Pancreas Allograft Biopsies with Positive C4d Staining and Anti-Donor Antibodies Related to Worse Outcome for Patients


Corresponding author: Hanneke de Kort,


C4d+ antibody-mediated rejection following pancreas transplantation has not been well characterized. Therefore, we assessed the outcomes of 27 pancreas transplantation patients (28 biopsies), with both C4d staining and donor-specific antibodies (DSA) determined, from a cohort of 257 patients. The median follow-up was 50 (interquartile range [IQR] 8–118) months. Patients were categorized into 3 groups: group 1, patients with minimal or no C4d staining and no DSA (n = 13); group 2, patients with either DSA present but no C4d, diffuse C4d+ and no DSA or focal C4d+ and DSA (n = 6); group 3, patients with diffuse C4d+ staining and DSA (n = 9). Active septal inflammation, acinar inflammation and acinar cell injury/necrosis were significantly more abundant in group 3 than in group 2 (respective p-values: 0.009; 0.033; 0.025) and in group 1 (respective p-values: 0.034; 0.009; 0.002). The overall uncensored pancreas graft survival rate for groups 1, 2 and 3 were 53.3%, 66.7% and 34.6%, respectively (p = 0.044). In conclusion, recipients of pancreas transplants with no C4d or DSA had excellent long-term graft survival in comparison with patients with both C4d+ and DSA present. Hence, C4d should be used as an additional marker in combination with DSA in the evaluation of pancreas transplant biopsies.


Pancreas transplantation is a widely accepted treatment choice for diabetic patients. By December 31, 2003, more than 21000 pancreas transplants had been performed worldwide (1). The 1-year survival of simultaneous pancreas kidney transplantation (SPKT) has greatly improved over the last decade, with rates of immunological graft failure of only 2% during 1996–2002 in technically successful SPKT. Rejection can occur through the immunological pathways of acute cell-mediated rejection (ACR) or antibody-mediated rejection (AMR).

C4d is a complement degradation product of complement factor C4, which is a footprint of the classical pathway of the complement cascade. C4d can be identified in tissue, and is a reliable marker of AMR in transplanted kidneys, intestines, heart and lungs (2–6). The typical histopathological and clinical findings in AMR are less well defined in pancreas than in other transplanted organs, and the role of C4d in relation to AMR is only beginning to be described. In a recent study, C4d+ immunolabeling of interacinar capillaries was significantly associated with donor-specific antibodies (DSA) and dysfunction of the pancreas allograft (7).

The diagnostic relevance and therapeutic consequences of positive C4d staining have been investigated extensively in relation to kidney transplantation. For other solid organ transplantations, knowledge of AMR and the role of C4d varies. In small bowel transplants, for instance, the relationship between C4d+ staining and AMR is less pronounced than in the kidney (3). For pancreas transplantation, which is often performed simultaneously with kidney transplantation, very little is known about the significance of C4d positivity. In the current study, we illustrate the C4d staining patterns in transplanted pancreases, the role of C4d in establishing a diagnosis of AMR, and their relationship with the long-term outcome of graft recipients from the combined experiences of two transplantation centers.

Patients and Methods

Patient selection

The electronic records of the Department of Pathology at the Leiden University Medical Center (LUMC) and the Department of Pathology at the University of Maryland Medical Center (UMMC) were searched resulting in the identification of 257 patients with pancreas allograft transplantation; 109 from LUMC (between 1991 and 2001) and 148 from UMMC (between 2000 and 2007). Of the 257 patients, 56 biopsies were available for this study (37 from UMMC and 19 from LUMC), and 27 patients had DSA information. From the LUMC, 26 preoperative pancreas transplant biopsies were available which were used to establish pretransplantation reference values for C4d staining of pancreas tissue. Clinical information on donor variables (gender and age at time point of death); recipient variables (age at time of transplantation, gender, re-transplantation, pregnancy and transfusion history and need for pre-emptive dialysis); transplantation-related factors (panel-reactive antibodies, human leukocyte antigen (HLA)-A, -B and -DR mismatches, cold ischemia time and induction therapy); and posttransplantation features (immunosuppressive regimen, acute rejection history, rejection treatment, time and indication for biopsy, cause of allograft loss, cause of death and follow-up in months of the pancreas allografts) were obtained from UMMC and LUMC medical records after approval for the study by the respective Institutional Review Boards.


All pancreas transplant biopsies were performed because of clinically suspected graft rejection, except for 1 protocol biopsy (Table 1). From all biopsies, 4-μm sections were cut and stained with hematoxylin and eosin (H&E).

Table 1.  Patient and donor demographics, clinical parameters, biopsy indication and treatment in relation to pancreas allograft function
Patient No.GAge at TxSexRe-TxPreg-nancyDonor sex, agePRA (%)Transfusion till biopsyPre-emptiveCIT (hour)Tx typeHLA mmInduction therapyIndication for biopsyTreatment prior to biopsyBiopsy (days)Treatment after biopsyFollow-up (months)Graft outcome
  1. ACS = abdominal compartment syndrome; Alm = alemtuzumab; ATG = anti-thymocyte globulin; Bas = basilixumab; biopsy (days), number of days after transplantation that biopsy was taken; CIT = cold ischemia time; creat = creatinin; Dac = daclizumab; G = group; GL = graft loss; HbA1c = hemoglobin A1c; HLA mm, human leukocyte antigen mismatches; MI = myocardial ischemia/infarction; MP = methyl prednisolone; OKT3 = muromonab-CD3; PAK = pancreas after kidney transplantation; PDFG = patient death with functioning graft; peripancreatic fluid = peripancreatic fluid collection; PP/IVIG = plasmapheresis/intravenous immunoglobulins; PRA = panel reactive antibodies; pre-emptive = pre-emptive therapy (to be on dialysis prior to Tx); PTA = pancreas transplantation alone; PTLD = posttransplant lymphoproliferative disease; SPKT = simultaneous pancreas kidney transplantation; Tx = transplantation.

 1139MNoF 174YesNo18:25SPKT5 creat↑,amylase↑, proteinuriaMP69 105,8PDFG
 2142FNo0F 360YesNo10:53SPKT4 creat↑,amylase↑MP, ATG 41MP135,3Functioning
 3138MNoF 210YesNo15:01SPKT4 amylase↑, proteinuriaMP, ATG, MP122  111,3Functioning
 4133MNoM 2111 YesNo11:35SPKT3 creat↑, amylase↑MP15 120,2PDFG (PTLD)
 5130MNoF 350YesNo11:35SPKT5 hyperglycemiaMP, ATG 31MP132,9Functioning
 6131FYes0M 4456 NoneNo12:41SPKT6ATGamylase↑, lipase↑ 132  11,4Functioning
 7142MNoF 420NoneNo19:08SPKT6ATGperipancreatic fluid  9 18,3Functioning
 8132MYesF 160NoneNo14:40PAK3ATGblood counts↑, fever, edema 18ATG x114,2Functioning
 9147MNoM 370NoneNo10:00SPKTAlmperipancreatic fluid 32 3,1Functioning
10132MNoM 2125 NoneNo19:30SPKT5ATGfever, edema 27 19Functioning
11148MNoM 312NoneNo15:18SPKT6ATGamylase↑, lipase↑ 915 MP43Functioning
12147MYesF 1069 NoneNo11:00PAK3Basamylase↑, lipase↑ 96MP5Functioning
13137FYes0M 160NoneNo26:43PAK5Basamylase↑, lipase↑ 1620  MP x2 + ATG x795,4Functioning
14256FNo2M 410YesYes15:00SPKT protocol 319  58,8Functioning
15228FNo1M 310NoneNo13:35SPKT6Basamylase↑, lipase↑ 83PP/IVIg x64,7Functioning
 1239MNoF 174YesNo18:25SPKT5 amylase↑MP, ATG481  105,8PDFG
16231MNoM 350YesYes12:37SPKT6OKT3creat↑, amylase↑, hyperglycemiaMP16ATG146Functioning
17247MNoM 124YesYes9:00SPKT3 amylase↑, HbA1c↑ 169 MP, ATG, MP75,5Functioning
18246MNoM 225YesYes12:52SPKT4 amylase↑MP, ATG222  136,6Functioning
19333MNoM 490YesYes15:40SPKT3 amylase↑, hyperglycemiaMP12ATG57,5PDFG (MI)
20341MNoM 2818 YesNo11:25SPKT6OKT3creat↑,amylase↑MP+OKT3, ATG21 169,5Functioning
21336FNo1M 3044 Yes5:31SPKT4OKT3amylase↑, proteinuria, edemaMP, ATG, MP161  7,2GL (rejection)
22328MNoF 363YesNo13:30SPKT4 blood glucose fluctuatingMP29 155,7Functioning
23340MYesM 1625 NoneNo24:52PAK5ATGamylase↑, lipase↑, hyperglycemia 174 ATG no MP, Alm21,7GL (rejection)
24350FNo2−1865 NoneNo26:30SPKT5Almperitonitis 10 3,9Functioning
25335MNoM 2717 NoneNo18:25PTA3AlmACS  1Pancreatectomy0,1GL (rejection)
26336FNo0M 244NoneNo11:00SPKT5Almcreat↑, amylase↑ 35PP/IVIg1,9GL (rejection)
27341FYes0M 3264 NoneNo19:41PAK6ATG, Dacleukocytosis, peripancreatic fluid 11 21,7Functioning

Immunohistochemical C4d staining

At LUMC, C4d immunohistochemical staining was performed on freshly cut formalin-fixed 4-μm paraffin sections of pancreas. Sections were deparaffinized, antigen was retrieved with TRIS/EDTA, pH 9.0, and the sections were incubated with a 1:25 dilution of rabbit anti-C4d polyclonal antibody (MP products, Biomedica, Austria) for 30 min. Subsequently, endogenous peroxidase was blocked and the EnVision™ Detection Systems Peroxidase/DAB, Rabbit/Mouse (DAKO cytomation, Glostrup, Denmark) was used for development.

At UMMC, C4d immunostaining was performed on 5-μm formalin-fixed paraffin sections using an automated immunohistochemical stainer (Ventana Medical Systems, Tucson, AZ) and labeled with a 1:50 dilution of anti-C4d rabbit polyclonal antibody (American Research Products, Inc., Belmont, MA). Target detection was performed using an indirect biotin-avidin system incorporating diaminobenzidine. Endogenous peroxidase quenching and biotin blocking were performed on-instrument with kit reagents (Ventana Medical Systems, Tucson, AZ). All stained sections were counterstained with hematoxylin for orientation.


The H&E-stained sections from the pancreas biopsies were scored according to the Banff working proposal of 2008 (8). Positive C4d staining patterns in interacinar capillaries were scored (Figure 1C). Diffuse C4d positivity was defined as >50% positively staining interacinar capillaries, focal C4d positivity was 5% to 50% positively staining interacinar capillaries, and minimal positive staining was <5% positively staining interacinar capillaries or a negative staining pattern. Table 2 lists the scorings of all individual parameters in relation to graft outcome.

Figure 1.

C4d staining patterns. (A) Negative/minimal staining. (B) Focal positive staining. (C) Diffuse positive staining in interacinar capillaries. Bar represents 5 μm.

Table 2.  C4d, DSA, specific histological and Banff scores for pancreas graft biopsies Thumbnail image of

Immunosuppressive therapy

At LUMC, standard maintenance therapy consisted of prednisone (P), cyclosporine (CsA) and azathioprine until May 1995. Transplant recipients after May 1995 received P, CsA and mycophenolate mofetil (MMF). Three patients received a 10-day course with 5 mg/day OKT3 as induction therapy. Acute rejection episodes were treated according to a standard protocol consisting of methylprednisolone (MP) 1 g intravenously for 3 consecutive days for the first rejection episode; a 10-day course of antithymocyte globulin (ATG) at a dosage of 5 mg/kg guided by absolute lymphocyte counts at the second rejection episode and again MP for the third rejection episode. None of the patients were treated for AMR.

Immunosuppressive protocol at UMMC included induction therapy with ATG (rabbit) (years 2000–2006) dose of 7–10 mg/kg or alemtuzumab (Alm) (2006–current) 30 mg. Steroid bolus of P 500 mg was administered intra-operatively and tapered to either low dose (before 2004) or off over 21 days (after 2004). Maintenance therapy included tacrolimus and MMF. Rejection therapy was tailored based on clinical suspicion, C4d staining and presence of detected antibody; and could include additional steroid therapy, ATG, plasmapheresis (PP) and intravenous immunoglobulin (IVIg).

DSA assessment

At LUMC, patient serum samples from 14 of 19 biopsies from 18 patients (1 patient had 2 biopsies) 1 month posttransplantation were procured from −80°C storage. A LAT™ mixed class I & II ELISA (One Lambda, Canoga Park, CA) was carried out to detect the presence of HLA class I and class II antibodies. The ELISA was conducted according to protocol with OD readouts at 630 nm. Afterward, seven positive patients were assessed for the specificity of the antibodies with a complement-dependent cytotoxicity (CDC) test.

At UMMC, DSA were identified in the patient serum samples at the time of biopsy using the Luminex 100 IS System (Luminex, Austin, TX). High-definition single beads were used with a mean intensity of fluorescence (MFI) cutoff value for one single antibody at MFI 1000. When three antibodies were present, for example, the MFI cutoff was 3000.

Statistical analyses

Raw data were processed using descriptive statistics and graphical representations. For uncensored graft survival, graft failure was defined as resuming insulin therapy or patient death with functioning graft. For censored graft survival, graft failure was defined as resuming insulin therapy; patient death with functioning graft was considered as lost to follow-up. Kaplan–Meier survival curves were compared with the log-rank test for uncensored and censored graft loss. Significance for all tests was set at 0.05. In Table 2 and Figure 2 the differences between groups, C4d−/C4d+ and DSA negative or positive were assessed by means of the chi-squared test, with *p < 0.05 and **p < 0.001. Statistical calculations were performed using SPSS 16.0 for Windows (SPSS, Chicago, IL).

Figure 2.

Bar graph of specific histological scores as displayed inTable 2. This figure gives an indication of the abundance of various lesions, regardless of extent of the lesion. Percentages are based on adequate biopsy samples (see Table 2). The differences between groups was assessed, with *indicating a p-value of <0.05.


Patient demographics

The mean age of the recipients was 46 ± 9 years, 45% were female and 91% were Caucasian. All the recipients had diabetes and 48% had hypertension as co-morbidity. The mean HLA mismatch was 4.4 ± 1.4. The median follow-up was 50 (IQR 8–118) months. Table 1 gives an overview of patient and donor demographics, clinical parameters, biopsy indication and treatment in relation to pancreas allograft function. No significant differences were found between the three groups except for a tendency to have higher PRA in group 3 and no pre-emptive therapy in group 1.

Hematoxylin and eosin staining

Fifty-six pancreas biopsies had H&E slides of which 24 showed normal histology. Four biopsies showed features of indeterminate rejection, 9 showed grade I (mild) acute cell-mediated rejection (ACR), 12 showed grade 2 (moderate) ACR and 7 showed grade 3 (severe) ACR. No chronic rejection was present in 37 pancreas biopsies, grade I chronic rejection was present in 10, grade II chronic rejection in 7, grade III chronic rejection in 1 and chronic ACR was present in 1.

Table 2 shows data on specific histomorphological lesions in 28 biopsies of patients whose DSA were determined, divided into three groups: patients with minimal or no C4d staining and no DSA were placed in group 1 (n = 13), six patients (2 patients with diffuse C4d+ and absent DSA, 2 patients with focal C4d+ staining and DSA present and 2 patients with minimal or absent C4d and DSA present) were placed in group 2 and nine patients with diffuse C4d+ staining with DSA present were placed in group 3. Frequency of ACR did not differ between the groups. Between group 1 and 2, there was no significant difference in the distribution of lesions. Active septal inflammation, acinar inflammation and acinar cell injury/necrosis were present significantly more often in C4d+/DSA+ patients (group 3) than in either group 1 or 2. Group 3 differed from group 1 on additional lesions, namely ductitis (p = 0.02), necrotizing arteritis (p = 0.042), capillaritis (p = 0.004) and edema (p = 0.013). Positive C4d staining correlated significantly with the presence of ductitis, venulitis, capillaritis, acinar inflammation, acinar cell injury/necrosis and edema. In the presence of DSA, capillaritis and acinar cell injury/necrosis were significantly found more often. Figure 2 shows the incidence of all scored histomorphological parameters.

C4d staining

There were 56 pancreas biopsies stained for C4d. Sixteen biopsies showed a diffuse staining pattern, 3 showed focal staining (Figure 1B) and 36 showed minimal or no staining. Of the 26 pretransplant pancreas graft biopsies, only one showed diffuse positivity for C4d. All others showed minimal or negative C4d staining (Figure 1A).

Donor-specific antibodies

Twenty-seven patients were tested for DSA, of which 9 had anti-HLA class I, 7 had anti-HLA class II and 15 had no antibodies. Two different methods were used to assess DSA: at UMMC, Luminex was used, at the LUMC, CDC was used.

Groups and outcome

Twenty-seven patients (1 patient had 2 biopsies for a total of 28 biopsies) with C4d and DSA information were categorized into three groups (Table 2). Patients with minimal or no C4d staining and no DSA were placed in group 1 (n = 13), six patients (2 patients with diffuse C4d+ and absent DSA, 2 patients with focal C4d+ staining and DSA present and 2 patients with minimal or absent C4d and DSA present) were placed in group 2 and nine patients with diffuse C4d+ staining with DSA present were placed in group 3. Figure 3A shows uncensored overall pancreas graft survival in groups 1, 2 and 3, which were 53.3%, 66.7% and 34.6%, respectively. The overall group p-value was significant for the three groups at the end of follow-up (p = 0.044). Figure 3B shows censored overall pancreas graft survival in groups 1, 2 which were 100% and group 3, which was 51.9%. The overall group p-value was significant for the three groups at the end of follow-up (p = 0.006). All graft failures due to rejection were found in group 3. Graft failures associated with AMR occurred during early follow-up (up to 2 years after transplantation).

Figure 3.

Uncensored (A) and censored (B) pancreas graft survival with C4d and DSA. Kaplan–Meier curve and descriptive table of uncensored (A) and censored (B) pancreas graft survival in group 1 patients with minimal or no C4d staining and no DSA, group 2 either negative/minimal C4d or focal C4d+ staining in the presence of DSA or focal C4d+ staining in the absence of DSA and group 3 diffuse C4d+ in the presence of DSA.


In this study, we found heterogeneous C4d staining patterns in pancreas transplant biopsies taken at various time points after transplantation. In the recent Banff classification of pancreas transplant biopsies, C4d staining patterns have a role in combination with histologically proven acute rejection of the pancreas and confirmed DSA (8). In our study, the outcomes of patients who had C4d and DSA information available were assessed. Patients with diffuse C4d positivity and concurrent DSA had poor graft survival, which was most likely related to AMR. Patients who had either DSA or C4d alone did well, similar to patients with no DSA or C4d staining detected in the biopsies. Although the number of patients is relatively small, these results suggest that the occurrence of either focal C4d positivity or DSA alone are of uncertain clinical significance and may not pose additional risk of poor outcome to the pancreas graft. The possible reason for C4d positivity in the absence of HLA DSA may be due to the presence of major histocompatibility complex class I chain-related gene A (MICA) or glutathione-S-transferase T1 (GSTT1) (9), or due to the timing of DSA assessment.

C4d+ AMR following pancreatic transplantation has not been well characterized. In 2006, Melcher et al. was the first to report a patient with a SPKT who developed AMR with donor-specific HLA-DR allo-antibodies and who had positive C4d staining in the pancreas 1 month after a SPKT. The renal transplant biopsy from this patient, taken at day 10, was also positive for C4d (10). A year later, Carbajal et al. reported a patient with pancreas after kidney transplantation (PAK) who developed AMR of the pancreas with a C4d positive biopsy. The kidney, which was from a different donor, remained unaffected, but the pancreas graft was lost (11). In the present study, a small number of simultaneously taken kidney biopsies were available with rather heterogeneous findings in the kidney and pancreas biopsies in relation to C4d staining patterns and Banff classification. Unfortunately, DSA information was scarce in this group, and therefore, no firm conclusions could be drawn from these data. Gaber reported one case of a C4d positive surveillance biopsy taken 2 weeks after transplantation from a patient who was sensitized by two previous islet transplants and who had detectable anti-HLA antibodies (12). Pascual et al. reported on 13 patients with an acute rejection of the pancreas, of which 2 were biopsy-proven and stained positive for C4d. In one patient, two pancreas biopsies were taken. DSA were negative at the time of the first biopsy, which was C4d positive. Three months later, a second biopsy was not only C4d positive but DSA positive as well (6). In a recent study by Torrealba et al., the potential role of C4d in pancreas transplant biopsies in the diagnosis of AMR was described in a group of 18 patients (7). In our study, an assessment was performed to identify specific histological lesions correlating with AMR in the pancreas, next to C4d and DSA assessment in relation to graft outcome. Although we have a relatively small number of biopsies, results from our study give an indication that septal inflammation, acinar inflammation, acinar cell injury/necrosis, capillaritis and edema are associated with AMR.

In the recent Banff classification of pancreas transplant biopsies (8), C4d staining patterns have a role in diagnosing AMR in combination with histologically proven acute rejection of the pancreas and confirmed DSA. As a clinical guideline, the results of our study emphasize the importance of taking into account a combination of C4d staining pattern, histological diagnosis and the presence of DSA before a diagnosis of AMR of the pancreas is made. The limitations of this study include the use of retrospective data and biopsies, and heterogeneity in the clinical indications for performing the transplant biopsies. The number of patients was too small to describe the clinical phenotype associated with AMR. Nevertheless, nearly all biopsies were taken within 1 year after transplantation, and the heterogeneity of the histopathological findings was helpful in establishing whether C4d could have an additional role in the work-up of pancreas transplant biopsies. The methods used to determine DSA levels in this study differed per institution, which may raise concern for the analysis and conclusions reached. Nevertheless, both methods are clinically used and accepted for diagnosis and subsequent treatment. DSA assessments were not always performed at the time of biopsy. However, AMR is a fluid state, and DSA levels are known to fluctuate over time. Therefore, it is uncertain what time point would be optimal to identify those patients developing an AMR.

In conclusion, C4d is a reliable tissue marker in combination with concurrent DSA in the evaluation of pancreas transplant biopsies. The finding that C4d positivity in pancreas transplant biopsies can occur in pancreas allografts with a good clinical outcome places doubt on the absolute value of C4d. Therefore, C4d staining patterns in pancreas transplant biopsies must be interpreted in combination with information about DSA and the histopathological lesions classified according to the Banff criteria. This clinical approach is similar to the standard procedures used to evaluate renal transplant biopsies. In this study, patients with diffuse C4d+ staining in the pancreas biopsy with concurrent DSA had poor graft survival, most likely related to AMR. The occurrence of either focally C4d+ staining or DSA poses no additional risk of a poor outcome of the pancreas graft in this study, although the sample number was relatively small. Further studies are needed to determine the value of the presence of a single marker.