JOURNAL OF PATHOLOGY, VOL. 178: 190-194 (1996)

TISSUE PLASMINOGEN ACTIVATOR PRODUCTION BY MONOCYTES IN VENOUS THROMBOLYSIS

K . s. soo, A. D. R. NORTHEAST*, L. c. HAPPERFIELD?, K. G. BURNAND" AND L. G. BOBROW?

Tuniour Immunology Unit, Imperial Cancer Research Fund and Department of Surgery, University College London Medical School, London, U. K.; *Surgical Unit, St Thomas' Hospital, London, U K.; tlmrnunohistology Laboratory, Imperiul Cancer

Research Fund, London, U. K.

SUMMARY Laminated occlusive thrombus was induced in the rat inferior vena cava (IVC) by a distal stenosis and injection of thrombin.

Immunocytochemistry was performed on serial cryostat sections of the thrombus for tissue plasminogen activator (tPA) and a variety of phenotype markers for mononuclear cells. There was little tPA in 2-day-old thrombus. However, tPA was present in significant quantities in 1- and 2-week-old thrombus. Most of the staining for tPA was associated with monocytes, which had infiltrated the thrombus in large numbers. No caval endothelium was seen in these sections. By 4 weeks, the IVC had re-canalized and new endothelium had formed; tPA staining was weakly positive in the endothelium and smooth muscle. In situ hybridization with a digoxigenin-labelled RNA probe confirmed the monocytes as the main source of tPA. This study shows that large numbers of infiltrating monocytes are present in venous thrombosis and that they are the main source of tPA.

KEY WORDS-tiSSUe plasrninogen activator; thrombolysis; monocytes

INTRODUCTION

Fibrinolytic activity in the form of tPA has been demonstrated in the vasa vasorum, in the endothelium, and in small amounts in smooth muscle of the vein wall.'4 The endothelium is traditionally regarded as the main source of tPA.'-'

Previous work in a rat model has indicated that the functional activity of tPA in the vessel wall of the thrombosed IVC is significantly lower than in non- thrombosed controls for up to 21 days after the onset of thrombosis.' This was not surprising, since the endo- thelium disappeared under the thrombus soon after it was formed. However, tPA levels in the thrombus itself rose progressively from the time of thrombus formation for up to 2 weeks," which could only be explained by a source other than the vascular endothelium. In this study, we set out to investigate the source of tPA in this rat IVC thrombosis model. We examined the role of monocytes, because thrombosis is accompanied by an influx of mononuclear cells with monocytes being the major component.

Peripheral blood monocytes belong to the mono- nuclear phagocytic system and originate in the bone marrow from monoblasts and promonocytes.""2 Monocytes leave the circulation within 1-3 days and subsequently differentiate into various types of tissue macrophage under the influence of various micro- environmental factors. l 3 l 5 In response to inflammation or injury, the adhesiveness of the vascular endothelium for monocytes is increased. The recruitment of large numbers of blood monocytes to the tissue satisfies the increased demand for tissue macro phage^.'^.'^ Throm- bosis is associated with a large monocytic infiltrate.I7

Addressee for correspondence: K. S. Soo, DNAX Research Institute, 901 California Avenue, Palo Alto, CA 94304. U.S.A.

CCC 0022-341 7/96/020190-05 0 1996 by John Wiley & Sons, Ltd

Monocytes adhere rapidly to damaged endothelium" and may differentiate into macrophages with a large number of secretory products and biological activities.I9 Although monocytes have been shown to produce tPA in this has not been demonstrated in vivo.

MATERIALS AND METHODS

Rat thrombus model Thrombosis was produced in the IVC of male

Carnforth Sprague Europe rats with a mean weight of 300g using the St Thomas' Each rat was anaesthetized with intramuscular fentanyl (Evans) and amylobarbitone sodium (Lilly). Median laparotomy was carried out and the IVC was mobilized. A ligature was placed around the IVC just below the left renal vein, creating a stenosis. The proximal IVC just above the iliac veins was clamped and 21U of rat thrombin (Sigma) in Ringer's solution was injected into the IVC below the ligature via a 25 G needle. The clamp was released after 5 min and flow in the IVC resumed. Reproducible laminated thrombus developed in the TVC below the ligature within 72 h. A negative control was provided by performing the surgical dissection and applying the ligature in an identical rat, the IVC being injected with Rigner's solution without thrombin. Ten rats were used, five in the experimental group and five as controls. The rats were killed on day 0, day 2, week 1, week 2, and week 4.

Zmmunocytochemistry

Cryostat sections 5 p m thick were prepared from the rat IVC. Positive controls for tPA staining were pro- vided by a tPA secreting rat cell line A15A5 (European Collection of Animal Cell Cultures. Porton Down,

Received 28 November 1994 Accepted I 0 April 1995

191 tPA PRODUCTION BY MONOCYTES IN VENOUS THROMBOLYSIS

U.K.) which was pelleted, flash frozen, and cryostat sectioned. Haematoxylin and eosin (H & E) staining was performed on all representative sections to ensure morphological integrity. Positive controls for the mono- nuclear cell phenotype markers were provided by cryo- stat sections of rat spleen, lymph node, and thymus. Negative controls were achieved by omitting the primary antibody. Monoclonal mouse anti-recombinant tPA antibody MA-16D2 was used for the tPA staining.23 The monoclonal mouse anti-rat B-cell-specific leucocyte common antigen antibody MRC 0X3324 and the mono- clonal mouse anti-rat CD3 antibody IF425 (Serotec) were used as B-cell and T-cell markers, respectively. Monocytes were identified using the monoclonal mouse anti-rat antibody ED126 (Serotec). The monoclonal mouse anti-rat CD45 (leucocyte common antigen) anti- body MRC OX127 was used to identify leucocytes. Some of the sections were also stained for CDlIWCD18 using monoclonal mouse anti-rat antibodies ED728 (Serotec) and ED828 (Serotec). Staining was carried out using the alkaline phosphatase/anti-alkaline phosphatase (APAAP) method, with Fast Red as the substrate.29 Sections were counterstained in Mayer’s haematoxylin and mounted.

In situ hybridization

The cDNA template fragment of 427 bases3’ was supplied in the EcoRl site of the pGEM-1 vector (Promega) for transcription of antisense rat tPA RNA. Due to the location of the EcoRl site at the end of the multiple cloning site in pGEM-1, it was impossible to generate sense RNA from the tPA/pGEM- 1 plasmid for use as negative control. The fragment was subcloned into the EcoRI site of the pBluescript SK+ vector (Stratagene). Linearization of the new tPNBluescript plasmid with BamHI restriction enzyme (Pharmacia) in conjunction with T7 RNA polymerase (Boehringer Mannheim) enabled the generation of antisense RNA. Linearization with Hind111 restriction enzyme (Pharmacia) in conjunction with T3 RNA polymerase (Boehringer Mannheim) resulted in the generation of sense RNA. Sense and antisense RNA probes were prepared by using a digoxigenin- 1 1 -dUTP transcription system (Boehringer Mannheim), following the manufac- turer’s protocol. Labelling was confirmed by a Northern blot. The RNA probe was electrophoresed against a RNA molecular weight ladder in a formaldehyde de- naturing gel. After electrophoresis, the RNA was blotted onto a Hybond-C nylon filter (Amersham) and the RNA ladder (BRL Gibco) was marked on the filter under UV illumination. The labelled probe was visualized with the alkaline phosphatase immunodetection method. The filter was incubated with an alkaline phosphatase- conjugated sheep anti-digoxigenin antibody (Boehringer Mannheim) and developed with nitro-blue tetrazoiiud 5-bromo-4-chloro-3-indolyl phosphate (NBTIBCIP) (Sigma). Probe labelling was confirmed on the basis of RNA staining corresponding to the correct molecular weight. The quantity of RNA produced was calculated from photo-absorbance at 260 nm.

Five-micrometre-thick cryostat sections were cut onto slides that were coated with aminopropyltriethoxysilane

(Sigma). Positive controls were provided by the A15A5 cell line. Negative controls were achieved by using a sense RNA probe. In situ hybridization was carried out using a published protocol (Boehringer Mannheim) modified for use with cryostat sections. Probes were visualized with the alkaline phosphatase immunodetec- tion method. Sections were incubated with an alkaline phosphatase-conjugated sheep anti-digoxigenin anti- body (Boehringer Mannheim) and visualized by NBT/ BCIP (Sigma).

RESULTS Gross morphology

Haematoxylin and eosin-stained sections of the rat IVC at day 0 showed the formation of a thrombus in the IVC, obliterating its lumen. The caval endothelium remained intact. At 2 days, the thrombus still occupied the lumen of the IVC but the underlying endothelium had disappeared (Fig. la). At I week, thrombolysis was underway and by 2 weeks, there were multiple areas of re-canalization within the IVC, lined by newly formed endothelium. At 4 weeks, the thrombus had resolved and new endothelium had formed round the lumen. All thrombus-containing sections were distinguished by a heavy infiltrate of mononuclear cells within the thrombus (Fig. la).

Immunocytochemistry

Immunocytochemistry for T-cell and B-cell markers was largely negative in all sections, but all the mononu- clear cells in the thrombi were strongly positive for ED 1, which showed that they were monocytes (Fig. l b ). The same cells were also positive for CD45 (MRC 0 x 1 ) and CD 1 1 blCD 18 (ED7 and ED8), which confirmed their leucopoietic origins. Endothelial cells, when present, were weakly positive for tPA (Fig. lc) and did not stain with any of the leucocyte markers. The most striking feature was the overwhelmingly positive tPA staining shown by the monocytes in sections of the day 2, week 1, and week 2 thrombi (Fig. Id). Strong tPA expression was associated with the mononuclear cell infiltrate and not with the endothelium. Thrombolysis was com- plete by week 4, when minimal tPA was demonstrated (Fig. le).

In situ hybridization

In situ hybridization echoed the results seen with immunocytochemistry. The antisense RNA probe dem- onstrated strong cytoplasmic localization in the mono- cytes within the day 2, week 1, and week 2 thombi (Fig. lf). The sense control was completely negative. The majority of the tPA message came from the monocytes.

DISCUSSION

In normal blood vessels, tPA is present in the endothelium, the vasa vasorum, and at low levels in the

192 K. S. SO0 ET AL.

Fig. 1-(a) H & E stained section of c a w with thrombus on day 2. The lumen was completely obliterated by a thrombus with a large mononuclear cell infiltrate. The endothelium under the thrombus had disappeared and only the underlying smooth muscle layer was seen. (h) Monocyte staining with ED1 antibody in a week 1 thrombus. The majority of the mononuclear cells infiltrating the thrombus were positive for this marker. (c) tPA staining in a normal cava showing slight positivity in the smooth muscle but not in the endothelium of the cells. (d) tPA staining in a week 1 thrombus demonstrating positivity in the mononuclear cell infiltrate. (e) tPA staining in a week 4 thrombus showing its virtual absence in the vessel wall, which had a thickened smooth muscle layer. (f) In siru hybridization using an antisense R N A probe for tPA in a week 1 thrombus, showing strong mRNA staining in the monocytes within the thrombus

smooth muscle cells.5 The vascular endothelium has traditionally been described as the main source of tPA and it has been suggested that the vessel wall is the main regulatory site of the fibrinolytic system.3' However, the sustained level of fibrinolytic activity demonstrated in the thrombus for up to 2 weeks after thrombus formation9 when no endothelium was demonstrable indicated that such a hypothesis is incomplete. As re-endothelization could only occur in the presence of re-canalization, which required thrombolysis and thus tPA, the endothelium was most unlikely to be the only source of tPA in the system.

We set out to look at the distribution of tPA in the thrombus using immunocytochemistry and found it to be predominantly associated with the mononuclear cell infiltrate. Phenotyping on serial sections showed these cells to be monocytes, based on their positive staining with the monocyte-specific ED1 antibody, CD45 leuco- cyte common antigen antibody, and CDlIbKD18 anti- bodies (ED7 and ED8). CD45 is expressed exclusively on leucopoietic cells and its presence reinforced the observation that the mononuclear cell infiltrate in the thrombus did not contain any endothelial component. The ED7 and ED8 antibodies stain specifically a

tPA PRODUCTION BY MONOCYTES IN VENOUS THROMBOLYSIS 193

membrane antigen CDI 1 b/CD18 which is present on monocytes, dendritic cells, and granulocytes. There was no staining of the normal endothelium by any of the leucocyte markers. We would ideally like to confirm the absence of endothelial cells within the mononuclear infiltrate by using a specific rat endothelial marker; unfortunately, none was available which did not cross- react with monocytes/macrophages. In situ hybridization demonstrated tPA message in these monocytes and confirmed them as the major source of tPA within the thrombus.

A role for immune cells in haemostasis and thrombo- sis is not new.32,33 However, it appears that the inflam- matory and the coagulative responses associated with thrombosis are linked.Is Many observations suggest that cytokines mediate the recruitment of mononuclear cells to areas of vessel damage and thrombosis. We have demonstrated that the monocytes in venous thrombus play an active part in thrombolysis by secreting tPA. It is likely that monocytes are actively recruited to sites of thrombosis. Soon after thrombus formation, the adjacent endothelium is lost and monocytes begin to accumulate. Damaged endothelium is capable of secret- ing interleukin-8 and monocyte chemotactic protein-I , which are both attractants for m~nocytes . '~ Further- more, platelet aggregation, an essential feature of thrombosis, results in the release of platelet factor 4, platelet-derived growth factor, p-thromboglobulin, epidermal growth factor, and transforming growth factor$, which are all chemotactic for various leuko- cytes.

The clinical use of recombinant tPA in therapeutic thrombolysis is now well e~ tab l i shed .~~ tPA suffers from a short half-life and continuous infusion of large quan- tities over several hours is required to achieve the desired thrombolysis. The most undesirable complication of this therapy is h a e m ~ r r h a g e . ~ ~ Our observation that mono- cytes in the venous thrombus are mainly responsible for local tPA production raises the possibility of alternative thrombolytic therapy and the prospect of reduced com- plications from haemorrhage. Controlled induction of tPA production by monocytes is a novel approach which might achieve this aim.

This is the first demonstration of tPA production by monocytes in vivo. Monocytes clearly play an important role in the mediation of venous thrombolysis. A clearer understanding of the factors which regulate their recruit- ment to the thrombus and the control of tPA production may be important in furthering the knowledge of thrombolysis.

ACKNOWLEDGEMENTS

We would like to thank Dr H. R. Lijnen for the MA-16D2 antibody, Dr T. Ny for the rat tPA cDNA fragment, and Dr D. W. Mason for the MRC OX1 and MRC OX33 antibodies. We thank Professor P. C. L. Beverley for discussions.

REFERENCES I . Todd AS, Nunn A. The biochemical localisation of fibrinolytic activator.

J Puthd 1959: 7 8 281-283.

2.

3.

4.

5.

6.

7.

8.

9.

10.

1 1 .

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29

30

31

Noordhoek Hegt V. Relations between activation and inhibition of fibri- nolysis in the wall of human arteries. Thromh Rrs 1977: 3 8 407-419. Noordhoek Hegt V. Distribution and variation offibrinolytic activity in the walls of human arteries and veins. Huemostusis 1976; 5 355-372. Warren BA. Fibrinolytic properties of vascular endothelium. 5 r J Exp Pothol 1963; 44: 365-372. Nachman RL, Hajjar KA. Endothelial cell fibrinolytic assembly. Ann N Y Acud Sci 1991; 614 240-249. Ljugner H, Holmberg H, Kjaelgaard A, Nilsson IM, Astedt B. Immuno- logical characterisation of plasminogen activators in the human vessel wall. J Clin Pathol 1983; 3 6 1046-1049. Larsson A, Aatedt B. lmmunohistochemical localisation of tissue plasmino- gen activator and urokinase in the vessel wall. J Clin Pufhol 1985; 38: 140-145. I

Northeast ADR, Eastham D, Gaffney PJ, Burnand KG. The anatomical distribution of fibrinolytic activity in normal rat blood vessels. Fihrfnol~y.sls 1990; 4(S3): 140. Northeast ADR, Burnand KG. The response of the vessel wall to throm- bosis: the in vivo study of venour thrombolysis. Ann NYAcudSci 1992: 667: 127-140. Northeast ADR, Gaffney PJ, Burnand KG. Urokinase production in organising thrombus. Br J Surg 1992; 79 1228. van Furth R, Diesselhoff-den Dulk MMC. The kinetics of pronionocytcs and monocytes in the bone marrow. J Ex" Mrd 1970; 132 81 3-828. Goud TJLM, Schotte C, van Furth R. Proliferative characteristics of monoblasts grown in eitro. J E.up Mrd 1975: 142 1200-1217. Issekutz TB, lssekutz AC, Movat HZ. The in vivo quantitation and kinetics of monocyte migration into acute inflammatory tissue. Am J Pafhol 1988: 103: 47-55. van Furtb R, Cohn ZA. The origin and kinetics of mononuclear phagocytes. J Exp Med 1968; 128: 415435. Johnston RB. Monocytes and macrophages. N Engl J M d 1988; 318 147-752. Ross R. The pathogenesis of atherosclerosis-an update. N Engl J Mrd 1986; 314 488-500. Doutremepuich C, Toulemonde F, De Seze 0, Pereira F, Anne MC, Doutremepuich F. Influence of heparin and low molecular weight heparin liaction of the leukocytes in an experimental venous thrombosis. J Leuk 5iol 1987; 42: 628-63 I. Wakefield TW, Greenfield LJ, Rolfe MW, el ol. Inflammatory and pro- coagulant mediator interactions in an experimental baboon model of venous thrombosis. Thromh Huenio.stosiu 1993; 69 164-172. Nathan CF. Secretory products of macrophages. J Clin Invesr 1987; 7 9 3 19-326. Hart PH, Vitti GF, Burgess DR, Singleton DK, Hamilton JA. Human monocytes can produce tissue-type plasminogen activator. J Exp Mud 1989; 169 1509-1514. Hart PH, Burgess DR, Vitti GF, Hamilton JA. Interleukin-4 stimulates human monocytes to produce tissue-type plasminogen activator. Blood 1989; 7 4 1222-1225. Northeast ADR, Drain R, Heyderman E, Kenny E, Burnand KG. A new thrombolysis model with endothelial conservation for the study of the role of plasminogen activators. J Puthol 1989; 157: 175A. Lijnen HR, Webb PD, van Hoef B, et ul. Biochemical and biological properties of a recombinant tissue-type plasminogen activator derived from the rat JM1-229 cell line. T/~rornh Huemostusis 1992; 67: 239 247. Woollett GR, Barclay AN, Puklavec M, Williams AF. Molecular and antigenic heterogeneity of the rat leukocyte-common antigen from thymo- cytes and T and B lymphocytes. Eur J Imrnwzoi 1985; 1 5 168-173. Tanaka T, Masuko T, Yagita H, Tamura T, Hashimoto Y. Characterisation of a CD3-like rat T-cell surface antigen recognised by a monoclonal antibody. J lmmunol 1989; 142 2191-2795. Dijkstra CD, Dopp EA, Joling P, Kraal G. The heterogeneity of mono- nuclear phagocytes in lymphoid organs: distinct macrophage iubpopu- lations in the rat are recognised by monoclonal antibodies EDI, ED2 and ED3. Inimunology 1985; 54: 589-599. Woollett GR, Barclay AN, Puiklavec M, Williams AF. Molecular and antigenic heterogeneity of the rat leukocyte common antigen from thymo- cytes and T and B lymphocytes. Eur J Inimunol 1985; 15: 168-173. Dasmoiseaux JGMC, Dopp EA, Beelan RHJ, Dijkstra CD. Heterogeneity of macrophages in the rate evidenced by variability in determinants: two new anti-rat macrophage antibodies against a heterodimer of 160 and 95 kd (CDllblCDlR). J Leucocyfe B i d 1989; 4 6 556-564. Cordell JL, Falini B, Erber WN, u f ul. lmmunoenzymatic labelling of monoclonal antibodies using immune complexes of alkaline phospha- tase and monoclonal anti-alkaline phosphatase (APAAP complexes). J Histochern Cytochern 1984; 3 2 219-229. Peng XR, Hseuh AJ, Ny T. Transient and cell-specific expression of tissue-type plasminogen activator and plasminogen-activator inhibitor type I results in controlled and directed proteolysis during gonadotrophin- induced ovulation. Euv J BGchem 1993; 214 147-156. Chandler WL. A kinetic model or the circulatory regulation of tissue plasminogen activator. Thromh Hoemost 1991; 6 6 321-328.

'7v .LLJ 00s 's 'n P6 I