Address reprint requests to Dr. David J. Loskutoff, Committee for the Study of Vascular ...... Prendergast GC, Diamond LE, Dahl D, Cole MD: The c-myc-regulated ...
American journal of Pathology, Vol. 143, No. 3, September 1993 Copynight C) American Society for Investigative Pathology
Type 1 Plasminogen Activator Inhibitor Gene Expression Following Partial Hepatectomy
Jacob Schneiderman,*t Michael Sawdey,* Heather Craig,* Terri Thinnes,* Gerald Bordin,t and David J. Loskutoff* From the Committee for the Study of Vascular Biology* and Division of Vascular and 7Toracic Surge,yt The Scripps Research Institute, La Jolla, California; and the Department of Clinical Pathology,* Scripps Clinic, La Jolla, California
A murine model ofpartial hepatectomy (PH) was employed to investigate type 1 plasminogen activator inhibitor (PAI-1) gene expression in regenerating liver. Mice were anesthetized, and a portion of the left lobe of the liver was ligated and resected distal to the ligature, and at various times thereafter, total liver RNA was prepared and analyzed by Northern blotting. PH caused a transient increase in PAI-I messenger (m)RNA that was apparent within I to 2 hours after surgery, was maximal at 8 hours (ninefold increase over sham-operated controls), and then slowly declined. Analysis of discrete liver segments demonstrated much greater induction of PAI-I mRNA in the region adjacent to PH than in more distal regions. Further analysis ofthe adjacent tissue by in situ hybridization revealed that PAI-I mRNA was induced primarily in hepatocytes in the transition zone created by the occluding hemostatic ligature between viable and necrotic tissue. Expression of PAI-I mRNA could also be detected in this transition zone in capsular mesothelial ceUls, subcapsular hepatocytes, and venous endothelial ceUs bordering the area. A much weaker signal was evident in hepatocytes dispersed throughout the remaining intact lobes ofPH mice, and no signal was detected in the livers of sham-operated mice. These observations suggest that PAI-i may be of importance in local tissue remodeling events accompanying liver regeneration. (Am J Pathol
1993, 143:753-762)
Partial hepatectomy (PH) induces the liver to enter a regenerative stage. Regrowth of the ablated tissue involves both parenchymal and nonparenchymal
cells and is effected by humoral factors that act in a temporally and spatially defined manner to elicit hepatic regeneration (for review, see ref. 1). In the prereplicative phase of liver regeneration, PH results in the induction of cellular immediate early genes, such as the transcription factors c-fos2'3 and c-jun. For example, a 13-fold increase in hepatic c-jun messenger (m)RNA was detected 2 hours after PH in mice, whereas the mRNA of a related gene, jun B, was shown to be increased by approximately 50% during the same interval.4 The products of the c-fos and c-jun genes are known to form heterodimers (AP-1) capable of binding to specific DNA sequence elements in the promoter regions of target genes and stimulating their rate of transcription.5'6 Because the gene encoding the extracellular matrix-degrading enzyme transin contains an AP-1 site7 and the level of transin mRNA has been shown to rise following PH, AP-1 was postulated to mediate induction of the transin gene in regenerating liver.4 Type 1 plasminogen activator inhibitor (PAI-1) is the primary physiological inhibitor of both tissue type and urokinase type plasminogen activators (for review, see ref. 8). In vivo, PAI-1 negatively regulates thrombolysis by tissue-type plasminogen activator.9'10 More recently, it has been shown to be deposited within the extracellular matrix of cultured cells11'12 where it may function to inhibit extracellular proteolysis mediated by urokinase type plasminogen activator.13'14 In vitro, PAI-1 is synthesized by a variety of cells, including human hepatocytes15 and hepatoma cells. Its synthesis in hepatoma cells is positively regulated by glucocorticoids,16 and by transforming growth factor-,B (TGF-f3),17 epidermal growth factor,18 and interleukin-1 (11-1),19 agents implicated in the Supported by NIH grant HL47819 to DJL. Accepted for publication May 4, 1993. Address reprint requests to Dr. David J. Loskutoff, Committee for the Study of Vascular Biology (CVB-3), The Scripps Research Institute, 10666 North Torrey Pines Road, La Jolla, CA 92037. This is publication number 7824-CVB from The Scripps Research Institute. Present address of J. Schneiderman is Department of General and Vascular Surgery, Sheba Medical Center, Tel Hashomer 52621, Israel.
753
754
Schneiderman et al
AJP September 1993, Vol. 143, No. 3
control of liver regeneration.20 In the case of TGF-, induction of the PAI-1 gene has been shown to occur at the transcriptional level21 and may involve AP-1like elements present in the PAI-1 promoter.22 Collectively, these observations suggest that PAI-1 may be of importance in liver response to injury. In the studies described below, we employ a mouse model system to demonstrate that PAI-1 mRNA is induced in the liver following PH, primarily in specific regions adjacent to areas of necrosis in the remnant of the resected lobe. These data imply a linkage between PAI-1 gene expression and the remodeling of the extracellular matrix during liver regeneration.
Materials and Methods
Surgical Procedures Twelve-week-old male BalbC/ByJ x C57BI6/J mice (CB6; Scripps Clinic Rodent Breeding Colony) were employed for all experiments. The animals were housed and treated in compliance with the guidelines set forth in "Principles of Laboratory Animal Care" and "Guide for the Care and Use of Laboratory Animals" (NIH publication no. 80-23, revised 1985). Surgical procedures were performed under general anesthesia, which was introduced and maintained with Metofane (methoxyflurane; PitmanMoore, Mundeleine, IL). Mice undergoing partial hepatectomy underwent a midline laparotomy, through which the left lobe of the liver was delivered. A hemostatic ligature was applied at the base of the lobe, and the lobe distal to it divided with a scalpel. A narrow rim of devascularized tissue beyond the hemostatic ligature (referred to as the necrotic tissue in the remnant of the left lobe) was left unresected. The abdomen was closed with a running suture, and the mice were allowed to recover from anesthesia. Control animals underwent a sham operation including midline laparotomy, delivery, and gentle rubbing of the left lobe of the liver without resection. At various times after surgery (0, 1, 2, 4, 8, 16, 24, and 120 hours), individual mice were euthanized, and various tissues were harvested. Harvested segments of the liver (see Figure 2) were designated as follows: A) the left lobe of the liver (sham-operated animals) or its remains (subtotal left hepatectomy); B) a wedge from the right lobe of the liver; and C) the rest of the remaining liver, including the right, median, and caudate lobes. Small portions of each segment were collected and processed for in situ hybridization. The remainder of the segment was stored for Northern blot analysis
as detailed below. In control experiments, heart, kidney, and spleen were also harvested for Northern analysis.
Northern Blot Analysis For preparation of total RNA, freshly isolated tissues were snap-frozen and stored in liquid nitrogen. Total RNA was extracted by the acid guanidium thiocyanate-phenol-chloroform method,23 and its concentration determined by measurement of sample absorbance at 260 nm. For Northern blotting, total RNA samples (30 pg) were denatured, electrophoresed in formaldehyde-containing agarose gels, transferred to nylon membranes (Biotrans; ICN, Irvine, CA), and fixed by ultraviolet irradiation as previously described.24 Equal RNA loading and transfer were confirmed by visualization of the ethidium bromide-stained RNA in the nylon membrane under ultraviolet light. The blots were prehybridized for 30 minutes at 65 C in a solution consisting of 50
mmol/L piperazine-N,N'-bis[2-ethanesulfonic acid] (PIPES) pH 6.8, 100 mmol/L NaCI, 20 mmol/L Na2PO4, 30 mmol/L NaHPO4, 1 mmol/L ethylenediaminetetraacetic acid (EDTA), 5% (wt/vol) sodium dodecyl sulfate, and then hybridized for 16 hours at 65 C in the same solution containing approximately 106 cpm/ml radiolabeled probe. The blots were washed at 65 C with prewarmed 100 mmol/L NaCI, 10 mmol/L Na citrate, 5% (wt/vol) sodium dodecyl sulfate, and then autoradiographed at -80 C employing Kodak XAR-5 film with intensifying screens.
Probes For Northern blot hybridizations, a plasmid subclone of the murine PAI-1 complementary (c)DNA25 was obtained as a generous gift from Drs. L. Diamond and M. Cole (Princeton University, Princeton, NJ). The 0.5-kb PAI-1 cDNA insert, corresponding to the 5' terminal region of murine PAI-1 mRNA, was excised by digestion with the appropriate restriction enzymes and purified by preparative gel electrophoresis. A plasmid containing the cDNA for human glyceraldehyde 6-phosphate dehydrogenase (GAP) was obtained from American Type Culture Collection (Rockville, MD), and linearized before use. The probes were radiolabeled to specific activities >108 cpm/pg by the random primer technique,26 employing a[32P]dGTP (> 3,000 Ci/mmol; Amersham, Arlington Heights, IL). For in situ hybridization analysis, a fragment corresponding to bases 1 to 1,085 of the murine PAI-1 cDNA25 was subcloned into the
Type 1 Plasminogen Activator Inhibitor in Murine Liver
755 AJP September 1993, Vol. 143, No. 3
vector pGEM 3Z (Promega Biotec, Madison, WI) and linearized with the appropriate restriction enzyme before use. Sense and anti-sense PAI-1 riboprobes were synthesized by in vitro transcription of the linearized template with T7 or SP6 RNA polymerase, respectively, in the presence of a[35S]UTP (> 1,000 Ci/mmol; Amersham, Arlington Heights,
IL).
Densitometric Analysis Northern blot autoradiographic signals were quantitated by densitometric scanning, utilizing an Ultroscan XL laser densitometer (LKB, Bromma, Sweden). Densitometric data were expressed as absorbance units/mm2 (AU/mm2) of peak area. All densitometric readings were within the linear range of measurement, as established by previous control experiments comparing AU/mm2 with cpm obtained by direct radioanalytic imaging of the same blot.27 Densitometry units were obtained by dividing AU/ mm2 obtained for PAI-1 mRNA by the corresponding AU/mm2 obtained for GAP mRNA.
Tissue Preparation Liver pieces (4 x 4 mm) were fixed in fresh 4% (wt/ vol) paraformaldehyde overnight at 4 C, incubated in 15% (wt/vol) sucrose for 4 hours at 4 C, and frozen in optimum cutting temperature compound (Tissue Tek, Miles) before freezing at -70 C.
In Situ Hybridization In situ hybridizations were carried out essentially as
described.28 Tissue sections (4 p) were fixed with (wt/vol) paraformaldehyde in phosphatebuffered saline for 10 minutes at 4 C and deproteinated with 1 pg/ml proteinase K in 500 mmol/L NaCI, 10 mmol/L Tris-HCI, pH 8.0, for 10 minutes at 23 C. Prehybridization was performed for 1 hour at 42 C in prehybridization buffer, consisting of 50% (wt/vol) formamide, 0.3 mol/L NaCI, 20 mmol/L Tris-HCI, pH 8.0, 5 mmol/L EDTA, 1 x Denhardt's solution, 10% (wt/vol) dextran sulphate, and 10 mmol/L dithiothreitol, followed by hybridization for 16 hours at 55 C in prehybridization buffer supplemented with 0.6 x 107 cpm/ml 35S-labeled riboprobe and 2.5 mg/ml yeast transfer RNA. The sections were washed with 2x standard saline citrate (SSC), 10 mmol/L 2-mercaptoethanol, 1 mmol/L EDTA (2 x 5 minutes, 23 C), incubated with RNAse A (20 pg/ml in 500 mmol/L NaCI/10 mmol/L Tris-HCI) for 30 minutes 4%
at 23 C, washed with 2x SSC, 10 mmol/L 2-mercaptoethanol, 1 mmol/L EDTA (2 x 5 minutes, 23 C), and then washed at high stringency in 0.1x SSC, 10 mmol/L 2-mercaptoethanol, 1 mmol/L EDTA for 2 hours at 60 C. The sections were washed again in 0.5x SSC (4 x 10 minutes, 23 C), dehydrated in a graded alcohol series containing 0.3 mol/L NH4OAc, dried, coated with NTB2 emulsion (Kodak, diluted 1:1 in water), and exposed at 4 C for 3 to 12 weeks. The slides were developed in D19 developer (Kodak) for 2 minutes, fixed, washed in water, and counterstained with hematoxylin and eosin. To control for nonspecific hybridization, the above procedures were performed employing a sense riboprobe (data not shown).
Results To investigate expression of the PAI-1 gene in regenerating liver, PH was performed on adult male CB6 mice. Mice underwent midline laparotomy, followed by ligation and resection of the portion of the left lobe distal to the ligature (approximately 25% of the liver mass). As a control for the effects of surgery, mice underwent otherwise identical sham operations in which the left lobe of the liver was exposed but not ligated or removed. At selected time points following surgery (ie, 0, 1, 2, 4, 8, 16, 24, and 120 hours), both the liver and selected extrahepatic tissues (eg, heart, kidney, and spleen) were harvested for analysis. Total RNA was prepared from the entire liver (sham mice) or its remains (PH mice), and PAI-1 gene expression was analyzed by Northern blotting employing a murine PAI-1 cDNA probe (Figure 1A). Little if any PAI-1 mRNA was detected in the livers of untreated (0 hour) mice. However, in the livers of both PH- and sham-operated mice, PAI-1 mRNA increased significantly within 1 hour of surgery, reaching a maximum at 8 hours and remaining elevated through 24 hours. Correspondingly greater increases in PAI-1 mRNA levels were observed in the livers of PH mice throughout this period. By 120 hours (ie, 5 days), the level of PAI-1 mRNA had returned to basal levels in both groups. To quantitate the response of PAI-1 mRNA to PH, replicate time course experiments were performed and analyzed by Northern blotting as above. To control for variations in total RNA content between samples, the blots were rehybridized with a GAP probe. The autoradiographic signals for PAI-1 and GAP mRNAs were quantitated by densitometric scanning, and the densitometric values for PAI-1
756
Schneiderman et al
AJP September 1993, Vol. 143, No. 3
A
Partial Hepatctomy
Sham
Figure 1. Northern blot analysis of PAI-1 mRVA in murine liver follou'ing partial hepatectomy. A: Male CB6 mice were sacrificed at designated S~~~~~~~~~~~~~~~~~~~~~~~~~~~~~61 time intervals following partial hepatectomy (PH mice) or sham surgerv (sham mice). Total RNA uas extracted from the entire liver (sham mice) or its remains (PH mice) and 30 jig analyzed by Northern blotting nith a -2P-labeled murine PAI- I cDNA probe. The lenigth of the band shown (3 0 kb) was estimatedfrom its nzigration relative to the 28s and 18s ribosomal RNA bands visualized by ethidiunm staining. Th7e anitoradiogram was exposed for 48 hours. B: Relative fold increase in PAI-I mRNA in PH mice conipared to shanm mice. Replicate time course experiments were performed, anid PAI-1 and GAP mRNA levels were determined by Northern blotting as above. The autoradiographic signals uwere quantitated by densitometric scanning and the densitometric valnies for PAI-1 mRNA were normalized to those of GAP. The relative Jbld itncrease in PAI-i mRNA following PH was then calculated from the meatn of the replicate values obtained for each time point. All densitometric readings were within the linear range of measurenient, as established by previons control ex-
I
Time (Hra):
B 3
Iz
cc
. eE
.iSr .
0c ;
Time following Surgery (Hrs)
mRNA were normalized to those of GAP. The fold increase in PAI-1 mRNA in PH relative to sham mice was then calculated from the normalized data (Figure 1B). Some PH-specific induction of PAI-1 mRNA was evident within 1 to 2 hours. However, the majority of the increase was apparent at later times, rising to fivefold at 4 hours and ninefold at 8 hours. By 16 to 24 hours, the relative fold increase in PAI-1 mRNA had declined to approximately threefold or lower (data not shown). Experiments were performed to analyze this response within discrete regions of the liver (Figure 2). PH and sham mice were sacrificed at 8 hours following surgery, and specific segments of the liver
Mouse No.:
periments.27
depicted in Figure 2 were then removed. Segment A is the left lobe of the liver (sham mice), or its remains (PH mice); segment B is a wedge of tissue taken from the right lobe of the liver; and segment C is the sum of the right, median, and caudate lobes. Total RNA was prepared from each segment, and changes in PAI-1 and GAP mRNAs were assessed by Northern blotting as above. Quantitative analysis of the blot autoradiograms by densitometric scanning revealed that PAI-1 mRNA levels had increased an average of 17- and eightfold in segments A and B, respectively, in PH mice compared to the sham mice (ie, mice 1 and 2 vs. 5 and 6; Figure 2). The average level of PAI-1 mRNA in seg-
Partial Hepatectomy 1 2 3 4 U.Q
Sham 5
6
.
Liver S
-PA--1
-GAP Figure 2. Northern blot analysis of PAI-1 mRNA in murine liver following partial hepatectomy: segmental analvsis. PH or sham mice were sacrificed at 8 hours following surgery, and total RNA was prepared from the regions indicated in the schematic diagram of the liver shown at the left and analyzed by Northern blotting: A denotes the stump of the remaining left lobe after partial hepatectomy (shaded area), or the unresected left lobe in sham mice; B denotes a limited segment from the edge of the right hepatic lobe; C denotes the suim of the right, median, and caudate lobes (the latter tuo not illustrated in the schenie); A + B + C denotes the entire remaining liver tissuie afterpartial hepatectomy. The top auitoradiograni shou's the results uising the 32P-labeled murine PAI-1 cDNA probe. The bottom auitoradiogram shows the results obtained Jbllouwing rehybridization of the blot with a GAP probe. The size of the GAP mRNA species (:1.1 kb) was estimated from its migration relative to the 28s and 18s ribosomal RNA species visualized by ethidium broniide stainitng. The auitoradiograms uwere exposed for 48 hours. -
. I
I
I
I
-
,
Type 1 Plasminogen Activator Inhibitor in Murine Liver 757 AJP September 1993, Vol. 143, No. 3
nuclei and eosinophilic cytoplasm (Figure 4C). A few sinusoidal cells within the transition zone also exhibited a strong hybridization signal (Figure 4C), and positive venous endothelial cells could frequently be observed in large veins intersected by the transition zone (Figure 4, B and D). Expression of PAI-1 mRNA was also clearly evident within the viable subcapsular cells that border the region of necrosis in segment A (Figure 4E). This layer of cells probably retains its viability via the capsular microcirculation. The hybridization signal was present predominantly within the first and second layers of subcapsular hepatocytes closest to the hepatic mesothelium and to a lesser extent in surface mesothelial cells. Rare positive sinusoidal cells could also be observed within the viable subcapsular region. In the remaining viable portion of segment A, and throughout segment B, a less intense hybridization signal for PAI-1 mRNA was evident (Figure 2). This decrease in PAI-1 mRNA seemed to result from significantly fewer positive hepatocytes in these areas (Figure 4F). Occasionally, a mild subcapsular hybridization signal was also noted in segment B. A pattern of hybridization generally similar to that described above was detected at 16 hours after PH in both segments A and B. By 24 hours, the level of signal had declined markedly, with the exception of positive venous endothelial cells within the necrotic area in close proximity to the transition zone and capsular regions (Figure 4G). By 120 hours after PH, no positive signal could be discerned. No positive signal could be visualized at any time in liver sections of sham mice (Figure 4H), despite the presence of PAI-1 mRNA levels comparable to those of PH mice (eg, Figure 1; compare PAI-1 mRNA autoradiographic signals for 8-hour sham versus 24-hour PH mice). This finding may reflect a more diffuse induction of PAI-1 gene expression throughout the livers of sham mice (ie, more cells are induced but at levels undetectable by in situ
ments A + B + C (mice 3 and 4) after PH seemed to be intermediate between that observed for segments A and B alone, although some variability between animals existed. In the sham controls (mice 5 and 6), little difference in PAI-1 mRNA levels was evident between segments A and B, although variability between animals was again apparent. To examine the response of the PAI-1 gene in extrahepatic tissues to PH, total RNA was prepared from the heart, kidney, and spleen of PH- and sham-operated mice at various times after surgery and analyzed for PAI-1 mRNA as above. Although only data from the heart are presented (see Figure 3), similar results were obtained for both kidney and spleen (data not shown). Densitometric analysis of the data shown in Figure 3 revealed that the level of PAI-1 mRNA in the heart increased approximately twofold following surgery in both PH and sham mice. Thus, in contrast to the results obtained for the liver, no greater increase could be discerned in extrahepatic organs of PH mice compared to the shams. Histopathological examination of liver segment A at 8 hours after PH revealed an abrupt and welldelineated zone of transition between the viable and necrotic tissue (Figure 4A). Although the sinuses within this transition zone contained a few neutrophils, little if any infiltration by lymphocytes or macrophages was observed, and increased numbers of Kupffer cells were not evident. All of the hepatocytes within the ischemic tissue seemed to be necrotic, and the sinuses in this region contained Kupffer cells as well as erythrocytes in varying stages of degeneration. Analysis of segment A (obtained 8 hours after PH) by in situ hybridization revealed a strong positive hybridization signal for PAI-1 mRNA that extended along the length of the transition zone between viable and necrotic cells (Figure 4B). Examination of this region under higher power magnification confirmed that the majority of the positive cells were hepatocytes, identifiable by their large
Partial Hepatectomy ai~~~~~~~
*
F
Sham
Time (Hrs): Figure 3. Nothern1 blot analysis of PAI-1 mRNA in murine heart following partial hepa-
-PAI-i
tectomY. Mice were sacrificed at designated time intervals flblowing PH or sham surgery and total RNA (30 pig) was extracted from the heart and analyzed by Northern blotting as described for Figure 1. The auitoradiograms depicting PAI-1 and GAP mRNAs were exposedjfr hours rest,ectivelv. 5 davs and }f-, sk
-GAP
,
vr ,v
[-
"--
Wo=v
$t-
I
-
-w
Type 1 Plasminogen Activator Inhibitor in Murine Liver 759 AJP September 1993, Vol. 143, No. 3
hybridization). Sections analyzed with a sense PAI-1 riboprobe were also uniformly negative (data not shown).
Discussion We have employed a murine model system to investigate the response of the PAI-1 gene to partial hepatectomy. Surgical resection of the majority of the left lobe of the liver (approximately 25% of the total liver mass) resulted in a dramatic increase in hepatic PAI-1 mRNA levels. However, a similar although lower response was evoked by the surgical procedure itself. Both PH- and sham-operated mice exhibited increases in PAI-1 mRNA levels within 1 hour of surgery, rising to maximal levels at 8 hours and declining thereafter (Figure 1A). This response is consistent with the known acute phase behavior of PAI-1 in man. For example, clinical studies have established that the concentration of PAI-1 in blood may rise sharply following major surgery,29 trauma,30 or experimental administration of lipopolysaccharide.3132 The biosynthetic origin of acute phase PAI-1 is presently unclear. Most acute phase proteins are believed to originate from parenchymal cells of the liver, and previous studies have demonstrated induction of PAI-1 mRNA in these cells.33 Other studies employing mice34 and rats35 have implicated vascular endothelial cells in the liver and other organs as the site of increased PAI-1 synthesis following lipopolysaccharide treatment. Thus whereas the cellular origin of acute phase PAI-1 remains to be established, the data presented above indicate the liver is likely to represent a significant source of PAI-1 in acute phase reactions following major surgery. Relative to sham-operated mice, PH mice exhibited a consistently greater increase in hepatic PAI-1 mRNA levels in the period from 1 to 24 hours after surgery, with maximal induction (ie, > ninefold increase relative to sham mice) evident at 8 hours (Figure 1 B). This induction seemed to be specific to the liver because no such difference in PAI-1 gene
expression was evident in extrahepatic tissues (Figure 3). Analysis of the response of discrete liver segments to PH (Figure 2) revealed that PAI-1 mRNA levels were induced 17-fold over the sham controls in the remnant of the resected left lobe (segment A). In situ hybridization of this region at 8 hours after PH (Figure 4) revealed sharply delineated rows of PAI-1 mRNA-positive cells, consisting predominately of hepatocytes bordering the region of ischemic necrosis. PAI-1 mRNA was also localized to venous endothelial cells, capsular mesothelial cells, and sinusoidal cells proximal to the necrotic tissue. The marked induction of PAI-1 mRNA in this region of the liver at 8 hours after PH (Figure 2) would thus seem to result from a highly localized increase in its expression immediately adjacent to the area of necrosis. The nature of the stimulus for this increase is unknown. Induction of PAI-1 mRNA in hepatic parenchymal cells has recently been demonstrated following treatment of rats with the synthetic glucocorticosteroid dexamethasone.33 However, in human subjects given adrenocorticotropic hormone, no cause-effect relationship could be demonstrated between plasma PAI-1 activity and cortisol levels, despite elevations in plasma cortisol levels comparable to those observed postoperatively.29 The PAI-1 gene also is known to be responsive to inflammatory cytokines such as interleukin124 and tumor necrosis factor-a,24 which may mediate in part the response of target genes in acute phase reactions and gram negative sepsis.36'37 Histopathological evidence of inflammation (eg, infiltration of lymphocytes or macrophages, or increased numbers of Kupffer cells) was not observed coincident with the localized increase of PAI-1 mRNA at 8 hours after PH. Thus, whereas systemic release of these cytokines may account for the more generalized increase in PAI-1 mRNA observed in hepatocytes within more remote regions of the liver, they would not seem likely to account for its induction in the transition zone or subcapsular areas at 8 hours after PH. However, it
Figure 4. In SitL hybridization analysis of PAI-1 mRNA in murine liver tissue followitng partial hepatectomy. Liver tissues were collected 8 hours after PH and analyzed by either May-Gruenuald Giemsa staining (A) or by in situ hybridization with a -35S-labeled anti-sense PAI 1 riboprobe (B to H), as described in Materials and Methods. A: May-Gruenwald Giemsa stain of PH liver segment A at 8 hours after PH, demonstrating the transition zone ( 7Z) between viable and nonviable tissue. Viable ( V) hepatocytes exhibit darker staining relative to necrotic (N) cells. Note Kupffer cells (arrows) and abundant erythrocytes within necrotic tissue (magnification: 200X). B: In situ hybridization signalfor PAI-1 mRNA in the transition zone of segment A at 8 hours after PH. Note intersection of transition zone with large hepatic vein (HV) at bottom left, and positive signal in luminal endothelium (magnification: 10OX). C: Higher power magnification of transition zone region shown in B. Arrows denote positive hepatocytes; arrowheads denote rare positive sinusoidal cells (magnification: 400X). D: In situ hybridization signal for PAI-1 mRNA in hepatic vein (HV) of segment A at 8 hours after PH. Arrows indicate positive endothelial cells (magnification: 400X). E: High-power view of segment A at 8 hours after PH, demonstrating PAI-1 mRNA hybridization signal in the subcapsular region (magnification: 400X). Note positive hybridization signtal in capsular mesothelial cells (arrowhead) and subcapsular hepatocytes (arrows) (magnification: 400X). F: Scattered PAI-1 mRNA positive hepatocytes (arrou') in segment B 8 hours after PH ( magnification: 400X). G: High-power vieu' of segment A at 24 hours after PH, demonstrating a strong signal within endothelial cells lining hepatic vein (HV) (magnificatiotn: 400X). H: Negative sham liver at 8 hours after suirgery ( magnification: 400X).
760
Schneiderman et al
AJP September 1993, Vol. 143, No. 3
presently cannot be excluded that interleukin-1 and tumor necrosis factor-a may in part stimulate endothelial PAI-1 synthesis in these regions at later times (Figure 4G), as significant infiltration of neutrophils could be visualized in the vicinity of the vessels expressing PAI-1 mRNA by 24 hours after PH. Expression of the PAI-1 gene in both cultured hepatoma38 and endothelial cells21 is strongly stimulated by TGF-f3, which is released from platelets in response to physiological stimuli.39 The release and diffusion of TGF-,B from platelets trapped within the region of ischemic necrosis could thus be hypothesized to induce PAI-1 mRNA in bordering hepatocytes and vascular cells, consistent with the observed pattern of PAI-1 gene expression (Figure 4). Alternatively, because TGF-,B mRNA has been reported to be induced in regenerating liver within 4 hours after PH and this induction correlates with increased accumulation of mRNAs for several extracellular matrix proteins,40 the response of PAI-i mRNA within the transition zone and subcapsular space may reflect de novo synthesis and release of TGF-4 from nonparenchymal cells in this region. Induction of PAI-1 mRNA in vascular cells may otherwise reflect induction of PAI-1 synthesis by basic fibroblast growth factor, which is released from vascular cells under hypoxic conditions41'42 or by
hypoxemia.443 Extensive remodeling of the extracellular matrix is known to occur during liver regeneration. Transin, the rat homologue of human stromelysin, is a secreted metalloproteinase with a broad specificity for extracellular matrix components. The mRNA encoding transin was recently shown to be expressed in hepatocytes within areas subsequently undergoing necrosis due to CC14 intoxication.44 Interestingly, transin also participates in the activation of interstitial collagenase in cooperation with plasmin.45'46 Peripheral to regions of necrosis, the induction of PAI-1 may serve to limit extracellular plasmin generation by cell-surface-bound urokinase. 13 The increased synthesis of PAI-1 in these regions may thus counteract plasmin-mediated extracellular matrix degradation within viable tissue regions bordering necrotic areas. The PAI-1 and transin genes both contain functional AP-1 sites. Moreover, induction of the protooncogene products c-fos and c-jun, which together activate gene transcription from AP-1 sites,5'6 occurs rapidly following PH.4 In situ hybridization studies have revealed that expression of c-fos and c-jun mRNAs following CC14 injection is detected first within pericentral hepatocytes, which later un-
dergo necrosis.47 However, the expression of these and other proto-oncogenes subsequently spreads to adjacent areas and ultimately involves the entire liver.447 Thus, whereas the induction of PAI-i mRNA adjacent to regions of necrosis may initially involve AP-1-responsive elements in the PAI-1 gene, other factors must also regulate its expression within these regions. The identity of these factors remains to be determined.
References 1. Michalopoulos GK: Liver regeneration: molecular mechanisms of growth control. FASEB J 1990, 4:176187 2. Kruijer W, Skelly H, Botteri F, van der Putten H, Barber JR, Verma IM, Leffert HL: Proto-oncogene expression in regenerating liver is simulated in cultures of primary adult rat hepatocytes. J Biol Chem 1986, 261:79297933 3. Thompson NL, Mead JE, Braun L, Goyette M, Shank PR, Fausto N: Sequential protooncogene expression during rat liver regeneration. Cancer Res 1986, 46:3111 4. Alcorn JA, Feitelberg SP, Brenner DA: Transient induction of c-jun during hepatic regeneration. Hepatology 1990, 11:909-915 5. Angel P, Allegretto EA, Okino ST, Hattori K, Boyle WJ, Hunter T, Karin M: Oncogene jun encodes a sequence-specific trans-activator similar to AP-1. Nature 1988, 332:166-171 6. Chiu R, Boye WJ, Meck J, Smeal T, Hunter T, Karin M: The c-fos protein interacts with c-jun/AP-1 to stimulate transcription of AP-1 responsive genes. Cell 1988, 54: 541-552 7. Matrisian LM, Leroy P, Ruhimann C, Gesnel M-C, Breathnach R: Isolation of the oncogene and epidermal growth factor-induced transin gene: complex control in rat fibroblasts. Mol Cell Biol 1986, 6:1679-1686 8. Loskutoff DJ: Regulation of PAI-1 gene expression. Fibrinolysis 1991, 5:197-206 9. Hanss M, Collen D: Secretion of tissue type plasminogen activator and plasminogen activator inhibitor by cultured human endothelial cells, modulation by thrombin, endotoxin, and histamine. J Lab Clin Med 1987, 109:97-104 10. Schleef RR, Higgins DL, Pillemer E, Levitt JJ: Bleeding diathesis due to decreased functional activity of type 1 plasminogen activator inhibitor. J Clin Invest 1989, 83:1747-1752 11. Laiho M, Saksela 0, Andreasen PA, Keski-Oja J: Enhanced production and extracellular deposition of the endothelial-type plasminogen activator inhibitor in cultured human lung fibroblasts by transforming growth factor-f. J Cell Biol 1986, 103:2403-2410
Type 1 Plasminogen Activator Inhibitor in Murine Liver 761 AJP September 1993, Vol. 143, No. 3
12. Mimuro J, Schleef RR, Loskutoff DJ: The extracellular matrix of cultured bovine aortic endothelial cells contains functionally active type 1 plasminogen activator inhibitor. Blood 1987, 70:721-728 13. Cubellis MV, Andreasen P, Ragno P, Mayer M, Dano K, Blasi F: Accessibility of receptor-bound urokinase to type-1 plasminogen activator inhibitor. Proc NatI Acad Sci USA 1989, 86:4828-4832 14. Ellis V, Wun T-C, Behrendt N, Ronne E, Dano K: Inhibition of receptor-bound urokinase by plasminogenactivator inhibitors. J Biol Chem 1990, 265:9904-9908 15. Sprengers ED, Princen HMG, Kooistra T, van Hinsbergh VWM: Inhibition of plasminogen activators by conditioned medium of human hepatocytes and hepatoma cell line hep G2. J Lab Clin Med 1985, 105: 751-758 16. Heaton JH, Gelehrter TD: Glucocorticoid induction of plasminogen activator and plasminogen activator inhibitor messenger RNA in rat hepatoma cells. Mol Endocrinol 1988, 3:349-355 17. Fujii S, Sobel BE: Induction of plasminogen activator inhibitor by products released from platelets. Circulation 1990, 82:1485-1493 18. Lucore CL, Fugii S, Wun TC, Sobel BE, Billadello JJ: Regulation of the expression of type 1 plasminogen activator inhibitor in Hep G2 cells by epidermal growth factor. J Biol Chem 1988, 263:15845-15848 19. de Boer JP, Abbink JJ, Brouwer MC, Meijer C, Roem D, Voorn GP, Lambers JWJ, van Mourik JA, Hack CE: PAI-1 synthesis in the human hepatoma cell line Hep G2 is increased by cytokines-evidence that the liver contributes to acute phase behaviour of PAI-1. Thromb Haemost 1991, 65:181-185 20. Daily PO, Johnston GC, Simmons CJ, Moser KM: Surgical management of chronic pulmonary embolism. J Thorac Cardiovasc Surg 1959, 109:523-531 21. Sawdey M, Podor TJ, Loskutoff DJ: Regulation of type 1 plasminogen activator inhibitor gene expression in cultured bovine aortic endothelial cells: induction by transforming growth factor-f, lipopolysaccharide, and tumor necrosis factor-a. J Biol Chem 1989, 264: 10396-10401 22. Keeton M, Curriden SA, van Zonneveld AJ, Loskutoff DJ: Identification of regulatory sequences in the type 1 plasminogen activator inhibitor gene responsive to transforming growth factor f. J Biol Chem 1991, 266: 23048-23052 23. Chomczynski P, Sacchi N: Single-step method of RNA isolation by acid guanidinium thyiocynate-phenolchloroform extraction. Anal Biochem 1987, 162:156159 24. Schleef RR, Bevilacqua MP, Sawdey M, Gimbrone MA Jr, Loskutoff DJ: Interleukin 1 (IL-1) and tumor necrosis factor (TNF) activation of vascular endothelium: effects on plasminogen activator inhibitor (PAI-1) and tissue type plasminogen activator (tPA). J Biol Chem 1988, 263:5797-5803
25. Prendergast GC, Diamond LE, Dahl D, Cole MD: The c-myc-regulated gene mri encodes plasminogen activator inhibitor 1. Mol Cell Biol 1990, 10:1265-1269 26. Feinberg AP, Vogelstein B: A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal Biochem 1983, 132:6-13 27. Sawdey M, Loskutoff DJ: Regulation of murine type 1 plasminogen activator inhibitor gene expression in vivo. Tissue specificity and induction by lipopolysaccharide, tumor necrosis factor-a, and transforming growth factor-f. J Clin Invest 1991, 88:1346-1353 28. Wilcox JN, Smith KM, Schwartz SM, Gordon D: Localization of tissue factor in the normal vessel wall and in the atherosclerotic plaque. Proc NatI Acad Sci USA 1989, 86:2839-2843 29. Aillaud MF, Juhan-Vague I, Alessi MC, Marecal M, Vinson MF, Arnaud C, Vague PH, Collen D: Increased PAinhibitor levels in the postoperative period: no causeeffect relation with increased cortisol. Thromb Haemost 1985, 54:466-468 30. Kluft C, Verheijen JH, Jie AFH, Rijken DC, Preston FE, Sue-Ling HM, Jespersen J, Aasen AD: The postoperative fibrinolytic shutdown: a rapidly reverting acute phase pattern for the fast-acting inhibitor of tissuetype plasminogen activator after trauma. Scand J Clin Lab Invest 1985, 45:605-610 31. Colucci M, Paramo JA, Collen D: Generation in plasma of a fast-acting inhibitor of plasminogen activator in response to endotoxin stimulation. J Clin Invest 1985, 75:818-824 32. Suffredini AF, Harpel PC, Parrillo JE: Promotion and subsequent inhibition of plasminogen activation after administration of intravenous endotoxin to normal subjects. N Engl J Med 1989, 320:1165-1172 33. Konkle BA, Schuster SJ, Kelly MD, Harjes K, Hessett DE, Bohrer M, Tavassoli M: Plasminogen activator inhibitor-1 messenger RNA expression is induced in rat hepatocytes in vivo by dexamethasone. Blood 1992, 79:2636-2642 34. Keeton M, Eguchi Y, Sawdey M, Ahn C, Loskutoff DJ: Cellular localization of type 1 plasminogen activator inhibitor mRNA and protein in murine renal tissue. Am J Pathol 1993, 142:59-70 35. Quax PHA, van den Hoogen CM, Verheijen JH, Padro T, Zeheb R, Gelehrter TD, van Berkel TJC, Kuiper J, Emeis JJ: Endotoxin induction of plasminogen activator and plasminogen activator inhibitor type 1 mRNA in rat tissues in vivo. J Biol Chem 1990, 265:1556015563 36. Dinarello CA: Interleukin-1 and the pathogenesis of acute phase response. N Engl J Med 1984, 311:14131418 37. Beutler B, Cerami A: Cachectin: more than a tumor necrosis factor. N Engl J Med 1987, 316:379-385 38. van Zonneveld A-J, Curriden SA, Loskutoff DJ: Type 1 plasminogen activator inhibitor gene: functional analysis and glucocorticoid regulation of its promoter. Proc
762
Schneiderman et al
AJP September 1993, Vol. 143, No. 3
Natl Acad Sci USA 1988, 85:5525-5529 39. Assoian RK, Sporn MB: Type-beta transforming growth factor in human platelets: release during platelet degranulation and action on vascular smooth muscle cells. J Cell Biol 1986, 102:1217-1223 40. Jakowlew SB, Mead JE, Danielpour D, Wu J, Roberts AB, Fausto N: Transforming growth factor-,8 (TGF-f3) isoforms in rat liver regeneration: messenger RNA expression and activation of latent TGF-f3. Cell Regulation 1991, 2:535-548 41. D'Amore PA, Thompson RW: Mechanisms of angiogenesis. Annu Rev Physiol 1987, 49:453-464 42. Folkman J, Klagsburn M: Angiogenic factors. Science 1987, 235:442-447 43. Lynch D, Ansel P, Levene R: Effects of anoxia on gene expression in human endothelial cells. J Cell Biol 1989, 107:581 A 44. Herbst H, Heinrichs 0, Schuppan D, Milani S, Stein H:
Temporal and spatial patterns of transin/stromelysin RNA expression following toxic injury in rat liver. Virchows Arch 1991, 60:295-300 45. Fini ME, Karmilowicz MJ, Ruby PL, Beeman AM, Borges KA, Brickerhoff CE: Cloning of a complementary DNA for rabbit proactivator. A metalloproteinase that activates synovial cell collagenase, shares homology with stromelysin and transin, and is coordinately regulated with collagenase. Arthritis Rheumatol 1987, 30:1254-1264 46. He CS, Wilhelm SM, Pentland AP, Marmer BL, Grant GA, Eisen AZ, Goldberg GI: Tissue cooperation in a proteolytic cascade activating human interstitial collagenase. Proc Natl Acad Sci USA 1989, 86:2632-2636 47. Herbst H, Milani S, Schuppan D, Stein H: Temporal and spatial patterns of proto-oncogene expression at early stages of toxic liver injury in the rat. Lab Invest 1991, 65:324-333