Glucocorticoids in malignant lymphoid cells: Gene regulation and the minimum receptor fragment for lysis

J. Steroid Biochem.

Molec. Biol. Vol. 41, No. 3-8, pp. 273-282, 1992

0960-0760/92 $5.00 + 0.00

Printed in GreatBritain.All rightsreserved

Copyright© 1992PergamonPressplc

GLUCOCORTICOIDS IN MALIGNANT LYMPHOID CELLS: GENE REGULATION AND THE MINIMUM RECEPTOR FRAGMENT FOR LYSIS E. B. THOMPSON,*L. V. NAZARETH,R. THULASI,J. ASHRAF,D. HAgaotrg and B. H. JOHNSON University of Texas Medical Branch, Galveston, TX 77550, U.S.A. Summary--We have examined clones of human malignant lymphoid cells for markers that correlate with glucocorticoid-mediated cell lysis. In glucocorticoid-sensitive clones of CEM, a human T-ceU lymphoblastic leukemia line, two genes correlate with glucocorticoid-induced cell iysis. The glucocorticoid receptor (GR) itself is induced by standard glucocorticoids in sensitive clones and not in insensitive clones. The phenylpyrazolo-glucocorticoid cortivazol (CVZ) is capable of lysing several clones resistant to high concentrations of standard potent glucocorticoids. When these clones were tested for cortivazol responses, they were not only lysed by cortivazol but also showed induction of GR mRNA. Thus receptor induction appears to correlate with the lysis function of receptor in these cells. To determine what parts of the GR are required for lysis, we have mapped this function by transfecting and expressing GR and GR fragment genes in a GR-deficient CEM clone. Our results indicate that none of the known trans-activation regions of the GR are required. Removal of the steroid binding domain gives a fragment that is fully constitutive. Only one and one-half "Zn fingers" of the DNA binding region are required. We also find in CEM cells rapid suppression of the c-myc protooncogene, preceding growth arrest and cell lysis by glucocorticoids. This occurs only in clones possessing both intact receptors and lysis function. Thus the simple presence of GR alone is not sufficient to guarantee c-myc down-regulation. Introduction into the cells of c-myc driven by a promoter that does not permit suppression by glucocorticoids confers resistance to steroids. Furthermore, suppression of c-myc by antisense oligonucleotides also kills the cells. Therefore, c-myc appears to be a pivotal gene related both to ability of steroid to kill and to cell viability.

INTRODUCTION Glucocorticoids are important drugs in the treatment of certain leukemias and lymphomas. However, the basis for their use remains largely empirical. Better insight into the mechanisms by which they inhibit the growth of some types of lymphoid cells and markers are needed so as to predict their action. Simple quantitative measurement of GR, while of potential usefulness in certain types of lymphoid diseases, must be supplemented with evaluations of cell responses that indicate the functional inhibitory activity of those receptors. Such markers have proven extremely useful in other steroid controlled malignancies, such as breast cancer, in which measurement of estrogen receptor itself is complemented by such measurements of estrogen receptor function as induction of progesterone receptor or the pS2 gene. Proceedings of the lOth International Symposium of the Journal of Steroid Biochemistry and Molecular Biology, Recent Advances in Steroid Biochemistry and Molecular Biology, Paris, France, 26-29 May 1991.

*To whom correspondenceshould be addressed.

To study both glucocorticoid mechanism of action involved in the inhibition of lymphoid cell growth, and in an effort to identify markers for prediction of glucocorticoid activity, we have examined the expression of several genes in lines of malignant lymphoid cells. This report will focus on the G R and c-myc. Expression of the G R is regulated by glucocorticoids, and it may be induced or repressed in various cell types. We also are mapping the topology of the G R to establish the regions responsible for cell lysis of human leukemic cells. In addition we have found that the protooncogene c-myc, which is repressed by glucocorticoids, is important to cell viability. For these studies we have employed clones of the childhood acute lymphoblastic leukemia cell line CCRFCEM. In these cells, we find a correlation between cells whose growth is inhibited by glucocorticoids and the induction of the GR. In CEM clones, there is also a correlation between growth inhibition and down regulation of the c-myc oncogene. As to the functional topology

273

274

E. ]3. THOMPSONet al.

of the GR, we find that inhibition of cell growth and lysis can be achieved by fragments of the GR much smaller than the holoreceptor. Required is most, but not all of the DNA binding region specific for interaction with the cis-active DNA sequence known as the glucocorticoid response element (GRE). No other region of the GR is essential for evocation of CEM cell lysis, though one or more other regions of the receptor may augment the efficiency of the cell lysis function. MATERIALS AND METHODS

Materials Cortivazol (11/~,17~,21-trihydroxy-6,16~-dimethyl-2'-phenyl-2'-H-pregna-2,4,6-trieno [3,2c]pyrazol-20-one 21-acetate) was supplied by RousseI-UCLAF Co. (Paris, France) courtesy of Dr J. P. Raynaud. Dexamethasone(Dex) was obtained from Sigma Chemical Company. Other chemical and biochemical reagents were obtained from standard sources which have been listed previously [1, 2]. Cells Clones CEM-C7, ICR-27, 4R4, 3R43 and C1 were all cloned from CCRF-CEM, a cell line originally derived from a female child with acute lymphoblastic leukemia[3]. C7 and C1 were unselected parallel clones from the original line; the others are derivatives of C7 selected for resistance to high dose Dex [4-7]. Cell culture was carried out in RPMI 1640 medium, supplemented either with insulin, transferrin and selenium, or with fetal bovine serum, as described previously [8]. Cell growth and viability was followed by direct observation, counting in a Coulter counter, and by trypan blue exclusion with manual counting by hemocytometer as described [8]. Gene constructs MMTV-Stumyc was obtained from the laboratory of Dr E. A. Thompson, University of Texas Medical Branch, Galveston, TX [9]. The ER: GR cassette chimeric gene was provided by Dr P. Chambon, Strasbourg, France [10]. The progesterone receptor gene was supplied by Drs B. O'Malley, M. Tsai and W. Schrader, Baylor College of Medicine, Houston, TX[ll]. The other steroid receptor constructs were obtained from Drs R. M. Evans and S. M. Hollenberg, The Salk Institute, La Jolla, CA [12-14]. The oligomers employed for

direct cell uptake in antisense experiments were synthesized by the HBC&G Synthesis Laboratory of our department. Transfection of cells by electroporation Electroporation was carried out, with slight modification, from the method of Harbour et al. [15]. For steroid receptor transfections, logarithmically growing cells were suspended at 4 × 106 cells/0.8 ml [for c-myc transfections, 8 × 106 cells/0.8 ml] in calcium and magnesiumfree phosphate buffered isotonic saline, pH 7.5. To each such cell suspension in an electroporation cuvette was added 15/~g [20 pg for c-myc] of plasmid DNA containing the gene of interest. The cells were pulsed at 200 V and 500 #F capacitance for 12-14 ms. The cells were then resuspended at approximately 4 × 105/ml [8 x 105/ml for c-myc] in fresh RPMI 1640 with 5% fetal calf serum and reincubated at 37°C. Careful cell counts were taken then, and at appropriate subsequent time intervals for the duration of the experiment. RNA analysis RNA was isolated according to standard procedures[l]. Total RNA was subjected to Northern or Dot blot analysis, using appropriate probes and according to our previously described conditions [1, 16]. Hybridization was carried out as described, and all blots were normalized to standard invariant RNA markers, including ~-tubulin, ribosomal RNA, and/or total poly A + RNA, measured on the same filter as the variable mRNA. RESULTS

Induction of GR in human acute lymphoblastic leukemia cells At receptor-saturating concentrations, Dex causes lysis of cells of the wild-type clone CEMC7 but not of resistant CEM clones C1, ICR-27, 4R4 and 3R43. However, the unusual steroid CVZ has been shown to lyse both wild-type and Dex resistant clones [17]. CVZ binds to GRs in both wild-type cells and resistant cells containing mutant receptor [18]. It has been shown that Dex can induce GR mRNA in the sensitive clone C7, but not the resistant clone ICR-27 [19]. Since CVZ can lyse the Dex resistant CEM cells as well as Dex sensitive ones, we examined these cells for induction of GR mRNA. We first established that CVZ was an inducer of GR mRNA in the wild-type Dex sensitive clone C7,

Steroid receptor mediated lymphoid cell death comparing CVZ for inducing potency with triamcinolone acetonide (TA), another potent glucocorticoid. Induction of G R m R N A by both steroids in wild-type CEM-C7 cells is shown in Fig. 1. The results show that CVZ induced G R m R N A in a dose dependent manner and with greater potency than TA. Having established that CVZ evoked the standard glucocorticoid response of the receptor m R N A in wild-type cells, we examined the ability of the phenylpyrazolo glucocorticoid to induce G R m R N A in Dex resistant clones. The data in Fig. 2 shows the results from experiments indicating that indeed CVZ can induce G R m R N A in three clones of Dex resistant CEM cells, representing two different classes of receptor mutants. Slightly higher concentrations of the steroid were required for induction in the resistant cells than in the sensitive parental clone. ICR-27 cells are receptor deficient, having few G R binding sites detectable by standard assay methods. However, these cells do contain immunologically detectable GR, which appears to bind CVZ [18, 20]. 4R4 and 3R43 are clones of the "activation labile" class, containing G R which can bind Dex with fair affinity but which cannot retain the steroid during the activation step that converts the receptor to its D N A binding form. Phenotypically these cells appear to be defective in nuclear translocation[5]. In these experiments two other commonly used steroids, Dex and TA, were tested as well. Even at concentrations of 10 -6 M Dex and 10-7M TA, they failed to induce G R m R N A in the resistant clones (data not shown).

275

i0 4R4 ~-~ 3R4S

0"

-8

-9

10

M

10

-7

M

io M

CVZ C O N C E N T R A T I O N

Fig. 2. Induction of GR mRNA by CVZ in three Dex- and TA-resistant CEM clones. ICR-27 (open bars). Clones 4R4 (solid bars), and 3R43 (diagonal bars) were exposed to a range of concentrations of CVZ from 10- u to 10- 7M, and total RNA was prepared and analyzed for GR mRNA by Dot blot as in Fig. 1. Concentrations of < 1 0 - 9 M had no effect (not shown). The results from three experimentsare averaged and shown with SD representedby the error bars. Reproduced with permission from Ref. [1].

Regions of the human GR important for lymphoid cell lysis

The presence of the G R is clearly necessary for glucocorticoids to evoke cell lysis. Fundamental evidence for this has been the finding that in both rodent and human cell lines, selection for steroid resistance has resulted in cells containing mutant receptors [5, 6, 21, 22]. As noted above, we have shown that one such human cell clone (ICR-27) contains abnormal GR, with a deficit in functional binding sites [20]. Transfection with an expression vector containing the gene for holoGR restores ligand binding and the lytic response when Dex is administered to these cells [15]. Knowing that these cells are capable of being lysed by transCVZ fected G R followed by addition of ligand, we ~m TA 5 reasoned that they could be employed to map Z the regions important for the lysis function of e 4. c,) the receptor. Functional regions of the G R have already been mapped for their role in gene z 8. induction, and to a lesser extent, gene repres2. o sion, and this data has been reviewed [23, 24]. The major defined regions of the G R important for induction of genes are: (1) the site-specific 0 -DNA binding domain, amino acids 421--486, -11 10 - 7 ,o-'% ,0-°~ ,°-% 10 M presumably organized in the form of two "zinc STEROID C O N C E N T R A T I O N fingers" separated by a linker region of about Fig. 1. Induction of GR mRNA by CVZ and TA in 15 amino acids; (2) the ligand-binding region, wild-type CEM-C7 cells. Data represent the mean + SD from three experimentsin which total RNA was examined subsumed by amino acids 556-777; and (3) by Dot blot with three replicas in each experiment. The tau 1 (77-262) plus tau 2 (526-555), two regions probe was nick translation-labelled human GR eDNA. important for maximizing gene transcription Note: in this and Fig. 2, i-fold induction means no difference from basal controls. Reprinted with permission from [23-25]. These regions are diagrammed in Ref. [1]. the representation of the human G R in Fig. 3.

I

276

E.B.

THOMPSON et al. % LYSIS .@ 10 M STERO0 + 48-96 HOURS

@.24 HCIJRS

FOU.OMNG"rRN~SFECTIO~ o

100

200

300

40o

S00

800

700

I

I

I

I

I

I

I

I

T~ I

"1

1~\\\\\\\\\~1

I

k'x\\\\\\\\\'~

~

Tam 2 St~oid Binding

t½½~½{I

:-:-:-:-:-:-:-:-:-:-:-:-c

26

+/-4

420 4S1

}r"g:l IS0

I

kN\\\\\\\\~] ~\\\\\\\\\N~

4|7

I11~"1 ========================= 42S

I

N:i:i:i:i:i:!:!:!:i:i:i:]

490

~

~i:i:i:!:i:i:i:i:i:i:i:i:

422

I

~x\NN\NNNNN'RI

Ill'lift =========================

~\\\\\\\\\"~

-:------'~---- ============================

k,\\\\\\\\N~

. . . . . . ============================ ,.,,,,.

~,\\\\\~

H 4 H ~ gC:.':~:i:i:'::i:i:!:!:!:!:q

GTGI

I GTG3A

I

24

+/-5

FR

I

-

20 +/-1 s

Fig. 3. The GRE-specific DNA binding domain is necessary for lysis of transfeeted GR-deficient leukemic lymphoblasts. The rectangles represent the human GR (top) and various mutant constructs. Indicated are deletions or insertion mutations in the DNA binding domain of the GR, chimeric constructs in which the GR DNA binding domain is replaced with modified versions of the thyroid hormone receptor DNA binding domain, and the progesterone receptor. The numbers above each box correspond to the amino acid position in the protein sequence of the steroid receptor. The left diagonal hatches, the diamonds, the right diagonal hatches and the chevrons correspond to the tau I domain, the DNA binding domain, the tau 2 domain and the steroid binding domain, respectively. The percentage cell lysis or reduction in viable cell number, in the absence ( - ) or presence ( + ) of steroid, following transfeetion of ICR-27 cells with these steroid receptor constructs, is indicated in the figure. Values given represent the averages of maximum kill observed in at least three transfeetions (each done in triplicate) along with standard deviations. A " - " sign in the data columns indicates < 4 % cell lysis. Reproduced with permission from data in Ref. [2].

Also mapped, but not indicated in the diagram, is a short sequence of amino acids just carboxyterminal from the DNA binding region, which appear to be essential for the nuclear translocation of iigand-bearing GR [26]. Repression of genes may not require all these same functional domains [24]. ICR-27 cells were transfected with a variety of receptor constructs. The cells were then tested for their lytic response, in the presence and absence of steroid. The DNA binding region (or most of it) proved to be essential for lysis. This was shown in several ways. Deletions of the entire region or of either complete Zn finger (along with part of the interfinger sequence) abolished the response, as did an insertion mutation in a critical part of the

first Zn finger, mutant 422 (Fig. 3). Substitution of the estrogen response element (ERE)specific DNA binding region of the estrogen receptor with that of the GRE-specific DNA binding region of the GR produced a chimeric receptor which killed cells exposed to estrogen rather than glucocorticoid[15]. The holoreceptor for progesterone (PR), whose DNA binding region recognizes and binds to GRE sites, caused the cells to be lysed by progesterone (Fig. 3). Finally, two chimeric genes were obtained which had been constructed so that within the GR, the DNA binding site recognized only the ERE or both the ERE and the GRE. Only the latter caused cell lysis (GTG3A and GTG1, Fig. 3). Taken together, these data indicate that the GRE-specific DNA

Steroid receptor mediated lymphoid cell death

277

region (truncated two amino acids beyond the first pair of cysteines in the second Zn finger). Thus of the DNA binding domain, only the first Zn finger through that point is necessary for the effect we observed. Second, receptor constructs truncated at amino acids 532, 515, 491,488 and 465 (all of which lack the entire steroid binding domain) caused lysis even in the absence of Dex. Third, they caused cell death within 6-24 h of transfection, as against 48-96 h after transfection with holoreceptor in the presence of

binding region is essential to evoke lysis in these cells. Removal of the steroid binding domain created genes constitutively active in causing cell lysis. These clones provided several additional interesting insights. First, one of them further defined the requirement for the DNA binding region. Clone 465* (Fig. 4) codes for amino acids 1-465 of the human GR; thus it lacks the entire steroid binding region plus most of the second Zn finger of the DNA binding

% LYSIS -8

lO M DEXAME'n..IA3Oe~E + 48-96

6-24

HOURS FC~OWIHG TI~I~FECTK3N 0

100

200

300

400

SO0

800

700

I

I

I

I

I

I

I

I

Tau 1

,

~

~\\\\\\\\~

f;;.;;t

S l ~ ' ~ Sindino

=========================== c

26 +/-4

FA:i:i:i:~:~:i:!:i:i:!:i:i:]

28 +,.2

===========================

28 ,.45

385

9

77

I

Tau 2

282

~

t.:;;;t

S32"

I

l~\\\\\\\\~

~½½t 8

22 +/-3.s

22 +/-3.s

29 +t-4

29 ,-4

23 +t.3

23 ,.3

815"

I

~\\\\\\\\\~

IlJ:~:ll 4g1"

I

I~\\\\\\\\\N'I

1;1;;i 488"

I

k~\\\\\\\\",~

El!ill

a39 +/-4

a39 +/. 4

l

I~\\\\\\\\\N

It:d

%

%

9

Ill

77

282

2O4

I~\\\\\N~1"'fl

+/-4

832"

18 +1-3

18 +/-3

27 --s.s

27 +,.s.s

2e .-4.s

2e +t-4s

SS0"

!:;:::1 N

[

+/.4

309

SS8"

tillill

gt

Fig. 4. The transcriptional activation and steroid binding domains of the GR are not required for leukemic cell kill. The human GR and mutant constructs having deletions in the amino- and carboxy-terminal domains of the GR are represented by the rectangular boxes according to the conventions used in Fig. 3. The percentage reduction in viable cell number, both in the absence ( - ) or presence ( + ) of Dex, following transfection of ICR-27 cells with these steroid receptor constructs is indicated. For constructs with the steroid binding domain deleted, equivalent kill was seen at 48-96 h with or without Dex. Values given represent the averages of maximum kill observed in at least three transfections (each done in triplicate) along with standard deviations. A " - " sign indicates < 4 % cell lysis. Superscript "a" indicates results significantly different from holoreceptor controls plus Dex (Student's t-test, P ~<0.005). For all the constructs with carboxy-terminal deletions, cells were followed for kill during the 6--96 h interval both in the presence or absence of Dex. There was equivalent cell kill in either case. Reprinted with permission from data in Ref. [2].

278

E.B. Thorn,sot4et al.

steroid. And fourth, they were unexpectedly efficient in killing (Fig. 4). All were at least equal in effect to transfected holoreceptor plus added hormone. Receptor constructs truncated at amino acids 488 and 465 actually were more potent than the holoreceptor in lysing cells (Student's t-test, P ~<0.005). The requirement for the GRE-specific DNA binding region, or key parts of it, strongly suggests that the receptor fragment produced is acting in the nucleus. Note, however, that all the truncated genes produce peptide lacking the nuclear localization sequence. Possibly these peptides are small enough to pass through nuclear pores without need of the mechanisms requiring the nuclear localization signal. The portion of the GR amino-terminal to the DNA binding domain proved to be non-essential for evocation of cell lysis. Deletion of just the tau I domain (A77-262) or nearly the entire amino-terminal sequence (A9-385) left a hormone-activated receptor with about the same potency as the holoreceptor (Fig. 4). However mutants with deletions both carboxy- and amino-terminal from the DNA binding domain suggest that part of the amino-terminal region may augment the lytic response. Thus, mutants A77-262/550" and A204-309/550" were slightly more active than A9-385/532" (Fig. 4). Also, the most active clones were the constitutive ones 488* and 465", which both contained the entire amino-terminal sequence, 1-420. Whether this reflects the action of a specific augmenting sequence or simply a total net size effect remains to be seen. Regulation o f c-myc and cytolysis of C E M cells

Because of the central role of protooncogenes in control of cell growth and differentiation, we selected a set of such genes for analysis and correlation with regulation of cell growth and lysis by glucocorticoids in CEM cell clones. Among the eight detectable oncogene and growth gene signals that were tested, only c-myc showed regulation by glucocorticoids. We have previously shown that glucocorticoids rapidly suppress c-myc only in clones which are growth inhibited and lysed by Dex. The c-myc mRNA levels are measurably diminished in 1 h after administration of steroid and reach a minimum by about 18 h, preceding the onset of cell lysis. In Dex resistant clones with mutations in GR, myc is not repressed. Even clone CEM-C1, which contains GR apparently identical to wildtype cells but which lacks some "lysis function"

preventing inhibition of its growth by Dex, does not repress c-mye [16]. From these results we proposed that c-myc suppression is a key step in the cascade of events leading to cell death of CEM cells. We hypothesized that overexpression of a c-mye gene in CEM-C7 cells should render them resistant to the cytolytic effects of Dex; normally this sensitive cell clone is completely killed by Dex in a sequence of events noticeably beginning 24 h after steroid administration. To test this hypothesis we transfected CEM-C7 cells with an expression vector containing MMTV-Stumye in which the mouse c-mye gene is under the control of the mouse mammary tumor virus long terminal repeat (LTR). This LTR is well known to contain specific binding sites (GREs) for the GR and to be able to mediate induction of downstream genes under its control. Several control conditions were employed, including electroporation without DNA, with the vector plasmid pBR322 as a nonspecific DNA, or with the plasmid pSV2neo as an alternative nonspecific vector. After electroporation, the CEM-C7 cells were treated with 1/~ M Dex and subsequently tested for remaining viable cells. After 48 h, the Dex-treated control cells showed approx. 60% cell loss; in these Dex sensitive cells this represents the usual degree of killing at this time point in these conditions. The cells transfected with MMTV-Stumye were significantly protected from the lethal effect of the glucocorticoid, showing a loss of only about 30% compared to controls. Figure 5 shows the results of one such experiment; the black bar graph indicates the percent inhibition of the Dex effect in the cells transfected with MMTV-Stumyc. There is a nearly 60% inhibition of the Dex lethal effect by the induced c-mye oncogene. In parallel experiments (data not shown) we found that p65, the major c-mye protein expressed in these cells, is maintained at relatively higher levels in the MMTV-Stumye transfected, Dextreated cells compared to controls. Thus the results of these experiments support the hypothesis that the repression of c-mye by glucocorticoids in these cells is a critical part of the process that leads to cell lysis. Maintaining c-myc at high levels prevents Dex induced cell death. A second prediction of the hypothesis that mye expression is a key step in glucocorticoid induced cell death is that any agent which represses c-myc expression sufficiently should kill CEM cells. To test this prediction we employed antisense c-myc oligonucleotides. It is

Steroid receptor mediated lymphoid cell death MMTV-Stumyc

MMTV

_.~ ~,

P1

80

48h

~ 60 •7

e

a)

279

mentary to vesicular stomatitis virus (VSV) mRNA had essentially no effect on viability. In summary, results from transfections both with the MMTV-Stumyc expression vector and antisense oligonucleotides suggest that c-myc gene expression is essential for continuing viability of these cells and that c-myc down regulation is a critical point in induction of cell death by glucocorticoids.

£ >

40

o

~ 2o

MMTV-mye Fig. 5. Sustained expression of c-myc protects against Dex induced lysis of CEM-C7 cells. CEM-C7 cells (1 x 107) were transfected with control D N A vector or steroid inducible MMTV-Stumyc expression vector. 24 h after transfection, cells were treated with 1 # M Dex or ethanol vehicle and cell viability was followed at intervals for 72 h. Shown is the maximum percentage inhibition of cytolysis in MMTVStumyc transfected cells (m), which occurred at 48 h of Dex treatment. The cytolytic effect of Dex in cells transfected with control D N A vector was normalized to 100% and percent inhibition due to transfection with MMTV-myc was calculated (ordinate). Inset: shown are the data upon which the percent inhibition of cytolysis due to MMTV-myc transfection is based. Transfection of CEM-C7 cells with MMTV-Stumyc or control DNA vector was carried out in quadruplicate. Cell viabilities were determined by the reduction in cell number in Dex-treated cells compared to untreated cells, expressed as percentages. The error bars represent standard deviations. Bar diagram: outline of the transfected MMTV-Stumyc gene, carried in the plasmid pA9 (not shown). MMTV-LTR (MMTV) is shown as I~; exons 1, 2, 3 of c-myc by []. Transcription start sites are shown by arrows. The white region between MMTV and myc exon 1 is the native myc promoter. Other white bands are introns and 3' DNA from the myc gene.

known that antisense oligomers are taken up by cells from culture medium, and it has been shown that antisense oligonucleotides complementary to certain regions of c-myc mRNA form a duplex with that mRNA and arrest translation, thus depleting both mRNA and c-myc protein levels in the cells [27]. Accordingly, varying concentrations of antisense oligomers were added to CEM-C7 cells, and cell viability was determined at various times thereafter. Antisense c-myc oligomers significantly reduced the viability of CEM-C7 cells, such that 48 h after their addition there was greater than 90% cell kill (Fig. 6). Incubation with a nonspecific antisense oligomer compli-

DISCUSSION

In these experiments we describe our results concerning the mechanism of action of glucocorticoids in induction of leukemic cell death. We have found a correlation in CEM cells between induction of GR mRNA and induction of cell lysis. We have confirmed the results of Eisen et al. that glucocorticoids induce GR mRNA in sensitive cells [19]. We have further shown that CVZ, a phenylpyrazolo glucocorticoid that can bind to the mutant GRs of Dex resistant CEM cells, also induces GR mRNA in them. We have also noted that Dex induces GR mRNA and protein in a myeloma cell line sensitive to glucocorticoids, but not in its insensitive sister line [28]. The mechanism of this induction and its relevance to steroid-induced cell death is not clear. It is true that glucocorticoids suppress GR mRNA in a variety of cell lines including normal lymphocytes, but on the other hand there are data showing that increased quantities of GR in a given cell line result in an increased degree of responsiveness [29, 30]. It is possible that the induction of the GR mRNA and protein in cells may contribute to the effectiveness of steroids by providing a positive feedback pathway which amplifies the effect of the hormone on the lytic process. As to the mechanism of the regulation of the GR genes, the GR promoter/regulatory region of the human GR gene has recently been cloned and sequenced[31]. In that study we found that the promoter of the GR is that of a typical "housekeeping gene", containing no TATA box or CAAT box, but a large number of GC boxes and several transcription start sites. We found no classic, complete GRE sites, although two half-sites were present at approximately -2.5 kb. Two potential negative GREs were also found in the same region. Further studies on various cells of malignant origin examining the mechanism of regulation of the glucocorticoid gene and its relationship to growth sensitivity are warranted.

280

E.B. THOMPSONet al. EXON 1

EXON 2

EXON 3

OLIGOvIER CON(3. 4,tzlVl

r.'::_~:J~l~ ¢-myc ~ ~ti-~'~

S~t~e

AUG C C C

% LYSIS AT CUC A A C

GUU

oli~r~

3'TAC

GGG G A G T T G

CAA

vsv M

5'TTG

GGA TAA

TTA

CAC

> <

48h

90 4

Fig. 6. Induction of cell lysis by antisense c-myc oligomer in CEM-C7 cells: CEM-C7 cells (6 x 105 cells/ml) in growth medium with 5% serum were incubated with 4 ~M of an antisense oligomer complementary to c-myc mRNA. As controls, cells were incubated with 4 # M VSV andsense oligomer or with no oligomer. Total cell number and cell viability were determined at intervals for 48 h. Maximum effect was observed at 48 h; therefore, the viable cell number of cells with no oligomer at 48 h was taken as 100% and percent viable cells remaining after addition of antisense oligomer or VSV oligomer were calculated and expressed as percent lysis. Results shown are averages from one experiment in duplicate. Top, diagram of the c-myc gene as described in Fig. 5. The exons from the gene assembled as mRNA are shown below, with the coding sequences in solid black. The initiation codons and the next four codons of the c-myc RNA are shown above the sequence of the antisense oligomer and the non-specific VSV sequence employed.

That the GR has a central role in the growth regulatory effects of glucocorticoids is clear. The immunosuppressive action of glucocorticoids includes two effects: block of cell proliferation and induction of cell lysis. Glucocorticoids mediate their direct effects on target cells by binding to a specific intracellular GR. Although steroid-resistant cells may contain GR, absence of specific glucocorticoid binding invariably has been associated with resistance to steroid [6, 21, 22, 32]. This led to the view that hormone-receptor complex formation is a key step in glucocorticoid-induced cell death. The importance of GR in the cytolytic process has been demonstrated in a variety of cell lines with glucocorticoid sensitive and resistant variants, such as $49, W7, and CCRF-CEM cells [5, 6, 21, 22]. Hence the hypothesis for glucocorticoid-induced lymphocytolysis has been that on binding to a specific receptor, glucocorticoids induce so called "lysis genes" critical for lysis. So far the mechanism(s) with which glucocorticoids induce cell lysis have not been elucidated. We suggest that this hypothesis be modified to include that glucocorticoids may cause lysis by repression of genes essential for cell viability. We have mapped the topology of the GR for evoking cell lysis [2, 15] and our results include several striking points: (1) At least most of the DNA binding region in the receptor specific for GRE sites is critical. The larger part of the second Zn finger plus the steroid binding region can be eliminated and leave a fully active,

constitutive gene for lysis. Specifically, truncation mutant 465* stops two amino acids beyond the first two cysteines in the second Zn finger in the DNA binding region of the human GR. This clone when transfected into glucocorticoid insensitive CEM cells is actually more potent than the holoreceptor plus Dex. However, the first Zn finger appears to be absolutely essential, as may be the interfinger region. (2) When the steroid binding region is removed, the resulting receptors are constitutive. This is also true for gene induction; however there is a significant quantitative difference. In the case of the cell lysis response, the truncated genes are actually in many cases as potent as haloreceptor plus steroid and in some cases more so. Thus we find that 465* and 488* are statistically significantly more potent than Dex. Furthermore, removing the tau I region, important for evoking maximum trans-activation, only results in a slight decrease in GR's killing efficiency. This may be due to a threshold effect. The constitutive activity of the truncated gene may produce an event, i.e. induction or repression of critical genes, sufficient to cause them to be altered beyond an essential threshold level, leading to cell lysis. Since cell death is an all-or-none response, such a threshold system, whether it is induction or suppression of genes, could account for our results. A minimal but critical change in gene expression could produce cell death. However if this be so, the minimal induction required may be very small. If the response is the same as that of MMTV:CAT, clone 465*

Steroid receptor mediated lymphoid cell death

281

had only 1% of wild-type inductive activity [12], herein. Thanks also to Jackie Orozco for help in preparation of this manuscript. Supported in part by National Institutes whereas in our system it is more active for lysis of Health, National Cancer Institute, grant CA-41407-06 and in collaboration with The Walls Foundation. than holoGR. Other regions of the GR, notably the aminoterminal sequences preceding the DNA binding REFERENCES region, appear to be nonessential for cell lysis, although a portion of the upstream sequence 1. Ashraf J., Kunapuli S., Chilton D. and Thompson E. B.: Cortivazol mediated induction of glucocorticoid may augment the lytic effect. This finding apreceptor messenger fibonucleic acid in wild-type and pears to be in contrast with the results from dexamethasone resistant human leukemic (CEM) cells. J. Steroid Biochem. Molec. Biol. 38 (1991) 561-568. mutants in the mouse GR. Cells of the $49 line selected for resistance to high dose Dex were 2. Nazareth L. V., Harbour D. V. and Thompson E. B.: Mapping the human glucocorticoid receptor for shown to include clones which contain trunleukemic cell death. J. Biol. Chem. 266 (1991) 12976-12980, cated receptors lacking a large portion of the amino-terminal region preceding the DNA 3. Foley G. E., Lazarus H., Farber S., Uzman B. G., Boone B. A. and McCarthy R. E.: Continuous culture binding site. Analysis of these genes suggests of human lymphoblasts from peripheral blood of a child with acute leukemia. Cancer 18 (1965) 522-529. that at least in the mouse $49 cells the upstream region is essential for cell lysis [33]. The appar4. Norman M. R. and Thompson E. B.: Characterization of a glucocorticoid-sensitive human lymphoid cell line. ent paradox between this result and our findings Cancer Res. 37 (1977) 3785-3791. in the human system may have several expla5. Schmidt T. J., Harmon J. M. and Thompson E. B.: "Activation-labile" glucocorticoid-receptor complexes nations, which will require direct experimenof a steroid-resistant variant of CEM-C7 human lymtation to test. phoid cells. Nature, Lond. 386 (1980) 507-510. Cell proliferation requires the expression of 6. Harmon J. M. and Thompson E. B.: Isolation and characterization of dexamethasone-resistant mutants c-myc; it often is diminished when cells are from human lymphoid cell line CEM-C7. Molec. Cell. caused to differentiate and enter G0134-36]. Biol. 1 (1981) 512-521. Suppression of the c-myc protooncogene seems 7. Zawydiwski R., Harmon J. M. and Thompson E. B.: Glucocorticoid-resistant human acute lymphoblastic critically important in CEM cells for cell lysis. leukemic cell line with functional receptor. Cancer Res. Its repression precedes lysis and is evoked by 43 (1983) 3865-3873. glucocorticoids only in cells which show sensi8. Chilton D., Johnson B. H., Danel-Moore L., Kawa S. and Thompson E. B.: Increased glucocorticoid retivity to them, whether those cells have funcsponses of CD4 + T-cell clonal lines adapted to serumtioning or mutant GRs as a basis for resistance. free media. In Vitro Cell. Dev. Biol. 26 (1990) 561 570. Two tests of the importance of c-myc support 9. Stewart T. A., Pattengale P. W. and Leder P.: Spontaneous mammary adenocarcinomas in transgenic mice the hypothesis: (1) Introduction of c-myc conthat carry and express MTV/myc fusion genes. Cell trolled by a promoter that does not permit 38 (1984) 627~37. suppression protects the cells against Dex- 10. Green S. and Chambon P.: Oestradiol induction of a glucocorticoid-responsive gene by a chimeric receptor. induced cell lysis. (2) Repression of c-myc by Nature, Lond. 325 (1987) 75-78. antisense oligonucleotides kills the cells. We 11. Conneely O. M., Dobson A. A., Tsai M-J., Beattie conclude that this protooncogene in this parW. G., Toft D. D., Huckaby C. S., Zarucki T., Schrader W. T. and O'Malley B. W.: Sequence and expression ticular leukemic cell line is of high importance of a functional chicken progesterone receptor. Molec. for the Dex effect. It is noteworthy that c-myc Endocr. 1 (1987) 517-525. repression by Dex has also been observed in 12. Hollenberg S. M., Giguere V., Segui P. and Evans R. M.: Colocalization of DNA binding and transcriptwo different mouse lymphoma cell lines, $49 tional activation functions in the human glucocorticoid and P1798 [37, 38]. Thus in three independently receptor. Cell 49 (1987) 39-46. derived lymphoid malignancies in two different 13. Giguere V., Hollenberg S. M., Rosenfeld M. G. and Evans R. M.: Functional domains of the human glucospecies, c-myc down regulation appears to corcorticoid receptor. Cell 46 (1986) 645~52. relate with Dex induced cell death. It may be 14. Umesono K. and Evans R. M.: Determination of target that this protooncogene, a known important gene specificity for steroid/thyroid hormone receptors. Cell 57 (1989) 1139-1146. element in controlling cell growth and differen15. Harbour D. V., Chambon P. and Thompson E. B.: tiation, has a critical role in sustaining viability Steroid mediated lysis of lymphoblasts requires the in these leukemic cells. Correlations of lysis DNA binding region of the steroid hormone receptor. by GR fragments with regulation of the c-myc 16. J. Steroid Biochem. 35 (1990) 1-9. Yuh Y. S. and Thompson E. B.: Glucocorticoid protooncogene have not yet been carried out effect on oncogene/growth gene expression in human but are of obvious importance. T-lymphoblastic leukemic cell line: CCRF-CEM: Acknowledgements~Our thanks to colleagues S. M. Hollenberg, R. M. Evans, B. O'Malley, M. Tsai, W. Schrader and P. Chambon for providing the steroid receptor clones used

specific c-myc mRNA suppression by dexamethasone. J. Biol. Chem. 264 (1989) 10,904-10,910. 17. Harmon J. M., Schmidt T. J. and Thompson E. B.: Non-glucocorticoid receptor-mediated effects of the

282

18. 19.

20.

21.

22.

23. 24. 25. 26. 27.

28.

E.B. TrlOMPSONet al. potent glucocorticoid deacylcortivazol. Cancer Res. 42 (1982) 2110-2114. Srivastava D. and Thompson E. B.: Two glucocorticoid binding sites on the human glucocorticoid receptor. Endocrinology 127 (1990) 1770-I 778. Eisen L., Elsasser M. and Harmon J.: Positive regulation of the glucocorticoid receptor in human T-cells sensitive to the cytolytic effect of glucocorticoids. J. Biol. Chem. 263 (1988) 12,044-12,048. Harmon J. M., Elsasser M. S., Eisen L. P., Urda L. A., Ashraf J. and Thompson E. B.: Glucocorticoid receptor expression in 'receptorless' mutants isolated from the human leukemic cell line CEM-C7. Molec. Endocr. 3 (1989) 734-743. Stevens J., Stevens Y.-W. and Haubenstock H.: Molecular basis of glucocorticoid resistance in experimental and human leukemia. In Biochemical Action of Hormones (Edited by G. Litwack). Academic Press, New York, Vol. X (1983) pp. 383-466. Sibley C. H. and Yamamoto K. R.: Mouse lymphoid cells: mechanisms of resistance to glucocorticoids. In Glucocorticoid Hormone Action (Edited by J. D. Baxter and G. G. Rousseau). Springer, Berlin (1979) pp. 357-376. Evans R. M.: The steroid and thyroid hormone receptor superfamily. Science, N.Y. 240 (1988) 889-895. Evans R. M.: Molecular characterization of the glucocorticoid receptor. Recent Prog. Horm. Res. 45 (1989) 1-27. Hollenberg S. M. and Evans R. M.: Multiple and cooperative trans-activation domains of the human glucocorticoid receptor. Cell 55 (1988) 899-906. Picard D. and Yamamoto K. R.: Two signals mediate hormone-dependent nuclear localization of the glucocorticoid receptor. EMBO Jl 6 (1987) 3333-3340. Holt J. T., Redner R. L. and Nienhius A. W.: An oligomer complementary to c-myc mRNA inhibits proliferation of HL-60 promyelocytic cells and induces differentiation. Molec. Cell. BioL 8 (1988) 963-973. Gomi M., Moriwaki K., Katagiri S., Kurata Y. and Thompson E. B.: Glucocorticoid effects on myeloma cells in culture: correlation of growth inhibition with

29. 30.

31.

32.

induction of glucocorticoid receptor mRNA. Cancer Res. 50 (1990) 1873-1878. Gehring U., Mugele K. and Ulrich J.: Cellular receptor levels and glucocorticoid responsiveness of lymphoid cells. Molec. Cell Endocr. 36 (1984) 107-113. Vanderbilt J. N., Miesfeld R., Maler B. A. and Yamamoto K. R.: Intracellular receptor concentration limits glucoeorticoid-dependent enhancer activity. Molec. Endocr. 1 (1987) 68-74. Zong J., Ashraf J. and Thompson E. B.: The promoter and first, untranslated exon of the human glucocorticoid receptor gene are GC rich but lack consensus glucocorticoid receptor element sites. Molec. Cell. Biol. 10 (1990) 5580-5585. Thompson E. B. and Harmon J. M.: Glucocorticoid receptors and glucocorticoid resistance in human leukemia in vivo and in vitro. In Steroid Hormone Resistanee (Edited by G. P. Chrousos, D. L. Loriaux and M. B. Lipsett). Plenum, New York (1986) pp. I 11-127.

33. Dieken E. S., Mecse E. U. and Miesfeld R. L.: nf glucocorticoid receptor transcripts lack sequences encoding the amino-terminal transcriptional modulatory domain. Molec. Cell. Biol. 10 (1990) 4574-4581. 34. Cole M. D.: The myc oncogene: its role in transformation and differentiation. A. Rev. Genet. 20 (1986) 361-384. 35. Ohalsson R. I. and Pffeifer-Ohalsson S. B.: Cancer genes, proto-oncogenes, development. Cell Res. 173 (1987) 1-16. 36. Alitalo K., Koskinen P., Makela T. P., Saksela K., Sistonen L. and Winqvist R.: myc oncogenes: activation and amplification. Biochim. Biophys. Acta 907 (1987) 1-32. 37. Eastman-Reks S. B. and Vedeckis W. V.: Glucocorticoid inhibition of c-myc, c-myb, c-Ki ras expression in a mouse lymphoma cell line. Cancer Res. 46 (1986) 2457-2462. 38. Forsthoefel A. M. and Thompson E. A.: Glucocorticoid regulation of transcription of c-myc cellular protooncogene in P1798 cells. Molec. Endocr. I (1987) 899-907.