- Email: [email protected]
Effect of minoxidil on pre- and postconfluent keratinocytes
Phartnacology and preclinical studies Effect of minoxidil on pre- and postconfluent keratinocytes Joseph Kubilus, Ph.D., Joseph C. Kvedar, M.D., and Howard P. Baden, M.D. Boston, MA Studies in our laboratory have shown that minoxidil prolongs the life of keratinocytes in culture and extends the time after confluence that cells can be subcultured. These data suggest that the drug reduces the rate at which cells are lost from the germinative pool and hence slows senescence. In a dose-response study with minoxidil, the maximal effect of the drug was seen at doses from 6 to 12 j.Lg/ml; however, activity could be detected at doses below 1 j.Lg/mI. Cells subcultured during log growth failed to demonstrate that minoxidil increased the total number of generations attainable under these conditions, although as expected, epidermal growth factor extended the life span of cells. When the experiments were repeated in a keratinocyte cell line that does not require a fibroblast feeder layer, the same results were obtained, indicating that the difference observed between log phase and postconfluence growth cannot be explained by the presence of fibroblasts. Minoxidil's effect on postconfluent cells was blunted by the addition of cholera toxin to the medium, suggesting that elevation of cyclic adenosine monophosphate cannot be a mechanism, although reduction of cyclic adenosine monophosphate is a possibility. Finally, maintaining keratinocytes at a 20 to 40 mM calcium concentration greatly reduced the ability of postconfluent cells to be subcultured, in comparison with the normal calcium concentration of 2 mM. That minoxidil almost completely reversed this inhibitory effect suggests it may work by preventing cross-linking by transglutaminase, which is activated by elevated calcium concentrations. (J AM ACAD DERMATOL 1987;16:648-52.)
Minoxidil, whether taken orally for the treatment of hypertension or applied to the skin, has been shown to increase the thickness and length of hair. 1-3 It is currently being developed as a treatment for androgenetic alopecia, although there are reports that it may also be useful in the management of alopecia areata. 4 Minoxidil's mechanism of action on the hair follicle is unknown, but the From the Harvard Medical School Department of Dermatology, Massachusetts General Hospital. Reprint requests to: Dr. Howard P. Baden, Harvard Medical School Department of Dermatology, Massachusetts General Hospital, Warren 5, Boston, MA 02114.
648
drug's vasodilation property suggests that increased hair growth is brought about by enhanced blood flow. Indeed, minoxidil does increase blood flow when applied to the skin, but it does so at higher concentrations than are needed to stimulate hair growth. s However, such studies were unable to measure blood flow to the hair follicle. There have been two reports that minoxidil may directly influence the growth of keratinocytes. 6 ,7 We have previously reported6 that minoxidil enhanced the survival of keratinocytes in culture and increased the length of time that cells could be held at conflUence before being passaged (subcultured). This effect on keratinocytes could be con-
Volume 16 Number 3, Part 2 March 1987
Effect of minoxidil on pre- and postcorifluent keratinocytes
Table I. Effect of different concentrations of minoxidil on the passage of normal human keratinocytes* (/Lg/ml)
No. of cells per 35-mm dish x 10-5
72.0 18.0 12.0 6.0 2.0 0.8 0.0
2.42 2.91 5.63 5.81 3.24 3.32 0.00
Minoxidil
*Second-passage cells were maintained at confluence for 4 weeks with different concentrations of minoxidil and then plated into fresh medium in the absence of minoxidil. The total number of cells were counted after 5 days. Data are the mean of three 35-rnm dishes.
sidered analogous to maintaining hair bulb cells in the anagen phase, thus allowing them to grow for longer periods, one of the therapeutic responses observed with minoxidil treatment. Cohen et aF have also shown an effect of minoxidil on cultured murine keratinocytes. In their report, minoxidil treatment of primary cultures resulted in the emergence of cells with altered morphology that could be repassaged. They concluded that minoxidil could act as a mitogen for keratinocytes. In the present study, we further explored minoxidil's effect on cultured cells and showed that its action differed on cells in the log and postconfluent phases of growth. Possible mechanisms of action are proposed, particularly minoxidil's effect on calcium metabolism. MATERIALS AND METHODS Epidennal cells were isolated from newborn human infant foreskins and cultured on mitomycin C-treated murine 3T3 cells with the use of Dulbecco's modified Eagle's medium containing 20% fetal calf serum with 0.4 fLg/m1 of hydrocortisone and 50 !-Lg/ml of gentamicin, as described previously.8,9 In addition, an immortalized line of human keratinocytes (NORAH 1) was used. 10 These cells could be passaged nine times in the absence of 3T3 feeder cells. Epidermal growth factor (Collaborative Research, Inc., Lexington, MA), cholera toxin (Sigma Chemical Co., St. Louis, MO), and minoxidil (provided by The Upjohn Company, Kalamazoo, MI) were dissolved directly in culture medium, and these solutions, after sterilization by filtra-
649
Table II. Growth of keratinocytes with and without minoxidil or epidermal growth factor* No. of cells per 35-mm dish x 10- 6 Passage
Control
Minoxidil
EGF
1 2 3
1.31 1.23 1.02 1.04 0.63 0.17
1.22
1.24 1.03 0.91 0.76 0.28
1.33 1.22 1.42 1.25 1.13 1.26 1.34 1.27
4
5 6 7 8
EGF: Epidermal growth factor. *The experiment was initiated with normal human keratinocytes obtained from primary cultures that had reached confluence. Each subculture was allowed to grow for 7 days. The cells were then suspended with trypsin and EDTA and counted electronically. One hundred thousand keratinocytes per 35-mm dish were used to initiate the next passage irrespective of the total number of cells obtained. Data are the mean of three 35-mm dishes. The concentration of minoxidil was 6 fLg/ml and of EGF, 10 ng/m!.
tion, were added. to the cells as needed for a particular experiment. These drugs were always added 3 days after plating to be certain that the number of cells that sat down were the same in control and treated cultures. At the time of passage, cultures were treated with 0.02% ethyleriediaminetetraacetic acid (EDTA) to remove fibroblasts, and keratinocytes were then removed in 0.05% EDTA and 0.05% trypsin, centrifuged, and suspended in medium. Subcultures were initiated with 100,000 keratinocytes per 35-mm dish with or without 140,000 mitomycin C-treated 3T3 cells. Growth of cultures was detennined by electronic counting of cell suspensions, RESULTS In our earlier studies on the length of time human keratinocytes could be held at confluence and then passaged, the concentration of minoxidil used was 20 IJ-g/ml. For determination of the optimal concentration of minoxidil for this effect, a doseresponse study was undertaken (a representative experiment is shown in Table I). Optimal drug concentration was between 6 and 12 IJ-g/ml. However, an effect was observed at less than 1 I-1g/ml. Similar results were found with four different cell lines.
650
Journal of the American Academy of Dennatology
Kubilus et at
Table ITI. Effect of minoxidil and epidermal growth factor on growth of NORAH 1 cells in the absence of 3T3 cells*
Table IV. Effect of minoxidil on passage of normal human keratinocytes in the presence of cholera toxin* No. of cells per 35-mm dish x 10- 5
No. of cells per 35·mm dish x 10. 0 Culture conditions
Control Minoxidil
EGF
se 8
SC 1
0.92 1.24 1.93
1.05 0.82 1.87
0.62 0.51 1.24
EOF: Epidermal growth factor; SC: subculture. *NORAH I cells in passage 54 were plated in the presence of 10 ng/ml BOF and 10- 9 M cholera toxin but without 3T3 feeder cells. They were then repeatedly subcultured in these conditions. After the first, third, and eighth subcultures, aliquots of the cells were grown in cholera toxin-free medium in the absence (control) or presence of minoxidil or EOF. After 7 days the cells were suspended and counted. Data are the mean of three 3S-mm dishes. The concentration of minoxidil was 6 JLg/ml.
Since enhanced survival occurred when cells were maintained at confluence, it was of interest to determine whether cells could be carried for an increased number of generations by passing them just before confluence (i.e., during the log phase of growth). Since epidermal growth factor was shownto have this effect, it was used as a positive contro1. ll As shown in Table II, minoxidil did not increase the number of times cells could be passed in comparison with control cultures. Epidermal growth factor, on the other hand, did allow an increased number of passages, and the experiment was stopped after the eighth passage, although the cells were growing vigorously. Identical results were observed in four different cell lines. The 3T3 feeder layers were present in the cultures listed in Table II but not in cultures from the earlier studies, showing that minoxidil prolonged the time that cells could be maintained at confluence. To determine whether this apparent difference in response was related to the presence of 3T3 cells, we used a human keratinocyte line that could be grown in the absence of 3T3 cells for at least nine passages. As shown in Table III, this system is an easy test of the capacity of cells to remain in the germinative pool, as evidenced by epidermal growth factor's increasing the number of cells present after 7 days of growth. Results
Culture conditions
Experiment 1
Experiment 2
Control medium Minoxidil Minoxidil + CT CT
1.92 3.40 1.98 1.02
.92 2.10 .85 .48
CT: Cholera toxin. *1n experiments 1 and 2, different lines of second-passage cells were maintained at confluence for 4 weeks in the media indicated and then plated into control medium containing neither rninoxidil nor cholera toxin. The number of cells were counted after 5 days. Data are the mean of three 35-mm dishes. The concentration of minoxidil was 6 JLg/ml and of cholera toxin 10- 9 M.
with minoxidil were identical to those with control medium. Thus minoxidil did not appear to maintain cells in the germinative pool during the log phase of growth. Since the effect of minoxidil appeared to be restricted to cells that had passed log-phase growth and were at confluence, we directed our attention tei possible mechanisms at that stage of growth. The addition of cholera toxin to human cultured keratinocytes raises cyclic adenosine monophosphate levels and reduces the doubling time for cells in log-phase growth. 12 When cholera toxin was maintained in cultures during confluence, it had an adverse effect on the ability of cells to be passed (Table IV). Cholera toxin reduced the number of cells in the cultures that were passed after being held at confluence for 4 weeks, in both the presence and the absence of minoxidil. That this result is an effect of elevated cyclic adenosine monophosphate levels rather than of cholera toxin itself is suggested by a similar but less pronounced effect when isoproterenol is used in place of cholera toxin (data not shown). The inhibitory effect of cholera toxin could be prevented by removing it as the cells approach confluence. However, experiments directed at learning how long it was necessary for the drug to be present before its effect became irreversible were inconclusive. Since an elevated calcium concentration has been shown to enhance terminal differentiation of
Volume 16 Number 3, Part 2 March 1987
Effect of minoxidil on pre- and postconfluent keratinoCytes
651
cultured keratinocytes, 13,14 we studied cells grown at different calcium concentrations in the presence or absence of minoxidil. In this culture system, cells grew poorly at concentrations below 2 mM of calcium, an effect not altered by the addition of minoxidil. Cells grown in control medium did well at 2, 5, 10, '20, and 40 mM of calcium but varied in their capacity to be passaged (Table V). In the presence of minoxidil, better growth was observed at all concentrations of calcium. Minoxidil was particularly effective at higher concentrations of calcium, where marked inhibition was observed in the control cultures. The experiment was duplicated with the same results.
Table V. Effect of minoxidil on passage of hllman keratinocytes grown at varying calcium concentrations*
DISCUSSION
*Second-passage cells were maintained at confluence at the various calcium concentrations for 3 weeks. They were then plated into fresh medium in the absence of minoxidil in 2 mM calcium and the number of cells counted after 5 days. All minoxidil-treated plates grew to confluence with longer incubation, but none of the control plates except the 2 mM calcium did so. Data are the mean of three 35-mm dishes. The concentration of minoxidil was 6 /Lg/ml.
Results of a previous report suggested that minoxidil, like epidermal growth factor, delays the senescence of cells, as reflected in the prolonged time cells could be held at confluence and continue to be passaged. 6 It is not clear why holding keratinocytes at confluence induces senescence, since the same is not true for other types of cells such as fibroblasts. However, senescence can be prevented in keratinocytes grown at very low calcium concentrations, where they do not stratify but grow as a monolayer. 13.14 How or whether stratification is the key factor in the aging of confluent cells, or whether some other factor is involved, remains to be determined, The optimal concentration at which minoxidil retarded senescence was between 6 and 12 J.1g/ml, although some effect could be detected at less than 1 J.1g/ml. A similar concentration was reported to be mitogenic for mouse keratinocytes. 7 Although we have not determined whether any of the minoxidil was bound to fetal calf serum, such binding does not occur in human ser.um, That minoxidil does not extend the number of times cells can be passed during log-phase growth was an unexpected observation and suggests that the drug does not keep more cells in a viable state, a result opposite from that with cells studied at postconfluence. With epidermal growth factor the slowing of senescence was seen in cells studied both in log-phase growth and at confluence, a finding that favors a common site of action in both situations. All these data further support the hy-
No. of celIs :t: SD per 35-mm dish x 10.- 5
Control
2 5 10
Minoxidil
20 40 2 5 10
20 40
3.45 3.42 3.05 2.53 1.35 8.30 8.26 8.14 7.36 7.32
± ± ± ± ± ± ± ± ± ±
0.25 0.27 0.24 0,26 0.22 0.52 0.49 0,53 0.47 0.50
pothesis that minoxidil and epidermal growth factor exert their effects by means of different mechanisms. Conclusions derived from culture conditions cannot readily be extrapolated to explain the action of minoxidil in vivo. However, the postconfluent state of cell development is in many ways similar to the anagen phase of hair growth. Therefore it is tempting to speculate that minoxidil works by prolonging the time of anagen growth, which would then lead to the production of longer hairs. The difference in the effect of cholera toxin on keratinocytes in the log and postconfluent phases was striking, Green l2 suggested "that in log-phase growth, cholera toxin could extend its effect by shortening the doubling time of the dividing fraction or by retarding the onset of senescence. The difference we observed between log and postconfluent phase growth may be related to a different effect of cyclic adenosine monophosphate at these two phases, or it may be that cholera toxin does not elevate cyclic adenosine monophosphate levels in postconfluent cells. We have not directly measured cyclic adenosine monophosphate in the cells and cannot unequivocally distinguish between these possibilities. However, isoproterenol, which
Journal of the American Academy of Dermatology
652 Kubilus et al
also raises cyclic adenosine monophosphate levels, had a similar effect. IS Since cholera toxin interferes with the effect of minoxidil on postconfiuent cultures, rninoxidil may not exert its effect through elevation of the cyclic nucleotide. Instead, rninoxidil may lower cyclic adenosine monophosphate levels, since its combined action with cholera toxin produces the same result as does control medium (Table IV). This topic requires additional study. Finally, minoxidil's protective effect against high calcium levels in culture medium (Table V) points to a possible mechanism of action. There is evidence that cross-linking by transglutaminase is one of the important steps in aging,16 and calcium is known to activate the enzyme. It is possible that minoxidil is able to retard senescence by interfering with an action of calcium. REFERENCES 1. Zappacosta AR. Reversal of baldness in a patient receiving minoxidiJ for hypertension [letter]. N Eng1 J Med 1980;303: 1480. 2. De Villez RL. Topical minoxidil therapy in hereditary androgenetic alopecia. Arch Dermato1 1985; 121:197202. 3. Olsen EA, Weiner MS, Delong ER, et a1. Topical minoxidiJ in early male pattern baldness. J AM ACAD DERMATOL 1985;13:185-92. 4. Fiedler-Weiss VC, West DP, Buys CM, et al: Topical rninoxidil dose-response effect in alopecia areata. Arch Dermatol 1986;122:180-2. 5. Wester RC, Maibach HI, Guy RH, et a1. Minoxidil stimulates cutaneous blood flow in human balding scalps: pharmacodynamics measured by laser Doppler velocimetry and photopulse plethysmography. J Invest Dermatol 1984;82:515-7. 6. Baden HP, Kubilus J. Effect of minoxidil on cultured keratinocytes. J Invest Dermatol 1983;81:558-60. 7. Cohen RL, Alves MEAF, Weiss VC, et al: Direct effects of minoxidil on epidermal cells in culture. J Invest Dermatol 1984;82:90-3.
8. Rheinwald JG, Green H. Serial cultivation of strains of human epidermal keratinocytes: the formation of keratinizing colonies from single cells. Cell 1975;6:331-43. 9. Kubilus J, Macdonald MJ, Baden HP. Epidermal proteins of cultured human and bovine keratinocytes. Biochim Biophys Acta 1979;578:484-92. 10. Baden HP, Kubi1us J, Kvedar J, et a1. Isolation and characterization of a spontaneously arising long-lived line of human keratinocytes (NORAH 1). In Vitro (in press). 11. Rheinwald JG, Green H. Epidermal growth factor and the multiplication of cultured human epidermal keratinocytes. Nature 1977;265:421-4. 12. Green H. Cyclic AMP in relation to proliferation of the epidermal cells: A new view. Ce111978;15:801-11. 13. Hennings H, Michael D, Cheng C, et al. Calcium regulation of growth and differentiation of mouse epidermal cells in culture. Cell 1980;19:245-54. 14. Boyce ST, Ham RG. Calcium-regulated differentiation of normal human epidermal kcratinocytes in chemically defined clonal culture and serum-free serial culture. J Invest Dermatol 1983;81(suppl):33-50. 15. Powell JA, Duell EA, Voorhees JJ. Beta-adrenergic stimulation of endogenous epidermal cyclic AMP formation. Arch Dermatol 1971;104:359-65. 16. Selkoe OJ, Ihara Y, Salazar FJ. Alzheimer'S disease: insolubility of partially purified paired helical filaments in sodium dodecyl sulfate and urea. Science 1982; 215:1243-5.
DISCUSSION
Question. Could minoxidil chelate calcium and/or alter its configuration? Could that explain its protective effect? Answer. I haven't thought about that part of it. It didn't occur to us that minoxidil might be a calcium chelator; however, it would be easy to look at that possibility. Calcium metabolism is a key pathway in the epidermis. It regulates a number of cellular metabolic pathways, turning things off and on, and it's not surprising that this effect occurs. The answer to your question is: I don't know, but it would be easy to find out.
Minoxidil, whether taken orally for the treatment of hypertension or applied to the skin, has been shown to increase the thickness and length of hair. 1-3 It is currently being developed as a treatment for androgenetic alopecia, although there are reports that it may also be useful in the management of alopecia areata. 4 Minoxidil's mechanism of action on the hair follicle is unknown, but the From the Harvard Medical School Department of Dermatology, Massachusetts General Hospital. Reprint requests to: Dr. Howard P. Baden, Harvard Medical School Department of Dermatology, Massachusetts General Hospital, Warren 5, Boston, MA 02114.
648
drug's vasodilation property suggests that increased hair growth is brought about by enhanced blood flow. Indeed, minoxidil does increase blood flow when applied to the skin, but it does so at higher concentrations than are needed to stimulate hair growth. s However, such studies were unable to measure blood flow to the hair follicle. There have been two reports that minoxidil may directly influence the growth of keratinocytes. 6 ,7 We have previously reported6 that minoxidil enhanced the survival of keratinocytes in culture and increased the length of time that cells could be held at conflUence before being passaged (subcultured). This effect on keratinocytes could be con-
Volume 16 Number 3, Part 2 March 1987
Effect of minoxidil on pre- and postcorifluent keratinocytes
Table I. Effect of different concentrations of minoxidil on the passage of normal human keratinocytes* (/Lg/ml)
No. of cells per 35-mm dish x 10-5
72.0 18.0 12.0 6.0 2.0 0.8 0.0
2.42 2.91 5.63 5.81 3.24 3.32 0.00
Minoxidil
*Second-passage cells were maintained at confluence for 4 weeks with different concentrations of minoxidil and then plated into fresh medium in the absence of minoxidil. The total number of cells were counted after 5 days. Data are the mean of three 35-rnm dishes.
sidered analogous to maintaining hair bulb cells in the anagen phase, thus allowing them to grow for longer periods, one of the therapeutic responses observed with minoxidil treatment. Cohen et aF have also shown an effect of minoxidil on cultured murine keratinocytes. In their report, minoxidil treatment of primary cultures resulted in the emergence of cells with altered morphology that could be repassaged. They concluded that minoxidil could act as a mitogen for keratinocytes. In the present study, we further explored minoxidil's effect on cultured cells and showed that its action differed on cells in the log and postconfluent phases of growth. Possible mechanisms of action are proposed, particularly minoxidil's effect on calcium metabolism. MATERIALS AND METHODS Epidennal cells were isolated from newborn human infant foreskins and cultured on mitomycin C-treated murine 3T3 cells with the use of Dulbecco's modified Eagle's medium containing 20% fetal calf serum with 0.4 fLg/m1 of hydrocortisone and 50 !-Lg/ml of gentamicin, as described previously.8,9 In addition, an immortalized line of human keratinocytes (NORAH 1) was used. 10 These cells could be passaged nine times in the absence of 3T3 feeder cells. Epidermal growth factor (Collaborative Research, Inc., Lexington, MA), cholera toxin (Sigma Chemical Co., St. Louis, MO), and minoxidil (provided by The Upjohn Company, Kalamazoo, MI) were dissolved directly in culture medium, and these solutions, after sterilization by filtra-
649
Table II. Growth of keratinocytes with and without minoxidil or epidermal growth factor* No. of cells per 35-mm dish x 10- 6 Passage
Control
Minoxidil
EGF
1 2 3
1.31 1.23 1.02 1.04 0.63 0.17
1.22
1.24 1.03 0.91 0.76 0.28
1.33 1.22 1.42 1.25 1.13 1.26 1.34 1.27
4
5 6 7 8
EGF: Epidermal growth factor. *The experiment was initiated with normal human keratinocytes obtained from primary cultures that had reached confluence. Each subculture was allowed to grow for 7 days. The cells were then suspended with trypsin and EDTA and counted electronically. One hundred thousand keratinocytes per 35-mm dish were used to initiate the next passage irrespective of the total number of cells obtained. Data are the mean of three 35-mm dishes. The concentration of minoxidil was 6 fLg/ml and of EGF, 10 ng/m!.
tion, were added. to the cells as needed for a particular experiment. These drugs were always added 3 days after plating to be certain that the number of cells that sat down were the same in control and treated cultures. At the time of passage, cultures were treated with 0.02% ethyleriediaminetetraacetic acid (EDTA) to remove fibroblasts, and keratinocytes were then removed in 0.05% EDTA and 0.05% trypsin, centrifuged, and suspended in medium. Subcultures were initiated with 100,000 keratinocytes per 35-mm dish with or without 140,000 mitomycin C-treated 3T3 cells. Growth of cultures was detennined by electronic counting of cell suspensions, RESULTS In our earlier studies on the length of time human keratinocytes could be held at confluence and then passaged, the concentration of minoxidil used was 20 IJ-g/ml. For determination of the optimal concentration of minoxidil for this effect, a doseresponse study was undertaken (a representative experiment is shown in Table I). Optimal drug concentration was between 6 and 12 IJ-g/ml. However, an effect was observed at less than 1 I-1g/ml. Similar results were found with four different cell lines.
650
Journal of the American Academy of Dennatology
Kubilus et at
Table ITI. Effect of minoxidil and epidermal growth factor on growth of NORAH 1 cells in the absence of 3T3 cells*
Table IV. Effect of minoxidil on passage of normal human keratinocytes in the presence of cholera toxin* No. of cells per 35-mm dish x 10- 5
No. of cells per 35·mm dish x 10. 0 Culture conditions
Control Minoxidil
EGF
se 8
SC 1
0.92 1.24 1.93
1.05 0.82 1.87
0.62 0.51 1.24
EOF: Epidermal growth factor; SC: subculture. *NORAH I cells in passage 54 were plated in the presence of 10 ng/ml BOF and 10- 9 M cholera toxin but without 3T3 feeder cells. They were then repeatedly subcultured in these conditions. After the first, third, and eighth subcultures, aliquots of the cells were grown in cholera toxin-free medium in the absence (control) or presence of minoxidil or EOF. After 7 days the cells were suspended and counted. Data are the mean of three 3S-mm dishes. The concentration of minoxidil was 6 JLg/ml.
Since enhanced survival occurred when cells were maintained at confluence, it was of interest to determine whether cells could be carried for an increased number of generations by passing them just before confluence (i.e., during the log phase of growth). Since epidermal growth factor was shownto have this effect, it was used as a positive contro1. ll As shown in Table II, minoxidil did not increase the number of times cells could be passed in comparison with control cultures. Epidermal growth factor, on the other hand, did allow an increased number of passages, and the experiment was stopped after the eighth passage, although the cells were growing vigorously. Identical results were observed in four different cell lines. The 3T3 feeder layers were present in the cultures listed in Table II but not in cultures from the earlier studies, showing that minoxidil prolonged the time that cells could be maintained at confluence. To determine whether this apparent difference in response was related to the presence of 3T3 cells, we used a human keratinocyte line that could be grown in the absence of 3T3 cells for at least nine passages. As shown in Table III, this system is an easy test of the capacity of cells to remain in the germinative pool, as evidenced by epidermal growth factor's increasing the number of cells present after 7 days of growth. Results
Culture conditions
Experiment 1
Experiment 2
Control medium Minoxidil Minoxidil + CT CT
1.92 3.40 1.98 1.02
.92 2.10 .85 .48
CT: Cholera toxin. *1n experiments 1 and 2, different lines of second-passage cells were maintained at confluence for 4 weeks in the media indicated and then plated into control medium containing neither rninoxidil nor cholera toxin. The number of cells were counted after 5 days. Data are the mean of three 35-mm dishes. The concentration of minoxidil was 6 JLg/ml and of cholera toxin 10- 9 M.
with minoxidil were identical to those with control medium. Thus minoxidil did not appear to maintain cells in the germinative pool during the log phase of growth. Since the effect of minoxidil appeared to be restricted to cells that had passed log-phase growth and were at confluence, we directed our attention tei possible mechanisms at that stage of growth. The addition of cholera toxin to human cultured keratinocytes raises cyclic adenosine monophosphate levels and reduces the doubling time for cells in log-phase growth. 12 When cholera toxin was maintained in cultures during confluence, it had an adverse effect on the ability of cells to be passed (Table IV). Cholera toxin reduced the number of cells in the cultures that were passed after being held at confluence for 4 weeks, in both the presence and the absence of minoxidil. That this result is an effect of elevated cyclic adenosine monophosphate levels rather than of cholera toxin itself is suggested by a similar but less pronounced effect when isoproterenol is used in place of cholera toxin (data not shown). The inhibitory effect of cholera toxin could be prevented by removing it as the cells approach confluence. However, experiments directed at learning how long it was necessary for the drug to be present before its effect became irreversible were inconclusive. Since an elevated calcium concentration has been shown to enhance terminal differentiation of
Volume 16 Number 3, Part 2 March 1987
Effect of minoxidil on pre- and postconfluent keratinoCytes
651
cultured keratinocytes, 13,14 we studied cells grown at different calcium concentrations in the presence or absence of minoxidil. In this culture system, cells grew poorly at concentrations below 2 mM of calcium, an effect not altered by the addition of minoxidil. Cells grown in control medium did well at 2, 5, 10, '20, and 40 mM of calcium but varied in their capacity to be passaged (Table V). In the presence of minoxidil, better growth was observed at all concentrations of calcium. Minoxidil was particularly effective at higher concentrations of calcium, where marked inhibition was observed in the control cultures. The experiment was duplicated with the same results.
Table V. Effect of minoxidil on passage of hllman keratinocytes grown at varying calcium concentrations*
DISCUSSION
*Second-passage cells were maintained at confluence at the various calcium concentrations for 3 weeks. They were then plated into fresh medium in the absence of minoxidil in 2 mM calcium and the number of cells counted after 5 days. All minoxidil-treated plates grew to confluence with longer incubation, but none of the control plates except the 2 mM calcium did so. Data are the mean of three 35-mm dishes. The concentration of minoxidil was 6 /Lg/ml.
Results of a previous report suggested that minoxidil, like epidermal growth factor, delays the senescence of cells, as reflected in the prolonged time cells could be held at confluence and continue to be passaged. 6 It is not clear why holding keratinocytes at confluence induces senescence, since the same is not true for other types of cells such as fibroblasts. However, senescence can be prevented in keratinocytes grown at very low calcium concentrations, where they do not stratify but grow as a monolayer. 13.14 How or whether stratification is the key factor in the aging of confluent cells, or whether some other factor is involved, remains to be determined, The optimal concentration at which minoxidil retarded senescence was between 6 and 12 J.1g/ml, although some effect could be detected at less than 1 J.1g/ml. A similar concentration was reported to be mitogenic for mouse keratinocytes. 7 Although we have not determined whether any of the minoxidil was bound to fetal calf serum, such binding does not occur in human ser.um, That minoxidil does not extend the number of times cells can be passed during log-phase growth was an unexpected observation and suggests that the drug does not keep more cells in a viable state, a result opposite from that with cells studied at postconfluence. With epidermal growth factor the slowing of senescence was seen in cells studied both in log-phase growth and at confluence, a finding that favors a common site of action in both situations. All these data further support the hy-
No. of celIs :t: SD per 35-mm dish x 10.- 5
Control
2 5 10
Minoxidil
20 40 2 5 10
20 40
3.45 3.42 3.05 2.53 1.35 8.30 8.26 8.14 7.36 7.32
± ± ± ± ± ± ± ± ± ±
0.25 0.27 0.24 0,26 0.22 0.52 0.49 0,53 0.47 0.50
pothesis that minoxidil and epidermal growth factor exert their effects by means of different mechanisms. Conclusions derived from culture conditions cannot readily be extrapolated to explain the action of minoxidil in vivo. However, the postconfluent state of cell development is in many ways similar to the anagen phase of hair growth. Therefore it is tempting to speculate that minoxidil works by prolonging the time of anagen growth, which would then lead to the production of longer hairs. The difference in the effect of cholera toxin on keratinocytes in the log and postconfluent phases was striking, Green l2 suggested "that in log-phase growth, cholera toxin could extend its effect by shortening the doubling time of the dividing fraction or by retarding the onset of senescence. The difference we observed between log and postconfluent phase growth may be related to a different effect of cyclic adenosine monophosphate at these two phases, or it may be that cholera toxin does not elevate cyclic adenosine monophosphate levels in postconfluent cells. We have not directly measured cyclic adenosine monophosphate in the cells and cannot unequivocally distinguish between these possibilities. However, isoproterenol, which
Journal of the American Academy of Dermatology
652 Kubilus et al
also raises cyclic adenosine monophosphate levels, had a similar effect. IS Since cholera toxin interferes with the effect of minoxidil on postconfiuent cultures, rninoxidil may not exert its effect through elevation of the cyclic nucleotide. Instead, rninoxidil may lower cyclic adenosine monophosphate levels, since its combined action with cholera toxin produces the same result as does control medium (Table IV). This topic requires additional study. Finally, minoxidil's protective effect against high calcium levels in culture medium (Table V) points to a possible mechanism of action. There is evidence that cross-linking by transglutaminase is one of the important steps in aging,16 and calcium is known to activate the enzyme. It is possible that minoxidil is able to retard senescence by interfering with an action of calcium. REFERENCES 1. Zappacosta AR. Reversal of baldness in a patient receiving minoxidiJ for hypertension [letter]. N Eng1 J Med 1980;303: 1480. 2. De Villez RL. Topical minoxidil therapy in hereditary androgenetic alopecia. Arch Dermato1 1985; 121:197202. 3. Olsen EA, Weiner MS, Delong ER, et a1. Topical minoxidiJ in early male pattern baldness. J AM ACAD DERMATOL 1985;13:185-92. 4. Fiedler-Weiss VC, West DP, Buys CM, et al: Topical rninoxidil dose-response effect in alopecia areata. Arch Dermatol 1986;122:180-2. 5. Wester RC, Maibach HI, Guy RH, et a1. Minoxidil stimulates cutaneous blood flow in human balding scalps: pharmacodynamics measured by laser Doppler velocimetry and photopulse plethysmography. J Invest Dermatol 1984;82:515-7. 6. Baden HP, Kubilus J. Effect of minoxidil on cultured keratinocytes. J Invest Dermatol 1983;81:558-60. 7. Cohen RL, Alves MEAF, Weiss VC, et al: Direct effects of minoxidil on epidermal cells in culture. J Invest Dermatol 1984;82:90-3.
8. Rheinwald JG, Green H. Serial cultivation of strains of human epidermal keratinocytes: the formation of keratinizing colonies from single cells. Cell 1975;6:331-43. 9. Kubilus J, Macdonald MJ, Baden HP. Epidermal proteins of cultured human and bovine keratinocytes. Biochim Biophys Acta 1979;578:484-92. 10. Baden HP, Kubi1us J, Kvedar J, et a1. Isolation and characterization of a spontaneously arising long-lived line of human keratinocytes (NORAH 1). In Vitro (in press). 11. Rheinwald JG, Green H. Epidermal growth factor and the multiplication of cultured human epidermal keratinocytes. Nature 1977;265:421-4. 12. Green H. Cyclic AMP in relation to proliferation of the epidermal cells: A new view. Ce111978;15:801-11. 13. Hennings H, Michael D, Cheng C, et al. Calcium regulation of growth and differentiation of mouse epidermal cells in culture. Cell 1980;19:245-54. 14. Boyce ST, Ham RG. Calcium-regulated differentiation of normal human epidermal kcratinocytes in chemically defined clonal culture and serum-free serial culture. J Invest Dermatol 1983;81(suppl):33-50. 15. Powell JA, Duell EA, Voorhees JJ. Beta-adrenergic stimulation of endogenous epidermal cyclic AMP formation. Arch Dermatol 1971;104:359-65. 16. Selkoe OJ, Ihara Y, Salazar FJ. Alzheimer'S disease: insolubility of partially purified paired helical filaments in sodium dodecyl sulfate and urea. Science 1982; 215:1243-5.
DISCUSSION
Question. Could minoxidil chelate calcium and/or alter its configuration? Could that explain its protective effect? Answer. I haven't thought about that part of it. It didn't occur to us that minoxidil might be a calcium chelator; however, it would be easy to look at that possibility. Calcium metabolism is a key pathway in the epidermis. It regulates a number of cellular metabolic pathways, turning things off and on, and it's not surprising that this effect occurs. The answer to your question is: I don't know, but it would be easy to find out.