Reconstitution of human skin in vitro using readily available sources: Novel studies of tumour promotion

Toxic. in Vitro Vol. 8, No. 4, pp. 655~57, 1994

~

Pergamon

0887-2333(94)E0060-7

Copyright © 1994 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0887-2333/94 $7.00 + 0.00

RECONSTITUTION OF H U M A N SKIN IN VITRO USING READILY AVAILABLE SOURCES: NOVEL STUDIES OF TUMOUR PROMOTION A. VERMA and A. J. SHAW Department of Toxicology, The School of Pharmacy, University of London, 29-39 Brunswick Square, London WCIN lAX, UK Abstract--As part of our development of an assay for skin tumour promoters we have standardized a three-dimensional model of human skin. The important features of this model for our purposes are (a) that it is characteristic of human skin, (b) that the developing epidermis can be seeded with "initiated" keratinocytes, and (c) that the calcium level of the medium can be controlled. The procedure used is an amalgamation of the methodologies of Asselineau and Prunieras (1984), Lenoir et al. (1988) and Rosdy and Clauss (1990). This model does not rely on a source of human skin and can be reproduced in any laboratory equipped with basic cell culture apparatus. The dermal equivalent consists of human embryonic fibroblasts and rat tail collagen. The overlying keratinocytes are derived from the outer root sheath of human hair follicles inserted into the collagen-fibroblast gel. Stratification is induced by culturing the epidermal cells at the air interface of a serum-free medium. The model has been characterized by histological staining, electron microscopy and immunostaining.

Introduction The reconstruction of human skin in culture is now becoming widespread with numerous methodologies available (Skin Pharmacology, 1990). However, the commonest hurdle when setting up these models is still the need for a source of human skin. To circumvent this problem we have amalgamated two well studied methods for reconstituting skin (Asselineau and Prunieras, 1984; Lenoir et al., 1988), and have produced a simple skin model that uses readily available human cell sources, namely hair follicle keratinocytes t~d GM 10 embryonic fibroblasts. This stratified model has been characterized by us, using histological staining and indirect immunofluorescence, and displays an intact stratum corneum and the correct suprabasal location of keratins KI/K10. This culture system could be reproduced in any standard cell culture laboratory and its relative inexpense and adaptability in comparison with commercially available models make it particularly attractive for toxicological and pharmacological studies.

An assay for tumourpromoters We are proposing to use this model as a tool for assaying and studying human skin tumour promoters and antipromoters. Classicaly in initiationpromotion protocols, skin tumours do not appear in mice treated with a small topical dose of a chemical carcinogen, for example dimethylbenz[a]anthracene, unless the treated skin is subsequently exposed to a Abbreviation: TPA = 4-fl-phorbol-12-myristate-13-acetate.

series of doses of a tumour promoter, for example 4-/~-phorbol-12-myristate-13-acetate (TPA). No reduction in tumour yield is seen if the onset of promotion is delayed for several months after the initiating event. Thus, initiated cells remain dormant in the skin until they are induced to expand clonally by the application of the promoting agent. This clonal expansion is a critical event since it greatly increases the probability that an initiated cell will undergo secondary genetic changes to produce a cancer cell (Yuspa et al., 1991). The failure of initiated cells to expand clonally in the absence of tumour promoters results from the growth restriction enforced by the surrounding normal keratinocytes (Hennings et al., 1991). Promoters overcome this inhibition by favouring the growth of initiated cells over normal cells by (a) selectively inducing normal cells to differentiate terminally, (b) being more cytotoxic to the normal cells, and/or (c) being more mitogenic to the initiated cells. Because any chemical-induced growth advantage bestowed on the initiated cells should become apparent in a co-culture alongside normal keratinocytes, we are proposing this as the basis for an assay for tumour promoters.

Cell lines We are currently investigating three cell lines (HK2, HK-3 and HK-4) to see whether they are suitable for inclusion as initiated cells in our co-culture assay for human skin tumour promoters. These cell lines are all human keratinocytes that have been immortalized by transfection with a mutant subgenomic SV40

655

656

A.

Table 1. Effect of TPA on the

VERMA and

plating efficiencyof four human

cell

lines Plating Cell

line

HK-2 HK-3 HK-4 A43 I

efficiency(% i. SD; 5 no

Control 19.6 16.3 21.8 17.7

f t t +

TPA

10.4 I3.3 16.0 rt 4.6 14.5 k 2.3 ND

3.0 1.3 2.3 3.7

n = 3-5) 20 “M TPA 8.1 * 16.5i 13.3 i I .2 *

2.8 1.9 2.8 0.8

gene using an amphotropic retroviral vector (M. O’Hare, Institute of Cancer Research). They are all cultured in serum-containing keratinocyte growth medium as described by Rheinwald (1989) with the omission of cholera toxin, which retards their growth. HK-2 and HK-4 cells resemble normal keratinocytes in culture whereas HK-3 cells are fibroblastic in appearance. The plating efficiency of all three cell lines in the presence of a 3T3 feeder layer (Table 1) was significantly greater than for primary adult keratinocytes (l-4%) and comparable with that of the human squamous carcinoma cell line, A431 (Table I). Plating efficiencies for HK-2, HK-3 and HK-4 cells were dramatically reduced in the absence of a 3T3 feeder layer, being 4.6% +O.l. 3.8% +0.6 and 4.6% + 0.5, respectively. The capacity of these three cell lines to differentiate normally was assessed by culturing them on collagen-fibroblast lattices at an air-liquid interface. Normal keratinocytes are induced to differentiate terminally under such conditions and form a multilayered ‘tissue’ that expressed various differentiation markers (including Keratin KIO) and has a well formed cornified layer. The HK-4 cells produced a continuous multilayered culture of 10-i 5 cells in depth. The culture was reminiscent of psoriatic skin in that cell nuclei could be seen in the upper epidermal layers and the cornified layer was poorly formed. The HK-2 cells also produced a continuous multilayered culture but of only 2-5 cells in depth. However, the cornified layer was more established than with the HK-4 cell cultures. The HK-3 cells produced a very erratic multilayer and the cells had clearly migrated through the collagen lattice. No cornified layer was seen. All three cell lines immunolabelled positively for keratin RIO in these air-liquid interface cultures. To test whether these immortalized keratinocytes were more resistant to the effects of TPA than normal keratinocytes, we dete~ined their plating efhciency in the presence of a 3T3 feeder layer and two concentrations of TPA (5 and 20 no) (Table I). TPA Table 2. Effect of TPA on the cornitied envelope formation of HK cells

envelope formation 1% of total cells)

Ccroified

Cell line HK-2 HK-3 HK-4

Control

I “M TPA

5 “M I-PA

20 “M TPA

I.1 0.3 0.5

1.6 0.3 2.3

0.8 0.6 2.6

0.9 1.0 I.8

A. J. SHAW

was added 24 hr after seeding the cells and the concentration was kept constant throughout the following I3 days. The plating efficiency of the HK-3 cell line was unaffected by the presence of 20 nM TPA whereas plating efficiencies of the HK-2 and HK-4 cell lines were reduced to 41.3% and 47.8% of control values, respectively. Parkinson et al. (1983) reported that the plating efficiency of normal human keratinocytes is reduced to 20% of control values after a 24-hr exposure to IO nM TPA. Thus, all the HK cell lines are more resistant than normal keratinocytes to the growth-inhibiting effects of TPA. To ascertain whether this was due to a reduced responsiveness of these cell lines to the terminal differentiation stimulus of TPA, we determined the extent of cornified envelope formation in these cell lines with and without TPA (Table 2). All three cell lines displayed the potential to form cornified envelopes but at a lower ratio than normal human keratinocytes (2.43.7%; Parkinson et al., 1993). Only cell line HK-4 was induced to differentiate terminally by TPA and only by a four--five-fold increase, which is substantially less than the response of normal keratinocytes (I 5-30-fold; Parkinson et nl., 1993). These results indicate that HK-3 cells, and to a lesser extent HK-2 and HK-4 cells, should have a growth advantage over normal keratinocytes when exposed to TPA. This advantage is due to these cells displaying a reduced responsiveness to the terminal differentiation stimulus of TPA. Because the HK-2 and HK-4 cell lines behaved fairly similarly to normal keratinocytes in both submerged and air-liquid cultures, we are continuing to study their potential as ‘initiated’ cells in our tumour promotion model. We are currently assessing their ability to expand clonafly when seeded amongst a confluent layer of normal human keratinocytes in the presence and absence of TPA. Acknowledgments-A. Verma is supported by an SERCCASE studentship with Fisons Pha~aceuti~ls. We are also grateful to Shell Research UK. the Nuffield Foundation, Ford Motors and E. Merck for their financial contributions, and Mike O’Hare (Institute of Cancer Research) for the loan of the HK cell series.

REFERENCES

Asseiineau D. and Prunieras M. (1984) Reconstruction of “simplified” skin: control of fabrication. Brirish Journal cf Dermatology 111 (Suppl. 27), 219-222. Skin Pharmaeo~ogy (1990) Reconstructed human skin. Skin Pharmacology 3, 6 1- 148. Hennings H., Lowry D. T. and Robinson V. A. (1991)

Coculture of neoplastic and normal keratinocytes as a modei to study tumour promotion. Skin Pharmacology 4 (Suppl. I), 79-84. Lenoir M-C., Bernard B. A., Pautrat G., Darmon M. and Shroot B. (1988) Outer root sheath cells of human hair follicle are able to regenerate a fully differentiated epidermis in oitro. De~~e~opment~fBiology 130, 610-620. Parkinson E. K., Grabham P. and Emmerson A. (1983) A

Reconstituted skin to study tumour promotion subpopulation of cultured human keratinocytes which is resistant to the introduction of terminal differentiation-related changes by phorbol-12-myristate-13-acetate: evidence for an increase in the resistant population following transformation. Carcinogenesis 4, 857-861. Rheinwald J. G. (1989) Methods for clonal growth and serial cultivation of normal human epidermal keratinocytes and mesothelial cells. In Cell Growth and Division--

657

a Practical Approach. Edited by R. Baserga. pp. 81-94. IRL Press, Oxford. Yuspa S. H., Kilkenny A., Cheng C., Roop D., Hennings H., Kruszekski F., Lee E., Strickland J. and Greenhalgh D. A. (1991) Alterations in epidermal biochemistry as a consequence of stage-specific genetic changes in skin carcinogenesis. Environmental Health Perspectives 93, 3 10.