Serial Cultivation of Single Keratinocytes from the Outer Root Sheath of Human Scalp Hair Follicles

A lain Limat, Ph.D. and Friedrich K. N aser, Ph.D. Cosmital SA, Marl y, Switze rl and (Resea rch Compa ny of We ll a AG, Darmstadt, F. R.G.)

A m e tho d fo r the isolatio n of o uter root sheath k eratino­ h air follicle (yielding approximately 1. 5 X 104 cells) con­ cytes from pluck ed human hair follicles and for their sub­ sistentl y gave rise to a confluent 35-mm culture dish (with sequent cultivatio n h as been develop ed. T he selective tryp­ approximately 1.5 X 106 cells) w ithin abo ut 2 weeks. The sinizatio n of o uter root sheath k eratinocytes provided a o u ter root sh eath k eratinocytes can be serially p assaged for sin gle cell su spension of d efined o rigin within the h air fol­ up to 3 times and also cryopreserved . J In vest Derm ato! licle. T h e 3T3 feede r layer technique suppo rts su stained 87:485- 488, 1986 growth of these cells in that as little as o ne single pluck ed

lucked hair folli cles were fi rst used as a convenient bi­ the fo llicles were transferred in to a 35-mm Petri dish. The re­ opsy material fo r the study of geneti c disorders and for maining liquid was sucked off and 10 fo llicles covered with 200 diagnosti c purposes. T hereafter, it beca me evident that ILl 0.1 % trypsin (1: 250) and 0.02% EDTA in phosphate-buffered hair fo llicles provide an interesting model system of saline (PBS) without Ca + + and M g + + (pH 7.2) and incubated epithelial ce ll s for more fundamenta l biomedi ca l re­ for 10 min at 3rc. T hereafter, a sin gle cell suspension was ob­ Psearch [1] . tained by vigorously pipetting the fo llicles in D M EM supple­ In recent years several meth ods have been devel oped for es­ mented with 10% newborn calf serum. For SO I11.e experiments, tablishing primary cul tures of human hair follicle cell s. Weterings the remaining keratinocytes still ad hering to the follicles were et al [2] were first able to grow human hair follicle keratinocytes sequentiall y removed by 2 additional trypsinization steps of 10 by expl anting plucked scalp hair follicles on bovine eye lens ca p­ min and 30 min, res pecti vel y. T he dissociated keratinocytes were sules. Si milarl y, Wells succeeded in explanting plucked human centrifuged fo r 10 min at 200 g and cells from one fo llicle resus­ hair follicles di rectl y on tiss ue cul ture plastic [3]. Recently, we pended in 2 ml of complete culture medium each. T he culture became interested in the study of epi theli al- mesenchymal inter­ medium consisted of 3 parts of D M EM (Seromed) , containing actions, especiall y concerning the hair fo llicle biology. For this sodium pyru vate and 1 g/liter glucose, and 1 part of Ham 's F-1 2 pu rpose, we aimed to initiate primary cultures of disaggregated (Seromed), supplemented w ith 10% fetal calf serum (Gibco), 5 hair follicle cells in order to in crease the proli fera ti ve ca pacity of ILg/ml in sulin (S igma), 0.4 ILg/ml hydrocortisone (S igma), 0.135 the primary hair keratinocytes and to get a more defined and mM adenine, 2 nM trii odothyronine, 0. 1 nM choleratoxin (S igma), standardized cul ture system . 10 ng/llll epidermal growth factor (Seragen) , and antibiotics (50 In this report, we describe a method for the growth of prim ary U / ml penicillin and 50 IL g/ml streptomycin) modified after Gilfix cultures starting from dissociated human hair follicles and for and Green [4]. T he keratinocytes were seeded at a density of2-3 2 their serial cul tivati on. T he method consists essentiall y in dis­ X 103 cells/clll2 on a 3T3 fee der layer (2 X 10" cells/cm ) (ATCC aggregating the outer root sheath keratinocytes with tryp­ CCL 92) prepared according to Rheinwald and Green [5]. T he sin/EDT A and by plating them on a pre form.ed 3T3 fee der layer. 3T3 cell proli fera ti on had been halted by a treatment with 6 ILg/ml Cells fro m only one single hai r fo llicle give rise to a confluent mitom ycin C (Sigm a) fo r 8 h. 35-mm Petri dish within about 2 weeks with a multipli cation The cultures were incubated at 37°C in a humidified atm osphere 3 factor of 10 and 12 es timated cell generati ons. of 5% CO2 in air and the medium changed tw ice a week. As a rule, the fee der layer was renewed once a week after selective M ATERI ALS AN D M ETH ODS rem oval of the 3T3 cells using 0.02% EDT A [5]. C ell Isolation and Cultivation Hair follicles were plucked from the scalp and the bulk of th e hair shafts cut off. T he fo lli cles Subculture After 2-3 weeks in primary culture, the residual were immersed in D ulbecco's modified Eagle's medium (D MEM ) 3T 3 cells were removed with 0.02% EDTA and the keratinocytes buffered with 25 mM H EPES (N-2-hydroxyethylpiperazine-N '- detached and disaggregated to si ngle cells by incubating with 2-ethanesulfonic acid) and supplemented with 400 U /ml penicillin 0.1 % trypsin (1: 250) and 0.02% EDTA in PBS (pH 7.2) at 37°C and 400 ILg/ml streptomycin. T hose in anagen phase, indicated for 15-30 min. T he dispersed cells were replated in cul ture me­ by the visible bulb and intact outer root sheath, were ca refull y dium at a density of 50 cells to 104 cells/cm2 on a fres h 3T3 feeder selected under the dissecting microscope. After 2 additional rinses, layer.

Cryopreservation For storage in liquid nitrogen the cells were released from culture dishes as described' above. After 1 wash Manuscrip t received October 16, 1985; accepted for publi ca tion March with culture medium, 106 cells were res uspended in 1 ml culture 28, 1986. medium containing 10% glycerol, transferred into a cryotube Reprint requ ests to: Alain Limat, Ph. D., Cos mital SA, Rtc de Chcsa li es (Nunc), and placed within a styrofoam container in a - 80°C 21, CH-1723 Marl y, Sw itze rl and . Abbrevi ations: freeze r. After 24 h the tubes were transferred to a liquid nitrogen DMEM: Dul becco's modi fied Eagle's med ium ta nk. To recover the cells fro m frozen storage, the tubes were PBS : phos phate-buffe red sa line warmed to 37°C and the suspension diluted in culture m edium .

0022-202X/86/lG03.50 Copyri ght © 1986 by The Society for In vestigative Derm atology, Inc.

485 486 L1MAT AND NOSER THE JOURNAL OF INVESTIG ATIVE DERMATOLOGY

Figure 1. Plucked adult human scalp hair follicle. A , Whole hai r follicle with its usu­ all y almost intact outer root sheath in the portion between the upper part of the hair bulb and the presumptive level of the se­ baceous gland which is normally absent. Unstained, bar = 0.5 mm. B, Transversal secti on of the middle third of a hair follicle. The peripheral layer of the columnar basa l cells is often preserved over an extensive portion of the circumferen ce. H&E, bar = 50 /-Lm. A

The cell s were washed and then inoculated at a density of 2 x final enzym atic digestion of 30 min removed the remainder of 104 cells/cm 2 onto a feeder layer. the outer root shea th cell s, leavin g behind the denuded inner root sheath (Fig 4) . Histologic Examination Confluent cultures of hair follicle ke­ ratinocytes grown in culture dishes were detached as epithelial Cell Culture Primary cultures of hair follicle keratinocytes could sheets according to Green et al. [6] and Watt [7]. Briefly, the be routinely es tablished by plating cells disaggregated from one cultures were incubated with Dispase II (Boehringer Mannheim) single hair follicle on a preformed 3T3 feeder layer in a 35-mm at 1. 2 units/ml in PBS. After about 15 min at 37°C, the epithelial culture dish or from 10 follicles on a 3T3 feeder layer in a 100- sheet started to separate from the plasti c surface. The detachment mm culture dish. The seedin g efficiency of the keratinocytes at could be accelerated by carefully lifting the sheet with tweezers. the initiation of the primary cultures was 2.6 ± 0.5% (n = 4) . Pieces of the sheet were punched out, stretched over a Whatman After 2-3 days in culture, groups of2-4 keratinocytes surrounded N o.1 fi lter paper,· and washed with PBS. They were th en fixed by 3T3 cell s were visible. These cells formed quickly expanding in Bouin's solu tion for 24 h and processed for wax embedding. colonies pushing away the 3T3 cells. The keratinocytes were of Seri al sections were prepared at 7 J.l.m and stained with hema­ similar small size and closely packed in a polygonal epitheloid toxylin and eosin . arrangement (Fi g 5) . As long as the colonies were of limited size, In order to facilitate their handling the plucked hair follicles the weekly renewal of the feed er layer proved to be beneficial. were embedded in a solution of 3% agar in PBS prewarmed at Otherwise, the cell s at the margin of the colonies enlarged and 40°C prior to the fixation. The subsequent processing of th e hair eventually ceased to divide in the absence of a feeder layer, as it follicles was carried out as described above. has been noted for interfollicular epidermal keratinocytes [8]. When reaching a larger size the colonies started to pile up and to stratify in their center, leading occasionally to the formation of domes RESULTS (Fig 6). These cultures were confluent after 2-3 weeks of incu­ Cell Isolation The plucked anagen hair follicles from adult bation. As approximately 107 cell s could be generated from 10 human scalp exhibited an almost intact epithelia l outer root sheath fo llicles in a 100-ml11 dish, the number of actively growin g cells (Fig lA,B) . Only the outer root sheath portion at the level of the in creased about WOO-fold during the primary culture li fetime. infundibulum and around the upper part of the hair bulb was Due to the virtual absence of dermal components in the plucked usually lacking. In most cases the lower part of the hair bulb and hair follicle and to the renewal of the feed er layer, an overgrowth the sebaceous gland remained in the tissue. by fibroblasts has never been encountered. O uter root shea th The disaggregation of the outer sheath keratinocytes with 0.1 % keratinocytes originating frol11 1110re internal layers removed by trypsin and 0.02% EDTA for 10 min yielded a mean cell number the second trypsinization step also consistently gave rise to cul­ of 1.5 ± 0.4 x 104 (n = 7) per follicle. Comparative studies of tures displaying similar morphology but w ith a seeding efficiency the untreated and disaggregated follicles indica ted that cells from of 1.1 ± 0.5% (n = 3). T his result indicates that more axially the external layers of the outer root shea th had been removed located outer sheath cells still retain their proliferative ca pacity in (Figs lA,B, 2A,B) . vitro. Furthermore, even cell s disaggregated by the third tryp­ When the remaining follicle was trypsinized for an additional sinization step were still able to attach and proliferate in culture. 10 min, 0.5 ± 0.25 x 104 (n = 11) cells were additionall y obtained A cross-section through a confluent culture detached from the per follicle. The thickness of the residual outer root sheath de­ plastic by means of dispase treatment reveals the multilayered pended on individual differences in the diameter of the plucked structure with the polygonal basal cell s underneath and the more hair follicles and ranged from roughly half of the outer sheath flattened cells of the upper layers (Fig 7). cell s still present (Fig 3A,B) to their almost complete removal. A As for the epidermal keratinocytes [8], the hair follicle kerat-

Figure 2. Plucked adult human scalp hair follicle fo llowin g a treatment with 0.1 % tryp­ sin and 0.02% EDTA for 10 min. A , Whole trcated hair fo llicle. T he extent of the cell rc­ moval can vary over the follicle length leadin g sometimes to an uneven thinning of the hai r follicle. Unstained, bar = 0.5 mm. B, Trans­ versal section of the middle third of a treated hair fo llicle. At this level, usuall y about half of the outer root sheath cells have been re­ moved. H&E, bar = 50 /-Lm . VOL. 87. N O.4 OCTOBEH 1986 C ULT UHES OF HUMAN HAm FOLLICLE C ELLS 487

Figure 3. Plucked adult human scalp hair follicle follo wing a treatment with 0.1 % tryp­ sin and 0.02% EDTA for 20 min. T he cor­ I responding pictures (A, unstained, bar = 0.5 mm; and B, H&E, bar = 50 J.Lm ) show that the outer root shea th cells have been removed to a grea t extent. A - B

Figure 4. Plucked adult human scalp hair follicle following an ultimate treatment with 0.1 % trypsin and 0.02% EDTA for 50 min.

...... ~"""...-...... ~~\.: ...... T he corres ponding pi ctures (A, unstained, bar = 0.5 mm; and B, H &E, bar = 50 J.Lm ) ---J,.: ·"~ I""''''''' show that all the outer root sheath cells have been removed leaving behind the denuded inner root shea th on (A) and the denuded hair shaft on (B) . ...- ~ ___~~ B

inocytes were best subcultured within 2 weeks aft er plating, be­ from one single hair follicle routinely gave rise to a confluent fore individual colonies had enlarged too much. T he subculti­ culture in a 35-mm dish after about 2 weeks. T he combined effect vation o n a fresh feeder layer enabled clonal seeding densities (50 of the medium compositi on and of the feeder layer ,!p parently cells/cm 2) with a cloning efficiency in the range of 0.5-5%. Until promoted rapid cell growth and, concomitantly, delayed the ter­ now cell s could be successfully passaged 3 times. minal differentiation. M ost cells isolated by the m ethod reported Cells recovered from liquid nitrogen storage and plated in the here ori ginated from the outer root sheath portion located be­ presence of a 3T3 feeder layer showed no morphologic differences tween the opening of the sebaceous gland and the upper hair bulb. compared with the nonfrozen cells. The seeding efficiency of the Thus, it is very unlikely that any of the keratinocytes growing keratinocytes recovered from cryogenic storage was in the range in culture originated fro m the infundibulum where the outer root of 1-5%. shea th is indistinguishable from interfollicular epidermis. In ad­ dition, the results obtained by sequential trypsinization of the DISC USSIO N outer root shea th suggest that cells from all layers retain a high Previously, growth of human hair keratinocytes in primary cul­ capacity for proliferation in culture. This behavior contras ts to tures has been reported using explant cultures of plucked hair the situation in the outer root shea th which is believed to be a follicles on bovine eye lens ca psules [2] or on plastic [3]. U sing rather static structure as judged fro m morphologic observations. this approach, major drawbacks were rather limited cell nL1l11bers However, both necrotic cells and mitotic figures are occasionall y and cultures ofIow reproductivity. Preliminary attempts to grow found in the outer shea th, indicatin g that cell division and cell dispersed hair keratinocytes on different biologic substrates (rat death occur at a well balanced, but low level w ithin this tissue tail collagen, fibronectin, extracellular m atrix secreted by bovin e [10]. corneal endothelial cells [9] failed (unpublished results). The cell s In vivo, the columnar basal cell s, instead of forming hemides­ neither attached nor spread out under these conditions. However, mosomal contacts, send out minute basal cytoplas mic processes this impairment could be ci rcumvented by adopting the 3T3 co­ through the surrounding basement mem brane [11 ,12]. Moreover, culture conditions es tablished by Rheinwald and Green [5] . Cells the outer sheath cells of the lower two-thirds of the fo llicle have

Figure 5. Primary hair follicle keratinocytes surrounded by mitomycin Figure 6. Primary hair follicle keratinocytes, 18 days after plating. Phase C-treated 3T3 cells,S days after plating. Phase contrast, bar = 40 J.L111 . contrast, bar = 40 J.L1I1 . 488 L1MAT AND N OSER TI-IE JO URNAL O F INVESTIGATIV E DERMATOLOGY

cell s w hi ch can be further expanded to about 108 actively growing celJ s by 2 passages, this might provide an alternative source for autologous keratinocytes for the treatment of burns.

We are illliebted to Dr. N . E. FlI scllig, Delltscil es Kre b~r~rs c/lllllgsze lltrllll1 , Heide/berg, Jo r stilllltiatillg disCII ssiolls 111/(1 Jor critica lly I'cadillg til e llralll/script. We also wisil to til allkjaqllelill e Clrarri crc (md Marie-Clallde RotzetterJor skillfili 1II 1d expert tec/mical assistall ce.

REFERENCES 1. Vermorken AJM, Weterings PJJM , de Bruyn C HMM, Geerts Sj: Human hai r follicl es in biomedical resea rch. Current Problems in Dermatology, vol. 9. Edited by JWH Mali . Basel, S Karger, 1981, pp 50-82 Figure 7. T ransversal section through a confluent hair follicl e kerati­ 2. Weterin gs PJJM, Vermorken AJM, Bloemendal H: A method for nocyte culture detached with dispase. The very co mpact arrange ment of culturing human hair follicl e cell s. Br J DermatoI104:1-5, 1981 th e basa l cells has probabl y been enh anced by a co ntraction phenomenon 3. Well s J: A simple technique for es tablishing cultures of epithelial occurring durin g the detachment procedu re. H&E, bar = 25 jLm . cell s. Br J DermatoI107:481- 482, 1982 4. Gilfi x BM , Green H: Bioassay of retinoids using cultured human conjuncti va l keratinocytes. J Cell PhysioI11 9: 172-174, 1984 been described to be virtually devoid of desm osomes except at 5. Rhein wald JG, Green H: Se ri al cultivation of strains of human epi­ the contacts w ith cell s of the Henle's layer w here these structures derma l keratinocy tes: th e formation of keratinizin g co lonies from are abundant [11]. Another characteristic of the o uter sheath cells sin gle cell s. Cell 6:331-344, 1975 is thei r hi g h content of g lycogen g ranules [11 , 12]. In expl ant 6. Green H, Kehinde 0, Thomas J: Growth of cultured human epi­ cultures, however, numero us desmosom es arc fo und, as shown dermal ceUs into multiple epithelia suitable for grafting. Proc Nat! Aca d Sc i USA 76:5665-5668, 1979 by Weterings et al [1 3]. In additio n, g lycogen accuJ11ulatiQns have been dem o nstrated mainly in the basal cell s, w hereas the keratin 7. Watt FM : Selective mi gration of terminally differentiating cells from polypeptide pattern resembled that of the o uter root sheath in th e ba sa l layer of cultured human epidermis. J Cell Bioi 98:16-31, 1984 situ [13; manuscript in preparatio n] . W e arc currently investi­ 8. Rh ein wa ld JG: Serial cultivation of normal human epidermal kerat­ gating the role of different culture conditions for the m odulatio n in ocytcs. Meth ods Cell Biology 21A:229-254, 1981 of phenotypic expressio n, such as the effect o f dermis o n the differentiatio n of o uter root sheath cell s cu ltured at the air-liquid 9. Gospodarowicz D: The co ntrol of mammalian cell proliferation by growth fa ctors, ba sement lamina, and lipoproteins. J In ves t Der­ interphase. T hese studies might help to elucidate to what extent matol 81 (suppl):40s-50s, 1983 these cell s retain in vitro their orgin al pro perties o r shift toward 10. Montagna W, Van Scott EJ: The anatomy of the hair follicle, The a more epidermal- like phenotype. Such a finding would not be Biology of Hair Growth. Edited by W Montagna , RA Elli s. New suprising since the regeneratio n of epid ermis after wounding has York/ London, Aca demjc Press , 1958, pp 39-64 been di scussed to proceed fro m the o uter root shea th of intact 11 . Pu ccinell i V A, Caputo R, Ceccarelli B: T he structure of human hair hair fo llicles. fo llicle and hair shaft: an el ectron microscope study. ltal Dermatol In summary, the technique described here enables the gener­ 108:453-498, 1967 ation of a large amount of hair follicle cells rea dily available for 12. Forslind B: Die wachsende Ana genphase , Haar und Haarkrankhei­ biomedical research and diagnosis, provided the initial presence ten. Edited by CE Orfanos. Stuttga rt/ New York, Gustav Fischer of the forei gn feeder celJ s is not prejudicial. In particul ar this might Verl ag, 1979, pp 51 -75 contribute to the elucidation of the biologic role of o uter root 13. Wcterings PJJM , Verh agen H, Wirtz P, Vermorken AJM: Differ­ sheath cell s in induction and m aintenance of hair growth and entiation of human sca lp hair follicl e kcratinocytes in culture. Vir­ differentiation. Since o ne sin gle follicle gives rise to 106 viable chows Arch. [Cell Pathol] 45: 255-266, 1984