Polyomavirus Infection of Renal Allograft Recipients

Histology and Urine Cytology

Morphologic changes in tissue samples (25 biopsies and two graft nephrectomies) from five patients who contracted PV disease were studied. PV disease is defined as a morphologically manifest renal infection with cytopathic signs accompanied by varying degrees of interstitial inflammatory cell infiltrates and functional impairment. Thirteen tissue samples (11 biopsies and two graft nephrectomies) were obtained during the course of persistent PV disease (up to 330 d after the initial diagnosis). Fourteen biopsies were taken before the initial manifestation, i.e., before the first histologic diagnosis of PV infection (including negative workup by immunohistochemistry [IHC]; see below).

For comparison, 108 renal allograft biopsies from patients lacking any morphologic evidence of PV disease were evaluated (including negative workup by IHC; see below). All of the 133 biopsies were performed due to deterioration of graft function at similar time intervals posttransplantation. Thirteen distinct histologic criteria were analyzed and semiquantitatively graded on a scale of 1 (minimal change) to 4 (marked change): extent of mononuclear cell infiltrate in the interstitium and tubulitis; vascular rejection (VR) with necrosis; transplant endarteritis; sclerosing VR (“chronic vascular rejection”); transplant glomerulitis; transplant glomerulopathy; extent of interstitial fibrosis; ischemic tubular damage; cyclosporine-induced tubular damage; toxic arteriolopathy; arteriolosclerosis; glomerulonephritis (immune complex-mediated); and pyelonephritis. A diagnosis of acute renal allograft rejection was based on the criteria of the Cooperative Clinical Trials in Transplantation classification scheme (27,28). The histologic evaluation was performed according to standard protocols by light microscopy (hematoxylin and eosin, periodic acid-Schiff, trichrome, methenamine silver stains) accompanied by IHC and electron microscopy. For routine IHC, a standard panel of antibodies was used (directed against IgG, IgM, IgA, complement factors C3, C4, C5b-9, C1q, and fibrin).

Voided urine samples were analyzed at time of biopsies as a standard patient management procedure. Urine cytology reports were retrospectively evaluated from 483 renal transplant recipients (including five patients with PV disease) managed in Basel between 1985 and 1997. The presence of PV-infected intranuclear inclusion cells, so-called “decoy cells,” as a morphologic marker for viral replication was semiquantitatively recorded as absent, scant (1 to 4 per 10 high-power fields [HPF]), or abundant (>5 per 10 HPF). From all patients with abundant decoy cell excretion (>5 per 10 HPF in at least one urine sample), renal allograft biopsies were (re-) examined (including IHC; see below) to search for cytopathic signs of a PV graft infection.

Immunohistochemical Detection of Viruses

PV was detected by indirect IHC on urine cytology, and tissue samples (fresh frozen and formalin-fixed). Several mouse monoclonal antibodies (mAb) were used to detect different PV strains: (1) mAb detecting both BK and JC strains (Readysysteme, Bad Zurzach, Switzerland); (2) mAb detecting JC strain (Readysysteme); (3) mAb detecting the SV40 large T antigen (Calbiochem/Oncogene Research Products, Cambridge, MA). The latter antibody reacts with primate (SV40) and human PV (BK and JC strains) and works after formalin fixation, using the microwave technique for antigen retrieval (Tris-HCl buffer, pH 10.5, 15 min at 100°C, followed by overnight incubation with the primary antibody at 4°C, dilution 1:3200). All staining procedures were performed according to a previously described protocol with the avidin-biotin-peroxidase technique (29). DAB (diaminobenzidine tetrahydrochloride) and AEC (3-amino,9-ethylcarbazole) were used as chromogens. All biopsies were also analyzed for the presence of CMV, Epstein-Barr virus (EBV), herpes simplex virus (HSV) or adenovirus antigens (CMV [IEA]: mouse monoclonal antibody, Biosoft, Varilhes, France; CMV [EA] and EBV [LMP, ZEBRA, EBNA]: mouse monoclonal antibodies, Dako, Glostrup, Denmark; HSV [types 1 and 2]: rabbit polyclonal antibody, Dako, Glostrup, Denmark; adenovirus: mouse monoclonal antibody, Chemicon, Temecula, CA).

PCR Studies

The genome detection of polyomavirus JC strain and BK strain used a modified semi-nested format based on a previous report (30). The outer primers 5′-AAG TCT TTA GGG TCT TCT AC-3′ and 5′-GTG CCA ACC TAT GGA ACA GA-3′ amplified a fragment of 176 bp common to both viruses. The inner primer pair combined the common 5′-AAG TCT TTA GGG TCT TCT AC-3′ with either the BK-specific primer 5′-GAG TCC TGG TGG AGT TCC-3′ to obtain a BK fragment of 149 bp, or the JC-specific primer 5′-GAA TCC TGG TGG AAT ACA-3′ yielding a JC fragment of 146 bp. Both were confirmed by dideoxy sequencing (H.H.H., personal observation). DNA was prepared by protease K digestion and purification on silica spin columns (QIAmp, Qiagen, Basel, Switzerland) from the following sources: one nephrectomy from a patient with manifest PV disease and voided urine from five patients with PV disease.

A total of 0.5 μg of DNA was used per reaction. Amplification was performed after preincubation (37°C, 10 min) for uracyl-N-glycosylase decontamination followed by 94°C at 30 s, 50°C at 45 s, and 72°C at 60 s (30 cycles, for both the outer and the inner amplification). Ten-microliter aliquots of the inner PCR product were separated on NuSieve:Agarose (3%:1%) electrophoresis gels. The threshold of detection was determined to be five genomes. Two parallel runs were analyzed for each result, and if equivocal, repeated. The results were interpreted as positive (2 of 2 runs positive); weakly positive (1 of 2 runs positive, viral load close to detection threshold); negative (2 of 2 runs negative).

PV Disease

Patients. Five patients (mean age at transplantation, 46 ± 2 yr; range, 27 to 57; four men and one woman) showed a morphologically manifest renal allograft infection due to polyomavirus (BK strain). PV infection became manifest and was histologically diagnosed several months after transplantation (9 ± 2 mo, mean ± SD; range, 7 to 11; all between October 1996 and June 1997). All biopsies before the initial diagnosis of PV disease (n = 14) lacked cytopathic signs of viral infection by light microscopy or IHC. Three patients had received a cadaver kidney, and two received organs from living unrelated donors. All of the five patients had a complicated posttransplantational course with recurrent rejection episodes. Before a manifest infection was diagnosed, immunosuppression had been switched in all patients from cyclosporine to high-dose tacrolimus rescue therapy for 5 ± 3 mo (mean ± SD) (Figure 1). Patients were kept on tacrolimus; three patients were switched back to cyclosporine 98 ± 46 d (mean ± SD; range, 49 to 140 d) after the initial diagnosis of PV disease (Figure 1).

Cytopathic Changes. PV disease was diagnosed histologically. The morphologic characterization was based on light microscopy, immunohistochemistry and electron microscopy. The diagnostic hallmark was the detection of intranuclear viral inclusions, which were exclusively found in epithelial cells (with the exception of podocytes). All tissue samples (n = 13) from patients with PV disease (n = 5) showed viral inclusions. Three types of intranuclear inclusion bodies were seen by light microscopy with similar frequencies along the entire nephron (Figure 2): (1) a homogeneous ground-glass appearance; (2) a central eosinophilic granular appearance with a (mostly) incomplete halo; (3) a more homogeneous, finely granular pattern. Cells with viral changes were often (but not always) enlarged with polymorphic nuclei. Frequently, tubular cells were rounded-up, necrotic, and extruded from the epithelial cell layer into tubular lumens causing marked denudation of basement membranes (Figure 3). Although cytopathic signs were seen along the entire nephron, they were often abundant in distal tubular segments and collecting ducts. Sporadically, infected cells were noted in the parietal epithelium lining Bowman's capsule (Figure 4). By IHC, PV (BK strain) was demonstrated (Figure 5; nuclear staining profile: AB detecting SV40 - positive; AB detecting BK/JC - positive; AB detecting JC - negative). Positive staining was also noted in tubular cells inconspicuous by light microscopy. IHC did not reveal PV in nonepithelial cells and podocytes. By electron microscopy, intranuclear viral particles were granular, sometimes arranged as crystalloids (Figure 6). They measured between 35 and 40 nm in diameter, a size characteristic for PV (4,7,12,18). Viral particles were occasionally detected adherent to the external surface of cell membranes (Figure 7). In the cytoplasm, viral particles were only rarely noted by electron microscopy (Figure 7) or IHC. Ultrastructurally, PV was not seen in nonepithelial cells and podocytes. In three biopsies (two patients), small glomerular crescents with inclusion-bearing cells were seen (Figure 8). Crescent formation affected between 10 and 20% of glomeruli. There was no evidence of an immune complex-mediated glomerulonephritis. In the renal pelvis and ureters (two nephrectomies), viral inclusion bodies were identified in superficial transitional cells by light microscopy and additionally in the basal cell layer by IHC. Transitional cells almost exclusively showed intranuclear inclusions of the ground-glass type (in contrast to nephrons), closely resembling decoy cells in urine cytology (Figure 9).

CMV, EBV, HSV, or adenovirus were not found. By standard immunohistochemistry using a routine panel of antibodies directed against immunoglobulins and complement factors, a specific staining pattern was not present.

Identification of BK Virus by PCR (Tissue, Urine). One nephrectomy specimen (Figure 10) and urine samples from five patients were studied by PCR. All examined samples revealed abundant amplicons of 149 bp indicative of BK virus genomes. JC virus was not detected.

PV Infection and Interstitial Inflammation. In seven biopsies (7 of 13; 54%), PV infection of tubular epithelial cells was associated with widespread inflammatory cell infiltrates in the interstitium and tubulitis. Inflammatory infiltrates were predominately composed of lymphocytes and histiocytes in combination with some polymorphonuclear leukocytes and plasma cells. Focally, plasma cells were abundant. In four biopsies (4 of 13; 31%), the infiltrate was sparse with minute tubulitis (≤1 nonatrophic tubular cross section involved). The infiltrating cells were randomly distributed and not primarily associated with virally affected tubules (Figure 11a). Mononuclear cell infiltrates and tubulitis were often noted in areas lacking viral inclusion bodies. Some infected tubules, especially in the medulla, did not show an adjacent inflammatory response (Figure 11b). In two biopsies (2 of 13; 15%), there was negligible interstitial inflammation.

Morphology of Persistent PV Disease. After the initial histologic diagnosis of PV disease had been made, eight subsequent tissue samples were examined during the course of persistent graft infection (range, 40 to 330 d postdiagnosis; six biopsies and two graft nephrectomies; five patients). All tissue samples showed the three types of viral inclusion bodies, severe focal tubular injury, and varying degrees of inflammatory cell infiltrates (see above). Chronic changes increased: (1) marked interstitial fibrosis (three patients; six samples); (2) marked arteriolosclerosis (one patient, one sample); (3) marked “cyclosporine-type” arteriolopathy (two patients, two samples); (4) intimal fibrosis (“chronic vascular rejection;” two patients, two samples).

In one patient, PV-associated crescent formation was observed in two subsequent biopsies (50 d and 330 d after the initial diagnosis of PV disease). The percentage of crescents remained unchanged over time (<20%); however, they became increasingly fibrotic with pronounced layering of the basement membrane of Bowman's capsule associated by an inflammatory cell infiltrate (Figure 12).

The gross appearance of two graft nephrectomies revealed kidneys slightly reduced in size with a smooth, focally finely granular outer surface. The parenchyma was homogenously ochre brown; the corticomedullary junction was ill-defined.

Decoy Cells in the Urine. Urine samples from all patients with persistent PV disease revealed decoy cells with ground-glass type intranuclear inclusions (Figure 9) positive for PV by IHC and electron microscopy. Decoy cells were typically excreted in large numbers. Small numbers (<5 per 10 HPF) were only sporadically noted during the course of disease. The excretion of decoy cells vanished in one patient after transplant nephrectomy (starting 2 d postoperatively; follow-up: 4 mo). In two patients tested, urine samples revealed large numbers of decoy cells 10 and 5 mo before manifestation of PV disease when the corresponding biopsies lacked cytopathic signs (see above).

Differences between PV Disease Group and Control Subjects

Decoy cell excretion was a typical finding in patients with PV disease. To evaluate the specificity of decoy cell excretion, urine cytologies from 483 renal allograft recipients (including five patients with PV disease) were analyzed. Abundant decoy cells were found in 28 recipients (6%) and scant decoy cells in 72 (15%). PV disease was associated with abundant decoy cell excretion in five of 28 (18%) recipients; twenty-three of twenty-eight (82%) were free of disease (no cytopathic changes in graft biopsies by light microscopy or IHC).

Tissue samples from patients with and without PV disease (n = 27 versus n = 108) differed in only two histologic parameters from 13 analyzed (P < 0.05, Fisher exact test) (Table 1): Interstitial cellular rejection and transplant glomerulitis were more frequently found in patients with PV disease (60 and 29%) than without (35 and 4%, respectively). Interstitial rejection and transplant glomerulitis were particularly frequent in the disease group before PV infection became manifest.

Outcome

Renal function in patients with PV disease deteriorated quite rapidly. Two patients (2 of 5; 40%) lost their grafts within 40 and 113 d postdiagnosis of infection. The serum creatinine level of the remaining three patients was 484 ± 326 μmol/L (mean ± SD; 11 mo posttransplantation). In comparison, the 23 transplant recipients with abundant decoy cells in urine cytology but without PV disease had a mean serum creatinine level of 207 ± 177 μmol/L (mean ± SD; 57 mo posttransplantation; correlation with disease group P = 0.01, t test). Two of the latter 23 patients (9%) had lost their grafts (correlation with disease group P < 0.01, χ² test).