Loss of expression of protectin (CD59) is associated with complement membrane attack complex deposition in myocardial infarction.

BACKGROUND: Protectin (CD59) is a recently discovered inhibitor of the complement membrane attack complex (MAC). In the present study we investigated expression of protectin in human heart and examined the relationship between MAC deposition and protectin in myocardial infarction. EXPERIMENTAL DESIGN: Myocardial tissue specimens were obtained at autopsy from patients who had died of myocardial infarction (n = 10) or other causes (n = 5). MAC and protectin were detected by indirect immunofluorescence microscopy analysis in the heart sections by using antibodies against individual components of MAC, MAC neoantigens and protectin. Myocardial protectin was purified by affinity chromatography and compared with the previously characterized erythrocyte and urinary protectins by sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis, N-terminal amino acid sequencing, and testing its ability to bind to the terminal complement complex. The possible glycophosphoinositol-type anchorage of protectin in the heart was examined by treating myocardial sections with glycophosphoinositol-specific phospholipase C. RESULTS: Immunoblotting and immunofluorescence analysis showed expression of protectin in the sarcolemmal membranes of normal myocardium. Protectin purified from normal human heart tissue had the same molecular weight and N-terminal amino acid sequence as CD59 purified from urine. In sucrose density ultracentrifugation analysis it was observed to bind efficiently to the SC5b-8 complex. In normal myocardium the expression of CD59 was sensitive to treatment with glycophosphoinositol-specific phospholipase C. The expression of CD59 was lost or clearly diminished in infarcted lesions aged 1-14 days. Loss of CD59 expression was accompanied by concomitant deposition of the MAC within the CD59-negative lesions. In border areas between an infarcted lesion and normal tissue, CD59 often appeared in small vesicles, suggesting shedding as a possible mechanism for its removal. CONCLUSIONS: Glycophosphoinositol-anchored CD59 is expressed in the sarcolemmal membranes of normal heart but lost from infarcted myocardium. Acquired loss of resistance to autologous complement and subsequent complement attack may thus be involved in the pathophysiology of myocardial infarction.     

Regulation of complement membrane attack complex formation in myocardial infarction.

Recent studies have suggested that the complement (C) system is involved in the development of tissue injury of myocardial infarction. As it is not known why the strictly controlled C system starts to react against autologous heart tissue, we have analyzed the expression of various membrane regulators of C (CR1, DAF, MCP, CD59, C8 binding protein) and the pattern of deposition of C components and plasma C regulators (C4b binding protein and vitronectin) in normal (n = 7) and infarcted (n = 13) human myocardium. In the infarcted myocardium deposits of the C membrane attack complex (MAC) were observed by immunofluorescence microscopy, and lesions resembling the transmembrane channels of MAC were detected by transmission electron microscopy. CD59 and C8 binding protein were strongly expressed by muscle cells of normal myocardial tissue. Little or no CR1, MCP, and DAF was observed on these cells. The assembly of MAC was accompanied by the deposition of vitronectin (S-protein) and C4b binding protein in the infarcted areas of myocardium. In accordance with our earlier results the expression of CD59 but not of C8 binding protein was clearly diminished in the lesions. The results show that C8 binding protein, vitronectin, and C4b binding protein do not prevent complement attack against the infarcted myocardium but rather become codeposited with the MAC. Ischemia-induced transformation of nonviable cells into complement activators, acquired loss of resistance to the MAC by shedding of CD59, and recruitment of multifunctional serum proteins by MAC could thus constitute a general process aimed at the clearance of injured tissue.     

Co-deposition of clusterin with the complement membrane attack complex in myocardial infarction.

Clusterin is a multi-functional plasma glycoprotein that has been shown to inhibit formation of the complement membrane attack complex (MAC) by preventing the association of terminal complement complexes with target cell membranes. Recent studies have suggested that complement activation is involved in the development of tissue injury of myocardial infarction. In this study we observed that clusterin is selectively deposited in the infarcted areas of human myocardium. Clusterin deposits were observed in the heart tissue of 10 patients whose infarcted lesions were 8 hr to 14 days old, but not in patients who died from other causes. Clusterin co-localized with the MAC on the surface of damaged cardiomyocytes. In normal myocardium only endothelial lining of blood vessels occasionally stained positive for clusterin. The 80,000 MW clusterin was also detected by Western blot analysis in extracts of myocardial infarction lesions, but only faintly in extracts of normal heart. As clusterin has apparently failed in protecting myocardium against complement-mediated cell injury its main role might be to participate in the clearance of damaged and necrotic tissue together with the MAC.     

Evaluation of the membrane attack complex of complement for the detection of a recent myocardial infarction in man

The diagnosis of an acute myocardial infarction (MI) can be cumbersome for pathologists. Even with a positive nitroblue tetrazolium (NBT) reaction, haematoxylin and eosin (H&E) evaluation of the myocardial tissue can remain inconclusive. Early signs presumed diagnostic for myocardial infarction, such as hypereosinophilia, waviness, and contraction band necrosis, have to be considered non-specific and are probably reversible signs of ischaemia. Several studies implicate the complement system, and especially complement factor C9, as part of the membrane attack factor (MAC), in cardiomyocyte damage during MI. In a post-mortem study on well-documented cardiological autopsies, we evaluated the use of complement factor C9 immunostaining as a marker for the detection of very recent MI. Forty-three tissue samples from 40 patients were obtained from the left ventricular free wall only, a region that can be specifically attributed to one corresponding coronary artery. As some patients presented with MIs of various stages in that perfusion area, in total 57 observations were possible. C9 immunostaining specifically detected irreversibly damaged (=infarcted) cardiomyocytes, as is implied by the lytic activity of C9/MAC binding to cell membranes. Most interesting was the group of clinically suspected, NBT-positive MIs resulting from very recent myocardial ischaemia. In this population, where H&E evaluation by (cardio-) experienced pathologists was not conclusive, C9 immunostaining clearly pointed towards myocardial infarction in 47% of the cases. In conclusion, C9 immunostaining, routinely practicable in the pathology laboratory, has an additional value in discriminating between reversible ischaemia and infarcted cardiomyocytes in very early MIs.     

IgM colocalises with complement and C reactive protein in infarcted human myocardium

AIMS: Reperfusion of ischaemic myocardium after acute myocardial infarction (AMI) can induce ischaemia/reperfusion (I/R) injury, as a result of local activation of the complement system. C reactive protein (CRP) is involved in this activation. This study analysed the potential role of IgM in complement activation in the infarcted human myocardium. METHODS: Immunochemical analysis was carried out on heart specimens from 59 patients who died from AMI. Serial slides of frozen tissue from the infarction site were stained for IgM, complement factors C3d and C5b-9 (membrane attack complex), and CRP. RESULTS: IgM deposits were found on the plasma membrane, cross striations, and in the cytoplasm of jeopardised cardiomyocytes in infarcts of one to five days duration. IgM depositions were remarkably similar to those of CRP and both complement factors. The relative staining intensities of IgM and CRP varied greatly among patients. CONCLUSIONS: Similar to CRP, IgM targets complement locally to jeopardised cardiomyocytes in the human heart after AMI. Localisation patterns and relative staining intensities suggest that IgM and CRP recognise similar epitopes in the ischaemic heart, but that the relative contribution of each protein to complement activation in the ischaemic myocardium differs among patients.     

Time course of complement activation and inhibitor expression after ischemic injury of rat myocardium.

Activation of the complement (C) system has been documented in both experimental and clinical studies of myocardial infarction, but the exact time course and mechanisms leading to C activation have remained unclear. Our earlier postmortem study on human beings showed that formation of the membrane attack complex (MAC) of C was associated with loss of CD59 (protectin), an important sarcolemmal regulator of MAC, from the infarcted area. The recent discovery of a rat analogue of CD59 has now allowed the first experimental evaluation of the temporal and spatial relationship between C component deposition and loss of CD59 in acute myocardial infarction (AMI). After ligating the left coronary artery in rats the earliest sign of C activation, focal deposition of C3, was observed at 2 hours. Deposition of the early (C1, C3) and late pathway (C8, C9) components in the AMI lesions occurred at 3 hours. Glycophosphoinositol-anchored rat CD59 was expressed in the sarcolemmal membranes of normal cardiomyocytes. In Western blot analysis extracts of normal rat heart CD59 appeared as a band of 21 kd of molecular weight under nonreducing conditions. Loss of CD59 in the AMI lesions was observed in association with deposits of MAC from day one onward. Our results show that C activation universally accompanies AMI in vivo. It is initiated within 2 hours after coronary artery obstruction via deposition of C3, which may be due to generation of the alternative pathway C3 convertase in the ischemic area. Deposition of C1 and late C components also starts during the early hours (2 to 4 hours) after ischemia. Subsequent loss of the protective CD59 antigen may initiate postinjury clearance of the irreversibly damaged tissue.     

Detection of a soluble form of the complement membrane attack complex inhibitor CD59 in plasma after acute myocardial infarction

Activation of the complement system has been documented in both experimental and clinical studies of acute myocardial infarction (AMI). Our earlier immunohistochemical studies have shown that the deposition of the membrane attack complex (MAC) of complement is associated with the loss of protectin (CD59), a glycosyl-phosphatidylinositol (GPI)-anchored sarcolemmal regulator of MAC, from the human and rat infarcted myocardium. In this study we detected, using an enzyme immunoassay (EIA), CD59 in the plasma of AMI patients at a concentration of 23.0+/-8.4 ng/ml (mean +/- SD; n = 17) at 4 h and 27.3+/-11.8 ng/ml (n = 24) at 24 h after AMI. Both values were significantly higher than in healthy controls (7.8+/-6.4 ng/ml; n = 20; P<0.001). The amount of CD59 correlated with the level of soluble terminal complement complexes (SC5b-9; r = 0.84; P<0.01) in the plasmas of AMI patients. Our results suggest that myocardial damage leads to release of CD59 from the sarcolemmal cell membranes during AMI.     

C-reactive protein activates complement in infarcted human myocardium

Circulating levels of C-reactive protein (CRP) constitute a cardiovascular risk marker. Immunohistochemical studies have revealed co-localization of CRP and activated complement in human infarcted myocardium suggesting CRP to enhance inflammation in ischemic myocardium by inducing local complement activation. The aim was to establish whether CRP activates complement in infarcted human myocardium and to assess the relationship between this activation and the duration of infarction. Myocardial tissue samples from 56 patients that had died from acute myocardial infarction were evaluated. Specimens were taken from infarcted as well as noninfarcted sites of the heart. CRP-mediated complement activation was assessed by immunohistochemistry and by measuring levels of complement, CRP, and CRP-complement complexes, specific markers for CRP-mediated activation, in homogenates of the heart. Infarctions of 12 hours to 5 days had significantly more extensive depositions of complement and CRP and contained significantly more CRP, activated complement, and CRP-complement complexes than infarctions that were less than 12 hours old. Levels of CRP complexes correlated significantly with CRP and complement concentrations in the infarctions, as well as with the extent of complement and CRP depositions as measured via immunohistochemistry. Specific activation products of CRP-mediated activation of complement are increased in infarcts of more than 12 hours in duration and correlate with the extent of complement depositions. Hence, CRP seems to enhance local inflammatory reactions ensuing in human myocardial infarcts of more than 12 hours duration.     

Localization of 20-kD homologous restriction factor (HRF20) in diseased human glomeruli. An immunofluorescence study.

The 20-kD homologous restriction factor (HRF20), which is identical to CD59, is a membrane-associated protein which inhibits the reaction of C9 to form membrane attack complex (MAC) of homologous complements. In various human glomerular diseases deposition of complement components is frequently seen and MAC is reported to associate with immune deposits. Using a specific monoclonal antibody, 1F5, against HRF20, we attempted to study the localization of HRF20 in human glomerulonephritides and to compare the localization of HRF20 with those of immune deposits and MAC. The frozen sections of kidney specimens were fixed in acetone at room temperature before staining. In normal kidneys and kidney specimens from the patients with minimal change nephrotic syndrome, membranous nephropathy, and IgA nephropathy, HRF20 was strongly localized in the peritubular capillaries and along Bowman's capsules. A weaker but well-defined staining was obtained in the mesangial area and faint staining was seen along the glomerular capillary walls. In contrast, glomerular capillary walls were rather strongly stained in the cases with diffuse lupus nephritis which had subendothelial dense deposits. These data suggest that HRF20 (CD59) is present in the human glomeruli and its expression is enhanced under certain conditions such as lupus nephritis.     

Repair of infarcted myocardium using mesenchymal stem cell seeded small intestinal submucosa in rabbits

Myocardial infarction (MI) remains a common and deadly disease. Using tissue-engineered cardiac grafts to repair infarcted myocrdium is considered to be a therapeutic approach. This study tested the feasibility of using MSCs-seeded SIS to repair chronic myocardial infarction in a rabbit model. MI in rabbits was created by ligation of the left anterior descending artery. BrdU-labeled mesenchymal stem cells (MSCs) were seeded on the small intestinal submucosa and cultured for 5–7 days prior to implantation. Four weeks after myocardial infarction, cardiac grafts were implanted onto the epicardial surface of infarcted myocardium. Four weeks after implantation of the membranes, a serial of tests including echocardiography, hemodynamics, histology and immunohistochemistry were undertaken to evaluate the effect of the implanted grafts on recovery of the infarcted myocardium. It was shown that left ventricular contractile function and dimension, the capillary density of the infarcted region, and myocardial pathological changes were significantly improved in rabbits implanted either SIS or MSCs-seeded SIS. But the MSCs-seeded SIS was more effective. Immunofluorescence staining demonstrated the migration of Brdu-labeled MSCs from the membrane into the infarcted area and their differentiation to cardiomyocytes and smooth muscle cells. Taken together, these results suggest that MSCs-seeded SIS can be used to repair chronic myocardial infarction, which enhances myocardial regeneration.     

Postmortem detection of inapparent myocardial infarction

Two methods of detecting early inapparent myocardial infarcts have been studied and their value in diagnostic practice compared.The better method proved to be the determination of the potassium to sodium ratio (ionic ratio) which falls in infarcted tissue within minutes of the onset of anoxia. The second method was nitro blue tetrazolium staining of gross sections of myocardium which revealed any infarct older than three and a half hours. As staining is dependent upon enzyme activity, the latter method is disturbed by autolysis. It was shown, on the other hand, that the ionic ratio (K(+)/Na(+)) was not affected by autolysis and was therefore well suited to forensic practice.Sixteen non-infarcted control hearts, plus the nine from cases of sudden death due to causes other than myocardial infarction, all yielded high ionic ratios (K(+)/Na(+)), average 1.4, and stained normally with tetrazolium (the normal controls). Positive control was provided by 20 histologically proven infarcts of which the ionic ratios (K(+)/Na(+)) were all low (average 0.7). Histochemical staining with tetrazolium delineated infarcted areas in each case.In a series of 29 sudden deaths, a cause of death other than myocardial infarction was found at necropsy in nine, mentioned above as normal controls. The remaining 20 hearts were not infarcted histologically, but were shown to be infarcted by examination of the ionic ratios (K(+)/Na(+)). These ratios were low (average 0.8) including three borderline ratios. Confirmatory evidence of infarction included nitro blue tetrazolium staining which revealed infarcts in 10 of the 20 cases, and clinical and necropsy observations.The ionic ratio (K(+)/Na(+)) decreases as the age of the infarct increases for at least 24 hours. Thereafter as healing proceeds, the ratio gradually reverts to normal. Thus, previous infarction and replacement fibrosis do not significantly alter the ionic ratio (K(+)/Na(+)). Nor is it changed by left ventricular hypertrophy, the presence of congestive cardiac failure, or digitalis therapy.It is suggested that macroscopic tetrazolium staining is a useful screening test for early inapparent myocardial infarcts. In cases where no infarct is delineated with that method estimation of the ionic ratio (K(+)/Na(+)) should be carried out on myocardium removed from standard areas on the anterior and posterior left ventricular walls.     

Rapid transformant selection by human complement using HRF20 (CD59) cDNA as a selection marker

The 20 kDa homologous restriction factor (HRF20, CD59) prevents formation of membrane attack complexes of human complement. Transfection of the cDNA of HRF20 to heterologous cells, which are sensitive to human complement, thereby renders these cells resistant to human complement attack. This property can then be used as a selection marker of transfection. In addition, the procedure is short and can be performed within 8 h. Since only 1 h is required for incubation with complement, the damage to the transfected cells is limited compared with damage resulting from selection for drug resistance which requires 10 days or more of incubation in the presence of harmful agents.     

Soluble complement receptor 1 protects the peripheral nerve from early axon loss after injury

Complement activation is a crucial early event in Wallerian degeneration. In this study we show that treatment of rats with soluble complement receptor 1 (sCR1), an inhibitor of all complement pathways, blocked both systemic and local complement activation after crush injury of the sciatic nerve. Deposition of membrane attack complex (MAC) in the nerve was inhibited, the nerve was protected from axonal and myelin breakdown at 3 days after injury, and macrophage infiltration and activation was strongly reduced. We show that both classical and alternative complement pathways are activated after acute nerve trauma. Inhibition of the classical pathway by C1 inhibitor (Cetor) diminished, but did not completely block, MAC deposition in the injured nerve, blocked myelin breakdown, inhibited macrophage infiltration, and prevented macrophage activation at 3 days after injury. However, in contrast to sCR1 treatment, early signs of axonal degradation were visible in the nerve, linking MAC deposition to axonal damage. We conclude that sCR1 protects the nerve from early axon loss after injury and propose complement inhibition as a potential therapy for the treatment of diseases in which axon loss is the main cause of disabilities.     

Autologous cardiomyocyte transplantation using a biodegradable polymer scaffold

BACKGROUND: To achieve a more reliable way of transplanting cardiomyocytes, we conducted an autologous cardiomyocyte transplantation using a biodegradable scaffold, instead of a syringe injection, as a vehicle for transporting cells in an ovine myocardial infarction model. MATERIALS AND METHODS: A myocardial infarction was created in sheep using sequential ligation of the homonymous artery and its diagonal branch. Autologous cardiomyocytes from the right ventricular infundibulum were cultured and seeded onto a biodegradable polymer scaffold. Three months after creating myocardial infarction, the two animals were re-anesthetized and cardiomyocyte-seeded scaffolds were implanted in the infarcted area. The animals were kept alive for a further month, and then sacrificed for postmortem heart examinations. Light microscopic analysis and an immunohistochemical study for myoglobin were performed. RESULTS: On postmortem gross examinations, the polymer scaffolds were visible in the background of well-demarcated thin-walled anteroseptal myocardial infarcts. Microscopic analysis showed abundant myoglobin-stained cells between the fiber strands of the polymer scaffolds. However, there is a possibility that some of these cells might have been giant cells reacting to foreign material. CONCLUSION: The transplantation of cultured autologous cardiomyocytes into an infarct region using a biodegradable scaffold instead of syringe injection provides another promising option for cardiomyocyte transplantation, which warrants further study.     

Characterization of homologous restriction factor (HRF20) in human skin and leucocytes.

Homologous restriction factor with a molecular weight of 20 kD (HRF20) is a membrane protein that inhibits assembly of the membrane attack complex of homologous complement. Distribution of HRF20 in normal human skin was studied. The plasma membrane of keratinocytes was stained, and the intensity of the staining pattern was higher in the basal cell layer than in the granular layer. Endothelial cells of blood vessels in the dermis were also stained. The molecular weight of HRF20 on erythrocytes and epidermis is 16 kD, determined by Western blot analysis. Those of polymorphonuclear cells and lymphocytes appeared as two bands, a major band of 20 kD and a minor band of 16 kD. Susceptibility of HRF20 to phosphatidylinositol-specific phospholipase C (PIPLC) was examined. After PIPLC treatment of the sections, HRF20 was not detected on the epidermis and was very slightly expressed on the blood vessels. These results indicate that HRF20 attaches to keratinocytes and blood vessels via phosphatidylinositol, regulating the formation of membrane attack complexes of homologous complement on the cell membrane.     

CR2-mediated activation of the complement alternative pathway results in formation of membrane attack complexes on human B lymphocytes

Normal human B lymphocytes activate the alternative pathway of complement via complement receptor type 2 (CR2, CD21), that binds hydrolysed C3 (iC3) and thereby promotes the formation of a membrane-bound C3 convertase. We have investigated whether this might lead to the generation of a C5 convertase and consequent formation of membrane attack complexes (MAC). Deposition of C3 fragments and MAC was assessed on human peripheral B lymphocytes in the presence of 30% autologous serum containing 4.4 mM MgCl2/20 mM EGTA, which abrogates the classical pathway of complement without affecting the alternative pathway. Blockade of the CR2 ligand-binding site with the monoclonal antibody FE8 resulted in 56 +/- 13% and 71 +/- 9% inhibition of the C3-fragment and MAC deposition, respectively, whereas the monoclonal antibody HB135, directed against an irrelevant CR2 epitope, had no effect. Blockade of the CR1 binding site with the monoclonal antibody 3D9 also resulted in a minor reduction in MAC deposition, while FE8 and 3D9, in combination, markedly reduced deposition of both C3 fragments (91 +/- 5%) and C9 (95 +/- 3%). The kinetics of C3-fragment and MAC deposition, as well as the dependence of both processes on CR2, indicate that MAC formation is a consequence of alternative pathway activation.     

Histopathology of Early Myocardial Infarcts: A New Approach

The histopathology of human myocardial infarcts is reviewed in a series of 46 cases ranging from sudden death to a clinical age of 3 days. A set of histopathologic features is described whereby the diagnosis of acute myocardial infarction can be made, even in cases of sudden death, on routine sections and even after considerable autolysis. This is primarily a stretching and waviness of the myocardial fibers, especially at the border of the infarcted area. Its mechanism is probably twofold: the rythmical pull exerted by the normal myocardium against the infarcted paralyzed area and the outward bulging of this area at every systole. On the basis of human material alone, it may be inferred that this pattern develops very rapidly: surely less than 1 hour and perhaps a few minutes after the local circulation has failed.     

Myocardial structural response following administration of a fatty diet to rats.

Myocardial infarctions in rats were induced by dietary means. Feeding a high-fat diet for a minimum of eight weeks was required for the onset of ventricular infarction; however, distinctive myocardial abnormalities preceding the terminal event were clearly recognizable.Within four weeks of exposure to the fatty diet, the morphologic hallmarks of the initial myocardial response consisted of (1) a gradual and progressively increasing deposition of lipids within myocardiocytes, and (2) mobilization of macrophages.With continued administration of the diet, the most conspicuous morphologic manifestation of the dietary effect on the heart consisted of an extensive invasion of macrophages into the interstitial spaces within the muscle mass.The terminal event, i.e. infarcted areas of varying extent and severity, involving diffusely ventricular sites, was only rarely associated with the unequivocal presence of mural thrombi. Significantly, however, the obstruction of microvascular channels with widely-distended, lipid-engorged macrophages was strikingly evident. These findings indicate that dietary fats stimulate the reticulo-endothelial system, and that the induction of myocardial infarcts is a consequence of excessive macrophagic activity.     

Immunohistochemical study of the membrane attack complex of complement in IgA nephropathy

Summary The localization of the membrane attack complex of complement (MAC) was examined in the normal human kidneys and in biopsy specimens from patients with primary IgA nephropathy by immunofluorescent and immunoelectron microscopies. Immunofluorescent staining for MAC was significantly more intense than in the normal kidneys, and was observed in the mesangium and occasionally along the glomerular capillary walls of 22 of 30 patients with IgA nephropathy. By dualstaining, the MAC deposits were generally concordant with the deposits of IgA, C3, C5 and C9, or of IgG, when present. C1_q or C4 was infrequently observed in the glomeruli. Immunoelectron microscopy revealed various staining patterns of glomerular MAC deposition; homogeneous fine-granular staining beneath the glomerular basement membrane (GBM) in the paramesangial zone, patchy staining within the mesangial electron dense deposits (EDD), and ring-shaped or ribbon-like staining, associated with the striated membrane structures (SMS), in the matrix of the mesangium, GBM and tubular basement membrane (TBM). This study suggests that the terminal complement system is activated, mainly by an alternative complement pathway mechanism, in the mesangium of IgA nephropathy, and is associated with the paramesangial lesion and EDD. MAC deposition in glomerular SMS may also result from in situ activation rather than trapping from the circulation. There was little correlation between glomerular MAC deposition and proteinuria or renal histology of patients with IgA nephropathy.     

Species-specific restriction of complement by HRF20 (CD59) generated by cDNA transfection.

The 20-kDa homologous restriction factor (HRF20, CD59) is a phosphatidyl inositol-anchored membrane glycoprotein that inhibits the formation of human complement membrane attack complexes. The cDNA of HRF20 was transfected into Chinese hamster ovary (CHO) cells resulting in expression of human HRF20 protein on the cell surface anchored via glycosylphosphatidyl inositol. The transfected CHO cells were resistant to human complement-mediated cell killing. However, the cells remained sensitive to rat and guinea pig complement. Therefore, species specificity between HRF20 and complement is maintained in HRF20 generated on the CHO cells following transfection with HRF20 cDNA.