Retinoic acid–induced apoptotic pathway in T-cell lymphoma

Experimental Hematology 28 (2000) 1441–1450

0301-472X/00 $–see front matter. Copyright © 2000 International Society for Experimental Hematology. Published by Elsevier Science Inc.PII S0301-472X(00)00546-4

Retinoic acid–induced apoptotic pathway in T-cell lymphoma:Identification of four groups of genes with differential biological functions

Kao-Chung Wang

a,e

, Ann-Lii Cheng

b

,Shuang-En Chuang

c

, Hey-Chi Hsu

a

, and Ih-Jen Su

a,d

a

Graduate Institute of Pathology, National Taiwan University Medical School, Taipei,Taiwan and

b

Department of Medicine, National Taiwan University Medical School, Taipei, Taiwan;

c

Division ofCancer Research, National Health Research Institute, Nan-Kang, Taiwan;

d

Department of Pathology and Institute of MolecularMedicine, National Cheng Kung University Medical School, Tainan, Taiwan;

e

Core Laboratory, Pohai Hospital, Lo-Tung, I-Lan, Taiwan

(Received 3 February 2000; revised 8 August 2000; accepted 11 August 2000)

Objective.

Retinoic acid (RA) has been used to induce the regression of refractory T-cell lym-phoma. In vitro and in vivo studies have shown that RA exerts this effect through the induc-tion of apoptosis. This study was designed to investigate the molecular pathway of RA-inducedapoptosis in T-lymphoma cell lines.

Materials and Methods.

RA-induced apoptosis was verified by morphology, flow cytometry,and DNA ladder analysis. Differential display method using a combination of 12 poly(A)-anchored primers and 20 arbitrary primers was adopted for gene cloning. Total RNAs wereextracted from H9 cell line at 0, 6, 12, and 24 hours after

All-trans

RA (ATRA) treatment andthe serial expression patterns of the candidate fragments were recognized. The cloned genefragments were then analyzed and confirmed by Northern blot analysis on H9 and SR786 celllines.

Results.

ATRA-induced apoptosis of T-cell lymphoma was protein synthesis–dependent. Theexecution or irreversible phase of apoptosis appeared to occur at 6–12 hours of RA treatment.Among the 60,000 arbitrarily displayed bands, 25 of 250 candidate fragments were selectedfor further cloning and sequencing. A total of 14 clones could be matched to known genes andwere categorized into four groups: A) transcription factors: prothymosin, CA150, p78 serine/threonine kinase, IL-1

b

–stimulating gene, glucocorticoid receptor, MLN64/CAB1, gastrin-binding protein, and polypeptide from glioblastoma; B) chaperone: 90 kDa heat shock pro-tein; C) ion channel: chloride channel protein 3; and D) cytoskeleton: cytovillin2/ezrin and vi-mentin. Another two clones of genes were of unrecognized functions. The remaining 11 clonesbelonged to unmatched or novel genes. The expression of these genes varied, either upregu-lated or downregulated, in response to ATRA treatment.

Conclusion.

RA-induced apoptosis may involve a cascade of genes that are related to tran-scription regulation, stress response, housekeeping, and the execution of apoptosis. The clarifi-cation of the RA-induced apoptotic pathway will help us to understand the molecular mecha-nism of cancer differentiation agents. © 2000 International Society for ExperimentalHematology. Published by Elsevier Science Inc.

Keywords:

Retinoic acid—Lymphoma—Apoptosis—Molecular pathway

Introduction

Retinoic acid (RA) and its derivatives have been demon-strated to be effective in the treatment and chemopreventionof selected human cancers [1]. The most well-documentedexample is the treatment of acute promyelocytic leukemia

(APL) by

All-trans

RA (ATRA) [2–4]. In addition to APL,RA has also been successfully used to salvage aggressiveperipheral T-cell lymphoma and cutaneous Ki-1 (CD30)lymphoma [5–7], particularly in refractory Epstein-Barr vi-rus–associated T-cell lymphoma, which expresses drug re-sistance phenotype [8]. In vivo and in vitro studies havedemonstrated that RA exerts these effects through the in-duction of differentiation and/or apoptosis [4,6–8]. In addi-tion to its therapeutic effect, RA has also been demonstrated

Offprint requests to: Ih-Jen Su, M.D., Ph.D., Department of Pathology,National Taiwan University Hospital, 7 Chung San South Road, Taipei,Taiwan; E-mail: [email protected]

1442

K.-C. Wang et al./Experimental Hematology 28 (2000) 1441–1450

to be effective in suppressing premalignant oral lesions andin preventing the development of second primary cancers inthe upper aerodigestive tract in patients with head and neckcancers [9–10].

Although RA exerts potential therapeutic and chemopre-ventive effects in human malignancies, the underlying mo-lecular mechanism remains poorly understood. In the pastyears, a few studies have been performed to investigatethe role of several known apoptosis-related genes in RA-induced apoptosis. The RA-induced apoptosis involvesprotein kinases, steroid receptors, the

bcl

-2 family, and on-cogenes [11–14]. In order to clarify whether other novelgenes or gene families may also be involved in the RA-induced apoptotic pathway, we adopted the differentialdisplay method to clone the RA-induced apoptosis–relatedgenes in a RA-sensitive T-cell lymphoma line H9. Throughthe serial expression patterns of the candidate genes at thedefined time course of RA treatment, we identified fourgroups of genes with differential biologic functions. Thesegenes probably represent a cascade of molecular and bio-logic events in the RA-induced apoptotic pathway andshould help to explain the molecular mechanism of cancerdifferentiation agents.

Materials and methods

Cell lines

Two RA-sensitive T-cell lymphoma cell lines, H9 (ATCCHTB176) and SR-786, were used in this study. H9 was a clonal de-rivative of HuT78 (ATCC TIB161, HTLV-I negative CD4

1

T-cell),a human T-cell line derived from leukemic T cells of a patient withSezary syndrome. SR-786 was derived from a CD30 (Ki-1)-posi-tive cutaneous anaplastic cell line with chromosomal aberration oft(2;5) (6). In this study, H9 cells were used in the differential dis-play study, the data obtained from which were further confirmed inSR786. The cells were maintained in RPMI 1640 medium with10% fetal bovine serum, 1% penicillin & streptomycin, and kept in37

8

C, 5% CO

2

, water jacked incubator.

DrugsAll-trans

retinoic acid (ATRA) (R-2625, Sigma Chemical Co., St.Louis, MO, USA) was used in this study. A stock solution of 20mM was prepared by dissolving in dimethylsulfoxide (DMSO).The stock solution was aliquoted and frozen at

2

70

8

C for no morethan one month before use. A 100-fold dilution working solutionof ATRA was prepared by diluting the stock solution with culturemedium. The final concentration of DMSO was below 0.1%.

In order to study whether the RA-induced apoptosis was pro-tein synthesis–dependent, the protein synthesis inhibitor cyclohex-imide (C-6679, Sigma Chemical Co.) was used in this study. Thepreparation of cycloheximide was similar to RA except that thesolvent was changed to ethanol and the concentration of stock so-lution was 10 mM.

Demonstration of ATRA-induced apoptosisMorphologic observations.

For the evaluation of morphologicchanges, the cultured cells were harvested and cytospinned at 1200

cpm. The cytospinned smears were then stained with a modifiedWright’s stain procedure [15] and read under light microscopy.

Flowcytometric analysis.

H9 cells were harvested and adjusted toabout 2

3

10

6

cells/mL. One mL of cells was washed once and re-suspended in 200

m

L of phosphate buffered saline (PBS). Ice coldethanol (800

m

L) was added to the suspension dropwise whilemixing the resuspended cells. The fixed cells were kept in

2

20

8

Cfor at least 30 minutes. The fixation solution was replaced by 0.5mL of prestain buffer (0.1% Triton X-100 in PBS), and kept in thedark at room temperature for at least 30 minutes. The prestainingbuffer was replaced by 680

m

L of PBS. One hundred

m

L propid-ium iodide (PI) solution and 20

m

L RNase A stock (final concen-tration of PI 50

m

g/mL and RNase A 50

m

g/mL) was added to re-suspended cells and kept in the dark at room temperature for 30minutes. FL2-A was measured by the Becton Dickinson FACScanand analyzed by MODFIT software (Verity Software House, Inc.,Topsharn, ME). All treatments were repeated three times.

DNA extraction and DNA fragmentation analysis.

A simple andefficient method of DNA fragmentation analysis to evaluate apopto-sis was adopted [16]. The DNAs were extracted from H9 cells after0, 6, 12, 24, and 48 hours of ATRA treatment. The H9 cells werewashed with PBS and pelleted by centrifugation. The cell pelletswere then resuspended gently but quickly for 10 seconds with 100

m

L lysis buffer (1% NP-40 in 20 mM EDTA, 50 mM Tris-HCl, pH7.5), followed by centrifugation for 5 minutes at 1600

g.

The super-natants were collected and brought to 1% SDS and treated for 2hours with 1/10 volume RNase A (stock concentration 10 mg/mL)at 56

8

C, followed by digestion with 1/10 volume proteinase K (stockconcentration 20 mg/mL) for at least 2 hours at 37

8

C. After additionof 100

m

L 10 M ammonium acetate, the DNAs were precipitated by750

m

L ethanol. The extracted DNAs were then redissolved in 20

m

L TE buffer, and separated by electrophoresis in 1% agarose gels.

Differential display methodto clone the ATRA-induced apoptotic genes

The differential display method was done, as previously described[17], with slight modification. Arbitrary primers were designed tohave a GC content of 50% and were inspected for the absence ofobvious hairpin structures.

ATRA treatment and RNA extraction.

The H9 cells were treatedwith ATRA and harvested at 0, 6, 12, and 24 hours. A parallelgroup without ATRA was used as controls.

Total cellular RNAs were extracted as previously described[18], with slight modification. A commercial reagent (Tri-reagent,Sigma) was used according to the manufacturer’s instructions. TheRNA pellets were redissolved in formamide.

Synthesis of first-strand cDNA and PCR amplification.

Two

m

g oftotal RNAs were used for the synthesis of first-strand cDNA accord-ing to the procedures previously described [17]. For PCR amplifica-tion, the reaction mixture included 1

m

L of the cDNA product, 2

m

Lof 10X PCR buffer (500 mM KCl, 100 mM Tris pH8.3, 20 mMMgCl

2

), 1.5

m

L of 0.1 mM dNTP, 1.25

m

L of 20 mM 5

9

primer, 1

m

L of

a

-32P dATP (800

m

Ci/

m

L, NEN Dupont (Boston, MA), 1:5dilution), 0.5

m

L of Taq polymerase (Boerhinger Mannheim, India-napolis, IN, USA), and 12.75

m

L DEPC-water. The PCR was per-formed using a thermocycler (Perkin-Elmer model 480, Branchburg,NJ, USA) programmed to conduct four cycles (94

8

C for 45 seconds,41

8

C for 60 seconds, and 72

8

C for 60 seconds) and then followed by


1443

18 cycles (94

8

C for 45 seconds, 60

8

C for 45 seconds, and 72

8

C for120 seconds). Reaction products were resolved in standard 6% poly-acrylamide sequencing gels containing 7 M urea. Gels were driedonto Whatman 3MM paper and exposed to Kodak XAR5 film (East-man Kodak Co., Rochester, NY) overnight.

Selection of the potential candidate clones by serial expressionpatterns.

DNA bands showing a serial expression pattern of eitherupregulation or downregulation at different time courses were se-lected for cloning and sequencing. The candidate bands were thenexcised from the gel and placed into a 1.5 mL tube with 1 mL ofTE buffer. The tube was placed in a boiling water bath for 10 min-utes and left at room temperature overnight. Five

m

L from this ex-traction was used for PCR reamplification.

Cloning and sequencing of the candidate fragments.

After ream-plification of the cut-out bands, the PCR products were cloned intoblue-white selective plasmids (pGEM T vector, Promega Co., Mad-ison, WI, USA) for sequencing. The white colonies were recon-firmed by PCR reaction with T7 & Sp6 primers for the presence ofinserts. The bands with a molecular weight corresponding to theoriginal cut-out samples from differential display were chosen foreither automated or manual sequencing. Automated sequencingwas performed using the Autosequencer 377 (Perkin Elmer’sApplied Biotechnique Inc., Foster City, CA, USA). Sample prepa-rations for auto-sequencing were according to manufacturer’s in-structions. The manual sequencing was performed using the Seque-nase version 2.0 DNA sequencing kit (United States Biochemical,Cleveland, OH, USA).

Sequence analysis and gene identification.

The Wisconsin Pack-age Version 9.1 (Genetics Computer Group (GCG), Madison, WI,USA) was used to analyze these differential display fragments.The FastA program was used instead of Blast program because ofits special gap ability, which is more sensitive to locate the match-ing sequences. Every sequence was analyzed by FASTA software(Genetic Computer Group, Madison, WI) in batch.

Northern blot analysis of the cloned sequences

To confirm the cloned sequences or genes involved in ATRA-induced apoptosis, Northern blot analysis was performed inATRA-treated H9 and SR-786 cells. Twenty-five

m

g of totalRNAs obtained from each treatment (0, 6, 12, 24, and 48 hours, re-spectively) of ATRA-treated H9 and SR-786 cells were separatedin 1.2% agarose/formaldehyde gel. The total RNAs were thentransferred to a nylon membrane by mass flow of 20X SSC over-night. The membrane was baked at 80

8

C for at least 30 minutes.The Northern membrane was then prehybridized and hybrid-

ized with appropriate probes of the cloned sequences. A GAPDHgene was used simultaneously in each study to provide an internalloading control of RNA quantitation. The expression intensity ofeach gene at different time course was quantified by phosphorim-ager Fuji BAS1500 system and plotted as the histogram by the ra-tio of the expression level of selected gene to GAPDH.

Results

ATRA-induced apoptosis in T cell lymphoma line H9Morphological observation.

ATRA at an IC50 concentra-tion of 0.2

m

M induced apoptosis in H9 cells. The cell mor-

phology showed no significant changes at 6 hours of ATRAtreatment. At 12–48 hours of ATRA treatment, the cell sizeshrank significantly and nuclear chromatin became con-densed (Fig. 1). Fragmented nuclei could be occasionallyobserved at 12–48 hours after ATRA treatment. In parallelstudies without ATRA treatment, no significant changes ofcell morphology could be observed up to 48 hours. There-fore, a low concentration of ATRA induced significant mor-phologic changes in H9.

Flowcytometric analysis.

H9 cells with propidium iodidelabeled showed the typical G0/G1 peak and G2/M peak atdouble fluorescence before 9 hours of ATRA treatment.However, the sub-G1 peak, representative of apoptotic cells,emerged and the G2/M peak disappeared at 12 hours (Fig.

Figure 1. Morphological changes of the cytospinned H9 cells after treat-ment with ATRA for 0, 6, 12, and 24 hours (A, B, C, D), respectively. Thecell sizes shrank dramatically and nuclear chromatin became condensed(arrow) at 12–24 hours after ATRA treatment. Fragmented nuclei couldoccasionally be demonstrated (arrow head). (3400, modified Wright’sstain.) In parallel studies without ATRA treatment, the cell morphologyshowed no significant changes up to 48 hours.

1444


2). MODFIT software analysis revealed that 60–70% of cellsunderwent apoptosis after 12 hours of ATRA treatment.

DNA fragmentation analysis.

DNA extracted from H9 cellsexposed to ATRA displayed a characteristic series of nucleoso-mal ladder pattern after agarose gel electrophoresis. The ladderpattern appeared faintly at 3–6 hours after ATRA treatmentand became readily recognized at 12 hours (data not shown).

ATRA-induced apoptosis was protein synthesis–dependent: an irreversible apoptosis process occurred at 6–12 hours

H9 cells were treated with ATRA and followed by the addi-tion of cycloheximide at 0, 6, 12, 24, and 48 hours of ATRAtreatment. The addition of cycloheximide could inhibit the ap-opotsis for up to 6 hours of ATRA treatment (Fig. 3). How-ever, after 12 hours of ATRA treatment, the addition of cyclo-heximide failed to arrest the apoptosis process, suggesting

Figure 2. Propidium iodide–labeled flowcytometric histograms of H9 cells at 0, 6, 9, 12, 24, 48 hours after ATRA treatment are shown. The changes of peaksrepresentative of G0/G1 and G2/M are serially observed. The shaded areas represent the fractions of cells in S phase. A sub-G1 peak showed up at 12 hoursand the fractions representative of apoptosis ranged from 60 to 70% at 12–48 hours.


1445

that the critical timing for the cellular commitment to apopto-sis occurred at 6–12 hours. The following strategy of differen-tial display study to clone the ATRA-induced genes related toapoptosis in H9 cells was therefore based on this observation.

Identification of ATRA-induced apoptotic genes bydifferential display based on the serial expression patterns

In this study, we used 12 anchored primers and 20 arbitraryprimers to display a total of 250

3

12

3

20

5

60,000 bandsfrom a typical 6% polyacrylamide gel. Judging from the se-rial expression patterns of either upregulation or downregu-lation (Fig. 4), a total of 250 candidate fragments were iden-tified to be potentially RA-responsive genes. Among them,25 candidate fragments were strictly time course–dependentand selected for further gene amplification, cloning, and se-quencing. After matching to the gene bank database, the se-quences of 14 fragments could be matched to the four fol-lowing known gene groups (Table 1):

Group 1:

transcription factor, including prothymosin-

a

(proT-

a

), putative transcription factor CA150, p78serine/threonine kinase, gene regulated by stimulationof chondrocytes with interleukin-1 (IL-1)-

b

, gluco-corticoid receptor (GR)-

a

, MLN64/CAB1, 78 KDagastrin-binding protein, polypeptide secreted from hu-man glioblastoma cells.

Group 2:

chaperone, the 90-kDa heat-shock protein(hsp 90).

Group 3:

ion channel, chloride channel protein 3(CLCN3).

Group 4:

cytoskeleton, cytovillin2/ezrin, vimentin(HuVim3).

Two fragments were matched to genes of unknown func-tion: A fragment matched to human chromosome 17 cloneHCIT 305D20 (GenBank accession number AC004098);the second fragment matched to “Human BAC cloneRG104I04 from 7121-7q22” (gene bank accession numberAC000119).

The sequences of the remaining 11 bands, which included8 upregulated and 3 downregulated clones, could not bematched to any known gene by FastA program and remainedto be characterized in the future. The 11 fragments have beendeposited to the gene bank for references (EST GenBank Ac-cession Number BE491950 to BE491960). The traditionalgenes involved in apoptosis such as p53,

bcl

-2, and caspaseswere not identified in this differential display method.

Northern blot analysis andexpression pattern of the cloned genes

The expression of the 25 cloned sequences in this studywere all reconfirmed to be functional genes by Northern

Figure 3. DNA fragmentation analysis to demonstrate the inhibitory effect of cycloheximide on the ATRA-induced apoptosis. The addition of cycloheximideto the culture medium could effectively arrest the ATRA-induced apoptosis before 6 hours but failed after 12 hours as evaluated by the presence of DNA lad-der patterns. The results suggest that the apoptosis program has already proceeded to an irreversible stage at 12 hours of ATRA treatment.

1446


blot analysis in H9 and SR-786 cells, indicating that the dif-ferential display method and the strategy we adopted in thisstudy were reasonably reliable. The upregulated genes in-cluded prothymosin-

a

, p78, and IL-1

b

stimulating gene(Fig. 5A), chloride channel protein 3, CA-150 transcriptionfactor, cytovillin/ezrin, BAC GR104104 (Table 1). Thedownregulated genes included GR-

a

, MLN64/CAB1, gas-trin-binding protein, polypeptide from glioblastoma cells,90 kDa heat shock protein, vimentin, HTGs phase 2HCIT305D20. The Northern expression patterns and therelative levels expressed as histogram of four representativegenes are presented in Figure 5B. For IL-1

b

-stimulatinggene, the upregulation of gene expression was immediate,reached the peak level at 6 hours, and remained at high leveluntil 48 hours of ATRA treatment. For transcription factorCA150, there was a biphasic response with an immediateupregulation at 6 hours, followed by a decline at 12–24hours. For heat shock protein 90 and vimentin, the activitydeclined sharply after 6 hours and reached the lowest levelat 12–24 hours. The expression patterns of the cloned genes

were similar in SR-786 cells for each gene (data not shown),indicating that RA may mediate apoptosis through a similaror identical molecular pathway in T-cell lymphoma lines. Theexpression of the remaining genes is summarized in Table 1.

Discussion

In this study, we performed a systematic approach by usingthe differential display method for gene cloning, and suc-cessfully identified four groups of genes associated withRA-induced apoptosis in two T-cell lymphoma lines. Thedata thus obtained will provide valuable information to aidin understanding the molecular cascades of the apoptoticpathway. The apoptotic program has been executed througha cascade of biologic and molecular events which may in-clude an initial and specific triggering pathway, followed bya second phase of stress response, a third phase of determi-nation or housekeeping pathway, and a final common exe-cution pathway [19]. Once the cells enter the determinationor housekeeping phase, the apoptosis process will become

Figure 4. A representative blot of differential display analysis. In each blot, eight lanes are displayed, representing the expression of candidate genes without(first four lanes, 0, 6, 12, 24 hours) or with ATRA treatment (lanes 5–8, 0, 6, 12, 24 hours). The first four lanes were used as control. By the expression pat-tern, candidate ATRA-induced genes were selected to be cloned and sequenced. Only bands showing an expression pattern of either upregulation or downreg-ulation were selected. An example (boxed area) is shown with magnification at the upper right portion. The solid arrow head represents a downregulated bandwhereas the open arrow head represents an upregulated band.


1447

irreversible. Based on the cycloheximide experiment in thisstudy, the determination or irreversible phase appeared tooccur at 6–12 hours of ATRA treatment, because the treat-ment of cycloheximide failed to arrest the apoptotic processafter this time period.

Since the therapeutic effect of RA has been demonstratedto be closely related to the inducibility of RAR-

a

gene ex-pression in lymphoid and APL cells [20–21] and theATRA-induced apoptosis will be arrested in the presence ofRAR-

a

antagonist [14, Wang et al., unpublished observa-tion], the upregulation of RAR-

a

is likely the initiatingevent for the subsequent signaling process related to RA-induced apoptosis [22]. The RAR-

a

represents a transcrip-tion factor that may bind to a specific sequence at the pro-moter region of RA-responsive genes [4]. Whether thepromoter region of the cloned genes contained the RAR-binding or responsive elements remains to be investigated.

The transcription regulatory genes we cloned in thisstudy may represent the early immediate genes followingRAR-

a

activation [23]. Genes that may belong to this cate-gory include prothymosin, transcription factor CA150, andIL-1

b

-stimulating gene based on the time course of generegulation and their biologic functions. In the prostaticgland studies, the

a

-prothymosin has been shown to involvein apoptosis and the expression was only enhanced duringprogrammed cell death and not during proliferation inducedby androgen replacement [24]. Bustelo et al. [25] examinedthe prothymosin expression in rat and found that prothy-mosin was required throughout all stages of the cell cycle.However, the same gene was found overexpressed, con-comitant with c-myc expression, and was restricted to tumornodules [26]. In this study, the prothymosin was upregu-lated by ATRA at an early stage of apoptosis (6 hours), sug-gesting that prothymosin plays a role at the early phase of

apoptosis. One interesting finding in this study was the up-regulation of IL-1

b

-stimulating gene in RA-induced apop-tosis. The activation of IL-1

b

has been reported to inhibitapoptosis [27]. No information regarding the role of IL-1

b

-stimulating gene in apoptosis is available in the literature; itshould be studied in the future.

The stress response genes may include hGR and hsp90.The heat shock protein family have been demonstrated to beinvolved in the regulation of apoptosis [28–29]. The hsp90we cloned in this study has been reported to correlate withsex steroid receptor status in endometrial carcinomas [30].hGR-

a

–specific antibodies can coprecipitate hGR andhsp90, suggesting that hGR is complexed with hsp90 [31].Interestingly, these two genes we cloned in this study wereboth downregulated by ATRA. The hGR and hsp90 iso-forms have been proposed as the core components of mam-malian stress response [32]. Downregulation of both genesby ATRA may facilitate the apoptosis pathway. It will beinteresting to test whether these two genes are also involvedin RA-induced apoptosis in cell lines other than T cells.

The determination or housekeeping phase genes may in-clude gastrin-binding protein and chloride-channel proteinCLCN3. Gastrin-binding protein is a fatty-acid-oxidizingenzyme, related to the energy supply of cells, and is abun-dant in many tissues of mouse, rat, and human. Gastrin-binding protein has been verified as a gastrin receptor [33].Since this gene was downregulated at a later stage (12hours) of ATRA-induced apoptosis in this study, the down-regulation of this gene may shut down part of the energysupply in cells and lead to cell death. The chloride-channelprotein CLCN3 we cloned in RA-induced apoptosis is inter-esting, since this gene is involved in the pathogenesis ofcystic fibrosis (CF). CF is caused by mutations in the CFTRgene. Maiuri et al. [34] reported inappropriately high DNA

Table 1. Groups of cloned DNA fragments by differential display method in all-trans retinoic acid–induced apoptosis

Fragment ID% Identity toknown gene

Regulationby ATRA Reference

Group 1: transcriptional regulatory factorProthymosin-a 00-3H 96.9 up Bustelo et al. [25], Wu et al. [26]CA150 12-1F 97.8 up Sune et al. [42]p78 serine/threonine kinase 16-2H 94.2 up GenBank M80359IL-1 stimulating gene 13H 98.9 up GenBank A52238GR-a 01-4C 94.6 down de Castro et al. [31]MLN/CAB1 03F 96.9 down Watari et al. [43]Gastrin-binding protein 01-1C 96 down Monstein et al. [33]Polypeptide from glioblastoma 09M 97.3 down GenBank E06811

Group 2: chaperoneHeatshock protein 90 19-1A 94.9 down Nanbu et al. [30]

Group 3: ion channelChloride channel protein 3 19-2M 92.9 up Borsani et al. [44]

Group 4: cytoskeletonCytovillin 2/ezrin 00-2H 94.6 up Turunen et al. [45]Vimentin 05M 99 down Honke and Wada [46]

1448 K.-C. Wang et al./Experimental Hematology 28 (2000) 1441–1450

fragmentation in various CF epithelia, suggesting the in-volvement of ion channel gene in apoptosis.

The cytoskeleton proteins cytovillin/ezrin and vimentinmay represent the late determination phase genes, since thedisruption of the cell membrane or cytoskeleton will lead to

an irreversible apoptosis process [35]. The downregulationof vimentin was a relatively late event, with the lowest levelgene expression at 12 hours of ATRA treatment in thisstudy, corresponding to the irreversible stage of ATRA-induced apoptosis. It is interesting to observe the downregu-

Figure 5. (A) Northern blot analysis of four representative genes regulated by retinioc acid (RA) after 0, 6, 12, 24, and 48 hours with (left panel) or without(right panel) ATRA treatment. The GAPDH gene was used as internal control of the loaded RNAs. The IL-1 stimulating gene and CA150 were upregulated at6 hours, while the hsp90 and vimentin were downregulated and reached bottom level at 12–24 hours. (B) Histograms representative of the relative intensity ofNorthern blot levels, as quantified by phosphorimager Fuji BAS1500 system, of four representative genes are shown. Different expression patterns wereobserved for different genes.

K.-C. Wang et al./Experimental Hematology 28 (2000) 1441–1450 1449

lation of cytoskeleton genes in RA-induced apoptosis, sincethe ruffling or blebing of the cell membrane is a characteris-tic feature of apoptosis and may involve the disruption ofcytoskeleton proteins such as vimentin and tubulin [36].

Differential display is a recently developed gene cloningtechnique [37,17]. This method uses only a fraction of RNAfrom the target pool and also requires less labor than othergene cloning methods. The major disadvantage of thismethod, however, is the false positivity of the cloned frag-ments due to the inconsistency of RNA sources and the lowspecificity of the arbitrary primer pairs in PCR reactions. Toovercome this disadvantage, serial sample profiles to dis-play the expression pattern were adopted in this study. Theverification of the cloned sequences as the true functioninggenes by Northern analysis further confirms the advantagesand superiority of the differential display method and thestrategy we adopted in this study. The 25 genes we identi-fied, however, did not include some traditional apoptosis-related genes such as fas, bcl-2/bax, and the caspase familygenes, probably due to the small number of clones (25 outof 250) selected for sequencing in this study. Preliminarystudies revealed that p21, bcl-2, and caspase 1 were in-volved in the RA-induced apoptosis in T lymphoma lines(K.C. Wang and I.J. Su, unpublished data).

By differential display method, RA has been demon-strated to induce specific gene expression in different cellsystems. A human homolog of the murine Ly-6 family, RIG-E,and a Src-like adapter protein have been identified in RA-induced differentiation of APL line NB4 (38,39). A putativeG protein–coupled receptor, RAIGI, and an insulin growthfactor–like binding protein, mac 25, were identified in epi-thelial cells lines (40,41). Whether these RA-inducible genesin myeloid and epithelial cells are also involved in T-celllymphoma lines, or vice versa, should be investigated in thefuture. The results will help to clarify the molecular cascadesof RA-induced apoptotic pathways in different cell systems.

In summary, this study has identified four groups of RA-regulated genes with differential biologic functions by a spe-cially designed differential display method. This study fur-ther proves the versatility of the differential display methodfor gene cloning. The expression of these cloned genes withdifferent time intervals may provide useful information onthe molecular cascades of the apoptotic pathway in RA-induced apoptosis. The knowledge regarding the mechanismof apoptosis should contribute to our understanding of themolecular mechanism of cancer differentiation agents.

AcknowledgmentsThis research was supported by grants from National Health Re-search Institute (Dr. I.J. Su and Dr. A.L. Cheng).

References1. Miller WH (1998) The emerging role of retinoids and retinoic acid me-

tabolism blocking agents in the treatment of cancer. Cancer 83:1471

2. Huang ME, Ye YC, Chen SR, et al. (1988) Use of all-trans retinoicacid in the treatment of acute promyelocytic leukemia. Blood 72:567

3. Fenaux P (1993) The role of all-trans retinoic acid in the treatment ofacute promyelocytic leukemia. Acta Haematol 1:22

4. Robertson KA, Emami B, Mueller L, Collins SJ (1992) Multiplemembers of the retinoic acid receptor family are capable of mediatingthe granulocytic differentiation of HL-60 cells. Mol Cell Biol12:3743

5. Chow JM, Cheng AL, Su IJ, Wang CH (1991) 13-cis-retinoic acid in-duces cellular differentiation and durable remission in refractory cu-taneous Ki-1 lymphoma. Cancer 67:2490

6. Cheng AL, Su IJ, Chen CC, et al. (1994) Use of retinoic acids in thetreatment of peripheral T-cell lymphoma: a pilot study. J Clin Oncol12:1185

7. Chou WC, Su IJ, Tien HF, et al. (1996) Clinicopathologic, cytoge-netic, and molecular studies of 13 Chinese patients with Ki-1 ana-plastic large cell lymphoma. Cancer 78:1805

8. Su IJ, Cheng AL, Tsai TF, Lay JD (1993) Retinoic acid–induced apop-tosis and regression of a refractory Epstein-Barr Virus–containing Tcell lymphoma expressing multidrug -resistance phenotype. Br JHaematol 85:826

9. Jetten AM, Kim JS, Sacks PG, et al. (1990) Inhibition of growth andsquamous cell differentiation markers in cultured human head andneck squamous carcinoma cells by all-trans retinoic acid. Int J Cancer45:195

10. Lingen MW, Polverini PJ, Bouck NP (1998) Retinoic acid and inter-feron-a act synergistically as antiangiogenic and antitumor agentsagainst head and neck squamous cell carcinoma. Cancer Res 58:5551

11. Katagiri K, Yokoyama KK, Yamamoto T, Omura S, Irie S, KatagiriT (1996) Lyn & Fgr protein-tyrosine kinases prevent apoptosis dur-ing retinoic acid–induced granulocytic differentiation of HL-60 cells.J Biol Chem 271:11557

12. Benedetti L, Grignani F, Scicchitano BM, et al. (1996) Retinoid-induced differentiation of acute promyelocytic leukemia involvesPML-RAR-a-mediated increase of type II transglutamase. Blood87:1939

13. Darwiche N, Scita G, Jones C, et al. (1996) Loss of retinoic acid re-ceptors in mouse skin and skin tumors is associated with activation ofthe ras oncogene and high risk for premalignant progression. CancerRes 56:4942

14. Sandarsen A, Claypool K, Mehta K, Lopez-Berestein G, CabanillasF, Ford RJ Jr (1997) Retinoid-mediated inhibition of cell growth withstimulation of apoptosis in aggressive B-cell lymphomas. CellGrowth Differ 8:1071

15. Hsu CY, Luh KT (1995) Cytology of pulmonary Fusobacterium nu-cleatum infection. A case report. Acta Cytol 39:114

16. Herrmann M, Lorenz HM, Voll R, Grunke M, Woith W, Kalden JR(1994) A rapid and simple method for the isolation of apoptotic DNAfragments. Nucleic Acids Res 22:5506

17. Linskens MH, Feng J, Andrews WH, et al. (1995) Cataloging alteredgene expression in young and senescent cells using enhanced differ-ential display. Nucleic Acids Res 23:3244

18. Bourlet T, Omar S, Grattard F, Pozzetto B (1997) Detection of cox-sakievirus B3 in intestinal tissue of orally-infected mice by a stan-dardized RT-PCR assay. Clinical & Diagnostic Virology 8:143

19. Bullock G, Ray S, Reed J, et al. (1995) Evidence against a direct rolefor the induction of c-jun expression in the mediation of drug-inducedapoptosis in human acute leukemia cells. Clin Cancer Res 1:559

20. Smith MA, Parkinson DR, Cheson BD, Friedman MA (1992) Retin-oids in cancer therapy. J Clin Oncol 10:839

21. Su IJ, Lay JD, Cheng AL (1994) Modulation of RAR-a, growth fac-tor genes, and oncogenes in RA-induced apoptosis of Ki-1 lymphomacell line. Int J Oncol 4:1089

22. Mangelsdorf DJ, Ong ES, Dyck JA, Evans RM (1990) Nuclear recep-tor that identifies a novel retinoic acid response pathway. Nature345:224

1450 K.-C. Wang et al./Experimental Hematology 28 (2000) 1441–1450

23. Steller H (1995) Mechanisms and genes of cellular suicide. Science267:1445

24. Furuya Y, Isaacs JT (1993) Differential gene regulation during pro-grammed death (apoptosis) versus proliferation of prostatic glandularcells induced by androgen manipulation. Endocrinology 133:2660

25. Bustelo XR, Otero A, Gomez Marquez J, Freire M (1991) Expressionof the rat prothymosin a gene during T-lymphocyte proliferation andliver regeneration. J Biol Chem 266:1443

26. Wu CG, Boers W, Reitsma PR, van Deventer SJ, Chamuleau RA(1997) Overexpression of prothymosin a, concomitant with c-myc,during rat hepatic carcinogenesis. Biochem Biophys Res Commun233:817

27. William R, Watson G, Rotstein OD, Parodo J, Bitar R, Marshall JC(1998) The IL-1b-converting enzyme (caspase-1) inhibits apoptosisof inflammatory neutrophils through activation of IL-1b. J Immunol161:957

28. Beresford PJ, Jaju M, Friedman RS, Yoon MJ, Lieberman J (1998) Arole for heat shock protein 27 in CTL-mediated cell death. J Immunol161:161

29. Urayama S, Musch MW, Retsky J (1998) Dexamethsasone protec-tion of rat intestinal epithelial cells against oxidant injury is mediatedby induction of heat shock protein 72. J Clin Invest 102:1860

30. Nanbu K, Konishi I, Mandai M, et al. (1998) Prognostic significanceof heat shock proteins HSP70 and HSP90 in endometrial carcinomas.Cancer Detect Prev 22:549

31. de Castro M, Elliot S, Kino T, et al. (1996) The non-ligand bindingb-isoform of the human glucocorticoid receptor (hGR b): tissue lev-els, mechanism of action, and potential physiologic role. Mol MedToday 2:597

32. Vamvakopoulos NC, Mayol V, Margioris AN, Chrousos GP (1992)Lack of dexamethasone modulation of mRNAs involved in the glu-cocorticoid signal transduction pathway in two cell systems. Steroids57:282

33. Monstein HJ, Nylander AG, Hakanson R (1997) Widespread tissueexpression of gastrin-binding-protein mRNA. Eur J Biochem246:502

34. Maiuri L, Raia V, De Marco G, et al. (1997) DNA fragmentation is afeature of cystic fibrosis epithelial cells: a disease with inappropriateapoptosis? FEBS Lett 408:225

35. Kondo T, Takeuchi K, Doi Y, Yonemura S, Nagata S, Tsukita S

(1997) ERM (ezrin/radixin/moesin)-based molecular mechanism ofmicrovillar breakdown at an early stage of apoptosis. J Cell Biol139:749

36. Jordan MA, Wilson L (1998) Microtubules and actin filaments: dy-namic targets for cancer chemotherapy. Curr Opin Cell Biol 10:123

37. Liang P, Pardee AB (1992) Differential display of eukaryotic mes-senger RNA by means of the polymerase chain reaction. Science257:967

38. Mao M, Yu M, Tong JH, et al. (1996) RIG-E, a human homolog ofthe murine Ly-6 family, is induced by retinoic acid during the differ-entiation of acute promyelocytic leukemia cell. Proc Natl Acad Sci US A 93:5910

39. Ohtsuki T, Hatake K, Ikeda M, et al. (1997) Expression of Src-likeadaptor protein mRNA is induced by all-trans retinoic acid. BiochemBiophys Res Commun 230:81

40. Cheng Y, Lotan R (1998) Molecular cloning and characterization of anovel retinoic acid–inducible gene that encodes a putative G protein–coupled receptor. J Biol Chem 273:35008

41. Swisshelm K, Ryan K, Tsuchiya K, Sager R (1995) Enhanced ex-pression of an insulin growth factor–like binding protein (mac25) insenescent human epithelial cells and induced expression with retinoicacid. Proc Natl Acad Sci U S A 92:4472

42. Sune C, Hayashi T, Liu Y, Lane WS, Young RA, Garcia-Blanco MA(1997) CA150, a nuclear protein associated with the RNA poly-merase II holoenzyme, is involved in Tat-activated human immuno-deficiency virus type I transcription. Mol Cell Biol 17:6029

43. Watari H, Arakane F, Moog-Lutz C, et al. (1997) MLN64 contains a do-main with homology to the steroidogenic acute regulatory protein (StAR)that stimulates steroidogenesis. Proc Natl Acad Sci U S A 94:8462

44. Borsani G, Rugarli EI, Taglialatela M, Wong C, Ballabio A (1995)Characterization of a human and murine gene (CLCN3) sharing simi-larities to voltage-gated chloride channels and to a yeast integralmembrane protein. Genomics 27:131

45. Turunen O, Winqvist R, Pakkanen R, Grzeschik KH, Wahlstrom T,Vaheri A (1989) Cytovillin, a microvillar Mr. 75,000 protein. cDNAsequence, prokaryotic expression, and chromosomal localization. JBiol Chem 264:16727

46. Honke K, Wada Y (1997) Regulation of vimentin expression andprotease-mediated vimentin degradation during differentiation of hu-man monocytic leukemia cells. Jpn J Cancer Res 88:484

Retinoic acid–induced apoptotic pathway in T-cell lymphoma

Documents

Transcript of Retinoic acid–induced apoptotic pathway in T-cell lymphoma