Regulation of Thrombomodulin Expression by All-trans Retinoic Acid

Regulation of Thrombomodulin Expression by All-trans Retinoic Acid and Tumor Necrosis Factor-a: Differential Responses in Keratinocytes and

Endothelial Cells By Thomas J. Raife, Elaine M. Demetroulis, and Steven R. Lentz

Thrombomodulin is a cell-surface anticoagulant glycopro- tein expressed by vascular endothelial cells and epidermal keratinocytes. Thrombomodulin expression in endothelial cells is regulated by retinoic acid and tumor necrosis factor- a (TNF), agents that also modulate epidermal differentiation. We examined thrombomodulin function and regulation of thrombomodulin expression by all-trans retinoic acid (ATRA) and TNF in human keratinocytes and endothelial cells. Untreated keratinocytes and endothelial cells expressed thrombomodulin of comparable activity and apparent thrombin affinity. Incubation of keratinocytes with 10 pmol/L ATRA for 24 hours increased thrombomodulin activity 5.4 f 0.9-fold (mean 5 SE), with equivalent increases observed in thrombomodulin protein (5.5 f 2.1-fold) and

HROMBOMODULIN IS A membrane-bound glyco- protein that functions as an endogenous anticoagulant

by altering the substrate specificity of the serine protease thrombin.’ When bound to thrombomodulin, thrombin’s procoagulant activities are inhibited, and thrombin and thrombomodulin form a catalytic complex that efficiently activates the anticoagulant protein C. Activated protein C enzymati- cally degrades coagulation factors V, and VIIIa, resulting in decreased thrombin production and inhibition of coagulation.

Thrombomodulin is expressed on the luminal surface of endothelium and syncitiotrophoblast, where it functions to limit coagulation to sites of vascular injury.’ Thrombo- modulin is also expressed by several types of nonvascular cells, including mesothelium3 and spinous layer keratinocytes of squamous epitheli~m.~” During embryonic development, thrombomodulin is expressed in both vascular and nonvascular sites, including developing lung bud, heart valves, and neur~epithelium.~.~ Homozygous germline dis- ruption of the thrombomodulin gene in mice results in lethality early in embryonic development.” These observations raise the possibility that thrombomodulin may function dif- ferently on vascular and nonvascular cells.

Expression of thrombomodulin in cultured endothelial cells is downregulated by inflammatory cytokines such as tumor necrosis factor (TNF) and interleukin-l.””s In contrast, exposure of endothelial cells to retinoic acid results in upregulation of thrombomodulin transcription and in abroga- tion of the downregulating effects of inflammatory cytokine^.'^.'' Although epidermal differentiation is known to be modulated by TNF and retinoic acid,I8.l9 the effect of these agents on thrombomodulin expression in keratinocytes is unknown.

In human keratinocytes, thrombomodulin expression increases when the concentration of calcium ions in the culture medium is increased from 0.07 mmoVL to l .4 mrn~VL.~ This regulatory effect of calcium may be important physiologi- cally, as an extracellular [Ca”] gradient may exist in the epidermis.”.” In addition to regulating thrornbornodulin expression, Ca” may modulate thrombomodulin function be- cause thrombomodulin isolated from different cell types var- ies in thrombin affinity and [Ca”] dependence of protein C

T

Blood, Vol 88, No 6 (September 15). 1996 pp 2043-2049

mRNA (4.2 f 1.2-fold). Incubation of keratinocytes with 1.0 nmol/L TNF markedly increased expression of keratinocyte transglutaminase, but had no effect on thrombomodulin activity, protein, or mRNA. In endothelial cells, ATRA produced a small increase in thrombomodulin activity (1.9 5 0.1-fold), and incubation with TNF for 24 hours decreased thrombomodulin activity 83% f 7%. The activity profile of keratinocyte thrombomodulin exhibited a distinct maximum near 1.0 mmol/L Ca2+. These results demonstrate that keratinocyte thrombomodulin is regulated by retinoids and CBZf, but not by TNF, and that regulation of thrombomodulin expression differs in keratinocytes and endothelial cells. 0 1996 by The American Society of Hematology.

activation.** However, the effect of [Ca”] on the activity of keratinocyte thrombomodulin has not been determined.

In this study, we measured the thrombin affinity of thrombomodulin in keratinocytes and endothelial cells, and compared the effects of all-trans retinoic acid (ATRA) and TNF on thrombomodulin expression in these cell types. In addition, we determined the [Ca”] dependence of protein C activation by thrombin in the presence of keratinocyte thrombomodulin. The results indicate that keratinocytes and endothelial cells express thrombomodulin with similar apparent thrombin affinity, but that ATRA is a more potent stimulus of thrombomodulin expression in keratinocytes than in endothelial cells. In contrast to ATRA, TNF markedly inhibited thrombomodulin expression in endothelial cells, but not in keratinocytes. The activity profile of keratinocyte thrombomodulin exhibited a distinct maximum near 1.0 mrnol/L Ca’+, a concentration that may support optimal activity in the spinous layer of epidermis.

MATERIALS AND METHODS

Materials. ATRA was purchased from Sigma Chemical CO (St Louis, MO), and TNF was purchased from Genzyme Corp (Cam- bridge, MA). Human thrombin was purchased from Enzyme Re- search Laboratories (South Bend, IN). Human protein C and anti-

From the Departments of Pathology and Internal Medicine, Uni- versity of Iowa College of Medicine, Iowa City; and the Veterans Affairs Medical Center, Iowa City, IA.

Submitted December 14, 1995; accepted May I , 1996. Supported in pari by National Institutes of Health Grants No.

HL-07344 and DK-25295, the Department of Veterans Affairs, and American Cancer Society Grant No. IN-I22P, administered through the University of Iowa Cancer Center.

Address reprint requests to Steven R. Lentz, MD, PhD, Depart- ment of Internal Medicine, C303 CH, The University of Iowa, Iowa City, IA 52242.

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 I996 by The American Society of Hematology. oooS-4971/96/8806-0012$3.00/0

2043

For personal use only.on April 4, 2019. by guest www.bloodjournal.orgFrom

http://www.bloodjournal.org/

http://www.bloodjournal.org/site/subscriptions/ToS.xhtml

2044 RAIFE, DEMETROULIS, AND LENTZ

thrombin 111 were generous gifts of Dr Hans Peter Schwarz (Immuno AG, Vienna, Austria) and Dr Douglas Tollefsen (Washington Uni- versity, St Louis, MO), respectively. Rabbit lung thrombomodulin was purchased from American Diagnostica, Inc (Greenwich, CT). Recombinant TM456 was purified from tissue culture medium as described previously.” Porcine intestinal heparin was obtained from Elkins-Sinn, Inc (Cherry Hill, NJ), and S-2366 was purchased from Kabi Pharmacia Hepar, Inc (Franklin, OH).

Cell culture. Human umbilical vein endothelial cells (HUVEC) and neonatal human foreskin keratinocytes were purchased from Clonetics Corporation (San Diego, CA). All cells were studied within 2 to 4 passages after primary culture. HUVEC were cultured in endothelial growth medium containing 2% fetal bovine serum (FBS) as described previo~s ly .~ Keratinocytes were cultured in serum-free keratinocyte growth medium (KGM) containing 0.07 mmol/L calcium chloride. When keratinocyte cultures reached 75% confluency, the medium was changed to KGM containing 1.4 m m o l n calcium chloride, and the cells were cultured for up to 96 hours in the presence of various concentrations of ATRA or TNF. Control cultures were incubated with either 0.1% dimethyl sulfoxide (DMSO; the vehicle for ATRA) or 0.3% phosphate-buffered saline (PBS; the vehicle for TNF).

Protein C activation. Protein C activation was measured using a two-stage assay described previ~us ly .~ Cells were washed three times with PBS and lysates were prepared in 20 mmol/L Tris-HCI, pH 8.0,0.6% Triton X-100, and 100 mmol/L NaCI. The total protein concentration of cell lysates was measured by a modified Bradford assay (Bio-Rad Laboratories, Richmond, CA). For measurement of thrombomodulin cofactor activity, lysates were incubated for 30 minutes at 37°C with 2.6 nmol/L human thrombin, 0.84 pmol/L human protein C, and various Concentrations of CaCI,. The reaction was stopped by addition of a mixture of antithrombin 111 (25 pgI mL) and heparin (25 U/mL), and the amidolytic activity of activated protein C was measured with the chromogenic substrate S-2366. All assays were performed in replicate (n = 3 to 6) and statistical comparisons were made using the two-tailed Student’s t-test.

For determination of apparent equilibrium dissociation constant the initial rate of protein C activation was measured in the

presence of fixed concentrations of protein C (0.84 pnol/L) and CaCl, (2.5 mmol/L), and increasing concentrations of thrombin (0 to 50 nmol/L). &(app) was defined as the concentration of thrombin that produced half-maximal thrombomodulin cofactor activity, as determined by nonlinear regression analysis (Sigmaplot, Jandel Sci- entific, San Rafael, CA).

Immunoblot analysis. Cells were washed three times with PBS and lysates were prepared in 20 mmoUL Tris-HC1, pH 8.0, 0 . 6 6 Triton X-100, 100 m m o l n NaCl, 3.0 mmol/L CaCI,, and 10 mmol/ L iodoacetamide. After centrifugation at 12,000g for 5 minutes, supernatant fractions containing 50 pg of total protein were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS- PAGE) under nonreducing conditions. Immunoblot analysis was performed as described previously‘ using mouse antihuman thrombomodulin monoclonal antibody TM1009 (Dako Corp, Carpinteria, CA) or mouse antihuman keratinocyte transglutaminase monoclonal antibody B.C1 (kindly provided by Dr Robert H. Rice, University of California, Davis, CA)24 with enhanced chemiluminescence detec- tion.

Nuclease SI protection analysis. Plasmids containing cDNA in- serts for human thrombomodulin (pUC 19TM 12)’’ and human y actin (pHFyA-l, provided by Dr L. Kedes, University of Southern California, Los Angeles, CA)” were linearized by digestion with MluI and BglII, respectively. The linearized plasmids were treated with calf intestinal alkaline phosphatase, and end-labeled with [ Y ~ ~ P ] A T P and T4 polynucleotide kinase. Total cellular RNA was isolated from cultured cells by acid guanidinium thiocyanate-phenol-

chloroform extraction (Tri-reagent; Molecular Research, Inc. Cincin- nati. OH). The end-labeled thrombomodulin and actin plasmids were hybridized overnight at 55°C with S0 p g or S p g of total cellular RNA. respectively, and analyzed by denaturing gel electrophoresi\ as described previously.”

Densitometry. Laser densitometry of nuclease S1 protection au- toradiographs and immunoblots was performed using a 300s corn- puting densitometer with ImageQuant v5.0 software (Molecular Dy- namics, Sunnyvale, CA).

RESULTS

Previous studies have shown that keratinocytes and endothelial cells contain thrombomodulin with similar apparent molecular mass and specific cofactor To determine whether keratinocyte thrombomodulin and endothelial cell thrombomodulin differ in affinity for thrombin, we measured the apparent dissociation constant in a protein C activation assay. In three separate experiments, Kd,+,p, for thrombin ranged from 5.3 2 0.3 to 8.4 t 0.5 nmol/L with keratinocyte thrombomodulin (Fig 1A) and from 3. I t 0.4 nmol/L to 5.6 2 1.3 nmol/L with HUVEC thrombomodulin (Fig IB). These values, which are similar to those reported for other preparations of human thrombomodulin,2x indicate that thrombin binds to keratinocyte thrombomodulin and endothelial cell thrombomodulin with similar affinity.

Effecn cf ATRA and TNF on tkromhomodulin cojc1ctor activiv. The effect of ATRA on the activity of keratinocyte thrombomodulin was measured in protein C activation assays with lysates prepared from keratinocytes incubated with various concentrations of ATRA for 24 hours (Fig 2A). ATRA produced a dose-dependent increase in thrombomodulin cofactor activity, with maximum activity after incubation with 20 pmol/L ATRA. Incubation with higher concentrations of ATRA caused cell toxicity. The time course of ATRA-induced increase in thrombomodulin activity is shown in Fig 2B. When keratinocytes were cultured for 96 hours in the presence of 10 ,umol/L ATRA, thrombomodulin cofactor activity peaked at 24 hours, and then declined grad- ually. As reported previously,” a more modest increase in thrombomodulin activity was observed when keratinocytes were incubated with 1.4 mmol/L calcium in the absence of ATRA.

In agreement with previous studies,”.” incubation of HU- VEC with 10 pmol/L ATRA for 24 hours produced a small but significant increase in thrombomodulin activity (Fig 3A). However, the increase in thrombomodulin activity induced by this concentration of ATRA was considerably greater in keratinocytes (5.4 L 0.9-fold) than in HUVEC (1.9 f 0. I - fold). In contrast to ATRA, TNF produced a marked decrease in thrombomodulin activity in HUVEC, but had little effect on thrombomodulin activity in keratinocytes. Incubation of HTJVEC with 1.0 nmoVL TNF for 24 hours decreased thrombomodulin activity by 83% 2 7% (Fig 3B). In several experiments, incubation of keratinocytes with TNF for up to 5 days resulted in no significant change in thrombomodulin expression compared with control keratinocytes incubated with PBS.

Effects of ATRA und TNF on thrombomodulin protein and mRNA. In endothelial cells, ATRA and TNF regulate thrombomodulin expression primarily at the level of tran-




DIFFERENTIAL REGULATION OF THROMBOMODULIN 2045

h

f E 9 Q W

n

2.0

1.5

1 .o

0.5

0.0 0 10 20 30 40

0.8 I I

I B h

f Q v

9 n

0.6

0.4

0.2

0.0 0 10 20 30 40

Thrombin (nM)

Fig 1. Determination of for keratinocyte and endothelial cell thrombomodulin. The thrombin dependence of protein C activation was measured in the presence of (A) keratinocyte or (61 HUVEC lysates. Values f SD for klSp, were 8.4 f 0.5 for keratinocytes and 3.5 f 0.5 nmol/L for HUVEC.

scription, producing equivalent changes in thrombomodulin protein and To determine effects of ATRA and TNF on thrombomodulin protein and mRNA in keratinocytes, immunoblot and nuclease S1 protection analyses were performed. Similar amounts of thrombomodulin protein (Fig 4) and mRNA (Fig 5) were detected in keratinocytes incubated for 24 hours with 1.0 nmol/L TNF or vehicle (PBS or DMSO). Keratinocytes incubated for 24 hours with 10 pmoV L ATRA contained markedly increased amounts of thrombomodulin protein (Fig 4) and thrombomodulin mRNA (Fig 5) compared with keratinocytes incubated with TNF, PBS, or DMSO. No treatment-related changes in the level of actin mRNA were observed (Fig 5). Mean ATRA-induced increases in thrombomodulin protein (5.5 ? 2.1-fold in five experiments) and thrombomodulin mRNA (4.2 +. 1.2-fold in two experiments) were similar in magnitude to increases in thrombomodulin activity observed in ATRA-treated keratinocytes (Fig 3A).

Although cultured keratinocytes have been reported to express functional TNF receptors,'x the failure of TNF to alter thrombomodulin expression raised the possibility that keratinocytes may lack responsiveness to TNF under these culture conditions. To test this possibility, immunoblot analysis was performed using a monoclonal antibody to keratinocyte transglutaminase, a marker of keratinocyte differentiation that is synthesized in the granular layer of ep ide~mis .~~ In agreement with previous we detected increased expression of transglutaminase after keratinocytes were cultured for 3 days in KGM containing 1.4 mmol/L Ca2' and either PBS or DMSO (Fig 6). The increase in keratinocyte transglutaminase was inhibited by addition of 10 pmol/L ATRA, and enhanced by addition of 1.0 nmol/L TNF, con- firming that keratinocytes respond to TNF under these cul-

80 A

0 10-7 104 10-5 10-4

ATRA (M)

100

80

60

40

20

0 0 20 40 60 80 100

Time (h) Fig 2. Effect of ATRA on thrombomodulin activity in keratino-

cytes. (A) Keratinocytes were incubated for 24 hours in KGM containing 1.4 mmol/L Ca" and the indicated concentrations of ATRA. (B) Keratinocytes were incubated for the indicated times in KGM containing 1.4 mmol/L C a z + and either 10 pmollL ATRA (0) or 0.1% DMSO (m). Thrombomodulin cofactor activity was measured in a two-stage protein C activation assay. Values represent the mean f SE of three determinations.




2046 RAIFE, DEMETROULIS, AND LENTZ

ture conditions. In contrast, thrombomodulin expression was enhanced by ATRA and unaffected by TNF (Fig 6).

In this experiment (Fig 6), peak induction of thrombomodulin protein was observed 3 days after addition of ATRA (lane 8). In other experiments, peak induction of thrombomodulin was observed at various times between 1 and 3 days after addition of ATRA. More rapid induction of thrombomodulin expression was observed with keratinocyte monolayers that were more than -70% confluent, and slower induction was observed with less confluent cultures.

Calcium-dependence of thrombomodulin activity. To determine the calcium-dependence of keratinocyte thrombomodulin activity, protein C activation assays were performed with keratinocyte lysates in the presence of up to 10.0 mmol/L CaCI2. The activity profile of keratinocyte thrombomodulin was compared with that of rabbit lung thrombomodulin and TM456, a soluble recombinant

40

20

0

A * T

HK HUVEC

6 , B

I

HK

r

HUVEC Fig 3. Differential effects of ATRA and TNF on thrombomodulin

activity in keratinocytes and endothelial cells. (A) Keratinocytes (HK) or HUMC were incubated for 24 hours with either 0.1% DMSO (0) or 10pmollL ATRA (m). (E) HK or HUVEC were incubated for 24 hours with either 0.3% PES (0) or 1.0 nmollL TNF (m). Thrombomodulin cofactor activity was measured in a two-stage protein C activation assay. Values represent the mean f SE of six (A) or three (B) determinations ( *P < .Ol).

n

200 -

97-

46 -

30- 1 2 3 4

Fig 4. Effects of ATRA and TNF on thrombomodulin protein in keratinocytes. Keratinocytes were incubated for 24 hours in KGM containing 1.4 mmollL Ca" and either 0.3% PBS (lane l), 0.1% DMSO (lane 2). 1.0 nmollL TNF (lane 3). or 1OpmollL ATRA (lane 41. Cell lysates containing 50 g g of total protein were subjected t o SDS- PAGE under nonreducing condkions, and immunoblots were performed using an antithrombomodulin monoclonal antibody. The Mr of protein standards is indicated at left. The position of thrombomodulin ( T M I is indicated at right.

thrombomodulin fragment that comprises the fourth through sixth EGF-like domains of human thrombomodulin.2' Keratinocyte thrombomodulin demonstrated a biphasic activity profile, with activity increasing sharply as the CaC& concentration was increased from 0 to 1.0 mmol/L, and activity decreasing at higher CaC12 concentrations (Fig 7A). TM456 also produced a biphasic activity profile, with peak activity at a slightly lower CaC12 concentration (0.5 mmol/L) than keratinocyte thrombomodulin (Fig 7B). Rabbit thrombomodulin exhibited a simple hyperbolic activity profile, with maximal activity at saturating concentrations of calcium ions (Fig 7C).

To determine whether stimulation with ATRA altered the calcium-dependence or thrombin affinity of keratinocyte thrombomodulin, keratinocytes were incubated with or with- out I O pnol/L ATRA for 24 hours, and thrombomodulin activity was measured in the presence of either 1.0 or 4.0 mmoln CaCI2. The ratio of thrombomodulin activities at 1.0 mmol/L and 4.0 mmol/L CaC12 was 2.8 for ATRA- treated keratinocytes, compared with 3.8 for untreated keratinocytes, 7.1 for TM456, and 0.4 for rabbit thrombomodulin. The &(app) for thrombin of thrombomodulin in ATRA-treated keratinocytes was 5.2 ? 1.8 nmoVL, which is similar to that of thrombomodulin in untreated keratinocytes (Fig 1A).




DIFFERENTIAL REGULATION OF THROMBOMODULIN

DISCUSSION

Thrombomodulin was identified 15 years ago as an endothelial cell cofactor for protein C activation." Recent evi- dence from several laboratories has established that thrombomodulin is expressed not only by vascular endothelial cells, but also by nonvascular cells such as keratin~cytes."~ AI- though the function of thrombomodulin in nonvascular loca- tions is poorly understood, its pattern of expression during embryogenesis suggests a possible role in cellular development and differentiation.'." This possibility is supported by the observation that homozygous deficiency of thrombomodulin produces an embryonic lethal phenotype in mice."

Previous studies have shown that thrombomodulin expression correlates strongly with epidermal differentiation in vivo, and with calcium-induced keratinocyte differentiation in vitro.4." This study demonstrates that keratinocyte thrombomodulin and endothelial cell thrombomodulin have similar cofactor activity and thrombin affinity, but differ in their responses to regulatory stimuli. Unlike thrombomodulin expression in endothelial cells, thrombomodulin expression in

0 v) z n

a c l- <

L L z l-

+" TM

- actin

l 2 3 4 Fig 5. Effects of ATRA and TNF on thrombomodulin mRNA in

keratinocytes. Keratinocyes were incubated for 24 hours in KGM containing 1.4 mmol/L Ca2' and either 0.1% DMSO (lane l), 10 pmol lL ATRA (lane 2). 1.0 nmol/L TNF (lane 3). or 0.3% PES (lane 4). Total cellular RNA was hybridized with probesfor human thrombomodulin (TM) or actin, and analyzed by nuclease S1 protection. The actin doublet in lane 2 was not observed in several other experiments.

v)

L m

0 v) t 0

L Z

2047

nnnn Days: 1 3 1 3 1 3 1 3

200 - 97 - 69 -

. - - "" -.. .- _,.."__ ̂b TG

200 - 97 - 69 -

.-<a-.

46 - 1 2 3 4 5 6 7 8

Fig 6. Effects of ATRA and TNF on transglutaminase and thrombomodulin in keratinocytes. Keratinocytes were incubated for 1 or 3 days in KGM containing 1.4 mmol/L Ca2' and either 0.3% PES (lanes 1 t o 21, 0.1% DMSO (lanes 3-41. 1.0 nmol/L TNF (lanes 5 t o 6). or 10 pmol/L ATRA (lanes 7 to 8). Cell lysates containing 50 pg of total protein were subjected to SDS-PAGE under nonreducing conditions, and immunoblots were performed using either antitransglutaminase (upper gel) or antithrombomodulin (lower gel) antibodies. The Mr of protein standards is indicated at left. The positions of transglutaminase (TGl and thrombomodulin (TM) are indicated at right. The additional bands of high apparent molecular mass detected with antithrombomodulin antibody (lane 81 were not observed in several other experiments.

keratinocytes was stimulated markedly by ATRA, but was unaffected by TNF.

Retinoids regulate epidermal differentiation in a complex manner. Based on in vitro studies, as well as observations in animals with vitamin A deficiency (which produces epithelial changes such as hyperkeratosis and squamous meta- plasia of transitional and columnar epithelium), it has been suggested that retinoic acid may inhibit epidermal differentiation.2'J However, recent data from mice with targeted over- expression of dominant-negative retinoic acid receptor genes indicate that retinoids may be necessary for normal epidermal The observation that ATRA is a potent stimulator of thrombomodulin expression in keratinocytes suggests that retinoids may play a physiological role in the regulation of thrombomodulin expression during squamous differentiation.

Dittman et aIz7 have identified functional retinoic acid response elements within the 5"flanking region of the human thrombomodulin gene. These investigators suggested that the thrombomodulin promoter may be more responsive to retinoids in cells with low basal levels of thrombomodulin expression than in cells with high basal levels of thrombomodulin expression. Our observations in keratinocytes are consistent with transcriptional regulation of the thrombomodulin gene by ATRA, in that increased thrombomodulin




2048

0.2

0.1

0.0

0.4

0.2

0.0

1 .o

0.5

I A

0 2 4 6 8 1 0

0 2 4 6 8 1 0

C

0.0 +l 0 2 4 6 8 1 0

CaC12 (mM)

Fig 7. [Ca2’1 dependence of thrombomodulin activity. Protein C activation was measured in the presence of the indicated concentrations of CaCI,. (A) Keratinocyte thrombomodulin, (B) recombinant human TM456, (C) rabbit lung thrombomodulin. Values represent the mean of duplicate determinations.

activity in ATRA-treated keratinocytes was associated with increased levels of thrombomodulin mRNA. However, de- spite similar levels of basal thrombomodulin expression, we observed a greater response to ATRA in keratinocytes than in HUVEC. This suggests that responsiveness of the thrombomodulin promoter to retinoic acid is regulated indepen- dently of the basal transcription rate.

The failure of TNF to alter expression of keratinocyte thrombomodulin while markedly stimulating expression of keratinocyte transglutaminase also suggests differential responsiveness of the thrombomodulin promoter. Recent studies indicate that an Ets-like element within the proximal thrombomodulin promoter may be involved in TNF-mediated transcriptional inhibition in endothelial ~ e l l s . ~ ’ . ~ ~ This sequence appears to recognize a novel nuclear factor that is present in endothelial cells.” Whether this factor is responsi- ble for TNF-mediated inhibition of thrombomodulin tran-

RAIFE, DEMETROULIS, AND LENT2

scription has not been determined, and it is not known i f this factor is present in keratinocytes.

Protein C contains a high-affinity CaZ+ site that must be occupied for efficient activation by the thrombidthrombo- modulin complex.’ The effect of Ca2+ on protein C activation is also influenced by the glycosylation state of thrombomodulin. Isoforms of thrombomodulin that lack chondroitin sulfate (CS) exhibit peak cofactor activity at [Ca”] below 1 .O mmol/L, whereas isoforms containing CS are most active at saturating [Ca”].’ Studies with cultured endothelial cells indicate that a variable fraction of human thrombomodulin js devoid of CS,22 implying paradoxically that protein C activation may be suboptimal at plasma [Ca”]. With both untreated and ATRA-treated keratinocytes, we observed maximal thrombomodulin activity at 1.0 mmol/L Ca”, which suggests that keratinocyte thrombomodulin does not contain large amounts of CS. However, the possibility that CS may have been lost or modified during preparation of cell lysates cannot be excluded.

Unlike endothelial cell thrombomodulin, keratinocyte thrombomodulin may be exposed to extracellular [Ca”] lower than that found in plasma. In vivo studies in mice suggest that an extracellular [Ca”] gradient exists within the stratified squamous epithelium of the epidermis, with lower [Ca”] in the basal and spinous layers, and higher [Ca”] in the granular layer.*’,*’ Thrombomodulin is selec- tively expressed in the spinous layer: where relatively low extracellular [Ca”] may support optimal cofactor activity.

In summary, we have shown that the thrombin affinity and basal level of expression of thrombomodulin are similar in cultured keratinocytes and endothelial cells, but that regulation of thrombomodulin expression differs in these two cell types. Keratinocyte thrombomodulin is upregulated strongly by ATRA, but is unaffected by TNF. In contrast, endothelia1 cell thrombomodulin is upregulated weakly by ATRA, and downregulated strongly by TNF. Therefore, clin- ical settings associated with the production or administration of retinoids or inflammatory cytokines may produce differential effects on thrombomodulin expression in blood vessels and epidermis. These data provide additional support for the hypothesis that thrombomodulin has distinct biological functions in vascular and nonvascular cells.

ACKNOWLEDGMENT

The authors thank Yan Chen and Anji Newell for technical assis- tance, and Drs Kathi Madison and Warren Piette for helpful com- ments.

REFERENCES 1 . Esmon CT: Thrombomodulin as a model of molecular rnecha-

nisms that modulate protease specificity and function at the vessel surface. FASEB J 9:946, 1995

2. Dittman WA, Majerus PW: Structure and function of thrombomodulin: A natural anticoagulant. Blood 75:329, 1990

3. Boffa M, Burke B, Haudenschild CC: Preservation of thrombomodulin antigen on vascular and extravascular surfaces. J Histochem Cytochem 35: 1267, 1987

4. Raife TJ, Lager DJ, Madison KC, Piette WW, Howard EJ, Sturm M”, Chen Y, Lmtz SR: Thrombomodulin expression by hu-




DIFFERENTIAL REGULATION OF THROMBOMODULIN 2049

man keratinocytes, Induction of cofactor activity during epidermal differentiation. J Clin Invest 93:1846, 1994 5. Mizutani H, Hayashi T, Nouchi N, Ohyanagi S, Hashimoto K,

Shimuzu M, Suzuki K: Functional and immunoreactive thrombomodulin expressed by keratinocytes. J Invest Dermatol 103:825, 1994

6. Lager DJ, Callaghan ET, Worth SF, Raife TJ, Lentz SR: Cellu- lar localization of thrombomodulin in human epithelium and squamous malignancies. Am J Pathol 146:933, 1995

7. Jackson DE, Mitchell CA, Bird P, Salem HH, Hayman JA: Immunohistochemical localization of thrombomodulin in normal human skin and skin tumours. J Path01 175:421, 1995

8. Imada S, Yamaguchi H, Nagumo M, Katayanagi S, Iwasaki H, Imada M: Identification of fetomodulin, a surface marker protein of fetal development, as thrombomodulin by gene cloning and functional analysis. Dev Biol 140:113, 1990

9. Ford VA, Wilkinson JE, Kennel SJ: Thrombomodulin distribu- tion during murine development. Roux Arch Devel Biol 202:364, I993

IO. Healy AM, Rayburn HB, Rosenberg RD, Weiler H: Absence of the blood-clotting regulator thrombomodulin causes embryonic lethality in mice before development of a functional cardiovascular system. Proc Natl Acad Sci USA 92850, 1995

1 1. Nawroth PP, Handley DA, Esmon CT, Stem DM: Interleukin 1 induces endothelial cell procoagulant while suppressing cell-surface anticoagulant activity. Proc Natl Acad Sci USA 83:3460, 1986

12. Nawroth PP, Stem DM: Modulation of endothelial cell hemo- static properties by tumor necrosis factor. J Exp Med 163:740, 1986

13. Conway EM, Rosenberg RD: Tumor necrosis factor sup- presses transcription of the thrombomodulin gene in endothelial cells. Mol Cell Biol 8:5588, 1988

14. Moore KL, Esmon CT, Esmon NL: Tumor necrosis factor leads to the internalization and degradation of thrombomodulin from the surface of bovine aortic endothelial cells in culture. Blood 73:159, 1989

15. Lentz SR. Tsiang M, Sadler .E: Regulation of thrombomodulin by tumor necrosis factor-alpha: Comparison of transcriptional and posttranscriptional mechanisms. Blood 77542, 1991

16. Ishii H, Horie S, Kizaki K, Kazami M: Retinoic acid counter- acts both the downregulation of thrombomodulin and the induction of tissue factor in cultured human endothelial cells exposed to tumor necrosis factor. Blood 80:2556, 1992

17. Miyake S, Ohdama S, Tazawa R, Aoki N: Retinoic acid pre- vents cytokine-induced suppression of thrombomodulin expression on surface of human umbilical vascular endothelial cells in vitro. Thromb Res 68:483, 1992

18. Pillai S, Bilke DD, Eessalu TE, Aggarwal BB, Elias PM: Binding and biological effects of tumor necrosis factor alpha on cultured human neonatal foreskin keratinocytes. J Clin Invest 83:816, 1989

19. Fuchs E, Green H: Regulation of terminal differentiation of cultured human keratinocytes by vitamin A. Cell 25:617, 1981

20. Menon GK, Grayson S, Elias PM: Ionic calcium reservoirs in mammalian epidermis: Ultrastructural localization by ion-capture cytochemistry. J Invest Dermatol 84508, 1985

21. Filvaroff E, Calautti E, Reiss M, Dotto GP: Functional evi- dence for an extracellular calcium receptor mechanism triggering tyrosine kinase activation associated with mouse keratinocyte differentiation. J Biol Chem 269:21735, 1994

22. Lin J, McLean K, Morser J, Young TA, Wydro RM, Andrews WH, Light DR: Modulation of glycosaminoglycan addition in natu- rally expressed and recombinant human thrombomodulin. J Biol Chem 269:25021, 1994

23. Lentz SR, Chen Y, Sadler JE: Sequences required for thrombomodulin cofactor activity within the fourth epidermal growth factor-like domain of human thrombomodulin. J Biol Chem 268: 15312, 1993

24. Thacher SM, Rice RH: Keratinocyte-specific transglutaminase of cultured human epidermal cells: Relation to cross-linked envelope formation and terminal differentiation. Cell 40:685, 1985

25. Gunning P, Ponte P, Okayama H, Engle J, Blau H, Kedes L Isolation and characterization of full-length cDNA clones for hum- man alpha, beta, and gamma-actin mRNAs: Skeletal but not cyto- plasmic actins have an amino-terminal cysteine that is subsequently removed. Mol Cell Biol 3:787, 1983

26. Lentz SR, Sadler JE: The molecular basis of thrombomodulin function, in Giddings JC (ed): Thrombin, Thrombomodulin, and the Control of Hemostasis, Austin, TX, RR Landes, 1994, p 91

27. Dittman WA, Nelson SC, Greer PK, Horton ET, Palomba ML, McCachren SS: Characterization of thrombomodulin expression in response to retinoic acid and identification of a retinoic acid response element in the human thrombomodulin gene. J Biol Chem 269:16925, 1994

28. Esmon CT, Owen WG: Identification of an endothelial cell cofactor for thrombin-catalyzed activation of protein C. Roc Natl Acad Sci USA 78:2249, 1981

29. Darmon M, Blumenberg M: Retinoic acid in epithelial and epidermal differentiation, in Darmon M, Blumenberg M (eds): Mo- lecular Biology of the Skin: The Keratinocyte. San Diego, CA, Aca- demic, 1993, p 181

30. Imakado S, Bickenbach JR, Bundman DS, Rothnagel JA, Attar PS, Wang X, Walczak VR, Wisniewski S, Pote J, Gordon JS, Heyman RA, Evans RM, Roop DR: Targeting expression of a dominant-negative retinoic acid receptor mutant in the epidermis of transgenic mice results in loss of barrier function. Genes Dev 9:317, 1995

31. Saitou M, Sugai S, Tanaka T, Shimouchi K, Fuchs E, Kaki- zuka A: Inhibition of skin development by targeted expression of a dominant-negative retinoic acid receptor. Nature 374: 159, 1995

32. Andersen B, Rosenfeld MG: New wrinkles in retinoids. Na- ture 374:118, 1995

33. Yu K, Morioka H, Fritze LMS, Beeler DL, Jackman RW, Rosenberg RD: Transcriptional regulation of the thrombomodulin gene. J Biol Chem 267:23237, 1992

34. von der Ahe D, Nischan C, Kunz C, Otte J, Knies U, Oder- wald H, Wasylyk B: Ets transcription factor binding site is required for positive and TNF alpha-induced negative promoter regulation. Nucleic Acids Res 215636, 1993

35. Otte J, Heischmann A, von der Ahe D: Regulation of the thrombomodulin gene promoter and cloning of the potential regulatory gene. Thromb Haemost 73: 1263, 1995




1996 88: 2043-2049

TJ Raife, EM Demetroulis and SR Lentz keratinocytes and endothelial cellsand tumor necrosis factor-alpha: differential responses in Regulation of thrombomodulin expression by all-trans retinoic acid

http://www.bloodjournal.org/content/88/6/2043.full.htmlUpdated information and services can be found at:

Articles on similar topics can be found in the following Blood collections

http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requestsInformation about reproducing this article in parts or in its entirety may be found online at:

http://www.bloodjournal.org/site/misc/rights.xhtml#reprintsInformation about ordering reprints may be found online at:

http://www.bloodjournal.org/site/subscriptions/index.xhtmlInformation about subscriptions and ASH membership may be found online at:

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


http://www.bloodjournal.org/content/88/6/2043.full.html

http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests

http://www.bloodjournal.org/site/misc/rights.xhtml#reprints

http://www.bloodjournal.org/site/subscriptions/index.xhtml





Documents

Regulation of Thrombomodulin Expression by All-trans Retinoic Acid