From bloodjournal.hematologylibrary.org by guest on October 20, 2011. For personal use only.

1996 87: 657-665

Anticoagulant effects of retinoic acids on leukemia cells

T Saito, T Koyama, K Nagata, R Kamiyama and S Hirosawa

Information about reproducing this article in parts or in its entirety may be found online at: http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#repub_requests Information about ordering reprints may be found online at: http://bloodjournal.hematologylibrary.org/site/misc/rights.xhtml#reprints Information about subscriptions and ASH membership may be found online at: http://bloodjournal.hematologylibrary.org/site/subscriptions/index.xhtml

Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American Society of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036. Copyright 2011 by The American Society of Hematology; all rights reserved.

From bloodjournal.hematologylibrary.org by guest on October 20, 2011. For personal use only.

Anticoagulant Effects of Retinoic Acids on Leukemia Cells

By Takako Saito, Takatoshi Koyama, Kaoru Nagata, Ryuichi Kamiyama, and Shinsaku Hirosawa We have recently found that all-trans retinoic acid (ATRA) upregulates thrombomodulin (TM) and downregulates tissue factor (TF) expression in acute myelogenous leukemia (AML) M3 cells (NB41and acute monoblastic leukemia cells (U9371 (Koyama et al, B/ood84:3001.1994). We have further investigated the effects of ATRA on leukemic cells freshly isolated from patients at diagnosis. Increase of TM antigen O M2 was documented in all AML cells: M (n = l), (n = 5). M3 (n = 3). M4 (n = 3). M5 (n = 31, and M6 (n = 1). Decrease of TF antigen was observed in 4 M2.1 M4, and all M3 and M5 patients. However, no TM and TF antigens were detected in all chronic lymphocytic leukemia cells (n = 3) with or without ATRA treatment. Changes of TM and TF antigen levels were associated with those of TM and TF cofactor levels on thecell surface. A stereoisomer of RA, 9 4 s RA, is a high-affinity ligand for the RA receptors (RARs) and the retinoid X receptors, although ATRA and another isomer, 1 3 4 s RA, solely bind t o RARs. We have also studied the effects of 9 4 s RA and 1 3 4 s RA on the expressions of TM and TF in NB4 and U937 cells. A relatively wide range of 9cis RA concentrations (0.01 t o 1 pmol/L) compared with ATRA was optimal for prolongation of normal plasma-based f e recalcification time (reduction o c lsurface TF activity), decrease of TF antigen, and increase of TM antigen on the surface and in the lysates of NB4 and U937 cells. Western blot analysis under nonreducing conditions showed that both ATRA and 9-cis RA markedly induced the prominent band at 75 kD of TM and reduced the band at 45 kD of TF. Northern blot analysis has shown similar changes of mRNA levels, which indicates that RAs regulate TM and TF expression in leukemic cells at transcriptional levels. Anticoagulant effects of ATRA, ie, upregulation of TM expression and downregulation of TF expression, are applied not only t o established cell lines of specific subtypes (M3 and M51 but also t o more universal AML (most cases of M3 and M5 and a part of the other types of AML) cells freshly isolated from patients. 9 4 s RA may be more effective than ATRA as an inducer of differentiation of AML M3 cells and as an anticoagulant agent for patients with certain types of AML as well. 0 1996 by The American Society of Hematology. thrombin (4,000 NIHU/mg protein) were from Sigma (St Louis, MO). 9-cis RA was kindly provided by F. Hoffmann-La Roche Ltd (Basel, Switzerland). Recombinant human soluble TM, which possesses the entire extracellular domain, was a gift from Asahi Chemical Industry CO (Fuji, Japan). Human placenta-derived F (Thromborel S ) was from Behringwerke AG (Marburg, Germany). The chromogenic substrate S2266 was from Kabi Vitrum (Stockholm, Sweden). Human protein C (inactivated and activated) and human antithrombin I11 were generously provided by the ChemSero-Therapeutic Research Institute (Kumamoto, Japan) and Green Cross (Osaka, Japan), respectively. Monoclonal mouse antihuman TM IgGs, KA-3 and KA-4, were obtained by established hybridomas techniques as previously reported? KA-3 and KA-4 recognize different epitopes of TM. KA-3 binds to N-terminal part of the epidermal growth factor-homology domain and KA-4 recognizes the serine/ threonine-rich polypeptide core near the primary thrombin binding site of TM molecule.6 Monoclonal mouse antihuman TF IgGs, 6B4 and 5G9, were provided by Corvas International Inc (Greenwich, CT). Polyclonal antihuman TM antibody was from Teijin Ltd (Tokyo, Japan). All other chemicals were reagent-grade products,

CUTE PROMYELOCYTIC leukemia (APL), acute myelogenous leukemia (AML) M3 according to the French-American-British classification, leads to disseminated intravascular coagulation syndrome (DIC) with considerably high frequency. DIC is often exacerbated by treatment with cytotoxic drugs. Recently, all-trans retinoic acid (ATRA) has been introduced for treatment of APL and has become practically and consistently effective in inducing complete remission. ATRA induces not only terminal differentiation of malignant cells but also rapid amelioration of DIC associated with marked hyperfibrinolysis. We have recently found that ATRA upregulates thrombomodulin (TM) and downregulates tissue factor (TF) expression in APL (NB4) and monoblastic leukemia (U937) cell lines. Furthermore, distinct expressions of TM and TF in human leukemic cells and their regulation by ATRA were shown. Since we have suggested a potential efficiency of ATRA as a preventive and therapeutic agent for DIC in AML M3 and acute monoblastic leukemia, AML MS, we were urged to study ATRA effects on expressions of TM and TF in primary leukemic cells freshly isolated from patients with AML. The biologic effects of ATRA are mediated by its binding to the retinoic acid receptors (RARs). 9-cis RA, which is a stereoisomer of RA, has been found to be a high-affinity ligand for RARs and retinoid X receptors (RXRs). ATRA has high binding affinity to RARs but very low-affinity to RXRs. 9-cis RA is slightly more potent inducer of differentiation of a AML M2 cell line, HL60, in vitro3 and does not induce its own catabolism. These features may distinctly be advantageous for 9-cis RA for treatment of patients with AML M3 and other diseases. Initial clinical trial of 9 4 s RA for treatment of APL is ongoing! In this study, we have examined the regulation of TM and TF expressions in patients leukemic cells and have compared the effects of 9cis RA with those of ATRA and 13-cis RA on leukemic cell lines, NB4 and U937.
MATERIALS AND METHODS Reagents. Chemicals were purchased from the following suppliers. ATRA, 13-cis RA, porcine mucosal heparin, andhuman aBlood, Vol 87, No 2 (January 15). 1996: pp 657-665

From School of Allied Health Sciences and the First Department of Internal Medicine, Tokyo Medical and Dental University, and the Department of Hematology, Musashino Red Cross Hospital, Tokyo, Japan. Submitted June 5, 1995; accepted August 30, 1995. Supported in part by a research grantfrom the Ministry of Education, Science, Sports and Culture and by a grant for intractable diseases from the Ministry of Health and Welfare of Japan. Presented in part in a Symposium at the XVthCongress of International Society on Thrombosis and Haemostasis, Jerusalem, Israel, June 1995. Address reprints requests to Takatoshi Koyama, MD, The First Department of Internal Medicine, Tokyo Medical and Dental University, 1-5-45, Yushima,Bunkyo-ku, Tokyo, 113 Japan. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. section 1734 solely to indicate this fact. 0 1996 by The American Society of Hematology.
0006-4971/96/8702-0040$3.00/0
657

From bloodjournal.hematologylibrary.org by guest on October 20, 2011. For personal use only.
658

SAlTO ET AL

" 3

M5b-1 M2-l

M3-1

M6 M2-4 M3-2 M2-3 M2-5

M2-2 M O
CLL-l 2,3 ALL-IIP

ATRA

ATRA-

.-

ATRA

(-l

(+l

(-1

(+l

Fig 1. Effects of ATRA on total TM and TF antigens in leukemic cells freshly isolatedfrom patients. Leukemiccellsisolated from patients were incubated with ATRA (1 pmol/L) for 24 hours. Total TM (A) and TF (B) antigen levds were shown on ordinates. 'DIC was documented or suspected.

and reagents were purchased from Wako Pure Chemicals (Osaka, Japan), unless otherwise indicated. Cell culture. The following human leukemic cell lines were used in this study. NB4 (APL; AML M3) was kindly provided by Dr M. Lanotte (HBpital Saint Louis, Paris, France). NB4 is characterized by the presence of chromosomal translocation t( 15; 17) withthe rearrangements of RARa and promyelocytic leukemia (PML) genes as shown in the most cases with APL and the presence of PMLA RARa fusion pr0tein.7.~ patient from whom NB4 was established suffered from DIC at diagnosi~.~ U937 wasprovided by the Japanese Cancer Research Bank (Tokyo, Japan). ATRA, 9-cis R A , and 13cis RA (MS) were dissolved in absolute ethanol to a concentration of 8.32 m m o K and further diluted into growth medium at the desired concentrations so that the final ethanol concentration in the culture media was less than 0.1 %. The control leukemia cells (those incubated without RAs) were exposed to an appropriate amount of ethanol in the culture medium. Whether primary leukemia cells were analyzed immediately after isolation or after 24 hours of culture, no significant differences of TM and TF expressions were generally observed. All the procedures involving RAs were performed in subdued lights. Leukemic cells freshly isolated from patients. Peripheral blood was drawn from 16 patients with AML (MO; M2-1, -2, -3, -4, and -5; M3-1, -2, and -3; M4-1, -2, and -3; M5a; M5b-1and -2; and

M6), 2 patients with acute lymphoblastic leukemia (ALL; ALL-l and -2) and 3 patients with chronic lymphocytic leukemia (CLL; CLL- 1, -2, and -3). Peripheral blood samples were obtained at diagnosis before the start of chemotherapy and the mononuclear fraction was isolated by sedimentation on Ficoll-Hypaque density gradients. Blood samples were selected for study only when they were shown to contain more than 85% blasts. Mononuclear cells were then washed three times in phosphate-buffered saline (PBS) and suspended in RPM1 1640 medium (Nissui Pharmaceutical CO, Tokyo, Japan) supplemented with 10% fetal calf serum. Measurement of TM and TF antigens. Leukemic cells were incubated with agents for the times indicated and washed with PBS three times. The cell numbers were counted and adjusted. Cells were then extracted for 30 minutes with 0.5% Triton X-100 in PBS at 4'C. Cell debris was removed by centrifugation at 12,000g for 20 minutes. Cell lysates were stored at -80C until the assay. Total TM antigen in cell lysates was measured by enzyme-linked immunosorbent assay (ELISA) using an antihuman TM monoclonal antibody U - 3 and a polyclonal antihuman TM antibody as reported before."' Total TF antigen in cell lysates was also measured by ELISA using antihuman TF monoclonal antibodies 6B4 and 5G9 as already reported." Cell surface TM cofactor activity. TM cofactor activity on cell surface was measured as described previously.' Surface TM cofactor activity was determined by the hydrolysis of chromogenic substrate

From bloodjournal.hematologylibrary.org by guest on October 20, 2011. For personal use only.
OF ANTICOAGULANT EFFECTS

A
kD

kD

662009766451
ATRA(-) 45-

31 22 -

2
(+)

7
(-)

8
(+)

9 1 0 1 1 2 3 1

(-1

(+l (-1

(+l

(-1

(+l
rec.TM

(-1

(+)

placenta

M4-3

M5b-2

M3-1

M5b-1

M5a

TF

Fig 2. Western blot analysis of patients' leukemic cell preparationswith anti-TM andanti-TF antibodies. Leukemic cells (10'1 were lyredin detergent and the preparations were subjected t o sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE1under nonreducing conditions. After transfer t o lmmobilon (Millipore Corp, Bedford, MA), immunoblotting wasaccomplished with a monoclonal anti-TM antibody KA-3 (A) and a monoclonal anti-TF antibody 5G9 (B). Each set of leukemic cells was treated with p n o l / L ATRA for 24 hours. The prominent 1 band at approximately75 kD represents TM (A). As a control, soluble recombinant TM was used (A, lane 11). A monoclonal anti-TF antibody 5G9 identified thenonreduced form of TF as a band at approximately45 kD (B). As a control, placenta TF was used (B, lane 31.

S-2266. The results were expressed as Am0D4,,.Jminitue or as a percentage of the initial velocity of activated protein C formation (100% for the formation with the cell surface TM under basal conditions). Controls of the cells in the absence of thrombin and protein C were treated similarly and no activation of protein C was observed. Cell surface TF cofactor activity: anai.vsis of procoagulant acrivity in clotting assay. Leukemic cell suspensions were adjusted to 1 X 107/mL in PBS. Cells ( I X IO6) were added to 0.1 mL ofpooled human normal plasma. After incubation at 37C for 3 minutes, 0.1 mL of 25 mmol/L calcium chloride was added and plasma recalcification time was determined with a semiautomatic coagulometer CA1 0 0 (Sysmex, Kobe, Japan). Because our previous report has already shown that the procoagulant activities on the surface of NB4 and

U937 cells determined as described above are deduced from TF expression, prolongation of recalcification time ismainly due to downregulation of TF expression by RAs.' TF cofactor activity was quantitated by reference to standard curves (log-log plot) constructed with human placenta TF and TF activity, that yielded a 50-second recalcification time was defined as 1 UlmL. Analysis o surface TM and TF antigens byflow. cvtometric anal?f sis. After incubation for 24 hours with or without I pmol/L ATRA, 9 4 s RA, and 1 3 4 s RA, NB4 cells and U937 cells were washed with PBS and IO" cells were resuspended in I mL PBS. The suspensions were incubated with an inhibitory monoclonal anti-TF antibody 6B4 or an anti-TM antibody KA-4 at a concentration of 10 &mL for 30 minutes at 4C. and standard indirect immunofluorescence

A
300 M4-3
1

.
*S
0

200-

MSb-2

E
M4-2
MS1

.E .W
c

(D Fig 3. Cell surface TM and TF 100activitiesof patients' leukemic cells modulated by ATRA. Leukemic cell samples freshly obtained from several patients were incubated with 1 pmol/L ATRA for 24 hours. Cell surface TM (AI and TF IBI activities were I determinedfor 10' suspended ATRA cells as described in the Materials and Methods.

3 "
T

l
ATRA

ATRA

CLL-1

ATRA

(-1

(+l

(-1

(+l

From bloodjournal.hematologylibrary.org by guest on October 20, 2011. For personal use only.
660 SAITO ET AL

150 -

A
cn
c

100-

10-

l -

l& l&

9 4 s RA (PM)

l&

lo"

C
IOW -

z 3
E
3

ATRA

E
oooOciaRA

T
8

2
t

E
500-

j
RA

JNB4

U .

U .

I us37

10-4

10-3

IO-*

10-1

I 5 @M)

(-1

all-bm. 9-ci8 RA(+) RA(+)

134s RA(+)

Fig 4. Effects of RAs on total TM and TF antigens in NB4 cell lysates and cell surface TF activity of NB4 and US37 cells. Concentration dependency of S-cis RA on the TM (A) and TF (B) antigenexprsssion in NB4 cell lysates and cell surfaceTF activity of NB4 cells (C) were determined. NB4 cells were incubated with various concentrations 9of cis RA (O.OOO1 to B pmollL) for 24 hours. Furthermore, the effects of ATRA, S-cis R , and 13A TF cis RA on cell surface activity were compared (Dl. NB4 and US37 cells were incubated with 1 pmollL RAs ( A M , S-cis R , A and 13-cis R A ) for 24 hours.

techniques were used for the analysis of surface TM and TF antigens." Immunoblorting analysis. Western blot analyses of TM and TF in cell lysates were performed essentially as described previously.' KA-3 and horseradish peroxidase-conjugated rabbit antimouse IgG antibody were used for the detection of TM. Horseradish peroxidaseconjugated anti-TF monoclonal antibody 5G9 was usedfor the detection of TF. The same amount of lysates from lo6 cells was applied to each lane. Northernblotting. Total cytoplasmic RNAwas prepared from NB4 and U937 cells after treatment with ATRA, 9 4 s RA, or 13cis RA for 5 hours by the acid guanidium thiocyanate-phenol-chloroform method. Isolated RNA was analyzed in Northern blotting using Digoxigenin Luminescent Detection Kit (Boehringer Mannheim, Mannheim, Germany). A T t h l l l I-Nhe I fragment in the coding region of the human TM gene and a TF cDNA probe were kind gifts from Dr K. Nawa and Dr K. Wakita, respectively (Molecular Biology Research Laboratory, Daiichi Pharmaceutical CO, Tokyo, Japan), and were used as hybridization probes. The TF cDNA probe was prepared by a polymerase chain reaction (PCR) method from a cDNA library of tumor necrosis factor a-stimulated human umbilical vein endothelial cells.'*
RESULTS

Effects of ATRA on TM and TF antigens in leukemic cells When untreated, TM and freshly isolated frompatients. TF antigen levels in freshly isolated leukemic cells were varied in the range of 0 to 865 ng TM protein and 0 to 9.24 ng TF protein per lo7 cells. After incubation with ATRA, total TM antigen was increased in all 16 AML cells (Fig

1A). In contrast, ATRA decreased TF antigen levels in I1 of 16 AML cells (M2-l, -2, -4, and -5; M3-1, -2, and -3; M4-3; M5a; and M5b-1 and -2; Fig 1B). Both TM and TF expression in all CLL cells was not detected before and after ATRA treatment. In 2 patients with ALL, TM expression in the leukemic cells was not detected before and after treatment, whereas TF expression was documented in 1 patient. The cases in which DIC was documented or suspected are marked with an asterisk in Fig 1B. Western blot analysis for TM and TF in leukemic cells from patients. The presence of TM and TF molecules in the patients' leukemic cells was determined by immunoblotting with the antibodies specific for human TM and TF. The monoclonal antibody KA-3 identified the nonreduced form of TM as a prominent band at approximately 75 kD. After treatment of the leukemic cells with ATRA, TM markedly increased in all AML cells (M3-1, M4-3, and all M5 [M5a and M5b-1 and -21 cells) examined (Fig 2A, lanes 2, 4, 6, 8, and 10).The monoclonal antibody against human TF 5G9 identified the nonreduced form of TF in M3-l cells as a band at approximately 45 kD (Fig 2B, lane 1). TF of the leukemic cells was markedly decreased after treatment with ATRA (Fig 2B, lane 2). Cell surface TM and TF activities o patients' leukemic f cells modulated by ATRA. Effects of ATRA on Th4 and TF activities on the surface of leukemic cell samples freshly obtained from several patients are shown in Fig 3. After treatment with ATRA, TM activities were increased on the

From bloodjournal.hematologylibrary.org by guest on October 20, 2011. For personal use only.
ANTICOAGULANT EFFECTS OF RETlNOlCACIDS
661

B
NB4
c

so}

U937

1
l

1
4
4
*B LL

I
3
1
T

1. control 2 all-rrens(+) .

3. 9 4 s (+) 4 134s (+) .

"1
Fig 5. Effects o RA. on total TM and TF antigens f in NB4 andU937 c e l l s . NB4 (A and C) andU937 (B and D) cells were incubated with l WmollL ATRA (Iana 2). 9-cia RA (lana 31, and 134s RA (lane 4) for 24 houn. Lane lrepfwents the control. TotalTM (A and B) and TF (C and D) antigen levels are shown on the ordinates.

NB4
T

U937
1. control 2 all-rrans(+) . 3 %cis (+) . 4 1 3 4 s (+) .

8 .g
m

P 5 e

h1l 2L
1 2 3 + m O 1 1 2 3 4

surface of all AML cells studied (Fig 3A). TF activities on the surface of 3 patients' cells (M3-1, M4-3, and M5b-2) were significantly reduced, whereas those of another 2 patients with CLL-1 and M4-2 were not significantly changed (Fig 3B). Effects of 9-cis RA on total T M and TF antigens in NB4 cells lysates (dose-dependency). The effects of a stereoisomer, 9-cis RA, on the expression of TM and TF in NB4 cells were examined. 9-cis RA increased TM antigen level while decreasing the TF antigen level dose-dependently. A relatively wide range of 9-cis RA concentrations, ie, 0.01 to 1 pmol/L, which is an easily attainable level in patients treated with 9-cis RA, induced an efficient decrease of TF antigen and an increase of TM antigen (Fig 4A and B). Cell su$ace TF activity of NB4 cells modulated by 9-cis RA and ATRA. Effects of 9-cis RA and ATRA on TF activity on the surface of NB4 cells were examined. As shown in Fig 4C, a relatively wide range of 9 4 s RA concentrations (0.01 to 1 pmol/L) was optimal for reduction of TF activity when compared with ATRA. Cell surface TF activity o NB4 and U937 cells modulated f by R A s . When NB4 and U937 cells were pretreated with 1 pmol/L of ATRA, 9-cis RA, or 13-cis RA for 24 hours,

the TF activities were reduced by all RAs (Fig 4D).9-cis RA is slightly more effective than ATRA and 13-cis RA. Effects of ATRA, 9-cis R A , and 13-cis RA on T M antigen in NB4 and U937 cells. In NB4 cells, ATRA, 9-cis RA, and 13-cis RA at a concentration of 1 pmoVL increased TM antigen to similar levels (Fig 5A). The TM antigen levels in U937 cells were increased by treatment with ATRA, 9-cis R A , and 13-cis RA. 9-cis RA was most effective for increase of the TM antigen level (Fig 5B). Effects of ATRA, 9-cis R A , and 13-cis RA on TF antigen in NB4 and U937 cells. ATRA, 9-cis RA, and 13-cis RA (1 pmoVL) reduced TF antigen levels in NB4 and U937 cells (Fig 5C and D). In both NB4 and U937 cells, 9-cis RA was slightly more potent than ATRA at 1 pmol/L, the concentration at which ATRA decreases expression of TF most efficiently. F C M analysis of T M and TF antigens on the cell sugace of NB4 and U937 modulated by RAs. We found 40.8% of the NB4 cells positive for the anti-TM antibody KA-4 by FCM analysis (Fig 6A). The percentage of positive cells for KA-4 was increased to 94.1%, 92.8%, and 94.1% by treatmentwith ATRA, 9-cis RA, and 13-cisRA, respectively (Fig 6B, C, and D). We found 66.4% of the NB4 cells to

From bloodjournal.hematologylibrary.org by guest on October 20, 2011. For personal use only.
662

SAITO ET AL

control

all-trans

Qcis

134s

TF

control

all-trans

Qcis

134s

Fig 6. Flow cvtometricanalysis of surface M and T TF antigensin NB4 and U937 cells modulatedby RAs. NB4 and U937cells were untreated (A and E) or treated with l pmollL ATRA (B and F1,9-cis RA (C and G), and 1 3 4 s RA (D and H). The flow cytometric analyses of both untreated and treated cells incubated with KA-4 (anti-TM: A, B, C, and D) or 684 (antiTF; E, F, G, and H) monoclonal antibodyfollowed by goat FITC-antimouse IgGwere shown. The abscissa shows fluorescence intensities and the ordinate shows the number cells. of

be positive for the anti-TF antibody 6B4 (Fig 6E), whereas the percentage of positive cells decreased to 9.5%, 9.0%, and 7.4% when treated with ATRA, 9-cis RA, and 13-cis RA, respectively (Fig 6F, G, and H). We also found 79.1 % of the untreated U937 cells to be positive for KA-4. When treated with ATRA,9-cis RA, and 13-cis RA, 99.2%, 99.5%, and 98.7% of U937 cells were positive for KA-4, respectively. Of the untreated U937 cells, 18.0% were positive for 6B4, whereas the percentage the positive cells decreased to 3.1%, 1.1%, and 3.7% when treated with ATRA, 9-cis RA, and 13-cis RA, respectively. Enhancement of sur.$ace TM activity in NB4 and U937 cells by 9-cis RA. Surface TM activity induced by 9-cis RA was determined by the acceleration of thrombin-catalyzed protein C activation. 9-cis RA increased the rate of protein C activation to about threefold in NB4 cells and to about 2.5-fold in U937 cells (Fig 7A and B). Basal AOD40s,m/min levels were 0.036 -C O.O03/min/5 X lo6 NB4 cells (34 2 2 ng activated protein C/106 cells) and 0.108 2 0.004/min/5 X lo6U937 cells (87.6 5 3 ng activated protein C/106cells). Western blot for TM: effects of RAs. After treatment of NB4 and U937 cells with 9-cis RA and 13-cis RA, a prominent band of TM at 75 kD was markedly increased to the level as high as when the cells were treated withATRA (data not shown).

Northern blot for TM and TF: effects o RAs. Increased f expression of specific mRNA for TM was detected as a single band at 3.7 kb when NB4 and U937 cells were treated with RAs (Fig 8A, lanes 2, 3, and 4). Expression of specific mRNA for TF found as a single band at 2.3 kb was markedly decreased when NB4 cells were treated with RAs (Fig 8B, lanes 2, 3, and 4). In leukemic cells freshly isolated from patients, changes of TM and TF mRNAs by ATRA were similar to those of NB4 and U937 cells (data not shown).
DISCUSSION

In the present study, we have shown that anticoagulant effects of RAs are inducible not only to specific subtypes (M3 and M5) of the established cell lines but also to certain AML subtypes (most cases of M3 and M5 and a part of the other types of AML) of the leukemic cells freshly isolated from patients. Constitutive expression levels of TM and TF were vaned in untreated leukemic cells that were freshly isolated. Although a rough correlation between relatively high TF antigen expression and the development of DIC seems to be present, as shown in Fig 1, anticoagulant glycoprotein TM is also expressed variedly in leukemic cells. Thus, development of DIC can be estimated by evaluating overall procoagulant and anticoagulant activities. Although overall procoag-

From bloodjournal.hematologylibrary.org by guest on October 20, 2011. For personal use only.
ANTICOAGULANT EFFECTS OF RETlNOlC ACIDS
663

B
U937 :TM

Fig 7. Changes in TM cofactor activity for protein C activation on NB4 and U937 cell surfaces after exposure to 9-cis RA (1 pmollL1. Surface TM activity was determined for 5 x lo6 suspended cells as described in the Materials and Methods. Results were expressed as the percentage of initial velocity of activated protein C (lane 2)compared with the control (lane 1) that was not treated with the reagent.

11
1 2

1. control 2 . 9 4 s RA(+)

ulant activity in leukemic cells may be controlled mainly by levels of TF and TM expressed, expressions of the other procoagulant and anticoagulant proteins should be considered to be involved in DIC. The presence of a unique procoagulant with cysteine proteinase activity distinct from TF in leukemic cells including NB4 cells has been reported.13 Differentiation therapy with RA induces complete remission in APL patients, with a rapid correction of the bleeding tendency. ATRA therapy exerts rapid improvement in coagulopathy and hyperfibrinolysis, with normalization of plasma fibrinogedfibrin degradation product (fragment E),

cross-linked fibrin degradation product (D-dimer fragment), thrombin-antithrombin 111 complexes, and plasmin-a2-plasmin inhibitor c ~ m p l e x . Several possible explanations for '~ marked reduction of hemorrhagic complications byATRA are proposed. ( I ) The differentiation process induced by ATRA does not exacerbate the release of procoagulant and fibrinolytic substances from the leukemic cell, as cytotoxic drugs do. (2) Rapid amelioration of DIC is induced ATRA by through downregulation of TF and cancer procoagulant expressions in APL cells and upregulation of TM expression in APL andendothelial cells.'.'' (3) Rapid reductionof elevated fibrinolytic and proteolytic activities in APL cells is exerted byATRA."." Upregulation ofTM antigen and activity expression by ATRA was shown in all primary leukemic cells from AML patients studied, although the levels of induction were different in each leukemic cells. ATRA decreased the level of TF antigen expression in 1 1 of 16 AML cells. Cell surface TF activities measured by recalcification time were reduced in 3 of 4 AML cells. These data suggest that upregulation of TM and downregulation of TF, as shown in NB4 and U937 cells, are major factors in the improvement of DIC. In this context, the past and thefuture clinical trials of RAs for AML and other malignant cells to induce differentiation should be evaluated from the coagulation andfibrinolysis points of view. However, in ALL and CLL cells, no TM expression was documented in the absence and presence of ATRA. In CLL cells, TF expression was not documented either. In 1 of 2 patients withALL, TF expression was documented, whereas downregulation of its expression by ATRA was not shown. These observations suggest that differentiation into mature lymphocytic cells implies the loss ofTMand TF inducibility but that premature lymphoblastic cells may express TF. TM may possess not only natural anticoagulant properties but also antifibrinolytic properties by acceleration of thrombin inactivation of single-chain urokinase-type plasminogen activator (U-PA).".'~Upregulation ofTMmay also reduce hyperfibrinolysis induced by U-PA in patients with APL,'" which leads to severe hemorrhagic complications. Antifibrinolytic activity of TM should be further evaluated. In this investigation, we have also shown upregulation of

A
kb

B
kb +TM
I.

5.32.8Fig 8. Northern blot analysis: effects of RAs. Twenty microgramsof total RNA was extracted from cultured NB4 and U937 cells after exposure to 1 pmol/L ATRA (lane 2). 9-cis RA (lane 3). and 13-cis RA (lane 4) for 5 hoursandelectrophoresedon a 1% agarose/formaldehydegel. The blot was probed with a digoxigenin-labeled cDNA fragment specifically encoding TM (A) or TF (B). Lane l 1 2 3 4 1 represents untreated control. RNAmolecular weight markers are given along the left margin. 1.9-

1.9-

l .o-

l .o-

4 1

NB4

U937

NB4

From bloodjournal.hematologylibrary.org by guest on October 20, 2011. For personal use only.
664 SAITO ET AL

TM expression and downregulation of TF expression in NB4 andU937 cell lines by stereoisomer RAs. 9-cis RA was slightly more effective than other types of RAs at 1 pmoV L, the concentration at which ATRA has its most optimal effect. In the experiment of RAs for the effects on total TM and TF antigens in the cell lysates and procoagulant activities on the surface of NB4 cells, 9-cis RA was especially more effective in a relatively wide range of concentrations (0.01 to 1 pmol/L) than was ATRA. Such concentrations are easily attainable level in patients treated with 9 4 s RA. In human umbilical vein endothelial cells, NB4 cells, and U937 cells, ATRA increases transcription of TM gene without changing TM mRNA stability, whereas cyclic AMP increases TM mRNA stability (our unpublished observations). RAs thus regulate TM expression in leukemia cells at transcriptional level, but it is not yetknown whether RAs change TF mRNA stability. Because ATRA upregulates TM expression in all primary AML cells, TM inducibility may not depend on cell differentiation by ATRA. The TM antigen level induced by ATRA in NB4 cells reaches a maximum level within a few days. Furthermore, a recent report has shown that TM induction in AML M2 cell line HL60 cells cultured with ATRA reflecting TM biosynthesis levels is independent of HL60 differentiation into neutrophilic cells. However, according to a recent report using RA-resistant NB4 cells, NB4 cells sensitivity to maturation by ATRA may be crucial to the loss of their procoagulant activity, including TF and cancer procoagulant activities. More detailed analysis of cell differentiation and TM/TF expressions may be necesS T .

ples. The excellent skillful technical assistance of Misako Shibakura is gratefully acknowledged.
REFERENCES

Retinoids derived from retinol (vitamin A) are critical regulators of vision, cell proliferation, differentiation, and embryonal morphogenesis. Intracellular isomerases further convert some intermediates to 9-cis, 1l-cis, and 13-cis RAs. 13-cis RA and 9 4 s RA are naturally occumng ligands of the nuclear RARs. The regulation of RA-target genes is caused by the mediation of the nuclear receptors that are capable of binding to RA-responsive elements (RAREs) of a direct repetition of two 5-PuGG(G/T)TCA-3. High-affinity binding of RARs to RAREs takes place only if a separate retinoid receptor family, RXRs, is present. ATRA binds to RARs with high affinity, but ATRA and 13-cis RA bind to RXRs with very low affinity. On the other hand, 9 4 s RA has high binding affinity to both RARs and RXRs. This may be one of the reasons why 9-cis RA is more effective than ATRA and 1 3 4 s RA in the regulation ofTM and TF in the leukemic cells. 9-cis RA may be slightly more effective than ATRA not only as an inducer of differentiation of AML M3 cells but also as an anticoagulant agent for treatment of patients with certain types of AML. Identification of specific pathways responsible for the anticoagulant effects of RAs, including ATRA and 9-cis RA, are intriguing and indispensable subjects for further investigation.
ACKNOWLEDGMENT

The authors are gratefully indebted to emeritus Prof Nobuo Aoki and Dr Shuji Tohda (Tokyo Medical and Dental University, Tokyo, Japan) for helpful discussions and supplying us some patient sam-

I . Koyama T, Hirosawa S, Kawamata N, Tohda S, Aoki N: Alltrans retinoic acid (ATRA) upregulates thrombomodulin and downregulates tissue factor expression in acute promyelocytic leukemia cells: Distinct expression of thrombomodulin and tissue factor in human leukemic cells Blood 84:300l, 1994 2. Heyman RA, Mangelsdorf DJ, Dyck JA, Stein RB, Eichele G, Evans RM, Thaller C: 9-cis retinoic acid is a high affinity ligand for the retinoid X receptor. Cell 68:397, 1992 3. Kizaki M, Ikeda Y, Tanosaki R, Nakajima H, Morikawa M, Sakashita A, Koeffler HP: Effects of novel retinoic acid compound, 9-cis-retinoic acid, on proliferation, differentiation, and expression of retinoic acid receptor-a and retinoid X receptor-a RNA by HL60 cells. Blood 82:3592, 1993 4. Miller WH Jr. Jakubowski A, Tong WP, Miller VA, Rigas JR. Benedetti F, Gill GM, Truglia JA, Ulm E, Shirley M, Warrell RP Jr: 9-cis retinoic acid induces complete remission but does not reverse clinically acquired retinoid resistance in acute promyelocytic leukemia. Blood 85:3021, 1995 5. Kimura S, Nagoya T, Aoki N:Monoclonal antibodies to human thrombomodulin whose binding is calcium dependent. J Biochem 105:478, 1989 6. Koyama T, Parkinson JF, Aoki N. Bang NU, Muller-Berghaus G, Preissner KT: Relationship between post-translational glycosylation and anticoagulant function of secretable, recombinant mutants of human thrombomodulin. Br J Haematol 78:515, 1991 7. Lanotte M, Martin-Thouvenin V, Najiman S, Balerini P, Valensi F, Berger R: NB4, a maturation inducible cell line with t(15; 17) marker isolated from a human acute promyelocytic leukemia (M3). Blood 77:1080, 1991 8. Dermime S, Grignani F, Clerici M, Nervi C, Sozzi G, Talamo GP, Marchesi E, Formelli F, Panniani G, Pelicci PG, GambacortiPasserini C: Occurrence of resistance to retinoic acid in the acute promyelocytic leukemia cell line NB4 is associated with altered expression of the pml/RARa protein. Blood 82:1573, 1993 9. Yamamoto K, Hirosawa S, Sakamaki H, Aoki N: Frequent rearrangements of retinoic acid receptor alpha gene and My1 gene, and rare mutations of RAS and F M S genes in acute promyelocytic leukemia. Am J Hematol 40:252, 1992 10. Obashi K, Hirokawa K, KomatsuN, Aoki N: The biosynthesis of thrombomodulin and its enhancement by dibutyryl CAMP in a human megakaryoblastic cell line, UT-7. Int J Hematol54:341, 1991 I 1. Koyama T, Nishida K, Ohdama S, Sawada M, Murakami N, Hirosawa S, Kuriyama R, Matsuzawa K, Hasegawa R,Aoki N: Determination of plasma tissue factor antigen and its clinical significance. Br J Haematol 87:343, 1994 12. Wakita K, Marumoto Y, Horiuchi T: Up-regulation of tissue factor mRNA in humanumbilical vein endothelial cells by calmodulin inhibitor W7. Thromb Res 73:177, 1994 13. Falanga A, Consonni R, Marchetti M, Mielicki WP, Rambaldi A, Lanotte M, Gordon SG, Barbui T: Cancer procoagulant inthe human promyelocytic cell line NB4 and its modulation by all-trans retinoic acid. Leukemia 8:156, 1994 14. Kawai Y, Watanabe K, Kizaki M, Murata M, Kamata T, Uchida H, Moriki T, Yokoyama K, Tokuhira M, Nakajima H, Handa M, Ikeda Y: Rapid improvement of coagulopathy by all-trans retinoic acid in acute promyelocytic leukemia. Am J Hematol 46:184, 1994 15. Tapiovaara H, Matikainen S, Hurme M, Vaheri A: Induction of differentiation of promyelocytic NB4 cells by retinoic acid is associated with rapid increase in urokinase activity subsequently downregulated by production of inhibitors. Blood 83: 1883, 1994

From bloodjournal.hematologylibrary.org by guest on October 20, 2011. For personal use only.
ANTICOAGULANT EFFECTS OF RETlNOlC ACIDS 665

16. Dombret H, Scrobohaci M-L, Daniel M-T, MiclCa J-M, Castaigne S, Chomienne C, Fenaux P, Degos L: In vivo thrombin and plasmin activities in patients with acute promyelocytic leukemia (APL): Effect of all-trans retinoic acid (ATRA) therapy. Leukemia 9: 19, 1995 17. DeMunk G, Groeneveld E, Rijken D: Acceleration of the thrombin inactivation of single-chain urokinase-type plasminogen activator (pro-urokinase) by thrombomodulin. J Clin Invest 88:1680, 1991 18. Molinari A, Giorgetti C, Lansen J, Vaghi F, Orsini G, Faioni EM, Mannucci PM: Thrombomodulin is a cofactor for thrombin degradation of recombinant single-chain urokinase plasminogen acti-

vator in vitro andin a perfused rabbit heart model. Thromb Haemost 67:226, 1992 19. Bennett B, Booth NA, Croll A, Duncan A, Dawson AA: The bleeding disorder in acute promyelocytic leukaemia: Fibrinolysis due to U-PA rather than defibrination. Br J Haematol 71:511, 1989 20. Kizaki K, Ishii H, Hone S , Kazama M: Thrombomodulin induction by all-trans retinoic acid is independent of HL-60 cells differentiation to neutrophilic cells. Thromb Haemost 72573, 1994 21. Falanga A, Consonni R,Oldani E, Marchetti M, GambacortiPasserini C, Lanotte M, Barbui T Effect of all-trans-retinoic acid on the procoagulant activity of retinoic acid-resistant human promyelocytic cell lines. Thromb Haemost 73:1280, 1995 (abstr)