Clinical and Pathological Features of Plasma Cell-Rich Acute Rejection After Kidney Transplantation

Abstract

Background 

Plasma cell-rich acute rejection (PCAR) is a rare type of allograft rejection characterized by the presence of mature plasma cells. In general, the prognosis of PCAR is poor, and its clinical and pathological features remain unclear.

Methods 

We performed a retrospective observational study and compared allograft survival between kidney transplant recipients who developed PCAR and those who did not develop PCAR. We further analyzed clinical and pathological risk factors for allograft failure in PCAR patients.

Results 

Of 1956 recipients, 40 developed PCAR. There was a higher prevalence of deceased donor transplants (27.5% vs 11.7%, P = 0.0059), longer median total ischemia time (99 minutes; interquartile range, 71-144 vs 77 minutes; interquartile range, 59-111; P = 0.0309), and lower prevalence of ABO-incompatible transplantation (7.5% vs 22.5%; P = 0.0206) in patients with PCAR than in those without PCAR.

Multivariate Cox regression analysis showed that development of PCAR was associated with allograft loss (hazard ratio, 8.03; 95% confidence interval, 3.89-14.80; P < 0.0001).

We classified PCAR according to the Banff 2015 criteria into a borderline change group, a T cell–mediated rejection (TCMR) group, an antibody-mediated rejection (AMR) or suspected of having AMR (AMR/sAMR) group, and a mixed rejection (TCMR/AMR) group. The AMR/sAMR group was associated with a lower rate of allograft survival without significant difference (log-rank test, P = 0.1692).

Conclusions 

The results indicated that PCAR was an independent risk factor for allograft loss. PCAR presented with all types of rejection in the Banff 2015 criteria, and AMR/sAMR was associated with poor allograft survival.

In Brief

Plasma cell-rich acute rejection is a rare and ill-defined category of acute rejection encountered in kidney transplant recipients. In this small series, the authors improve out knowledge regarding both clinical presentation and prognosis.

Plasma cell-rich acute rejection (PCAR) is a rare type of allograft rejection characterized by infiltration of mature plasma cells. Since the first report by David-Neto et al¹ in 1993, this type of rejection has been described in some studies and case reports and is now characterized by poor graft survival and refractoriness to treatment,^2,3 In general, the prognosis of PCAR is poor, and its clinical and pathological features remain unclear. In addition, our understanding of allograft rejection has improved, and we now categorize T cell–mediated rejection (TCMR) and antibody-mediated rejection (AMR) according to the latest Banff 2015 criteria.⁴ PCAR has been described as either Banff IA or IB (TCMR),⁵ but we recently reviewed some PCAR cases with histological features of AMR or mixed TCMR/AMR, with serological evidence of donor-specific antibody (DSA).⁶

To clarify the pathogenesis and etiology of PCAR, we investigated clinical and pathological characteristics.

MATERIALS AND METHODS

Patient and Study Setting

This was a retrospective observational study. We assessed 1956 kidney transplantation procedures and 9347 kidney allograft biopsies performed between 1999 and 2012 at our institution and related hospitals. The samples were collected in the Department of Pathology in our institution and diagnosed by experienced pathologists. The primary outcome was death-censored graft survival (defined as the date of initiation of either form of renal replacement therapy after first indication biopsy).

The patients were divided into a PCAR group and a non-PCAR group. The inclusion criteria for PCAR were diagnosis of allograft rejection according to the Banff 2015 classification⁴ and the presence of more than 10% plasma cells among all infiltrating cells in the cortex, counted in 10 serial, high-power fields using periodic acid-Schiff, periodic acid-methenamine-silver, and/or hematoxylin and eosin staining (Figure 1). As described previously,² plasma cells were easily identifiable by their “clock-face” nuclear chromatin, eccentric nuclei, amphophilic/eosinophilic cytoplasm, and paranuclear pale zone. We confirmed these cells were stained by CD138 in some cases.

FIGURE 1: Histological findings of PCAR in kidney allografts. Upper left, tubulointerstitial rejection with diffuse inflammatory infiltrates (PAS, ×200). Upper right, a magnified view of interstitial and tubular inflammation demonstrating plasma cells, lymphocytes, and some eosinophils (PAM, ×400). Lower left, infiltrated plasma cells (CD 138 staining, ×400). PAS, periodic acid-Schiff; PAM, periodic acid-methenamine-silver.

We excluded cases that were diagnosed with reflux nephropathy or pyelonephritis, or were biopsied after the induction of dialysis, or showed adenovirus or polyomavirus nephropathy using hematoxylin and eosin-based identification of viral inclusions⁷ and simian virus 40 immunohistochemistry (mouse monoclonal antibody, diluted 1:600, PAb416; Abcam plc, Cambridge, UK). Samples showing posttransplantation lymphoproliferative disorder (PTLD) nephropathy were also excluded by in situ hybridization for Epstein-Barr virus-encoded RNA and plasma cell monoclonality, as verified by immunostaining of immunoglobulin heavy chains (IgG, IgA, IgM) and light chains (kappa and lambda), when in situ hybridization for Epstein-Barr virus-encoded RNA was positive.

The procedures followed were in accordance with the Declaration of Helsinki and its revisions. This study was reviewed and approved by the local institutional review board (approval number 3506).

Banff Classification

All biopsies that met the above criteria were diagnosed with PCAR and classified by experienced pathologists using the Banff 2015 criteria.⁴ C4d staining was performed via immunohistochemistry on deparaffinized sections, as previously described,^8,9 and interpreted according to the recommendations of the Banff criteria.

We classified all biopsies of PCAR into 4 rejection types (borderline change, TCMR, AMR, and mixed TCMR/AMR).⁴ In brief, borderline changes were diagnosed in cases with foci of tubulitis (t1, t2, or t3) with minor interstitial inflammation (i0 or i1), or interstitial inflammation (i1, i2, i3) with mild tubulitis (t1). TCMR was diagnosed in cases with histological evidence of both significant interstitial inflammation (Banff interstitial inflammation score, i ≥ 2) and moderate to severe tubulitis (Banff tubular inflammation score, t ≥ 2). AMR was diagnosed in cases with histological evidence of microvascular inflammation (Banff glomerular inflammation score, g ≥ 1 and/or Banff peritubular capillary inflammation score, ptc ≥ 1), with linear C4d staining in peritubular capillaries (Banff C4d score, c4d ≥ 2), or moderate microvascular inflammation (g + ptc ≥ 2) without linear C4d staining in peritubular capillaries (c4d ≤ 1). In this report, we suspected AMR (sAMR) if the cases showed pathological features but did not show serologic evidence of DSA or if DSA status was unknown. Mixed TCMR/AMR was diagnosed in cases with histological evidence of both TCMR and AMR or sAMR.

Clinical Data

The clinical data of patients in the present study were collected from their hospital records. The regimen or dose of immunosuppression was changed by the primary clinician according to clinical status and the presence of infection (viral or bacterial), uncontrolled side effects, or pregnancy. We screened for polyomavirus nephropathy according to guidelines.¹⁰ We monitored urinary cytology every month; if decoy cells were present, we tested urine and/or blood for BK viral load and considered allograft biopsy.

Donor-Specific Antibodies

Patients were evaluated for HLA sensitization status using a complement-dependent cytotoxicity assay. Serum samples were examined for IgG antibody against HLA class I or II, using Flow PRA or LABScreen Mixed (One Lambda, Canoga Park, CA). DSA was identified on screening of HLA antibody-positive recipients using LABScreen single antigen beads (One Lambda). Mean fluorescent intensity values over 1000 were determined to be positive.

Data Analysis

All analyses were performed using the JMP software package (ver. 10.0.0; SAS Institute). Data were summarized as means ± SD, median and interquartile range (IQR), or frequencies. Categorical variables were analyzed with the χ² test, Fisher exact test, and the 2-group proportion test, whereas continuous variables were compared by using the paired t test, the Wilcoxon signed-rank test, the Mann-Whitney U test, the Kruskal-Wallis H test, or analysis of variance, as appropriate. Kaplan-Meier curves and log-rank tests were used to compare death-censored allograft survival rates. The Cox proportional-hazards model was used to calculate hazard ratios (HRs) and 95% confidence intervals (95% CIs) for death-censored allograft survival. All P values were 2-sided, and P values of 0.05 were considered to indicate statistical significance.

RESULTS

Population and Criteria

Between 1999 and 2012, 1,956 kidney transplantations and 9,347 kidney biopsies were performed (Figure 2). Fifty biopsy samples from 50 allografts matched the inclusion criteria and were diagnosed with plasma cell-rich infiltration; 10 cases were excluded (4 cases of reflux nephropathy, 3 cases of BK virus (BKV) nephropathy, 1 case of PTLD in the allograft, 1 case of drug-induced nephropathy, and 1 case in which biopsy was performed after initiation of hemodialysis). Forty biopsy samples from 40 patients were diagnosed with PCAR, and the other 1,916 patients were defined as non-PCAR patients. The prevalence of PCAR was 2.0% of all allografts.

FIGURE 2: The flow criteria of PCAR patients through the study. A total of 9347 kidney biopsies of 1956 allograft kidneys were evaluated. Fifty biopsy samples matched the inclusion criteria, and 10 cases were excluded. Forty biopsies were diagnosed with PCAR.

Clinical Demographics of PCAR and non-PCAR Patients

The baseline characteristics of the patients at kidney transplantation are shown in Table 1. There was no significant difference between the PCAR and non-PCAR patients in terms of age, sex, primary kidney disease, donor status, warm ischemia time, prior history of transplantation, HLA mismatch, positive prevalence of crossmatching, or induction or maintenance immunosuppression regimen. The prevalence of deceased donor transplants was higher in PCAR patients (27.5%) than in non-PCAR patients (11.7%, P = 0.0059). The median total ischemia time was longer in PCAR patients (99 minutes; IQR, 71-144) than that in non-PCAR patients (77 minutes; IQR, 59-111; P = 0.0309). The prevalence of ABO-incompatible transplantation was lower in PCAR patients (7.5%) than that in non-PCAR patients (22.5%; P = 0.0206).

TABLE 1: Demographic and clinical characteristics

Clinical Outcome of PCAR and Non-PCAR Patients

Allograft survival in patients with and without PCAR is shown in Figure 3. PCAR patients showed significantly lower death-censored allograft survival than that in non-PCAR patients, according to the Kaplan-Meier analysis (log-rank test P < 0.0001).

FIGURE 3: Graft survivals of PCAR and non-PCAR patients. The Kaplan-Meier curves showed death-censored allograft survival time after kidney transplantation of patients with PCAR or without PCAR (non-PCAR).

Multivariate Cox regression analysis was performed to identify risk factors for allograft loss (Table 2). Development of PCAR was associated with allograft loss (HR, 8.03; 95% CI, 3.89-14.80; P < 0.0001), as was deceased donor transplantation (HR, 1.99; 95% CI, 1.20-3.20; P = 0.0086).

TABLE 2: Risk factor of allograft loss according to Cox regression analysis

Onset Time of PCAR

The onset time of PCAR is shown in Figure 4. PCAR developed in the early period after transplantation; the number of cases increased after 1 year and then began to slow down. The median time to rejection onset was 605 days (1.6 years), with a range of 21 days to 18.8 years after transplantation.

FIGURE 4: The cumulative incidence of PCAR. The cumulative incidence of PCAR showed PCAR onset period from transplantation.

Clinical Characteristics of PCAR Patients According to the Banff 2015 Criteria

We divided 40 cases of PCAR into 4 types: 4 were in the borderline change group, 14 in the TCMR group, 9 in the AMR/sAMR group, and 13 in the mixed-type rejection (TCMR/AMR) group. The clinical characteristics of each type are shown in Table 3. There were no significant differences between the groups in age, sex, ABO-incompatible transplantation, prevalence of DSA, and serum creatinine at baseline or at biopsy. A prior history of biopsy-proven rejection was noted in 9 of the 40 PCAR patients, with a higher rate in those with AMR/sAMR (44.4%) than in other types; this was not statistically significant (P = 0.2218). In addition to rejection, interstitial fibrosis/tubular atrophy was diagnosed in 2 cases, recurrence of IgA nephropathy in 2 cases, and recurrence of focal segmental glomerulosclerosis in 1 case before PCAR onset. Duration from transplantation to onset of PCAR was longer in the AMR/sAMR group (median, 1384 days; IQR, 643-3162) than that in the other groups, without statistical significance (P = 0.0807).

TABLE 3: Demographic and clinical characteristics of PCAR among rejection types of Banff 2015 classification

Of the 40 PCAR patients, 35 received treatment for rejection. There were no differences in steroid (either oral or intravenous) or anti–T-cell treatment (including gusperimus, anti-thymoglobulin, and muromonab-CD3) between the PCAR types. Anti–B-cell treatment (including rituximab, plasma exchange, and intravenous immunoglobulin) was used significantly more in the TCMR/AMR groups (84.6%) than in the other groups (P = 0.0092).

Pathologic Characteristics of PCAR Patients According to the Banff 2015 Criteria

The pathological characteristics of each type are shown in Table 4. Interstitial inflammation scores (i) and tubulitis scores (t) were significantly higher in borderline change, TCMR, and TCMR/AMR groups than in the AMR/sAMR group. The g scores, pt, and c4d staining scores were significantly higher in the AMR/sAMR and TCMR/AMR groups than in the borderline change and TCMR groups. Chronic glomerulopathy (cg), arteriolar hyalinosis (ah), hyaline arteriolar thickening (aah), interstitial fibrosis (ci), and tubular atrophy scores (ct) were significantly higher in the AMR/sAMR group than in the other groups. There were no differences in vasculitis (v) or vascular fibrous intimal thickening scores (cv) between the groups.

TABLE 4: Pathological characteristics of PCAR among rejection types of Banff 2015 classification

Seven cases had tertiary lymphoid organs (TLOs) in the cortical interstitial area, but there was no difference in prevalence between the groups (P = 0.4254).

Allograft Survival in PCAR Patients According to the Banff 2015 Criteria

Allograft survival in PCAR patients according to the Banff 2015 criteria is shown in Table 3 and Figure 5. Among all PCAR cases, the 1-year allograft survival was 86.4% and the median allograft survival time was 2478 days. According to classification, the median allograft survival time in the borderline change and TCMR groups was longer than the observation time. The AMR group had a shorter median allograft survival time than the other groups, without a significant difference (P = 0.1661). Kaplan-Meier analysis showed that patients in the AMR/sAMR group had a lower rate of allograft survival than other groups, without a significant difference (log-rank test, P = 0.1692).

FIGURE 5: Graft survival of PCAR among rejection types in Banff classification. The Kaplan-Meier curves showed death-censored allograft survival time after biopsy of patients in each group.

Multivariable Cox regression analysis was performed to identify risk factors associated with allograft survival other than those in the Banff 2015 classification, as shown in Table 4. The analysis showed that serum creatinine level at biopsy was associated with allograft loss (HR, 1.87; 95% CI, 1.07-3.35; P = 0.0274). Recipient age, biopsy time since transplantation, and detection of DSA were not independently associated with allograft loss.

DISCUSSION

The etiology of plasma cell infiltration in allografts is not well-understood. In settings of allograft organ transplantation, B cells receive activating signals by binding with antigens and activated by interacting with CD4+ T cells in germinal center, and finally differentiate into plasma cells or memory B cells.¹¹ Some plasma cells migrate into bone marrow and live long, but usually they did not infiltrate in allograft organs.

Previous clinical observational studies indicated that infiltration of plasma cells in allografts was associated with drug hypersensitivity, infection,¹ or PTLD nephropathy.² Moreover, reflux nephropathy and BKV nephropathy cases also showed plasma cell infiltration in our cohort. After we excluded these, the remaining cases were considered “plasma cell-rich acute rejection (PCAR),” as reported by Charney et al.⁵ The possibility that these excluded cases were falsely considered PCAR should be noted.

Risk factors for developing PCAR have been considered. Among baseline characteristics in our study, deceased donor transplantation, longer total ischemia time, and non–ABO-incompatible transplantation were more frequent in PCAR patients than in non-PCAR patients. Longer duration from transplantation to rejection onset,¹² tapering or withdrawal of immunosuppression,¹³ and nonadherence to immunosuppression^14,15 were also reported to be associated with PCAR. Our cohort included 4 cases of chronic immunosuppression underdosing, and 9 cases of change in immunosuppression regimen before PCAR. We also found recurrent IgA nephropathy in 2 cases, and recurrent focal segmental glomerulosclerosis in 1 case, but could not confirm their association with PCAR.

PCAR showed all histological types of rejection according to the Banff 2015 criteria. Among the 40 PCAR patients, the AMR/sAMR group showed the worst allograft survival. The AMR/sAMR type was associated with longer duration from transplantation to rejection onset, greater prior history of acute rejection, greater prevalence of DSA, and higher ci and ct scores. These factors might result in a worse outcome. We also examined the prevalence of TLOs, which often develop at sites of transplant allografts and are considered germinal center reactions, resulting in anti-HLA-producing plasma cells and memory B cells.¹⁶ TLOs have also reportedly been associated with allograft tolerance,^17,18 but showed no association with allograft survival in our study.

Among our PCAR cases, 3 had intimal arteritis. Wu et al¹⁹ reported that intimal arteritis could be found in all types of rejection, and that intimal arteritis associated with AMR had worse outcomes. However, we did not find an association between intimal arteritis (v score) and PCAR cases.

The prognosis of PCAR was poor in our study. The development of PCAR was a risk factor for allograft loss regardless of background status (HR, 8.03; 95% CI, 3.89-14.80; P < 0.0001). Meehan et al² reported a 1-year allograft survival rate in PCAR of 56%, and Desvaux et al³ reported a 40% rate among those refractory to immunosuppression therapy. The reason for better allograft survival in our study was not clear, but it is possible that the exclusion criteria were different and that more intensive treatment in our study improved the prognosis of PCAR.

Allograft survival differed from that reported in the Banff classification and the AMR/sAMR group showed the worst prognosis, without a statistical difference (log-rank test, P = 0.1692). Abbas et al²⁰ reported that 64% of PCAR cases had detectable DSA and cases with DSA had worse allograft survival. Detectable DSA was not an independent risk factor for allograft loss in our cohort, but DSA at the time of rejection was only assessed in 60% of cases. The onset time of acute rejection has been reported to affect the clinical outcome, regardless of the Banff classification,²¹ but late onset was not an independent risk factor for allograft failure in our cohort.

Various treatment regimens were used for PCAR because this was a retrospective cohort study. Most of the cases were treated with more than 1 therapy. Many physicians chose therapies on the basis of the clinical and histological findings in each case, depending on the presence of DSA or the Banff classification score.

There were several limitations. This was a single-center, retrospective study, the sample size (n = 40) was small, the PCAR definition has not been confirmed, the treatments differed in each case, and DSA at the time of rejection was assessed in only 60% of cases.

In conclusion, PCAR was an independent risk factor for allograft loss. PCAR showed all types of rejection according to the Banff 2015 classification, and the AMR/sAMR group was associated with poor allograft survival.

TABLE 5: Risk factor of allograft loss in patients with PCAR according to Cox regression analysis

ACKNOWLEDGMENTS

The authors would like to express their gratitude to Mr. Shigeru Horita, Mr. Hideki Nakayama, Ms. Mayuko Ono (Department of Pathology, Tokyo Women’s Medical University, Tokyo, Japan) for their excellent technical assistances, Dr. Junki Koike (Department of Pathology, Kawasaki City Tama Hospital, Kanagawa, Japan), Eri Imai (Department of Nephrology, Itabashi Chuo Hospital, Tokyo, Japan), Junichi Hoshino (Department of Nephrology, Toranomon Hospital, Tokyo, Japan) and Koki Mise (Nephrology, Diabetology and Endocrinology Department, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama, Japan) for their valuable suggestions, and all staffs in department of Nephrology, Tokyo Metropolitan Health and Medical Treatment Corporation Okubo Hospital (Tokyo, Japan) for their clinical support and variable suggestion in this study. The authors also would like to thank Editage (www.editage.jp) for English language editing.