Proc. Natl. Acad. Sci. USAVol. 77, No. 5, pp. 2936-2940, May 1980Medical Sciences

Induction of differentiation of the human promyelocytic leukemiacell line (HL-60) by retinoic acid*

(retinoids/myeloid differentiation/hematopoiesis)

T. R. BREITMAN, STUART E. SELONICKt, AND STEVEN J. COLLINSfLaboratory of Tumor Cell Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20205

Communicated by Marshall Warren Nirenberg, January 31, 1980

ABSTRACT The HL-60 cell line, derived from a patient withacute promyelocytic leukemia, proliferates continuously insuspension culture and consists predominantly (>90%) ofpromyelocytes. These cells can be induced to differentiate tomorphologically and functionally mature granulocytes by in-cubation with a wide variety of compounds, including butyrateand hypoxanthine and polar planar compounds such as dimethylsulfoxide and hexamethylene bisacetamide. We have now foundthat retinoic acid (all-trans-retinoic acid) induces differentiation(as measured morphologically and by the ability to reduce ni-troblue tetrazolium) of HL-60 at concentrations as low as 1 nM.Maximal differentiation (approximately 90%) occurs at 1 MuM,a concentration 1/500th to 1/160,000th the concentrations ofbutyrate (0.5 mM) and dimethyl sulfoxide (160 mM) that promotea similar increase in differentiation. Continuous exposure toretinoic acid is necessary for optimal differentiation, with thepercentage of mature cells in the culture directly related to thelength of time of exposure to retinoic acid. Retinoic acid and13-cis-retinoic acid are equally effective in inducing differen-tiation of HL-60. Retinol (vitamin A), retinal, and retinyl acetateare approximately 1/1000th less potent. This study suggests thatretinoids could provide a therapeutic tool in the treatment ofacute myeloid leukemia, a disease that has been looked uponas primarily involving a block in myeloid differentiation, andindicates that retinoids, in addition to their well-characterizedinvolvement in epithelial cell differentiation, may also be in-volved in the differentiation of certain hematopoietic cells.

Retinoids (vitamin A and its natural and synthetic analogs)influence the normal differentiation of epithelium. Retinoiddeficiency in rats leads to premalignant lesions in a wide varietyof epithelial tissue that are reversed by the administration ofretinoids (1). Retinoids inhibit carcinogenesis in vio (2, 3) andsuppress malignant transformation in vitro (4, 5). Certain ofthese compounds inhibit the growth of transformed cells (6, 7)and induce differentiation of mouse embryonal carcinoma cellsin vitro (8-10). The antineoplastic effects of retinoids suggestthat these compounds could be used therapeutically for thechemoprevention of cancer (11).The HL-60 cell line, derived from a patient with acute

promyelocytic leukemia (12), proliferates continuously insuspension culture and consists predominantly of promyelo-cytes. These cells manifest a transformed phenotype, as shownby karyotypic abnormalities, continuous proliferation for over3 years in culture, tumorigenicity in nude mice, and morpho-logical, histochemical, and karyotypic similarity to the patient'sleukemic cells (13). However, these cells can be induced toterminally differentiate to morphologically mature granulo-cytes by incubation with a wide variety of compounds, in-cluding butyrate, hypoxanthine, actinomycin D, and polarplanar compounds such as dimethyl sulfoxide (Me2SO) andhexamethylene bisacetamide (HMBA) (14, 15). Moreover, these

induced HL-60 cells have many of the functional characteristicsof normal peripheral blood granulocytes, including phagocy-tosis, complement receptors, chemotaxis, and the ability toreduce nitroblue tetrazolium (NBT) (14, 16, 17). Thus, this cellline provides a unique system for studying human myeloiddifferentiation in vitro.

In the present report we describe the induction of terminaldifferentiation of HL-60 cells by retinoic acid. This compoundinduces differentiation of HL-60 cells at concentrations 1/500thto 1/160,000th the concentrations of other inducers. This studysuggests a new approach to the therapy of some human leuke-mias and indicates that retinoids may be involved in the dif-ferentiation of certain hematopoietic tissues.

MATERIALS AND METHODSCells. The HL-60 cell line has been maintained in continous

suspension culture for almost 3 years (12); passages 15 through50 were used for these experiments. Cells (2 X 105 cells per ml)were grown in Corning polystyrene tissue culture flasks con-taining growth medium consisting of equal volumes of Dul-becco's modified Eagle's minimal essential medium and Ham'sF12 medium and supplemented with 1.2 g of NaHCO3 perliter, 15 mM Hepes, and 10% (vol/vol) heat inactivated fetalcalf serum (Flow Laboratories, McLean, VA). Substitution ofRPMI 1640 (GIBCO) supplemented with 10% inactivated fetalcalf serum for the above medium gave identical results. Cellcounts were determined with a Coulter Counter (CoulterElectronic, Hialeah, FL), and viability was assessed by trypanblue dye exclusion. Morphologic assessment of the cells wasmade on Cytospin slide preparations stained with Wright-Giemsa. Differential counts were then performed under lightmicroscopy on a minimum of 200 cells.

Chemicals. Retinoic acid, retinol, retinal, and retinyl acetatewere obtained from Sigma and stored at -20°C; 13-cis-retinoicacid was a gift of Hoffman-La Roche. 12-O-Tetradecanoyl-phorbol 13-acetate, obtained from the National Institutes ofHealth Carcinogen Repository, was stored at -20°C in acetoneat 200 ,ug/ml. NBT was obtained from Aldrich. HMBA was agift of Roberta Reuben.

Induction of Differentiation. All retinoids were dissolvedin 95% (vol/vol) ethanol at concentrations varying from 10-3to 10-7 M, and the solutions were diluted 1000-fold into thegrowth medium such that the final ethanol concentration was

Abbreviations: Retinoic acid, all-trans-f3-retinoic acid; Me2SO, di-methyl sulfoxide; NBT, nitroblue tetrazolium; HMBA, hexamethylenebisacetamide.* A portion of this work was presented at the Annual Meeting of theAmerican Federation for Clinical Research, May 10-12, 1980,Washington, DC.

t Present address: Department of Medicine, Johns Hopkins Hospital,601 North Broadway, Baltimore, MD 21205.

i Present address: Department of Medicine, V.A. Hospital, 4435Beacon Avenue South, Seattle, WA 98108.

2936

The publication costs of this article were defrayed in part by pagecharge payment. This article must therefore be hereby marked "ad-vertisement" in accordance with 18 U. S. C. §1734 solely to indicatethis fact.

Dow

nloa

ded

by g

uest

on

Oct

ober

6, 2

021

Proc. Natl. Acad. Sci. USA 77 (1980) 2937

..a

FIG. 1. Morphology of uninduced and retinoic acid-inducedHL-60 cells. Cytospin slide preparations of suspension cell culturesstained with Wright-Giemsa. (X1100.) (Upper) Cells cultivatedwithout inducer consist of promyelocytes with characteristic cyto-plasmic granules, large nuclei, and prominent nucleoli. (Lower) Cellscultured in 1 ,M retinoic acid for 5 days consist of metamyelocytesand banded and segmented neutrophils.

no higher than 0.1%. Varying the concentration of ethanol from0.1% to 0.0001% had no effect on the extent of retinoic acid-induced differentiation. Cells were incubated at 37°C in ahumidified atmosphere of 5% CO2 in air. All manipulationswere performed in subdued light.NBT Reduction. NBT reduction was assayed as described

(16). Cells (2 X 106 per ml of medium) were incubated for 25min at 37°C with an equal volume of 0.2% NBT dissolved inDulbecco's phosphate-buffered saline containing 200 ng offreshly diluted tetradecanoylphorbol acetate per ml. The per-

0060_/_60~~~~~~~~~

0~~~~~~0~~~~~~~~

40-

0

20- 0

X O// 10 7'

-*Control

0 2 4 6Time in culture, days

FIG. 2. Differentiation of HL-60 cells induced by various con-centrations of retinoic acid. Cells were seeded at 2 X 105 per ml.Numbers identifying each curve are the molar concentrations of ret-inoic acid. Differentiation was determined by NBT reduction.

centage of cells containing intracellular blue-black formazandeposits was then determined on Wright-Giemsa-stained Cy-tospin slide preparations. At least 200 cells were assessed.

RESULTSMorphological Changes in HL-60 Cells Induced by Reti-

noic Acid. In the absence of an inducer, cultured HL-60 cellsare predominantly promyelocytes, with 3-10% of the cells moremature granulocytes (myelocytes, metamyelocytes, and bandedand segmented neutrophils). Retinoic acid induced strikingchanges in the morphology of HL-60 cells characteristic ofterminal differentiation of myeloid cells (Fig. 1). After 6 daysof incubation in 1 ,uM retinoic acid, over 90% of HL-60 cellsresembled mature granulocytes. The extent of this differen-tiation as assessed by differential counts indicates that withretinoic acid there was a "shift to the right" when comparedwith other HL-60 inducers, including Me2SO and. sodiumbutyrate (Table 1). Thus, retinoic acid induces relatively ex-tensive morphological differentiation of HL-60 at concentra-tions some 1/500th to 1/160,000th the concentrations of otherinducers.

Table 1. Differential counts of HL-60 after incubation with various inducing compoundsMyeloid cell type, % of total cells

Banded SegmentedInducer Myeloblasts Promyelocytes Myelocytes Metamyelocytes neutrophils neutrophils

None 3 84 12 1 0 0Me2SO (160 mM) 0 4 63 14 10 9HMBA (2.0 mM) 0 9 27 41 18 5Butyrate (0.5 mM) 1 13 70 15 1 0Retinoic acid (1 IAM) 0 6 46 18 13 17

HL-60 cells were incubated under standard conditions with the indicated concentration of inducer. After 6 days, differentialcounts were performed on Wright-Giemsa-stained Cytospin slide preparations of the cell suspensions.

Medical Sciences: Breitman et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

6, 2

021

2938 Medical Sciences: Breitman et al.

Functional Changes in HL-60 Cells Induced by RetinoicAcid. NBT, a water-soluble dye, is reduced to insoluble intra-cellular blue-black formazan by phagocytizing or mem-brane-stimulated granulocytes (18). Morphologically matureHL-60 cells, but not the immature promyelocytes, reduce NBTwhen stimulated by tetradecanoylphorbol acetate (16, 17). TheNBT reaction provides a sensitive and easily quantitated HL-60differentiation marker, eliminating observer subjectivity as-sociated with morphological assessment alone. After 5 days ofincubation in various concentrations of retinoic acid, the per-centage of HL-60 cells reducing NBT varied with retinoic acidconcentration, reaching a maximum of 95% at 1 ,M (Fig. 2).HL-60 cells treated with retinoic acid also exhibited the ca-pacity to phagocytose Candida albicans (data not shown).Decreased HL-60 Cell Growth in Presence of Retinoic

Acid. With the addition of retinoic acid, alterations in the rateof cell growth were apparent by 3 days (Fig. 3). With 1 AMretinoic acid, growth ceased by day 4. These cells no longerproliferated even when resuspended in growth medium withoutadded retinoic acid. Thus, as expected, morphologically andfunctionally mature HL-60 cells induced by retinoic acid havealso lost the capacity for further proliferation. The high per-centage of viable cells (Fig. 3) eliminated the possibility thatall or a major part of the retinoic acid-induced differentiationwas a result of selective enrichment for differentiated cells.

Induction of HL-60 Differentiation by Other Retinoids.Retinoic acid and 13-cis-retinoic acid were equally effectiveand superior to other retinoids in inducing differentiation ofHL-60 (Fig. 4). Retinol (vitamin A), retinal, and retinyl acetatewere approximately 1/1000th as potent as retinoic acid.

' ' / Control20 - lo-9-

10-8

0(86%)7E 7-n {6%

6 o- % 010-6_6 0(70%)

05

4

3 -

2

, ,0 2 4 6

Time in culture, daysFIG. 3. Growth of HL-60 cells in various concentrations of reti-

noic acid. Cell counts were determined with a Coulter Counter.Numbers identifying each curve are the molar concentrations of ret-inoic acid. Numbers in parentheses are percentage of viable cells onday 5 as determined by trypan blue exclusion.

to

6060

0

40-

a-

20-

10-9 lo-8 10 10-6Retinoid concentration, M

FIG. 4. Differentiation of HL-60 cells induced by various con-centrations of retinoids. Cells were seeded at 2 X 105 per ml. Differ-entiation was determined by NBT reduction after 5 days of growthin the indicated concentration of retinoid. 0, All-trans-retinoic acid;0, 13-cis-retinoic acid; ,, retinol; o, retinyl acetate; *, retinal.

Exposure Time Necessary for Optimal HL-60 Differen-tiation. To determine how long an exposure to retinoic acid wasrequired for HL-60 cells to differentiate, we incubated cellswith 1 ,uM retinoic acid for various time intervals, washed them,and then resuspended them in fresh growth medium withoutretinoic acid. After a total of 5 days of incubation, the per-centage of cells reducing NBT was assessed. Continuous ex-posure to retinoic acid was necessary for optimal differentiation,with the percentage of mature cells in the culture directly re-lated to the time of retinoic acid exposure (Fig. 5). In otherexperiments, an increase in differentiation was noted after only3 hr of exposure.

DISCUSSIONRetinoic acid induces morphological and functional terminaldifferentiation of thehuman promyelocytic leukemia cell lineHL-60. Maximal differentiation (approximately 90%) occursat a retinoic acid concentration of 1 ,uM, a concentration 1/500th to 1/160,000th the concentrations of sodium butyrate (0.5mM), HMBA (2.0 mM), and Me2SO (160 mM) that promote asimilar increase in differentiation (Table 1). Moreover, theextent of morphological differentiation of retinoic acid-treatedHL-60 cells is as great as that found with any other inducer(Table 1).

Although other retinoids increase HL-60 cell differentiation,retinoic acid is clearly superior to retinol (vitamin A), retinal,and retinyl acetate (Fig. 4). This compound has been shown tohave the same high activity with respect to retinol analogs inother culture systems, including reversal of keratinization inretinoid-deficient tracheal epithelium (19), inhibition of growthof transformed cells (6), and enhanced production of plas-minogen activator in mouse embryonal carcinoma cells (8, 10).HL-60 cells may provide a similar convenient in vitro system

Proc. Natl. Acad. Sci. USA 77 (1980)

Dow

nloa

ded

by g

uest

on

Oct

ober

6, 2

021

Proc. Natl. Acad. Sci. USA 77 (1980) 2939

Cuca0

co

4-c

._

0a

60F

40F

20h

0 2 4 6Time of retinoic acid exposure, days

FIG. 5. Differentiation of HL-60 cells after various exposure timesto retinoic acid. Cells (2 X 105 per ml) were incubated in 1 .M retinoicacid for the indicated period of time, washed, and resuspended ingrowth medium without added retinoic acid. After 5 days of totalincubation, differentiation was determined by NBT reduction. 0 and0, Data from two independent experiments.

for assessing the activity of various natural and synthetic reti-noids and for studying the biological effects of retinoid antag-onists.The mechanism by which retinoic acid enhances differen-

tiation and exerts its antineoplastic effect on HL-60 and othercells is unknown. An appealing possibility is that in HL-60 ret-inoic acid suppresses the activity of an endogenously producedpolypeptide growth factor in a manner analogous to theblockage by retinoids of the phenotypic cell transformationproduced by sarcoma growth factor in normal rat kidney fi-broblasts (20). Other mechanisms of action, such as exerting aneffect in the nucleus after binding to a highly specific cyto-plasmic receptor (cellular retinoic acid-binding protein) (21)and promoting glycosylation of membrane glycoproteins (22),cannot be ruled out. It is also unknown whether the mechanismof induction by retinoic acid is the same as that of polar com-pounds such as Me2SO and HMBA; in initial experiments theeffects of combinations of retinoic acid and Me2SO were ad-ditive and not synergistic, suggesting that these two inducersmay act by similar pathways.The ability of retinoic acid to convert proliferating HL-60

leukemic promyelocytes to terminally differentiated, func-tionally mature granulocytes suggests that this compound couldprovide a new therapeutic tool in the treatment of acute mye-loid leukemia, a disease that has been looked upon as primarilyinvolving a block in myeloid differentiation (23). Indeed, 13-cis-retinoic acid has been used for treatment of certain skindisorders (24, 25). If the in vitro studies described here reflectpotential behavior of this drug in vivo, then serum levels as highas 1 ttM might be necessary over a prolonged period of time toinduce maximal differentiation of the leukemic cells. Althoughthis level is pharmacologically obtainable (26, 27), it is unknown

what toxicity would result from prolonged in vivo exposure.In comparison to other retinoids, retinoic acid is more toxic totracheal cartilage in organ cultures (19). However, 13-cis-ret-inoic acid is less toxic than retinoic acid (28, 29) and in clinicalstudies 13-cis-retinoic acid has shown relatively little toxicity(24, 25).Leukemic cells other than HL-60 may not be susceptible to

induction of differentiation by retinoic acid. In preliminaryexperiments we have not been able to induce the productionof hemoglobin in Friend erythroleukemic cells or in K-562 (30)(a cell line derived from a patient with chronic myelogenousleukemia in blast crisis) with the addition of retinoic acid. Bothof these cell lines produced hemoglobin when incubated withthe appropriate inducer, Me2SO for Friend leukemic cells (31)and butyrate for K-562 (32). Retinoic acid also did not inducegranulocytic differentiation of KG-1, a cell line derived froma patient with acute myelogenous leukemia (33). Thus, HL-60is unusual among leukemic cells in exhibiting such a pro-nounced differentiative response to retinoic acid.

Although HL-60 cells are of leukemic origin, the morpho-logical, functional, and biochemical similarities between thechemically induced HL-60 cells and normal peripheral bloodgranulocytes (16, 17) suggest that differentiation induction ofHL-60 cells reflects molecular events occurring in normalgranulocytic maturation. Thus, it is possible that retinoids, inaddition to their well characterized involvement in epithelialcell differentiation, are also involved in the differentiation ofcertain hematopoietic cells.

We are grateful to Dr. R. C. Gallo for his support and to Beverly R.Keene and Ersell Richardson for their excellent technical assistance.Kathy White is thanked for her expert secretarial assistance.

1. Wolbach, S. B. & Howe, P. R. (1925) J. Exp. Med. 42, 753-777.

2. Moon, R. C., Grubbs, C. J., Sporn, M. B. & Goodman, D. G. (1977)Nature (London) 267,620-621.

3. Bollag, W. (1974) Eur. J. Cancer 10, 731-737.4. Harisiadis, L., Miller, R. C., Hall, E. J. & Borek, C. (1978) Nature

(London) 274, 486-487.5. Merriman, R. L. & Bertram, J. S. (1979) Cancer Res. 39,

1661-1666.6. Lotan, R. & Nicolson, G. L. (1977) J. Natl. Cancer Inst. 59,

1717-1722.7. Meyskens, F. L., Jr. & Salmon, S. E. (1979) Cancer Res. 39,

4055-4057.8. Strickland, S. & Mahdavi, V. (1978) Cell 15,393-403.9. Jetten, A. M. & Jetten, M. E. R. (1979) Nature (London) 287,

180-182.10. Rizzino, A. & Crowley, C. (1980) Proc. Natl. Acad. Sci. USA 77,

457-461.11. Sporn, M. B., Dunlop, N. M., Newton, D. L. & Smith, J. M. (1976)

Fed. Proc. Fed. Am. Soc. Exp. Biol. 35, 1332-1338.12. Collins, S. J., Gallo, R. C. & Gallagher, R. E. (1977) Nature

(London) 270,347-349.13. Gallagher, R., Collins, S., Trujillo, J., McCredie, K., Ahearn, M.,

Tsai, S. Metzgar, R., Aulakh, G., Ting, R., Ruscetti, F. & Gallo,R. (1979) Blood 54, 713-733.

14. Collins, S. J., Ruscetti, F. W., Gallagher, R. E. & Gallo, R. C.(1978) Proc. Natl. Acad. Sci. USA 75,2458-2462.

15. Collins, S. J., Bodner, A., Ting, R. & Gallo, R. C. (1980) Int. J.Cancer 25, 213-218.

16. Collins, S. J., Ruscetti, F. W., Gallagher, R. E. & Gallo, R. C.(1979) J. Exp. Med. 149, 969-974.

17. Newberger, P., Chovaniec, M., Greenberger, J. & Cohen, H.(1979) J. Cell Biol. 82, 315-322.

18. Baehner, R. L., Boxer, L. A. & David, J. (1976) Blood 48, 309-313.

19. Sporn, M. B., Dunlop, N. M., Newton, D. L. & Henderson, W.R. (1976) Nature (London) 263, 110-113.

I0

0

80k

i

Medical Sciences: Breitman et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

6, 2

021

2940 Medical Sciences: Breitman et al.

20. Todaro, G. J., DeLarco, J. E. & Sporn, M. B. (1978) Nature(London) 276, 272-274.

21. Chytil, F. & Ong, D. E. (1979) Fed. Proc. Fed. Am. Soc. Exp. Biol.38,2510-2514.

22. DeLuca, L. M., Bhat, P. V., Wiodzimierz, S. & Adamo, S. (1979)Fed. Proc. Fed. Am. Soc. Exp. Biol. 38,2535-2539.

23. Gallo, R. C. (1973) in Modern Trends in Human Leukemia II,eds. Neth, R., Gallo, R., Spiegelman, S. & Stohlman, F., Jr. (J. F.Lehmanns, Munich, West Germany), pp. 227-237.

24. Peck, G. L., Olsen, T. G., Yoder, F. W., Strauss, J. S., Downing,D. T., Pandya, M., Butkus, D. & Arnaud-Battandier, J. (1979) N.Engl. J. Med. 300,329-333.

25. Peck, G. L., Olsen, T. G., Butkus, D., Pandya, M., Arnaud-Bat-tandier, J., Yoder, F. & Lewis, W. R. (1979) Proc. Am. Assoc.Cancer Res. 20, 225 (abstr.).

Proc. Nati. Acad. Sci. USA 77 (1980)

26. Frolik, C. A., Tavela, T. E., Peck, G. L. & Sporn, M. B. (1978)Anal. Biochem. 86,743-750.

27. Puglisi, C. V. & DeSilva, J. A. F. (1978) J. Chromatog. 152,421-430.

28. Hixson, E. J. & Denine, E. P. (1978) Toxicol. Appl. Pharmacol.44,29-40.

29. Nettesheim, P. & Williams, M. L. (1976) Int. J. Cancer 17,351-357.

30. Lozzio, C. B. & Lozzio, B. B. (1975) Blood 45,321-334.31. Friend, C., Scher, W., Holland, J. G. & Soto, T. (1971) Proc. NatI.

Acad. Sci. USA 68,378-382.32. Andersson, L. C., Mikko, J. & Gahmberg, C. G. (1979) Nature

(London) 278, 364-365.33. Koeffler, H. P. & Golde, D. W. (1978) Science 200, 1153-

1154.

Dow

nloa

ded

by g

uest

on

Oct

ober

6, 2

021