Full text of DOI:10.1053/ajkd.2001.27693

Tubular Phenotypic Change in Progressive Tubulointerstitial Fibrosis
in Human Glomerulonephritis
Kiichiro Jinde, MD, David J. Nikolic-Paterson, PhD, Xiao R. Huang, PhD, Hideto Sakai, MD,
Kiyoshi Kurokawa, MD, Robert C. Atkins, DSc, and Hui Y. Lan, MD, PhD
●
There is much debate over the origins of fibroblast-type cells that accumulate in interstitial fibrosis. A controver-
sial hypothesis, supported by data from animal and cell-culture studies, is that fibroblast-type cells can derive from
tubular epithelial cells by a process of epithelial-mesenchymal transdifferentiation. However, to date, no evidence
supports this postulate in human glomerulonephritis. This study sought to provide evidence that tubular epithelial
cells can undergo phenotypic change toward a fibroblast-like cell in human glomerulonephritis. One hundred
twenty-seven open renal biopsy specimens from patients with minimal change disease (MCD), immunoglobulin A
(
IgA) nephropathy, and rapidly progressive glomerulonephritis (RPGN) were examined for tubular phenotypic
change by two-color immunohistochemistry using the criteria of de novo expression of ␣-smooth muscle actin
␣-SMA), a myofibroblast marker; loss of the epithelial marker cytokeratin; and collagen production. In normal
(
human kidney and MCD, tubular epithelial cells expressed cytokeratin with no evidence of ␣-SMA staining.
However, in 36 of 90 cases of IgA nephropathy and 9 of 18 cases of RPGN, small numbers of tubular epithelial cells
in areas of fibrosis showed de novo ␣-SMA expression, accounting for 0.4% ؎ 0.2% (IgA nephropathy) and 3.8% ؎
1
.5% (RPGN) of cortical tubules. An intermediate stage of phenotypic change was observed in some cuboidal
؉
epithelial cells that expressed both cytokeratin and ␣-SMA. Tubules containing ␣-SMA–positive (␣-SMA ) cells also
stained for collagen types I and III, suggesting that tubular cells undergoing phenotypic change have an active role
in the fibrotic process. There also was a marked increase in transforming growth factor-␤1 (TGF-␤1) tubular
expression in areas with interstitial fibrosis, including tubules with phenotypic change. There was a highly
؉
significant correlation between tubular ␣-SMA expression and interstitial fibrosis, interstitial ␣-SMA myofibroblast
accumulation, deposition of collagen types I and III, tubular TGF-␤1 expression, and renal dysfunction. In
conclusion, this study provides evidence that tubular epithelial cells can undergo phenotypic change toward a
myofibroblast-like phenotype on the basis of de novo ␣-SMA expression, loss of cytokeratin, and de novo collagen
staining. These data, although not conclusive, provide the first support for the hypothesis that transdifferentiation
of tubular epithelial cells has a role in progressive renal fibrosis in human glomerulonephritis.
©
2001 by the National Kidney Foundation, Inc.
INDEX WORDS: Glomerulonephritis; transdifferentiation; epithelial; myofibroblast; fibrosis; ␣-smooth muscle actin
(
␣-SMA); cytokeratin; transforming growth factor-␤ (TGF-␤).
UBULOINTERSTITIAL fibrosis is be-
lieved to be the common final pathway
A crucial issue for our understanding of how
interstitial fibrosis develops is knowing the ori-
gin of interstitial myofibroblasts. The conven-
tional view is that these fibroblast-type cells
originate by differentiation of resident renal fibro-
T
leading to end-stage renal failure in glomerulone-
1
phritis. Renal fibroblasts are a heterogeneous
population that is difficult to define in situ be-
cause of the lack of a specific panfibroblast
antigen. Most studies of renal fibrosis have fo-
cused on a subset of fibroblasts called myofibro-
blasts, defined by their expression of ␣-smooth
From the Department of Nephrology, Tokai University
School of Medicine, Isehara, Japan; Department of Nephrol-
ogy, Monash Medical Centre; Department of Medicine,
Monash University, Clayton, Australia; and the Department
of Medicine, University of Hong Kong, Queen Mary Hospi-
tal, Hong Kong.
1
,2
muscle actin (␣-SMA). Myofibroblasts are a
major site of synthesis of such matrix molecules
as collagen types I and III in fibrotic lesions in
the kidney and fibrotic changes in other tis-
Received January 22, 2001; accepted in revised form May
4, 2001.
2
-6
sues, although it should be remembered that
not all matrix-producing cells are myofibro-
blasts. There is excellent correlation between the
appearance of interstitial ␣-SMA–positive (␣-
Supported in part by Committee of Research and Confer-
ence Grants grant no. 98/33704 from The Hong Kong
University.
Address reprint requests to Hui Y. Lan, MD, PhD, Associ-
ate Professor of Medicine, Department of Medicine-Nephrol-
ogy, Baylor College of Medicine, One Baylor Plaza, Alkek
Bldg N520, Houston, TX 77030. E-mail: hlan@bcm.tmc.edu
ϩ
SMA ) myofibroblasts and the development of
interstitial fibrosis in human and experimental
glomerulonephritis, and interstitial ␣-SMAimmu-
nostaining is the best prognostic indicator of
©
0
2001 by the National Kidney Foundation, Inc.
272-6386/01/3804-0006$35.00/0
doi:10.1053/ajkd.2001.27693
3
,4,7-17
disease progression.
American Journal of Kidney Diseases, Vol 38, No 4 (October), 2001: pp 761-769
761

762
JINDE ET AL
blasts or migration of pericytes or smooth muscle
obtained at the time of biopsy (Table 1). Normal human
kidney was obtained from the uninvolved pole of carcinoma
nephrectomies (4 cases) and unmatched donor organs (2
cases) from the Department of Nephrology, Monash Medi-
cal Centre (Clayton, Australia).
1
8
cells into the interstitium. However, another
possibility exists: the controversial hypothesis
that tubular epithelial cells transdifferentiate into
myofibroblasts. Epithelial-mesenchymal transi-
tion occurs in fetal development and some patho-
Histological Evaluation
1
9-22
logical states,
but there is relatively little
Biopsy tissue was fixed in 10% neutral buffered formalin
and sections (2 ␮m) were stained with hematoxylin and
eosin or periodic acid–Schiff (PAS). The percentage of
cortical tubulointerstitium showing fibrosis was scored on
evidence to support such a mechanism in renal
fibrosis. Phenotypic and morphological evidence
that tubular epithelial cells can transdifferentiate
into fibroblast-like cells comes from cell-culture
27
PAS-stained sections by point-counting on coded slides.
2
3-25
studies
and studies of mouse and rat models
Antibodies
2
6-29
of kidney disease.
However, to date, no data
The following antibodies were used in this study: mouse
monoclonal antibody (mAb) 1A4 that specifically recog-
nized the smooth muscle ␣-actin isoform (Sigma Chemical
Co, St Louis, MO), pan cytokeratin mouse mAb AE1/AE3
support the hypothesis that tubular epithelial-
myofibroblast transdifferentiation has a role in
interstitial fibrosis in human glomerulonephritis.
To this end, we sought evidence that tubular
epithelial cells can undergo phenotypic change
toward a myofibroblast-type cell in biopsy tissue
from patients with glomerulonephritis.
(Dako Corp, Carpenteria, CA), mouse antihuman collagen
type I mAb I-8H5 (Fuji Chemical Industries, Toyama, Ja-
pan), mouse antihuman collagen type III mAb III-53 (Fuji
Chemical Industries), and rabbit antihuman transforming
growth factor-␤1 (TGF-␤1; Santa Cruz Biotechnology, Santa
Cruz, CA). Negative controls used isotype-matched mouse
mAb or normal rabbit IgG. In addition, peroxidase-conju-
gated goat antimouse immunoglobulin, mouse peroxidase-
conjugated antiperoxidase complexes (PAPs), alkaline phos-
phatase–conjugated goat antimouse and goat antirabbit
immunoglobulins, and mouse and rabbit alkaline phos-
phatase–conjugated antialkaline phosphatase complexes
METHODS
Patients
Open renal biopsy specimens from 127 patients from the
Department of Nephrology, Tokai University School of
Medicine (Isehara, Japan), were examined on this study. The
study was approved by the University Review Board. Dis-
ease category was based on histological examination of
biopsy specimens. The cohort consisted of 19 cases of
minimal change disease (MCD), 90 cases of immunoglobu-
lin A (IgA) nephropathy, and 18 cases of rapidly progressive
glomerulonephritis (RPGN; Table 1). Patients with IgA
nephropathy were graded according to the histological clas-
(APAAPs) were purchased from Dakopatts (Glostrup, Den-
mark).
Immunohistochemistry
Immunohistochemistry staining for ␣-SMA and collagen
types I and III used a three-layer PAP method, as previously
2
7,31
described.
Paraffin sections (4 ␮m) were dewaxed. For
3
0
sification of Clarkson et al and divided into two groups.
Group 1 included grades I (minimal alteration) and II (pure
mesangial proliferation). Group 2 included grades III (focal
and/or diffuse mesangial proliferation and Ͻ50% crescents)
and IV (focal and/or diffuse mesangial proliferation and
collagen staining, sections were treated with 10 minutes of
microwave oven heating in 400 mL of 0.01 mol/L of sodium
citrate, pH 6.0, at 2,450 MHz and 800 W, followed by
incubation with 0.05% trypsin for 30 minutes at 37°C.
Sections then were preincubated for 20 minutes in 10% fetal
calf serum (FCS) and 10% normal goat serum to block
nonspecific binding, incubated with a mouse mAb for 60
5
0% to 70% crescents). Data for serum creatinine level,
creatinine clearance, and total urinary protein excretion were
Table 1. Patient Renal Function Data
Serum Creatinine (mg/mL)
Creatinine Clearance (mL/min)
Urinary Protein (g/day)
MCD (n ϭ 19)
0.72 Ϯ 0.31
0.79 Ϯ 0.18
90.7 Ϯ 28.2
89.3 Ϯ 23.4
57.0 Ϯ 28.3*
4.0 Ϯ 4.0
0.31 Ϯ 0.33*
1.55 Ϯ 1.48*
1.88 Ϯ 1.98
IgA nephropathy group 1 (n ϭ 40)
IgA nephropathy group 2 (n ϭ 50)
RPGN (n ϭ 18)
†
†
1.59 Ϯ 1.39
‡
§ꢀ
*§ꢀ
†‡
4.32 Ϯ 2.86
23.4 Ϯ 22.5
NOTE. Data expressed as mean Ϯ SD.
*
P Ͻ 0.001 versus MCD.
†
P Ͻ 0.05 versus IgA nephropathy group 1.
‡
§
ꢀ
P Ͻ 0.01 versus MCD by analysis of variance.
P Ͻ 0.001 versus IgA nephropathy group 1 by analysis of variance.
P Ͻ 0.001 versus IgA nephropathy group 2 by analysis of variance.

TUBULAR DIFFERENTIATION IN HUMAN GN
763
minutes, and washed three times in phosphate-buffered
saline. After inactivating endogenous peroxidase in 0.3%
hydrogen peroxide in methanol, sections were labeled sequen-
tially with peroxidase-conjugated goat antimouse IgG and
mouse PAP, followed by development with diaminobenzi-
dine to produce a brown color. Sections were then counter-
stained with PAS and mounted.
(Fig 1A). No staining of tubular epithelial cells
or glomeruli was apparent in normal human
kidney.
Tubular Phenotypic Change in Human
Glomerulonephritis
Immunostaining for cytokeratin and TGF-␤1 used a three-
layer APAAP method. This was performed alone or after a
three-layer PAP stain, described previously. Briefly, sections
were microwaved for 10 minutes. Sections then were prein-
cubated in block solution as described previously and incu-
bated with a mouse or rabbit antibody for 60 minutes,
followed by alkaline phosphatase–conjugated goat anti-
mouse or goat antirabbit IgG and mouse or rabbit APAAP.
Finally, sections were developed with 5-bromo-4-chloro-3-
indolyl phosphate and nitroblue tetrazolium salt (Roche
Diagnostics Australia Pty Ltd, Castle Hill, NSW, Australia)
to produce a dark blue/purple color. Sections were counter-
stained with PAS without hematoxylin and coverslipped in
an aqueous mounting medium. Negative controls included
the use of irrelevant isotype-matched antibodies in the first
or second round of staining. In some cases, serial sections
were stained with individual antibodies and compared with
double staining.
Tubular phenotypic change was examined on
the basis of de novo expression of myofibroblast
marker, ␣-SMA. No tubular expression of ␣-SMA
was detected in the 19 cases of MCD (Fig 2B).
However, de novo expression of ␣-SMA protein
was identified in small numbers of tubules in 36
of 90 cases of IgA nephropathy and 9 of 18 cases
of RPGN. Tubular ␣-SMA expression was re-
stricted to areas of tubular damage and fibrosis.
Both intact and damaged tubules showed ␣-SMA
staining. Figure 1(B) shows ␣-SMA staining of
large cuboidal tubular epithelial cells in an area
of tubulointerstitial damage with mild fibrosis.
This ␣-SMA expression in tubules clearly is
caused by large cuboidal epithelial cells, rather
ϩ
than the possibility of interstitial ␣-SMA myo-
Quantitation of Immunohistochemistry
Tubules were identified by morphological characteristics
on the basis of PAS staining of the tubular basement mem-
brane. Tubular expression of ␣-SMA was defined as contain-
ing one or more ␣-SMAϩ cells and expressed as a percent-
fibroblasts infiltrating into the tubule (Fig 1B).
Tubular ␣-SMA expression was an infrequent
event. Quantitation of the percentage of tubules
expressing ␣-SMA is shown in Fig 2 and com-
pared with the degree of interstitial fibrosis and
age of total cortical tubules (500 to 1,000 cells) examined.
Interstitial ␣-SMAϩ myofibroblasts were counted in high-
ϩ
number of interstitial ␣-SMA myofibroblasts.
power fields (original magnification ϫ400) covering the
A second way to examine tubular phenotypic
change is to look for the loss of a specific
epithelial marker, such as cytokeratin. Figure 1C
shows that cytokeratin is expressed by all tubules
in normal human kidney. Double immunostain-
ing provided clear evidence of tubular pheno-
typic change. Figure 1D through F shows the
presence of tubular cells that express both cyto-
keratin and ␣-SMA and tubular cells that express
2
entire cortical interstitium using a 0.02-mm graticule fit in
the eyepiece of the microscope, and data are expressed as
ϩ
␣
-SMA cells per square millimeter. The area of positive
staining for collagen types I and III in the interstitium was
determined by point counting and expressed as a percentage
of interstitial involvement. Tubular staining of TGF-␤1 was
assessed in a semiquantitative fashion as follows: 0, no
staining; 1, less than 10% of tubules stained; 2, 10% to 25%
of tubules stained; 3, 25% to 50% of tubules stained, and 4,
greater than 50% of tubules stained. All scoring was per-
formed on coded slides.
␣-SMAbut have lost cytokeratin expression. The
presence of cuboidal epithelial cells expressing
both markers represents an intermediate stage in
the process of phenotypic change. Such double
staining was usually present in epithelial cells
that maintained a cuboidal appearance with rela-
tively normal tubular morphological characteris-
tics (Fig 1D). Both cuboidal epithelial cells and
flattened epithelial cells that expressed ␣-SMA
but lacked cytokeratin were seen (Fig 1D through
F). The flattened rim of elongated cells in dam-
aged tubules that express ␣-SMA, but not cyto-
keratin, suggests a change to a mesenchymal
phenotype (Fig 1D).
Statistical Analysis
Data from quantitation of immunohistochemistry staining
was compared by one-way analysis of variance using Graph-
Pad Prism 3.0 (GraphPad Software, San Diego, CA). Corre-
lation analysis for parametric data used Pearson’s single
correlation coefficient, whereas analysis of nonparametric
data used Spearman’s correlation coefficient.
RESULTS
Normal Human Kidney
Immunohistochemistry staining identified the
presence of ␣-SMA protein in smooth muscle
cells of blood vessels in normal human kidney

764
JINDE ET AL
Fig 1.

TUBULAR DIFFERENTIATION IN HUMAN GN
765
Collagen Production by Tubular Epithelial
Cells in Human Glomerulonephritis
de novo ␣-SMA expression and renal dysfunc-
tion and interstitial fibrosis, including interstitial
ϩ
␣
-SMA myofibroblasts, tubular expression of
A functional characteristic of myofibroblasts
is the production of extracellular matrix proteins,
TGF-␤1, and interstitial deposition of collagen
types I and III (Table 2). A highly significant
correlation was also evident between the number
3
-5
particularly collagen types I and III. Weak
immunohistochemistry staining of collagen types
I and III was seen in the interstitium in normal
human kidney, but this was completely absent
from tubular epithelial cells (not shown). An
increase in interstitial deposition of collagen types
I and III was evident in IgA nephropathy and
RPGN (Fig 2). In addition, some tubules in areas
with marked interstitial collagen deposition con-
tained small numbers of epithelial cells that
showed strong staining for collagen types I and
ϩ
of interstitial ␣-SMA myofibroblasts, intersti-
tial collagen deposition, renal dysfunction, and
interstitial fibrosis (Table 2).
DISCUSSION
This study has identified phenotypic change in
tubular epithelial cells in areas of tubular damage
and fibrosis in human glomerulonephritis. This is
based on immunohistochemistry staining show-
ing de novo expression of the mesenchymal
marker ␣-SMA and loss of the epithelial marker
cytokeratin by cuboidal tubular epithelial cells.
In addition, these phenotypic changes were asso-
ciated with tubular collagen production.
III. Figure 3A shows a tubule in an area of
fibrosis in which several ␣-SMAϩ cells are
present. Sections from the same tubule show the
presence of cells positively stained for collagen
type I and type III (Fig 3B and C).
Although transdifferentiation of epithelial cells
into mesenchymal cells has been described in
Tubular TGF-␤ Production
The growth factor TGF-␤1 has been impli-
cated in the process of interstitial fibrosis.
fetal development and a number of pathological
3
2,33
19-22
conditions,
there is little evidence to support
TGF-␤1 was detected in a small number of
tubules in MCD and was upregulated signifi-
cantly in IgA nephropathy and RPGN (Fig 2F).
Areas of severe fibrosis showed strong tubular
TGF-␤1 expression. Strong TGF-␤1 staining of
the occurrence of tubular epithelial cell transdif-
ferentiation into fibroblast-like cells in kidney
disease. This is a difficult hypothesis to prove for
several reasons. First, should tubular epithelial
cells transdifferentiate into fibroblast-like cells
and then migrate into the interstitium, it is no
longer possible to identify such cells as having a
tubular origin. Second, areas of interstitial fibro-
sis contain many fibroblasts in very close proxim-
ity to tubular epithelial cells. Therefore, it is
difficult to distinguish between interstitial fibro-
blasts that have entered damaged tubules from
ϩ
a tubule that contains ␣-SMA epithelial cells is
shown in Fig 3A and D.
Correlation of Tubular ␣-SMA Expression With
Renal Function and Histological Damage
There was a highly significant correlation be-
tween the percentage of cortical tubules showing
4
™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™
Fig 1. Immunohistochemistry staining showing tubular phenotypic changes in human glomerulonephritis. (A)
Normal human kidney showing ␣-SMA expression (red/brown) in vascular smooth-muscle cells of an artery, but no
staining is seen in tubules or glomeruli. (B) IgA nephropathy group 1 showing a focal area of tubulointerstitial
؉
fibrosis. Note the presence of ␣-SMA cells within the damaged area (*) that have cuboidal epithelial-like
morphological characteristics (arrows), representing an early stage of tubular phenotypic change. (C-F) Sections
were double stained for the myofibroblast marker ␣-SMA (red/brown) and the epithelial marker cytokeratin
(blue/purple). (C) Normal human kidney showing cytokeratin staining in all tubular epithelial cells, whereas
glomeruli are negative. (D) IgA nephropathy group 2 showing a focal area of tubulointerstitial fibrosis. An
intermediate stage of phenotypic change can be seen in tubular epithelial cells coexpressing cytokeratin and
␣-SMA (arrows). Tubular epithelial cells that have lost cytokeratin expression and gained ␣-SMA expression are
shown by arrowheads. This includes tubular epithelial cells with a cuboidal appearance and tubules with a flattened
appearance suggestive of epithelial differentiation into a myofibroblast-like phenotype. (E, F) RPGN showing areas
of severe tubular fibrosis. Several damaged tubules show colocalization of cytokeratin and ␣-SMA staining
؉
(
arrows), indicating an early stage in phenotypic change, whereas other ␣-SMA tubular epithelial cells can be seen
that have lost cytokeratin staining (arrowheads). (G) No staining is seen in an area of tubular damage using an
isotype-matched negative control antibody. Sections were counterstained with PAS (minus hematoxylin). (Original
magnification [A, C, G] ؋100, [B, D, E] ؋250, [F] ؋400.)

766
JINDE ET AL
Fig 2. Quantitation of tubular ␣-SMA expression and interstitial fibrosis in human glomerulonephritis. (A)
؉
Cortical interstitial fibrosis. (B) Percentage of tubules with ␣-SMA staining. (C) Interstitial ␣-SMA myofibroblasts.
D) Interstitial deposition of collagen type I. (E) Interstitial deposition of collagen type III. (F) Semiquantitative score
(
of tubular TGF-␤1 expression. Data expressed as mean ؎ SEM. Abbreviations: IgAN1, IgA nephropathy group 1;
a
IgAN2, IgA nephropathy group 2. *P < 0.05 versus MCD. **P < 0.01 versus MCD. ***P < 0.001 versus MCD. P < 0.05
b
c
versus IgA nephropathy group 1. P < 0.01 versus IgA nephropathy group 1. P < 0.001 versus IgA nephropathy
#
##
###
group 1. P < 0.05 versus IgA nephropathy group 2. P < 0.01 versus IgA nephropathy group 2. P < 0.001 versus
IgA nephropathy group 2. P by analysis of variance with (A-E) Tukey’s multiple comparison test for parametric data
and (F) Kruskal-Wallis test with Dunn’s multiple comparison for nonparametric data.

TUBULAR DIFFERENTIATION IN HUMAN GN
767
Fig 3. Immunostaining of serial sections showing colocalization of tubular ␣-SMA, collagen, and TGF-␤1 in an
area of fibrosis in RPGN. (A) De novo ␣-SMA expression by tubular epithelial cells indicating phenotypic change
(
arrows). Production of collagen types I (B) and III (C) by cells with elongated morphological characteristics (arrows)
؉
in close association with ␣-SMA cells seen in (A). (D) TGF-␤1 staining of tubular epithelial cells in this area of
fibrosis. Note that some tubular epithelial cells show elongated morphological characteristics (arrows). (Original
magnification ؋400.)
tubular epithelial cells that may have become
fibroblasts. The most readily identifiable step in a
process of epithelial-mesenchymal transdifferen-
tiation would be early phenotypic changes that
occur while the tubular epithelial cell still main-
tains its distinctive cuboidal appearance and be-
fore tubular integrity is lost. This phenotypic
change was identified in the current study by
double immunostaining for ␣-SMA and cytoker-
atin. A key aspect of this identification was the
use of PAS counterstaining to identify tubular
basement membrane. This allowed a clear distinc-
tion between ␣-SMA staining of intact tubules
from the closely associated interstitial ␣-SMA
myofibroblasts. This is probably the reason, to-
gether with the infrequent occurrence of such
cells, that tubular ␣-SMA staining has not been
reported in previous studies of human glomerulo-
nephritis.
Although phenotypic changes observed in the

768
JINDE ET AL
Table 2. Correlation Analysis of Tubular ␣-SMA Expression and Renal Function and Histological Damage
Serum
Creatinine Clearance
mg/mL)
Creatinine
Urine
Protein
(g/d)
Interstitial
Fibrosis
(%)
Interstitial
Tubular
TGF-␤1
(0-4)
Interstitial
Collagen I Collagen III
Interstitial
␣-SMAϩ
␣-SMAϩ
(
(mL/min)
tubules (%) (cell/mm2)
(%)
(%)
␣
-SMAϩ tubules (%)
0.419
Ͻ0.0001
Ϫ0.294
0.0011
Ϫ0.01
NS
0.469
Ͻ0.0001
—
—
0.317
0.0003
0.262
0.45
0.41
P
0.0029 Ͻ0.0001 Ͻ0.0001
ϩ
Interstitial ␣-SMA
2
(
cells/mm )
0.510
Ͻ0.0001 Ͻ0.0001
Ϫ0.519
Ϫ0.063
NS
0.692
Ͻ0.0001
0.317
0.0003
—
—
0.381
0.516
0.612
P
Ͻ0.0001 Ͻ0.0001 Ͻ0.0001
NOTES. Results shown as correlation coefficient. All analyses used Pearson’s correlation coefficient, except for tubular
TGF-␤ and FGF-2, which used Spearman’s correlation coefficient. Analysis was performed on results from the entire 127
patients.
Abbreviation: NS, not significant.
current study are not de facto proof of epithelial-
mesenchymal transdifferentiation in human glo-
merulonephritis, they support the hypothesis that
transdifferentiation has a role in interstitial fibro-
sis in glomerulonephritis. A key functional char-
acteristic of myofibroblasts is collagen produc-
tion. The association of collagen production and
phenotypic change in tubular epithelial cells is
consistent with the process of transdifferentiation,
although again, this does not constitute proof of
transdifferentiation. However, phenotypic changes
observed in this study of human glomerulonephri-
tis are important because they validate the rel-
evance of animal studies in which evidence of
tubular epithelial-mesenchymal transdifferentia-
tion has been presented. Studies of mouse models
of antitubular basement membrane disease and
polycystic kidney disease have shown pheno-
typic evidence of epithelial-mesenchymal trans-
differentiation based on tubular epithelial cell
sion. There was a highly significant correlation
between tubular TGF-␤1 expression and de novo
␣-SMA expression in tubular epithelial cells.
Although these results do not prove that TGF-␤1
induces tubular phenotypic change, they are con-
sistent with cell-culture studies showing that
TGF-␤1, either alone or in combination with
epithelial growth factor (EGF), can induce mes-
enchymal transdifferentiation of tubular epithe-
2
3,24
lial cell lines.
In summary, this study provides the first evi-
dence that tubular epithelial cells can undergo
phenotypic change toward a myofibroblast-like
cell during progressive renal fibrosis in human
glomerulonephritis on the basis of de novo
␣-SMA expression, loss of cytokeratin, and de
novo collagen staining. These events were re-
stricted to areas of fibrosis. In particular, identifi-
cation of cuboidal tubular epithelial cells coex-
pressing ␣-SMA and cytokeratin is believed to
represent the earliest stage of phenotypic change
in tubular epithelial cells. Although not conclu-
sive, these studies support the hypothesis that
tubular epithelial-myofibroblast transdifferentia-
tion has a role in progressive interstitial fibrosis
in human glomerulonephritis.
2
6,29
expression of fibroblast-specific protein-1.
We observed tubular epithelial-myofibroblast
transdifferentiation in the rat remnant kidney
based on de novo expression of ␣-SMA by
tubular epithelial cells and morphological charac-
teristics with the presence of characteristic actin
microfilaments and dense bodies within tubular
REFERENCES
27
epithelial cells with intact tight junctions. These
results were confirmed in a separate study of the
1
. Strutz F, Muller GA: Interstitial pathomechanisms un-
derlying progressive tubulointerstitial damage. Kidney Blood
Press Res 22:71-80, 1999
2
8
rat remnant kidney. In addition, differentiation
of cultured tubular epithelial cells into fibroblast-
2
. Badid C, Mounier N, Costa AM, Desmouliere A: Role
2
3-25
like cells has been reported.
of myofibroblasts during normal tissue repair and excessive
scarring: Interest of their assessment in nephropathies. His-
tol Histopathol 15:269-280, 2000
What factors regulate this phenotypic change
of tubular epithelial cells? TGF-␤1 has been
3
1
. Tang WW, Van GY, Qi M: Myofibroblast and ␣ (III)
3
2,33
implicated in promoting renal fibrosis.
The
collagen expression in experimental tubulointerstitial nephri-
tis. Kidney Int 51:926-931, 1997
4. Goumenos D, Tsomi K, Iatrou C, Oldroyd S, Sungur
current study showed that tubular ␣-SMA stain-
ing was associated with tubular TGF-␤1 expres-

TUBULAR DIFFERENTIATION IN HUMAN GN
769
A, Papaioannides D, Moustakas G, Ziroyannis P, Moun-
tokalakis T, El Nahas AM: Myofibroblasts and the progres-
sion of crescentic glomerulonephritis. Nephrol Dial Trans-
plant 13:1652-1661, 1998
and mesenchyme: Normal, pathological, and experimentally
induced. Am J Kidney Dis 26:678-690, 1995
21. Gressner AM: Transdifferentiation of hepatic stellate
cells (Ito cells) to myofibroblasts: A key event in hepatic
fibrogenesis. Kidney Int Suppl 54:S39-S45, 1996
22. Tang L, Tanaka Y, Marumo F, Sato C: Phenotypic
change in portal fibroblasts in biliary fibrosis. Liver 14:76-
84, 1994
5
. Hogemann B, Gillessen A, Bocker W, Rauterberg J,
Domschke W: Myofibroblast-like cells produce mRNA for
type I and III procollagens in chronic active hepatitis. Scand
J Gastroenterol 28:591-594, 1993
6
. Zhang HY, Gharaee-Kermani M, Zhang K, Karmiol S,
23. Okada H, Danoff TM, Kalluri R, Neilson EG: Early
role of Fsp1 in epithelial-mesenchymal transformation. Am J
Physiol 273:F563-F574, 1997
24. Fan J-M, Ng Y-Y, Hill PA, Nikolic-Paterson DJ, Mu
W, Atkins RC, Lan HY: Transforming growth factor-␤
regulates tubular epithelial-myofibroblast transdifferentia-
tion in vitro. Kidney Int 56:1455-1467, 1999
Phan SH: Lung fibroblast alpha-smooth muscle actin expres-
sion and contractile phenotype in bleomycin-induced pulmo-
nary fibrosis. Am J Pathol 148:527-537, 1996
7
. Alpers CE, Hudkins KL, Gown AM, Johnson RJ:
Enhanced expression of “muscle-specific” actin in glomeru-
lonephritis. Kidney Int 41:1134-1142, 1992
8
. Alpers CE, Hudkins KL, Floege J, Johnson RJ: Human
25. Healy E, Leonard M, Madrigal-Estebas L, O’Farrelly
C, Watson AJ, Ryan MP: Factors produced by activated
leukocytes alter renal epithelial cell differentiation. Kidney
Int 56:1266-1269, 1999
renal cortical interstitial cells with some features of smooth
muscle cells participate in tubulointerstitial and crescentic
glomerular injury. J Am Soc Nephrol 5:201-209, 1994
9
. Goumenos DS, Brown CB, Shortland J, El Nahas AM:
26. Strutz F, Okada H, Lo CW, Danoff T, Carone RL,
Tomaszewski JE, Neilson EG: Identification and character-
ization of a fibroblast marker: FSP1. J Cell Biol 130:393-
405, 1995
27. Ng YY, Huang TP, Yang WC, Chen ZP, Yang AH,
Nikolic-Paterson DJ, Atkins RC, Lan HY: Tubular epithelial-
myofibroblast transdifferentiation in progressive tubulointer-
stitial fibrosis in 5/6 nephrectomized rats. Kidney Int 53:864-
876, 1998
28. Frazier KS, Paredes A, Dube P, Styer E: Connective
tissue growth factor expression in the rat remnant kidney
model and association with tubular epithelial cells undergo-
ing transdifferentiation. Vet Pathol 37:328-335, 2000
29. Okada H, Ban S, Nagao S, Takahashi H, Suzuki H,
Neilson EG: Progressive renal fibrosis in murine polycystic
kidney disease: An immunohistochemical observation. Kid-
ney Int 58:587-597, 2000
Myofibroblasts, predictors of progression of mesangial IgA
nephropathy? Nephrol Dial Transplant 9:1418-1425, 1994
1
0. Hewitson TD, Becker GJ: Interstitial myofibroblasts
in IgA glomerulonephritis. Am J Nephrol 15:111-117, 1995
1. Roberts IS, Burrows C, Shanks JH, Venning M,
1
Mcwilliam LJ: Interstitial myofibroblasts: Predictors of pro-
gression in membranous nephropathy. J Clin Pathol 50:123-
1
27, 1997
2. Essawy M, Soylemzoglu O, Muchaneta-kubara EC,
1
Shortland J, Brown CB, El Nahas AM: Myofibroblasts and
the progression of diabetic nephropathy. Nephrol Dial Trans-
plant 12:43-50, 1997
1
3. Badid C, Desmouliere A, McGregor B, Costa AM,
Fouque D, Hadj Aissa A, Laville M: Interstitial alpha-
smooth muscle actin: A prognostic marker in membranous
nephropathy. Clin Nephrol 52:210-217, 1999
1
4. Zhang G, Moorhead PJ, El Nahas AM: Myofibro-
30. Clarkson AR, Woodroffe AJ, Aarons L: IgA nephrop-
athy and Henoch-Sch o¨ nlein purpura, in Schrier RW,
Gottschalk CW (eds): Diseases of the Kidney (ed 6). Boston,
MA, Little, Brown, 1996, pp 1645-1670
blasts and the progression of experimental glomerulonephri-
tis. Exp Nephrol 3:308-318, 1995
1
5. Hewitson TD, Wu HL, Becker GJ: Interstitial myofi-
broblasts in experimental renal infection and scarring. Am J
Nephrol 15:411-417, 1995
31. Lan HY, Mu W, Nikolic-Paterson DJ, Atkins RC: A
novel, simple, reliable, and sensitive method for multiple
immunoenzyme staining: Use of microwave oven heating to
block antibody crossreactivity and retrieve antigens. J Histo-
chem Cytochem 43:97-102, 1995
32. Border WA, Noble NA, Yamamoto T, Harper JR,
Yamaguchi Yu, Pierschbacher MD, Ruoslahti E: Natural
inhibitor of transforming growth factor-beta protects against
scarring in experimental kidney disease. Nature 360:361-
364, 1992
1
6. Diamond JR, van Goor H, Ding G, Engelmyer E:
Myofibroblasts in experimental hydronephrosis. Am J Pathol
46:121-129, 1995
7. Yang N, Wu LL, Nikolic-Paterson DJ, Ng Y-Y, Yang
1
1
W-C, Mu W, Gilbert RE, Cooper ME, Atkins RC, Lan HY:
Local macrophage and myofibroblast proliferation in progres-
sive renal injury in the rat remnant kidney. Nephrol Dial
Transplant 13:1967-1974, 1998
1
8. Gabbiani G: The cellular derivation and the life span
of the myofibroblast. Pathol Res Pract 192:708-711, 1996
9. Ekblom P, Weller A: Ontogeny of tubulointerstitial
cells. Kidney Int 39:394-400, 1991
0. Hay ED, Zuk A: Transformations between epithelium
33. Isaka Y, Akagi Y, Ando Y, Tsujie M, Sudo T, Ohno N,
Border WA, Noble NA, Kaneda Y, Hori M, Imai E: Gene
therapy by transforming growth factor-beta receptor-IgG Fc
chimera suppressed extracellular matrix accumulation in
experimental glomerulonephritis. Kidney Int 55:465-75, 1999
1
2