Brain Research, 556 (1991) 1-5 Elsevier Science Publishers B.V. ADONIS 0006899391168195 BRES 16819
Research Reports
Regional distribution of thrombomodulin in human brain Vicky L.Y. Wong 1'2, Florence M. Hofman 2, Hidemi Ishii 3 and Mark Fisher 1 Departments of lNeurology and ZPathology, University of Southern California School of Medicine, Los Angeles, CA 90033 (U.S.A.) and JDepartment of Clinical Biochemistry, Teikyo University, Sagamiko, Kanagawa (Japan) (Accepted 26 February 1991)
Key words: Thrombomodulin; Brain; Endothelium; Infarct; Coagulation; Stroke; Thrombosis
Thrombomodulin, an integral membrane protein with important anticoagulant properties, was evaluated for its distribution in blood vessels ,of human brain. Adjacent sections from frozen human brain were stained for von Willebrand Factor (an endothelium marker) or thrombomodulin. The number of thrombomodulin-positive blood vessels was expressed as a percentage of all blood vessels in a specific area. The density ratio was compared among cortical and subcortical regions in the human brain. Our results indicated that thrombomodulin density in neocortex, cerebellum, medulla and hippocampus was similar. There was, however, significantly less thrombomodulin in putamen (P < 0.01), pous (P < 0.001) and mesencephalon (P < 0.001) compared to neocortex. The regional distribution of thrombomodulin may prove to be helpful for understanding the development of thrombotic diseases of the brain.
INTRODUCTION
MATERIALS AND METHODS
T h r o m b o m o d u l i n is an integral m e m b r a n e protein with a molecular weight of 75,000 D a s,19,27. T h r o m b o -
Tissue samples
modulin has b e e n localized to arteries, veins, capillaries and lymphatics o f nearly every h u m a n tissue and is n o r m a l l y p r e s e n t on endothelial cells 21. The role of t h r o m b o m o d u l i n in n o r m a l anticoagulation is critical to understanding hemostasis. T h r o m b o m o d u l i n binds t h r o m b i n and changes the substrate specificity of thromb i n - t o w a r d protein C. The t h r o m b i n - t h r o m b o m o d u l i n complex activates p r o t e i n C 20,000 times m o r e rapidly than t h r o m b i n alone 7. A c t i v a t e d protein C in turn inactivates clotting factors Va and V I I I a 1~,2a and plasminogen activator inhibitor type 129, thus shifting the hemostatic balance to anticoagulation. A l t h o u g h the distribution of t h r o m b o m o d u l i n is wides p r e a d t h r o u g h o u t the body, t h r o m b o m o d u l i n has been r e p o r t e d to be low o r absent from e n d o t h e l i u m in h u m a n brain ~7'~s'2~, while present in rabbit brain 4,5. The present study addresses the question w h e t h e r there is any specific p a t t e r n of t h r o m b o m o d u l i n distribution in brain. We r e p o r t immunohistochemical evidence that t h r o m b o m o dulin is p r e s e n t in h u m a n brain and has a p a t t e r n of high and low distribution in 12 distinct brain regions. O u r results indicate that ports, mesencephalon and p u t a m e n exhibit a relative deficiency of t h r o m b o m o d u l i n as c o m p a r e d to neocortex.
Tissue from 10 normal human brains were used in this study; the specimens were obtained at autopsy with postmortem intervals ranging 4-14 h, age range 35-79 years; there were 8 males and 2 females (Table IA). Tissues were obtained from the Alzheimer Disease Research Center of Southern California, University of Southern California, Los Angeles, CA. Neuropathologic examination revealed no significant pathologic abnormalities, including no evidence of small vessel disease. Brain specimens were cut, snap frozen in liquid nitrogen and stored at -70 °C or in fiquid nitrogen. Cryostat sections were serially cut at 7/~m, left to air dry overnight, and then fixed with acetone for 5 min.
Antisera Polyclonal. Rabbit anti-human thrombomodulin IgG was prepared as previously de~2ribed16, using 120/~g of isolated placental thrombomodulin in Bacto complete Freud's adjuvant as the immunogen. The specificity of the purified antibody has been confirmed by Western blot analysis and inactivation of thrombomodulin activity1~. In addition, the activity of 12 ng purified human thrombomodulin was inhibited 50% by 150 ng rabbit anti-human thrombomedulin IgG, whereas the activity of 36 ng purified rabbit thrombomodulin was not inhibited even with 3/~g of rabbit anti-human thrombomodulin IgG. Monoclonal. Mouse anti-human thrombomodulin IgG was prepared as previously described16, using 20/~g of isolated placental thrombomodulin in complete Freud's adjuvant as the immunogen. Specificity of the mouse monoclonal antibody was confirmed by Western blot analysisTM. The monoclonal antibody was used to stain selected brain sections using the alkaline phosphatase staining method in order to confirm the polyclonal reagent staining specificity.
lmmunoperoxidase staining reagents and procedure Rabbit anti-yon Willebrand Factor (vWF) (Dako, Santa Barbara,
Correspondence: M. Fisher, University of Southern California, Department of Neurology, 2025 Zonal Street, Los Angeles, CA 90033, U.S.A.
TABLE I Thrombomodulin distribution Clinical variables in human patients used for thrombomodulin study (A) and regional distribution of thrombomodulin in human brain (B). P.M. time: postmortem time in hours. (,4) Case characteristics Clinical variables
Case number
Age Sex P.M. time
1
2
3
4
5
6
7
8
9
10
71 F 4
47 M 8.5
64 M 10
38 M 9
72 F 6.5
79 M 10
55 M 9.5
38 M 8.5
54 M 10
35 M 14
(B) Thrombomodulin-vWF ratios Brain regions
Case number 1
Cortical areas
Hippocampus Putamen Cerebellum Mesencephalon Pons Medulla
4 6 9 10 21 41
0.93
2
3
4
1.10
1.00
0.65 1.08 0.82 1.01
0.72 0.55
Mean + S.D.
0,98 0.77 0.94 0.84 0.80 0.85 0.56 0.59 0.43 0.99
5
6
7
8
9
10
0.89 1.03
0.92 1,10
0,88
0.95 +_0.12
0.90 0.51 0.80
1.03 0.71 1.00
0.85 0.80 0.80
1.02 0.94 0.94 0,74
0.37 0.54
0.76 0.88
0.90
0.95
0.90 + 0.75 + 0.85 + 0.69 + 0.53 + 0.85 +
1.00
1.00
1.02
0.80
0.66
0.57 0.55
0.86
0.10 0.14' 0.17 0.12" 0.15" 0.18
*P < 0.01, *P < 0.001.
CA) was used at a dilution of 1:1000; rabbit anti-human prostatic epithelial antigen, an irrelevant antibody control, was used at 1:300. Rabbit anti-human thrombomodulin antibody was used at a dilution of 1:500. Biotinylated goat anti-rabbit immunoglobulin (Vector, Burlingame, CA) and avidin-biotin peroxidase complex (ABC, Vector) were used at manufacturer's prescribed concentrations. Amino-ethyl carbazole (AEC) was used as the substrate. Immunoperoxidase staining was performed as previously described 13. Acetone-fixed tissue sections, cut within 48 h, were rehydrated with phosphate-buffered saline (PBS, pH 7.4) for 10 rain, incubated with 0.3% HzO z, and washed with PBS. Nonspecific staining was then blocked with 5% solution of normal goat serum (Cedarlan Lab, Hornby, Ontario, Canada) for 15 min at 25 °C. Slides were subsequently blotted dry, and incubated with primary antibody for 30 rain at 25 °C in a humidified chamber. After each incubation, sections were washed in PBS and incubated with 1% bovine serum albumin (Fisher, Richmond, CA) for 5 rain at 25 °C. After the primary antibody, slides were incubated with biotin-labeled secondary antibody for 30 min, and then with ABC for 15 min. The sections were finally incubated with AEC for 10 min, counterstained in Mayer's hematoxylin for 3 min, and mounted in glycerol and PBS. Alkaline phosphatase staining reagents and procedure Mouse anti-human thrombomodulin antibody was used at a dilution of 1:1000. Alkaline phosphatase anti-alkaline phosphatase was purchased from Dako (Santa Barbara, CA); alkaline phosphatase substrate kit and levamisole were products of Vector (Burlingame, CA). Alkaline phosphatase staining was performed according to the instruction of the manufacturer. Acetone-fixed tissue sections, cut within 48 h, were rehydrated with Tris-buffered saline (TBS, 0.05 M Tris, 0.15 M NaC1, pH 7.6) for 10 min. Slides were then blotted dry and incubated with monocional anti-thrombomodulin for 45 min at
25 °C. After primary antibody incubation, slides were incubated with rabbit anti-mouse immunoglobulin, and then with alkaline phosphatase mouse anti-alkaline phosphatase immune complex for 30 rain at 25 °C. The sections were incubated with alkaline phosphatase substrate with levamisol as inhibitor of endogenase alkaline phosphatase for 30 min in the dark. Sections were then counterstained and mounted in the manner described above. Quantitation Adjacent sections were stained and examined under the light microscope at high power (400x magnification). A grid was placed over the tissue section and the number of stained blood vessels within a 1 mm 2 area was counted. The number of blood vessels per mm 2 ranged from 20 to 300. Three corresponding areas in vWF-stained sections and in thrombomodulin-stained sections were counted and compared. The density of thrombomodulin was expressed as a ratio and calculated as follows: The total number of thrombomodulin-stained blood vessels/mm2 in 3 one-mm 2 areas The total number of vWF-stained blood vessels/mm 2 in 3 one-mm 2 areas To determine variability from section to section, adjacent sections from 7 brain regions were stained for the same antibody (vWF or thrombomodulin); the density ratio of adjacent sections with the identical antibody had a mean -+ S.D. value of 1.06 + 0.02. These controls indicated that the number of blood vessels within adjacent sections do not vary substantially. The thrombomodulin density ratio of different brain regions was compared and the statistical differences between regions were analyzed using unpaired Student's t-tests.
RESULTS Human brain specimens, stained with anti-vWF, demonstrated intense labeling of endothelial cells in both large and small blood vessels in different parts of the brain. Figs. 1A and 2A show staining of vessels (5-10/zm in width) in cortical area 21 and putamen, respectively. Adjacent sections stained with anti-thrombomodulin are shown in Fig. 1B (Table IB, Case 3) and 2B (Table IB, Case 6), respectively. Staining with anti-thrombomodulin was limited to endothelial cells (Fig. 3). Thrombomodulin monoclonal antibody also stained endothelial cells in multiple brain regions (Fig. 4). Cortical and subcortical regions were examined with both anti-thrombomodulin and anti-vWF. Thrombomodulin density ratios among the 17 neocortical regions were similar and these ratios were thus presented as composite. The mean neocortical value 0.95 + 0.12 (Table 1B), was compared to the ratio of each of the other regions. The hippocampus, cerebellum and medulla exhibited ratios of 0.90 + 0.10, 0.85 + 0.17 and 0.85 _+ 0.18, respectively (Table IB). In contrast, the pons, mesencephalon and putamen showed ratios of 0.53 + 0.17, 0.69 + 0.12 and 0.75 + 0.14, respectively (Table
Fig. 2. Immunostaining of putamen: adjacent sections stained with
anti-vWF (A, 1:1000) and polyclonal anti-thrombomodulin (B) antibody (1:500); counterstained with hematoxylin. Bar = 100/~m.
IB), which were significantly less than that of the neocortex. Similar thrombomodulin densities could be demonstrated in adjacent gray and white matter of tissue samples. For example, the thrombomodulin density ratio for gray and white matter in pons was 0.55 and 0.54, respectively, from case 1, 0.43 and 0.50, respectively, from case 4, and 0.76 and 0.60, respectively, from case 8. Thrombomodulin levels in different brain regions varied independent of age, gender or postmortem interval (Table IA).
Fig. 1. Immunostaining of area 21: adjacent sections stained with
anti-vWF (A, 1:1000) and polyclonal anti-thrombomodulin (B) antibody (1:500); counterstained with hematoxylin. Bar = 100/~m.
Fig. 3. Immunostaining of blood vessel from area 6 with polyclonal anti-thrombomodulin antibody (1:500). Bar = 10/~m.
4
Fig. 4. Monoclonal anti-thrombomodulin antibody (1:1000) stained blood vessels in pons; counterstained with hematoxylin. Bar = 100 /~m.
To further investigate the differing thrombomodulin densities, we evaluated the relationship between thrombomodulin density ratio and mean vWF-stained blood vessel number. Thrombomodulin density ratio was inversely associated with the number of vWF-stained blood vessels from all brain regions from the 10 cases, with simple correlation coefficient r -- -0.31 (P < 0.03). This suggests that increasing vessel density in brain may be associated with a reduced proportion of vessels staining with thrombomodulin. However, in putamen, mesencephalon and pons, the mean vWF-stained blood vessel numbers (160.6 + 81.0, 164.7 _+ 64.8, 180.2 + 69.5, respectively) were not significantly different from that of cortex (151.2 _+ 72.1) indicating that relatively low thrombomodulin ratios in these 3 subcortical areas were not secondary to vessel density. DISCUSSION In this study, we have demonstrated thrombomodulin staining of endothelium in blood vessels in various regions of the brain. The distribution of thrombomodulin varied among regions. Hippocampus, cerebellum, medulla along with neocortical regions exhibited comparable thrombomodulin staining; significantly less thrombomodulin-positive vessels were found in pons, putamen REFERENCES 1 Omitted. 2 Omitted. 3 Conway, E.M. and Rosenberg, R.D., Tumor necrosis factor suppresses transcription of the thrombomodulin gene in endothelial cells, MoL Cell. BioL, 8 (1988) 5588-5592. 4 DeBault, L.E., Esmon, N.L., Olson, J.R. and Esmon, C.T., Distribution of the thrombomodulin antigen in the rabbit vasculature, Lab. Invest., 54 (1986) 172-178. 5 DeBault, L.E., Esmon, N.L., Smith, G.P. and Esmon, C.T., Localization of thrombomodulin antigen in rabbit endothelial cells in culture, Lab. Invest., 54 (1986) 179-187.
and mesencephalon, compared to neocortex. The expression of thrombomodulin in the brain may be regulated. In endothelial cell culture, binding of thrombin to thrombomodulin causes internalization of the complex, degradation of thrombin and subsequent recycling of thrombomodulin to the surface 22. Additionally, thrombomodulin can be down-regulated by endotoxin, tumor necrosis factor (TNF) and interleukin123'25"26. For example, T N F has been reported to regulate thrombomodulin by suppressing its transcription 3, along with producing degradation and internalization of thrombomodulin 24. Whether these cytokines play important roles in regulating the distribution of thrombomodulin in human brain awaits further investigation. Recent experimental studies further emphasize the anti-thrombotic role of thrombomodulin 12'2°. Preinjection of anti-thrombomodulin in mice followed by thrombin injection helped potentiate the lethal effect of thrombin while preinjection of thrombomodulin prolonged the survival period 2°. Gomi et al. further localized fibrin deposits in large and small blood vessels of lungs in mice receiving thrombin but not in large blood vessels of animals preinjected with recombinant thrombomodulin 12. These results help demonstrate the role of thrombomodulin in preventing thrombosis. The significance of our results showing an irregular thrombomodulin distribution in the brain may be related to the topography of ischemic cerebral vascular disease. Lacunar infarcts (microinfarcts) are mostly absent from cerebral cortex and cerebellum but are more commonly found in the putamen and pons 9'1°. The pons is also one of the most c o m m o n subcortical regions for infarcts of all sizes TM. These results raise the possibility of an association between relative thrombomodulin deficiency and ischemic cerebral vascular disease. Acknowledgements. This study was supported by Grants NS 20989 (to M.E) and EY 08144 (to EM.H.) from the National Institutes of Health. The authors gratefully acknowledge the Alzheimer Disease Research Center of Southern California funded by Grant P50AG05142 from the National Institute of Aging. We are also grateful to Dr. David Hinton for helpful discussions. Mrs. Florence Miyagawa and Mrs. Lisa Doumak provided skillful secretarial assistance.
6 Esmon, C.T., The role of protein C and thrombomodulin in the regulation of blood coagulation, J. Biol. Chem., 264 (1989) 4743-4746. 7 Esmon, C.T. and Owen, W.G., Identification of an endothelial cell cofactor for the thrombin-catalyzed activation of protein C, Proc. Natl. Acad. Sci. U.S.A., 78 (1981) 2249-2252. 8 Esmon, N.L., Owen, W.G. and Esmon, C.T., Isolation of a membrane-bound cofactor for thrombin-catalyzed activation of protein C, J. Biol. Chem., 257 (1982) 859-864. 9 Ferrand, J., Essai sur l'h~mipldgie des vieillards. Les lacunes de desintegration cgrgbrale, Paris, Thesis, 1902. 10 Fisher, C.M., Lacunes: small, deep cerebral infarcts, Neurology, 15 (1965) 774-784.
11 Fulcher, C.A., Gardiner, J.E., Griffin, J.H. and Zimmerman, T.S., Proteolytic inactivation of human factor VIII procoagulant protein by activated protein C and its analogy with factor V, Blood, 63 (1984) 486-489. 12 Gomi, K., Zushi, M., Honda, G., Kawahara, S., Matsuzaki, O., Kanabayashi, T., Yamamoto, S., Maruyama, I. and Suzuki, K., Antithrombotic effect of recombinant human thrombomodulin on thrombin-induced thromboembolism in mice, Blood, 75 (1990) 1396-1399. 13 Hofman, EM., Billing, R.J., Parker, J.W. and Taylor, C.R., Cytoplasmic as opposed to surface la antigens expressed on human peripheral blood lymphocytes and monocytes, Clin. Exp. Immunol., 49 (1982) 355-363. 14 Hughes, W., Dodgson, M.C.H. and Machennan, D.C., Chronic cerebral hypertensive disease, Lancet, 2 (1954) 770-774. 15 Omitted. 16 Ishii, H., Nakano, M., Tsubouchi, J., Ishikawa, T., Uchiyama, H., Hiraishi, S., Tahara, C., Miyajima, Y. and Kazama, M., Establishment of enzyme immunoassay of human thrombomodulin in plasma and urine using monoclonal antibodies, Thrombosis Haemostasis, 63 (1990) 157-162. 17 Ishii, H., Nakano, M., Tsubouchi, J., Kazama, M. and Majerus, P.W., Distribution of thrombomodulin in human tissue and characterization of thrombomodulin in plasma, Acta Haematol. Japan, 51 (1988) 1228-1233. 18 Ishii, H., Salem, H.H., Bell, C.E., Laposata, E.A. and Majerus, P.W., Thrombomodulin, an endothelial anticoagulant protein is absent from the human brain, Blood, 67 (1986) 362-365. 19 Jackman, R.W., Beeler, D.L., Van DeWater, L. and Rosenberg, R.D., Characterization of a thrombomodulin eDNA reveals structural similarity to the low density lipoprotein receptor, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 8834-8838. 20 Kumada, T., Dittman, W.A. and Majerns, P.W., A role for thrombomodulin in the pathogenesis of thrombin-induced thromboembolism in mice, Blood, 71 (1988) 728-733.
21 Maruyama, I., Bell, C.E. and Majerus, P.W., Thrombomodulin is found on endothelium of arteries, veins, capillaries, and lymphatics and on syncytiotrophoblast of human placenta, J. Cell Biol., 101 (1985) 363-371. 22 Maruyama, I. and Majerus, P.W., The turnover of thrombin: thrombomodulin complex in cultural human umbiblicai vein endothelial ceils and A549 lung cancer cells. Endocytosis and degradation of thrombin, J. Biol. Chem., 260 (1985) 1543215438. 23 Moore, K.L., Anderoli, S.P., Esmon, N.L., Esmon, C.T. and Bang, N.U., Endotoxin enhances tissue factor and suppresses thrombomodulin expression of human vascular endothelium in vitro, J. Clin. Invest., 79 (1987) 124-130. 24 Moore, K.L., Esmon, C.T. and Esmon, N.C., Tumor necrosis factor leads to the internalization and degradation of thrombomodulin from the surface of bovine aortic endothelial cells in culture, Blood, 73 (1989) 159-165. 25 Nawroth, EP., Handley, D.A., Esmon, C.T. and Stern, D.M., Interleukin 1 induces endothelial cell procoagulant while suppressing cell surface anticoagulant activity, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 3460-3464. 26 Nawroth, P.P. and Stern, D.M., Modulation of endothelial cell hemostatic properties by tumor necrosis factor, J. Exp. Med., 164 (1986) 740-745. 27 Salem, H.H., Maruyama, I., Ishii, H. and Majerus, P.W., Isolation and characterization of thrombomodulin from human placenta, J. Biol. Chem., 259 (1984) 12246-12251. 28 Suzuki, K., Stenflo, J., Dahlback, B. and Theodorsson, B., Inactivation of human coagulation factor V by activated protein C, J. Biol. Chem., 258 (1983) 1914-1920. 29 Van Hinsbergh, V.W.M., Bertina, R.M., Van Wijngaarden, A., Van Tilburg, N.H., Emeis, J.J. and Haverkate, E, Activated protein C decreases plasminogen activator inhibitor activity in endothelial cell-conditioned medium, Blood, 65 (1985) 4 ~ 451.
Research Reports
Regional distribution of thrombomodulin in human brain Vicky L.Y. Wong 1'2, Florence M. Hofman 2, Hidemi Ishii 3 and Mark Fisher 1 Departments of lNeurology and ZPathology, University of Southern California School of Medicine, Los Angeles, CA 90033 (U.S.A.) and JDepartment of Clinical Biochemistry, Teikyo University, Sagamiko, Kanagawa (Japan) (Accepted 26 February 1991)
Key words: Thrombomodulin; Brain; Endothelium; Infarct; Coagulation; Stroke; Thrombosis
Thrombomodulin, an integral membrane protein with important anticoagulant properties, was evaluated for its distribution in blood vessels ,of human brain. Adjacent sections from frozen human brain were stained for von Willebrand Factor (an endothelium marker) or thrombomodulin. The number of thrombomodulin-positive blood vessels was expressed as a percentage of all blood vessels in a specific area. The density ratio was compared among cortical and subcortical regions in the human brain. Our results indicated that thrombomodulin density in neocortex, cerebellum, medulla and hippocampus was similar. There was, however, significantly less thrombomodulin in putamen (P < 0.01), pous (P < 0.001) and mesencephalon (P < 0.001) compared to neocortex. The regional distribution of thrombomodulin may prove to be helpful for understanding the development of thrombotic diseases of the brain.
INTRODUCTION
MATERIALS AND METHODS
T h r o m b o m o d u l i n is an integral m e m b r a n e protein with a molecular weight of 75,000 D a s,19,27. T h r o m b o -
Tissue samples
modulin has b e e n localized to arteries, veins, capillaries and lymphatics o f nearly every h u m a n tissue and is n o r m a l l y p r e s e n t on endothelial cells 21. The role of t h r o m b o m o d u l i n in n o r m a l anticoagulation is critical to understanding hemostasis. T h r o m b o m o d u l i n binds t h r o m b i n and changes the substrate specificity of thromb i n - t o w a r d protein C. The t h r o m b i n - t h r o m b o m o d u l i n complex activates p r o t e i n C 20,000 times m o r e rapidly than t h r o m b i n alone 7. A c t i v a t e d protein C in turn inactivates clotting factors Va and V I I I a 1~,2a and plasminogen activator inhibitor type 129, thus shifting the hemostatic balance to anticoagulation. A l t h o u g h the distribution of t h r o m b o m o d u l i n is wides p r e a d t h r o u g h o u t the body, t h r o m b o m o d u l i n has been r e p o r t e d to be low o r absent from e n d o t h e l i u m in h u m a n brain ~7'~s'2~, while present in rabbit brain 4,5. The present study addresses the question w h e t h e r there is any specific p a t t e r n of t h r o m b o m o d u l i n distribution in brain. We r e p o r t immunohistochemical evidence that t h r o m b o m o dulin is p r e s e n t in h u m a n brain and has a p a t t e r n of high and low distribution in 12 distinct brain regions. O u r results indicate that ports, mesencephalon and p u t a m e n exhibit a relative deficiency of t h r o m b o m o d u l i n as c o m p a r e d to neocortex.
Tissue from 10 normal human brains were used in this study; the specimens were obtained at autopsy with postmortem intervals ranging 4-14 h, age range 35-79 years; there were 8 males and 2 females (Table IA). Tissues were obtained from the Alzheimer Disease Research Center of Southern California, University of Southern California, Los Angeles, CA. Neuropathologic examination revealed no significant pathologic abnormalities, including no evidence of small vessel disease. Brain specimens were cut, snap frozen in liquid nitrogen and stored at -70 °C or in fiquid nitrogen. Cryostat sections were serially cut at 7/~m, left to air dry overnight, and then fixed with acetone for 5 min.
Antisera Polyclonal. Rabbit anti-human thrombomodulin IgG was prepared as previously de~2ribed16, using 120/~g of isolated placental thrombomodulin in Bacto complete Freud's adjuvant as the immunogen. The specificity of the purified antibody has been confirmed by Western blot analysis and inactivation of thrombomodulin activity1~. In addition, the activity of 12 ng purified human thrombomodulin was inhibited 50% by 150 ng rabbit anti-human thrombomedulin IgG, whereas the activity of 36 ng purified rabbit thrombomodulin was not inhibited even with 3/~g of rabbit anti-human thrombomodulin IgG. Monoclonal. Mouse anti-human thrombomodulin IgG was prepared as previously described16, using 20/~g of isolated placental thrombomodulin in complete Freud's adjuvant as the immunogen. Specificity of the mouse monoclonal antibody was confirmed by Western blot analysisTM. The monoclonal antibody was used to stain selected brain sections using the alkaline phosphatase staining method in order to confirm the polyclonal reagent staining specificity.
lmmunoperoxidase staining reagents and procedure Rabbit anti-yon Willebrand Factor (vWF) (Dako, Santa Barbara,
Correspondence: M. Fisher, University of Southern California, Department of Neurology, 2025 Zonal Street, Los Angeles, CA 90033, U.S.A.
TABLE I Thrombomodulin distribution Clinical variables in human patients used for thrombomodulin study (A) and regional distribution of thrombomodulin in human brain (B). P.M. time: postmortem time in hours. (,4) Case characteristics Clinical variables
Case number
Age Sex P.M. time
1
2
3
4
5
6
7
8
9
10
71 F 4
47 M 8.5
64 M 10
38 M 9
72 F 6.5
79 M 10
55 M 9.5
38 M 8.5
54 M 10
35 M 14
(B) Thrombomodulin-vWF ratios Brain regions
Case number 1
Cortical areas
Hippocampus Putamen Cerebellum Mesencephalon Pons Medulla
4 6 9 10 21 41
0.93
2
3
4
1.10
1.00
0.65 1.08 0.82 1.01
0.72 0.55
Mean + S.D.
0,98 0.77 0.94 0.84 0.80 0.85 0.56 0.59 0.43 0.99
5
6
7
8
9
10
0.89 1.03
0.92 1,10
0,88
0.95 +_0.12
0.90 0.51 0.80
1.03 0.71 1.00
0.85 0.80 0.80
1.02 0.94 0.94 0,74
0.37 0.54
0.76 0.88
0.90
0.95
0.90 + 0.75 + 0.85 + 0.69 + 0.53 + 0.85 +
1.00
1.00
1.02
0.80
0.66
0.57 0.55
0.86
0.10 0.14' 0.17 0.12" 0.15" 0.18
*P < 0.01, *P < 0.001.
CA) was used at a dilution of 1:1000; rabbit anti-human prostatic epithelial antigen, an irrelevant antibody control, was used at 1:300. Rabbit anti-human thrombomodulin antibody was used at a dilution of 1:500. Biotinylated goat anti-rabbit immunoglobulin (Vector, Burlingame, CA) and avidin-biotin peroxidase complex (ABC, Vector) were used at manufacturer's prescribed concentrations. Amino-ethyl carbazole (AEC) was used as the substrate. Immunoperoxidase staining was performed as previously described 13. Acetone-fixed tissue sections, cut within 48 h, were rehydrated with phosphate-buffered saline (PBS, pH 7.4) for 10 rain, incubated with 0.3% HzO z, and washed with PBS. Nonspecific staining was then blocked with 5% solution of normal goat serum (Cedarlan Lab, Hornby, Ontario, Canada) for 15 min at 25 °C. Slides were subsequently blotted dry, and incubated with primary antibody for 30 rain at 25 °C in a humidified chamber. After each incubation, sections were washed in PBS and incubated with 1% bovine serum albumin (Fisher, Richmond, CA) for 5 rain at 25 °C. After the primary antibody, slides were incubated with biotin-labeled secondary antibody for 30 min, and then with ABC for 15 min. The sections were finally incubated with AEC for 10 min, counterstained in Mayer's hematoxylin for 3 min, and mounted in glycerol and PBS. Alkaline phosphatase staining reagents and procedure Mouse anti-human thrombomodulin antibody was used at a dilution of 1:1000. Alkaline phosphatase anti-alkaline phosphatase was purchased from Dako (Santa Barbara, CA); alkaline phosphatase substrate kit and levamisole were products of Vector (Burlingame, CA). Alkaline phosphatase staining was performed according to the instruction of the manufacturer. Acetone-fixed tissue sections, cut within 48 h, were rehydrated with Tris-buffered saline (TBS, 0.05 M Tris, 0.15 M NaC1, pH 7.6) for 10 min. Slides were then blotted dry and incubated with monocional anti-thrombomodulin for 45 min at
25 °C. After primary antibody incubation, slides were incubated with rabbit anti-mouse immunoglobulin, and then with alkaline phosphatase mouse anti-alkaline phosphatase immune complex for 30 rain at 25 °C. The sections were incubated with alkaline phosphatase substrate with levamisol as inhibitor of endogenase alkaline phosphatase for 30 min in the dark. Sections were then counterstained and mounted in the manner described above. Quantitation Adjacent sections were stained and examined under the light microscope at high power (400x magnification). A grid was placed over the tissue section and the number of stained blood vessels within a 1 mm 2 area was counted. The number of blood vessels per mm 2 ranged from 20 to 300. Three corresponding areas in vWF-stained sections and in thrombomodulin-stained sections were counted and compared. The density of thrombomodulin was expressed as a ratio and calculated as follows: The total number of thrombomodulin-stained blood vessels/mm2 in 3 one-mm 2 areas The total number of vWF-stained blood vessels/mm 2 in 3 one-mm 2 areas To determine variability from section to section, adjacent sections from 7 brain regions were stained for the same antibody (vWF or thrombomodulin); the density ratio of adjacent sections with the identical antibody had a mean -+ S.D. value of 1.06 + 0.02. These controls indicated that the number of blood vessels within adjacent sections do not vary substantially. The thrombomodulin density ratio of different brain regions was compared and the statistical differences between regions were analyzed using unpaired Student's t-tests.
RESULTS Human brain specimens, stained with anti-vWF, demonstrated intense labeling of endothelial cells in both large and small blood vessels in different parts of the brain. Figs. 1A and 2A show staining of vessels (5-10/zm in width) in cortical area 21 and putamen, respectively. Adjacent sections stained with anti-thrombomodulin are shown in Fig. 1B (Table IB, Case 3) and 2B (Table IB, Case 6), respectively. Staining with anti-thrombomodulin was limited to endothelial cells (Fig. 3). Thrombomodulin monoclonal antibody also stained endothelial cells in multiple brain regions (Fig. 4). Cortical and subcortical regions were examined with both anti-thrombomodulin and anti-vWF. Thrombomodulin density ratios among the 17 neocortical regions were similar and these ratios were thus presented as composite. The mean neocortical value 0.95 + 0.12 (Table 1B), was compared to the ratio of each of the other regions. The hippocampus, cerebellum and medulla exhibited ratios of 0.90 + 0.10, 0.85 + 0.17 and 0.85 _+ 0.18, respectively (Table IB). In contrast, the pons, mesencephalon and putamen showed ratios of 0.53 + 0.17, 0.69 + 0.12 and 0.75 + 0.14, respectively (Table
Fig. 2. Immunostaining of putamen: adjacent sections stained with
anti-vWF (A, 1:1000) and polyclonal anti-thrombomodulin (B) antibody (1:500); counterstained with hematoxylin. Bar = 100/~m.
IB), which were significantly less than that of the neocortex. Similar thrombomodulin densities could be demonstrated in adjacent gray and white matter of tissue samples. For example, the thrombomodulin density ratio for gray and white matter in pons was 0.55 and 0.54, respectively, from case 1, 0.43 and 0.50, respectively, from case 4, and 0.76 and 0.60, respectively, from case 8. Thrombomodulin levels in different brain regions varied independent of age, gender or postmortem interval (Table IA).
Fig. 1. Immunostaining of area 21: adjacent sections stained with
anti-vWF (A, 1:1000) and polyclonal anti-thrombomodulin (B) antibody (1:500); counterstained with hematoxylin. Bar = 100/~m.
Fig. 3. Immunostaining of blood vessel from area 6 with polyclonal anti-thrombomodulin antibody (1:500). Bar = 10/~m.
4
Fig. 4. Monoclonal anti-thrombomodulin antibody (1:1000) stained blood vessels in pons; counterstained with hematoxylin. Bar = 100 /~m.
To further investigate the differing thrombomodulin densities, we evaluated the relationship between thrombomodulin density ratio and mean vWF-stained blood vessel number. Thrombomodulin density ratio was inversely associated with the number of vWF-stained blood vessels from all brain regions from the 10 cases, with simple correlation coefficient r -- -0.31 (P < 0.03). This suggests that increasing vessel density in brain may be associated with a reduced proportion of vessels staining with thrombomodulin. However, in putamen, mesencephalon and pons, the mean vWF-stained blood vessel numbers (160.6 + 81.0, 164.7 _+ 64.8, 180.2 + 69.5, respectively) were not significantly different from that of cortex (151.2 _+ 72.1) indicating that relatively low thrombomodulin ratios in these 3 subcortical areas were not secondary to vessel density. DISCUSSION In this study, we have demonstrated thrombomodulin staining of endothelium in blood vessels in various regions of the brain. The distribution of thrombomodulin varied among regions. Hippocampus, cerebellum, medulla along with neocortical regions exhibited comparable thrombomodulin staining; significantly less thrombomodulin-positive vessels were found in pons, putamen REFERENCES 1 Omitted. 2 Omitted. 3 Conway, E.M. and Rosenberg, R.D., Tumor necrosis factor suppresses transcription of the thrombomodulin gene in endothelial cells, MoL Cell. BioL, 8 (1988) 5588-5592. 4 DeBault, L.E., Esmon, N.L., Olson, J.R. and Esmon, C.T., Distribution of the thrombomodulin antigen in the rabbit vasculature, Lab. Invest., 54 (1986) 172-178. 5 DeBault, L.E., Esmon, N.L., Smith, G.P. and Esmon, C.T., Localization of thrombomodulin antigen in rabbit endothelial cells in culture, Lab. Invest., 54 (1986) 179-187.
and mesencephalon, compared to neocortex. The expression of thrombomodulin in the brain may be regulated. In endothelial cell culture, binding of thrombin to thrombomodulin causes internalization of the complex, degradation of thrombin and subsequent recycling of thrombomodulin to the surface 22. Additionally, thrombomodulin can be down-regulated by endotoxin, tumor necrosis factor (TNF) and interleukin123'25"26. For example, T N F has been reported to regulate thrombomodulin by suppressing its transcription 3, along with producing degradation and internalization of thrombomodulin 24. Whether these cytokines play important roles in regulating the distribution of thrombomodulin in human brain awaits further investigation. Recent experimental studies further emphasize the anti-thrombotic role of thrombomodulin 12'2°. Preinjection of anti-thrombomodulin in mice followed by thrombin injection helped potentiate the lethal effect of thrombin while preinjection of thrombomodulin prolonged the survival period 2°. Gomi et al. further localized fibrin deposits in large and small blood vessels of lungs in mice receiving thrombin but not in large blood vessels of animals preinjected with recombinant thrombomodulin 12. These results help demonstrate the role of thrombomodulin in preventing thrombosis. The significance of our results showing an irregular thrombomodulin distribution in the brain may be related to the topography of ischemic cerebral vascular disease. Lacunar infarcts (microinfarcts) are mostly absent from cerebral cortex and cerebellum but are more commonly found in the putamen and pons 9'1°. The pons is also one of the most c o m m o n subcortical regions for infarcts of all sizes TM. These results raise the possibility of an association between relative thrombomodulin deficiency and ischemic cerebral vascular disease. Acknowledgements. This study was supported by Grants NS 20989 (to M.E) and EY 08144 (to EM.H.) from the National Institutes of Health. The authors gratefully acknowledge the Alzheimer Disease Research Center of Southern California funded by Grant P50AG05142 from the National Institute of Aging. We are also grateful to Dr. David Hinton for helpful discussions. Mrs. Florence Miyagawa and Mrs. Lisa Doumak provided skillful secretarial assistance.
6 Esmon, C.T., The role of protein C and thrombomodulin in the regulation of blood coagulation, J. Biol. Chem., 264 (1989) 4743-4746. 7 Esmon, C.T. and Owen, W.G., Identification of an endothelial cell cofactor for the thrombin-catalyzed activation of protein C, Proc. Natl. Acad. Sci. U.S.A., 78 (1981) 2249-2252. 8 Esmon, N.L., Owen, W.G. and Esmon, C.T., Isolation of a membrane-bound cofactor for thrombin-catalyzed activation of protein C, J. Biol. Chem., 257 (1982) 859-864. 9 Ferrand, J., Essai sur l'h~mipldgie des vieillards. Les lacunes de desintegration cgrgbrale, Paris, Thesis, 1902. 10 Fisher, C.M., Lacunes: small, deep cerebral infarcts, Neurology, 15 (1965) 774-784.
11 Fulcher, C.A., Gardiner, J.E., Griffin, J.H. and Zimmerman, T.S., Proteolytic inactivation of human factor VIII procoagulant protein by activated protein C and its analogy with factor V, Blood, 63 (1984) 486-489. 12 Gomi, K., Zushi, M., Honda, G., Kawahara, S., Matsuzaki, O., Kanabayashi, T., Yamamoto, S., Maruyama, I. and Suzuki, K., Antithrombotic effect of recombinant human thrombomodulin on thrombin-induced thromboembolism in mice, Blood, 75 (1990) 1396-1399. 13 Hofman, EM., Billing, R.J., Parker, J.W. and Taylor, C.R., Cytoplasmic as opposed to surface la antigens expressed on human peripheral blood lymphocytes and monocytes, Clin. Exp. Immunol., 49 (1982) 355-363. 14 Hughes, W., Dodgson, M.C.H. and Machennan, D.C., Chronic cerebral hypertensive disease, Lancet, 2 (1954) 770-774. 15 Omitted. 16 Ishii, H., Nakano, M., Tsubouchi, J., Ishikawa, T., Uchiyama, H., Hiraishi, S., Tahara, C., Miyajima, Y. and Kazama, M., Establishment of enzyme immunoassay of human thrombomodulin in plasma and urine using monoclonal antibodies, Thrombosis Haemostasis, 63 (1990) 157-162. 17 Ishii, H., Nakano, M., Tsubouchi, J., Kazama, M. and Majerus, P.W., Distribution of thrombomodulin in human tissue and characterization of thrombomodulin in plasma, Acta Haematol. Japan, 51 (1988) 1228-1233. 18 Ishii, H., Salem, H.H., Bell, C.E., Laposata, E.A. and Majerus, P.W., Thrombomodulin, an endothelial anticoagulant protein is absent from the human brain, Blood, 67 (1986) 362-365. 19 Jackman, R.W., Beeler, D.L., Van DeWater, L. and Rosenberg, R.D., Characterization of a thrombomodulin eDNA reveals structural similarity to the low density lipoprotein receptor, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 8834-8838. 20 Kumada, T., Dittman, W.A. and Majerns, P.W., A role for thrombomodulin in the pathogenesis of thrombin-induced thromboembolism in mice, Blood, 71 (1988) 728-733.
21 Maruyama, I., Bell, C.E. and Majerus, P.W., Thrombomodulin is found on endothelium of arteries, veins, capillaries, and lymphatics and on syncytiotrophoblast of human placenta, J. Cell Biol., 101 (1985) 363-371. 22 Maruyama, I. and Majerus, P.W., The turnover of thrombin: thrombomodulin complex in cultural human umbiblicai vein endothelial ceils and A549 lung cancer cells. Endocytosis and degradation of thrombin, J. Biol. Chem., 260 (1985) 1543215438. 23 Moore, K.L., Anderoli, S.P., Esmon, N.L., Esmon, C.T. and Bang, N.U., Endotoxin enhances tissue factor and suppresses thrombomodulin expression of human vascular endothelium in vitro, J. Clin. Invest., 79 (1987) 124-130. 24 Moore, K.L., Esmon, C.T. and Esmon, N.C., Tumor necrosis factor leads to the internalization and degradation of thrombomodulin from the surface of bovine aortic endothelial cells in culture, Blood, 73 (1989) 159-165. 25 Nawroth, EP., Handley, D.A., Esmon, C.T. and Stern, D.M., Interleukin 1 induces endothelial cell procoagulant while suppressing cell surface anticoagulant activity, Proc. Natl. Acad. Sci. U.S.A., 83 (1986) 3460-3464. 26 Nawroth, P.P. and Stern, D.M., Modulation of endothelial cell hemostatic properties by tumor necrosis factor, J. Exp. Med., 164 (1986) 740-745. 27 Salem, H.H., Maruyama, I., Ishii, H. and Majerus, P.W., Isolation and characterization of thrombomodulin from human placenta, J. Biol. Chem., 259 (1984) 12246-12251. 28 Suzuki, K., Stenflo, J., Dahlback, B. and Theodorsson, B., Inactivation of human coagulation factor V by activated protein C, J. Biol. Chem., 258 (1983) 1914-1920. 29 Van Hinsbergh, V.W.M., Bertina, R.M., Van Wijngaarden, A., Van Tilburg, N.H., Emeis, J.J. and Haverkate, E, Activated protein C decreases plasminogen activator inhibitor activity in endothelial cell-conditioned medium, Blood, 65 (1985) 4 ~ 451.
