Distinct In Vivo Roles of Colony-Stimulating
Factor-1 Isoforms in Renal Inflammation
This information is current as
of June 18, 2017.
Mei-Huei Jang, Deborah M. Herber, Xinnong Jiang, Sayan
Nandi, Xu-Ming Dai, Geraldine Zeller, E. Richard Stanley
and Vicki R. Kelley
J Immunol 2006; 177:4055-4063; ;
doi: 10.4049/jimmunol.177.6.4055
http://www.jimmunol.org/content/177/6/4055
Subscription
Permissions
Email Alerts
This article cites 47 articles, 21 of which you can access for free at:
http://www.jimmunol.org/content/177/6/4055.full#ref-list-1
Information about subscribing to The Journal of Immunology is online at:
http://jimmunol.org/subscription
Submit copyright permission requests at:
http://www.aai.org/About/Publications/JI/copyright.html
Receive free email-alerts when new articles cite this article. Sign up at:
http://jimmunol.org/alerts
The Journal of Immunology is published twice each month by
The American Association of Immunologists, Inc.,
1451 Rockville Pike, Suite 650, Rockville, MD 20852
Copyright © 2006 by The American Association of
Immunologists All rights reserved.
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
References
The Journal of Immunology
Distinct In Vivo Roles of Colony-Stimulating Factor-1 Isoforms
in Renal Inflammation1
Mei-Huei Jang,* Deborah M. Herber,* Xinnong Jiang,* Sayan Nandi,† Xu-Ming Dai,†
Geraldine Zeller,* E. Richard Stanley,† and Vicki R. Kelley2*
M
acrophage (M␾)3-rich infiltrates are a hallmark of inflammation in a broad range of diseases. We have determined that activated M␾ mediate apoptosis of renal
resident cells, most notably tubular epithelial cells (TEC) (1).
Since activated M␾ accumulate in the kidney during nephritis,
these leukocytes are instrumental in the destruction of the kidney.
With this in mind, we suspected that growth factors that support
M␾ are instrumental in kidney injury and prime therapeutic target
candidates to combat renal inflammation.
CSF-1, also known as M-CSF, is the primary regulator of M␾
survival, proliferation, and differentiation (2, 3) and is involved in
M␾-mediated nephritis. We identified a strong association between CSF-1, M␾, and inflammation in MRL-Faslpr mice that
share features with human lupus (4). Specifically, we determined
*Laboratory of Molecular Autoimmune Disease, Renal Division, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA 02115; and †Department of
Developmental and Molecular Biology, Albert Einstein College of Medicine, Bronx,
NY 10461
Received for publication April 18, 2006. Accepted for publication June 28, 2006.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
1
This work was supported by National Institutes of Health Grants 36149 and 52369
(to V.R.K.) and CA32551 (to E.R.S.), Albert Einstein College of Medicine Cancer
Center Grant 5P30-CA13330, an American Society of Hematology Fellow Scholar
Award (to X.-M.D.), and a Leukemia and Lymphoma Society Special Fellow Award
(to X.-M.D.).
2
Address correspondence and reprint requests to Dr. Vicki R. Kelley, Brigham and
Women’s Hospital, Harvard Institutes of Medicine, 4 Black Fan Circle, Boston, MA
02115. E-mail address: [email protected]
3
Abbreviations used in this paper: M␾, macrophage; TEC, tubular epithelial cell;
WT, wild type; UUO, unilateral ureteral obstruction; sgCSF-1, secreted glycoprotein
CSF-1; spCSF-1, secreted proteoglycan CSF-1; csCSF-1, cell surface CSF-1; osteopetrotic, Csf1op/Csf1op; ChS, chondroitin sulfate; sppCSF-1, secreted proteoglycan
precursor of CSF-1; sgpCSF-1, secreted glycoprotein precursor of CSF-1; TgCS, cell
surface transgene; TgSPP, secreted proteoglycan precursor transgene; TgSGP, secreted glycoprotein precursor transgene; EGFP, enhanced GFP; CL, contralateral;
␤-gal, 5-bromo-4-chloro-3-indolyl-␤-D-galactopyranoside; LTL, lotus tetragonolobus
lectin; DBA, dolichos biflorus aggulutinin; iNOS, inducible NO synthase; DC, dendritic cell.
Copyright © 2006 by The American Association of Immunologists, Inc.
that CSF-1 expression is increased in the circulation and kidney
before overt renal pathology and becomes more abundant with
advancing nephritis (5). We have shown that renal resident cells,
most notably TEC, are the principal source of CSF-1 during lupus
nephritis (6) and that M␾ and T cells localize in intrarenal sites
rich in CSF-1 (6). Gene transfer of CSF-1 into the kidney in autoimmune-prone mice recruits M␾ and initiates inflammation (7).
Furthermore, CSF-1-deficient MRL-Faslpr mice are protected
from nephritis and the systemic illness characteristic of the MRLFaslpr wild-type (WT) strain (8). Similarly, in unilateral ureteral
obstruction (UUO) in which blocking the flow of urine results in a
florid M␾ infiltration into the renal interstitium that leads to tubular damage and interstitial fibrosis (9 –11), we determined that renal injury is dependent on CSF-1 (12). In the CSF-1 null mice,
fewer M␾ accumulate within the kidney, and moreover, they do
not proliferate and fewer are activated than in WT mice. The net
result is a decrease in TEC apoptosis. Therefore, the consequence
of eliminating CSF-1 is that there are fewer harmful M␾ in the
kidney to cause tissue damage. Taken together, the findings with
MRL-Faslpr and UUO mice indicate that CSF-1 mediates M␾dependent immune and nonimmune incited kidney injury.
Three distinct isoforms of CSF-1 have been identified. The fulllength, primary CSF-1 transcript encodes a membrane-spanning
precursor protein from which either a secreted glycoprotein
(sgCSF-1) or a secreted proteoglycan (spCSF-1) are cleaved in the
secretory vesicle. Splicing out of the region encoding the proteolytic cleavage sites and the glycosaminoglycan addition site from
this transcript creates an mRNA that encodes the biologically active membrane-spanning, cell surface glycoprotein (csCSF-1) (2).
Evidence suggests that these CSF-1 isoforms have both shared and
unique features based on stability and availability. Although
csCSF-1 is relatively stably expressed with a half-life on the cell
surface of ⬃7 h (13, 14), it has a limited range since this isoform
requires cell-cell contact to exert its effects on M␾ (2, 3). The two
secreted isoforms of CSF-1 comprise circulating CSF-1, which has
0022-1767/06/$02.00
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
CSF-1, the major regulator of macrophage (M␾) development, has three biologically active isoforms: a membrane-spanning, cell
surface glycoprotein, a secreted glycoprotein, and a secreted proteoglycan. We hypothesized that there are shared and unique roles
of individual CSF-1 isoforms during renal inflammation. To test this, we evaluated transgenic mice only expressing the cell surface
or precursors of the secreted CSF-1 isoforms for M␾ accumulation, activation, and M␾-mediated tubular epithelial cell (TEC)
apoptosis during unilateral ureteral obstruction. The only difference between secreted proteoglycan and secreted glycoprotein
CSF-1 isoforms is the presence (proteoglycan) or absence (glycoprotein) of an 18-kDa chondroitin sulfate glycosaminoglycan. We
report that 1) cell surface CSF-1 isoform is sufficient to restore M␾ accumulation, activation, and TEC apoptosis to wild-type
levels and is substantially more effective than the secreted CSF-1 isoforms; 2) the chondroitin sulfate glycosaminoglycan facilitates
M␾ accumulation, activation, and TEC apoptosis; 3) increasing the level of secreted proteoglycan CSF-1 in serum amplifies renal
inflammation; and 4) cell-cell contact is required for M␾ to up-regulate CSF-1-dependent expression of IFN-␥. Taken together,
we have identified central roles for the cell surface CSF-1 and the chondroitin sulfate chain on secreted proteoglycan CSF-1 during
renal inflammation. The Journal of Immunology, 2006, 177: 4055– 4063.
4056
Materials and Methods
Mice
CSF-1-deficient osteopetrotic mice and littermate control mice (⫹/Csf1op
or ⫹/⫹) (referred to as WT) were backcrossed onto the FVB/NJ background for at least 10 generations. The plasmid bearing the full-length
CSF-1 driven by the CSF-1 promoter and the first intron (20) was used in
the construction of the cell surface (TgCS), secreted proteoglycan precursor (TgSPP), and secreted glycoprotein precursor (TgSGP) transgenes (17,
19). The breeding, genotyping and nomenclature of these transgenic mice
have been described previously (17, 19, 20). The following transgenic lines
were used: 1) TgN(FLCsf1)Ers10/⫹; Csf1op/Csf1op (referred to as TgC/⫹);
2) TgN(CSCsf1)Ers5/⫹;Csf1op/Csf1op (referred to as TgCS/⫹); 3)
TgN(SPPCsf1)Ers7/⫹;Csf1op/Csf1op (referred to as TgSPP/⫹); 4)
TgN(SPPCsf1)Ers2/⫹;Csf1op/Csf1op (referred to as TgSPP-2x/⫹); 5)
TgN(SGPCsf1)Ers4/⫹;Csf1op/Csf1op (referred to as TgSGP/⫹); 6)
TgN(SGPCsf1)Ers2/⫹;Csf1op/Csf1op (referred to as TgSGP-2x/⫹); and 7)
TgN(Csf1-Z)Ers7/⫹ (referred to as TgZ) in which lacZ expression is driven
by the same Csf1 promoter first intron sequence used to construct the other
transgenes (20). These lines were housed and bred in the pathogen-free
animal facilities of the Albert Einstein College of Medicine and Harvard
Medical School. Transgenic mice (C57BL/6 ⫻ CBA)F1 expressing the
enhanced green fluorescence protein (EGFP) under the control of the
CSF-1R (c-fms) promoter and first intron (Tgfms-EGFP), referred to as
MacGreen transgenic mice, were provided by Dr. D. A. Hume (University
of Queensland, Brisbane, Australia) (29). The MacGreen mice were bred
and housed at Harvard Medical School. We fed all mice a normal laboratory chow diet; however, since the CSF-1-deficient (Csf1op/Csf1op) strain
lacks incisors, this chow was provided as a powder. We evaluated similar
numbers of females and males in all experiments. The use of mice in this
study was reviewed and approved by the Standing Committee on Animals
in the Harvard Medical School in adherence to the National Institutes of
Health Guide for the care and use of laboratory animals.
Unilateral ureteral obstruction
UUO was performed on adult mice as described previously (30). Mice
were sacrificed 3 days after each UUO experiment, and the obstructed and
contralateral (CL) kidneys were removed. After bisecting the kidneys, portions were reserved for immunostaining, flow cytometry, and light microscopy as described below.
CSF-1 and CSF-1R expression
CSF-1 concentrations in the serum were measured using a CSF-1 radioimmunoassay that detects only biologically active CSF-1 as described previously (18, 31). TgZ mice were used to report the number, location, and
cell type of CSF-1-expressing cells in kidneys. To identify ␤-galactosidase
encoded by lacZ, we sectioned (10 ␮m) snap-frozen kidneys, and fixed
them in cold paraformaldehyde (2%). These sections were washed with
PBS, rinsed with distilled water, and incubated with a solution containing
5-bromo-4-chloro-3-indolyl-␤-D-galactopyranoside (␤-gal) at 37°C as described previously (20). This yields a blue reaction product with ␤-galactopyranoside. To determine the location and type of tubules expressing
CSF-1, we stained the same section for the presence of ␤-gal with lotus
tetragonolobus lectin (LTL) (Vector Laboratories), which binds to proximal tubules, and dolichos biflorus aggulutinin (Vector Laboratories), which
binds to distal tubules and collecting ducts (32, 33). For this purpose,
following the development of ␤-gal, these sections were washed with PBS
and incubated (1 h) with fluorescein-labeled LTL (1/400) and rhodaminelabeled DBA (1/400). We determined the relative CSF-1 expression in the
tubules by counting the number of ␤-gal faintly and intensely stained tubules, and the percentage of LTL or DBA staining tubules expressing ␤-gal
faintly and intensely were counted in 10 randomly selected low power
(⫻100) fields using coded slides. To determine CSF-1 expression in glomeruli, we counted the number of ␤-gal staining cells per glomerulus in 10
randomly selected glomeruli using coded slides.
To identify the number and location of c-fms bearing cells in the kidney,
we excised kidneys from MacGreen transgenic mice. We fixed these kidneys in paraformaldehyde (4%) for 2 h, followed by overnight incubation
in sucrose (18%) at 4°C. The tissues were embedded in Tissue-Tek OCT
(Sakura Finetek), snap frozen, sectioned (4 ␮m), mounted (VectaShield
fluorescence mounting medium; Vector Laboratories), and analyzed using
a Nikon Eclipse E1000 upright fluorescence microscope. We analyzed sequential sections for the presence of EGFP (c-fms) and CD68. The number
of EGFP- and CD-68-bearing cells were enumerated in 10 randomly selected high-power (⫻400) fields within the interstitium and in 10 randomly
selected glomeruli using coded slides. To determine whether these CSF1R-bearing M␾ were adjacent to proximal tubules during UUO, we stained
the same sections from MacGreen obstructed kidneys for LTL.
Identification of M␾ by immunostaining
Kidneys were snap frozen, embedded in Tissue-Tek OCT, and the blocks
were sectioned using a cryostat. M␾ were stained for the presence of CD68
using a rat anti-mouse CD68 Ab (1/100 dilution; Serotec) with the immunoperoxidase technique, as reported previously (34). The number of M␾ in
the kidney was determined by counting the CD68-positive cells in 20 randomly selected high-power fields within the interstitium and 20 randomly
selected glomeruli using coded slides.
Flow cytometry
Following systemic perfusion with cold PBS, we excised the obstructed
kidneys. To prepare kidney single-cell suspensions, we gently pressed the
kidneys through a cell strainer using a rubber syringe plunger. We treated
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
a half-life in the circulation of only ⬃10 min (15). However, circulating CSF-1 can reach many distal sites (2, 16 –19). We suspected that the site of expression, stability, and concentration of
each CSF-1 isoform is instrumental in mediating renal
inflammation.
To determine the shared and unique roles of the individual
CSF-1 isoforms during renal inflammation, it is critical to dissect
the action of each isoform in vivo. The full-length CSF-1 mRNA
encodes all three CSF-1 isoforms (19, 20). Using a transgene encoding this mRNA driven by the 6.5-kb nucleotide sequence that
resides upstream of the coding region of the CSF-1 gene, transgenic mice were created on the CSF-1-deficient osteopetrotic
(Csf1op/Csf1op) mouse background that restored the normal tissuespecific and developmental expression of CSF-1 (20). With this in
mind, we created transgenic mice in which the same 6.5-kb driver
was used to drive expression of precursors of the individual CSF-1
isoforms in a normal tissue-specific and developmental pattern (17,
19). We then induced renal inflammation in these mice using UUO
and evaluated the impact of M␾-dependent events that result in
renal injury. To confirm that the CSF-1 driver efficiently regulates
the expression of individual CSF-1 isoforms in transgenic strains,
we demonstrated restoration of renal inflammation in CSF-1-deficient mice expressing a transgene in which this promoter drives
expression of full-length CSF-1.
The secreted proteoglycan isoform of mouse CSF-1 contains a
single 18-kDa chondroitin sulfate (ChS) chain per subunit that is
covalently attached to the protein backbone at Ser276 (13, 19).
Although the bioactivities of many cytokines (e.g., basic growth
factor-1, fibroblast growth factor, and hepatocyte growth factor)
and chemokines (e.g., platelet factor-4, stromal factor-1, IL-8, neutrophil-activating peptide-2, growth-related oncogene-␣, and
RANTES) are regulated by glycosaminoglycans (21–28), CSF-1 is
the only cytokine/growth factor reported to exist as a proteoglycan,
and the glycosaminoglycan attachment site consensus sequence of
Ser-Gly-X-Gly/Ala in CSF-1 is highly conserved. In addition, the
spCSF-1 and sgCSF-1 contribute approximately equally to the total circulating CSF-1, i.e., each generates ⬃50% (14). The important role of the ChS chain in in vivo signaling by secreted CSF-1
was established recently in studies using mice that exclusively express either the secreted proteoglycan precursor of CSF-1 (sppCSF-1) or the corresponding glycoprotein precursor (sgpCSF-1)
in which the glycosaminoglycan attachment site was mutated (19).
The secreted CSF-1 in these strains only differs in the absence, or
presence, of the ChS chain. Therefore, by using sppCSF-1 and
sgpCSF-1 transgenic mice expressing a portion or the total amount
of circulating CSF-1 in homozygous WT Csf1 (⫹/⫹) mice, we
have been able to explore the role of the spCSF-1 and sgCSF-1
isoforms during renal inflammation. We now report that csCSF-1
is sufficient for renal inflammation and have identified an unexpected role for spCSF-1.
ROLE OF CSF-1 ISOFORMS IN RENAL INFLAMMATION
The Journal of Immunology
Detection of apoptotic TEC using the TUNEL assay
We prepared sections (4 ␮m) from formalin-fixed (10%) paraffin-embedded blocks. Apoptotic cells were identified in the kidneys by enzymatic in
situ labeling of apoptosis-induced DNA strand breaks (TUNEL method)
FIGURE 1. CSF-1 expression is primarily in proximal tubules and is increased in the kidney during UUO.
Csf1 promoter-lacZ (TgZ/⫹) reporter mice were used to
determine the expression pattern of CSF-1 in the kidney
during UUO. Fluorescence-labeled lectins were used to
identify the renal tubules; fluorescein-labeled LTL
stains proximal tubules (apical aspect), whereas rhodamine-labeled DBA stains distal tubules and collecting
ducts. The same sections were stained for the expression
of ␤-gal (CSF-1), LTL, and DBA (A, B, and C, respectively). The level of ␤-gal staining (predominantly nuclear) was recorded as intense (thick arrowhead) and
weak (thin arrows) (A). ␤-gal is expressed by proximal
tubules and is not expressed by distal tubules or collecting ducts (A–C). Note arrowheads (A and B) indicate
colocalization, and clear arrows (A and C) indicate lack
of colocalization. Hatched bars in F indicate percentage
of LTL⫹ tubules staining intensely for ␤-gal, and open
bars indicate percentage of LTL⫹ tubules weakly staining for ␤-gal. The amount of ␤-gal staining, and the
number of proximal tubules expressing ␤-gal (intense
plus weak) increases in the obstructed kidney (76%), as
compared with the CL kidney (46%) (A, D, and F, upper
panel). Similarly, the number of ␤-gal-expressing cells
in glomeruli increases in the obstructed kidney (A, inset,
arrows; F, lower panel) as compared with the CL kidney
(D, inset, arrow; F, lower panel). Of note, the number of
tubules expressing ␤-gal increases in the CL kidneys
(46%) as compared with unmanipulated kidneys (23%)
(D–F). The number of ␤-gal staining tubules (F, upper
panel) was counted in 10 randomly selected low-power
fields (⫻100). The number of ␤-gal-staining cells per glomerulus (F, lower panel) was counted in 10 randomly selected glomeruli. Values are means ⫾ SEM, ⴱ, p ⬍ 0.05.
The data are representative of three experiments.
using the TdT FragEL DNA Fragmentation Detection kit (Oncogene), according to the manufacturer’s instructions. We identified apoptotic TEC by
morphological criteria and counted the number of apoptotic TEC in 10
random high-power fields using coded slides.
Statistical analysis
The data are presented as means ⫾ SEM. The Mann-Whitney U test was
used to test significance. Differences were considered statistically significant for comparisons of data sets yielding p values ⱕ 0.05.
Results
CSF-1 is preferentially expressed in proximal tubules and is
increased in obstructed kidneys
To identify CSF-1-expressing cells in the kidney during renal inflammation, we compared the intensity of ␤-gal expression in obstructed, nonobstructed CL, and unmanipulated kidneys of Csf-1
promotor lacZ (TgZ) reporter mice (Fig. 1). Using fluorescencelabeled tubule-specific markers (LTL identifying proximal tubules,
DBA identifying distal tubules and collecting ducts) on the same
sections prepared for ␤-gal expression, we have localized CSF-1
intense (arrowheads) and weak (thin arrows) expression primarily
in the proximal tubules (Fig. 1, A and B). By comparison, we
detected weak staining within a few cells in glomeruli (Fig. 1A,
inset) in obstructed kidneys. In contrast, we did not detect ␤-gal
expression in distal tubules or collecting ducts in these kidneys
(clear arrow, Fig. 1, A and C). There was an increase in the number
of ␤-gal-expressing cells and amount of ␤-gal in the obstructed
kidney as compared with the CL kidney (Fig. 1, A and D). The
total ␤-gal staining (76%, intense staining (o)), plus weak staining
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
these cell suspensions with ACK lysis buffer (BioSource International) to
remove RBC, washed them with PBS twice, and resuspended them in
FACS buffer (PBS, 5% FBS, and 0.09% NaN3).
To determine M␾ activation, we evaluated the expression of cell surface
markers, including CD23, CD69, and Iak and Iab on M␾ as follows: cells
were stained with PE-labeled anti-mouse CD23 Ab, anti-mouse CD69 Ab
or 1/1 mixture of anti-mouse Iak and anti-mouse Iab Ab, washed with
FACS buffer, then permeabilized with saponin, and stained with FITClabeled anti-mouse CD68 (Serotec). To identify intracellular markers characteristic of M␾ activation, we measured the expression of inducible NO
synthase (iNOS), IFN-␥, and TNF-␣. iNOS was detected by permeabilizing cells with saponin and staining with FITC-labeled anti-iNOS Ab (BD
Pharmingen) and PE-labeled anti-CD68 Ab (Serotec). To detect IFN-␥ and
TNF-␣, we stimulated cells with leukocyte activation mixture with Golgi
plug (BD Pharmingen) for 4 h at 37°C, washed with FACS buffer, permeabilized with saponin, and then stained with PE-labeled anti-IFN-␥ or antiTNF-␣ Abs (eBioscience) and FITC labeled anti-CD68 Ab, respectively.
To determine the percentage of CD68 bearing cells that are dendritic
cells (DC), we dual stained for the presence of CD68 and the DC marker,
CD11c. For this purpose, kidney cells were stained with PE-labeled antimouse CD11c (BD Pharmingen) at a final concentration of 5 ␮g/ml for 30
min on ice, washed twice with FACS buffer, permeabilized with saponin,
and then stained with FITC-labeled anti-mouse CD68. Cells were incubated with Abs for 30 min on ice in the dark. After incubating with antiCD68 Ab, the cells were washed with FACS buffer to allow resealing of
the membrane and fixed in paraformaldehyde (1%).
Isotype-matched Abs (BD Pharmingen or eBioscience) were included in
each experiment as controls. We evaluated 10,000 cells using a BD Biosciences FACSCalibur, and the data were analyzed with CellQuest
software.
4057
4058
ROLE OF CSF-1 ISOFORMS IN RENAL INFLAMMATION
(䡺), in the proximal tubules of the obstructed kidneys was increased as compared with the CL kidneys (46%) (Fig. 1F, upper
panel). In addition, there were far more (⬎4-fold) ␤-gal intensely
stained proximal tubules in the obstructed kidney as compared
with the CL kidney (Fig. 1F, upper panel). Similarly, we detected
an increase (⬎4-fold) in ␤-gal expression in glomeruli of the obstructed kidney as compared with the CL kidney (Fig. 1F, lower
panel). It is of interest that the level of ␤-gal expression in proximal tubules in the CL kidneys (6% intense, 40% weak) was
greater than in unmanipulated kidneys (3% intense, 21% weak)
(Fig. 1, D–F, upper panel). Taken together, CSF-1 is expressed
mainly by proximal tubules and, to a lesser degree, by glomeruli,
and both the amount of CSF-1 and the number of tubules expressing CSF-1 increase during renal inflammation.
line), in the obstructed kidneys as compared with the CL kidneys
(Fig. 2, A–G). Thus, the CD68-bearing cells expressing c-fms localize predominantly in the interstitium adjacent to proximal tubules, the main source of CSF-1, during renal inflammation.
Some mature DC bear c-fms (35). To distinguish CD68-expressing M␾ and DC in the kidney during UUO, we identified DC by
dual staining kidney cells for the presence of CD68 and CD11c.
We concluded that few (6.0 ⫾ 1.2% SEM, n ⫽ 6) of the CD68bearing cells in the obstructed kidney and even fewer (1.9 ⫾ 0.5%
SEM, n ⫽ 5) in the CL kidney are DC. Given the paucity of
CSF-1R-bearing DC detected in the kidney during obstruction, we
conclude that M␾ are the overwhelming majority of CSF-1R-bearing leukocytes that infiltrate the kidney during UUO. Therefore,
we will refer to the CD68-bearing leukocytes as M␾.
c-fms expressing M␾ increase in the kidney during UUO
TgC expression restores renal inflammation in CSF-1-deficient
mice
TgC expression in Csf1op/Csf1op mice drives normal tissue-specific and developmental expression of the full-length CSF-1 precursor, resulting in expression of all three CSF-1 isoforms and
corrects all the major aspects of the Csf1op/Csf1op phenotype (17,
19, 20). However, it was not clear that this promoter region driving
expression of full-length CSF-1 would be sufficient to restore the
expression of CSF-1 during pathologic events. To answer this
FIGURE 2. c-fms-expressing M␾ are increased during UUO and localize adjacent to proximal tubules. Sequential kidney sections from Csf-1r promoterEGFP (MacGreen) mice provided a pattern of c-fms and CD68 as a means of identifying M␾. EGFP-expressing M␾ were increased in the obstructed kidney
as compared with the CL kidney (A–G). Note that the vast majority of CD68⫹ M␾ expressing EGFP (arrows) are not in glomeruli (encircled), but rather
within the renal interstitium adjacent to proximal tubules identified with LTL staining (E). F and G, The number of leukocytes expressing EGFP in 10
randomly selected high-power fields (⫻400) and 10 glomeruli were counted and the data graphed. Values are means ⫾ SEM, ⴱ, p ⬍ 0.05. The data are
representative of three experiments.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
Using the Csf-1r promoter-EGFP (MacGreen) mice to report
CSF-1R (c-fms) expression and sequential sections (Fig. 2, A and
B, C and D) to detect CD68 (immunostaining), we concluded that
cells expressing CD68 and c-fms localized to the same site in the
obstructed and CL kidneys during UUO. We noted an increase in
EGFP-expressing M␾ predominantly in the renal interstitium, adjacent to proximal tubules (LTL) (Fig. 2, A–D) and, to a lesser
extent within and surrounding glomeruli (circumscribed with a
The Journal of Immunology
FIGURE 3. M␾ accumulation in the TgC/⫹, TgCS/⫹, and TgSPP/⫹,
but not TgSGP/⫹ obstructed kidneys, are fully restored to the WT level.
The number of M␾ in the TgC/⫹ and TgCS/⫹ mice was fully restored to
WT levels. By comparison, the number of M␾ was only partially restored
(⬃74%) in the TgSPP/⫹ mice but was fully restored in TgSPP-2x/⫹ mice
possessing WT levels of circulating CSF-1. In contrast, the number of M␾
did not rise above Csf1op/Csf1op levels in TgSGP/⫹ and was only partially
restored (⬃65%) in TgSGP-2x/⫹ mice. Note: these transgenic CSF-1
strains are on the Csf1op/Csf1op background. The number of M␾, defined
by the expression of CD68, were counted in (A) the interstitial (20 randomly selected high-power field (HPF)) and (B) glomeruli (20 randomly
selected HPF) regions. Values are the mean ⫾ SEM, ⴱ, p ⬍ 0.001 vs WT
mice; ⴱⴱ, p ⬍ 0.001 vs Csf1op/Csf1op mice. The data represent a combination of 11 experiments.
Individual CSF-1 isoforms have differing roles in regulating M␾
accumulation during UUO
The csCSF-1 isoform is sufficient to restore intrarenal M␾ accumulation to WT levels. Restoration of csCSF-1 normalizes several but not all aspects of development in Csf1op/Csf1op mice (17).
For example, M␾ densities are restored in some (e.g., kidney cortex), but not all, tissues (e.g., adult kidney medulla, spleen, or
liver). Therefore, our goal was to determine whether csCSF-1 was
sufficient to foster the intrarenal accumulation of M␾ during renal
inflammation. We detected greater numbers of intrarenal M␾ in
the interstitium (Fig. 3A) and glomeruli (Fig. 3B) in the TgCS/⫹
mice relative to their numbers in Csf1op/Csf1op mice during UUO.
The number of M␾ within the interstitium and glomeruli in the
TgCS/⫹ strain were fully restored to the level detected in WT mice
(Fig. 3, A and B). Thus, the transmembrane CSF-1 isoform displayed on the cell surface is sufficient to restore intrarenal M␾
accumulation to WT levels during renal inflammation.
The proteoglycan isoform of secreted CSF-1 facilitates the accumulation of M␾. The precursors of the secreted CSF-1 isoforms, sppCSF-1 and sgpCSF-1, differ in that the former contains
two 18-kDa ChS chains per full-length CSF-1 dimer. To determine
whether the ChS chain facilitates M␾ accumulation during renal
inflammation, we compared the number of intrarenal M␾ in the
TgSPP/⫹ and TgSGP/⫹, Csf1op/Csf1op, and the WT strains during
UUO (Fig. 3, A and B). We detected more M␾ in the renal interstitium and glomeruli in TgSPP/⫹ as compared with the Csf1op/
Csf1op mice. However, the number of M␾ within each area of the
kidney remained lower as compared with the WT strain. In contrast, the number of intrarenal M␾ in the obstructed kidneys of
TgSGP/⫹ mice did not rise above the level detected in the Csf1op/
Csf1op strain. Furthermore, we detected far fewer M␾ in the renal
interstitium and glomeruli in the TgSGP/⫹ as compared with
the TgSPP/⫹ mice during UUO (Fig. 3, A and B). Thus, since
TgSPP/⫹ and TgSGP/⫹ mice possess equivalent concentrations of
circulating CSF-1 (Table I), the ChS chain in the spCSF-1 is instrumental in facilitating the accumulation of M␾ in the kidney
during inflammation. However, the cell surface isoform is more
effective than either secreted isoform in mediating M␾ accumulation in the kidney during UUO.
Restoring serum spCSF-1 and sgCSF-1 to WT levels increases
M␾ in the kidney, and the ChS chain facilitates this process.
The spCSF-1 and sgCSF-1 isoforms each constitute approximately
half the total amount of serum CSF-1 in WT mice. Since the circulating CSF-1 concentration in the TgSPP/⫹ and TgSGP/⫹ mice
is only one-half that of the WT (Csf1op/⫹) mice (Table I), it is
possible that both the concentration and/or type of secreted CSF-1
determine the extent of M␾ accumulation during renal inflammation. To correct for the reduction in total circulating CSF-1 in the
transgenic TgSPP/⫹ and TgSGP/⫹ mice, we selected lines in
which serum CSF-1 concentration was equivalent to the concentration in WT (⫹/⫹) mice (referred to as TgSPP-2x/⫹ and TgSGP2x/⫹) (Table I). Compared with the number of M␾ in the
TgSPP/⫹ and TgSGP/⫹ kidneys, we detected a rise in the numbers of M␾ in the TgSPP-2x/⫹ and TgSGP-2x/⫹ kidneys during
UUO (Fig. 3, A and B). Thus, restoring the amount of CSF-1 in the
circulation to the WT (⫹/⫹) levels by increasing either the secreted proteoglycan or the secreted glycoprotein precursors increases the numbers of M␾ in the kidney during renal inflammation. However, while the intrarenal M␾ accumulation was restored
to WT levels in the TgSPP-2x/⫹ mice, the numbers of M␾ in the
kidney of TgSGP-2x/⫹ mice, although greater than in the Csf1op/
Csf1op strain mice, remained below the WT level. Taken together,
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
question, we analyzed the accumulation and activation of M␾, as
well as M␾-mediated TEC apoptosis during renal inflammation
induced by UUO. We compared TgC/⫹ and WT kidneys 3 days
after UUO, since at this time intrarenal M␾ accumulation is florid
and hydronephrosis is less severe as compared with later time
points (12). Consistent with previous findings, we detected twice
the numbers of M␾ accumulated in WT-obstructed kidneys (interstitium, glomeruli) as compared with Csf1op/Csf1op-obstructed
kidneys (Fig. 3) (12). Using these WT and Csf1op/Csf1op mice as
reference points, we determined that CSF-1 is restored in the
TgC/⫹ mice during renal inflammation. The extents of intrarenal
M␾ accumulation (Fig. 3, A and B), activation (Fig. 4), and apoptosis of TEC (Fig. 5) are substantially higher in the TgC/⫹ as
compared with the Csf1op/Csf1op mice and indistinguishable from
the WT strain during UUO. Thus, TgC functionally reconstitutes
the expression of CSF-1 during renal inflammation incited by
UUO. Taken together, our data imply that the promoter we are
using effectively drives the expression of the individual CSF-1
isoforms in transgenic strains during renal inflammation.
4059
4060
ROLE OF CSF-1 ISOFORMS IN RENAL INFLAMMATION
FIGURE 4. Distinct roles of individual CSF-1 isoforms in regulating the frequency of activated M␾ in the
kidney during UUO. The frequency of M␾-expressing
cell surface markers (CD23, CD69, Ia; A) and intracellular markers (iNOS and IFN-␥; B) characteristic of activated M␾ were evaluated by flow cytometry. The expression level of each marker in the WT mice is defined
as 100%, and the frequency of M␾ activation in the
individual transgenic lines and Csf1op/Csf1op mice are
expressed as percentages relative to the WT mice. Note:
these transgenic CSF-1 strains are on the Csf1op/Csf1op
background. Values are the mean ⫾ SE, ⴱ, p ⬍ 0.05 vs
WT mice; ⴱⴱ, p ⬍ 0.05 vs Csf1op/Csf1op mice. The data
are a combination of three experiments.
Distinct roles of individual CSF-1 isoforms in regulating the
frequency of activated M␾ in the kidney during UUO
We have previously established that activated, but not resting M␾,
induce TEC apoptosis (1) and that CSF-1 mediates intrarenal M␾
activation during UUO (12). To determine the shared and unique
roles of the individual CSF-1 isoforms in M␾ activation during
renal inflammation, we compared the frequency of activated M␾ in
the TgCS/⫹, TgSPP/⫹, TgSGP/⫹, Csf1op/Csf1op, and WT kidneys
during UUO. For this purpose, we evaluated cell surface (e.g.,
CD23, CD69, and Ia) and intracellular (e.g., iNOS, IFN-␥, and
⌻NF-␣) markers of activation and measured the expression of each
marker in the transgenic mice as a percentage of the WT level
(defined as 100%; Fig. 4).
The csCSF-1 isoform is sufficient to restore the frequency of
activated M␾ to WT levels, while the sgCSF-1 isoform is far less
effective. The frequency of activated M␾ in TgCS/⫹ mice was
dramatically higher than in Csf1op/Csf1op mice and indistinguishable from the frequency in TgC/⫹ and WT mice (Fig. 4). Thus, the
csCSF-1 isoform alone is sufficient to restore the frequency of
activated M␾ to WT levels during renal inflammation. By comparison, the frequency of activated M␾ in the TgSGP/⫹ mice is
substantially less than in TgCS/⫹ mice. While the percentage of
M␾ activation markers (e.g., CD23, CD69, Ia, iNOS, and TNF-␣)
in TgSGP/⫹ mice was greater than in Csf1op/Csf1op mice, only the
expression of CD23, and not the other markers, was restored to the
WT level. Furthermore, increasing the circulating concentration of
Table I. Serum CSF-1 levels in CSF-1 isoform transgenic lines
FIGURE 5. CSF-1 isoforms have differing roles in regulating M␾-mediated TEC apoptosis that parallel the accumulation of activated M␾ during UUO. The numbers of apoptotic TEC in the TgC/⫹ and TgCS/⫹ mice
are fully restored to the WT level. The numbers of apoptotic cells in the
TgSPP/⫹ mice are partially restored (⬃57% of WT). However, full restoration occurred in TgSPP-2x/⫹ mice expressing WT (⫹/⫹) levels of
circulating CSF-1 (Table I). In contrast, the numbers of apoptotic cells in
TgSGP-2x/⫹ mice expressing equivalent levels of circulating CSF-1 were
only partially restored (52%). Note: these transgenic strains are on the
Csf1op/Csf1op background. The number of apoptotic TEC were counted in
20 randomly selected high-power field (HPF). Values are the mean ⫾
SEM, ⴱ, p ⬍ 0.001 vs WT mice; ⴱⴱ, p ⬍ 0.005 vs Csf1op/Csf1op. The data
represent a combination of seven experiments.
WT (⫹/⫹)
WT (Csf1op/⫹)
Csf1op/Csf1op
TgCS/⫹
TgSPP/⫹
TgSPP/⫹⫺2x
TgSGP/⫹
TgSGP/⫹⫺2x
Unmanipulated
CSF-1 (ng/ml)a
UUO Strain
CSF-1 (ng/ml)b
17.8 ⫾ 3.65
12.8 ⫾ 3.65
0.4 ⫾ 0.53d
0.3 ⫾ 0.40d,e
5.9 ⫾ 1.46d
15.0 ⫾ 3.81
6.2 ⫾ 2.30d
15.5 ⫾ 2.70
NDc
8.1 ⫾ 2.27
0.6 ⫾ 0.94d
0.1 ⫾ 0.00d,e
ND
10.1 ⫾ 1.14
ND
10.4 ⫾ 1.90
n ⱖ 6/group; values are means ⫾ SD.
n ⫽ 3– 8/group; values are means ⫾ SD.
ND, not done.
d
Value of p ⬍ 0.001 compared with WT (⫹/⫺).
e
csCSF-1 expression was detected on fibroblasts at levels that were intermediate
between the levels on WT (⫹/⫹) and WT (Csf1op/⫹) fibroblasts (1).
a
b
c
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
increasing the concentration of the secreted CSF-1 precursor isoforms to WT levels facilitates M␾ accumulation during renal inflammation, but the presence of the covalently linked ChS chain on
this cytokine is needed to obtained an optimal response.
The Journal of Immunology
Individual CSF-1 isoforms have differing roles in regulating
M␾-mediated TEC apoptosis that parallel the accumulation of
activated M␾ during UUO
Activated M␾ release NO and other mediators that induce TEC
apoptosis (12, 36). We previously established that the numbers of
intrarenal activated M␾ correlate with the extent of TEC apoptosis
and that CSF-1 mediates this sequence of events during renal inflammation (12). To determine the unique and shared roles of individual CSF-1 isoforms in renal inflammation, we evaluated the
impact of these isoforms on M␾-mediated TEC apoptosis. For this
purpose, we counted the number of apoptotic TEC in the TgCS/⫹,
TgSGP/⫹, TgSPP/⫹, Csf1op/Csf1op, and WT mice during UUO.
The csCSF-1 isoform is sufficient to restore M␾-mediated TEC
apoptosis. We detected a substantial increase in apoptotic TEC in
the kidney in TgCS/⫹ as compared with the Csf1op/Csf1op mice. In
fact, the numbers of apoptotic TEC were restored to WT levels
(Fig. 5). Since the number of activated M␾ dictates the extent of
TEC apoptosis, this finding is consistent with the ability of
csCSF-1 to restore M␾ accumulation and activation to WT levels
(Figs. 2 and 4). Taken together, the csCSF-1 isoform is instrumental in mediating M␾-mediated TEC apoptosis.
Increasing the concentration of the secreted CSF-1 isoforms enhances M␾-mediated TEC apoptosis, but the presence of the
ChS chain greatly facilitates this process. To determine whether
the secreted CSF-1 isoforms regulate kidney injury during inflammation, we compared TEC apoptosis in TgSPP/⫹, TgSGP/⫹, WT,
and Csf1op/Csf1op mice during UUO (Fig. 5). We determined that
the numbers of intrarenal apoptotic TEC in TgSGP/⫹ mice did not
rise above the level in the Csf1op/Csf1op mice. In contrast, the
number of apoptotic TEC in TgSPP/⫹ mice increased above the
level in the Csf1op/Csf1op mice, albeit it not to the WT level. Thus,
the covalently linked ChS chain of spCSF-1 facilitates CSF-1 mediated renal injury but does not entirely restore it to the WT level
during inflammation.
To investigate whether increasing these secreted CSF-1 isoforms in the circulation amplifies the number of apoptotic TEC, we
determined the number of apoptotic TEC in the TgSPP-2x/⫹ and
TgSGP-2x/⫹ mice. We detected a substantial increase in the number of apoptotic TEC in the TgSGP-2x/⫹ and TgSPP-2x/⫹ mice as
compared with the TgSGP⫹ and TgSPP/⫹ mice, respectively
(Fig. 5). However, although the number of TEC that were apoptotic in the TgSPP-2x/⫹ mice was fully restored to WT levels, the
number in the TgSGP-2x/⫹ mice only reached ⬃65% of the WT
level (Fig. 5). Thus, restoring the serum spCSF-1 to WT levels, at
approximately twice its concentration in the circulation of WT
mice, fully restores M␾ accumulation, activation, and M␾-mediated TEC apoptosis to WT levels during renal inflammation. In
contrast, similar concentrations of sgCSF-1 do not fully restore
M␾ accumulation, activation, and M␾-mediated TEC apoptosis.
Discussion
We used obstructive nephropathy as a model of renal inflammation
to study the functions of different isoforms of CSF-1. Using a
transgenic complementation approach, we have shown that the
membrane-bound csCSF-1 isoform alone in mice lacking CSF-1 is
sufficient to restore M␾ accumulation, activation, and M␾-mediated TEC apoptosis to the WT levels during renal inflammation
and that it is substantially more effective than either the spCSF-1
or the sgCSF-1 isoforms. However, spCSF-1 is more effective than
sgCSF-1 in stimulating M␾ accumulation, M␾ activation, and
M␾-mediated TEC apoptosis during renal inflammation. As the
transgenes encoding these two secreted forms only differ in point
mutations affecting glycosaminoglycan attachment (19), these observations point to a central role of the CSF-1 ChS proteoglycan.
In addition, we determined that the CSF-1-dependent up-regulation of IFN-␥ in M␾ required cell-cell contact between csCSF-1
and the M␾ CSF-1R. Taken together, there are unique and shared
roles of the individual CSF-1 isoforms that are instrumental in
mediating renal inflammation.
We have pinpointed the expression of CSF-1 in the kidney primarily to the proximal tubules and, to a lesser degree within glomeruli, and localized CSF-1R-expressing M␾ to areas adjacent to
these CSF-1-rich sites. Furthermore, we show that CSF-1 is upregulated in the obstructed kidney as compared with the CL kidney. This finding is consistent with our prior data that CSF-1 is
up-regulated in the kidney with advancing renal injury in the
MRL-Faslpr strain (5, 6). While our findings confirm a prior report
that CSF-1 is increased during UUO, we did not identify CSF-1 in
all tubules in the cortex and tubules in the medulla (37). The difference in apparent CSF-1 expression may be related to the detection techniques used (in situ hybridization, Ab methods, morphology), and/or species (rat) (37) as compared with our studies that
use a mouse transgenic line expressing a reporter gene (␤-gal)
driven by the CSF-1 promoter, combined with lectins known to
specifically bind to differing tubule types, and morphology. Furthermore, we now report that there is an increase in the number of
proximal tubules expressing low-level CSF-1 in the CL kidney as
compared with normal kidneys. This implies that molecules are
released from the inflamed obstructed kidney into the circulation
and that these stimulate the expression of CSF-1 in the CL kidney.
This is in keeping with a series of kidney transplant experiments in
our laboratory establishing that circulating factors in the autoimmune milieu of the MRL-Faslpr mouse induce CSF-1 in the kidney
and are critical for maintaining CSF-1 expression (7, 38). Clearly,
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
sgCSF-1 ⬎ 2-fold (TgSGP-2x/⫹ mice) did not enhance the frequency of activated M␾ in the kidney during UUO (Table I). Thus,
the csCSF-1 isoform is sufficient to restore the frequency of activated M␾ to WT levels, while the sgCSF-1 isoform is far less
effective.
The ChS chain facilitates the restoration and the frequency of
activated M␾ to WT levels. We determined that the spCSF-1 isoform that possesses a ChS chain more effectively restores the frequency of activated M␾ in the kidney as compared with the ChSdeficient sgCSF-1 isoform during UUO (Fig. 4). The frequencies
of CD23, CD69, Ia, iNOS, and TNF-␣ expressing M␾ in TgSPP/⫹
mice were restored to the WT levels. Thus, the ChS attached to
CSF-1 facilitates M␾ activation. Furthermore, the frequency of
activated M␾ was not higher in the TgSPP-2x/⫹ mice possessing
WT (⫹/⫹) levels of circulating CSF-1 (Table I). This indicates
that spCSF at a concentration of approximately one-third the circulating concentration of WT (⫹/⫹) mice (Table I) is able to restore M␾ activation during renal inflammation. By comparison,
the sgCSF-1 was much less efficient (Fig. 4). This finding suggests
that the ChS chain on the proteoglycan CSF-1 facilitates M␾
activation.
Cell-cell contact is required for CSF-1 to increase the frequency
of M␾ expressing IFN-␥ in the kidney. Interestingly, neither the
spCSF-1 nor the sgCSF-1 isoform increased the frequency of IFN␥-expressing M␾ (Fig. 4). The percentage of M␾ generating
IFN-␥ in the TgSGP/⫹ and TgSPP/⫹ mice, or in the TgSGP-2x/⫹
and TgSPP-2x/⫹ mice, did not rise above the level in the Csf1op/
Csf1op strain. This is in sharp contrast to the findings in the
TgCS/⫹ strain in which the frequency of M␾ expressing IFN-␥
was restored to the WT level. These data suggest that cell-cell
contact is required for CSF-1 to increase the frequency of M␾
expressing IFN-␥ in the kidney.
4061
4062
M␾ from the TgSPP/⫹ mice fail to express IFN-␥, these M␾
restore TEC apoptosis, suggesting the IFN-␥ expression by M␾ is
not essential for the induction of apoptosis of TEC.
Several mechanisms could account for the differential functions
seen with different isoforms in renal inflammation. These include
differences in half-life/stability, activity, localization, and cell signaling. For example, the csCSF-1 has a longer half-life (⬃7 h in
cultured cells) (13) and can be expected to have a longer half-life
in vivo than the secreted isoforms (⬃10 min) (15) and therefore
would have greater bioavailability. In addition, csCSF-1 and the
secreted CSF-1 isoforms may differentially signal through the
CSF-1R. For example, the cell surface isoform of stem cell factor,
whose receptor, like CSF-1, is a member of the platelet-derived
growth factor receptor family, mediates more prolonged signaling
and delayed receptor internalization than the secreted SCF isoforms (47). In conclusion, in light of the many diseases mediated
by M␾ in a CSF-1R-dependent manner (48), dissecting the molecular mechanisms responsible for the function of individual
CSF-1 isoforms in renal inflammation offers a strategy to identify
therapeutic approaches for other M␾-mediated diseases.
Acknowledgments
We acknowledge Dr. Richard Stevens, Harvard Medical School, for the
discussions concerning proteoglycan biology, and editorial suggestions,
and Dana Xu, Dr. Xiao-Hua Zong, and Ranu Basu for technical assistance.
Disclosures
The authors have no financial conflict of interest.
References
1. Tesch, G. H., A. Schwarting, K. Kinoshita, H. Y. Lan, B. J. Rollins, and
V. R. Kelley. 1999. Monocyte chemoattractant protein-1 promotes macrophagemediated tubular injury, but not glomerular injury, in nephrotoxic serum nephritis. J. Clin. Invest. 103: 73– 80.
2. Pixley, F. J., and E. R. Stanley. 2004. CSF-1 regulation of the wandering macrophage: complexity in action. Trends Cell. Biol. 14: 628 – 638.
3. Stanley, E. R., K. L. Berg, D. B. Einstein, P. S. Lee, F. J. Pixley, Y. Wang, and
Y. G. Yeung. 1997. Biology and action of colony-stimulating factor-1. Mol. Rep.
Dev. 46: 4 –10.
4. Rubin Kelley, V., R. D. Bloom, M. A. Yui, C. Martin, and D. Price. 1994. Pivotal
role of colony stimulating factor-1 in lupus nephritis. Kidney Int. Suppl. 45:
S83–S85.
5. Yui, M. A., W. H. Brissette, D. C. Brennan, R. P. Wuthrich, and
V. E. Rubin-Kelley. 1991. Increased macrophage colony-stimulating factor in
neonatal and adult autoimmune MRL-lpr mice. Am. J. Pathol. 139: 255–261.
6. Bloom, R. D., S. Florquin, G. G. Singer, D. C. Brennan, and V. R. Kelley. 1993.
Colony stimulating factor-1 in the induction of lupus nephritis. Kidney Int. 43:
1000 –1009.
7. Naito, T., H. Yokoyama, K. J. Moore, G. Dranoff, R. C. Mulligan, and
V. R. Kelley. 1996. Macrophage growth factors introduced into the kidney initiate renal injury. Mol. Med. 2: 297–312.
8. Lenda, D. M., E. R. Stanley, and V. R. Kelley. 2004. Negative role of colonystimulating factor-1 in macrophage, T cell, and B cell mediated autoimmune
disease in MRL-Faslpr mice. J. Immunol. 173: 4744 – 4754.
9. Diamond, J. R. 1995. Macrophages and progressive renal disease in experimental
hydronephrosis. Am. J. Kidney Dis. 26: 133–140.
10. Diamond, J. R., S. D. Ricardo, and S. Klahr. 1998. Mechanisms of interstitial
fibrosis in obstructive nephropathy. Semin. Nephrol. 18: 594 – 602.
11. Klahr, S., and J. Morrissey. 2002. Obstructive nephropathy and renal fibrosis.
Am. J. Physiol. 283: F861–F875.
12. Lenda, D. M., E. Kikawada, E. R. Stanley, and V. R. Kelley. 2003. Reduced
macrophage recruitment, proliferation, and activation in colony-stimulating factor-1-deficient mice results in decreased tubular apoptosis during renal inflammation. J. Immunol. 170: 3254 –3262.
13. Price, L. K., H. U. Choi, L. Rosenberg, and E. R. Stanley. 1992. The predominant
form of secreted colony stimulating factor-1 is a proteoglycan. J. Biol. Chem.
267: 2190 –219.
14. Price, L. 1992. Biosynthetic studies of membrane associated and secreted cell
CSF-1. Doctoral dissertation, Albert Einstein College of Medicine of Yeshiva
University, New York, p. 150.
15. Bartocci, A., D. S. Mastrogiannis, G. Migliorati, R. J. Stockert, A. W. Wolkoff,
and E. R. Stanley. 1987. Macrophages specifically regulate the concentration of
their own growth factor in the circulation. Proc. Natl. Acad. Sci. USA 84:
6179 – 6183.
16. Cecchini, M. G., M. G. Dominguez, S. Mocci, A. Wetterwald, R. Felix,
H. Fleisch, O. Chisholm, W. Hofstetter, J. W. Pollard, and E. R. Stanley. 1994.
Role of colony stimulating factor-1 in the establishment and regulation of tissue
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
it will be intriguing to identify which molecules downstream of
CSF-1 in the inflammatory cascade regulate this M␾ growth
factor.
It is striking that csCSF-1 fully restores M␾ accumulation during renal inflammation. CSF-1 is a chemoattractant (39) and therefore may be responsible for recruiting M␾ into the kidney during
inflammation. It is possible that tubules exclusively expressing
csCSF-1 establish a gradient of cleaved, locally released CSF-1
capable of attracting M␾ into the kidney. It is also possible that
other cytokines/chemokines, released by activated TEC during inflammation, are instrumental in attracting M␾ bearing the CSF-1 R
toward tubular cells expressing CSF-1 on their surface and that
csCSF-1 expressed on TEC in the kidney stabilizes M␾ localization by mediating the adhesion of CSF-1R expressing M␾ to TEC.
In this regard, the membrane isoform of csCSF-1 and the CSF-1R
mediate adhesion between stromal cells (expressing csCSF-1) and
hemopoietic cells (bearing the CSF-1R) (40). The exact interrelationship and roles of csCSF-1, chemokines, and adhesion in renal
inflammation are currently being investigated.
Our studies demonstrate that the ChS chain of spCSF-1 is instrumental in mediating M␾-induced destruction of the kidney.
The covalently linked glycosaminoglycan on the secreted cytokine
appears to be essential for its full CSF-1-mediated accumulation
and activation of M␾ in the kidney during inflammation that leads
to M␾-dependent apoptosis of renal tubules. Among possible explanations for this effect, the ChS chain could cause CSF-1 to bind
to a positively charged protein in the kidney’s extracellular matrix,
thereby concentrating and extending the bioavailability of
spCSF-1 (41). Alternatively, as many chemokines and cytokines
are rapidly degraded by proteases, the ChS chain could prevent the
rapid proteolytic inactivation of CSF-1 by a membrane protease, as
has been shown for the heparan sulfate protection of stromal cellderived factor-1 (27). Third, while spCSF-1 has previously been
shown to have the same affinity for the CSF-1R as sgCSF-1 by
radioreceptor assay (13), experiments with human spCSF-1 suggest that the ChS glycosaminoglycan chain and the C-terminal
sequence containing it have negative effects on CSF-1 stimulated
proliferation but may also increase its stability (42). Thus, apart
from the possible effects of the ChS chain on CSF-1 stability, it is
possible that CSF-1/CSF-1R-dependent signaling in M␾ is modulated via interactions of the ChS chain with either extracellular
matrix or a M␾ cell surface molecule. The exact mechanism responsible for the role of the ChS chain of spCSF-1 on secreted
CSF-1 in fostering M␾-mediated renal injury remains to be
explored.
We previously established that CSF-1 mediates M␾ activation
during renal inflammation (12). We now report the shared and
individual roles of CSF-1 isoforms in mediating M␾ activation
during UUO as defined by cell surface markers (CD23, CD69, and
Ia) and secreted proteins (iNOS, TNF-␣, and IFN-␥). The cell
surface isoform of CSF-1 fully restores M␾ activation as defined
by these criteria. This is in keeping with the impact of this isoform
on M␾ accumulation and M␾-mediated apoptosis of TEC. Unexpectedly, M␾ activation (with the exception of IFN-␥) was fully
restored to WT levels in the spCSF-1 transgenic line expressing
only one-third of the WT (⫹/⫹) level. These data suggest that the
ChS is a potent facilitator of M␾ activation. Interestingly, csCSF-1
alone, and neither of the secreted CSF-1 isoforms, resulted in
IFN-␥ production by M␾ in the obstructed kidney. This indicates
that cell-cell contact is required for CSF-1 to mediate the up-regulation of IFN-␥ in M␾. Since IFN-␥ regulates a myriad of immune responses that promote as well as suppress (43– 46) disease,
the biological consequence of M␾ that are activated and yet do not
generate IFN-␥ remains to be explored. However, even though the
ROLE OF CSF-1 ISOFORMS IN RENAL INFLAMMATION
The Journal of Immunology
17.
18.
19.
20.
21.
22.
23.
25.
26.
27.
28.
29.
30.
31.
32. Murata, F., S. Tsuyama, S. Suzuki, H. Hamada, M. Ozawa, and T. Muramatsu.
1983. Distribution of glycoconjugates in the kidney studied by use of labeled
lectins. J. Histochem. Cytochem. 31: 139 –144.
33. Barresi, G., G. Tuccari, and F. Arena. 1988. Peanut and Lotus tetragonolobus
binding sites in human kidney from congenital nephrotic syndrome of Finnish
type. Histochemistry 89: 117–120.
34. Moore, K. J., T. Naito, C. Martin, and V. R. Kelley. 1996. Enhanced response of
macrophages to CSF-1 in autoimmune mice: a gene transfer strategy. J. Immunol.
157: 433– 440.
35. MacDonald, K. P., V. Rowe, H. M. Bofinger, R. Thomas, T. Sasmono,
D. A. Hume, and G. R. Hill. 2005. The colony-stimulating factor 1 receptor is
expressed on dendritic cells during differentiation and regulates their expansion.
J. Immunol. 175: 1399 –1405.
36. Muhl, H., K. Sandau, B. Brune, V. A. Briner, and J. Pfeilschifter. 1996. Nitric
oxide donors induce apoptosis in glomerular mesangial cells, epithelial cells and
endothelial cells. Eur. J. Pharmacol. 317: 137–149.
37. Isbel, N. M., P. A. Hill, R. Foti, W. Mu, L. A. Hurst, C. Stambe, H. Y. Lan,
R. C. Atkins, and D. J. Nikolic-Paterson. 2001. Tubules are the major site of
M-CSF production in experimental kidney disease: correlation with local macrophage proliferation. Kidney Int. 60: 614 – 625.
38. Wada, T., T. Naito, R. C. Griffiths, T. M. Coffman, and V. R. Kelley. 1997.
Systemic autoimmune nephritogenic components induce CSF-1 and TNF-␣ in
MRL kidneys. Kidney Int. 52: 934 –941.
39. Webb, S. E., J. W. Pollard, and G. E. Jones. 1996. Direct observation and quantification of macrophage chemoattraction to the growth factor CSF-1. J. Cell Sci.
109 (Pt. 4): 793– 803.
40. Zheng, G., Q. Rao, K. Wu, Z. He, and Y. Geng. 2000. Membrane-bound macrophage colony-stimulating factor and its receptor play adhesion molecule-like
roles in leukemic cells. Leuk. Res. 24: 375–383.
41. Suzu, S., T. Ohtsuki, M. Makishima, N. Yanai, T. Kawashima, N. Nagata, and
K. Motoyoshi. 1992. Biological activity of a proteoglycan form of macrophage
colony-stimulating factor and its binding to type V collagen. J. Biol. Chem. 267:
16812–16815.
42. Partenheimer, A., K. Schwarz, C. Wrocklage, E. Kolsch, and H. Kresse. 1995.
Proteoglycan form of colony-stimulating factor-1 (proteoglycan-100). Stimulation of activity by glycosaminoglycan removal and proteolytic processing. J. Immunol. 155: 5557–5565.
43. Diaz-Gallo, C., and V. R. Kelley. 1993. Self-regulation of autoreactive kidneyinfiltrating T cells in MRL-lpr nephritis. Kidney Int. 44: 692– 699.
44. Schwarting, A., K. Moore, T. Wada, G. Tesch, H. J. Yoon, and V. R. Kelley.
1998. IFN-␥ limits macrophage expansion in MRL-Faslpr autoimmune interstitial
nephritis: a negative regulatory pathway. J. Immunol. 160: 4074 – 4081.
45. Teixeira, L. K., B. P. Fonseca, B. A. Barboza, and J. P. Viola. 2005. The role of
interferon ␥ on immune and allergic responses. Mem. Inst. Oswaldo Cruz
100(Suppl. 1): 137–144.
46. Pestka, S., C. D. Krause, and M. R. Walter. 2004. Interferons, interferon-like
cytokines, and their receptors. Immunol. Rev. 202: 8 –32.
47. Miyazawa, K., D. A. Williams, A. Gotoh, J. Nishimaki, H. E. Broxmeyer, and
K. Toyama. 1995. Membrane-bound Steel factor induces more persistent tyrosine
kinase activation and longer life span of c-kit gene-encoded protein than its soluble form. Blood 85: 641– 649.
48. Chitu, V., and E. R. Stanley. 2006. Colony-stimulating factor-1 in immunity and
inflammation. Curr. Opin. Immunol. 18: 39 – 48.
Downloaded from http://www.jimmunol.org/ by guest on June 18, 2017
24.
macrophages during postnatal development of the mouse. Development 120:
1357–1372.
Dai, X. M., X. H. Zong, V. Sylvestre, and E. R. Stanley. 2004. Incomplete
restoration of colony-stimulating factor 1 (CSF-1) function in CSF-1-deficient
Csf1op/Csf1op mice by transgenic expression of cell surface CSF-1. Blood 103:
1114 –1123.
Stanley, E. R. 1979. Colony-stimulating factor (CSF) radioimmunoassay: detection of a CSF subclass stimulating macrophage production. Proc. Natl. Acad. Sci.
USA 76: 2969 –2973.
Nandi, S., M. P. Akhter, M. F. Seifert, X. M. Dai, and E. R. Stanley. 2006.
Developmental and functional significance of the CSF-1 proteoglycan chondroitin sulfate chain. Blood 107: 786 –795.
Ryan, G. R., X. M. Dai, M. G. Dominguez, W. Tong, F. Chuan, O. Chisholm,
R. G. Russell, J. W. Pollard, and E. R. Stanley. 2001. Rescue of the colonystimulating factor 1 (CSF-1)-nullizygous mouse (Csf1op/Csf1op) phenotype with
a CSF-1 transgene and identification of sites of local CSF-1 synthesis. Blood 98:
74 – 84.
Aviezer, D., D. Hecht, M. Safran, M. Eisinger, G. David, and A. Yayon. 1994.
Perlecan, basal lamina proteoglycan, promotes basic fibroblast growth factorreceptor binding, mitogenesis, and angiogenesis. Cell 79: 1005–1013.
Yayon, A., M. Klagsbrun, J. D. Esko, P. Leder, and D. M. Ornitz. 1991. Cell
surface, heparin-like molecules are required for binding of basic fibroblast growth
factor to its high affinity receptor. Cell 64: 841– 848.
Catlow, K., J. A. Deakin, M. Delehedde, D. G. Fernig, J. T. Gallagher,
M. S. Pavao, and M. Lyon. 2003. Hepatocyte growth factor/scatter factor and its
interaction with heparan sulphate and dermatan sulphate. Biochem. Soc. Trans.
31: 352–353.
Mikhailov, D., H. C. Young, R. J. Linhardt, and K. H. Mayo. 1999. Heparin
dodecasaccharide binding to platelet factor-4 and growth-related protein ␣: induction of a partially folded state and implications for heparin-induced thrombocytopenia. J. Biol. Chem. 274: 25317–25329.
Proudfoot, A. E., T. M. Handel, Z. Johnson, E. K. Lau, P. LiWang,
I. Clark-Lewis, F. Borlat, T. N. Wells, and M. H. Kosco-Vilbois. 2003. Glycosaminoglycan binding and oligomerization are essential for the in vivo activity of
certain chemokines. Proc. Natl. Acad. Sci. USA 100: 1885–1890.
Spillmann, D., D. Witt, and U. Lindahl. 1998. Defining the interleukin-8-binding
domain of heparan sulfate. J. Biol. Chem. 273: 15487–15493.
Sadir, R., A. Imberty, F. Baleux, and H. Lortat-Jacob. 2004. Heparan sulfate/
heparin oligosaccharides protect stromal cell-derived factor-1 (SDF-1)/CXCL12
against proteolysis induced by CD26/dipeptidyl peptidase IV. J. Biol. Chem. 279:
43854 – 43860.
Kuschert, G. S., F. Coulin, C. A. Power, A. E. Proudfoot, R. E. Hubbard,
A. J. Hoogewerf, and T. N. Wells. 1999. Glycosaminoglycans interact selectively
with chemokines and modulate receptor binding and cellular responses. Biochemistry 38: 12959 –12968.
Sasmono, R. T., D. Oceandy, J. W. Pollard, W. Tong, P. Pavli, B. J. Wainwright,
M. C. Ostrowski, S. R. Himes, and D. A. Hume. 2003. A macrophage colonystimulating factor receptor-green fluorescent protein transgene is expressed
throughout the mononuclear phagocyte system of the mouse. Blood 101:
1155–1163.
Kawada, N., T. Moriyama, A. Ando, M. Fukunaga, T. Miyata, K. Kurokawa,
E. Imai, and M. Hori. 1999. Increased oxidative stress in mouse kidneys with
unilateral ureteral obstruction. Kidney Int. 56: 1004 –1013.
Stanley, E. R. 1985. The macrophage colony-stimulating factor, CSF-1. Methods
Enzymol. 116: 564 –587.
4063