Polyomavirus Allograft Nephropathy: Sequential Assessment of Histologic Viral Load, Tubulitis, and Graft Function Following Changes in Immunosuppression

Authors


*Corresponding author: Parmjeet S. Randhawa, randhawapa@msx.upmc.edu

Abstract

Our initial cases of polyoma virus allograft nephropathy (PVAN) received pulse steroids due to anxiety about concomitant acute rejection triggered by the presence of tubulitis. However, our current policy is to reduce immunosuppression in all cases. The aim of this study was to determine whether clinical follow-up in these patient categories shows any differences in: (a) histologic viral load, (b) grade of tubulitis, and (c) graft function. Reduced viral load assessed within 8 weeks was seen in 4/20 (20.0%) biopsies treated initially by increased immunosuppression, compared to 15/19 (83.3%) biopsies treated with reduced immunosuppression (p = 0.001, Fisher's exact test). Yet, >70% reversal of the rise in serum creatinine occurred in only 3/19 (15.8%) and 1/19 (5.3%) patients, respectively, in these two groups. Improved tubulitis was seen in 11/20 (55%) of biopsies treated with steroids, despite the lack of beneficial effect on serum creatinine in 12/19 (63.1%) instances. In biopsies not treated with any change in immunosuppression, the serum creatinine remained stable in 1/5 (20%) and worsened in 4/5 (80%) biopsies. These data demonstrate that in biopsies with PVAN and tubulitis, reduced immunosuppression is more effective in lowering viral load than steroid therapy. Lack of parallelism between viral load, tubulitis grade, and serum creatinine illustrates a complex interplay of viral and alloimmune factors leading to graft injury.

Introduction

Polyomavirus nephropathy (PVAN) is an increasingly recognized complication in kidney transplant recipients (1–5). The histology is characterized by viral inclusions, interstitial inflammation, and tubulitis. The significance of tubulitis in these cases has been a subject of some debate. Most clinicians accept it as a part of viral-induced parenchymal injury, and reduce immunosuppression after a diagnosis of PVAN. However, others contend that tubulitis is a manifestation of concurrent acute cellular rejection, and advocate increased immunosuppression (6). In an attempt to clarify this controversy, we performed a study to determine the effect of changes in immunosuppression on: (a) histologic viral load, (b) grade of tubulitis, and (c) serum creatinine as a marker of graft function. The series presented includes our initial cases, who were briefly treated with steroids because of certainty about concomitant acute rejection, and our subsequent cases treated by reduced immunosuppression from the outset, despite the presence of concomitant tubulitis.

Materials and Methods

The diagnosis of PVAN was provisionally made by histologic documentation of polyomavirus inclusions in the tubular epithelium accompanied by varying degrees of interstitial inflammation. The diagnosis was confirmed in all cases by immunohistochemistry or in-situ hybridization, as previously reported (2). Viral infection localized to the nuclei in the tubular epithelium. Both proximal and distal tubules were affected. The number of infected cells varied from 1 to 10 per tubular cross-section, with 1–70% of the sampled renal parenchyma showing evidence of viral cytopathic effect. All biopsies were reviewed in detail by a pathologist and graded for tubulitis and interstitial inflammation using Banff 1997 criteria (7). Changes in these histologic parameters were assessed both as Banff scores, and, when the score was numerically unchanged, by a ‘gestalt approach’. The viral load, assessed by light microscopy, immunohistochemistry, and in-situ hybridization, was graded semi-quantitatively as grade 0, 1 + (less than 5 infected cells), 2 + (5–15 infected cells), or 3 + (>15 infected cells). Follow-up biopsies were recorded as complete viral clearance, improved, unchanged, or worse compared to the index specimen. In 12 biopsies the tissue viral load was also determined using a published real-time quantitative PCR assay (8). Clinical data were abstracted from the medical records. For each biopsy studied, we obtained a baseline creatinine (4 weeks prior to biopsy), a peak serum creatinine, a post-therapy creatinine (2 weeks after therapy), and the most recent serum creatinine, to assess the current graft status. Changes in immunosuppression in relation to each biopsy were individually analyzed. If these changes led to >70% reversal of the rise in serum creatinine, the biopsy was classified as ‘complete response’. Graft dysfunction episodes with 30–70% reversal qualified for ‘partial response’, <30% reversal as ‘stable creatinine’, and rise in creatinine >30% as ‘worse’. A Banff score of t0 was designated as ‘complete response’ with regard to tubulitis. This study was approved by the University of Pittsburgh Institutional Review Board (IRB protocol # 000622).

Results

The 66 biopsies studied here were obtained from 31 patients with polyoma virus allograft nephropathy. There were 12 females and 19 males, with a mean age of 47 ± 13 years (range 21–72, median 50) years. The native kidney diseases that led to transplantation were diabetes mellitus (9/31, 29%), chronic glomerulonephritis (5/31, 16.1%), hypertension (6/31, 19.1%), focal segmental glomerulosclerosis (1/31, 3.2%), Alport's syndrome (2/31, 6.5%), Henoch Schonlein purpura (1/31, 3.2%), dysplastic kidney (2/31, 6.5%), polycystic kidney (1/31, 3.2%), reflux (2/31, 6.5%), and systemic lupus erythematosus (2/31, 6.5%). The time of onset of viral nephropathy was 11 ± 6.8 months (range 2–32, median 11) months after transplantation. The mean and median baseline serum creatinine prior to the diagnosis of PVAN was 2.0 ± 1 mg/dL and 2.6 mg/dL, respectively. At the time viral nephropathy was diagnosed, the mean and median serum creatinine were 2.6 ± 1 mg/dL and 2.6 mg/dL, respectively. Eleven patients (35.5%) lost their grafts, five underwent allograft nephrectomy, and one was re-transplanted. Chronic allograft nephropathy was the primary cause of graft loss in all cases. However, persistent viral DNA and antigens indicative of ongoing viral nephropathy were also demonstrable in five patients. One patient showed no viral DNA or antigens in the tissue, but viruria could be detected by PCR. The probability of graft survival 1 year after diagnosis was 53.8%. Excluding patients whose grafts had failed, the most recent serum creatinine, measured 95.4 ± 102 weeks after the diagnosis of PVAN, was 3.6 ± 1.5 mg/dL.

Tables 1 and 2 summarize data on clinico-pathologic correlations presented in 3 categories, depending on whether the diagnosis of PVAN led to an initial increase, decrease, or no change in overall immunosuppression. Table 1 focuses on 45 biopsies, each of which was followed by a repeat biopsy obtained within 8 weeks. In this group, 20 biopsies were treated with increased immunosuppression, 19 biopsies with decreased immunosuppression, and 6 biopsies with no change in immunosuppression. Table 2 presents an analysis of 21 biopsies with regard to long-term changes in viral load, tubulitis, and serum creatinine. In this group, 11 biopsies were treated with increased immunosuppression, and 10 biopsies with decreased immunosuppression. The smaller number of data points in this table reflects the lack of availability of a follow-up sample for histologic examination.

Table 1.  Changes in immunosuppression correlated with short-term effects on viral load, tubulitis, and serum creatinine
 Initial increaseDecreaseNo change
 CRPRNRCRPRNRCRPRNR
 StableWorse StableWorse StableWorse
  1. *p = 0.001 CR + PR Increased vs. Decreased Immunosuppression (Fisher's Exact Test).

  2. Abbreviations: CR-complete response; PR-partial response; NR-no response. Parentheses indicate per cent values.

Viral load3/20 [15]*1/20 [5]*7/20 [35]9/20 [45]7/19 [41.2]8/19 [42.1]0/19 [0]4/19 [21.1]1/6 [16.7]2/6 [33.3]03/6 [50]
Tubulitis2/20 [10]9/20 [45]5/20 [25]4/20 [20]05/19 [26.3]4/19 [21.1]10/19 [52.6]01/6 [16.7]05/6 [83.3]
Creatinine3/19 [15.8]4/19 [21.1]1/19 [5.3]11/19 [57.9]1/19 [5.3]3/19 [5.8]2/19 [10.5]13/19 [68.4]001/5 [20]4/5 [80]
Table 2.  Long-term effects on viral load, tubulitis, and serum creatinine
 Initial increase in immunosuppressionDecrease in immunosuppression
 CRPRNRCRPRNR
 StableWorse StableWorse
  1. No statistically differences were found amongst different biopsy categories.

  2. Abbreviations: CR-complete response; PR-partial response; NR-non response. Parentheses indicate per cent values.

Viral load8/11 [72.8]2/11 [18.2]0/11 [0]1/11 [9.1]7/10 [70]2/10 [20]0/10 [0]1/10 [10]
Tubulitis3/11 [27.3]6/11 [54.5]1/11 [9.1]1/11 [9.1]0/10 [0]5/10 [50]2/10 [20]3/10 [30]
Creatinine3/10 [30]2/10 [20]0/10 [0]5/10 [50]0/9 [0]2/9 [22.2]1/9 [11.1]6/9 [66.6]

PVAN biopsies treated with initial transient increase in immunosuppression

The increased immunosuppression in these cases was temporary, and was prompted by the presence of tubulitis, which is conventionally regarded as a marker of acute cellular rejection in the allograft kidney. In an attempt to separate the short-term consequences of steroid therapy and the subsequent long-term effects of reduced immunosuppression (after steroid therapy had failed), data were separately analyzed for follow-up biopsies performed <8 weeks (Table 1, n = 20) and >8 weeks (Table 2, n = 11) after the first diagnosis of PVAN.

In biopsies performed within 8 weeks of diagnosis (Table 1), serum creatinine showed a complete therapeutic response in 3/19 (15.8%) episodes of renal dysfunction, a partial response in 4/19 (21.1%) episodes, and no response in 12/19 (63.2%) patients. The ‘no response’ category included one patient in whom creatinine remained stable, and 11 patients in whom the serum creatinine worsened after initial therapy. Changes in viral load generally paralleled the trends in serum creatinine. Thus, a decreased viral load (partial response) was seen in 1/20 (5%), and clearance of virus (complete response) in an additional group of 3/20 (15%) samples. The percentage of biopsies showing viral clearance or decreased viral load was significantly lower in patients treated with an initial increase in immunosuppression, compared with patients whose immunosuppression was reduced at the outset (20% vs. 83.3%, p = 0.004, Fisher's exact test). Partial or complete resolution of tubulitis was observed in a higher proportion of cases treated with an initial increase in immunosuppression, compared to biopsies where immunosuppression was reduced at the outset (55% vs. 26.3%). However, an associated improvement in serum creatinine (complete or partial response) was seen in only 7/19 (36.8%) biopsies, and it was of a transient nature.

In biopsies performed beyond 8 weeks of initial diagnosis, after reduction of immunosuppression (Table 2), serum creatinine showed a complete therapeutic response in 3/10 (30.0%) episodes of renal dysfunction, partial response in 2/10 (20%) episodes, and deterioration in the remaining 5/10 (50.0%) episodes. As a rule, the viral load was substantially improved. Thus, decreased viral load occurred in 2/11 (18.2%), and complete clearance of virus in 8/11 (72.8%) of samples. The latter group included all five biopsies with complete or partial creatinine response. Tubulitis showed complete resolution in 3/11 (27.3%) and partial improvement in 6/11 (54.5%).

PVAN biopsies treated with no change in immunosuppression (Table 1, n = 6)

All of these biopsies were performed within 8 weeks of the initial diagnosis of PVAN to assess the progression of viral nephropathy (Table 1). Serum creatinine was stable in 1/5 (20.0%), and worse in 4/5 (80%) cases. No data were available for one biopsy. The grade of tubulitis improved in 1/6 (16.7%), and worsened in 5/6 (83.3%) biopsies. Decreased viral load was demonstrable in 2/6 (33.3%) of follow-up biopsies, while 1/6 (16.7%) biopsies showed complete viral clearance. The remaining 3/6 (50%) samples showed increased viral cytopathic effect in the tubular epithelium.

PVAN biopsies treated with decreased immunosuppression from the outset

Serum creatinine measured within 8 weeks of biopsy showed: (a) complete therapeutic response in 1/19 (5.3%) episodes of renal dysfunction, (b) partial response in 3/19 (15.8%) episodes, (c) stable values in 2/19 (10.5%), and (d) progressive rise in the remaining 13/19 (68.4%) episodes (Table 1, n = 19). There was decreased viral load in 8/19 (42.1%) and complete clearance of virus in 7/19 (41.2%) samples.

In biopsies performed >8 weeks after the index specimen (Table 2, n = 10), serum creatinine showed: (a) complete therapeutic response in 0/9 (0%) episodes of renal dysfunction, (b) partial response in 2/9 (22.2%) episodes, (c) stable values in 1/9 (11.1%), and (d) progressive rise in the remaining 6/9 (66.6%) episodes. Decreased viral load was demonstrable in 2/10 (20.0%) and complete clearance of virus in 7/10 (70.0%) of samples. Tubulitis showed improvement in 5/10 (50.0%) of follow-up biopsies. The lack of improvement or worsening of tubulitis in half of the patients, despite reduction in viral load, suggests that reduction of immunosuppression was complicated by the development of irreversible viral or alloimmune injury to the graft.

In 12 biopsies, we correlated changes in immunosuppression with intra-tissue viral concentrations measured by real-time quantitative PCR. Four biopsies were treated by increased immunosuppression, and in 2 instances, this led to increased viral load in a follow-up biopsy. The remaining 2 follow-up biopsies showed no significant change in viral load, within the limits of assay precision. Seven biopsies were treated with decreased immunosuppression, and 6 of these showed decreased viral load, while the seventh biopsy showed no significant change in tissue viral concentration. One biopsy was not treated by any change in immunosuppression, and in this case, a follow-up biopsy showed increased viral load.

Discussion

PVAN is a complication of excessive immunosuppression in kidney transplant patients. Most clinicians reduce the dose of immunosuppressive drugs after the diagnosis is made, but some have advocated a brief course of steroids if tubulitis is present. The purpose of this study was to determine the effect of changes in immunosuppression on: (a) histologic viral load, (b) grade of tubulitis, and (c) serum creatinine as a marker of graft function. Clinical response as assessed by serum creatinine did not show statistically significant differences between the patients treated with an initial increase or decrease in immunosuppression. Thus, complete reversal of serum creatinine within 8 weeks of biopsy was seen in 3/19 (15.8%) of biopsies treated by increased immunosuppression, and in 1/19 (5.3%) of biopsies treated with decrease in immunosuppression (Table 1). However, the majority of biopsies treated with steroids showed either no change (1/19 = 5.3%), or actual worsening (11/19 = 57.9%) of serum creatinine. Even when improvement in serum creatinine occurred following steroid therapy, it was of a transient nature.

Biopsies treated with no change in immunosuppression showed a persistently high viral load in 3/6 (50.0%) of the cases (Table 1). Increased immunosuppression led to persistent or worsened viral cytopathic effect in 16/20 (80%) of biopsies within 8 weeks. In contrast, reduction of immunosuppression led to viral clearance in 7/10 (70%) of biopsies after long-term follow-up (Table 2). These observations form the basis of our current policy to reduce the immunosuppression, whenever a diagnosis of PVAN is made. Nonetheless, the clinical course of individual patients is variable, and reduction of viral load did not always translate into improved graft function, probably due to irreversible chronic allograft nephropathy (9), and perhaps the confounding effect of alloimmune injury. Long-term follow-up did not show any statistically significant differences between patients who did or did not receive a brief initial course of steroid therapy (Table 2). It is also worth mentioning that viral clearance was observed even in a few patients getting no therapy or steroid therapy for presumed concomitant rejection.

A somewhat paradoxical observation made during this study is that partial or complete resolution of tubulitis was observed in 11/20 biopsies treated with initial increase in immunosuppression (Table 1). This suggests that tubulitis in these biopsies was, at least partially, related to an associated component of alloimmune injury. In this context, it is pertinent to recall that Hirsch et al. have recently reported 4 cases of PVAN with concurrent acute cellular rejection, which responded to steroid therapy (10). These patients had been diagnosed relatively early during the course of a study which mandated prospective monitoring of urine and blood for BKV infection. In contrast, Limaye and colleagues have documented a patient with acute cellular rejection and BKV viremia, where the administration of steroids led to a fall in serum creatinine, but a progressive increase in viral load, culminating in viral nephropathy (11). This variable response of PVAN to steroid therapy highlights our lack of complete understanding of all the clinical variables that affect viral replication in the allograft kidney. We postulate that host cellular and humoral immunity play an important role in determining ultimate clinical outcome in individual patients. Further exploration of this notion would require the development of immunologic assays directed against specific viral antigens.

Acknowledgments

Betul Celik was supported by Turkish Ministry of Health Grant # 6436–07/17/01. Parmjeet S. Randhawa was supported by NIH grant R21 AI51227-01. Abhay Vats was supported by the National Kidney Foundation of Western Pennsylvania.

Ancillary