Treatment with Anti-MHC-Class-II Antibody Postpones Kidney Allograft Rejection in Primates but Increases the Risk of CMV Activation
Abstract
Treatment of kidney graft recipients with antibodies that may specifically suppress the anti-donor response would be an ideal situation to prevent graft rejection. MHC class-II-specific antibodies and, in particular, DR specific antibodies have often been proposed as treatment to prevent antigen presentation, and thus graft destruction. Here we report an attempt to prevent graft rejection using a humanized MHC class-II-specific monoclonal antibody CDP855 in a cynomolgus monkey kidney graft model. A modest delay in graft rejection was observed when the antibody was given only on days 0, 1 and 2 after transplantation. Unexpectedly 50% of the animals succumbed of a viral infection, most likely CMV in two of three cases, prior to graft rejection in the first week post-transplantation. We speculate that the antibody treatment triggered CMV activation, possibly as a consequence of the activation of factors such as NF-κb by the interaction of the antibody and its target cells.
Introduction
In the recent years an increasing number of immunosuppressive drugs have become available for the prevention of graft rejection. This has improved graft survival. Unfortunately, the routinely used drugs have serious side effects, which cause morbidity and do not prevent late graft failure due to chronic vascular pathology. This has stimulated in recent years the search for rejection therapies that will allow graft function with a minimum of immunosuppression, preferably to be given only around the time of transplantation. The most promising approaches in non-human primates have been rigorous T-cell depletion combined with other immunosuppression (1–3) and blockade of co-stimulation (4). Another approach that might be additive to these approaches is to specifically block the APC-T-cell interaction by using MHC class-II-specific antibodies. It has been shown in rodents (5–11), as well as in monkeys (12–14) that anti-MHC-class-II antibodies can exert an immunosuppressive effect both in transplantation as well as in auto immune models (15). However, also serious adverse effects have been reported to occur after in vivo administration of anti-class-II-antibodies in primates (12,13) and rats (5). These consisted of acute intravascular coagulation, which was dose dependent, and acute shock-like signs resembling a cytokine release syndrome (8,12,13).
In this study we investigated the safety and efficacy of CDP855, a humanized, IgG4 version of the murine L243 antibody in a cynomolgus monkey kidney allograft model.
Materials and Methods
Animals
Naive, captive-bred 4–6 kg cynomolgus monkeys (Macaca fascicularis) were purchased from a commercial breeding station and housed at the Biomedical Primate Research Centre. All procedures were performed in accordance with guidelines of the Animal Care and Use Committee installed by Dutch law. All animals were typed for Mafa-A, B and DR antigens by serology (16). Disparity for DR locus antigens was confirmed by DRB typing. Recipients were disparate for at least one Mafa-DR antigen and at least one Mafa-A and -B antigen with the donor. The recipient–donor pairs were compatible for ABO-antigens (17). In addition, the stimulation index of the one-way mixed lymphocyte reaction of the recipient cells directed against the donor antigens was positive (>3, see below).
Antibody and treatment
In the pre-transplantation safety experiments, the murine DR specific antibodies were used: L243(IgG2a) (18). The Fab′2 fragments, as well as humanized versions of this MHC-DR specific antibody, L235E (chimeric human IgG1) and CDP855 (chimeric human IgG4) were used for safety assessment. CDP855 was used in the transplantation study and was given intravenous (IV) over a period of 60 min. On the day of transplantation (day 0) the total dose of 10 mg/kg of CDP855 was administered just before the reperfusion of the transplanted kidney. On days 1 and 2 the test substance was given IV over a period of 60 min. The animals were sedated for dosing using ketamine (±10 mg/kg) on days 1 and 2.
Plasma levels of CDP855 were not measured in the animals in this study. However, PK studies performed in healthy animals following three consecutive daily doses of 10 mg/kg CDP855 showed that at this dose the 24-h trough levels were around 1 μg/mL. This level is most likely sufficient to block the DR molecules present in the recipients for a short period.
Kidney transplantation
Heterotopic kidney allotransplantation with bilateral nephrectomy was performed as described previously (19,20). The clinical condition of the animals was monitored by daily visual inspection and by frequent hematological and clinical chemistry blood values determined in a local clinical laboratory (SSDZ, Delft). Transplant rejection was monitored by serum urea and creatinine. Rejection was not treated. When serum creatinine was >1000 μmol/L or when the clinical condition deteriorated, the animals were euthanized and necropsy was performed. For histological examination, tissues from the necropsy were formalin-fixed and sections were stained with hematoxylin and eosin (H&E), periodic acid Schiff and a silver impregnation stain (Jones). Histomorphological evaluation of allograft rejection was performed according to the Banff classification (21). The Mann-Whitney U test was used as statistical analysis for the graft survival.
Immunological determinations
Mixed lymphocyte reactions were performed to evaluate the inhibitory effects of the antibody preparations, and were performed between prospective donor and recipient animals to select MLR positive combinations. Peripheral blood mononuclear cells (PBMC) were obtained using lymphocyte separation medium (LSM; ICN Biomedical, Aurora, OH). A total of 105 responder PBMC were co-cultured in 96-well round-bottom plates with 105 irradiated (30 Gy) stimulator PBMC in 150 μL RPMI1640 medium supplemented with 25 mM Hepes, 2 mM L-glutamine, 10% heat inactivated fetal calf serum, 20 μM 2-mercaptoethanol, 100 U/mL penicillium and 200 μg/mL streptomycin (all obtained from Invitrogen, Breda, the Netherlands). The antibodies were added to the cultures at the concentrations indicated in the results section. The cultures were incubated at 37°C in 5% CO2 for 5 days, the last 20 h in the presence of 0.5 μCi 3H-thymidine in 50 μL. The cells were harvested on filters and radioactivity was measured in a matrix 9600 beta counter (Packard, Meridan, CT).
Subset analyses were performed using a whole blood assay and analysis by flow cytometry using a FACScan (Becton Dickinson (BD), Immunocytometric Systems, San Jose, CA). All procedures were performed at 4°C. The antibodies used were: FN18/FITC (CD3, BPRC), Leu-3A/PE (CD4, BD), Leu-2A/PE (CD8, BD), HLA-DR/PE (BD) and Leu-16/PE (CD20, BD). FACS Lysing Solution (BD) was added according to the instructions of the manufacturer. Flow cytometry was performed on the stained cells with forward and perpendicular light scatter gates set on the lymphocyte fraction. Data were stored on discs and subsequently analyzed using the BD FACScan™ program. For analysis, the unstained control cell population was used for quadrant marker placement.
CMV IgM and IgG levels were determined by ELISA by a local clinical laboratory (SSDZ, Delft). A human CMV negative reference serum was used as control. A ratio between the OD of the monkey serum and the reference serum was calculated, and if this ratio was higher than 1.1, the monkey was considered to be CMV antibody positive and thus a potential carrier of CMV.
Results
In vitro effects in monkeys
To test the capacity of CDP855 to block the alloresponse in monkeys, its inhibitory effects were determined in mixed lymphocyte cultures of rhesus, cynomolgus monkeys and in human cultures. Eight MLR positive responder–stimulator cell combinations were tested for each species in two independent experiments per species. Maximal inhibition of the MLR was observed at concentrations of 1000 ng/mL or higher for all three species. The level of inhibition was slightly higher in the human cultures as compared to the monkey cultures (Figure 1), but this was not significant. The IgG1 form of the antibody (L235E) and the Fab′2 fragments were equally inhibitory (results not shown).
Inhibition of MLR of human cells (open circles), rhesus monkey cells (open triangles) and cynomolgus monkey cells (closed symbols). 0% inhibition: no antibody added to the cultures. Inhibition percentage is 100 − (blocked response/medium response × 100).
Safety aspects in monkeys
MHC class-II-specific monoclonal antibodies have been reported to cause serious side effects. Although such antibodies have been administered without problems in many studies, some antibodies may cause intravascular coagulation, most likely caused by direct stimulation of the vascular endothelium (12,13). CDP855 is derived from the murine antibody L243 (22). The murine form (IgG2a) of the antibody was given to rhesus monkeys at a dose of 10 mg/kg (bolus IV injection, single dose) and at 0.1 mg/kg (bolus IV injections, daily for 10 days). Two animals received the high dose of antibody. These two animals both developed acute cardiopulmonary distress and died of severe pulmonary edema within 20–25 min of the start of the infusion of the antibody. The lower dose was well tolerated for the total duration of the infusion period of 10 days. Thus, the side effects were dose dependent. Most likely, the side effects were also dependent on the Fc part of the antibody as Fab′2 fragments given at 0.6 to 6 mg/kg did not cause these side effects (Table 1). Subsequently chimeric forms of the antibody were developed, which have a human Fc tail. The IgG4 form of this antibody, CDP855, does not show significant binding to human, rhesus monkey or cynomolgus Fc receptors, and thus is less likely to cause serious side effects. The antibody was first tested in rhesus monkeys at 1 mg/kg. This was well tolerated. Administration of this antibody at 10 mg/kg antibody given over a 1-h period to cynomolgus monkeys did not result in any clinical side effects.
| Species | Fc isotype | Dose and frequency | N | Result |
|---|---|---|---|---|
| Rhesus monkey | Murine IgG2a | 10 mg/kg, once | 2 | Acute toxicity (shock), death after ±20 min |
| Rhesus monkey | Murine IgG2a | 0.1 mg/kg/day, 10 days | 1 | No significant side effects |
| Rhesus monkey | L243 Fab2 | 0.6 mg/kg, once | 1 | No significant side effects |
| Rhesus monkey | L243 Fab2 | 2 mg/kg, once | 1 | No significant side effects |
| Rhesus monkey | L243 Fab2 (none) | 6 mg/kg, once | 2 | No significant side effects |
| Rhesus monkey | CDP855 | 1 mg/kg, once | 2 | No significant side effects |
| Human IgG4 | ||||
| Cynomolgus monkey* | CDP855 | 10 mg/kg/day | 6 | No significant side effects |
| Human IgG4 | 3 days | |||
- *Animals the same as in transplantation study.
Graft survival and CMV infection
Untreated control animals rejected their graft within 7 days (rejection days 4, 6, 6, 8) with signs of severe cellular rejection (23). Table 2 shows the animal survival and the cause of death for all six CDP855-treated animals. Three animals showed delayed graft rejection at 12–15 days post transplantation. Three animals had a short survival, but this was not due to graft rejection as evidenced by the absence of significant pathological changes in the grafts indicative for rejection. All three animals had signs of viral infections, which either directly or indirectly may have caused the death of the animals. Two animals (455 and 141) had an inflammatory process at the site of ureter–bladder anastomosis. This most likely caused a blockade of the ureter and the animals passed very little urine the days just before autopsy. In both the animals evidence for an active CMV infection was found in the urinary bladder (typical CMV inclusions). Animal 141 also had severe peritonitis. This could have been caused by urine leaking into the abdomen. Animal 583 had no signs of CMV infection on histopathology, but died of pneumonia, which was most likely of viral origin: as there was no evidence of bacterial infection (no gram positive staining, the pleural fluid was sterile upon culture) and in the lung syncytial cells were found. In a group of 36 cynomolgus monkeys that received a kidney transplant under the same experimental conditions, and also received more potent immunosuppression (sirolimus and everolimus) (23) in the 2 years prior to this study, not one early post-transplantation deaths occurred due to CMV disease or viral pneumonia. CMV antibodies were determined in serum samples before transplantation and at autopsy. All six animals had IgG antibodies before transplantation indicating that all recipients were latently infected with CMV. No IgM antibody titers could be detected. The two animals that developed a proven CMV infection (455 and 141) showed a slight decrease in CMV IgG antibodies, as well as animal 091, while the other animals showed a slight increase in CMV IgG antibodies (Table 3).
| Monkey | Day necropsy | Banff score | Histology kidney graft | Main diagnosis | |||
|---|---|---|---|---|---|---|---|
| g | i | t | v | ||||
| 583* | 6 | 0 | 1 | 0 | 0 | No rejection | Pneumonia (viral) |
| 455 | 7 | 2 | 0–1 | 0 | 0 | No rejection | Ureter obstruction due to CMV infection |
| 141 | 8 | 0 | 0–1 | 0 | 0 | No rejection | Ureter obstruction due to CMV infection; peritonitis |
| 43* | 12 | 3 | 2 | 3 | 1 | Grade IIA acute rejection | Acute kidney graft rejection |
| 91 | 13 | 3 | 2 | 3 | 0–1 | Grade IB acute rejection | Acute kidney graft rejection |
| 377 | 15 | 1 | 2 | 2 | 0 | Grade IA acute rejection | Acute kidney graft rejection |
- g: glomerulitis; i: interstitial inflammation; t: tubulitis; v: intimal arteritis. 0 = no changes; 1, 2, 3: increasing Changes.
- *These monkeys received prophylactic ganciclovir.
| Monkey | Prophylactic ganciclovir | Day autopsy | Diagnosis | CMV IgM antibody* | CMV IgG antibody* | ||
|---|---|---|---|---|---|---|---|
| Day 0 | Autopsy | Day 0 | Autopsy | ||||
| 583 | yes | 6 | Pneumonia (viral) | 0.4 | 0.4 | 3.4 | 4.8 |
| 455 | 7 | CMV infection | 0.4 | 0.5 | 2 | 1.5 | |
| 141 | 8 | CMV infection | 0.4 | 0.4 | 1.4 | 0.7 | |
| 43 | yes | 12 | Acute kidney graft rejection | 0.4 | 0.4 | 1.8 | 3.2 |
| 91 | 13 | Acute kidney graft rejection | 0.4 | 0.4 | 1.6 | 1.1 | |
| 377 | 15 | Acute kidney graft rejection | 0.4 | 0.5 | 1.9 | 2.5 | |
- *The figures represent a ratio of the OD of monkey serum divided by the OD of a negative human reference serum. Values above 1.1 are considered positive.
Lymphocyte subset analysis showed that both T and B cells were lower for the first 4 days post-transplantation. B-cells (CD20 positive) remained depressed in all animals during the whole post-transplantation period (Figure 2). The number of DR positive cells, comprising B-cells, monocytes and possibly activated T cells, were low during the first week post-transplantation, but these increased again during the second week post-transplantation.
Mean percentage of CD4, CD20 and DR positive cells after CDP855 treatment as determined by immunofluorescence and FACS analysis.
Discussion
Antigen presentation after organ transplantation occurs by the direct presentation of donor MHC class-II molecules, as well as by indirect presentation of donor MHC peptides in recipient MHC class-II molecules to recipient T-cell receptors. The MLR is considered as an in vitro analogue of the in vivo rejection process. The proliferation measured in this assay depends a to a large extent on the direct presentation of stimulator MHC class-II antigens, as unrelated class-II identical individuals, differing for multiple minor antigens, will show a significantly reduced proliferation in such MLR assays, both in human (24) as well as in monkeys (25). Addition of MHC-DR specific antibodies are known to block the MLR presumably by preventing adequate recognition by the T-cells (26–28). Whether this is limited to the direct presentation pathway is unclear. It also seems likely that indirect presentation can be blocked as this is also mediated by class-II molecules. Blocking this process in vivo by infusing the recipient with anti-DR antibodies may thus prevent graft recognition and graft rejection. Early studies in rats have shown that recipient-specific but not donor-specific class-II antibodies delay graft survival, emphasizing that the indirect pathway of antigen recognition is important and can be blocked in allotransplantation (8). The results from this study show that the anti-DR antibody CDP855 is capable of blocking the alloresponse in vitro. The response was never reduced completely to background levels, indicating that perhaps other class-II antigens (such as DQ) play a minor role in antigen presentation as well.
Kidney graft rejection in three CDP855-treated animals was slightly but significantly delayed as compared to untreated control animals. The antibody treatment resulted in B-cell depletion from the circulation. This depletion could be the result from B-cell kill, or it could be that the homing pattern of the B-cells had changed. Interestingly, CD20 negative and DR positive cells reappeared earlier than the CD20 positive B-cells, indicating that these distinct cell populations responded differently to the treatment. These results are in agreement with an earlier report when the IgG1 isoform of the same antibody (gLE1) was used in cynomolgus monkeys. gLE1 prolonged kidney allograft survival to 14–28 days and resulted in a similar short B-cell depletion (29). As the treatment was given only at the time of transplantation, rejection was only delayed for a short period and the grafts were lost due to cellular rejection at 2 weeks post-transplantation. The mechanism by which the anti-MHC-class-II antibodies exert their effect remains unclear. When inhibition of the direct presentation pathway is the underlying mechanism, donor-specific anti-class-II antibodies should be able to inhibit. However, in rats using allele-specific anti-DR antibodies it has been demonstrated that the effects are primarily exerted on recipient, as donor-specific antibodies did not have a significant effect (10). Also in a xenotransplantation study, anti-DR treatment of the recipient was effective (9). Our own previous data in rhesus monkeys using other murine monoclonal antibodies showed some graft prolongation (30). The mechanism by which the anti-DR treatment did block in vivo remains unclear.
Three of the six animals treated with CDP855 developed a viral infection shortly after transplantation. Two of these animals had a histologically proven CMV infection. In the third animal, no typical CMV inclusion bodies were found. However, the clinical symptoms of this animal's disease fit with a CMV infection. A very high percentage of Macaques are latently infected with CMV. As all animals in this study had pre-existing anti-CMV antibodies, it is most likely that an activation of an endogenous latent CMV infection occurred. In immunosuppressed macaques a latent CMV infection is often activated and animals become productively infected and may develop lesions and diseases related to CMV infection (31). In immunosuppressed animals with inflammatory lesions not related to CMV infection, CMV infection appears to be attracted by the inflammatory process, replicates in the affected tissue and may aggravate the existing lesion (Kuhn, unpublished observations in SIV-infected macaques). To our knowledge, it has not been described earlier that anti-class-II antibodies can cause activation of a viral infection. CMV can be activated in organ transplant recipients due to the combination of an inflammatory response to the allograft combined with the immunosuppression. The allograft results in the activation of genes such TNF-α and IFN-γ, followed by up-regulation of transcription factors AP-1 and NF-κB. These in turn are capable of activation of latent CMV (32,33). The immunosuppression could then prevent an adequate attack of viremia. However, in this study, CDP855 was only marginally immunosuppressive as graft survival was only marginally prolonged, and it seems unlikely that the general immunosuppressive effect of CDP855 can be held responsible for the viral activation. In the same animal model using CsA, sirolimus and everolimus (23) we did not observe CMV disease or pneumonia in a large group of cynomolgus monkeys from the same origin. Although we cannot exclude that a ureter obstruction was responsible for an inflammatory process that subsequently triggered CMV disease in two animals, this seems not very likely in view of our previous experience in this model. We, therefore, hypothesize that the anti-DR antibody was more directly responsible for the re-activation of the virus. The antibody itself could not activate CMV in human CMV infected cells (Proesch and Volk, personal communication). However, it has been described that anti-DR and L243 in particular can trigger B cells to become activated and release TNF-alpha and activate NF-κB (34,35). Monkeys treated with various dosages of the antibody but that did not receive a kidney allograft have not developed CMV disease. Thus, both the transplantation procedure and the antibody treatment may have acted in concert resulting in CMV activation.
