IN VITRO CELLULAR & DEVELOPMENTAL BIOLOGY Volume 21, Number 3, Part I, March 1985 �9 1985 Tissue Culture Association, Inc.

S E L E C T I V E G R O W T H O F N O R M A L A D U L T H U M A N U R O T H E L I A L C E L L S IN S E R U M - F R E E M E D I U M

DAVID KIRK,' SUSUMU KAGAWA, 2 GUDRUN VENER, K. SHANKAR NARAYAN, Y. OHNUKI,

AND LAWRENCE W. JONES

Huntington Medical Research Institutes, 99 North El Molino Avenue, Pasadena, California 91101

{Received 10 February 1984; accepted 8 November 1984~

SUMMARY

A serum-free medium (HMRI-2) has been developed for the outgrowth and subculture of epithelial cells from normal adult human ureter and bladder. Medium HMRI-2 consists of Ham's MCDB 152 with double the amounts of the essential amino acids in Stock 1, low Ca 2§ (0.06 mM) and is supplemented with epithelial growth factor, 5 ng/ml; transferrin, 5 tag/ml; insulin, 5 tag/ml; ethanolamine and phosphoethanolamine, 0.1 mM each; hydrocortisone, 2.8 X 10 -6 M; and bovine pituitary extract, 126 tag protein/ml. The cultured cells showed ultrastructural markers of epithelial cells Iprekeratin fibers, tonofilaments, surface microvilli with glycocalyxj, exhibited ABO antigens, and had a normal human diploid karyotype. Primary cultures could be subcultured and also cryopreserved in HMRI-2 in liquid nitrogen. Cells in mass cultures showed a population doubling time of 40.5 -t- 4.5 h and had a maximum in vitro life span of 20 to 25 population doublings. It was observed that primary outgrowths, secondary cultures, and even cryopreserved cells all retained the capacity to respond to high Ca 2+ and serum by differentiation and desquamation. This study has resulted in the availability of easily obtainable serum-free epithelial cultures from normal adult human ureter and bladder. The useful in vitro life span of these cultures may be important in future studies of carcinogenesis.

Key words: normal human urothelium; serum-free medium; terminal differentiation; low calcium.

INTRODUCTION

Normal adult human epithelia, the cell type associated with about 95% of human cancers (carcinomask have traditionally been refractory to growth in monolayer cell culture. Studies with human keratinocytes have indicat- ed that this lack of proliferative capacity is due to a propensity of epithelial cells to differentiate terminally in conventional monolayer culture systems IlL Two inde- pendent but complementary approaches have radically improved the ability to culture human epithelia. These two strategies are: (a) adjustment of both the composition and level of each nutrient medium component {2), and ~b) replacement of serum by hormones and growth factors (3D. This joint approach has enabled the construction of specific media that selectively promote the proliferation, as opposed to the terminal differentiation, of specific cell types. Completely defined media have now been developed for several normal human epithelial cell types,

i To whom requests for reprints should be addressed. 2 Current address: Department of Urology, University

Tokushima, Kuramoto-Cho 2, Tokushima City 770, Japan. of

165

including epidermal keratinocytes ~4,5k bronchial epi- thelium { 6), and mammary epithelium { 7 }.

As a first step toward developing a fully defined medium for replicative human urothelial cell cultures, we have attempted to develop a serum-free cell culture system that used reduced amounts of undefined protein components. Such a model system would permit a direct experimental study on the mechanism of carcinogenesis in the human urothelium. Our early attempts to culture explants of normal adult human ureter and bladder in a conventional serum-containing medium iPFMR-12D re- suited initially in abundant epithelial outgrowth; howev- er, these outgrowths rapidly differentiated into squamous nondividing cells that detached from the culture dish. Because of the similarity of this process of terminal differentiation in our urothelial cultures to that reported for cultures of human and mice epidermis (8), medium found to reduce terminal differentiation in epidermal keratinocytes was investigated. In this paper, we report the isolation, replication, and characterization of epitheli- al cell cultures from normal adult human ureter and bladder using a modified MCDB 152 medium containing reduced Ca 2§ and supplemented with eight hormones/ growth factors and a small amount of an undefined bovine pituitary extract.

166 KIRK ET AL.

MATERIALS AND METHODS

Sources of urothelial tissue. Human ureter and bladder samples (ranging in age from neonatal to 73 yr) were obtained, either at autopsy or from surgery under approved guidelines, from Dr. T. Pretlow, University of Alabama, Birmingham, and from Dr. L. W. Jones in Pasadena. All tissue samples used were of normal histology. Autopsy specimens were obtained within 5 h after death. Specimens were transported to the laboratory at ambient temperature in PFMR-4 ~9) with or without 5% fetal bovine serum IFBS) and antibiotics (penicillin, 100 U/ml; kanamycin, 100 t~g/ml). Autopsy samples from Alabama were processed in our laboratory approximately 24 to 36 h after autopsy. Fresh surgical samples were processed within 2 h after removal.

Serum, hormones and growth factors. Endothelial cell growth supplement ~ECGS), platelet-derived growth factor (PDGF), and both fibroblast (FGF} and epidermal growth factors (EGF) were purchased from Collaborative Research, Waltham, MA. Transferrin (TF), insulin (IN), ethanolamine iEA), and phosphoethanolamine (PEAJ were obtained from Calbiochem-Behring Corp., La Jolla, CA. The ECGS and PDGF were dissolved in sterile distilled water as concentrated stocks. The EGF was prepared as a 1000-fold concentrate in HEPES buffered saline (HBS) containing 500 t~g/ml bovine serum albumin (BSA, essentially fatty acid free, catalog no. A 7511, Sigma Chemical Co., St. Louis, MO). The IN was dissolved in 0.1 mM HC1, and EA, PEA, and TF were dissolved in HBS at 100X. Hydrocortisone (HC, Steraloids, Inc., Wilton, NH) was used as a 1000-fold concentrate in ethanol. Bovine pituitary extract ~BPE) was prepared by homogenizing the pituitaries (Pel Freeze, Rogers, AR) in 0.15 M NaC1 198.7 g in 248 ml) for 5 rain and extracting overnight at 4 ~ C. The homogenate was centrifuged (5000 Xg, 15 min} and then the supernatant fluid was sedimented at 100 000 Xg for 1 h at 4 ~ C. The superuatant fluid was sequentially filtered through 0.8, 0.45, and 0.22 gm membrane filters for sterilization and then stored frozen at --70 ~ C. The protein concentration of the BPE extract was 12.6 mg/ml. The FBS iIrvine Scientific, Irvine, CA} was passed through Sephadex G10 at 45 ~ C as previously described ~6). The Ca 2§ concentration of whole FBS (3.625 mM) and the G10 stripped FBS (S-FBS, 0.033 mM) was determined by atomic absorption.

Nutrient media. Medium MCDB 152 (5~ and PFMR-4 (10~ were prepared in our laboratory. Medium PFMR-12 was PFMR-4-supplemented with 5% S-FBS, 5 ng/ml EGF, 5 ~g/ml IN, and 10 -7 M HC. The HMRI- I consist- ed of Ham's MCDB 152 with low Ca 2§ ~0.1 mM) supple- mented with EGF, 5 ng/ml; TF, 5 gg/ml; IN, 5 ~g/ml; EA and PEA, 0.1 mM each; HC, 2.8 X 10 -6 M; and bovine pi- tuitary extract, 126 gg protein/ml. The HMRI-2 is medium HMRI-1 with double the amounts of the essential amino acids (EAA) in MCDB 151 stock 1 (10} and the calcium adjusted to 0.06 mM.

Primary cultures. Tissue was minced with either scissors or scalpels to fragments measuring about 2 )< 2 mm. Fragments were allowed to attach to petri dishes

(LUX, Scientific, Newbury Park, CA) containing a thin film of medium. Additional medium was then added and replaced every 2 to 3 d.

Subcultures. Monolayers were dissociated with 1% polyvinylpyrrolidone (Calbiochem-Behring, La Jolla, CA), 0.02% [ethylene bisloxyethylenenitrilo)]-tetraacetic acid (Eastman Kodak Co., Rochester, NY) and 0.2% trypsin, 2X crystalline (Sigma Chemical Co., St. Louis, MO), in HBS referred to as PET (9). The cells were sedimented (500 Xg, 5 min) and resuspended in medium.

Cryopreservation. Cells were frozen in liquid nitrogen vapor at 2 X 106 cells/ml in HMRI-1 or HMRI-2 containing 7.5% dimethyl sulfoxide, then stored in liquid nitrogen (9).

Growth measurement. The population doubling time {PDT) was determined by plating 5 X 104 cells per 35-mm petri dish and counting the resultant cell number in a hemocytometer at di/ferent time intervals. Using values from the linear portion of the log plot, the PDT was calculated from the equation:

PDT = t2-- tj/ (Log2 x2 --Log2 x,) where x, = cell number at t, and x2 = cell number at t2. Clonal plating efficiency of Weigert's iron hematoxylin-stained cultures was expressed as the percentage of seeded ceils that produced colonies containing more than 10 cells.

Characterization. For scanning (SEM) and transmission iTEM) electron microscopy cells growing on plastic cover slips or dishes were fixed and processed as described previously (11). Screening for mycoplasma by SEM was uniformly negative (12). Conventional Giemsa staining was used for counting chromosome numbers. Modal chromosome number was obtained from counts on 50 well-spread metaphase cells. For karyotypic analysis, Q- and C-bandings were used. Details of these methods have been published (13). Type A red cell surface antigen was detected by a modification of the red cell adherence test t14). Formalin {10% }-fixed cells were washed three times with HBS and treated with the appropriate antiserum at room temperature for 15 min. After washing three times with HBS and once with 7% BSA in HBS, human red cells were added at a concentration of 1% Ivol/vol}. After 15 min at room temperature, cultures were inverted in HBS for 15 min to remove unbound red ceils. Cultures with bound red cells were fixed in 2% glutaraldehyde buffered by phosphate and containing 5.4% dextrose, and finally stained with hematoxylin and eosin.

Incorporation of tritiated thymidine ([3H]TdR). Cultures were pulsed with [3H]TdR (1 gCi/ml; sp act 74.9 Ci/mmol, New England Nuclear, Boston, MA) in thymidine-free medium as indicated. Hydroxyurea (10 raM, Calbiochem-Behring), an inhibitor of DNA synthe- sis, was added as a control. Cultures labeled with pH]TdR were washed three times with HBS. The cell layer was wiped off with a cotton pellet attached to a wooden stick. The cells enmeshed within the cotton pel- let were immersed in two changes of 5% trichloroacetic acid at 4 ~ C (30 min each) followed by two changes of ethanol 120 rain each at room temperature), and allowed to dry. The cotton pellets were transferred to counting vials where cellular material was dissolved in 0.4 ml Protosol

SERUM-FREE UROTHELIAL CELL CULTURE 167

(New England Nuclear). After 10 min at 37 ~ C, 10 ml of Aquasol scintillation cocktail {New England Nuclear) was added to each vial and the samples were counted in a Beckman LS-9000 liquid scintillation spectrometer.

RESULTS

Preliminary culture experiments. Initially, urotheUum was cultured as explants using PFMR-12 medium. A survey of such cultures obtained from normal urothelium has shown the success rate of establishing primary cultures to be 40% (8 out of 20) for bladder mucosa and 16.7% (1 out of 6) for ureter specimens. However, PFMR-12 promoted mixed epithelial and fibroblastic cultures as observed by phase contrast microscopy. The epithelial outgrowths (Fig. 1 a) in the mixed cultures soon enlarged and desquamated from the surface of the culture dish. The desquaminated cells were judged to be t e rmina l ly d i f fe ren t i a ted because they con ta ined abundant keratin as demonstrated by intense Rhodamine B staining {15) (data not shown) and could not be sub- cultured. Clearly, PFMR-12 was unsatisfactory as a medium for supporting the replicative growth of undiffer- entiated urothelial cells. Because of the striking similarity of this terminal differentiation to that of skin, we inves- tigated the efficacy of media developed for the culture of human skin cells in Ham's laboratory including MCDB 151 (10L MCDB 152 I4), and MCDB 153 (16L

Of these various formulations, exploratory experiments using MCDB 152 with seven factors {152-7F) were most promising. When the level of EGF was reduced, 152-7F was renamed HMRI-1 (Table 1L Explant cultures set up in HMRI-1 now resulted in the establishment of primary cultures with a success rate of 71.4% (5 out of 7} for bladder and 80% (4 out of 5) for ureter. Not only did the substitution of HMRI-1 for PFMR-12 improve the success rate in establishing primary outgrowths from urothelium by twofold, it also proved to be selective for epithelial outgrowth inasmuch as no fibroblast outgrowths were observed in this medium. In contrast to the tightly attached migrating margin of bladder epithelium in PFMR-12 {Fig. 1 a), either single migrating cells or loosely attached sheets of cells emerged from the same bladder tissue when explanted in HMRI-1 {Fig. 1 bL

Growth of normal adult urothelium in HMRI-I. A confluent primary culture of normal adul t ureter illustrates the classical closely apposed epithelial pattern (Fig. 2). Two types of epithelial cells were generally seen in both ureter and bladder cultures; the small type with relatively little cytoplasm and the larger granulated cells that apparently do not divide (Fig. 3L The declining proliferative capacity on subculture was associated with an increased proport ion of the larger terminal ly differentiated cells {Fig. 3). The small cell was thought to be responsible for repopulating the culture. Confluent cultures could be subcultured using P E T and reseeded in HMRI-1 medium at a plating density of l0 s per 60-mm dish. Cultures have been passed up to four times. A calculation of the cumulative increase in cell numbers with each subculture has shown ureter cells to have undergone 25.33 population doublings during the first

three culture passages, excluding the unknown number of population doublings that occurred during primary growth.

FIG. 1. Primary epithelial outgrowths from a 54-yr-old normal female human bladder explanted in PFMR-12 for 16 d tFig. 1 a) and in HMRI-1 for 8 d (Fig. 1 b). Phase contrast. X104. Bar = 100 ~m.

FIG. 2. Primary outgrowth of epithelial cells from a 71-yr-old normal female human ureter after 7 d in HMRI-1. Phase contrast. X104. Bar = 100tam.

FIO. 3. Epithelial cells derived from a 68-yr-old normal male human bladder after two subcultures in HMRI-1. Phase contrast. X104. Bar = 100 vm.

TABLE 1

EXPERIMENTAL MEDIA FOR NORMAL ADULT HUMAN UROTHELIAL CELLS

C O M P O N E N T S ~ PFMR-12 152-7F HMRI-1

Basal nutrient PFMR-4 MCDB 152 MCDB 152 Ca 2§ 10 -3 M 10 -4 M 10 -4 M EGF 5 ng/ml 50.0 ng/ml 5.0 ng/ml TF - 5.0/~g/ml 5.0 ~g/ml IN 5.0 #g/ml 5.0 #g/ml 5.0/~g/ml EA - 10 -4 M 10 -4 M PEA - 10 -4 M 10 -4 M HC 1.0 )< 10 -7 M 2.8 )< 10 -6 M 2.8 )< 10 -6 M BPE - 126 ~g/ml 126 gg/ml S-FBS 5.0% - BSA - 0.5 gg/ml 0.5 gg/ml mOSM/kg 280 312 315 pH b 7.4 7.4 7.4 pCO2 3.0% 3.0% 3.0%

~ Materials and Methods for abbreviations. hAs measured after equilibration with C02. Medium was brought to

this pH before addition of NaHC03.

168 KIRK ET AL.

6.0

t33

-4 5.0 -.J tU (.9

<5 C3 -4

4 . 0 I I I I I 2 4 6 8

DAYS FIG. 4. Growth curve of normal adult human ureter cells cul-

tured in HMRI-I. The cells were described in the legend to Fig. 2. Cell stocks, frozen in liquid nitrogen at the first passage, were thawed and seeded at 5 X 104 per 4 ml of HMRI-I in 60- mm petri dishes. Each point is the mean of triplicate cultures.

Information on growth rates was determined using cryopreseved primary human ureter outgrowths. Cells frozen at the first passage were thawed and plated (5 X 104 cells/60 mm dish) in 4 ml HMRI-1. Viability of the thawed cells was observed to be high because at least 95% of the cells stuck down and spread out on the dish after 24 h. At intervals, triplicate cultures were trypsinized and the cells counted. Growth was exponential for 6 d (Fig. 4L The average doubling time in HMRI-1 was 51.0 4- 6.6 h and represents a sevenfold increase in cell number in 6 d.

These thawed ureter epithelial cells were also capable of clonal growth in HMRI-1. Cells were seeded (5 X 103) in 4 ml HMRI-1 per 60 mm dish. After 7 d, cultures were stained with Weigert's iron hematoxylin and scored for number of clones. After cryopreservation, the first passage normal adult human urothelial cells cloned with an efficiency of 1.05% 4- 0.18.

Effect of selected factors on growth. The effect of selected growth factors on the growth of first passage normal adult ureter cells in HMRI- I was determined during the logarithmic growth phase 12 to 6 d, Fig. 4). Doubling the concentrations of the essential amino acids (EAAI of MCDB 151 increased growth 63.8% above the control value and decreased the PDT from 51 to 40.5 h (P < 0.10, Table 2L Although ECGS did not affect growth, it did improve the appearance of the cells and increased the proportion of small proliferative cells. Serum inhibited growth by one-third and increased the PDT to 63.2 h (P < 0.10), whereas commercial PDGF completely inhibited growth. In both cases, the decreased growth was

accompanied by an increase in enlarged squamous, termin- ally differentiated cells. Omission of BPE from HMRI-1 also reduced growth to half that of the control and nearly doubled the PDT ~control, 51.0 h; without BPE, 86.8 h, P < 0.051.

The effect of Ca 2§ on p t t I T d R incorporation in cultured bladder epithelial cells. Because of the known regulatory effect of calcium on the growth of other normal epithelial cell types, its effect on urothelial ceils was examined. Normal bladder was explanted in HMRI-1 supplemented with 50 ng/ml EGF and 100 gg/ml BSA. The outgrowth was resuspended with PET and subcuhur- ed in the above medium in multiweil dishes. Mter 24 h, various levels of Ca 2§ were added in fresh medium. Two days later the cultures were then pulsed for 21 h with [3H]TdR in fresh thymidine-free medium containing the appropriate Ca ~+ concentration ~Fig. 5). Incorporation of [3H]TdR as a function of Ca 2§ concentration showed a maximal response at or near 6 X 10 -'~ M Ca 2§ It is in- teresting to note that increasing the Ca ~§ to 1 mM (the Ca ~+ level in PFMR-12) inhibited the [~H]TdR incorpora- tion to about the same extent (97%) as did hydroxyurea, a known inhibitor of DNA synthesis (Fig. 5). Similar results were obtained using the normal ureter epithelial ceils described in Fig. 2 (data not shownL In the latter case, cell counts, rather than [3H]TdR incorporation, were used and it was observed that the maximal Ca 2§ response using both methods was virtually identical (6 X 10-~M~ for the two epithelial cell types.

Substitution of PFMR-4 for M C D B 152 as basal medium for HMRI-1. In HMRI-1 both high Ca ~§ (1 mM~ and serum ~1%) were inhibitory to ureter epithelial cell growth. Because PFMR-12 contains both Ca 2§ (1 raM) and serum (5%L it was questioned whether PFMR-4 Ibasal medium of PFMR-121 could substitute for MCDB 152 as basal medium for HMRI-1. Therefore, we prepared PFMR-4 with the same Ca ~§ level (0.1 mM} and factor

TABLE 2

EFFECTS OF DIFFFRENT ADDITIVES ON THE GROWTH OF NORMAL ADULT

HUMAN URETER EPITHELIAL CELLS* IN HMRI-1

Mean Cell Number Percent Additives X | 0 -4 Control b PD ' I ~, h Signilicance d

None 35.2 -4- 4.6 100 .0 51.0 + 6.6 - 2)< EAA" 57.7 + 6.4 163 .8 40.5 + 4.5 <0.10 ECGS, 200/~g/ml 31.0 4- 3.0 88.1 54.1 4- 5.3 N.S. MinusBPE 16.0 4- 0.2 45.5 86.8 4- 1.1 <0.05 S-FBS, 1% 23.8 4- 0.9 67.6 63.2 4- 2.3 <0.10 PDGF, 5 U/ml 3.9 4- 0.3 11.2 No growth <0.01

~ primary epithelial outgrowths were recovered from liquid N2 storage and plated in 4 ml HMRI-1 ~5 X 104 cells/60-mm dish). After 24 h, the medium was removed and the experimental media were added. Cultures were trypsinized and counted after 6 d incubation. Each value is the mean of triplicate cultures -4- SE.

bControl = 100%. ~ doubling time in hours. ~Compared to control ~Student's t test, P-value}. "Essential amino acids IEAAI in Stock 1, MCDB 151 ~10).

SERUM-FREE UROTHELIAL CELL CULTURE 169

..4

.4

50

40

30

20

I0

. . . . . T . . . . . , . . . . _ -~ e~_ e~t ._ .W, s,~,i~ . . . . I I

/ O - s / 0 - ~ / 0 - ~ 2 x / O - ~

C,~ '~" C ONCE N Tt~710N [MOLAR 1

FIG. 5. Effect of Ca 2§ concentration on incorporation of [~It] TdR in first pa~age bladder epithelial cells from a normal 59- yr-old female. A multiwell dish was inoculated with 6 )< 104 cells/well in l ml of HMRI-1 containing ~ ng/ml EGF and 100 pg/ml BSA. After 24 h the medium was replaced with media containing graded levels of Ca 2.. After a 2 d treatment, the cells were pulsed for 21 h with [3H]TdR 12 pCi/ml} in the appropriate Ca 2* thymidine-free test media. The counts per minute were corrected for background. Each point represents the mean of triplicate cultures -4- SE. The dotted line giw~'s the corrected CPM/well at 1 X 10 -4 M Ca ~+ in the presence of 10 mM hydroxyurea.

A surface antigens. However, the antigen was distributed nonrandomly and was localized on the perimeter of the cells. The chromosomal analyses of primary cultures from two bladders ~derived from a normal 68-yr-old male and a normal 54-yr-old female} showed them both to have normal human chromosome numbers. Fur ther Q- banding of the bladder cells from the male donor showed them to have a normal human karyotype.

DISCUSSION

This study demonstrated clearly that the type of cell cultured from either normal human ureter or bladder is dependent on the nutrient medium used. Whereas a serum-supplemented PFMR-4 ~PFMR-12) promoted a mixed culture of fibroblasts and differentiated epithelial cells, HMRI-1 selected specifically for replicative epithelial ceils. Both ureter and bladder epithelial cell cultures could be obtained routinely from adult donors in their sixth or seventh decade using HMRI-1 , and, furthermore, these cells could be cryopreserved in this serum-free medium with retention of cell viability upon thawing. Confirmation that these HMRI-1 cultures were normal epithelial ceils was demonstrated by the presence

supp lements as used for H M R I - 1 (Table 1). Normal adult ureter explants cultured in the modified PFMR-4 and HMRI-1 were initially morphologically very similar. However, after 2 to 3 wk, fibroblastic growth was observed in the PFMR-4 based medium but not in HMRI-1 . This indicated that the use of MCDB 152 as the basal medium was important for preserving the selective growth of human ureter epithelial cells.

Repression of fibroblast growth in HMRI-1. Ureter and bladder fibroblasts did not grow in HMRI-1. In the rare event where fibroblasts did migrate from explants in HMRI-1 they did not proliferate and were rapidly lost from the culture. When bladder fibroblasts grown in PFMR-12 were refed with HMRI-1 , they were observed to round up and detach from the culture dish after 24 h.

Characterization o/ cell cultures. Scanning electron microscopy of bladder epithelial cells has shown that the small, loosely apposed ceils growing in HMRI-1 have abundant surface microvilli {Fig. 6 A) whereas the larger, flat ceils had fewer surface structures {Fig. 6 B). Tonofilaments as well as extensive networks of prekeratin fibers were seen by T E M (Fig. 6 C, D). The surface microvilli also contained a mid-rib and glyeocalyx typical of epithelial cells (Fig. 6 D}. Desmosomes were rarely observed. Surface ABO antigens have been demonstrated on both primary and secondary bladder epithelial cultures. Two b ladder and two ureter cultures derived from Type A donors all exhibited Type

FIG. 6. Scanning and transmission electron micrographs of normal adult human bladder epithelium in HMRI-I. SEM of secondary culture from 54-yr-old female. A, Small loosely apposed ells with surface microvilli. )<554. B, Large, fiat cells with smooth surface membranes. )< 554. Transmission electron mierographs of a 7-d-old primary epithelial outgrowth from a 70-yr-old female. C, Extensive network prekeratin fibers. )<2548. D, High power of C showing detail of larekeratin fibers and microvilll. X8862. Bar = 10 pro.

170 KIRK ET AL.

of tonofilaments, prekeratin fibers, red blood cell surface antigens, and a normal human karyotype. As regards possessing specific in vivo features, the bladder cultures had a pronounced capacity for keratinization, but lacked both glycogen and an assymetrically thickened plasma membrane.

The HMRI-1 was derived from MCDB 152, the basal medium used to develop a defined medium capable of supporting the clonal growth of human keratinocytes (4L A recent report has also described the clonal growth of normal adult human bronchial epithelial cells in MCDB 151 supplemented only with defined factors and trace elements t6L I t is remarkable that epithelia of such diverse function (skin, bronchus, bladder) derived from all three primary germ layers lskin from ectoderm, bronchus from endoderm, bladder from mesoderml can all be cultured successfully in a similar basal medium, MCDB 151, with minor differences. The two most important considerations were: lal low Ca 2§ levels, and lb) replacement of serum by hormones and growth factors. It has previously been shown that the Ca 2§ concentration regulated the balance between prolifera- tion and terminal differentiation in several epithelial cell systems tl,4,6,17,18L Interestingly, the Ca 2. levels observed for optimal proliferation of cultured epithelium from human bronchus (6), skin (5L and bladder and ureter ~0.11, 0.3, and 0.06 mM, respectively~ were all substantially below the Ca 2§ level of conventional medium (1 mML The addition of stripped serum lcontaining negligible Ca 2§ to urothelial cultures also resulted in a switch from proliferation to morphological terminal differentiation. Similar findings have been reported for bronchial cells where it was con- cluded ~19~ that the active serum component was a con- taminant associated with commercial PDGF. This would imply that inhibition of urothelial ceils by commercial PDGF, as reported here, was also due to a contaminant and not to the P D G F itself. I t is also interesting that the normal human adult urothelial cells were consistently sensitive to both Ca 2§ and serum inhibition when grown as primary or secondary cultures, or even after cryopreser- vation. Although Ca ~§ and serum were the two key factors observed here to regulate differentiation of uro- thelial cells, the basal medium was found to be important for insuring selective growth of epithelial cells. Sub- stitution of MCDB 152 lbasal medium of HMRI-1) with a low Ca 2. version of PFMR-4 (basal medium of PFMR- 12) in H M R I - I resulted initially in both qualitatively and quantitatively similar epithelial cell growth. However, the selectivity for the epithelial ceils was soon lost, and the PFMR-4 substituted HMRI-1 gradually supported addition- al fibroblast growth.

I t should be emphasized that MCDB 151 was initially developed for the singular purpose of supporting clonal growth of keratinocytes (10). Our interests, however, were in developing mass cultures of urothelial cells suitable for expansion and cryopreservation. As may be expected, therefore, considering the obvious quantitative differ- ences in the requirements of clonal versus mass culture growth, we have obtained substantially greater growth by doubling the amounts of amino acids supplied in MCDB

151 Stock 1 I10L This modification has now become routine and, along with the incorporation of 0.06 m M Ca 2§ the medium has been renamed HMRI-2 . Omission of BPE, the only undefined additive, caused a significant reduction in growth as was found also for secondary cultures of keratinocytes (4L Addition of ECGS improved the appearance of the cell cultures without being growth stimulatory. However, it is possible that ECGS may be growth stimulatory in the absence of BPE, and in fact might be a first step toward reducing the BPE require- ment.

Recently, a report appeared describing the growth and subculture of normal human ureter and embryonic bladder (20L They used Ham's F12 medium supplemented with serum, insulin, transferrin, and hydrocortisone. Their description of the primary outgrowths, with respect to ultrastructure and terminal differentiation, was remarkably similar to our urothelial outgrowths when cultured in our high Ca 2§ and serum-containing medium, PFMR-12. Our study, therefore, confirms their finding that conventional serum- supplemented medium induced cellular differentiation as evidenced by stratification and desquamation of human urothelium in cell culture. However, unlike our experience with cultures in PFMR-12, they were able to pass their cultures serially in their F12 medium, although no data concerning the in vitro life span were given. This difference in the ability to pass the cells might well be due to the relatively lower Ca 2§ level in the F12 medium (0.3 mM~ compared to that in PFMR-12 (0.92 mML Our efforts to develop replicative cell cultures from normal human urothelium from ureter and bladder have resulted in the availability of easily obtainable serum-free cultures with a useful in vitro life span. These cultures provide a controlled experimental system for studying chemical carcinogenesis in normal human urothelium.

REFERENCES

1. Green, H. Terminal differentiation of cultured human epidermal cells. Cell 11: 405-416; 1977.

2. Ham, R. G. Survival and growth requirements of nontrans- formed cells. In: Baserga, R., ed. Handbook of experimental pharmacology, vol. 57. Heidelberg: Springer-Verlag; 1981: 13-88.

3. Sato, G. H. The role of serum in cell culture. In: Litwack, G., ed. Biochemical actions of hormones, vol. III. New York: Academic Press; 1975: 391-396.

4. Tsao, M. C.; Walthall, B. J.; IIam, R. G. Clonal growtll of normal human epidermal keratinocytes in a defined medium. J. Cell. Physiol. 110: 219-229; 1982.

5. Buyce, S. T.; Ham, R. G. Calcium-regulated differentiation of normal human epidermal keratinocytes in chemically defined clonal culture and serum-free serial culture. J. Invest. Derma- tol. 81: 33s-40s; 1983.

6. Lechner, J. F.; llaugen, A.; McClendon, I. A.; Pettis, E. W. Clonal growth of normal adult human bronchial epithelial cells in a serum-free medium. In Vitro 18: 633-642; 1982.

7. Hammond, S. L.; Ham, R. G.; Stampfer, M. R. Defined medium for normal human mammary epithelial cells. In Vitro 19: 252; 1983.

8. Yuspa, S. H.; Hawley-Nelson, P.; Stanley, J. R.; ltennings, It. Epidermal cell culture. Transplant. Proc. 12: 114-122; 1980.

9. Lechner, J. F.; Babcock, M. S.; Marnell, M.; Narayan, K. S.; Kaighn, n . E. Normal human prostate epithelial cell cultures. In: Harris, C. C.; Trump, B. F.; Stoner, G. D., eds. Methods

SERUM-FREE UROTHELIAL CELL CULTURE 171

in cell biology, vol. 2lB. New York: Academic Press; 1980: 195-225.

10. Peehl, D. M.; Ham, R. G. Clonal growth of human keratinocytes with small amounts of dialyzed serum. In Vitro 16: 525-~--A-0; 1980.

11. Kaighn, M. E.; Narayan, K. S.; Ohnuki, Y.; Lechner, J. F.; Jones, L. W. Establishment and characterization of a human prostatic carcinoma cell line (PC-3L Invest. Urol. 17: 11-23; 1979.

12. Phillips, D. M. Detection of mycoplasmal infection of cell cultures by electron microscopy. In: McGarrity, G. J.; Murphy, D. G.; Nichols, W. W., eds. Myeoplasma infection of cell cultures. New York: Plenum Press; 105-118; 1978.

13. Ohnuki, Y.; Marnell, M. M.; Babcock, M. S.; Lechner, J. F.; Kaighn, M. E. Chromosomal analysis of human prostatic adenocarcinoma cell lines. Cancer Res. 40: 524-534; 1980.

14. Yamase, H. T.; Powell, G. T.; Kosf, L. G. A simplified method of preparing permanent tissue sections for the erythrocyte adherence test. Am. Soc. Clin. Pathol. 75: 178-181; 1981.

15. Rheinwald, J. G. Serial cultivation of normal human epithelial keratinocytes. In: Harris, C. C.; Trump, B. F.; Stoner, G. D., eds. Methods in cell biology, vol. 21A, New York: Academic Press; 1980: 229-254.

16. Ham, R. G. Importance of the basal nutrient medium in the design of hormonally defined media. In: Sato, G. H.; Pardee, A. B.; Sirbasku, D. A., eds. Cold Spring Harbor conferences on cell proliferation, vol. 9. New York: Cold Spring Harbor Laboratory; 1982: 39-64.

17. Watt, F. M.; Green, H. Stratification and terminal differentia- tion of cultured epidermal cells. Nature 295: 434-436; 1982.

18. Babcock, M. S.; Marino, M. R.; Gunning, W. T.; Stoner, G. D. Clonal growth and serial propagation of rat esophageal epithelial cells. In Vitro 19: 403-415; 1983.

19. Lechner, J. F.; McClendon, I. A.; LaVeck, M. A.; Shamsuddin, A. M.; Harris, C. C. Differential control by platelet factors of squamous differentiation in normal and malignant human bronchial epithelial cells. Cancer Res. 43: 5915-5921: 1983.

20. Reznikoff, C. A.; Johnson, M. D.; Norback, D. It.,; Bryan, G. T. Growth and characterization of normal human urothelium in vitro. In Vitro 19: 326-343; 1983.

We fully acknowledge and appreciate the help of Dr. M. Edward Kaighn ~National Cancer Institute), who initiated this work. We thank Drs. John F. Lechner and Richard G. Ham for invaluable advice and for disclosing results before publication. Dr. B. M. Bishai's help with some experiments is gratefully acknowledged. We also thank M. Laura Hughes and Mr. Ray Pun for help with the photography. This work was supported by a grant from the National Cancer Institute (R01CA25089L Bethesda, MD.