 HEFs 0 0Stage 2 HEFs 7.0 15.4Stage 3 HEFs 10.0 87.5

Stage 1, HEFs with three to five suprabasal cell layers. Stage 2, HEFs with six to eight suprabasal cell layers. Stage 3, HEFs with > 8 suprabasal cell layers. No significant differences were found between carcinogen control and RA-carcinogen groups.

Figure 5. Mouse dorsal skin from the carcinogen control group.Stage 2 hyperplastic epidermal foci (HEFs). Preserved K1 expressionin all suprabasal cell layers in stage 2 HEFs. Avidin–biotin–peroxidasecomplex; Harris haematoxylin.

Figure 6. Mouse dorsal skin from the retinoic acid-carcinogen group.Stage 2 hyperplastic epidermal foci. Reduced K1 expression in suprabasalcell layers. Avidin–biotin–peroxidase complex; Harris haematoxylin.

292 © 2007 The Authors. Journal compilation © 2007 ESVD and ACVD.

García-Fernández et al.

in stage 2 HEFs. Comparable changes in keratinocytematuration, with a decrease in K1 and/or K10 related toRA action, have been demonstrated previously both invitro26–30 and in vivo after topical RA treatment of adultepidermis.17,30 Thus, the stimulatory action on cell prolif-eration and the inhibitory effect on cell differentiation ofHEFs attributed to prenatal RA exposure support a long-termeffect of this retinoid.

On the other hand, this study seemed to support atemporal change in the development of DMBA-TPA-inducedfocal hyperplastic lesions as has been cited in literature.14

This was suggested by a progressive increase in thenumber of epidermal cell layers in parallel with theenhancement of K6 immunostaining and loss of K1/K10expression and the regular presence of HEFs showingmixed 2/1 or 3/2 stages.

The results also support the view that the epidermalhyperplastic lesions induced by the carcinogenesis protocolmay be considered as a marker of potential skin tumouri-genicity as has been previously expressed.14,31 Hence thealtered expression patterns in differentiation and proliferation-related keratins in some advanced stage HEFs may beassociated with a basic defect in the programme of cellgrowth and differentiation and could be early indicators ofneoplastic transformation.32,33 If skin hyperplastic lesionshave pre-neoplastic value, a decreased number of HEFsdeveloped in prenatal RA exposure mice compared withmice not exposed to this retinoid seemed to corroboratethe chemopreventive activity of retinoids on mouse skincancer described following their topical application7,34 ororal administration.7,12,35 This inhibitory action on thedevelopment of HEFs associated with in utero RA exposuremay be related to the ability of retinoids to inhibit skintumour promotion by TPA pointed out in postnatalexposure.36,37 However, the earlier loss of cell differentiationin HEFs associated with prenatal RA exposure raisesdoubts on a complete chemopreventive effect of thisretinoid on skin carcinogenesis, and suggests that quanti-fication of lesions induced by a carcinogenesis regimealone is insufficient for assessment of skin cancer prevention,and that other parameters should also be evaluated.

Acknowledgements

The authors thank Dr Juan García Vieitez of the Depart-ment of Pharmacology and Toxicology (University of León,Spain) for advice on RA dosage.

Source of Support: Grants from the Junta de Castilla yLeón (España) (LE04/98).

References

1. Reynolds NJ, Fisher GJ, Griffiths EM et al. Retinoic acidmetabolites exhibit biological activity in human keratinocytes,mouse melanoma cells and hairless mouse skin in vivo. Journalof Pharmacology and Experimental Therapeutics 1993; 266:1636–42.

2. Craven NM, Griffiths CEM. Topical retinoids and cutaneousbiology. Clinical and Experimental Dermatology 1996; 21: 1–10.

3. García-Fernández RA, Pérez-Martínez C, Espinosa-Alvarez J et al.Mouse epidermal development: effects of retinoic acid exposurein utero. Veterinary Dermatology 2006; 17: 36–44.

4. García-Fernández RA, Pérez-Martínez C, Escudero-Díez A et al.Effects of utero retinoic acid exposure on mouse pelage hairfollicle development. Veterinary Dermatology 2002; 13: 157–63.

5. Sitzmann JH, Bauer FW, Cunliff WJ et al. In situ hybridizationanalysis of CRABPII expression in sebaceous follicles from 13-cisretinoic acid-treated acne patients. British Journal of Dermatology1995; 133: 241–8.

6. Peck GL, Olsen TG, Yoder FW. Prolonged remissions of cystic andconglobate acne with 13-cis retinoid acid. New England Journal ofMedicine 1979; 300: 329–33.

7. Verma AK. Retinoids in chemoprevention of cancer. Journal ofBiology Regulators and Homeostatic Agents 2003; 17: 92–7.

8. Hendrickx AG, Hummler H. Teratogenicity of all-trans-retinoic acidduring early embryonic development in the cynomolgus monkey(Macaca fascicularis). Teratology 1992; 45: 65–74.

9. Kochhar DM, Jiang H, Penner JD et al. Differential teratogenicresponse of mouse embryos to receptor selective analogs ofretinoic acid. Chemico-Biological Interactions 1996; 100: 1–12.

10. Rosa FW. Teratogenicity of isotretionin. Lancet 1983; 2: 513.11. Gensler HL, Sim DA, Bowden GT. Influence of the duration of

topical 13-cis-retinoic acid treatment on inhibition of mouse skintumor promotion. Cancer Research 1986; 46: 2767–70.

12. Verma AK, Dubik L, Ali M. Modulation of mouse skin tumorpromotion by dietary 13-cis-retinoic and alpha-di-fluoromethylor-nithine. Carcinogenesis 1986; 7: 1019–23.

13. De Luca LM, Sly L, Jones CS et al. Effects of dietary retinoic acidon skin papilloma and carcinoma formation in female SENCARmice. Carcinogenesis 1993; 14: 539–42.

14. Binder RL, Johnson GR, Gallagher PM et al. Squamous cellhyperplastic foci: precursors of cutaneous papillomas induced inSENCAR mice by a two-stage carcinogenesis regimen. CancerResearch 1998; 58: 4314–23.

15. Nelson KG, Slaga TJ. Keratin modifications in epidermis, papillomasand carcinomas during two-stage carcinogenesis in the SENCARmouse. Cancer Research 1982; 42: 4176–81.

16. Roop DR, Krieg TM, Mehrel T et al. Transcriptional control of highmolecular weight keratin gene expression in multistage skincarcinogesis. Cancer Research 1988; 48: 3245–52.

17. Eichner R, Kahn M, Capetola RJ et al. Effects of topical retinoidson cytoskeletal proteins: implications for retinoid effects onepidermal differentiation. Journal of Investigative Dermatology1992; 98 (2): 154–61.

18. Weiss RA, Eichner R, Sun TT. Monoclonal antibody analysis ofkeratin expression in epidermal diseases: a 48- and 56-kdaltonkeratin as molecular markers for hyperproliferative keratinocytes.Journal of Cell Biology 1984; 98: 1397–406.

19. Kopan R, Fuchs E. A new look into and old problem: keratins astools to investigate determination, morphogenesis, and differen-tiation in skin. Genes and Development 1989; 3: 1–15.

20. Council Directive 86/609 EEC (1986): Council Directive on theApproximation of Laws, Regulation and Administrative Previsionsof the Member States Regarding the Protection of Animals usedfor Experimental and other Scientific Purposes (86/609/EEC).

21. Auld J. Antigen unmasking in routinely processed paraffin sectionsby pressure cooking. United Kingdom National External QualityAssessment Schemes-ICC 1994; 17: 197–209.

22. Martín Andrés A, Luna del Castillo JD. Bioestadística Para LasCiencias de la Salud, 4th edn. Madrid, Spain: Norma SA, 1994:383–450.

23. Fuchs E, Byrne C. The epidermis: rising to the surface. CurrentOpinion in Genetics and Developmen 1994; 4: 725–36.

24. Huitfedt HS, Heyden A, Clausen OP et al. Altered regulation ofgrowth and expression of differentiation-associated keratins inbenign mouse skin tumors. Carcinogenesis 1991; 12: 2063–7.

25. Kopan R, Fuchs E. The use of retinoic acid to probe the relationbetween hyperproliferation-associated keratins and cell proliferationin normal and malignant epidermal cell. Journal of Cell Biology1989; 109: 295–307.

26. Ashton RE, Connor MJ, Lowe NJ. Histologic changes in the skinof the Rhino mouse (hrrh hrrh) induced by retinoids. Journal ofInvestigative Dermatology 1984; 82: 632–5.

© 2007 The Authors. Journal compilation © 2007 ESVD and ACVD. 293

Long-term RA action on chemically induced epidermal hyperplasia

27. Hieber AD, King TJ, Morioka S et al. Comparative effects ofall-trans beta-carotene versus 9-cis beta-carotene on carcinogen-induced neoplastic transformation and connexin 43 expression inmurine 10T1/2 cells and on the differentiation of human keratino-cytes. Nutrition and Cancer – An International Journal of 2000; 37:234–44.

28. Zou CP, Hong WK, Lotan R. Expression of retinoic acid receptor betais associated with inhibition of keratinization in human head and necksquamous carcinoma cells. Differentiation 1999; 64: 123–32.

29. Kunchala SR, Suzuki T, Murayama A. Photoisomerization ofretinoic acid and its influence on regulation of human keratinocytegrowth and differentiation. Indian Journal of Biochemica andBiophysica 2000; 37: 71–6.

30. Lichti U, Yuspa SH. Modulation of tissue and epidermal transglutam-inases in mouse epidermal cells after treatment with 12-O-tetradecanoylphorbol-13-acetate and/or retinoic acid in vivo andin culture. Cancer Research 1988; 48: 74–81.

31. Sisskin EE, Gray T, Barret JC. Correlation between sensitivity totumor promotion and sustained epidermal hyperplasia of miceand rats treated with 12-O-tetradecanoylphorbol-13-acetate.Carcinogenesis 1982; 3: 403–7.

32. Perkins W, Campbell I, Leigh IM et al. Keratin expression innormal skin and epidermal neoplasms demonstrated by a panel ofmonoclonal antibodies. Journal of Cutaneous Pathology 1992; 19:476–82.

33. Markey AC, Leigh IM, Lane B et al. Changes in keratin phenotypeassociated with malignant transformation of keratinocytes.Abstract British of Journal of Dermatology 1990; 123 (Suppl. 121).

34. Moon RC MR, Rao KJ VN. Retinoids and human cancer in exper-imental animals. In: Sporn MB, Roberts AB, Goodman DS, eds.The Retinoids. New York: Raven Press, 1994: 597–658.

35. Chen LC, Sly L, De Luca LM. High dietary retinoic acid preventsmalignant conversion of skin papillomas induced by a two-stagecarcinogenesis protocol in female SENCAR mice. Carcinogenesis1994; 15: 2383–6.

36. Verma AK, Shapas BG, Rice HM et al. Correlation of the inhibitionby retinoids of tumor promoter-induced mouse epidermalornithine decarboxylase activity and of skin tumor promotion.Cancer Research 1979; 39: 419–25.

37. Verma AK, Slaga TG, Wertz PW et al. Inhibition of skin tumorpromotion by retinoic acid and its metabolite 5,6-epoxyretinoicacid. Cancer Research 1980; 40: 2367–71.

Résumé Des souris Adult Naval Medical Research Institute (NMRI), après exposition prénatale à l’aciderétinoique (RA), ont été traitées avec une modalité standard en deux phases de carcinogénèse cutanée, pourcaractériser les foyers épidermiques hyperplasiques (HEFs), qui précèdent l’apparition des papillomescutanés, et pour étudier l’action à long terme in vivo du RA sur la peau de souris adulte traitée au DMBA(7,12 dimethyl benz[a]anthracene) et au TPA (12-O-tetradecanoylphorbol 13-acetate). Les résultats montrentque le RA administré aux souris gestantes a une action inhibitrice à long terme sur la différencitation cellulaireet sur le développement des lésions hyperplasiques apparaissant avant un cancer sur la peau adulte de leurprogéniture, et également un effet stimulant de la prolifération cellulaire sur ces lésions hyperplasiques.

Resumen Tras la exposición prenatal a ácido retinoico (RA), ratones adultos del Instituto Naval de InvestigaciónMédica fueron tratados con un régimen estándar de carcinogénesis de piel en dos fases, con el fin decaracterizar focos epidérmicos hiperplásicos que preceden a la aparición de papilomas cutáneos, y tambiénpara investigar el efecto in vivo a largo plazo del ácido retinoico en la piel de ratones adultos tratados conDMBA (7,12 dimetilbenzantraceno) y TPA (12-O-tetradecanoilforbol 13-acetato). Los resultados demuestranque el ácido retinoico administrado en ratonas gestantes inhibe a largo plazo la diferenciación celular y eldesarrollo de lesiones hiperplásicas que se producen previamente a la aparición de cáncer en la piel deanimales adultos procedentes de sus camadas, así como estimula la proliferación celular de estas lesioneshiperplásicas.

Zusammenfassung Adulte Mäuse des medizinischen Forschungsinstituts der Marine (NMRI) wurden,nachdem sie Retinolsäure pränatal ausgesetzt worden waren, mit einem standardisierten zweistufigen HautKanzerogenese Regime behandelt, um hyperplastische epidermale Fokusse (HEFs) zu charakterisieren,welche dem Auftreten von kutanen Papillomen vorausgehen, und um die Langzeitwirkung der Retinolsäureauf adulte Mäusehaut, die mit DMBA (7,12 Dimethyl benzanthracen) und TPA (12-O-tetradecanoylphorbol13-acetat) behandelt worden war, in vivo zu untersuchen. Die Ergebnisse zeigen, dass Retinolsäure, dieträchtigen Mäusen verabreicht wurde, eine langfristige Inhibitionswirkung auf die Zelldifferenzierung unddie Entwicklung von hyperplastischen Läsionen hatte, welche vor dem Karzinom auf der adulten Haut ihrerNachkommenschaft entstehen. Es zeigte sich auch eine anregende Wirkung auf die Zellproliferation dieserhyperplastischen Läsionen.