The Journal of Infectious Diseases

MAJOR ARTICLE

Glomerular Parietal Epithelial Cells Infection Is Associated

With Poor Graft Outcome in Kidney Transplant Recipients
With BK Polyomavirus–Associated Nephropathy

Downloaded from https://academic.oup.com/jid/article-abstract/219/12/1879/5288592 by F. Hoffmann-La Roche Ltd user on 27 November 2019

Xu-Tao Chen,1,a Shi-Cong Yang,2,a Wen-Fang Chen,2 Jun Li,1 Su-Xiong Deng,1 Jiang Qiu,1 Ji-Guang Fei,1 Rong-Hai Deng,1 Yan-Yang Chen,2 Pei-Song Chen,3
Yang Huang,1 Chang-Xi Wang,1,b and Gang Huang1,b
1
Organ Transplant Center, 2Department of Pathology, and 3Clinical Laboratory Department, First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China

Background.  The purpose of this study was to investigate the effect of BK polyomavirus (BKPyV infection of glomerular parietal
epithelial cells (GPECs) on graft outcome in kidney transplant recipients with BKPyV-associated nephropathy (BKPyVAN).
Methods.  A total of 152 kidney transplant recipients with BKPyVAN were divided into 31 with (GPEC-positive group) and 121
without (GPEC-negative group) BKPyV-infected GPECs. Clinicopathological characteristics and allograft survival were compared
between the groups.
Results.  The GPEC-positive group had more patients with advanced-stage BKPyVAN than the GPEC-negative group (P < .001).
At the last follow-up, the GPEC-positive group had a significantly higher serum creatinine level than the GPEC-negative group. The
graft loss rate in the GPEC-positive group was higher than that in the GPEC-negative group (32.3% vs 12.4%; P = .008). Kaplan-
Meier analysis showed that the graft survival rate in the GPEC-positive group was lower than that in the GPEC-negative group (log-
rank test, P = .004). Multivariate Cox regression analysis demonstrated that BKPyV infection of GPECs was an independent risk
factor for graft survival (hazard ratio, 3.54; 95% confidence interval, 1.43–8.76; P = .006).
Conclusions.  GPEC infection in patients with BKPyVAN indicates more-severe pathological damage and a rapid decline in
renal function. BKPyV infection of GPECs is an independent risk factor for allograft loss.
Keywords.  BK polyomavirus; renal transplantation; glomerular infection; pathology; graft survival.

BK polyomavirus (BKPyV)–associated nephropathy
(BKPyVAN) is a severe complication of renal transplantation,
affecting 1%–10% of kidney transplant recipients in the first
2 years after transplantation [1]. Immunosuppression following
kidney transplantation has been shown to increase the risk of
BKPyV reactivation in recipients [2–4]. At present, there is no
effective antiviral treatment for BKPyV infection, and recipients
with BKPyVAN have an elevated risk of graft loss [1]. It has
been reported that renal graft failure occurs in 10%–80% of
recipients within 2–3 years after diagnosis of BKPyVAN [1, 5].
The typical pathological features of BKPyVAN include
tubulointerstitial lesions, viral cytopathic damage to tubular epi-
thelial cells, and positive results of immunohistochemical (IHC)
nuclear staining for SV40 large T antigen [6–8]. In addition to
Received 25 November 2018; editorial decision 4 January 2019; accepted 9 January 2019;
published online January 12, 2019.
a
X.-T. C. and S.-C. Y. contributed equally to this work.
b
G. H. and C.-X. W. contributed equally to this work.
Correspondence: G. Huang, MD, PhD, #58 Zhongshan Rd 2, Guangzhou, Guangdong Province,
China, 510080 (huanggang_791021@163.com).
The Journal of Infectious Diseases®  2019;219:1879–86
© The Author(s) 2019. Published by Oxford University Press for the Infectious Diseases Society
of America. All rights reserved. For permissions, e-mail: journals.permissions@oup.com.
DOI: 10.1093/infdis/jiz022
tubular involvement, we have noticed that BKPyV infection
can be occasionally observed in the Bowman space in renal bi-
opsy specimens from patients with BKPyVAN. Nevertheless,
glomerular BKPyV infection received minimal attention, and
scant literature has described this phenomenon [9–11]. Celik
et al reviewed 124 biopsy specimens from 83 kidney transplant
recipients and identified 36 specimens with BKPyV infection
of the Bowman space [11]. However, the effect of glomerular
BKPyV infection on graft outcome is unknown.
There are 3 histological classifications of BKPyVAN, in-
cluding the original schema reported by the University of
Maryland [12] and subsequently modified by American
Society of Transplantation (AST) [8], as well as another system
proposed by the Banff Working Group [7]. These classifications
focus on several pathological features of BKPyVAN, such as
cytopathic effects, tubulitis, interstitial inflammation, intersti-
tial fibrosis, and tubular atrophy, which have been shown to
be associated with graft loss [6, 7, 13]. However, none of these
histological classifications suggests a clinical significance of glo-
merular infection in patients with BKPyVAN. Thus, the pur-
pose of this study was to examine the effect of BKPyV infection
of glomerular parietal epithelial cells (GPECs) on graft outcome
in kidney transplant recipients with BKPyVAN.

Glomerular infection in BKPyVAN  •  jid 2019:219 (15 June) • 1879

METHODS
Study Design and Patients
The medical records of 897 kidney transplant recipients who un-
derwent kidney allograft biopsies at our hospital between 2007
and 2017 were retrospectively reviewed. A flow diagram of pa-
tient inclusion is shown in Figure 1. A total of 159 patients who
had at least 1 biopsy specimen that was positive for anti–SV40
large T antigen on immunohistochemical (IHC) staining were
identified in the medical records. Seven patients were excluded:
3 patient was excluded because of an insufficient number of
glomeruli (ie, <7) in the biopsy specimen; 1, because of concur-
rent myeloma nephropathy; and 3, because of positivity for JC
virus viremia. The remaining 152 patients were divided into 31
with (GPEC-positive group) and 121 without (GPEC-negative
group) BKPyV-infected GPECs, based on the presence or ab-
sence of anti-SV40 large T antigen on IHC staining.
This study adhered to the tenets of the Declaration of Helsinki
and was approved by the ethics committee and research board
of our institution. Written informed consent was obtained from
each patient.
Data Collection
Patient demographic and clinicopathological data were col-
lected and compared between the 2 groups. Severe pneumonia
was defined as the requirement of any mechanical ventila-
tion. Delayed graft function was defined as the requirement
of any dialysis treatment within the first week after trans-
plantation. The end point of the analysis was the date of pa-
tient death or graft loss (ie, the date of dialysis resumption or
retransplantation) or the date of the most recent clinical data
(until December 2017).
Urine Cytology
Urinary cytological smears were stained by the Papanicolaou
method and evaluated for the presence of cells with intranuclear
Kidney transplantation n = 897
(allograft biopsy n = 1200)
PyVANa n = 159
GPEC positive n = 31
Figure 1.  Flow of patients through the study. GPEC, glomerular parietal epithelial cell; PyVAN, polyomavirus-associated nephropathy. aBetween 2007 and 2016.
1880 • jid 2019:219 (15 June) • Chen et al
viral inclusions (decoy cells), which were counted as the number
per 10 high-power fields [14].
Virological Studies
Urine and plasma BKPyV loads were quantitatively measured
by quantitative polymerase chain reaction (PCR) analysis (MJ
Research, Waltham, MA). Specimen collection and processing,
PCR primers, the TaqMan probe (targeting the BKPyV VP1
gene), the plasmid standard containing the targeted BKPyV
Downloaded from https://academic.oup.com/jid/article-abstract/219/12/1879/5288592 by F. Hoffmann-La Roche Ltd user on 27 November 2019
VP1 gene, amplification protocols, PCR precautions, and
quality assurance were performed as previously described [15].
Urine and plasma BKPyV loads are presented as BKPyV ge-
nome copies per milliliter. The limit of quantitation was 1000
copies/mL.
Diagnosis of BKPyVAN
Specimens were stained with hematoxylin and eosin, periodic
acid Schiff, and Masson silver trichrome stains for light micros-
copy. The diagnosis of polyomavirus-associated nephropathy
(PyVAN) was established by the presence of interstitial inflam-
mation, tubular atrophy, and interstitial fibrosis and the extent
of viral cytopathic changes in the tubular epithelial cells, and it
was confirmed by positive results of IHC nuclear staining with
anti-SV40 large T antigen monoclonal antibody, as previously
described [16]. The histological features of PyVAN were classi-
fied using the AST schema, and PyVAN was classified as stage A,
B, and C, based on the guidelines published by Hirsch et al [8].
Histological evaluation of the viral load was semiquantitative,
with findings reported as the percentage of tubules positive for
polyomavirus, using a 4-tier system (<10%, 10%–25%, 25%–
50%, and >50%) [15]. The diagnosis of BKPyVAN was con-
firmed on the basis of positive BK viruria and/or BK viremia,
in addition to positive results of immunostaining for SV40
large T antigen. Particular attention was paid to the presence
of anti-SV40 large T antigen staining in the Bowman space.
Inclusion criteria
• Positive SV40 large T antigen
Exclusion criteria
• Concurrent myeloma
nephropathy n = 1
• JC polyomavirus
infection n = 3
GPEC negative n = 121 • Glomeruli <7 n = 3
GPEC-positive was defined as at least 1 glomerulus positive for
anti-SV40 large T antigen. All slides were reviewed by 2 inde-
pendent pathologists, and histological scores were evaluated
using the Banff criteria [7, 17, 18].
Electron Microscopy
A Philips CM10 electron microscope (Philips, Eindhoven, the
Netherlands) was used to observe viral particles, especially in
tubular epithelial cells and the Bowman space [16]. Cells with
viral particles (diameter, approximately 45  nm) in the cyto-
plasm or nucleus were considered to be infected.
Modification of Immunosuppression and Follow-up
For patients with BKPyVAN, modification of immunosuppres-
sion included a reduction in the dosage of calcineurin inhibitor
and a switch from tacrolimus treatment to cyclosporine A  or
rapamycin treatment. For patients with severe infections, the
dosage of mycophenolic acid and/or calcineurin inhibitor was
reduced, or the medication(s) was discontinued. The adminis-
tration of oral glucocorticoids was not changed. Acute cellular
rejection was treated with pulse methylprednisolone with or
without polyclonal antibody. Antibody-mediated rejection was
treated with pulse methylprednisolone, intravenous immuno-
globulin, rituximab, and/or plasma exchange.
All patients were regularly followed up in the outpatient
clinic, and serum creatinine level, calcineurin inhibitor trough
level, and BKPyV load were monitored. Follow-up renal biopsy
was performed if the serum creatinine level was continuously
elevated or after treatment of BKPyVAN for 6–12 months.
Statistical Analysis
Continuous data are expressed as means ± standard deviations
and were compared by the Student independent t test. If nor-
mality was not assumed, the Mann-Whitney U test was use to
compare data between groups. Categorical data are presented as
numbers and percentages were and compared by the Pearson χ2
test or the Fisher exact test (if an expected value was ≤5). Two-
factor mixed-design analysis of variance (ANOVA; 1 variable
was time dependent) was used to compare serum creatinine
level at baseline, diagnosis, and at final follow-up. Point-biserial
correlation analysis was used to calculate the correlation co-
efficient between pathological scores and BKPyV infection
in the glomerulus. Survival analysis using the Kaplan-Meier
method and a Cox proportional-hazards regression model was
performed to evaluate the impact of BKPyV infection of GPECs
on allograft survival. To investigate the association of BKPyV
infection of GPECs and other independent variables with al-
lograft survival, univariate and multivariate regression models
were performed, and only variables significant in both univar-
iate and multivariate results were considered associated factors.
Results are reported as estimated hazard ratios (HRs) and 95%
confidence intervals (CIs). All statistical tests were 2-tailed,
and a P value of <  .05 was considered to indicate statistical
significance. All analyses were performed by using IBM SPSS,
version 20 (SPSS Statistics V20, IBM, Somers, NY).
RESULTS
Patient Characteristics
Patient characteristics are summarized in Table 1. The base-
line serum creatinine level in the GPEC-negative group
was significantly higher than in the GPEC-positive group
Downloaded from https://academic.oup.com/jid/article-abstract/219/12/1879/5288592 by F. Hoffmann-La Roche Ltd user on 27 November 2019
(P = .004, Table 1). However, there was no significant differ-
ence in other characteristics between the GPEC-positive and
GPEC-negative groups, including sex, age, etiology of end-
stage renal disease, duration and type of dialysis, immunosup-
pressive regimen, donor type, severe pneumonia, and delayed
graft function.
Time of BKPyV Infection of GPECs
The time of BKPyV infection of GPECs in the GPEC-positive
group is shown in Supplementary Figure 1. A significant pro-
portion of patients with BKPyV-infected GPECs were identified
within 24 months after transplantation, with others developing
BKPyV-infected GPECs up to 7 years after transplantation.
Comparison of BKPyV Load Between Groups
At the initial diagnosis of BKPyVAN, the number of decoy cells
and BKPyV loads in the urine and plasma were not significantly
different between the GPEC-positive and GPEC-negative
groups (all, P  >  .05, Supplementary Table 1). Treatments for
BKPyVAN were not significantly different between the 2 groups
(P = .491; Supplementary Table 2).
During follow-up, there was no significant difference in
the clearance of BKPyV DNA from blood (83.9% vs 92.6%;
P = .163) and urine (19.4% vs 22.3%; P = .721; Supplementary
Table 2) between the 2 groups.
Comparison of Graft Function Between Groups
The vast majority of patients in both groups underwent a renal
biopsy because of an elevated serum creatinine level (>30%
above baseline). The serum creatinine level at initial diagnosis
of BKPyVAN was significantly elevated, compared with the level
at a stable phase in both groups (P < .05 for both comparisons).
There was no significant difference in serum creatinine level at
the initial diagnosis of BKPyVAN between groups (P  =  .352;
Supplementary Table 1). However, the mean serum creatinine
level (±SD) at the last follow-up visit was significantly higher
in the GPEC-positive group, compared with that in the GPEC-
negative group (255.43  ±  150.71 vs 196.60  ±  112.50  μmol/L;
P = .041; Supplementary Table 2).
The median duration of follow-up was 11.2  months in the
GPEC-positive group and 19.6  months in the GPEC-negative
group. The rate of increase in the serum creatinine level was
greater in the GPEC-positive group than in the GPEC-negative
group (P < .001 for time effect; P = .003 for time × GPEC effect;
2-factor mixed-design ANOVA).
Glomerular infection in BKPyVAN  •  jid 2019:219 (15 June) • 1881
Table 1.  Demographic and Transplant Characteristics of 152 Patients With BK Polyomavirus (BKPyV)–Associated Nephropathy, by BKPyV Infection Status
in Glomerular Parietal Epithelial Cells (GPECs)

GPEC Negative GPEC Positive Total

Parameter (n = 121) (n = 31) (n = 152) P

Sex .568
 Male 77 (63.64) 18 (58.06) 95 (62.50)
 Female 44 (36.36) 13 (41.94) 57 (37.50)
Age at transplantation, y 38.78 ± 8.88 41.15 ± 9.71 39.27 ± 9.08 .195

Downloaded from https://academic.oup.com/jid/article-abstract/219/12/1879/5288592 by F. Hoffmann-La Roche Ltd user on 27 November 2019

ESRD etiology .500
  Chronic glomerulonephritis 78 (64.46) 23 (74.19) 101 (66.45)
  IgA nephropathy 17 (14.05) 4 (12.90) 21 (13.82)
 Other 26 (21.49) 4 (12.90) 30 (19.74)
Renal replacement therapy .601
  Peritoneal dialysis 29 (23.97) 6 (19.35) 35 (23.03)
 Hemodialysis 83 (68.60) 21 (67.74) 104 (68.42)
  Without dialysis 9 (7.44) 4 (12.90) 13 (8.55)
Immunosuppressive regimen
  Polyclonal antibody 64 (52.89) 18 (58.06) 82 (53.95) .606
  Monoclonal antibody 51 (42.15) 12 (38.71) 63 (41.45) .729
 Tac-MPA 120 (99.17) 31 (100.00) 151 (99.34) 1.000
 CsA-MPA 1 (0.83) 0 (0.00) 1 (0.66) 1.000
Donor type
 Living 22 (18.18) 5 (16.13) 27 (17.76) 1.000
 Nonliving 99 (81.82) 26 (83.87) 125 (82.24) 1.000
Baseline creatinine level at stable phase, μmol/L 129.52 ± 44.66 104.97 ± 25.85 124.51 ± 42.62 .004
Complication
  Delayed graft function 15 (12.40) 4 (12.90) 19 (12.50) 1.000
  Severe pneumonia 14 (11.57) 5 (16.13) 19 (12.50) .544
  Preceding TCMR/AMR 8 (6.61) 2 (6.45) 10 (6.58) 1.000

Data are no. (%) of participants or mean value ± SD.

Abbreviations: AMR, antibody-mediated rejection; CsA, cyclosporine A; ESRD, end-stage renal disease; IgA, immunoglobulin A; MPA, mycophenolic acid; Tac, tacrolimus; TCMR, T cell–
mediated rejection.

More patients developed a sustained increase in serum
creatinine level (>30% greater than that at diagnosis) in the
GPEC-positive group than in the GPEC-negative group
(41.9% vs 18.2%; P = .005; Supplementary Table 2). The abso-
lute value of the increase in serum creatinine level (calculated
as the level at the last follow-up visit minus the level first re-
corded) was significantly higher in the GEPC-positive group
than that in the GEPC-negative group (304.71  ±  329.89 vs
115.74 ± 217.63 μmol/L; P < .001). These results suggest that the
GPEC-positive group had worse graft function than the GPEC-
negative group.
Pathological Findings at Initial Diagnosis of BKPyVAN
Representative images of GPEC-positive biopsy specimens
under light microscopy and electron microscopy are shown in
Figure 2. Based on AST criteria, in the GPEC-positive group
there were no patients with stage A  disease, 26 (83.9%) with
stage B disease, and 5 (16.1%) with stage C disease. In the
GPEC-negative group, there were 11 patients (9.1%) with stage
A disease, 96 (79.3%) with stage B disease, and 14 (11.6%) with
stage C disease (Supplementary Table 3). The GPEC-positive
group had a higher proportion of patients with advanced
stage BKPyVAN than the GPEC-negative group (P  <  .001;
Supplementary Table 3). The incidence of periglomerular fi-
brosis in the GPEC-positive group was higher than that in
the GPEC-negative group (P  =  .045; Supplementary Table 3).
Positive anti-SV40 large T antigen staining in tubular epithelial
cells was observed in all biopsy specimens from both groups.
Notably, the mean Banff scores for the extent of SV40 large
T antigen staining, tubulitis, interstitial inflammation, tubular
atrophy, and interstitial fibrosis were all significantly higher in
the GPEC-positive group, compared with the GPEC-negative
group (P  <  .05 for all comparisons; Supplementary Table 3).
Correlation analysis revealed that positivity in GPECs was sig-
nificantly positively correlated with the extent of SV40 large T
antigen staining (r  =  0.52), tubulitis (r  =  0.17), interstitial in-
flammation (r = 0.30), tubular atrophy (r = 0.21), and interstitial
fibrosis (r = 0.28; P < .05 for all comparisons).
Pathological Findings During Follow-up
During follow-up, 12 of 31 patients in the GPEC-positive
group and 61 of 121 patients in the GPEC-negative group un-
derwent repeat biopsies (Supplementary Table 3). The median
time between the initial biopsy and last follow-up biopsy was

1882 • jid 2019:219 (15 June) • Chen et al


A B
C D
Figure 2.  Histological features in biopsy specimens from patients with BK
polyomavirus (BKPyV)–associated nephropathy with glomerular parietal epithe-
lial cell (GPEC) involvement. A, BKPyV-involved GPECs exhibited an enlarged and
ground-grass–like nucleus, accompanied with prominent intranuclear inclusions
(arrows; hematoxylin and eosin staining; original magnification  ×200). B,
Immunohistochemical staining showed nuclear positivity for SV40 large T antigen
(arrows) in GPECs (DAB staining; original magnification ×200). C, Electronic micros-
copy revealed electron-dense structures in the nucleus of GPECs (arrow; original
magnification ×1500). D, On high-power microscopy, these structures had a para-
crystalline arrangement, composing of naked and round viral particles that meas-
ured 45 nm in diameter (original magnification ×50 000).
14.35 months (range, 1.63–39.47 months) in the GPEC-positive
group and 16.45  months (range, 0.30–69.23  months) in the
GPEC-negative group.
The proportion of patients who developed global
glomerulosclerosis was similar between the 2 groups (2.5% vs
2.4%). In the GPEC-positive group, compared with findings for the
initial biopsy specimen, SV40 large T antigen staining, tubulitis,
and interstitial inflammation were decreased but tubular atrophy
and interstitial fibrosis were increased on the last biopsy spec-
imen (Supplementary Table 3). During follow-up, the degrees of
interstitial fibrosis (P = .043) and tubular atrophy (P = .040) were
significantly higher in the GPEC-positive group, compared with
findings in the GPEC-negative group (Supplementary Table 3).
Pathological changes in biopsy specimens from the first to the last
biopsy are summarized in Supplementary Table 4.
Association Between GPEC Positivity and Graft Survival
During follow-up, graft loss in the GPEC-positive group was
significantly greater than that in the GPEC-negative group
(32.3% vs 12.4%; P = .008; Supplementary Table 2). One patient
with graft function in the GPEC-positive group died because of
hepatic failure. The 5-year cumulative graft survival rate after
transplantation was 66.7% in the GPEC-positive group and
100.0% in the GPEC-negative group (P = .004). Kaplan-Meier
analysis showed that the death-censored graft survival rate
after initial biopsy was significantly lower in the GPEC-positive
group, compared with that in the GPEC-negative group (log-
rank test P  =  .004; Figure 3). The mean time (±SD) from di-
agnosis of BKPyVAN to graft loss was 7.4 ± 8.5 months in the
GPEC-positive group and 17.2  ±  20.5  months in the GPEC-
negative group (P = .179).
Univariate and multivariate Cox regression models were
performed to investigate risk factors for graft loss. As shown in
Table 2, BKPyV infection of GPECs was identified as an inde-
Downloaded from https://academic.oup.com/jid/article-abstract/219/12/1879/5288592 by F. Hoffmann-La Roche Ltd user on 27 November 2019
pendent risk factor for graft loss (HR, 3.54; 95% CI, 1.43–8.76;
P = .006). Serum creatinine level at diagnosis (HR, 1.01; 95%
CI, 1.00–1.01; P  <  .001) and tubular atrophy score (HR, 3.13;
95% CI, 1.21–8.12; P = .019) were also independent risk factors
for graft loss.
DISCUSSION
In this study, we investigated the effect of GPEC infection on
graft outcome in kidney transplant recipients with BKPyVAN.
Of the patients with BKPyVAN, 20.4% developed a GPEC in-
fection, most within 24 months after transplantation. At the last
follow-up, the GPEC-positive group had a significantly higher
serum creatinine level and ratio of sustained increase in serum
creatinine level. At diagnosis of BKPyVAN, the proportion of
patients with advanced-stage BKPyVAN was greater in the
GPEC-positive group. The graft loss rate in the GPEC-positive
group was significantly higher than that in the GPEC-negative
group. Kaplan-Meier analysis showed that the graft survival
rate in the GPEC-positive group was significantly lower than
that in the GPEC-negative group. Multivariate Cox regression
analysis demonstrated that GPEC infection was an independent
risk factor for graft loss. Taken together, these results suggested
1.0 GPEC
GPEC negative
GPEC positive
GPEC negative censored
GPEC positive censored
0.8
Cumulative survival (proportion)
0.6
0.4
0.2
0
0 25.0 50.0 75.0 100.0 125.0
Time after initial biopsy (months)
Figure 3.  Graft survivals among patients with (glomerular parietal epithelial cell
[GPEC]–positive group) and those without (GPEC-negative group) BK polyomavirus
(BKPyV)–infected GPECs. Kaplan-Meier curves show death-censored allograft sur-
vival times after initial biopsy in both groups. Log-rank test, P = .004.
Glomerular infection in BKPyVAN  •  jid 2019:219 (15 June) • 1883
Table 2.  Risk Factors for Graft Loss, According to Cox Regression Analysis

Univariate Multivariate

Parameter HR (95% CI) P HR (95% CI) P

BKPyV infection status in GPECs

 Negative Reference Reference
 Positive 3.75 (1.66–8.47) .002 3.54 (1.43–8.76) .006
Donor type
 Nonliving Reference …

Downloaded from https://academic.oup.com/jid/article-abstract/219/12/1879/5288592 by F. Hoffmann-La Roche Ltd user on 27 November 2019

 Living 0.58 (.17–1.94) .374 …
Serum creatinine level at diagnosis (mmol/L) 1.01 (1.005–1.011) <.001 1.01 (1.00–1.01) <.001
BKPyV viremia level at diagnosis (copies/mL) 1.00 (1.00–1.00) .946 …
Clearance of BKPyV DNA in blood
 No Reference Reference
 Yes 0.27 (.09–.80) .018 0.32 (.09–1.09) .068
Delayed graft function
 No Reference …
 Yes 2.69 (.98–7.37) .055 …
Preceding rejection episodes
 No Reference …
 Yes 1.36 (.56–3.31) .499 …
Rejection episodes after BKPyVAN .279
 No Reference …
  T cell–mediated rejection 4.31 (1.00–18.55) .050 …
  Antibody-mediated rejection 0.00 (.00–.00) .986 …
  Mixed cellular- and antibody-mediated rejection 0.00 (.00–.00) .990 …
Tubulitis score 0.86 (.52–1.43) .572
Tubular atrophy score 2.68 (1.59–4.52) <.001 3.13 (1.21–8.12) .019
Interstitial inflammation score 2.02 (1.25–3.26) .004 0.89 (.45–1.78) .742
Interstitial fibrosis score 2.10 (1.33–3.32) .001 0.53 (.22–1.23) .139

Abbreviations: BKPyV, BK polyomavirus; BKPyVAN, BK polyomavirus–associated nephropathy; CI, confidence interval; GPEC, glomerular parietal epithelial cell; HR, hazard ratio.

that glomerular BKPyV infection is associated with worse graft
function, more-advanced BKPyVAN stage, and a higher rate of
graft loss in kidney transplant recipients with BKPyVAN. To
the best of our knowledge, this is the first study reporting the
prognostic value of GPEC infection on graft outcome in kidney
transplant recipients with BKPyVAN.
Nankivell et al reported that a high level of viremia was asso-
ciated with a heavy tissue viral load [19]. Thus, we had assumed
that the GPEC-positive group, probably having a heavier tissue
viral load, might display a higher plasma BKPyV load than that
in the GPEC-negative group, but this turned out to be irrelevant.
On the other hand, hepatitis B virus (HBV) gene mutations were
reported to cause glomerular involvement and HBV-associated
glomerulonephritis [20]. Therefore, we thought it plausible that
specific variant strains of BKPyV might have greater pathogenic
potential and viral invasiveness, resulting in GPEC infection,
which might need further investigation.
Reducing the immunosuppression intensity is the primary
treatment for BKPyVAN [5]. In this study, changes in the im-
munosuppressive regimen, including reducing the dosage or
changing the immunosuppressant, were used in the treatment
of BKPyVAN. Clearance of BKPyV DNA in the blood was
observed in 92.6% of patients in the GPEC-negative group,
demonstrating the effectiveness of these treatment regimens. In
the GPEC-positive group, BKPyV replication also presented a
declining trend. In addition, clearance of BK viremia and BK
viruria occurred in 83.9% and 19.4% of patients, respectively.
These results suggest that, although GPEC positivity represents
a more severe BKPyV infection, patients with BKPyV-infected
GPECs still respond well to treatment.
Celik et al reported that BKPyV infection in the glomerular
epithelium was usually found in biopsy specimens with a high
viral load in the tubular epithelium [11]. In line with this obser-
vation, our pathological findings showed that BKPyV infection
of GPECs was accompanied by more-severe tubulointerstitial
inflammation, tubular atrophy, interstitial fibrosis, and a greater
extent of SV40 large T antigen staining in the tubular epithe-
lium, indicating that BKPyV-infected GPECs are associated
with more-severe pathological damage. Multivariate Cox re-
gression analysis showed that tubular atrophy score was an in-
dependent risk factor for graft loss. Study has shown that, in
severely BKPyV-infected tubules, necrosis of tubular cells and
tubular basement membrane denudation can be observed [13,
21]. Assessment of follow-up biopsy specimens in this study
showed that the degree of SV40 large T antigen staining and
tubulitis in the GPEC-positive group was reduced as compared

1884 • jid 2019:219 (15 June) • Chen et al

to that in initial biopsy specimens and was correlated with
clearance of BK viruria. This improvement may be attributed
to immunological reconstitution after reducing the immuno-
suppressant dosage. However, the extent of tubular atrophy
remained significantly higher in the GPEC-positive group
during follow-up, suggesting irreversible pathological damage.
Our previous study demonstrated that an elevated serum
creatinine level at diagnosis was associated with graft func-
tion decline 12  months after diagnosis and with long-term
graft loss in patients with BKPyVAN [16]. In the current
study, the mean serum creatinine level at diagnosis was
219.1 μmol/L in the GPEC-positive group and 198.6 μmol/L
in the GPEC-negative group, both of which were signif-
icantly higher than the baseline level. The majority of our
patients (92.0% in the GPEC-positive group and 90.2% in the
GPEC-negative group) had an increased serum creatinine
level at diagnosis, suggesting that most patients had severe
pathological damage. In addition, the increasing rate of in-
crease in the serum creatinine level was higher in the GPEC-
positive group, compared with that in the GPEC-negative
group, during follow-up, indicating more-severe BKPyV
infection and pathological damage in the GPEC-positive
group. Multivariate Cox regression analysis confirmed that
the serum creatinine level at diagnosis was an independent
risk factor for graft loss. In this study, the 5-year cumula-
tive allograft survival rate was 66.7% in the GPEC-positive
group and 100% in the GPEC-negative group. The 2-year al-
lograft failure rate for all patients was 30.0%, which was the
same as that reported by the Banff Working Group [7] but
higher than that (15.4%) reported by Drachenberg et al [13].
The GPEC-positive group had a significantly lower death-
censored graft survival rate, and BKPyV infection of GPECs
was identified as a risk factor for graft loss. These results
strongly suggest that GPEC infection by BKPyV predicts
worse graft outcome.
There are some limitations of this study. This was a retrospec-
tive study, and complete pathological data during follow-up
were not available for all patients.
In summary, this study found that the presence of BKPyV-
infected GPECs in renal biopsy specimens from kidney
transplant recipients with BKPyVAN indicates more-severe
pathological damage and is associated with a rapid decline in
renal function. Furthermore, GPEC infection is an independent
risk factor for graft loss.
Supplementary Data
Supplementary materials are available at The Journal of Infectious
Diseases online. Consisting of data provided by the authors
to benefit the reader, the posted materials are not copyedited
and are the sole responsibility of the authors, so questions or
comments should be addressed to the corresponding author.
Notes
Acknowledgments.  We thank all of the nurses at the
Department of Organ Transplantation, the First Affiliated
Hospital, Sun Yat-sen University, for their collaboration and
dedication to the patients.
Financial support.  This work was supported by the National
Natural Science Foundation of China (grant 81770749), the
Natural Science Foundation of Guangdong Province (grant
Downloaded from https://academic.oup.com/jid/article-abstract/219/12/1879/5288592 by F. Hoffmann-La Roche Ltd user on 27 November 2019
2017A030313710), and the Fundamental Research Funds for
the Central Universities (17ykpy29).
Potential conflicts of interest.  All authors: No reported
conflicts. All authors have submitted the ICMJE Form for
Disclosure of Potential Conflicts of Interest. Conflicts that the
editors consider relevant to the content of the manuscript have
been disclosed.
References
1. Sawinski D, Goral S. BK virus infection: an update on di-
agnosis and treatment. Nephrol Dial Transplant 2015;
30:209–17.
2. Schachtner  T, Babel  N, Reinke  P. Different risk factor
profiles distinguish early-onset from late-onset BKV-
replication. Transpl Int 2015; 28:1081–91.
3. Borni-Duval C, Caillard S, Olagne J, et al. Risk factors for BK
virus infection in the era of therapeutic drug monitoring.
Transplantation 2013; 95:1498–505.
4. Belliere  J, Kamar  N, Mengelle  C, et  al. Pilot conversion
trial from mycophenolic acid to everolimus in ABO-
incompatible kidney-transplant recipients with BK viruria
and/or viremia. Transpl Int 2016; 29:315–22.
5. Barth  H, Solis  M, Lepiller  Q, et  al. 45  years after the dis-
covery of human polyomaviruses BK and JC: time to speed
up the understanding of associated diseases and treatment
approaches. Crit Rev Microbiol 2017; 43:178–95.
6. Masutani  K, Shapiro  R, Basu  A, Tan  H, Wijkstrom  M,
Randhawa  P. The Banff 2009 working proposal for
polyomavirus nephropathy: a critical evaluation of its utility
as a determinant of clinical outcome. Am J Transplant 2012;
12:907–18.
7. Nickeleit V, Singh HK, Randhawa P, et al.; Banff Working
Group on Polyomavirus Nephropathy. The Banff
Working Group classification of definitive polyomavirus
nephropathy: morphologic definitions and clinical
correlations. J Am Soc Nephrol 2018; 29:680–93.
8. Hirsch  HH, Randhawa  P; AST Infectious Diseases
Community of Practice. BK polyomavirus in solid
organ transplantation. Am J Transplant 2013; 13 Suppl
4:179–88.
9. Meehan SM, Kraus MD, Kadambi PV, Chang A. Nephron
segment localization of polyoma virus large T antigen in
renal allografts. Hum Pathol 2006; 37:1400–6.
Glomerular infection in BKPyVAN  •  jid 2019:219 (15 June) • 1885
10. Sekulic M, Crary GS, Herrera Hernandez LP. BK polyomavirus
tubulointerstitial nephritis with urothelial hyperplasia in a
kidney transplant. Am J Kidney Dis 2016; 68:307–11.
11. Celik  B, Randhawa  PS. Glomerular changes in BK virus
nephropathy. Hum Pathol 2004; 35:367–70.
12. Drachenberg  RC, Drachenberg  CB, Papadimitriou  JC,
et al. Morphological spectrum of polyoma virus disease in
renal allografts: diagnostic accuracy of urine cytology. Am J
Transplant 2001; 1:373–81.
13. Drachenberg  CB, Papadimitriou  JC, Chaudhry  MR, et  al.
Histological evolution of BK virus-associated nephropathy:
importance of integrating clinical and pathological findings.
Am J Transplant 2017; 17:2078–91.
14. Nickeleit V, Klimkait T, Binet IF, et al. Testing for polyomavirus
type BK DNA in plasma to identify renal-allograft recipients
with viral nephropathy. N Engl J Med 2000; 342:1309–15.
15. Huang  G, Wang  CX, Zhang  L, et  al. Monitoring of
polyomavirus BK replication and impact of preemptive im-
munosuppression reduction in renal-transplant recipients
in China: a 5-year single-center analysis. Diagn Microbiol
Infect Dis 2015; 81:21–6.
1886 • jid 2019:219 (15 June) • Chen et al
16. Huang G, Wu LW, Yang SC, et al. Factors influencing graft
outcomes following diagnosis of polyomavirus -associated
nephropathy after renal transplantation. PLoS One 2015;
10:e142460.
17. Sis B, Mengel M, Haas M, et al. Banff ‘09 meeting report:
antibody mediated graft deterioration and implemen-
tation of Banff working groups. Am J Transplant 2010;
10:464–71.
Downloaded from https://academic.oup.com/jid/article-abstract/219/12/1879/5288592 by F. Hoffmann-La Roche Ltd user on 27 November 2019
18. B Kopp J. Banff classification of polyomavirus nephropathy:
a new tool for research and clinical practice. J Am Soc
Nephrol 2018; 29:354–5.
19. Nankivell  BJ, Renthawa  J, Sharma  RN, Kable  K,
O’Connell  PJ, Chapman  JR. BK virus nephropathy: histo-
logical evolution by sequential pathology. Am J Transplant
2017; 17:2065–77.
20. Lu H, Zhu H, Zhou J. S gene mutations of HBV in children
with HBV-associated glomerulonephritis. Virol J 2012;
9:59.
21. Menter  T, Mayr  M, Schaub  S, Mihatsch  MJ, Hirsch  HH,
Hopfer H. Pathology of resolving polyomavirus-associated
nephropathy. Am J Transplant 2013; 13:1474–83.