Differentiation of Erythroleukemic Cells In Vitro: Irreversible Induction by Dimethyl Sulfoxide (DMSO) 1 HARVEY D PREISLER Z T 3 AND MIRIAM GILADI Mount Sinaz School of Medicine, New York, New York
ABSTRACT The inclusion of DMSO in the media of suspension cultures of Friend erythroleukemia cells results in the erythroid differentiation of these cells. The studies reported here were directed towards answering two questions. (1) How long an exposure to DMSO is necessary to induce the differentiation of these cells; and (2) What is the fate of the differentiating cells when DMSO is removed from the medium. Exposure to DMSO for less than 24 hours failed to produce any detectable evidence of erythroid differentiation. On the other hand, culture in the presence of DMSO for 24 hours followed by culture in DMSOfree medium for four additional days produced a small but detectable increment in the proportion of benzidine positive cells in the culture. Once the differentiation of an individual cell was initiated, the process continued after removal of DMSO from the medium. The cell became progressively more differentiated as evidenced by increases in the intensity of benzidine staining as well as in the rate of heme synthesis and heme content. However, when cells which had been induced to differentiate by DMSO were cultured in DMSO-free medium for more than 3-4 days, they became vacuolated and apparently died. This latter phenomenon, as well as the more rapid proliferation of the undifferentiated cells in the culture, accounts for the observation that when new cultures are established fYom cultures which have been grown in the presence of DMSO for several days, the culture which results ultimately contains only differentiated cells.
Friend erythroleukemia cells growing in observation that even the most mature suspension culture can be induced to dif- cells in differentiating culture produced ferentiate along the erythroid pathway by leukemia virus (Sato et al., '71). The studthe addition of DMSO to their culture ies reported 'here demonstrate that the difmedia (Friend et al., '71). While complete ferentiation of these leukemic cells is a maturation to the level of the erythrocyte stable process which continues whether does not occur, many of the features of or not the inducer of differentiation is renormal erythroid maturation are demon- moved from the media. strable including the synthesis of heme (Friend et al., '71), globin (Boyer et al., METHODS '72, Ostertag et al., '72; Friend et al., '74), Cell cultures globin mRNA (Ross et al., '72; Preisler Friend erythroleukemia cells (clone et al.,'73a,b), as well as the development of responsiveness to erythropoietin (Preis- 745A) were used in these studies. These ler et al., '73) and the ability to form eryth- suspension culture cells were grown in roid colonies in vitro (Goldstein et al., '74). Eagle's basal medium supplemented with Initially Friend et al. ('71) reported that fetal calf serum to 15% (v/v) at 37°C in a exposure to DMSO for as little as one hour humidified atmosphere consisting of 5 % resulted in the subsequent differentiation COe and room air. The cells were seeded of these leukemic cells. However, a later at a concentration of 1 X 105 cells/ml and report (Friend et al., '73) Taised the pos- routinely passed every three or four days. sibility that removal of DMSO from the Received June 7, '74. Accepted Sept. 30, '74. medium might result in the dedifferentia1 This work was supported by USPHS Grant CA-12737, tion of the already differentiating. leukemic National Cancer Institute. Scholar of the Leukemia Society of America, Inc. cells. The stability of the differentiated Present address: Department of Medicine A, Roswell state was also brought into question by the Park Memorial Institute, Buffalo, New York 14203. J.
CELL PHYSIOL,85. 537-546.
HARVEY D. PREISLER AND MIRIAM GILADI
To induce erythroid differentiation, DMSO (2% v/v) was included in the culture media at the time of seeding.
Cell counts and morphologic studies The cells were counted in a hemocytometer. A cytocentrifuge was used to make slides for the morphologic studies. The slides were stained with benzidine (Stephenson et al., '71) and counterstained with Wright-Giemsa. A single observer rated the benzidine positivity of the cells on each slide according to the following criteria; negative = blue or light gray cytoplasm, 2 or weakly positive = gray-yellow to light yellow cytoplasm, 4 + or strongly positive = golden cytoplasm. Five hundred cells were studied on each slide. All of the slides from a single experiment were studied on the same day and the observer did not know the identity of any slide until all the slides in the group had been rated.
Assay of the rate of heme synthesis To assess the rate of heme synthesis, 12 ml was removed from a culture and distributed equally into three tissue culture plates. One-half pCi of 59Fe bound to mouse transferrin was added to each plate and after four hours of incubation the cells were harvested, washed twice with isotonic saline, and the heme extracted according to the method of Krantz (Krantz, '65). J9Fe incorporation into the heme was used as an indicator of the rate of heme synthesis by these cells. This technique has been demonstrated to provide an accurate assessment of heme synthesis by these erythroleukemic cells (Scher et al., '71). Quantitation of h e m e present in cell cultures Cells were pelleted by centrifugation and the supernatant discarded. The cells were then resuspended in buffer (NaC1 0.01 M , MgClz.6Hz0 0.003 M, Tris.HC1 0.01 M pH 7.4) and triton X 100 was added to 1% . After the lysate was centrifuged at 12,000 RPM in a Sorval refrigerated centrifuge to pellet the nuclei, the supernatant was removed and two dilutions made (usually 1:2 and 1:4) and 0.1 ml of each was assayed for heme content as described by Cartwright (Cartwright, '68). The heme content of each sample was determined using a sample of blood with a
known hemoglobin content as the standard. The mean of the calculated heme content of the undiluted specimen and of the two dilutions was calculated and used as the true heme value. RESULTS
Sequential studies (Ag. 1a,b) The purpose of these studies was to determine the shortest explosure time to DMSO which was necessary for the induction of differentiation of these cells. In the studies reported in figure l a cells were seeded in the presence of DMSO and aliquots were removed at daily intervals, slides made and proportion of benzidine positive cells determined. 180 160
-- t $140
120 ; ; D
"0 I 2 3 4 5 6 0 1 2 3 4 5 6 DAYS IN CULTURE
DAYS IN CULTURE PRIOR TO T R A IU S FE R
Fig. 1 Effect of growth in the presence of DMSO on the differentiation of Friend erythroleukemia cells. A, the appearance of benzidine positive cells when DMSO was present throughout the entire culture period. The proportion of benzidine positive cells in cultures to which DMSO was not added was always less than 1% . B, correlation between the duration of growth in the presence of DMSO and the degree of differentiation of the culture. Cells were cultured in the ]presence of DMSO for 1-4 days and then transfer:red to and grown in DMSO-free medium for 1 4 days so that the total duration of culture in each case was five days. The abscissa gives the duration of culture in the presence of DMSO and the ordinate gives the proportion of benzidine positive cells and the rate of heme synthesis on the fifth day of culture. Hence the longer the initial period of exposure to DMSO, the greater the degree of differentiation on day 5 of culture.
IRREVERSIBLE DIFFERENTIATION OF LEUKEMIC CELLS
In the studies illustrated in figure l b cells were seeded in the presence of DMSO and, at daily intervals through day four, 20 cc of cells were harvested by centrifugation, the supernatant carefully removed and the cells resuspended in 13 cc of fresh DMSO-free medium. Approximately 0.1 ml of the original DMSO-containing medium was present overlying the cell pellet and could not be removed prior to resuspension of the cells. Thus, the resuspension of the cells in fresh medium resulted in a dilution of the remaining DMSO by 1:130 resulting in a final DMSO concentration of 0.015% (v/v) in the cell suspension. This concentration of DMSO had no effect on the differentiation of these cells. In preliminary studies the cells which were growing in the presence of DMSO were washed once with medium prior to transfer to fresh medium. The washing was discontinued when we found that there was no change in the results when this step was omitted. Cells were harvested on the fifth day after the original seeding and an aliquot removed and used to determine the proportion of benzidine positive cells. The rest of the cells were resuspended in fresh medium and the rate of heme synthesis was measured as described above. Figure 1a illustrates the appearance of benzidine positive cells in suspension cultures which contain DMSO. During the first two days of culture there was no change in the proportion of benzidine positive cells. Between the end of the second and end of the third days of culture the proportion of benzidine positive cells increased from 2% to 46%. After the end of the third day of culture the proportion of benzidine positive cells continued to increase but the rate of increase was less than that which had occurred during the third day of culture. The majority of the benzidine positive cells in the culture were weakly positive on day 3 of culture. Thereafter the proportion of weakly positive cells declined while the proportion of strongly benzidine positive cells increased throughout the culture period. While the sudden appearance of benzidine positive cells which occurred during the third day of culture suggested that at this time erythroleukemic cells were uniquely sensitive to DMSO, inspection of
figure l b demonstrated, however, that this was not the case. The degree of differentiation of the culture five days after the initial seeding in the presence of DMSO was directly proportional to the duration of exposure to DMSO prior to transfer to DMSOfree medium. This was the case whether one considered the proportion of benzidine positive cells in the culture (for example 12% vs 56% for exposure to DMSO for 1 day or 4 days respectively), or the rate of heme synthesis on the fifth day (for example 20 CPM/lOfi cells for 2-day exposure vs 70 CPM/lOficells for 3 days exposure). Exposure of these cells to DMSO for less than 24 hours did not reproducibly result in any evidence of subsequent erythroid differentiation, whether measured by the proportion of benzidine positive cells or by the rate of heme synthesis by the culture. Day 2 studies (fig. 2, table 1) The studies illustrated in figure l b demonstrated that under the conditions employed the proportion of benzidine positive cells in the culture increased during growth in DMSO-free medium. While this suggested that DMSO was no longer necessary once an individual cell began to differentiate, the day 2 studies were undertaken to directly test this possibility. Cells were seeded in 40 ml cultures in the presence of DMSO. Two days later, the cells were harvested by centrifugation and resuspended in 40 ml of fresh medium as described above. Aliquots were removed daily for counting and morphological studies, and the entire culture was harvested five days after the original culture was seeded. In addition to cell counts and morphologic studies the amount of heme present in the culture and the rate of heme synthesis by the cells were measured on day 2 and day 5. Controls for these studies included cells grown in the presence or absence of DMSO for five days without a medium change as well as cells cultured in DMSO-free medium for two days and then transferred to DMSO-free fresh medium on day 2 and grown for three additional days. Figure 2a demonstrates that the total number of cells increased in an almost exponential fashion for the first three days after transfer to fresh medium. Figure 2b
HARVEY D. PREISLER AND MIRIAM GILADI
t NUMBER OF CELLS
BENZlDlNE POSITIVE CELLS (%1 TOTAL
-*---.. 2+ 4+
G I I I I I
NUMBER OF CELLS
0 1 2 3 4
D A Y S IN FRESH MEDIUM Fig. 2 The kinetics of appearance of differentiated cells after growth in the presence of DMSO for two days. The abscissa indicates the duration of culture in DMSO-free medium subsequent to the initial period of growth in the presence of DMSO. A , changes in the total number of cells during culture in fresh medium. B, alterations in the total proportion of benzidine positive cells as well as the proportion of strongly (4 ) and weakly (2+) benzidine positive cells during growth in fresh medium. C, alterations in the absolute number of benzidine positive cells. Note the substantial increase in the number of differentiating cells during culture in DMSO-free medium.
illustrates that the proportion of benzidine positive cells increased during the first day of growth in fresh medium from 1% to 32% and then slowly declined. While the decline in the total proportion of benzidine positive cells was parallelled by a decline in the proportion of weakly benzidine positive cells, the proportion of strongly benzidine positive cells increased from 0 % to 13% during the first three days of the transfer to fresh medium. Figure 2c illustrates that despite the decline in the proportion of benzidine positive cells noted in figure 2b, the absolute number of differentiating cells in the culture actually increased from 3 x 103 to 7 X 1 0 5 during the first three days of culture in fresh medium. Of particular interest was the observation that the absolute number of weakly benzidine positive cells increased during the first three days of growth in DMSO-free medium from 3 X 10" to 4 X 107. While this increase could have resulted either from the continuedl appearance of differentiated cells and/or replication of the weakly benzidine positive cells, the former probably played the major role since mitosis of differentiating cells was relatively uncommon. While the possibility that a minority of differentiating cells underwent dedifferentiation has not been absolutely ruled out, it is apparent that the decline in the proportion of benzidine positive cells after the first daly of growth in DMSO-free medium was ;I manifestation of the fact that the number of undifferentiated cells was increasing at a more rapid rate than was the number of differentiating cells. Table 1 reports the effect of growth in the presence or absence of DMSO for two days followed by growth in fresh medium for three days on the rate of heme synthesis and the amount of heme present in these cultures. While there was no detectable increase in the proportion of benzidine positive cells in cultures after growth in the presence of DMSO €or two days, there was a doubling of the amount of hemell06 cells and a 5-8 fold increase in the rate of heme synthesis/lOG cells. Cultures which had been exposed to DMSO for two days and then grown for three days in fresh medium contained 4-7 times as much heme as did control cultures which were either grown continuously in DMSO-free medium for five days
IRREVERSIBLE DIFFERENTIATION OF LEUKEMIC CELLS
T h e &ect of incubation w i t h DMSO f o r two d a y s o n t h e rate of heme synthesis trnd on t h e heme content of erythroleukemic cell cultures Duration of growth in DMSO-containing medium (days)
Duration of growth in DMSO-free media (days)
heme/ 10 ml
Exp. I Exp. I 1
Exp. I Exp. II
Exp. I Exp. II
Exp. I Exp. I1
cLg heme/ 1 0 8 cells
0.285 0.749 0.389 1.38 0.337 4.43
0.210 0.446 0.351 1.07 0.086 4.00
Group I 0
2 0 2 0 5 0 2 0
2 0 5 3
5' 0 2 0
0.71 0.52 1.9 5.0 2.75 1.6
1.02 0.65 1.7 2.8 5.1 1.4
19 76 5 63 11
11 55 2 85 12 199
15 81 1 20 2 69
7 75 1 108
2.6 3.2 7.4 33.0 9.1 71
17 1 79
2.2 2.9 6.0 29.8 4.4 56
Group 11 20 68 7 470
12 54 5 484
I Grown in DMSO-free medium for two days and then hansferred to DMSO-free fresh medium for three additional days. In Group I the 59Fe was added directly to the culture being assayed. In Group I1 cultures the cells were resuspended in fresh medium and preincubated for one hour before the addition of "Fe. 2 The cells present in 10 ml of suspension culture were pelleted by centrifugation and their heme content measured a s described in METHODS.
or which were grown in DMSO-free medium for two days followed by culture in fresh medium for three additional days. Calculation of heme content per 106 cells gave similar results. Measurement of 5gFe incorporation into heme revealed even greater differences between the rate of synthesis of cells exposed to DMSO for two days and those which were never exposed to DMSO. However, since iron pool sizes were not measured, the rate of 59Feincorporation can not be used as an absolute indicator of the true rate of heme synthesis. Naturally, culture which had been exposed to DMSO for only two days synthesized less and contained less heme than did cultures exposed to DMSO for five days. Two additional observations should be mentioned. First, the transfer of cells to fresh medium after two days of growth (whether in the presence or absence of DMSO) resulted in a much higher cell density after three additional days of growth than one usually sees after five continuous days of culture in the same medium under standard conditions. Secondly, the rate of heme synthesis per 106 cells was significantly higher in DMSOtreated cultures which were transferred to fresh medium one hour prior to the addition of 5"e (table 1).
This latter observation raised the possibility that serial transfer of these cells to fresh medium might result in a higher degree of differentiation than had been heretofore observed, and hence led to the next series of studies. Day 4 studies (figs. 3 , 4, table 2) Twenty ml cell cultures were grown in the presence of DMSO for four days, harvested and resuspended either in 20 or 75 ml of fresh medium. Aliquots were removed at daily intervals for counts and morphologic studies. At the time of transfer to fresh medium and at three and six days after transfer, the total amount of heme in the cultures was determined. Figure 3 illustrates that while the proportion of benzidine positive cells in the culture declined after transfer, the absolute number of benzidine positive cells was unchanged for three to four days and then subsequently declined. For these three to four days both the proportion of and the number of weakly benzidine positive cells declined while the reverse was true for the 4 benzidine positive cells. After three days of growth in fresh medium, many of the benzidine positive cells became vacuolated. The degree of vacuolation increased to the point that some of
HARVEY D. PREISLER AND MIRIAM GILADI
.--. NUMBER CELLS
BENZlDlNE POS IT1V E CELLS
NUMBER OF CELLS
0 1 2 3 4 5 6
0 1 2 3 4 5 1 6
D A Y S IN F R E S H M E D I U M Fig. 3 Effect of transfer of cultures containing significant numbers of differentiated cells to DMSO-fkee medium. Erythroleukemia cells which had been grown in the presence of DMSO for four days were resuspended in fresh medium at high (I) or low (11) cell densities and grown fnr six additional days. A, changes in total cell counts during culture. B , changes in the proportion of benzidine positive cells. C, alterations in the absolute number of benzidine positive cells in the cultures. While there was no change in the absolute number of benzidine positive cells in the culture, during the first 3 4 days of growth in fresh medium, the number of strongly benzidine positive cells increased.
the cells consisted of a thin rim of cytoplasm surrounding either multiple vacuoles or a single large vacuole (fig. 4). Because of the scant amount of cytoplasm, it was very difficult, particularly after five and six days of culture in fresh medium, to decide whether some cells were weakly or strongly benzidine positive. Simultaneous with the appearance of vacuolization, the total number of benzidine positive cells
and the number of strongly benzidine positive cells in the culture began to decrease. This decline was reflected .in a decrease, between days 3 and 6 of growth in fresh medium, in the amount of heme present in the culture (table 2). In the same cultures, undifferentiated cells either were not vacuolated or, in occasional instances, had one or two very small vacuoles. The progressive increase in the number
IRREVERSIBLE DIFFERENTIATION OF LEUKEMIC CELLS
Fig. 4 Vacuolization of differentiating erythroleukemic cells. Cells were cultured for two days in medium containing DMSO and then transferred to and grown in fresh DMSO-free medium for three days. The markedly vacuolated cells in the center and on the right were strongly benzidine positive and contained pyknotic nuclei. A portion of the vacuolated cytoplasm of another benzidine positive cell is seen in the extreme lower right-hand corner.
of strongly benzidine positive cells during the first three days of growth in fresh medium was reflected in a doubling in the amount of heme present in the culture (table 2). Similarly, the decline in the number of benzidine positive cells between days 3 and 6 of culture was reflected in the halving of the amount of heme present in the culture between days 3 and 6.
transferred to fresh medium. Under conditions of culture in which the cells were continuously exposed to DMSO, the earliest detectable evidence of differentiation was noted at the end of the second day of culture when there was an increase in the rate of heme synthesis in the culture and in the amount of heme present. In previous studies we were unable to detect any increase in the rate of heme synthesis DISCUSSION after continuous culture in the presence Several phenomena demonstrate that of DMSO for one day (unpublished obseronce initiated by DMSO, the differentia- vations). These observations are in accord tion of Friend erythroleukemia cells was with studies of the appearance of globin not dependent upon the continued pres- (Friend et al., '74), globin mRNA (Friend ence of DMSO. These phenomena included: et al., '74; Ross et al., '72) and the develop (1) the continued appearance of benzidine ment of responsiveness to erythropoietin positive cells in cultures growing in the (Preisler and Zanjani, '73). Morphologic absence of DMSO but which had been evidence of differentiation was not detectgrown in the presence of DMSO for two able until the end of the third day of culdays prior to transfer to DMSO-free me- ture. Studies utilizing growth in the presdium, and (2) the increase in the degree ence of DMSO followed by transfer to and of differentiation of the individual cells in growth in DMSO-free media have demoncultures which had been grown in the strated that some cells are irreversibly inpresence of DMSO for four days and then duced to differentiate after growth in the
HARVEY D. PREISLEH AND MIRIAM GILADI TABLE 2
Chanyes zn the h m e content of the cultures illustrated in figure 3
‘X change hemoglobin
per culture % S.E. _____
ln=6. 2 n = 3. Duration of culture in fresh medium after resuspension of cells previously cultured for four days in the presence of DMSO. The total amount of heme in the culture was measured at three different times: (1) at the time the cells were resuspended in fresh medium, (2, after three days of growth in fresh mediurn, and (3) after six days of growth in fresh medium. The 55 change in hemecontent was calculated as follows: (1) For three days culture in fresh medium: amount of heme present on day 3 - amount of heme present at time of initiation of culture in fresh medium x 100 amount of heme present at the time of initiation of culture in fresh medium ( 2 ) For six days culture in fresh medium: amount of heme present on day 6 - amount of heme present on day 3 of culture in fresh medium x 100 amount of heme present on day 3 of culture in fresh medium
presence of DMSO for as little as a single day. Furthermore, these studies suggest that the proportion of cells which were induced to differentiate was directly related to the duration of growth in the presence of DMSO. It was not clear why some cells took longer to be induced than others, but perhaps a cell must traverse a sensitive portion of the cell cycle while in the presence of DMSO if induction of differentiation is to occur. A previous study (Friend et al., ’71) reported that exposure to DMSO for one hour was sufficient to “trigger” hemoglobin synthesis when assayed four days later. We were unable to reproduce this observation even if we exposed the cells to DMSO for five hours and measured the proportion of benzidine positive cells or the rate of heme synthesis four or five days after exposure to DMSO. In our procedure, virtually no DMSO was present during the secondary culture of these cells. In the cited study, however, after the cells were exposed to DMSO they were transferred along with their DMSO-containing culture media to fresh media perhaps raising the concentration of DMSO in the fresh medium to levels which could support a small but significant degree of differentiation. When cells were cultured for four days in DMSO-containing medium and then grown for three additional days in fresh medium, there was a doubling in the amount of heme present in the culture.
This was probably a reflection of an increase in the heme content of those cells which were weakly benzidine positive at the time of transfer to fresh medium and which became strongly positive during culture in fresh medium. No cells, however, evidenced a higher degree of nuclear maturity and no cells matured to the erythrocyte level. Rather, with continued culture the benzidine positive cells became vacuolated and their numbers began to decline. This decline was probably a result of cell death since the degree of vacuolization of many of the cells appeared to be incompatible with life. Hence the reappearance of undifferentiated cells when cells from differentiated cultures are used to seed fresh cultures (Friend et al., ’71) results from two phenomena: (1) the death of some of the differentiatiated cells, and (2) the more rapid rate of replication of the undifferentiated cells as compared with the differentiated cells. It would therefore appear that these erythroleukemic cells can manifest some characteristics of erythroid differentiation but cannot, under these conditions, become mature erythrocytes. Perhaps the persistence of the leukemia virus in these cells prevents complete maturation. LITERATIJRE CITED Boyer, S. H . , K . D. Wuu, A. N . Moyes, R. Young, W. Scher, C. Friend, H. D. Preisler and A. Bank 1972 Hemoglobin biosynthesis in murine virus-
IRREVERSIBLE DIFFERENTIATION OF LEUKEMIC CELLS induced leukemic cells in vitro: structure and amounts of globin chains produced. Blood, 40: 823-835. Cartwright, G. 1968 Diagnostic Laboratory H e matology. Fourth ed. Grune and Stratton, New York, pp. 318-322. Friend, C., H. D. Preisler a n d W. Scher 1973 Studies on the control of differentiation of mur i n e virus-induced erythroleukemic cells. In: Current Topics in Developmental Biology. Vol. 8. A. Monroy and A. A. Moscona, e d s . Academic Press, New York, pp. 81-101. Friend, C., W. Scher, J. G. Holland a n d J . Sato 1971 Hemoglobin synthesis in murine virusinduced leukemic cells i n vitro: stimulation of erythroid differentiation by dimethyl sulfoxide. Proc. Nat. Acad. Sci. (U.S.A.), 68: 378-382. Friend, C., W . Scher and H. D. Preisler 1974 Hemoglobin biosynthesis in murine virus-induced leukemia cells in vitro. Ann. N. Y. Acad. Sci., 241 : 582-589. Goldstein, K., H. D. Preisler, J. D. Lutton and E. D. Zanjani 1974 Erythroid colony formation in vitro by dimethyl sulfoxide treated erythroleukemic cells. Blood, 44: 831-836. Krantz, S. B. 1965 T h e effect of erythropoietin o n h u m a n bone marrow cells in vitro. Life Sci., 4: 2393-2397. Ostertag, W., H. Melderes, G. Steinheider, G . Kluge a n d S. Duhe 1972 Synthesis of mouse haemoglobin a n d globin mRNA in leukaemic cell cultures. Nature (New Biol.), 239 231-234.
Preisler. H. D., D. Housman, W. Scher and C. Friend 1973a Effects of 5-bromo-2’-deoxyurid i n e on production of globin, messenger RNA by dimethyl sulfoxide-stimulated Friend leukemia cells. Proc. Nat. Acad. Sci. (U.S.A.), 70: 2956-2959. Preisler, H. D., W. Scher and C. Friend 1973b Polyribosome profiles and polyribosome associated RNA of Friend leukemia cells following DMSO-induced differentiation. Differentiation, 1 : 27-37. Preisler, H. D., and M. Giladi 1974 Erythropoietin responsiveness of differentiating Friend leukemia cells. Nature, 251 : 645. Ross, J., Y. Ikawa and P. Leder 1972 Globin messenger RNA induction during erythroid differentiation of cultured leukemic cells. Proc. Nat. Acad. Sci. (U.S.A.), 69: 3620-3623. Sato, T., C. Friend and E. d e Harven 1971 U1trastructural changes in Friend erythroleukemia cells treated with dimethyl sulfoxide. Can. Res., 31: 1402-1417. Scher, W . , J . G . Holland and C. Friend 1971 Hemoglobin synthesis in murine virus-induced leukemic cells in vitro. I. Partial purification a n d identification of hemoglobins. Blood, 37: 428437. Stephenson, J. R., A. A . Axelrad, D. I,. McLeod and M. M. Shreeve 1971 Induction of colonies of hemoglobin-synthesizing cells by erythropoietin in vitro. Proc. Nat. Acad. Sci. (U.S.A.), 68: 1542.
Note added in proof: We have recently found that when cells are incubated for two days in the presence of DMSO and then transferred to DMSO-free media which contains bromodeoxyuridine, the subsequent differentiation of the cells is inhibited. Hence, the continued differentiation of cells after exposure to DMSO and growth in DMSO-free media appears to be dependent upon the continued synthesis of globin mRNA.