Presence of Urinary Haufen Accurately Predicts Polyomavirus Nephropathy

Haufen: Characteristic Features

In urine samples, Haufen were defined as discrete, tightly clustered, cast-like aggregates of a minimum of six polyomaviruses with an unequivocal three-dimensional architecture (Figure 1, A and B). Ultrastructurally, polyomaviruses were easily identified by their characteristic regular capsid structure and a size of approximately 40 to 45 nm in diameter. By definition, Haufen did not contain cell membrane fragments or debris as major core components. Single polyomaviruses and two-dimensional flat groups of virions, sometimes arranged in small sheets, were not classified as Haufen and were not further analyzed. Urine samples categorized as “Haufen-positive” generally contained multiple small and large polyomavirus aggregates that occasionally reached sizes >100 virions. Haufen were typically identified within the first 10 min of grid examination, along with varying numbers of free, nonaggregated polyomaviruses in the background. Three of three Haufen-positive urine samples that were studied by immunogold labeling demonstrated a significant Tamm-Horsfall protein content in the polyomavirus aggregates (Figure 1C).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

(A through C) Three-dimensional cast-like polyomavirus aggregates, termed Haufen, in voided urine samples. (A) A small Haufen (nine virions). (B) A large Haufen (>100 virions). (C) Haufen are rich in Tamm-Horsfall protein (marked by the black dots representing gold particles). (A and B, negative staining EM, uranyl acetate counterstaining; C, immunogold labeling EM with a mouse monoclonal anti-human Tamm-Horsfall antibody, 5-nm gold particles). (D and E) BKN shows diffuse, virally induced acute tubular injury and a predominantly mononuclear inflammatory cell infiltrate in the interstitium. (E) A high-power view of a tubular cross-section with severe, virally induced epithelial cell injury and host cell lysis (long arrows); the arrowheads mark the tubular basement membrane. *Intratubular proteinaceous material rich in Tamm-Horsfall protein. (D and E: hematoxylin- and eosin-stained, formalin-fixed, and paraffin-embedded tissue sections). (F through I) Intratubular release of mature daughter virions subsequent to BKV replication in epithelial cells, a hallmark of BKN. (F and G) Immunohistochemical incubations to detect polyomavirus capsid proteins demonstrate abundant virions in tubular lumens (T). The viruses form small and large aggregates, especially in areas of host cell lysis; the arrowheads mark the tubular basement membranes. (G) A small intratubular (T) intensely staining Haufen (boxed area). (F and G, immunohistochemical incubations with a monoclonal mouse anti-polyomavirus antibody directed against the VP1 capsid protein). (H and I) EM illustrates a tubular cross-section with a sloughed epithelial cell undergoing lysis subsequent to the replication of polyomaviruses (long arrow; compare with E and F). Daughter virions are released into the tubular lumen (T) and are located as aggregates (boxed areas) in a microenvironment rich in Tamm-Horsfall protein. The arrowheads in H mark the tubular basement membrane. (H and I, uranyl acetate staining). Magnifications: ×125,000 in A through C; ×25 in D; ×150 in E; ×120 in F and G; ×2500 in H; ×15,000 in I.

Renal biopsies with BKN showed dense polyomavirus aggregates in injured renal tubular segments (Figure 1, D through I). These intratubular polyomavirus aggregates seemed to be washed out of the affected nephrons as “Haufen” (Figure 1G).

Interobserver Reproducibility/Agreement Rate of Haufen Detection

Thirty-eight of 40 representative urine samples (20 Haufen-positive and 20 Haufen-negative) re-analyzed by a second investigator (V.M.) in a so-called “blinded manner” were accurately reclassified (κ = 0.90; concordance rate 94.6%).

Effects of Long-Term Storage on Haufen

Eighteen Haufen-positive urine samples fixed in 2% paraformaldehyde were reexamined after 3 to 15 mo of storage at 4°C. All 18 samples demonstrated well-preserved polyomaviruses and typical Haufen without apparent ultrastructural changes.

Haufen in Urine Samples and Concomitant Signs of BKV Replication

Pre-BKN (from Time of Transplantation until 2 Wk before Initial Histologic Diagnosis).

A total of 92% (12 of 13) of urine samples from 10 patients were Haufen-negative. Only 8% (one of 13; patient 3; Figure 2) were Haufen-positive 8 wk before initial diagnosis of BKN. Plasma PCR: a total of 85% (11 of 13) of samples were positive, 15% (two of 13) above threshold (patients 3 and 14; Figure 2). Urine PCR: all samples were positive, 92% (12 of 13) above threshold. Decoy cells: all samples were positive, 77% (10 of 13) above threshold. Biopsies: two of two renal biopsies (patients 10 and 14; Figure 2) were negative for BKN. They were triggered by viremia (750 and 2.7 × 10⁴ BKV copies/ml), viruria (1.7 × 10⁹ and 4.29 × 10⁷ BKV copies/ml), and decoy cell shedding above threshold (>50 decoy cells per ThinPrep slide; Tables 1 and 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

The dynamics of Haufen shedding over time in voided urine samples from patients with BKN. Illustrated are 21 study group patients with biopsy-proven BKN; in 18 patients, plasma and urine samples were available at time of the initial diagnostic biopsy. ▴, Haufen-positive urine samples; ▵, Haufen-negative urine samples; B⁺, renal biopsies with BKN; B⁻, renal biopsies without BKN. Plasma BKV load levels are marked by rectangles and grouped into five categories: solid black rectangle with red circle = ≥1 × 10⁴ BKV copies/ml; solid black rectangle ≥1 × 10³ BKV copies/ml and <1 × 10⁴ BKV copies/ml; gray rectangle = 250 to 999 BKV copies/ml; open rectangle with diagonal line = <250 BKV copies/ml; open rectangle = undetectable BKV copies/ml.

Dynamics of Haufen Shedding

During the pre-BKN time period, samples from 10 patients were studied (Figure 2). Only one of 13 urine samples from one of 10 patients (patient 3) was Haufen-positive, accompanied by above threshold viremia (9.4 × 10⁴ viral copies/ml), viruria (6.4 × 10⁸ viral copies/ml), and decoy cell shedding (>100 decoy cells per ThinPrep slide). A biopsy, however, could not be performed until 8 wk later, when BKN was ultimately diagnosed. Twelve of 13 urine samples from nine of 10 patients collected during the pre-BKN period were Haufen-negative, including three consecutive negative samples from patient 10 with varying concomitant viremia. In five of 10 patients, Haufen-negative samples were obtained 3 to 4 wk before initial histologic diagnosis of BKN. At the time of BKN diagnosis, all 18 patients from whom urine samples were available for analysis were Haufen-positive (Figure 2).

During the follow-up period (3 to 120 wk after BKN diagnosis), samples from 20 of 21 patients were studied. Six of six patients with biopsy-proven persistent BKN and ten additional patients had one to 14 sequential Haufen-positive urine samples. For seven patients, Haufen-positive samples were collected in the setting of BKV plasma loads of <1 × 10³ copies/ml 3 to 48 wk after the initial BKN diagnosis. After one to five consecutive Haufen-positive samples, 13 of 20 patients, including two patients (9 and 13) with biopsy-proven BKN resolution, turned Haufen-negative. Three patients (3, 7, and 9) converted within 3 to 5 wk after initial diagnosis. In patients 2 and 7, conversion occurred in the setting of high viremia (>1 × 10⁴ BKV copies/ml) and viruria (>1 × 10⁷ BKV copies/ml) and in patients 1, 2, and 6, in association with decoy cell shedding above threshold. Once converted, all patients remained Haufen-negative during further follow-up (up to 105 wk) when signs of BKV replication steadily decreased, although never became undetectable in nine of 13 patients. All 139 control patients were Haufen-negative despite marked signs of BKV replication in some individuals (Figure 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

One renal transplant control group patient with marked polyomavirus replication and no BKN. Over months (weeks after transplantation plotted on x axis), this renal allograft recipient presented with very high viremia (▪), viruria (•), and decoy cell counts (total decoy cell numbers per ThinPrep slide given in boxed areas). Two allograft biopsies performed 52 and 96 wk after grafting showed no evidence of BKN (B⁻). Five urine samples were consistently Haufen negative (▵).

Haufen Shedding, BKV Replication, and BKN

The shedding of Haufen was not merely linked to the degree of viruria or viremia (e.g., decoy cell excretion, BKV loads in urine and plasma) (Table 3). Significant differences were found between the proportions of samples with high numbers of decoy cells, BKV gene copies in the urine, and plasma on the one hand and Haufen-positivity on the other hand (all differences P < 0.001; κ between 0.39 and 0.56; fair to modest agreement).

The shedding of Haufen was very tightly linked to intrarenal disease (BKN) (see Table 4 and Concise Methods section for definitions of BKN positivity and negativity; no significant difference between Haufen-positive/negative and BKN-positive/negative groups P = 0.317, κ = 0.98; excellent agreement; concordance rate 99%). The detection of Haufen had a sensitivity of 100%, a specificity of 99%, a positive predictive value (PPV) for BKN of 97%, and a negative predictive value (NPV) of 100%.

In contrast, intrarenal disease (BKN) was not tightly associated with viruria. The distribution of BKN-positive/negative groups and decoy cell–positive/negative groups or samples with high/low urine BKV gene copy numbers differed significantly (see Table 4 for definitions; all differences highly significant P < 0.001, κ = 0.25 and 0.34, respectively; fair agreement). A marginal difference was observed between the distribution of BKN-positive/negative groups and samples with high/low plasma BKV gene copy numbers (see Table 4 for definitions; P = 0.058, κ = 0.62; good agreement). The PPVs for BKN ranged between 40% (high decoy cell shedding) and 74% (high plasma BKV load levels), and NPVs ranged between 88% (plasma BKV load levels) and 100% (decoy cell shedding and urine BKV load levels; Table 4).

A subgroup analysis limited to samples obtained from renal allograft recipients for whom all tests were performed in parallel gave similar results (BKN-positive time points n = 30; BKN-negative time points n = 32; thresholds as listed in Table 4; Haufen: sensitivity 100%, specificity 97%, PPV 97%, NPV 100%; high decoy cell shedding: sensitivity 100%, specificity 19%, PPV 54%, NPV 100%; high plasma BKV load levels: sensitivity 70%, specificity 81%, PPV 78%, NPV 74%; high urine BKV load levels: sensitivity 100%, specificity 22%, PPV 55%, NPV 100%).

BKN Definition

According to Heptinstall's Pathology of the Kidney,¹² BKN was defined by morphologic changes in renal biopsies (i.e., histologic evidence of polyomavirus replication in the renal cortex and/or medulla with intranuclear viral inclusion bodies and/or positive immunohistochemical and/or in situ hybridization signals accompanied by varying degrees of tubular injury, inflammation, and fibrosis; disease stages A, B, or C). Typically, BKN follows a protracted disease course and is accompanied by general signs of polyomavirus replication (i.e., decoy cell shedding in the urine) and elevated BK polyomavirus load levels in the urine and plasma. General signs of polyomavirus replication, however, are not pathognomonic for BKN, because they can frequently also be found in patients without disease as asymptomatic and clinically insignificant events.^{7,12,13,16,24,29}

Haufen Definition

“Haufen” (after the German word for “cluster or stack”) were defined as three-dimensional, cast-like, dense polyomavirus aggregates in urine samples analyzed by EM.

Cytologic and Histologic Studies

All 337 urine samples were processed for cytology using standard liquid-based procedures (Cytyc/Hologic Corp., Marlborough, MA) and stained by the Papanicolaou method. Decoy cells were quantified as reported previously.^{7,11,16,17,22,29} Thirty-eight renal biopsies from 32 patients were processed for routine histology, including standard transmission EM studies in selected cases. Morphologic workup, diagnosis, and staging of BKN were conducted according to previous reports and the definitions given already.^7,11,12

All biopsy samples fulfilled the Banff ’97 criteria for tissue adequacy.³⁶ Immunohistochemical detection of polyomavirus-associated antigens was performed on all biopsies with a monoclonal mouse antibody directed against the SV40 large T antigen (Clone 416; Calbiochem, San Diego, CA) and a monoclonal mouse anti-polyomavirus antibody against the VP1 capsid protein (clone 597) following previous reports.^11,37 Transmission EM studies on renal biopsy samples were performed according to standard protocols.

Negative Staining Urine EM for the Detection of Haufen and Immunogold Labeling

Negative staining EM on voided urine samples was conducted according to previously published procedures.^25,34 Important parts of the protocol include an initial urine clarification step followed by a second viral concentration step as described previously.^24,25 Briefly, 10 ml of fixed urine was centrifuged at 12,500 rpm for 30 min (clarification step) to remove large cell fragments; the supernatant was subsequently centrifuged at 36,000 rpm for 1 h for viral concentration. After the last centrifugation round, the supernatant was mostly discarded, and the pellet including a minimal amount (400 to 500 μl) of covering supernatant (containing Haufen) was saved for further analysis. Negative staining was performed using a previously described standard two-step method (adsorption followed by negative staining using 1% uranyl acetate).^33,34 One grid was prepared for each sample. A detailed technical procedure manual will be published separately.

Polyomaviruses were identified by their characteristic ultrastructural morphology. Specimens were qualitatively categorized as either Haufen-positive (i.e., containing one or more tight, three-dimensional polyomavirus clusters) or Haufen-negative. In positive cases, multiple Haufen were generally identified within 5 to 10 min of screening at ×50,000 magnification. No attempt was made to quantify Haufen shedding. All Haufen-negative samples were examined for 30 min before the sample was categorized as negative. The analysis was performed by one investigator (H.K.S.), who was blinded to the clinical and histologic information. A subset of 40 representative urine samples were reevaluated by a second investigator (V.M.) in a blinded manner to test the interobserver reproducibility of Haufen detection.

Three Haufen-positive urine samples were processed and stained by immunogold labeling with a Tamm-Horsfall antibody using standard protocols (monoclonal mouse anti-human Tamm-Horsfall antibody [cat. no. CL1032A; Cedarlane Laboratories, Hornby, ON, Canada] and goat anti-mouse IgGFab′2 5-nm gold beads [cat. no. EMGFAF5; British Biocell Int., distributed in the United States by Ted Pella, Inc., Redding, CA]).

Eighteen Haufen-positive, fixed urine samples were stored for 3 to 15 mo at 4°C. The specimens were reevaluated by negative staining EM to study effects of long-term storage on the structural integrity of polyomaviruses and Haufen.

Quantitative PCR

Plasma and urine BKV loads were determined by a real-time quantitative PCR assay using the ABI PRISM 7900HT Sequence Detection System (Foster City, CA) with well-characterized probes and primers specific for BK virus.¹⁸ Real-time detection of PCR products was achieved with a fluorescence hydrolysis (TaqMan) probe. Primers and probes were purchased from TIB Molbiol LLC (Adelphia, NJ). The primer and probe sequences were as follows: BK virus forward 5′-GCAGCTCCCAAAAAGCCAAA-3′; BK virus reverse 5′-CTGGGTTTAGGAAGCATTCTA-3′; BK virus TaqMan probe 5′-6-FAM-ACCCGTGCAAGTGCCAAAACTACTAATAAAAGG-TAMRA-3′.

Quantitative linearity of the assay extended from 10 to 10⁹ measured copies of BKV equivalents correlating with a dynamic linear range of 250 to 2.5 × 10¹⁰ BKV copies/ml urine or plasma. Detectable BKV DNA below the lower limit of linearity (<250 BKV copies/ml) was classified as a low-positive result.

Statistical Analysis

Statistical analyses were performed with standard programs (Excel [Microsoft Corp., Redmond, WA] and SAS [SAS Institute, Cary, NC]). Using previously proposed cutoff levels for identifying patients at high risk for BKN (plasma PCR ≥1 × 10⁴ BK virus copies/ml; urine PCR ≥1 × 10⁷ BK virus copies/ml; urine cytology ≥10 decoy cells per ThinPrep slide), samples were categorized as either positive (above threshold level) or negative (below threshold level).^2,17,18,22

For statistical analysis, the term “BKN positive” was restricted to time points of biopsy-confirmed BKN (as defined already; both at time of initial BKN diagnosis and during follow-up/persistent disease). All urine and plasma samples collected at these BKN-positive time points were classified as either true positive (containing Haufen or showing BK virus loads/decoy cells above cutoff levels) or false negative (no Haufen or showing BK virus loads/decoy cells below cutoff levels).

The term “BKN negativity” was applied to time points of biopsy-proven absence of BKN (including all biopsies obtained during disease resolution), all control patients, and all time points of the pre-BKN period (span after transplantation up to 2 wk before the initial histologic diagnosis of BKN). All samples collected at these BKN-negative time points were classified as either true negative (containing no Haufen or showing BK virus loads/decoy cells below cutoff levels) or false positive (containing Haufen or showing BK virus loads/decoy cells above cutoff levels). The agreement rates among decoy cell shedding, urine and plasma BKV load levels, Haufen, and BKN were assessed using (1) McNemar test (significant differences reported as P < 0.05), (2) κ statistics, and (3) percentage of concordance.