Atypical hemolytic uremic syndrome (aHUS) is a thrombotic microangiopathy often caused by mutations in complement factor H (CFH), the main regulator of alternative complement pathway. Because CFH is produced mainly by the liver, combined liver–kidney transplantation is a reasonable option in treatment of patients with severe aHUS. We studied complement activation by monitoring activation markers during liver transplantation in two aHUS patients treated extensively with plasma exchange and nine other liver transplantation patients. After the reperfusion, a clear increase in all the activation markers except C4d was observed indicating that the activation occurs mainly through the alternative pathway. Concentration of SC5b-9 was higher in the hepatic than the portal vein indicating complement activation in the graft. Preoperatively and early during the operation, the aHUS patients showed highest C3d concentrations but otherwise their activation markers were similar to the other patients. In the other patients, correlation was found between perioperative SC5b-9 concentration and postoperative alanine aminotransferase and histological changes. This study explains why supply of normal CFH by extensive plasma exchange is beneficial before combined liver–kidney transplantation of aHUS patients. Also the results suggest that perioperative inhibition of the terminal complement cascade might be beneficial if enhanced complement activation is expected.
During the last decade, liver transplantation has become an attractive treatment option for a number of rare diseases. Some of these diseases are genetic disorders where mutations are found in proteins produced mainly by the liver. Liver transplantation can be used to permanently replace those genetically defective plasma proteins. We and others have previously described utilization of combined liver–kidney transplantation in one familial rare disease, atypical hemolytic uremic syndrome (aHUS), that is usually caused by a single missense mutation in one of the plasma proteins produced nearly exclusively by the liver.
Hemolytic uremic syndrome (HUS) is a thrombotic microangiopathy characterized by hemolytic anemia, microvascular thrombosis and kidney failure. There are two forms of HUS, a typical form that can occur in epidemics and an atypical form that is sporadic or familial. Typical HUS is associated with gastroenteritis caused by cytotoxin-producing strains of Escherichia coli, is often self-limiting, and rarely leads to complications or end-stage renal failure (ESRF). In contrast, aHUS is often progressive with recurrent attacks and leads often to ESRF. The pathogenesis of aHUS is closely linked to complement dysregulation since mutations in complement components or regulators are found in up to 60% of the cases (1,2). Most commonly, mutations are found in complement factor H (CFH) (3–6).
Complement is an innate immunity defense mechanism and is composed of a set of proteins in plasma and on cell surfaces. The purpose of this system is to directly destroy invading microbes or to assist in phagocytosis of different kinds of targets. Complement can be activated via three initiation pathways. The classical pathway is activated by recognition of surface-bound immunoglobulins with the component C1q. The lectin pathway is activated by mannose-binding lectin or ficolins. Only these pathways lead to generation of C4 cleavage products, some of which can be detected in plasma (e.g. C4d). The third pathway, the alternative pathway, is activated on any surface that is not specifically protected. It is the only pathway that leads to generation of plasma Bb fragment in addition to generation of several activation markers common for all the initiation pathways (C3a, C3d and iC3b). Finally, all the pathways lead to a common terminal cascade and its activation can be detected by appearance of C5b-9 complexes on the target surface or complexed with S-protein in plasma (SC5b-9). CFH is an essential complement regulator that inhibits the C3 activation steps in both plasma and on surfaces of self-cells while allowing the activation to occur on foreign targets. Function of CFH is impaired by mutations in aHUS (7–9). Loss-of-function mutations of two other complement regulators, plasma factor I and cell surface protein CD46, and gain-of-function mutations in complement components C3 and factor B have also been reported to be mutated in some aHUS patients (10–15). In addition, aHUS-associated mutations have been described in thrombomodulin and ADAMTS-13 (16).
Currently, aHUS associated with complement dysregulation is treated either with plasmapheresis/plasma exchange (PE) or solid organ transplantation (1). Very recently, also therapeutic anti-C5 antibody has been successfully used in a few cases (17–20). Although plasmapheresis and anti-C5 therapy are needed lifelong, the transplantation procedures can be curative. Kidney transplantation, however, is successful only in cases with dysfunction of membrane-bound complement regulators, such as CD46 (21–23). Recurrence rate of aHUS associated with CFH mutations is approximately 80% (22). Since CFH is produced mainly by the liver, a combined liver–kidney transplantation is a rational treatment option in the most severe cases with, for example, involvement of the central nervous system. First attempts of the combined transplantation in aHUS were unsuccessful (24–26), but since an intensive preoperative PE therapy was implemented, several successful cases have been reported (27–30). It was recently shown that isolated liver transplantation could also be used in management of aHUS associated with CFH mutations even before development of ESRF (31).
It is known that complement is activated during reperfusion of the transplanted liver since increased plasma C3a and SC5b-9 levels and decreased C3 level are found within an hour after the reperfusion (32). Combined liver–kidney transplantation seems to be a good curative treatment option for severe aHUS caused by CFH mutations, but only after intensive PE. Therefore, the main aims of this study were to analyze which of the three complement pathways is activated during the reperfusion and to characterize possible differences in complement activation during the transplantation of aHUS versus other transplant patients. The other aims of this study were to investigate if the activation occurs in the liver graft (not in the intestine) and to evaluate if any preoperative complement activation marker level could be used to predict enhanced intraoperative complement activation.
Materials and Methods
Initially, nine adult liver transplantation patients and two aHUS patients were recruited to the study (Table 1). Written informed consent was obtained from each patient before recruitment. An additional six patients were recruited to verify the correlation between SC5b-9 and postoperative ALT. The underlying reasons for liver transplantation in these six patients were polycystic liver disease, Wilson's disease, alcoholic liver disease, cirrhosis of unknown etiology and acute liver failure (n = 2). This study was approved by the ethics committee of Helsinki University Central Hospital (DNr. 33/13/03/02/08).
The clinical details of the aHUS case 1 (aHUS-1), a 16-year-old female with CFH R1215Q mutation and severe kidney and central nervous system involvement, have been described earlier (28). The other aHUS case (aHUS-2) has not been previously described. Briefly, the patient was a 21-year-old otherwise healthy Caucasian male admitted to Helsinki University Central Hospital due to anemia and acute kidney failure in early spring 2007. He was diagnosed with aHUS and genetic analysis revealed a heterozygous R1215Q mutation in CFH but no mutations in CFI, CD46 or CFB. At the initial admission, PE therapy (administrated daily for 2 weeks) suppressed the disease activity and the renal function recovered. Thereafter, he needed two more 10-day periods of daily PEs. The principal anticoagulant was aspirin (100 mg OD) and in addition a small dose of low molecular weight heparin was given with both PE and ultrafiltrations. Daily methylpredinisolone 40 mg was used for 3 months. At 4 months, rituximab therapy (33) was applied with no success. Attempts to prolong the interval between PEs were made without success and the patient ended up again on daily PEs and ultrafiltrations. Combined liver–kidney transplantation was performed 10 months after the initial admission. The intraoperative PE therapy followed the same pattern as used with the patient aHUS-1 (8 L PE before the operation and 3 L PE immediately after liver grafting but before the kidney operation).
All grafts were retrieved from brain-dead donors and were preserved in the University of Wisconsin solution at +4°C. After hepatectomy, the supra- and infrahepatic caval anastomoses and portal anastomosis were performed; the vena cava was cross-clamped during the anhepatic period without using venovenous bypass. With the suprahepatic caval vein clamped, the graft was first flushed with 1000 mL of Ringer's solution via portal vein followed by perfusion with <400 mL of portal venous blood. Subsequently, suprahepatic and portal veins were declamped followed by declamping of infrahepatic caval vein. In one patient (Pt 1), piggyback technique was used. Finally, the hepatic arterial anastomosis was completed followed by bile duct reconstruction. Immunosuppression was achieved by cyclosporine and methylprednisolone and postoperatively mycofenolate mofetil was administered.
Systemic arterial blood samples were drawn from radial artery at different time points: after the induction but before the operation was started, before the anhepatic period, before the reperfusion, 15 min after the reperfusion, 60 min after the reperfusion and 8 or 24 h after the reperfusion. Portal vein and hepatic vein samples were obtained by puncture at three time points: during the initial reperfusion, 15 and 60 min after the reperfusion. Samples were drawn to syringes and transferred immediately to EDTA containing tubes (final concentration 5 mM) placed on ice. Plasma was separated by centrifugation after the operation and stored at –70°C until analyses were performed.
Complement activation measurements
Concentrations of C3, C3a, C3d, iC3b, Bb, C4, C4d, SC5b-9 and CFH were measured from the EDTA plasma samples. C3a, iC3b, Bb, C4d and SC5b-9 concentrations were measured with enzyme immunoassay kits according to manufacturer's instructions (Quidel, San Diego, CA, USA). C3 and C4 concentrations were measured with standard immunoturbidometry using Hitachi LTD Modular equipment (Tokyo, Japan) and Tina-quant® C3c and C4 reagents (Roche Diagnostics Corp., Basel, Switzerland). EDTA plasma was treated with 11% (w/v) of PEG 6000 at 4°C for 60 min. After centrifugation at 3000 × g, the C3dg concentration was assessed in the supernatant by nephelometry (Beckman Coulter, Brea, CA, USA). Concentration of CFH was measured using an in-house ELISA method essentially as described earlier (34) but using 1% bovine serum albumin in phoshpate buffered solution (PBS) as the assay buffer and dilution 1:10 000 of the plasma samples.
Liver biopsies were taken to 10% formalin during the donor operation, just before flushing and cooling of the organ (prereperfusion sample) and correspondingly in the end of the recipient operation, just before the closure (postreperfusion sample). After standard hematoxylin–eosin staining histological changes such as increased neutrophil count and congestion were evaluated from the pre- and postreperfusion samples without knowing the results from complement analyses.
Statistical analyses were performed using GraphPad Prism version 5 (GraphPad Software, La Jolla, CA, USA). Due to the small sample size, only nonparametric tests were used. For the comparison of protein concentrations as a function of time, we used Friedman's test followed by Dunn's test for multiple comparisons. For comparing the differences between protein concentrations in arterial, hepatic and portal blood, we used Wilcoxon signed rank test. Spearman's test was used to estimate correlations between different variables. Student's t-test was used for comparing SC5b-9 concentration and histological changes.
Concentrations of complement activation markers in plasma were analyzed from a total of 11 liver transplant patients: two aHUS patients and nine other liver transplantation patients. When compared to the reference range of healthy individuals, nearly all the markers showed elevated concentrations in all the 11 patients (Figure 1). Plasma C3d was higher in the aHUS patients than any other patient. One of the aHUS patients (aHUS2) had also the highest Bb, iC3b and SC5b-9 levels. Among the other transplantation patients, it is noteworthy that both patients with acute liver failure had highest C4d concentrations (Figure 1).
Complement activation during reperfusion of non-aHUS patients
Activation of complement during the operation was monitored from arterial blood samples taken at certain time points. Before the reperfusion, i.e. before and during the anhepatic phase, plasma concentrations of all the markers remained essentially unchanged in the non-aHUS patients (Figure 2). Immediately after the reperfusion, levels of the C3 fragments, Bb and SC5b-9 were increased and peaked at 60 min after the reperfusion in all but two patients. During the reperfusion, plasma concentrations of C3 and C4 decreased slightly without any observed peaks. Clear elevation of C4d concentration was observed during the reperfusion only in one patient (Figure 2E). This patient (Pt3, indicated with a gray line in Figure 2) had exceptionally high peaks of, for example, C3a, Bb and SC5b-9. The amount of intraoperatively given blood products did not seem to influence complement activation (Table 1). These results indicate that complement was activated during the reperfusion mainly via the alternative pathway.
Activation of complement during the reperfusion was quickly terminated in all the patients as judged from the rapid decline in all the activation markers. Concentration of the markers returned close to or below the preoperative values within 24 h from the beginning of the reperfusion (Figure 2H). Even the two patients from whom only an 8-h sample could be obtained showed a clear decrease in the activation markers (patients Pt5 and Pt8). In patient Pt5, the decrease in C3a, iC3b and SC5b-9 concentrations within 7 h (from sample A60′ to sample A8h) was 93%, 69%, and 92%, respectively. Decrease in the C3d and Bb concentrations within 7 h was 28% and 44%, respectively, indicating longer half-lives for these markers. A statistically significant positive correlation could be found between peak SC5b-9 concentration and peak postoperative alanine aminotransferase in these patients (ALT, p = 0.0083, Spearman's test). To verify the clinical correlation between the peak intraoperative SC5b-9 and postoperative ALT concentrations, we tested six more patients (n = 15) resulting in even clearer correlation (p = 0.0006; Figure 3A). To further evaluate the effect of complement activation on the liver, we studied histological changes between pre- and postreperfusion liver biopsies. The patients could be divided into two groups, one with no histological changes and the other with increased amounts of neutrophils or congestion (Figure 3B). SC5b-9 level was statistically significantly higher in the group with histological changes than the group without (p = 0.028, Student's t-test, Figure 3B).
Complement activation in aHUS patients
Concentrations of C3a, iC3b, C3d, Bb and SC5b-9 were higher in the patient aHUS-1 than the non-aHUS patients (Figure 4). In this patient, concentrations of the C3 fragments and Bb were highest before the reperfusion. Some of these markers, especially SC5b-9, showed an increase before the reperfusion but all of them decreased upon reperfusion. In plasma of the patient aHUS-2, the profile of C3d was similar to the patient aHUS-1 in the beginning of the operation, but levels of the other fragments showed a similar profile to that found in the other transplantation patients. The differences in the activation pattern in the patient aHUS-1 could not be explained by the amount or type of given blood products. Operations of both the aHUS patients were essentially uneventful and during the follow-up period of 4 and 3 years both have lived normal life with functional grafts.
CFH concentration during the operation
Concentrations of CFH were also measured during the transplantation since this protein is a central complement regulator in plasma and mutated in our two aHUS patients. Plasma CFH concentration decreased slightly at the onset of reperfusion in the non-aHUS patients. The patient aHUS-1 showed constant CFH level during the operation. The patient aHUS-2 had low CFH at the beginning of the operation (below reference range) but thereafter the concentration first increased and after the reperfusion decreased slightly staying, however, within the reference range (Figure 5). When CFH concentrations from all the transplantation patients was compared to the complement activation products, no statistically significant correlation could be observed (data not shown).
Role of transplanted liver in complement activation
To address the impact of the intestinal vascular bed or the transplanted liver in the activation of the alternative pathway in our patients, we collected samples simultaneously from radial artery, portal vein and hepatic vein in the beginning and during the reperfusion. In the beginning of the reperfusion, concentrations of C3a, iC3b, C3d, Bb and C4d were significantly lower in the hepatic vein samples when compared to the arterial or portal samples (Figure 6). Although there was an increase in the concentrations of all the activation markers during the reperfusion, there were no significant differences between the arterial, portal and hepatic plasma samples except for higher concentration of SC5b-9 in hepatic plasma at 15 min after the reperfusion (p = 0.031 hepatic vs. arterial, p = 0.016 hepatic vs. portal) indicating that activation occurred in the liver (Figure 6F).
We have studied activation of complement during liver transplantation of two patients with aHUS and nine other patients and verified plasma SC5b-9 results with six additional patients. We detected clear complement activation in all the patients immediately after the reperfusion, as expected, demonstrated that the activation was mediated mainly by the alternative pathway and occurred mainly in the liver graft, found a correlation between plasma SC5b-9 and level of postoperative ALT, showed that higher plasma SC5b-9 is associated with histological changes in postreperfusion liver biopsies, and finally could show that the activation profile after reperfusion is very similar in extensively PE treated aHUS patients and the other liver transplant patients.
Complement is known to get activated during reperfusion of liver graft and our results are concordant with this (32). In ischemia-reperfusion injury of liver and other organs, central involvement of complement is obvious (35–38). In the previous studies, elevated C3a and SC5b-9 have been found in plasma and increased amounts of some complement activation products (SC5b-9 and C4d) have been detected in immunostainings of postreperfusion liver biopsies (39,40). In addition to the terminal complement activation marker SC5b-9, we analyzed early activation fragments with longer half-lives than C3a – iC3b, C3d and Bb. We also analyzed plasma level of C4d that describes activation via the classical and lectin pathways. All the markers except C4d were increased during the reperfusion and the highest rise was seen with SC5b-9, C3a and Bb (Figure 1). During reperfusion, the peak in complement activation occurred mainly via the alternative pathway as judged by an increase in plasma concentration of the alternative pathway activation markers but not in the concentration of the classical/lectin pathway activation marker C4d (Figure 2). The reason why we were unable to detect plasma C4d, although it has previously been detected in immunostainings of postreperfusion liver biopsies, is obscure (40). All the markers except C3d and Bb returned to baseline in all the patients within 24 h after the operation (Figure 2). This is expected since T1/2 of at least C3d in blood is longer than that of C3a and C5b-9 (41,42). These results indicate that complement gets activated only for a short period immediately after the reperfusion. Analysis of concentrations of the markers in portal vein and hepatic vein samples revealed that at least some of the activation occurs in the liver (Figure 6). However, interpretation of small differences can be compromised by the relatively small number of patients.
Even though the complement activation is short-term during liver transplantation, it can cause problems intraoperatively and be harmful to the transplanted liver. It has previously been shown that increased plasma C3a and SC5b-9 are associated with hemodynamic instability during reperfusion (43,44). Previous studies have also shown that increased C5b-9 staining in liver biopsy samples is associated with higher postoperative ALT and lower factor V levels in plasma (39) and that increased C4d staining in liver biopsy is associated with higher postoperative aspartate aminotransferase levels and poor initial function of the liver (40). These reports are concordant with our results on the positive correlation between peak postreperfusion plasma SC5b-9 concentration and both postoperative ALT and increased amounts of neutrophils or congestion in postreperfusion liver biopsies. The patient Pt5 showed the highest level of complement activation markers and was the only patient receiving a liver graft with moderate macrovesicular steatosis (confirmed to be 30% in pathological examination). This patient had also high postoperative ALT levels despite smooth perioperative course. In a previous study, liver graft steatosis was associated with higher complement activation and increased liver injury during reperfusion (45). This along with the early experiences with liver transplantation for aHUS patients highlights the importance of using only high-quality liver grafts for these patients to secure uneventful recovery.
The first liver transplantations to aHUS patients were not successful (24–26). One case resulted in a fatal primary nonfunction of the liver and autopsy revealed ischemic-coagulative necrosis, congestion, microvascular thrombosis and signs of complement activation in the liver (26). The main reason why aHUS patients with dysfunctional CFH are especially vulnerable for problems during liver transplantation is likely to be defective regulation of the alternative pathway activation. Liver cells have been reported to have restricted expression of membrane-bound complement regulators on their surfaces (46). This might cause them to be more dependent on circulating complement regulators and more vulnerable to complement-mediated injury. We observed a small decrease in plasma CFH concentration at the time of reperfusion indicating a possible consumption of this complement regulator (Figure 5).
Successful combined liver–kidney transplantations for patients with CFH mutations have been performed under protection of extensive PE therapy (27–30). Preoperative PE replenishes the aHUS patients in functional CFH and is the intended method to prevent problems during the surgery. The aHUS patients in our study received extensive PE before the operation and both had higher preoperative C3d and Bb concentrations than the other liver transplantation patients. The higher preoperative levels, however, did not predict the level of perioperative complement activation. The increased C3d in plasma is in line with our earlier findings that the aHUS patients also had increased amounts C3d on the surfaces of their erythrocytes (unpublished data). These findings are concordant with preoperatively low C3 in plasma of the aHUS patients since low C3 is often caused by C3 consumption following ongoing complement activation. For both aHUS patients, the course of the operation was uneventful and the grafts showed good function postoperatively, although the patient aHUS-1 showed higher concentrations of the activation markers early during the operation. The early signs of activation might have been due to a more active disease and labile situation before the surgery.
A new therapeutic monoclonal antibody against C5 (eculizumab) that blocks complement-mediated direct cytotoxicity has been introduced during the past few years. It is currently used mainly in the treatment of paroxysmal nocturnal hemoglobinuria (PNH), but could be exploited in several other clinical situations. A few aHUS patients with complicated disease have been given eculizumab with encouraging results (17–20). The long-term safety and efficacy of eculizumab in the treatment of aHUS need to be verified, but they provide a promising alternative treatment option to the sometimes ineffective PE therapy and the combined liver–kidney transplantation. In liver transplantation surgery, this therapeutic antibody might be suitable in cases where the graft is known to be steatotic or otherwise suboptimal. This is, however, a speculative suggestion and remains to be addressed in further studies. Similar to eculizumab, another monoclonal anti-C5 antibody, pexelizumab, might be used in limiting complement-mediated damage. It has been evaluated in several clinical trials but is not available for clinical use yet (47–49).
In conclusion, our study provides novel information about activation of complement during liver transplantation in general and gives an explanation to the need for extensive PE therapy before combined liver–kidney transplantation of patients with aHUS. According to our results, perioperative activation of complement cannot be predicted from the preoperative levels of activation markers but compromised quality of the transplanted liver (e.g. steatosis) seems to be associated with complement activation. Since the plasma level of the terminal complement complex SC5b-9 is associated with a postoperatively increased ALT level and accumulation of neutrophils or congestion in postreperfusion liver biopsies, it remains possible that therapeutic anti-C5 antibody such as eculizumab might be beneficial in transplantation of a suboptimal graft, but this remains to be studied.
The authors thank Dr Hanna Jarva for helping with the CFH assay and Marjatta Turunen and Kirsti Widing for technical assistance. The authors are also grateful for the whole surgical team and the HUSLAB laboratory personnel responsible for handling the samples. This study was financed by the Academy of Finland (project 128646), Sigrid Juselius Foundation and HUSLAB EVO-Research funds. AK is partially financed by the Helsinki Biomedical Graduate School.
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.