TIPS - February 1983
metabolism might be subject to large inter- with glyceryl trinitrate, on dihydro- availability will he greatly increased, drug individual variation, high doses of di- ergotamine absorption through the liver. The extraction may even approach zero, and hydroergot alkaloids would be injudicious area under the serum concentration-time systemic clearance will he lowered. without more knowledge of the phar- curve increased by 56% to 370%. This was BARRY G. WOODCOCK and macokinetics and concentrations attained in not caused by a change in absorption but NORBERT RIETBROCK patients. was due to a more rapid passage of drug Uncertainty in the selection of a suitable across the liver and therefore to a lower first Department of Clinical Pharmacology, Umversity dosage may have been the reason for the pass metabolism. The clinical consequence Clb~, Theodor-Stern-Kai 7, D-60~ Frankfurt elaborate dosage regimes sometimes of these altered pharmacokinetics was a rise F.R.G. recommended. One of these, as an illustra- of 30 mmHg in the systolic blood pressure. tion, involves administration of dihydro- In the absence of glyceryl trinitrate, the Reading list ergotoxinc as an oral solution where the initial change in systolic blood pressure was no 1 van den Berg, E., Walterbusch, G , Gotten, L., Rumpf, K. D , Otten, B and Frohlich, H (1982) dose is 10 drops and subsequent doses are different from that produced by placebo. Dtsch. Med. Wochenschr. 107,716-718 These results show that first pass extrac- 2 Bobik, A , Jennmgs, G. L , Skews, H., Esler, M. increased at the rate of 2 drops daily over 20 days. At the end of the treatment period the tion by the liver is a prime determinant of and McLean, A (1981) CIm. Pharmacol. Ther. dose is reduced by 2 drops daily in a corres- dihydroergot alkaloid bioavailability which 30, 673-679 ponding manner. Schemes like this are dif- can vary with changes in the physiological 3 Little, P. J., Jennmgs, G L., Skews, H. and Bobik, A. (1982)Br. J. Chn. Pharmacol. 13,785--790 ficult to rationalize and have been claimed conditions. Alterations in gastrointestinal blood flow and changes in liver anatomy 4 Woodcock, B. G., Loh, W., Habcdank, W D and to turn patients into hypochondriacs. Raetl:nxx:k, N. (1982) Clin. PharmacoL Ther. 32, The bioavailability study with glyceryl through disease will change drug absorp622 trinitrate and dihydroergotamine, referred tion and elimination. Patients with liver cir- 5 Woodcock, B. G. and Rietbrock, N. (1982)Dtsch. to earlier, is notable in that it includes an rhosis and hepatic shunts for example, will Med. Wochenschr. 107,996 examination of the effects of an increase in require much lower oral and intravenous 6 Rossignol, R and Eblgwel-lbru, M. (1980) Trends m Pharmacol. Sci. I, 287-289 liver blood flow, produced by pretreatment doses since dihydroergot alkaloid bio-
Current techniques T e c h n i q u e s o f interest to p h a r m a c o l o g i s t s greater than that of other tissues.
Human hair follicles in biochemical pharmacology and toxicology
Hair follicles and glucose-6-phosphate dehydrogenase deficiency
The usefulness o f the human hair follicle as a biopsy tissue for determining genotypes in inborn errors o f metabolism is due to two major facts: first, the human hair follicle can easily be removed from the donor without any harm or risk; second, due to the fact that most hair follicles are o f monoclonal origin they are highly suitable for determining X-linked diseases. In this review we show that the hair follicle may be o f importance for pharmacology and toxicology, both in rural areas o f developing countries and, as a result o f our recent work on culturing the tissue, as a tool in the struggle against diseases o f modern society. The use of human hair follicles as a biopsy material in studies in biochemical pharmacology and toxicology is of a recent date. Initially, hair follicles were employed for diagnostic purposes. Later on, it was found that they could also be used as a model system for more fundamental research. The purpose of this paper is to review work done with hair follicles, especially in the various areas of research in which our group is involved. No studies using hair follicles can he carried out without consideration of the hair growth cycle. The growth of a hair follicle starts from one, or a few, cells near the dermal papilla. They are therefore said to be of monoclonal origin. In the hair growth
cycle 3 main phases can he distinguished: (I) anagen, which is divided into follicle regeneration and an active hair growth stage; (2) catagen, the retrogressive stage; and (3) telogen, the inactive, resting period. For most studies, only hair follicles in the anagen phase of the growth cycle are used. A schematic representation of a hair follicle in the skin, showing the different substructures, is given in Fig. 1. The hair is produced by a relatively small mass of cells, the 'matrix', where cell division is continuously taking place. The other parts of the hair, except for the outer root sheath, develop from the matrix cells by differentiation processes. The synthetic capacity of the matrix cells is an order of magnitude
Glucose-6-phosphate dehydrogenase deficiency is the most common enzyme abnormality in man. It varies greatly in its prevalence throughout the world and the areas with highest incidence of this disease are the Mediterranean, South East Asia and parts of Africa. The world distribution closely follows that of malaria falciparum. Evidence suggests that the high frequency of the defective glucose-6-phosphate dehydrogenase gene in certain parts of the world is due to an increased resistance of heterozygotes to malaria. Glucose-6-phosphate dehydrogenase, the key enzyme in the hexose monophosphate shunt, catalyzes the oxidation of glucose-6-phosphate to 6-phosphogluconate (see reaction sequence). 6.Pbospho. ghiconate is oxidized by the enzyme 6-phosphogluconate dehydrogenase to ribulosc-5-phosphate and CO2. G6P + NADP ~ ---*6PG + NADPH (1) 6PG + NADP~ -* ribulose-5-phosphate + CO2 + NADPH (2) During these reactions NADPH is formed. This cofactor is necessary for the conver-
ElsevierBiomedicalPress 1983 0165- 6147/83/0000- 0000~01 00
T I P S - February 1983
56 henle's layer I cuhcle of inner sheath
radtxpth transientI /1~.~[ ,
zone ..... ]cellgrowth end
Fig. 1. A. Dtagrammatic representatton of the human hawfollicle in the skin. B Schemattc drawing showing various substructures of the human hairfollicle.
sion of methemoglobin into hemoglobin in the erythrocyte and for the formation o f reduced glutathione. Reduced glutathione is essential for the maintenance of cellular integrity, since it protects the SH-groups of membrane proteins, hemoglobin and e n z y m e s against oxidation. A deficiency of glucose-6-phosphate dehydrogenase m a y lead to acute hemolytic anemia upon ingestion of certain drugs and foods (Table I). The agents inducing hemolysis in such patients have been found to be predominantly drugs that either produce peroxides in vivo or act themselves as direct oxidants; they oxidize the G S H and " N A D P H present in the erythrocytes, resulting in oxidant damage to vital components of the cell and thereby cause an intravascular hemolysisk The elevated level of bilirabin after hemolysis can, in newborn h u m a n s , result in the development of severe neonatal jaundice and kemicterus. It is estimated that up to 1964 about 43% of all cases of kernicterus in Singapore were due to glucose-6phosphate dehydrogenase deficiency. In that year a screening program for this disease was started, resulting in an elimination of kernicterus due to glucose-6-phosphate dehydrogenase deficiency 2. The pattern of inheritance of this trait is that of a recessive X-linked defect. Since males carry the defect on their single X - c h r o m o s o m e (they are hemizygous), the disease is expressed phenotypically and they form the group sensitive to druginduced hemolysis. In female cells one of the two X - c h r o m o s o m e s is not active. This X-chromosome inactivation occurs early in embryonic life (and is transmitted to the daughter cells) and it is a s s u m e d that inactivation o f the paternal or maternal X-chromosome is at random (Lyon hypothesis). Female heterozygotes have intermediate values of glucose-6-phosphate
dehydrogenase activity because the normal gene on one X - c h r o m o s o m e and the defective allele on the other X-chromosome are both expressed statistically in a 50-50 distribution. Obviously, heterozygous females will pass the defect to half their sons. Affected males have normal sons because they pass only their Y-chromosome to them and they have heterozygous daughters because they pass the defective X-chromosome to all of them. It is o f great importance to detect heterozygous females (caniers) in order to protect their newborn children from kernicterus. The most convenient way to perform carrier detection of glucose-6-phosphate dehydrogenase deficiency is to use hair follicles. These have the advantage that most of t h e m are of monoclonal origin and
TABLE Ia. Drugs reported to reduce hemolysism subjects with glucose-6-phosphatedehydrogenasedeficiency Drug Drugs producing climcally significanthemolysis Acetanthde Dapsone Furazolidone Furaltadone Nitrofural Nitrofurantoln Sulfandamlde Sulfapyridme Sulfacetamide Salazosulfapyndme Sulfamethoxypyrldazme Thiazosulfone Qumidme Prnnaqume Pamaquine Pentaquine Quinocide Naphthaline Neoarsphenarmne Phenylhydrazine Toloniumchloride Trimlrotoluene
+++ ++ ++ ++ ++++ ++ +++ +++ ++ +++ ++ ++
+++ ++++ +++ + ++ ÷÷+ ++ +++ ++++
TABLE lb. Drugs reported to reduce hemolysts in subjects with glucose-6-phosphatedehydrogenasedeficiency Drug
Drugs reported as hemolyuc agents in some cases, but usually not producing clinicallysignificanthemolysis under normal conthtions (e g., in the absence of infectmn) Phenacetm Acetylsahcyhcacid Sulfadiazme Sulfafurazole AldesulfoneSodmm Chloramphemcol Nitrite Methyleneblue Ascorbic aod Dtrnercaprol Chloroqume Mepacrme
+ ± ++ + o, + + + + + -+ ±
+ ++ ++ +++
TIPS - February 1983
H e t e r o z y g o t e O0~O O000 (carrier)
Fig. 2. Heterozygote detection in glucose-6-phosphate dehydrogenase deficiency. Glucose-6-phosphate dehydrogenasel6-phosphogluconate dehydrogenase activtty ranos of single hair follicles from a healthy control, a patient with glucose-6-phosphate dehydrogenase deficiency and a heterozygote. Each dot represents an individual hair follicle.
as a consequence express only one X-chromosome. Therefore heterozygotes will have hairs both with and without ghicose-6-phosphate dehydrogenase activity; there will be no problem of overlap of glucose-6-phosphate dehydrogenase activity values between patients, carriers and normal individuals, as may occur with erythrocyte activity values. The assay for glucose-6-phosphate dehydrogenase is based on the measurement of NADPH formation. To correct for differences in hair follicle sizes the reference enzyme, 6-phosphoglnconate dehydrogenase, is measured. This has a similar distribution in the hair follicle to glucose6-phosphate dehydrogenase as determined by histochemical techniques3. The ratio of both enzyme activities in 25 individual hair follicles has been used to distinguish between patients, carders and normal individuals in Fig. 2. Upon storage, a gradual drop in the activity of both enzymes in hair follicles was found, but the rate of decrease in enzyme activity was about equal and the enzyme activity ratio was therefore not affected. Even after 10 days of storage a reliable enzyme activity ratio could be obtaineds. Therefore, hair follicles seem to be an attractive biopsy material for screening purposes: they can be collected and mailed without any precautions to central laboratories where the enzymes can be assayed. The hair follicle enzyme assay for detection of carriers may be of especial importance in developing countries where glucose-6-phosphate dehydrogenase deficiency is endemic. In many of these countries it is not possible to detect the disease
early after birth, since most babies are horn in rural areas where medical facilities are scarce. For these areas our method using hair follicles would be particularly suitable; hair follicles of pregnant women could be simply sent by mail to specialized centres so that the costs would remain relatively low. Those women with a defective X-chromosome could then give birth to their child under special medical attention so that kemicterus could be avoided. In a recent study we have shown that storage of hair follicles under conditions that are often encountered in tropical countries such as elevated temperatures and high humidity does not influence enzyme activities. Moreover, hair colour and greying of hair, irrespective of the age of the donor, are irrelevant to the assay t. Therefore we may conclude that hair follicles may be used for easy population screening for heterozygotes of ghicose-6-phosphate dehydrogenase deficiency. Hair follicles and the diseases of modern
society In the previous section we have shown how hair follicles might be useful in the
struggle against a disease occurring predominandy in developing countries. However, we also hope to use the human hair follicle in a comparable struggle against diseases of western society. In Table H it is shown that in 1978 in The Netherlands five classes of disease accounted for the deaths of most people. About 25% of people died of cancer and almost 50% of cardiovascular disease. We hope to use the human hair follicle in predicting predisposition for one type of cancer (bronchogenic carcinoma) and for atherosclerosis. It has been known for a long time now that the occurrence of both these diseases is strongly influenced by cigarette smoking5'6. In Table III the death rate from lung cancer over the last 30 years is shown. It is evident that the mortality resulting from this disease has tripled in almost all European countries. Studying the consumption of cigarettes over the last 50 years in the same European countries demonstrates that the increase in the consumption of cigarettes has preceded the increase in the numher of casualties of this disease. When the number of cigarettes consumed by a large group of smokers is studied, it appears that there is an almost linear relationship between the number of cigarettes smoked and the risk of dying from the lung cancer, Kreyburg type I (the type of cancer originating from the lung epithelium). For lung cancer of Kreyburg type II, originating from lung fibroblasts, there appears to be no such correlation5. However, for cancer of the larynx and the oral cavity a comparable relationship is found. The most important result for heavy smokers is that stopping this dangerous habit will, overa period of 15 years, reduce the risk of smokingrelated cancer to that of a smoker who has not smoked more than five cigarettes a day. With respect to cardiovascular disease, comparable results have been found 6. For smokers younger than 35 years of age, there is an increased risk of cardiovascular disease, but the fatal number of deaths due to cardiovascular attack is low. However, over the age of 45 the chances of dying from cardiovascular attack increase dramatically with the number of cigarettes consumed daily. Smoking more than 20 cigarettes a day increases this risk seven-told in this age
TABLEII. Causesof death(pe~entagedisuibmion)m The Netherlands(1978) Cause of death
Infectiveand parasiticdiseases Neoplasms ofrespmatotysystem of bceast Dtseasesof the ctrculatorysystem Diseasesof the resptratorysystem Diseasesof the digesnvesystem Accidents,etc.
05 28 0 11.3 43.5 8.0 2.8 6.0
0.5 24 7 I3 5I 47.3 5.6 3.5 5.1
T I P S - F e b r u a r y 1983
TABLEIII a Lungcancer.Deathrate for malesper 100 000 population
Austrta Belgium Denmark Republicof Ireland France F.R G. U.K. Italy TheNetherlands Norway Spain
1955-9 1960-4 1965-9 1970--4 1975
53 32 20 15 17 26 57 12 28
65 44 31 27 24 40 74 19 39 12 13
73 60 44 35 31 52 89 28 53 17 18
77 76 55 44 38 60 102 45 67 23 24
76 93 66 53 44 64 111 51 81 29 28
75 69 49 67 113 88 34
TABLEIII b. Lungcancer Deathrate for femalesper 100 000 population
Ausma Belgium Denmark Repubhcof Ireland France F.R.G U.K, Italy TheNetherlands Norway Spam
1955-9 1960--4 1965-9 1970-4 1975
8.6 5.6 5.2 50 52 5.2 10 0 3.6 3.6
10.2 6.1 6.5 7.5 5.4 6.4 11.7 44 3.7 3.8 3.8
group. Stopping smoking is again enormously important. One year after cessation of smoking the chance of dying from a cardiovascular attack diminishes by 50% and after 4 to 5 years the chances are back to that of the non-smoker. The mutagens in cigarette smoke can form covalent bonds with DNA resulting in a blocking of normal regulatory processes in the cell and sometimes in malignant growth. Most of the chemicals in cigarette smoke are not mutagenic per se, but they may be converted into mutagens in the body. The major enzyme system involved in this activation of pre-mutagenic substances is the cytochrome P-450 complex. This multi-enzyme complex hydroxylates a variety of foreign compounds and drugs as well as steroids. It is characterized by the following properties: (i) it is membrane-bound; (ii) it requires NADPH for enzymatic activity; (iii) it is inhibited by carbon monoxide. An example of an enzyme belonging to this class is aryl hydrocarbon hydroxylase (AHH), which metabolizes polycyclic aromatic hydrocarbons such as benzo(a)pyrene. The latter substance is present in cigarette smoke, charcoalcooked foods, city smog, etc. Hydroxylation reactions form epoxides that are the actual mutagenic substances. They can, however, be detoxified in the body. The
10.5 6.9 9.0 84 6.1 8.0 14.5 5.8 4.4 4.0 4.5
11.8 78 11 6 11.4 6.5 8.5 19.0 6.6 4.8 5.1 5.4
12.7 8.9 15.4 16.3 6.3 9.0 24.8 7.6 5.7 6.6 5.7
13.7 18 0 6.6 10.1 27.0 6.3 7.9
epoxides may be spontaneously converted into phenols, they may react with water to form dihydrodiols or alternatively they may react both enzymatically or nonenzymatically with glutathione after which they are excreted as mercapturic acids in the urine. Upon exposure of an individual to carcinogenic substances the activity of the enzyme metabolizing these substances is induced. The induced enzyme is very similar to the uninduced enzyme, but it has slightly different spectral properties. The uninduced enzyme has an absorbance maximum at 450 nm (and is therefore called cytochrome P-450), whereas the induced enzyme has an optical maximum at 448 nm. For this reason it is designated cytochrome P-448. Both cytochromes can be distinguished from each other by the fact that they are sensitive to different inhibitors. Cytochrome P-450 is inhibited by metyrapone, whereas cytochrome P-448 is inhibited by a a-naphthoflavone. In mice the inducibility of AHH is genetically determined. For example, in the mouse strain C57BL/6N the AHH activity is inducible, whereas it is not inducible in the DBA/2 mouse. It has also been demonstrated that for several strains of mice the AHH inducibility is related to the carcinogenic index (Fig. 3). Although in mice a clear relationship between inducibility of AHH and the carcinogenic index has been established, a
comparable relation in humans has never been irrefutably demonstrated. A reason for this phenomenon might be that most of the attempts to determine such a correlation have used lymphocytes. However, smoking-dependentlung cancer is a disease of the epithelium and it has been argued that the metabolism of polycyclic aromatic hydrocarbons differs in different cell-types7 and therefore the lymphocyte may not be a representative tissue for studying the relationship between smoking and epithelial neoplasms such as lung cancer. It was therefore necessary to look for a biopsy material that is easily available for largescale population screening and that is of epithelial origin. We found the human hair follicle fulfilled these conditions. There were, however, several problems to be solved. For studies on induction of AHH activity the cells of the tissue have to be cultured. However, this had never been tried with hair follicles at the time that we started our investigations. Secondly, the presence of the enzyme AHH had never been demonstraled in human hair follicles. The presence of AHH and of epoxide hydratase (EH) has now been demonstrated in this tissues,9. With regard to the tissue culture problem, only the 'feeder-cell' technique, developed by Green 1° for the culturing of keratinocytes existed when we started our work. However, this technique was of no use to us, since it was impossible to determine the necessary biochemical parameters in a system where two different types of cells were present together. Green developed his feeder-cell technique because keratinocytes did not grow on artificial substrates and since these cells of the skin are present on the underlying dermis, he hypothesized that in vitro the epithelial cells might also need
JL • t 25
Fig. 3. Relationship between the carcinogenic index for subcutaneous methylcholanthrene and the genetically mediated reduction of aryl hydrocarbon hydroxylase activity by methylcholanthrene for each of 14 inbred strains of mice (fromRef. 17).
TIPS - February 1983
,some of the metabolic activities of the underlying fibroblasts. We chose a different approach: althdugh in the skin keratinocytes and fibroblasts are both present, they are separated from each other by the basal membrane and our hope was that the feeder-cell technique in fact worked so well because the keratinocytes and the fibroblasts brought together were able to synthesize important components of such a basal membrane. We therefore tried to find an alternative for a basal membrane. Such a membrane had to fulfil two basic criteria: firstly, it had to he the nannal growth substrate of an epithelium; secondly, it had to have excellent properties with respect to light transparency in order to allow microscopic examination of the cells. We found the capsule of the eye lens (Fig. 4) to be a suitable membrane. This tissue consists of an epithelium that is continuously differentiating into fiber cells and the whole lens is enclosed in a capsule predominantly consisting of collagen. It is evident that the lens capsule fulfils both our criteria. There was still one problem: the capsule had to be fixed in such a way that cells of hair follicles could be placed on them. For this purpose we developed a new culture dish called Epicult; this is now commercially available* and allows hair follicles to be cultured 11. It even allows sub-cultivation of the cells by transplantation of small parts of the culture to other Epicults t2. Using this system we have been able to show that carcinogen metabolism is inducible in cultured human hair follicle cells and in another study we also demonstrated that the basal levels of carcinogen metabolism are genetically determined 13'14. In the future we hope to verify whether genetic predisposition to develop lung cancer as a result of cigarette smoking can be demonstrated with the use of this epithelial tissue. With respect to a causal relationship between cigarette consumption and cardiovascular disease the situation is less clear. However, since we have demonstrated that AHH is also present in cultured and fresh endothelial cells (Hukkelhoven, M. W. A. C., Vermorken, A. J. M., Vromans, E. and Bloemendal, H. unpublished observations) we suggest that carcinogen metabolism may also play a role in the' development of atherosclerotic plaques. *Sanbio, Nistelrode, The Netherlands.
Fig. 4. Schematic drawing o f the lens
Partition of benzo(a)pyrene in vitro with different blood lipoproteins results in a binding of about 60% of this mutagenic substance to the low-density lipoproteins tS. This is further evidence in favour of our hypothesis, especially since it has been shown that there is a correlation between the level of low density lipoproteins and the risk of cardiovascular disease, also, lowdensity lipoproteins are specifically taken up by endothelial cells. Our hypothesis, therefore, is that low-density lipoproteins transport aromatic hydrocarbons from the lung, where they are taken up from cigarette smoke, to the endothelial cells. In these cells the hydrocarbons are metabolized into epoxides. These epoxides may either he detoxified or be covalently bound to cellular macromolecules, including DNA. The resulting mutation might then cause an atherosclerotic plaque. The fact that plaques arise from mutations has been postulated before by Benditt et al. 16 The culture model that we have developed for the human hair follicle allows determination of enzyme activities of an easily available human biopsy tissue of epithelial origin/n vitro. It may enable identification of high-risk groups for lung cancer and atherosclerosis and might therefore help to motivate these groups to change their attitude to smoking.
C. H. M. M. and Oei, T. L. (19"/9)J. C/in. Chem. Clin. Biochem. 17, 32.5-329 4 Vennorken, A. J. M , Markslag, A. M. G., Goos, C. M. A. A , M~guel, A., Ramon, M. and Petitpierre, E (1982) C/in. Genct. 22, 57--61 5 Vermorken, A. J. M. and Bloemendal, H. (1979) in Turnout Markers: Impact and Prospects (E. Boelsma, and Ph. Riimke, eds), 305-324, Elseviea-/North-Holland Biomedical Press, Amsle~lam, New York and Oxford 6 Ball, K. andTurner, R (1974)Lancet fi, 822-826 7 Arias, S. A., Boobis, A. R., Felton, J. S., TIm~eirs~n, S. A. and Nebert, D W. (1977) J. Biol. Chem. 252, 4712-4721 8 Vermo~en, A J. M., Coos, C. M. A. A., Henderson, P. Th. and Bloemendal, H. (1979) Toxicology 14, 109-116 9 Httkkelhoven, M. W. A C , Vromans, E. W. M., Vermorken, A. J. M and Bloemendal. H. (1982) FEBSLett. 144, 104-108 10 Rhemwald, J G and Green, H. (1975) Cell 6, 331-334 I 1 Weterings, P. J. J. M., Vermorken, A. J. M. and Bloen,,ent~. H. (1981) Br. J. Dermatol. 104, 1-5 12 Weterings, P, J J. M , Vermorken, A. J. M. and Bloernendal, H. (1982) Exp. Cell Res. 139, 439--443 13 Hukkelhoven, M W A C , Vromans, E. W. M., Vetmorken, A J. M., van Dtepen, C B and Bloemendal, H. (1982)Anal. Biochem. 125, 370-373 14 Huldmlhoven, M. W. A. C., Vetmotken, A. J. M., Vromans, E. and Bloemendai. H. (1982) C/in. Genet. 21, 53-58 15 Ketterer, B. (1980)Med. Bull. 36, 71-78 16 Bemhtt, E. P and Benditt, J. M. (1973) Proc. Natl. Acad. Sct. U.S.A. 70, 1753-1756 17 Thorgelssoo, S S. and Nebert, D. W. (1977) Adv. Cancer Res. 25, 149-193
A. J M VERMORKEN
Research Unit for Cellular Differentiation and Transformation, Untversity o f Ntjmegen, Geert Grooteplein Noord 21, 6525 E Z Niimegen, The Netherlands. Ions Vermorken obtained tus Ph.D. m 1977from the University o f Nilmegen, The Netherlands. He is head o f the Research Unu for Cellular DifferentiaReadinglist 1 Desforges, J. F. (1976) N. Engl. J. Med. 294, tion and Transformation at the same univers'ty. His malor research interests are the biochemtstry and 1438-1440 2 Wong Hock Boon (1975) J Singapore Paedzat. pharmacology o f human skin. in most o f his work Soc. 17. 1-9 ' he uses the skin or its appendages as a b~opsy ttssue 3 Vermorken, A. J. M., Wtrtz, P., Spierenburg, for cancer studies. He is also involved m studies on G. Th., van Bennekom, C. A., de Btmyn, acne andpsoriasis.