Microcystin-LR-induced Ultrastructural Changes in Rats

Vet. Pathol. 279-15 (1990) Microcystin-LR-induced Ultrastructural Changes in Rats S. B. HOOSER, V. R. BEASLEY, E. J. BASGALL, W. W. CARMICHAEL, AND W. M. HASCHEK Departments of Veterinary Pathobiology and Biosciences, University of Illinois, Urbana, IL; and Department of Biological Sciences, Wright State University, Dayton, OH Abstract. The ultrastructure of hepatic, pulmonary, and renal lesions was evaluated in rats injected intraperitoneally with a lethal dose of microcystin-lr (MCLR, 160 wg/kg), a cyclic heptapeptide hepatotoxin produced by the blue-green algae, Microcystis aeruginosa. Hepatic lesions were first seen at 10 minutes post-dosing and consisted of mild widening of hepatocyte intercellular spaces centrilobularly. At 20 minutes post-dosing, hepatocyte plasma membrane alterations were more pronounced, consisting of plasma membrane invagination with formation of variably sized and shaped intracytoplasmic vacuoles, loss of microvilli along the sinusoidal face, and widespread, pronounced hepatocyte separation. By 30 minutes, the space of Disse was markedly widened. At 60 minutes post-dosing, centrilobular areas contained necrotic cells and apparently intact, isolated, organelles intermingled with erythrocytes and platelets. In less severely affected regions there was prominent hepatocyte rounding, and erythrocytes and platelets were present in the widened space of Disse. Large amounts of hepatocellular debris and intact hepatocytes were present in the pulmonary vasculature, while smaller amounts of debris were also seen in the glomerular and peritubular capillaries of the renal cortex. This study shows that initial lesions are confined to shape changes in the plasma membrane of hepatocytes. These changes are consistent with the hypothesis that microcystin-lr induces alterations in the hepatocyte cytoskeleton. Later changes consist of hepatocyte disassociation and necrosis, as well as endothelial damage, which allow release of hepatocytes and debris into the circulation with microembolism in lungs and kidneys. Key words: Kidney; liver; lung; Microcystis aeruginosa; microcystin LR; ultrastructure. Microcystin LR (MCLR, formerly microcystin A or cyanoginosin LR) is a cyclic heptapeptide hepatotoxin produced by the cyanobacterium, Microcystis aeruginosa. Deaths in livestock and wildlife, as well as human illness, have resulted from ingestion of water containing Microcystis aeruginosa. 2-10,16 In all of these species, the liver is the target organ, and the lesion consists of severe hepatic necrosis. Following administration of a lethal dose of MCLR, mice die within 60 to 90 minutes. Rats, however, survive 20 to 32 hours, which is similar to survival times reported in livestock. Because of this similarity, we chose to examine the development of MCLR-induced lesions in the rat. We have shown that lethal doses (LD,o = approximately 120 pg/kg) of purified hepatotoxin injected intraperitoneally in rats cause histologic lesions that be- gin 20 to 30 minutes post-admini~tration.~ Initially, there is disassociation and rounding of centrilobular hepatocytes rapidly followed by severe degeneration and necrosis. By 60 minutes, these changes have progressed to midzonal and periportal regions. In addition, there is a breakdown of sinusoidal endothelium and a loss of central veins with severe centrilobular hemorrhage. At this time, intact hepatocytes are seen in the pulmonary vasculature. At 9 hours post-dosing the renal cortical capillaries contain large amounts of eosinophilic, finely fibrillar material. Death occurs 20 to 32 hours post-admini~tration.~ Limited studies have characterized the ultrastructural effects of similar pentapeptide or hexapeptide hepatotoxins from M. aeruginosa. Scanning electron microscopy (SEM) of rat hepatocyte suspensions showed dose-dependent deformation and blebbing within 5 minutes of toxin administration. I 2 3 l 3 Sequential SEM and transmission electron microscopy (TEM) of mouse liver following intraperitoneal administration of an aqueous M. aeruginosa extract showed progressive breakdown of sinusoidal endothelium, disappearance of the space of Disse, damage to hepatocyte membranes, and necrotic changes in hepatocyte cy- topla~m.~ This is the only report in which lesions began periportally and progressed centrilobularly. In sheep given aqueous M. aeruginosa extracts, TEM 20 hours post-dosing showed centrilobular hepatic necrosis with aggregation of the endoplasmic reticulum, peripheral displacement of cellular organelles, and vacuolation of severely affected cells.9 Most recently, a sequential TEM study using purified MCLR in mice, at doses of 10 and 100 pg/kg intraperitoneally, described hepatic changes consisting of vesiculation and degranulation of the 9

10 Hooser et al. Fig. 1. Normal hepatocytes and endothelium (arrows) lining sinusoid(s). Hepatocyte microvilli are present in the space of Disse (D). Fig. 2. Normal hepatocytes, bile canaliculus (B) with adjacent tight junctions. rough endoplasmic reticulum, mitochondria1 swelling, and, at 100 pg/kg, an increase in intracytoplasmic membranous whorls.' There are marked differences in the response of mice and rats to MCLR. In mice, massive, centrilobular to midzonal hemorrhage and death occur within 60 to 90 minutes. In rats, although hepatic necrosis and hemorrhage occur within 60 minutes, the hemorrhage is not as severe as in mice, and rats survive 20 to 32 The purposes of this study were to 1) characterize the ultrastructural hepatic, pulmonary, and renal changes in rats following administration of MCLR; 2) correlate these changes with previously described light microscopic changes to determine the sequence of lesion development in rats; and 3) identify the subcellular site(s) of MCLR toxicity. Materials and Methods Male, 175-200 g Sprague-Dawley rats were obtained from Harlan Sprague-Dawley, Inc., Indianapolis, given commercial rat chow and water ad libitum, and kept on a 12-hour light/l2-hour dark cycle. They were allowed to acclimate for 2 weeks before use. Microcystin LR (MCLR) from Microcystis aeruginosa, laboratory strain 7820, was produced and purified (approximately 95% pure) in our (W. W. Carmichael) laboratory by techniques that involve methanol/butanol cell extraction, centrifugation, sephadex filtration, and purification by highperformance liquid chromatography. MCLR was dissolved in 0.9% NaCl prior to use. The rats were injected intrapentoneally with purified MCLR at a lethal dose of 160 pg/kg or with 0.9% NaCl (control). At each of the following time intervals, 5, 10, 20, 30, and 60 minutes post-dosing, two rats (for a total of ten) were anesthetized with ether (control rats were anesthetized at 60 min-

Microcystin-LR-induced Changes 11 Utes following saline administration), the thorax on each was opened, and a cannula was placed in the left ventricle. Whole body perfusion and fixation was done, using Tyrode s solution at 37 C, followed by cool 2.5% glutaraldehyde in 0.1 M isotonic cacodylate buffer and 3% sucrose (ph 7.2). Tyrode s solution and fixative were kept at a constant perfusion pressure of 120 mm Hg by use of a Cole-Palmer masterflex pump (Cole-Palmer Instrument Co., Chicago, IL). Samples of perfusion-fixed liver, lung, and kidney were minced into 1-mm cubes and placed in 2.5% glutaraldehyde with 0.1 M isotonic cacodylate buffer and 3% sucrose (ph 7.2) until subsequent processing was done (at least 7 days following perfusion). Tissue samples were washed in 0.1 M isotonic cacodylate buffer and 3% sucrose, post-fixed in 1% osmium tetroxide, rinsed, and dehydrated in a graded ethanol series of 10-1009 0. Final dehydration with 100% propylene oxide was followed by infiltration and embedding with epoxy. Thin sections of the specimens were made, mounted on copper grids, stained with uranyl acetate and lead citrate, and viewed with a JEOL 1 00-CX transmission electron microscope. Results In the liver, no ultrastructural changes were seen in the controls or at 5 minutes post-dosing in microcystin- LR (MCLR)-treated rats (Figs. 1, 2). Ten minutes after MCLR administration, there was mild widening of intercellular spaces between centrilobular hepatocytes with occasional invaginations of the plasma membrane. These changes progressed in extent and severity over time. Twenty minutes post-dosing, alterations in centrilobular hepatocyte plasma membranes were more pronounced, consisting of invagination with formation of variably sized and shaped cytoplasmic vacuoles. Blebbing and invagination of the entire hepatocyte plasma membrane and loss of microvilli along the sinusoidal face were present. In addition, hepatocyte separation was more pronounced and widespread. In areas with severe hepatocyte lesions, widening of sinusoidal endothelial fenestrae was present (Fig. 3). Alterations in other types of hepatic cells (Kupffer, Ito, bile ductular) were not seen at this time. Thirty minutes post-dosing in centrilobular regions, there was marked widening of the space of Disse and continued loss of hepatocyte sinusoidal microvilli. Blebbing and invagination of hepatocyte plasma membranes and widening of intercellular spaces between hepatocytes were more pronounced and widespread. In addition, bile canaliculi in affected areas were frequently dilated with blunting or loss of microvilli; however, tight junctions surrounding bile canaliculi remained intact. Sixty minutes post-dosing, centrilobular areas contained necrotic hepatic cells that lacked all or part of their plasma membranes but whose organelles appeared relatively normal. Free-floating, but intact, organelles (nuclei, mitochondria, and rough endoplasmic reticulum), together with erythrocytes and platelets were also seen in these areas (Fig. 4). Leukocytes were not seen. Endothelium was not recognizable in these areas. In the less severely affected midzonal and periportal regions, at 60 minutes post-dosing, there was increased severity of the lesions seen at 30 minutes, prominent hepatocytic rounding, and accumulation of erythrocytes within the space of Disse (Fig. 5). There was moderate, focal loss of the sinusoidal endothelium and occasional hepatocyte necrosis. Moderate whorling of the rough endoplasmic reticulum around normalappearing mitochondria was occasionally present (Fig. 6). Sixty minutes post-dosing, pulmonary vasculature contained cellular debris, including relatively intact cells with recognizable mitochondria, rough endoplasmic reticulum, and nuclei (Fig. 7). At 60 minutes postdosing, renal cortical glomerular and peritubular capillaries contained relatively intact cells identical to those in the pulmonary vasculature and small amounts of necrotic cellular debris, including cytoplasmic fragments and mitochondria (Fig. 8). Mitochondria and whorled rough endoplasmic reticulum that were present in the pulmonary and renal capillary debris and in intact cells were identical to mitochondria and whorled rough endoplasmic reticulum that were present in hepatocytes in the liver at 60 minutes. Pulmonary or renal damage associated with the microemboli was not apparent. Discussion The initial ultrastructural lesions in rats after microcystin LR (MCLR) administration are limited to alterations in the plasma membrane of the hepatocyte that are recognizable as early as 10 minutes post-dosing. Lesions seen in this study correlate well with the hepatocyte disassociation and rounding previously seen by light micro~copy.~ Loss of hepatocyte cell-to-cell contact, loss of sinusoidal microvilli, plasma membrane blebbing and invagination, and rounding of hepatocytes could be due to effects on the plasma membrane and/or some component of the hepatocyte cytoskeleton. Studies in our laboratory have shown marked alteration of actin filaments in primary cultures of hepatocytes exposed to MCLR.6 In cultures exposed to MCLR and stained with rhodamine-labeled phalloidin (which specifically binds to filamentous actin), hepatocyte actin filaments form thickened rays that extend into plasma membrane blebs and later form a single, dense aggregate in the cell. In contrast, cytochalasin B and phalloidin treatment of hepatocytes causes the formation of numerous, irregular aggregates of filamentous actin throughout the cells. In vivo, rats given a lethal dose of MCLR have hepatocyte actin filament aggregations

12 Hooser et al. Fig. 3. Hepatocytes and sinusoidal endothelium. Hepatocyte-hepatocyte separation, plasma membrane invagination, loss of hepatocyte microvilli in space of Disse (D), formation of intracytoplasmic vacuoles, and widening of sinusoidal endothelial fenestrae (arrowheads). Organelles appear normal; 20 minutes post-dosing. Fig. 4. Centnlobular hepatocytes and debris intermingled with erythrocytes (R). Hepatocyte nuclei and rough endoplasmic reticulum appear normal. Mitochondria appear condensed; 60 minutes post-dosing. * Fig. 5. Rounding and disassociation of hepatocytes. Erythrocytes and platelets in space of Disse (D); 60 minutes postdosing.

Fig. 6. Blebbing of hepatocyte plasma membrane and whorling of rough endoplasmic reticulum around mitochondria; 60 minutes post-dosing.

14 Hooser et al. Fig. 7. Hepatocyte debris (arrowhead) with whorled rough endoplasmic reticulum and mitochondria in pulmonary vessel; 60 minutes post-dosing. Fig. 8. Hepatocyte debris in glomerular capillary (arrowhead). Note similarity of rough endoplasmic reticulum and mitochondria to those in Figs. 5, 6, and 7; 60 minutes post-dosing. that correlate with the location and onset of light microscopic lesions.6 These findings correlate very well with the lesions seen by light and transmission electron microscopy. Hepatocyte disassociation and rounding seen with light microscopy and hepatocyte-to-hepatocyte disassociation, loss of hepatocyte microvilli, and hepatocyte plasma membrane invagination seen ultrastructurally are all compatible with structural changes in the cells caused by alterations in actin filaments. The fact that no other hepatocyte degenerative changes (nuclear, mitochondrial, or endoplasmic reticular) are seen until considerably after the plasma membrane shape changes lends credence to the hypothesis that actin filament changes precede other hepatocyte changes. In previous studies it was not clear whether the changes seen in the sinusoidal endothelium and the space of Disse were secondary to hepatocyte changes or if there was also a direct effect on sinusoidal endothelial cell^.^.'^ In our study hepatocyte changes preceded all other changes, which supports the hypothesis that MCLR primarily affects hepatocytes and that sinusoidal endothelial effects occur secondarily. Additional evidence for a direct effect on hepatocytes is that cholate, deoxycholate, bromosulfophthalein, and rifampicin provide protection to isolated rat hepatocytes exposed to MCLR.I3 This suggests that MCLR enters the cell by binding to a hepatocyte-specific, bile-acid carrier as do other low molecular weight cyclic peptides, such as phalloidin and s~matostatin.'~ These findings do not rule out, however, an additional effect on sinusoidal endothelial cells. The apparent lack of effect on the majority of hepatic cytoplasmic organelles is also of interest. Even following the loss of the plasma membrane, major hepatocyte organelles were still intact. The majority of mitochondria did not appear swollen; rough endoplasmic reticulum had not degranulated; and hepatic nuclei did not have clumping of chromatin and had not undergone pyknosis. This is in contrast to studies in mice where changes in intracellular organelles were among the primary lesions described. The presence of hepatocytes in the pulmonary vasculature, seen by light microscopy at 60 minutes, has been confirmed by transmission electron micro~copy.~ The presence of hepatic debris in renal cortical glomerular and peritubular capillaries also has been shown; however, hepatic debris in renal capillaries was observed in this study as early as 60 minutes post-dosing by electron microscopy, while it was not seen by light microscopy until 9 hours po~t-dosing.~ The release of hepatocyte microemboli from dam-

aged liver to the lungs and kidneys has only been reported following MCLR treatment. When disassociated hepatocytes and hepatic debris are released from the liver into the circulation, the larger particles such as cells are trapped in the pulmonary vasculature, but some debris passes into the systemic circulation and is deposited in renal capillaries. The accumulation of hepatic debris in pulmonary capillaries accounts for the previous report of atypical pulmonary thrombosis seen in mice.i4 No pulmonary damage was seen in rats. In contrast, it is possible that in animals with intravascular macrophages, such as sheep and pigs, phagocytosis of such particulate matter may result in release of macrophage enzymes with resultant damage to the microvasculature. Sheep given lethal doses of toxic cells of Microcystis aeruginosa had pulmonary edema, although hepatocyte microemboli were not rep~rted.~ We have shown that, in the rat, hepatic lesions resulting from the administration of lethal doses of MCLR occur rapidly and initially are confined to shape changes in the hepatocyte plasma membrane. We believe that these plasma membrane changes are due to alterations in hepatocyte actin filaments. Whether or not actin filament alterations are due to a direct interaction with MCLR or due to a secondary effect remains to be investigated. It seems likely, however, that after uptake of MCLR, alterations in hepatocyte actin filaments cause cell rounding and hepatocyte disassociation that leads to breakdown of sinusoidal endothelium and hepatic architecture. Disruption of the sinusoidal endothelium results in severe intrahepatic hemorrhage, and, coupled with hepatocyte disassociation, leads to the release of hepatocyte microemboli into the hepatic vein, the pulmonary vasculature, and then into the systemic circulation. Death in rats is due to hepatocyte necrosis, loss of hepatic function and eventual circulatory collapse. Acknowledgements These studies were supported in part by the United States Army Medical Research and Development Command contract DAMD 17-85-C-524 1. The views, opinions, and/or findings contained in this report are those of the authors and should not be construed as an official Department of the Army position, policy, or decision unless so designated by other documentation. The authors thank Ms. Leslie Waite and Ms. Joannie Dye for technical assistance. References 1 Dabholka AS, Carmichael WW: Ultrastructural changes in the mouse liver induced by hepatotoxin from the fresh- Request reprints from Dr. W. M. Haschek-Hock, Department Lincoln, Urbana, IL 61801 (USA). Microcystin-LR -induced Changes 15 water cyanobacterium Microcystis aeruginosa strain 7820. Toxicon 25:285-292, 1987 2 Elleman TC, Falconer IR, Jackson ARB, Runnegar MT: Isolation, characterization, and pathology of the toxin from a Microcystis aeruginosa (= Anacystis cyanea) bloom. Aust J Biol Sci 31:209-218, 1978 3 Falconer IR, Beresford A, Runnegar MT: Evidence of liver damage by toxin from a bloom of the blue-green alga, Microcystis aeruginosa. Med J Aust 1:5 11-5 14, 1983 4 Falconer IR, Jackson ARB, Langley J, Runnegar MT: Liver pathology in mice in poisoning by the blue-green alga, Microcystis aeruginosa. Aust J Biol Sci 34: 179-1 87, 1981 5 Galey FG, Beasley VR, Carmichael WW, Kleppe G, Hooser SB, Haschek WM: Blue-green algae (Microcystis aeruginosa) hepatotoxicosis in a herd of dairy cows. Am J Vet Res 48:1415-1420, 1987 6 Hooser SB, Beasley VR, Kuhlenschmidt MS, Haschek WM: Microcystin-LR induces morphologic and cytoskeletal hepatocyte changes in vitro and in vivo. The Toxicologist 9( 11): 127, 1989 7 Hooser SB, Beasley VR, Lovell RA, Carmichael WW, Haschek WM: Toxicity of microcystin-lr, a cyclic hepatopeptide hepatotoxin from Microcystis aeruginosa to rats and mice. Vet Pathol 26:246-252, 1989 8 Jackson ARB, McInnes A, Falconer IR, Runnegar MT: Toxicity for sheep of the blue-green alga, Microcystis aeruginosa. Toxicon 3(Suppl): 191-194, 1983 9 Jackson ARB, McInnes A, Falconer lr, Runnegar MT: Clinical and pathological changes in sheep experimentally poisoned by the blue-green alga, Microcystis aeruginosa. Vet Pathol 21: 102-1 13, 1984 10 Konst H, McKercher PD, Gorham PR, Robertson A, Howell J: Symptoms and pathology produced by toxic Microcystis aeruginosa NRC- 1 in laboratory and domestic animals. Can J Comp Med 29:221-228, 1965 11 Krishnamurthy T, Carmichael WW, Sarver EW: Toxic peptides from freshwater cyanobacteria (blue-green algae). I. Isolation, purification, and characterization of peptides from Microcystis aeruginosa and Anabaenaflosaquae. Toxicon 24:865-869, 1986 12 Runnegar MTC, Falconer IR: The in vivo and in vitro biological effects of the peptide hepatotoxin from the blue-green alga, Microcystis aeruginosa. S Afr J Sci 78: 363-366, 1982 13 Runnegar MT, Falconer IR, Silver J: Deformation of isolated rat hepatocytes by a peptide hepatotoxin from the blue-green alga, Microcystis aeruginosa. Naunyn Schmiedebergs Arch Pharmacol317:268-272, 1981 14 Slatkin DN, Stoner RD, Adams WH: Atypical pulmonary thrombosis caused by a toxic cyanobacterial peptide. Science 220:1383-1385, 1983 15 Ziegler K, Frimmer M: Molecular aspects of cytoprotection by modified somatostatins. Klin Wochenschr 64: 87-89, 1986 16 Zin LL, Edwards WC: Toxicity of blue-green algae in livestock. Bovine Pract 14: 151-153, 1979 of Veterinary Pathobiology, University of Illinois, 200 1 South