Abstract
The increasing global prevalence of diabetic nephropathy poses
substantial health and economic burdens. Currently, effective
anti-fibrotic therapies for managing kidney fibrosis associated with
chronic kidney disease are lacking. This study reveals corisin, a
microbiota-derived peptide, as a central driver in the progression of
diabetic kidney fibrosis. Corisin levels were found to be markedly
elevated in the serum of diabetic chronic kidney disease patients
relative to healthy controls, with strong correlations to advanced
disease stages and declining renal function. In a murine model of
kidney fibrosis, corisin levels were similarly heightened, directly
contributing to increased inflammation and worsening fibrosis and renal
impairment. Notably, the use of a monoclonal anti-corisin antibody
significantly reduced nephropathy severity in diabetic mice. Through
molecular dynamics simulations and experimental validation, we
demonstrated that corisin interacts with human serum albumin,
potentially enhancing its renal accumulation and pathological impact.
The pathogenic mechanism of corisin involves the acceleration of
cellular senescence and the induction of epithelial-mesenchymal
transition and apoptosis in kidney cells. These findings underscore the
critical role of corisin in progressive diabetic nephropathy and
suggest a promising new target for therapeutic intervention.
Subject terms: End-stage renal disease, Renal fibrosis, Microbiome
__________________________________________________________________
Here, the authors identify the microbiota-derived corisin as a driver
of diabetic kidney fibrosis via cellular aging and show that targeting
corisin with a monoclonal antibody alleviates disease in mice,
suggesting a potential therapeutic avenue.
Introduction
The rising prevalence of diabetes mellitus (DM) has emerged as a
critical global health challenge, imposing a substantial economic
burden on societies worldwide^[82]1–[83]3. Recent epidemiological data
underscore the alarming magnitude of this issue, revealing that 536.6
million individuals, constituting an estimated prevalence of 10.5%,
were affected by DM worldwide in 2021^[84]4. The prevalence is expected
to rise to 12.2%, affecting 783.2 million individuals, by 2045^[85]4.
The seriousness of the situation is compounded by the association of DM
with elevated morbidity and mortality rates. According to the World
Health Organization (WHO), there was a 3% increase in mortality rates
due to DM across different age groups between 2000 and 2019, resulting
in approximately 1.5 million deaths globally in 2019 directly
attributed to DM^[86]3. The primary culprits contributing to the high
morbidity and mortality rates associated with DM are microangiopathy
and macroangiopathy^[87]5–[88]10. Notably, diabetic nephropathy stands
out as one of the most prevalent complications, assuming the role of
the leading cause of chronic kidney disease and end-stage kidney
disease^[89]6,[90]9,[91]11,[92]12. This expanding diabetic population
and the associated complications, particularly diabetic nephropathy,
underscore the pressing need for comprehensive strategies to manage and
mitigate the impact of diabetes on global health.
The unique functional and histopathological features of diabetic
nephropathy include proteinuria, enhanced cell proliferation, elevated
matrix deposition in the mesangium and renal interstitium, tubular
atrophy, and thickening of the glomerular basement membrane, resulting
in glomerulosclerosis and tubulointerstitial fibrosis, collectively
contributing to a reduced glomerular filtration rate^[93]13,[94]14.
These fibrotic changes in the kidneys can lead to irreversible damage,
ultimately culminating in end-stage kidney disease. Currently, there is
no definitive cure for DM-associated kidney fibrosis. Treatment options
primarily revolve around controlling hyperglycemia, proteinuria, and
arterial hypertension, decreasing cardiovascular risk, alleviating
symptoms, and implementing renal replacement therapies in cases of
severely impaired kidney function. The primary effects of dysglycemia
on the kidneys include injury, activation, and/or apoptosis of
glomerular endothelial cells, podocytes, and tubular epithelial cells,
as well as overactivation of the intrarenal
renin-angiotensin-aldosterone system^[95]13,[96]15. Recent studies
suggest that alterations in the microbiome or dysbiosis may play a role
in the progression of diabetic nephropathy^[97]16–[98]18. Dysbiosis has
been linked to excessive production and accumulation of bacterial
metabolites, including urease, p-cresol, and acetate, potentially
contributing to oxidative stress, kidney cell apoptosis, inflammation,
and profibrotic activity^[99]19–[100]22. Notably, recent studies
demonstrated the presence of gut-derived bacteria within the kidneys of
spontaneously hypertensive rats, and patients with arterial
hypertension^[101]23. Dysbiosis has also been implicated in the sudden
exacerbation of renal failure or acute kidney injury^[102]24–[103]26.
Additionally, our recent research identified corisin, a
microbiota-derived peptide, as a potential contributor to podocyte and
renal tubular cell apoptosis, further implicating the microbiome in the
pathogenesis of kidney fibrosis^[104]27,[105]28.
Building upon this foundation, we hypothesized that corisin contributes
to the progression of kidney fibrosis, a key pathological feature of
diabetic CKD. Overall, our study demonstrates that elevated circulating
corisin levels correlate with disease severity and renal dysfunction in
DM patients and with increased kidney fibrosis in murine models.
Systemic administration of corisin was associated with kidney fibrosis,
while its inhibition attenuates the progression of diabetic nephropathy
in experimental animal models. Mechanistically, corisin interacts with
human serum albumin, facilitating its transport to the kidneys, where
it accelerates cellular senescence and induces apoptosis or
epithelial-mesenchymal transition, promoting fibrogenesis. These
findings underscore the potential of targeting the corisin pathway as a
promising strategy to mitigate the progression of diabetic nephropathy.
Results
High circulating levels of corisin in patients with diabetic CKD
We initially posited that the circulating concentration of
microbiota-derived corisin is dysregulated in patients with diabetic
CKD. To test this hypothesis, we quantified corisin levels in serum
from this patient population (Supplementary Table [106]1). Our analysis
revealed a marked elevation in serum corisin levels in patients with
diabetic CKD compared to healthy controls. Furthermore, serum corisin
levels were also significantly higher in patients with non-diabetic CKD
relative to healthy individuals. However, no significant difference was
observed between the diabetic and non-diabetic CKD groups. Notably,
serum corisin concentrations were significantly elevated in diabetic
CKD patients at stages G2 to G4 compared to those at stage G1
(Fig. [107]1A, B). Analysis of 24-hour urine collections demonstrated
no significant differences between the G1 and G2-G4 subgroups in either
the daily urinary excretion of corisin or albumin (Supplementary
Fig. [108]1).
Fig. 1. Increased serum levels of corisin in diabetic chronic kidney disease
patients.
[109]Fig. 1
[110]Open in a new tab
A Thirty-five patients were enrolled in the study. Thirteen patients
were excluded due to ongoing anticancer therapy, incomplete datasets,
or suboptimal sample quality. B Data from 20 healthy subjects served as
controls. Diabetic chronic kidney disease patients were allocated into
two groups: one with stage G1 (n = 16) and another with stages G2, G3,
and G4 (n = 19). Five patients with non-diabetic chronic kidney disease
were also included in the study. Data are expressed as mean ± SD.
Statistical significance was determined using one-way ANOVA followed by
the Newman–Keuls post hoc test for comparisons among three groups. C
Human corisin and corisin-like peptide sequences identified in the
urine of patients with diabetic chronic kidney disease (DM-CKD) and
healthy controls (HC). Staphylococcus species are the primary source of
corisin in diabetic patients with chronic kidney disease. Sequences
were aligned using DECIPHER, including two reference sequences (S.
nepalensis: GenBank [111]CP120099.1; S. haemolyticus: GenBank
[112]CP071512.1). Sequence IDs represent the closest species-level
match, accompanied by a unique identifier corresponding to the original
sequence. D A heatmap displaying read counts from sequence IDs,
representing unique peptide sequences from the original amplicons, was
used to normalize relative abundance across patients. The data are
plotted on a log^[113]10 scale using phyloseq. The source data are
available in the Source Data file.
An inverse correlation was observed between serum corisin levels and
estimated glomerular filtration rates (eGFR), while a direct
correlation was noted with systolic blood pressure and serum creatinine
and total cholesterol levels (Supplementary Fig. [114]2). We also
conducted univariate and multivariate analyses, using eGFR as the
dependent variable. Variables with a p-value of less than 0.05 in their
univariate relationship with eGFR were included in the multivariate
analysis. Serum corisin levels were independently and inversely
correlated with eGFR, consistent with reduced renal clearance
(Supplementary Table [115]2). In addition, baseline serum corisin
levels exhibited an increasing trend in patients with diabetic
nephropathy who experienced a greater decline in eGFR over a two-year
follow-up period (>4 mL/min/1.73 m²) compared to those with a lesser
decline (<4 mL/min/1.73 m²), suggesting a potential role of corisin in
the progression of kidney function deterioration in diabetic CKD
patients (Supplementary Fig. [116]3).
Staphylococcus species are the primary source of corisin in diabetic patients
with CKD
We have previously isolated corisin and corisin-like peptides from the
lung tissue of mice with lung fibrosis, identifying commensal or
pathogenic bacteria from the Staphylococcus species as their main
source^[117]27,[118]29. After confirming the presence of corisin
peptides in urine through immune assays, we aimed to determine the
source of these peptides in patients with diabetic CKD. We prepared DNA
from the urine of these patients and amplified DNA fragments encoding
corisin/corisin-like peptides. To develop PCR primers that specifically
amplify the DNA encoding the proapoptotic peptides in each urine
sample, we created a large alignment of the polypeptide sequences of
the IsaA-like transglycosylases from diverse bacteria. We identified
two highly conserved sequences flanking the proapoptotic peptides
(SVKAQF and WGTGSV), which facilitated the design of two primers,
Corisin-F (5’ATCAGTTAAAGCTCAATTC) and Corisin-R
(5’GCTACTGAACCAGTACCCCATG). This primer pair successfully amplified
~150 bp DNA fragments, the predicted size of the coding sequences for
the proapoptotic peptides. Subsequent DNA sequencing and bioinformatics
analysis of the corisin/corisin-like peptides in the urine samples
revealed that they match those present in diverse Staphylococcus
species with genomes reported in the Genbank database (Fig. [119]1C, D,
Supplementary Dataset [120]1). These results confirm the presence of
corisin and corisin-like peptides in human fluids and establish
Staphylococcus species as their primary source in diabetic CKD
patients.
High circulating and urine levels of corisin in mice with diabetic CKD
As kidney fibrosis is a hallmark of CKD, we sought to validate the
abnormal corisin levels observed in patients with diabetic CKD by
comparing plasma and urinary corisin levels between normal wild-type
(WT) mice and kidney-specific transforming growth factor-β1 (TGFβ1)
transgenic (TG) mice with kidney fibrosis. Plasma and urinary corisin
levels were significantly elevated in TGFβ1 TG mice with kidney
fibrosis compared to WT controls. In addition, plasma and urinary
levels of TGFβ1 were significantly elevated in TGFβ1 transgenic (TG)
mice compared to WT controls (Supplementary Fig. [121]4A–C). However,
no significant correlations were observed between corisin and TGFβ1
levels in either plasma (r = −0.09, p = 0.65) or urine (−0.18,
p = 0.37), suggesting that TGFβ1 may not directly regulate corisin
expression.
To determine whether DM further exacerbates the abnormal levels of
corisin, we measured and compared the levels of corisin in plasma and
urine samples collected from diabetic and non-diabetic TGFβ1 TG mice
with kidney fibrosis. We observed a significant increase in both plasma
and urine levels of corisin in diabetic TGFβ1 TG mice with kidney
fibrosis compared to their non-diabetic counterparts (Supplementary
Fig. [122]4D–F). To investigate whether similar alterations in
circulating corisin levels occur in a different experimental model of
DM, unilaterally nephrectomized WT mice, with or without diabetic
nephropathy, were prepared, and plasma corisin levels were measured. In
this model, unilateral nephrectomy accelerates the progression of
diabetic nephropathy. Consistent with findings from TGFβ1 transgenic
mice with spontaneous kidney fibrosis and TGFβ1 TG mice with diabetic
nephropathy, plasma corisin levels were significantly elevated in WT
mice with diabetic kidney fibrosis compared to non-diabetic WT controls
(Supplementary Fig. [123]4G–I). Overall, these observations suggest a
potential association between corisin and kidney fibrosis, particularly
in diabetic conditions.
Systemic administration of corisin is linked to increased kidney fibrosis
Given our observations of elevated corisin levels in
nephropathy-related conditions and the documented role of corisin in
organ fibrosis exacerbation, we posited that corisin might contribute
to nephropathy by accelerating the progression of kidney
fibrosis^[124]27,[125]30. To explore this hypothesis, we systemically
administered synthetic corisin to TGFβ1 transgenic mice with kidney
fibrosis, a model we have previously characterized^[126]31. Synthetic
corisin or control scrambled peptide was administered intraperitoneally
to the TGFβ1 TG mice thrice weekly for two weeks (Fig. [127]2A). The
results revealed that TGFβ1 TG mice treated with corisin exhibited a
significant increase in urinary levels of corisin, albumin, and
albumin-to-creatinine ratio. The plasma levels of blood urea nitrogen
(BUN) and the urinary levels of liver-type fatty acid-binding protein
(L-FABP) were also significantly elevated in TGFβ1 TG mice receiving
intraperitoneal corisin compared to their untreated counterparts
(Fig. [128]2B, C). As expected, plasma corisin levels were
significantly elevated in TGFβ1 TG mice receiving intraperitoneal
synthetic corisin compared to their counterparts receiving the
scrambled peptide. Quantitatively, the mean circulating corisin levels
in corisin-treated mice were 2.2-fold higher than those observed in
diabetic CKD patients and 1.6-fold higher than those in diabetic TGFβ1
transgenic mice. The circulating levels of lipopolysaccharide-induced
CXC chemokine (LIX/CXCL5), macrophage inflammatory protein-2 (MIP-2),
interleukin-1β (IL-1β), platelet-derived growth factor (PDGF), tissue
factor (TF), and plasminogen activator inhibitor-1 (PAI-1) were also
increased in the corisin-treated mice compared to those receiving the
scrambled peptide (Fig. [129]2D). Furthermore, renal dysfunction was
associated with significantly increased fibrotic changes in the
corisin-treated mice compared to those receiving the scrambled peptide
(Fig. [130]2E–I). These findings suggest that corisin may contribute to
the progression of chronic nephropathy by promoting inflammation and
fibrosis in the kidney.
Fig. 2. The administration of corisin induces acute kidney injury and
exacerbation of kidney fibrosis in TGFβ1 transgenic mice with pre-existing
renal injury.
[131]Fig. 2
[132]Open in a new tab
A Experimental plan for inducing acute kidney injury in TGFβ1
transgenic (TG) mice with pre-existing renal dysfunction. One group of
transforming growth factor β1 (TGFβ1) TG mice (TGFβ1 TG/corisin; n = 6)
received 5 mg/kg of body weight of synthetic corisin by intraperitoneal
injection every two days for two weeks, and another group (TGFβ1
TG/scr. peptide; n = 6) received a similar dose of scrambled peptide
following the same schedule and route of administration. B Urinary
corisin, albumin, and creatinine were measured as described under
materials and methods. Number of mice: TGFβ1 TG/corisin, n = 6; TGFβ1
TG/scr. peptide, n = 6. Data are presented as mean ± SEM. Statistical
significance was assessed using ANOVA with the Newman-Keuls test for
longitudinal data and an unpaired one-sided t-test for comparisons
between two groups. *p < 0.05 vs week 0; †p = 0.02 and ‡p = 0.01 vs
TGFβ1 TG/scrambled peptide group. C The urinary levels of kidney injury
molecule-1 (KIM-1), liver-type fatty acid-binding protein (L-FABP),
blood urea nitrogen (BUN), and creatinine were measured as described
under material and methods. Number of mice: TGFβ1 TG/corisin, n = 6;
TGFβ1 TG/scr. peptide, n = 6. Data are presented as mean ± SD.
Statistical significance was determined using a two-sided unpaired
t-test. D Plasma levels of lipopolysaccharide-induced CXC chemokine
(LIX/CXCL5), macrophage inflammatory protein-2 (MIP-2), interleukin-1β
(IL-1β), platelet-derived growth factor (PDGF), tissue factor (TF), and
plasminogen activator inhibitor-1 (PAI-1) were measured using enzyme
immunoassays. Normally distributed data are presented as mean ± SD,
while skewed data are expressed as the median with interquartile range.
Number of mice: TGFβ1 TG/corisin, n = 6; TGFβ1 TG/scr. peptide, n = 6.
Statistical significance was assessed using a two-sided unpaired t-test
for normally distributed data and the two-sided Mann-Whitney U test for
skewed data. E–I Masson’s trichrome and periodic acid–Schiff staining
of renal tissues from both groups of mice. Renal fibrosis was
quantified using WinROOF imaging software. Scale bars indicate 100 µm
in (D) and 50 µm in (F). Number of mice: TGFβ1 TG/corisin, n = 6; TGFβ1
TG/scr. peptide, n = 6. Data are presented as mean ± SD. Statistical
significance was determined using a two-sided unpaired t-test. The
source data are available in the Source Data file.
Corisin inhibition mitigates the progression of nephropathy under diabetic
conditions
Considering the potential involvement of corisin in the progression of
kidney fibrosis and the well-established association of DM as a risk
factor for progressive nephropathy and end-stage kidney disease, we
hypothesized that inhibiting corisin may alleviate kidney fibrosis
under diabetic conditions^[133]32. To test this hypothesis, we
evaluated whether administering a neutralizing anticorisin monoclonal
antibody (mAb) could mitigate the progression of nephropathy in TGFβ1
TG mice with DM. Through streptozotocin injection, DM was induced in
TGFβ1 TG mice with pre-existing renal dysfunction (Fig. [134]3A). The
diagnosis of DM in the mouse models was confirmed by blood glucose
levels and intraperitoneal glucose tolerance test results
(Fig. [135]3B). Upon confirmation of DM, one group of mice received the
anticorisin mAb, while another group received an IgG control,
administered intraperitoneally three times a week for eight weeks. The
mice were sacrificed at the end of the ninth week after streptozotocin
administration. Throughout the treatment period, urine and plasma
samples were collected sequentially from each experimental group. It is
noteworthy that, in a preliminary experiment, the degree of kidney
fibrosis was compared between DM TG mice treated with intraperitoneal
injections of control IgG and saline, revealing no significant
difference and thereby supporting the use of control IgG as an
appropriate negative control in this experiment (Supplementary
Fig. [136]5).
Fig. 3. Treatment with the anticorisin antibody inhibits the progression of
kidney fibrosis under diabetic conditions.
[137]Fig. 3
[138]Open in a new tab
A Experimental plan. Diabetic TGFβ1 TG mice were divided into the
following groups. One group (DM TG/anticorisin) received 10 mg/kg of
body weight of anticorisin mAb by intraperitoneal injection three times
a week for eight weeks, while another group (DM TG/control IgG)
received a similar dose of IgG control. Following the same sampling
schedule, a wild-type (WT) (n = 4) group was also included. B Blood
glucose levels throughout the experiment, intraperitoneal glucose
tolerance test (IPGTT), plasma creatinine, blood urea nitrogen (BUN),
urine albumin/creatinine ratio, and urinary liver-type fatty
acid-binding protein (L-FABP) levels. n = 5 for DM TG/control IgG and
DM TG/anti-corisin; n = 4 for WT/SAL. Normally distributed data are
presented as the mean, while skewed data are presented as the median.
Statistical significance was assessed using ANOVA followed by either
the Newman-Keuls or Dunn’s post hoc test, as appropriate; all tests
were two-sided. †p < 0.05 vs WT/SAL across the same week. C, D
Paraffin-embedded kidney tissue samples for collagen staining with
trichrome acid. n = 5 in DM TG/control IgG and DM TG/anticorisin
groups, and n = 4 in the WT/SAL group. Renal fibrosis was quantified
using WinROOF. Scale bars represent 500 µm. Data are presented as
mean ± SD. Statistical significance was assessed using ANOVA and the
Newman-Keuls test; all tests were two-sided. E, F Paraffin-embedded
kidney tissue samples for Periodic acid-Schiff (PAS) staining. Scale
bars represent 50 µm. n = 5 in DM TG/control IgG and DM TG/anticorisin
groups, and n = 4 in the WT/SAL group. Data are presented as mean ± SD.
Statistical significance was assessed using ANOVA and the Newman-Keuls
test; all tests were two-sided. G, H The plasma levels of
lipopolysaccharide-induced CXC chemokine/C-X-C motif chemokine 5
(LIX/CXCL5), interleukin-1β (IL-1β), matrix metalloproteinase-2
(MMP-2), platelet-derived growth factor (PDGF), total and active
transforming growth factor β1 (TGFβ1), connective tissue growth factor
(CTGF), and collagen I were measured by enzyme-linked immune assays.
n = 11 in DM TG/control IgG, n = 10 in DM TG/anticorisin groups, and
n = 4 in the WT/SAL group. Normally distributed data are presented as
mean ± SD, while skewed data are expressed as the median with
interquartile range. Statistical significance was assessed using ANOVA
followed by the Newman-Keuls or Dunn’s post hoc test, as appropriate;
all tests were two-sided. The source data are available in the Source
Data file.
By the 12th week of the study, TG mice treated with the anti-corisin
neutralizing mAb exhibited a significant reduction in plasma creatinine
and blood urea nitrogen (BUN) levels, along with a decreased urinary
albumin-to-creatinine ratio. Urinary levels of L-type fatty
acid-binding protein, a marker of tubular cell injury, were also
significantly improved compared to control mice (Fig. [139]3B).
Treatment with the anti-corisin neutralizing mAb markedly reduced
extracellular matrix deposition, as well as circulating levels of
chemokines (LIX/CXCL5), growth factors (total TGFβ1, active TGFβ1,
PDGF, CTGF), and collagen I (Fig. [140]3C–H). Furthermore, diabetic
mice treated with the anti-corisin mAb showed significantly decreased
renal deposition of corisin (Supplementary Fig. [141]6A, B;
Supplementary Fig. [142]7A, B; Supplementary Fig. [143]8A, B),
accompanied by reduced apoptosis in both glomerular and tubular
epithelial cells (Supplementary Fig. [144]9A, B). Consistent with
previous reports identifying macrophages as mediators of tissue injury,
fibrosis, and functional decline in diabetic chronic kidney disease
(CKD)^[145]33,[146]34, we observed increased macrophage infiltration in
diabetic mice with kidney fibrosis, which was significantly attenuated
in mice treated with the anti-corisin mAb (Supplementary Fig. [147]9C,
D). Overall, these findings suggest that corisin contributes to renal
tissue injury, inflammation, and functional impairment associated with
kidney fibrosis under diabetic conditions.
Enrichment of corisin-expressing Staphylococcus in fecal samples from
diabetic TGF-β1 transgenic mice and elevated circulating levels of I-FABP2, a
marker of intestinal epithelial injury, in patients with diabetic CKD
Having established the increased circulating levels of corisin and its
role in kidney fibrosis, we next investigated whether its potential
source resides within the gut microbiome. To this end, we employed the
previously described PCR primers, which selectively amplify a ~150 bp
DNA fragment encoding proapoptotic peptides, to analyze fecal samples
from WT mice and TGFβ1 TG mice with STZ-induced DM and kidney fibrosis.
Our analysis revealed significant enrichment in Staphylococcus species
expressing corisin in fecal samples from diabetic TGFβ1 TG mice
compared to control mice (Supplementary Fig. [148]10A, B). Gut
microbiota dysbiosis can increase intestinal permeability, promoting
the translocation of microbial products into the systemic
circulation^[149]35. To evaluate intestinal epithelial injury, we
measured circulating levels of intestinal fatty acid-binding protein 2
(I-FABP2) and found them to be significantly elevated in diabetic
patients with CKD compared to non-diabetic CKD and healthy controls
(Supplementary Fig. [150]10C). These findings collectively suggest that
under diabetic conditions, increased intestinal permeability and
microbial dysbiosis enhance the systemic release of corisin, indicating
that enhanced production contributes to elevated circulating levels.
Molecular dynamic simulation predicts corisin interaction with human serum
albumin
Following the demonstration that corisin may be implicated in the
development of diabetic nephropathy, we explored how microbiome-derived
corisin targets the kidneys. Serum albumin is known for its role as a
carrier for various endogenous substances, including those from the
microbiome, such as short-chain fatty acids and indole
derivatives^[151]36. We hypothesized that corisin interacts with serum
albumin to reach the kidneys and contribute to nephropathy
pathogenesis. To interrogate this hypothesis, we conducted molecular
dynamics simulations to analyze the interactions between corisin and
human serum albumin and bovine serum albumin. Human serum albumin is
composed of a single polypeptide chain containing 585 amino acid
residues, organized into three α-helical domains labeled I, II, and
III^[152]36–[153]38. The amino acid sequence of human albumin exhibits
a high degree of homology with bovine albumin, sharing 77.5%
identity^[154]39. We employed AlphaFold-Multimer^[155]40 to model the
configurations of corisin bound to human serum albumin (HSA) and bovine
serum albumin (BSA). The predictive analysis suggests that corisin
interacts with domains I and III of HSA, whereas it predominantly
associates with domain II of BSA (Fig. [156]4A, B). A 100 ns molecular
dynamics simulation was conducted on both HSA-corisin and BSA-corisin
complexes using OpenMM^[157]41. Subsequently, the interaction energies
of these complexes were quantified using the Linear Interaction Energy
(LIE) method in CPPTRAJ^[158]42. The results reveal lower interaction
energy for the HSA-corisin complex, with a significant difference of
approximately -22.3353 kcal/mol. These findings suggest that corisin
may form a stable complex with human serum albumin, indicating that,
like other microbiome-derived metabolites, albumin could serve as a
carrier for corisin to reach distant organs, including the kidneys.
Fig. 4. Corisin interacts with serum human albumin.
[159]Fig. 4
[160]Open in a new tab
A, B The predicted interaction structure of human serum albumin-corisin
and bovine serum albumin-corisin by molecular dynamic simulations.
Human serum albumin and bovine serum albumin are shown in surface
representation, and domains I, II, and III are shown in pink, yellow,
and ice blue, respectively. The corisin is shown in licorice
representation. C, D Western blotting showing the interaction of
corisin with recombinant human albumin (rhAlb) in vitro. A fixed
concentration of rhAlb was combined with concentrations of synthetic
corisin or scrambled peptide and incubated at room temperature for
10 min. The mixture underwent sodium dodecyl-sulfate polyacrylamide gel
electrophoresis (SDS-PAGE), followed by Western blotting using
anti-corisin monoclonal antibody (mAb) or anti-rhAlb antibody. The
experiment was independently repeated three times with similar results.
E Co-immunoprecipitation of corisin with rhAlb. Synthetic corisin was
incubated with rhAlb, treated with an anti-corisin mAb, and
immunoprecipitated using protein G agarose beads. rhAlb was not
immunoprecipitated by the anti-corisin mAb and protein G agarose beads
alone (lane 5). In the presence of corisin, rhAlb was successfully
co-immunoprecipitated by the anti-corisin mAb and protein G agarose
beads (lane 6). The experiment was independently repeated three times
with similar results. Ig, immunoglobulin; MW, molecular weight. F
Corisin interacts with human and mouse serum albumin. A predetermined
concentration of human serum albumin (HSA) or mouse serum albumin (MSA)
was incubated with synthetic corisin. The mixture was subjected to
SDS-PAGE and Western blotting using an anti-corisin monoclonal
antibody. The experiment was independently repeated three times with
similar results. G Corisin interacts with human serum albumin derived
from both patients with diabetic nephropathy and healthy individuals.
Serum samples were diluted (1:20) and incubated for 10 min with corisin
(4 µg) diluted in saline. As a control, a mixture containing corisin
(4 µg) and rhAlb (5 µg) was prepared. Each mixture was then subjected
to electrophoresis, followed by Western blot analysis using mAb
specific for corisin or human albumin. The experiment was independently
repeated three times with similar results. H, I Complex formation
between corisin and albumin in the urine of diabetic CKD patients.
Urine samples from 3 diabetic patients with chronic kidney disease
(CKD) and 3 healthy subjects were concentrated 5 times, and 5 µL of
each sample were subjected to SDS-PAGE, followed by Western blotting
using an anti-corisin mAb (left panel) or anti-human albumin antibody
(right panel). DM diabetes mellitus.
Experimental validation of corisin-human serum albumin interaction
We conducted Western blotting to validate the computational analysis
suggesting the interaction of corisin with human serum albumin. Various
concentrations of synthetic corisin or its scrambled peptide were
incubated with a defined concentration of recombinant human serum
albumin (rhAlb) at room temperature for 10 min. The mixtures were then
subjected to electrophoresis and Western blotting using an anticorisin
mAb. Samples containing corisin and rhAlb displayed multiple bands at
approximately 35, 46, 50, and 69 kDa, with band intensity increasing
proportionally with corisin concentration, indicating co-localization
and interaction of corisin with human albumin (Fig. [161]4C). In
contrast, no bands were observed in samples containing the scrambled
peptide. Membrane stripping followed by treatment with an anti-human
albumin antibody showed uniformly intense bands across all lanes
(Fig. [162]4D).
To confirm the co-localization of synthetic corisin with recombinant
human serum albumin during electrophoresis, we conducted mass
spectrometry. A specific quantity of synthetic corisin was dissolved in
either 0.5% rhAlb or physiological saline and incubated at room
temperature for 10 min. The samples were then subjected to
electrophoresis followed by Coomassie blue staining (Supplementary
Fig. [163]11). The stained bands corresponding to 35 kDa were excised,
and the sequences of the proteins and peptides were analyzed using mass
spectrometry. We compared the proteins and peptides present in the gels
loaded with human albumin, with or without corisin. The results,
including the list of identified proteins and peptides, are presented
in Supplementary Data Set [164]2. Mass spectrometry analysis revealed
that corisin coexisted with human albumin in the gel loaded with both
albumin and corisin, whereas corisin was absent in the gel loaded
solely with albumin. These observations further support the
co-localization and interaction of corisin with human albumin.
We conducted an immunoprecipitation assay to further corroborate the
formation of a complex between corisin and human albumin. Recombinant
human albumin and synthetic corisin were mixed and incubated at 37 °C
for 30 min. Subsequently, anticorisin mAb was added to the reaction
mixture and incubated at 4 °C for 30 min. Protein G agarose beads were
then added to the mixture and incubated further. After thoroughly
washing the protein G agarose beads, the precipitated proteins were
eluted by adding sodium dodecyl sulfate loading buffer and
centrifugation. The eluted proteins’ supernatant was subjected to gel
electrophoresis and stained with Coomassie blue. The results
demonstrated the co-precipitation of corisin with human albumin
(Fig. [165]4E, lane 6), providing additional evidence of the
interaction and complex formation between corisin and human albumin.
We also demonstrated through Western blotting that the corisin peptide
interacts with human and mouse serum albumin from a commercial source
and with serum albumin from patients with diabetic nephropathy and
healthy subjects (Fig. [166]4F, G), independently of pH conditions
(Supplementary Fig. [167]12A, B). In addition, our findings indicate
that the majority of corisin in serum exists in a bound form, with free
corisin being nearly undetectable (Supplementary Fig. [168]13).
Overall, the results of these validation experiments strongly support
the molecular interaction between albumin and corisin, suggesting a
potential role for albumin in facilitating the transport of corisin to
peripheral organs. This observation could have significant implications
for understanding the mechanisms by which microbiome-derived corisin
influences distant tissues and contributes to disease pathogenesis.
Enhanced proapoptotic activity of corisin in a human albumin-enriched
solution
Previous studies have demonstrated that the conformational structure,
stability, and functional properties of peptides can be significantly
enhanced through complex formation with albumin^[169]43–[170]48. Based
on these findings, and given our observation that corisin interacts
with human albumin, we hypothesized that synthetic corisin is
structurally unstable when isolated in solution. We further propose
that its interaction with human albumin improves the structural
conformation of corisin, thereby enhancing its stability and functional
activity. To validate this hypothesis, we utilized molecular dynamics
simulations to examine the folding patterns of synthetic corisin in
solution, initiating from both the native folded and unfolded states.
Employing Markov State Models, we captured the thermodynamics and
kinetics of distinct folded states, with the resulting free energy
landscape (Supplementary Fig. [171]14A) delineating the energy barriers
between different conformational states of the peptide.
We focused on the sole α-helix present between Val2 and Ser6 in the
native structure, using this inter-residue distance, approximately
3.2 Å in the native peptide, as a key metric to monitor the folding
progress. Additionally, the fraction of native contacts was employed as
another crucial metric. Together, both the inter-residue distance and
the fraction of native contacts provided a comprehensive depiction of
the peptide’s folding-free energy landscape. We observed three minima
in the landscape, each representing distinct conformational states
(Supplementary Fig. [172]14B). In state 1, the peptide exhibits a high
fraction of native contacts, about 0.78, with a Val2-Ser6 distance of
approximately 3 Å. In state 2, while the sole α-helix remains stable,
other peptide parts lose native contact. In state 3, the peptide loses
the α-helix structure entirely. This observation also provides insights
into kinetic transitions across these three states, indicating that
reaching state 1, characterized by a high ratio of interactions
resembling the native structure, requires a significantly longer time.
These results support our hypothesis that synthetic corisin alone is
impaired in adopting its native conformation in solution.
Further investigations included 150-nanosecond molecular dynamics
simulations to assess if corisin undergoes conformational changes upon
interacting with human serum albumin. The root-mean-square deviation
(RMSD) of corisin from its native structure, exceeding 2 Ångstroms
during this interaction, suggests significant conformational
alterations (Supplementary Fig. [173]14C). Subsequent experiments
tested whether this interaction enhances the proapoptotic efficacy of
synthetic corisin. We cultured normal renal proximal tubule epithelial
cells in varying concentrations of corisin diluted in 0.5% recombinant
human serum albumin suspended in saline and compared them with cells
cultured in equivalent concentrations of corisin in saline. The results
showed that corisin diluted in 0.5% recombinant human albumin induced
apoptosis in a dose-dependent manner, with a significantly higher
percentage of apoptotic cells than those treated with saline alone
(Supplementary Fig. [174]15A, B). Similar outcomes were observed in
human Caki-2 tubular epithelial cells and normal human podocytes
stimulated with corisin in the presence or absence of 0.5% recombinant
human albumin. Comparative flow cytometry analysis revealed that
corisin in 0.5% recombinant human albumin significantly increases the
percentage of apoptotic Caki-2 cells and podocytes compared to those
treated with corisin in saline (Supplementary Fig. [175]15C–F). These
findings suggest that the proapoptotic activity of corisin is
significantly potentiated in the presence of human albumin, likely due
to the interaction and complex formation between corisin and human
albumin. Building on these observations, we subsequently employed
corisin diluted in a human albumin solution for our in vitro studies.
In addition, the effect of corisin in 0.5% rhAlb on the expression of
anti-apoptotic factors was assessed. Corisin in rhAlb significantly
downregulated the expression of Bcl-2, Bcl-xL, apoptosis inhibitor 2
(cIAP2), and survivin in primary human renal tubular epithelial cells,
an effect that was reversed by anti-corisin mAb co-treatment
(Supplementary Fig [176]16A, B). Moreover, cell cycle analysis of both
human primary podocytes and renal primary tubular epithelial cells
demonstrated that corisin in 0.5% rhAlb, markedly increased the sub-G1
and G1 populations while decreasing the S-phase population
(Supplementary Fig. [177]17A–D). These alterations were ameliorated by
co-treatment with the anti-corisin monoclonal antibody. Consistent with
these findings, both cell proliferation and proliferation rate were
significantly reduced following corisin exposure (Supplementary
Fig. [178]17E, F). These findings indicate that corisin impairs cell
cycle progression and proliferation while promoting apoptosis in renal
epithelial cells.
Complex formation between corisin and albumin in the urine of diabetic CKD
patients
The demonstration that corisin interacts with human albumin and that
this interaction enhances its cell injury capacity prompted the
hypothesis that corisin forms complexes with albumin in the urine of
patients exhibiting albuminuria. To test this hypothesis, urine samples
from patients with diabetic CKD and from healthy controls were
subjected to electrophoresis, followed by Western blot analysis using
monoclonal antibodies against corisin and human albumin. Fragmentation
and polymeric complexes of albumin have been reported in urine of DM
patients^[179]49,[180]50. Western blot results revealed bands exceeding
60 kDa and 40 kDa in all urine samples from patients with diabetic
nephropathy but no bands in samples from healthy controls
(Fig. [181]4H, I). These findings suggest that corisin is indeed
complexed with albumin in the urine of patients with albuminuria,
potentially mediating the adverse effects associated with
albuminuria^[182]51–[183]53.
Corisin penetrates kidney cells via the albumin receptor cubilin to target
the mitochondria
Previous studies have demonstrated that corisin specifically targets
mitochondria^[184]27,[185]29. After establishing that corisin interacts
with human albumin, we investigated whether corisin penetrates kidney
cells through albumin receptors to target the mitochondria. We focused
on renal tubular epithelial cells and podocytes, as these cell types,
unlike mesangial cells, express albumin receptors. Initially, we
cultured Caki-2 tubular renal epithelial cells and primary human
podocytes with fluorescein isothiocyanate (FITC)-labeled corisin or a
FITC-labeled scrambled peptide dissolved in 0.5% human albumin.
Following the addition of a Mito-tracker, we assessed changes in
mitochondrial staining. Cells treated with FITC-labeled corisin
exhibited dual staining of green and red, indicating mitochondrial
targeting by corisin, unlike those treated with the scrambled peptide
(Fig. [186]5A–D).
Fig. 5. Corisin penetrates kidney cells via the human albumin receptor
cubilin to target mitochondria.
[187]Fig. 5
[188]Open in a new tab
A–D, Human Caki-2 cell lines and normal human primary podocytes were
cultured to near confluence, then treated with fluorescein
isothiocyanate (FITC)-labeled corisin (corisin-FITC) or FITC-labeled
scrambled peptide for 4 h. MitoTracker was added 30 min before fixation
to label mitochondria, and cells were subsequently imaged via
microscopy. The fluorescence intensities of F-actin and DAPI were
quantified using ImageJ. Scale bars represent 20 µm. n = 8 replicates.
Data are presented as mean ± SD. Statistical analysis was conducted
using a two-sided unpaired t-test. E–G, Normal human primary podocytes
were cultured for 48 h with siRNA targeting CD36, megalin, cubilin,
SPARC, and FCGRT, and the relative expression of albumin receptors was
evaluated. n = 4 replicates. Data are expressed as the mean ± SD.
Statistical comparison between each receptor-specific siRNA and its
corresponding control siRNA was performed using a two-sided unpaired
t-test. In a separate experiment, cells were cultured under the same
conditions and then treated with corisin-FITC, MitoTracker, and DAPI
and observed using fluorescence microscopy. The fluorescence
intensities of F-actin and DAPI were quantified using ImageJ, an
open-source software from the National Institutes of Health (NIH).
Scale bars represent 20 µm. n = 8 replicates. n = 12 in control and
n = 6 in remaining groups. Data are presented as mean ± SD. Statistical
analysis was conducted using ANOVA followed by Dunnett’s test; all
tests were two-sided. The source data are available in the Source Data
file.
Subsequently, we pretreated human podocytes with small interfering RNA
(siRNA) targeting several albumin receptors, including cubilin,
megalin, CD36, SPARC (secreted protein acidic and rich in cysteine),
and the neonatal Fc receptor (FcRn), which is encoded by the FCGRT
gene. After exposure to FITC-labeled corisin dissolved in 0.5% human
albumin, we observed significant inhibition of mRNA expression for all
tested albumin receptors by their respective siRNAs. However, only the
pretreatment with cubilin siRNA significantly suppressed the
mitochondrial targeting by corisin (Fig. [189]5E–G). These results
suggest that the albumin receptor cubilin mediates corisin’s
penetration into cells to specifically target the mitochondria.
Cellular entry of corisin in the absence of human albumin
To investigate whether corisin can penetrate cells independently of
albumin, normal primary renal tubular epithelial cells were cultured in
the presence of corisin-FITC diluted in either 0.5% recombinant human
albumin (rhAlb) or saline. The intracellular uptake of corisin was
tracked and quantified. The results demonstrated that corisin can
partially penetrate cells even in the absence of albumin, suggesting
the existence of alternative mechanisms for cellular entry
(Supplementary Fig. [190]18A, B). These findings indicate that
internalization of corisin may not be solely dependent on albumin
binding and raise the possibility that alternative mechanisms, such as
endocytosis, may contribute to its cellular uptake. Further
investigation is warranted to elucidate the precise pathways involved.
Corisin induces senescence of parenchymal renal cells to accelerate kidney
fibrosis
After elucidating how corisin targets renal tissues and penetrates
kidney cells, we investigated the potential mechanism by which corisin
could be involved in the progression of kidney fibrosis. Previous
studies have shown that corisin exacerbates pulmonary fibrosis and
acute lung injury by inducing mitochondria-mediated apoptosis in
alveolar epithelial cells and promoting an inflammatory
response^[191]27,[192]29,[193]54. Furthermore, the present study
recapitulated the apoptotic activity of corisin on normal human
podocytes and renal tubular epithelial cells (Supplementary
Fig. [194]15A–F), as previously reported^[195]28. Dysfunctional cell
death disrupts the glomerular filtration barrier and tubular integrity,
leading to proteinuria, reduced glomerular filtration rate (GFR), and
tubular dysfunction^[196]55,[197]56. Therefore, the apoptotic activity
of corisin on renal parenchymal cells may also be a critical
contributor to the pathogenesis of nephropathy.
Importantly, dysregulation of apoptosis is closely linked to senescence
and the development of age-related diseases, including kidney
fibrosis^[198]57,[199]58. As organisms age, the accumulation of
cellular damage and oxidative stress leads to increased apoptotic
activity, contributing to tissue degeneration and functional
decline^[200]57,[201]58. Our previous research has demonstrated that
corisin-induced apoptosis is associated with the activation of the
intrinsic mitochondrial apoptotic pathway and the elevated production
of reactive oxygen species^[202]28,[203]29. Based on the foregoing
observations, we hypothesized that corisin-associated apoptosis is part
of the aging process induced by corisin during kidney fibrosis. To test
this hypothesis, we treated primary human podocytes and human renal
tubular epithelial cells with corisin and evaluated the activity of
senescence-associated β-galactosidase (SAβGal), a marker of aging, and
the mRNA expression of senescence markers, including tumor protein p53
(TP53, p53), Ki-67 (MKI67), and the cyclin-dependent kinase inhibitors
CDKN2B (p15), CDKN2A (p16), CDKN1A (p21), and CDKN1B
(p27)^[204]59,[205]60. Additionally, we assessed components of the
senescence-associated secretory phenotype (SASP), such as matrix
metalloproteinase-12 (MMP-12) and osteopontin (SPP1)^[206]59,[207]60.
Corisin significantly enhanced the activity of SAβGal in human primary
podocytes, human renal primary tubular epithelial cells, and the Caki-2
tubular epithelial cell line compared to controls. This enhanced
activity was significantly reversed by culturing the cells in the
presence of the anticorisin mAb21A (Fig. [208]6A–F). The expression of
the senescence markers p16, p21, p27, p53, and SPP1 was significantly
increased by corisin in both human podocytes and renal tubular
epithelial cells compared to control cells. Conversely, Ki-67
expression was significantly decreased by corisin in both cell types.
Compared to controls, the relative mRNA expression of MMP-12 was
significantly increased by corisin treatment in podocytes but not in
tubular epithelial cells (Fig. [209]6G, H). The effects of corisin on
the mRNA expression of p21, p27, p53, and SPP1 were significantly
blocked by treating the cells with anticorisin mAb. Similarly, the
effect of corisin on the relative mRNA expression of Ki-67 was
significantly inhibited in human renal tubular epithelial cells in the
presence of anticorisin mAb (Fig. [210]6G, H). Abnormal nuclear
morphology is a hallmark of senescent cells^[211]61. In line with this,
human primary podocytes and renal primary tubular epithelial cells
cultured in the presence of corisin exhibited an increased nuclear
area, nuclear perimeter, and reduced circularity compared to controls.
Notably, these nuclear abnormalities were significantly ameliorated by
treatment with anti-corisin mAb (Supplementary Fig. [212]19A, B;
Supplementary Fig. [213]20A, B). We also evaluated the expression of
p21 (Fig. [214]6I, J), and p53 (Supplementary Fig. [215]21A, B) in
diabetic TGFβ1 TG mice with kidney fibrosis by immunohistochemistry and
found that corisin induced enhanced protein expression in diabetic mice
with kidney fibrosis compared to counterpart mice treated with
anticorisin mAb. Overall, these observations indicate that
corisin-associated senescence induction in renal parenchymal cells may
play an essential role in the progression of kidney fibrosis.
Fig. 6. Corisin induces increased expression of senescence markers in kidney
cells.
[216]Fig. 6
[217]Open in a new tab
A–F Induction of senescence-associated β-galactosidase (SAβGal)
activity by corisin. Human Caki-2 cells (left panels), human primary
renal tubular epithelial cells (middle panels), and normal human
primary podocytes (right panels) were cultured in the absence or
presence of 20 or 40 µg/mL of corisin, diluted in 0.5% recombinant
human albumin (rhAlb), or 40 µg/mL of corisin with anti-corisin
antibody. SAβGal activity was measured, and %SAβGal-positive cells was
counted across multiple high-power fields using immunofluorescence
microscopy. Scale bars represent 50 µm. Control, n = 6; corisin
(20 µg/mL), n = 6; corisin (40 µg/mL), n = 6; corisin (40 µg/mL) +
anticorisin (200 µg/mL), n = 6. Data are expressed as the mean ± SD.
Statistical analysis by ANOVA followed by the Newman-Keuls test; all
tests were two-sided. G–H Increased mRNA expression of
senescence-associated factors in kidney cells by corisin. Normal human
podocytes (n = 6) and normal human primary tubular epithelial cells
(n = 5) were cultured in the absence or presence of 20 and 40 µg/mL
corisin, diluted in 0.5% recombinant human albumin, or 40 µg/mL of
corisin with anti-corisin antibody for 48 h. mRNA expression levels of
cyclin-dependent kinase inhibitors p15 (CDKN2), p16 (CDKN2A), p21
(CDKN1A), p27 (CDKN1B), p53 (TP53), Ki-67 (MKI67), matrix
metalloproteinase-12 (MMP12), and secreted phosphoprotein 1 (SPP1, also
known as osteopontin) were evaluated by RT-PCR. Normally distributed
data are presented as mean ± SD, while skewed data are expressed as the
median with interquartile range. Statistical significance was assessed
using ANOVA with the Newman-Keuls or Dunn’s test; all tests were
two-sided. I, J Anticorisin antibody inhibits the expression of
senescence markers. Diabetes mellitus (DM) was induced in transforming
growth factor β1 (TGFβ1) transgenic (TG) mice by streptozotocin. The
mice were divided into a DM TG/control IgG group (n = 5) and a DM
TG/anticorisin group (n = 5). Wild-type (WT) mice injected
intraperitoneally with saline (SAL, n = 4) served as controls.
Paraffin-embedded kidney tissue sections were prepared for p21
immunofluorescent staining. Green immunofluorescent signals (red
arrows) indicate p21 expression. The p21-positive area was quantified
using WinROOF. Scale bars represent 100 µm. Data are presented as
mean ± SD. Statistical significance was evaluated by ANOVA followed by
the Newman-Keuls test; all tests were two-sided. The source data are
available in the Source Data file.
Enhanced expression of corisin and p21 in renal tissue from CKD patients
Previous research has demonstrated that p21 expression remains elevated
in the renal tissue of CKD patients, even after improvements in blood
glucose levels^[218]62. This persistent elevation is associated with
the severity of diabetic kidney disease and serves as a marker of
sustained renal damage^[219]62. In this study, we assessed the presence
of corisin, apoptotic cells, and p21 expression in renal tissue from
CKD patients using immunohistochemistry and compared it to normal renal
tissue. Consistent with our observations in mouse models of diabetic
TGFβ1 TG mice with kidney fibrosis, immunostaining for corisin, the
senescence markers p21, and apoptotic cells were significantly
increased in renal tissue from CKD patients compared to control tissue
(Supplementary Fig. [220]22A, B). These findings further support the
translational relevance of our mouse model results to human disease.
Corisin induces epithelial-mesenchymal transition in kidney parenchymal cells
Given the data above showing the potential involvement of corisin in
senescence, the general knowledge that senescence and
epithelial-mesenchymal transition (EMT) share common regulatory
pathways in fibrosis, the role of EMT in facilitating the escape of
cells from senescence, and the influence of the SASP on tissue
remodeling and fibrosis, we hypothesized that corisin might also induce
EMT in kidney parenchymal cells^[221]63–[222]65. To test this
hypothesis, we cultured human primary podocytes and human renal
proximal tubule epithelial cells in the presence of corisin and
evaluated markers of EMT. Podocytes and human renal proximal tubule
epithelial cells cultured with corisin displayed a spindle-shaped
morphology with significantly enhanced filamentous actin staining
compared to control cells. Additionally, the intensity of filamentous
actin was significantly reduced in cells treated with corisin plus the
anticorisin mAb (Fig. [223]7A–D). The mRNA expression of α-smooth
muscle actin (α-SMA), fibronectin, and collagen I was significantly
increased by corisin in both podocytes and renal tubular epithelial
cells compared to controls. Anti-corisin mAb significantly inhibited
the increased expression of α-SMA and fibronectin in podocytes, as well
as α-SMA, fibronectin, and collagen I in renal tubular epithelial
cells. Furthermore, the mRNA expression of TGFβ1 was significantly
elevated in podocytes treated with corisin, but this effect was
significantly reduced by the anti-corisin mAb (Fig. [224]7E, F). These
results suggest that corisin can directly promote fibrogenesis by
inducing EMT in kidney parenchymal cells.
Fig. 7. Corisin induces epithelial-mesenchymal transition in podocytes and
renal tubular epithelial cells.
[225]Fig. 7
[226]Open in a new tab
A–D Normal human primary podocytes or primary renal proximal tubular
epithelial cells (RPTEC) were cultured for 48 h under conditions
without corisin (medium containing 0.5% recombinant human albumin) and
with corisin at concentrations of 20 and 40 µg/mL, dissolved in 0.5%
recombinant human albumin, or 40 µg/mL of corisin with anti-corisin
antibody. Subsequently, cells were stained with Phalloidin-iFluor™ 488
and 4’,6-diamidino-2-phenylindole (DAPI). Fluorescence intensity of
F-actin and cell counts were quantified using ImageJ, a public domain
software from the National Institutes of Health (NIH). n = 4 in
experiments using podocytes and n = 6 in experiments using RPTEC. Scale
bars indicate 20 µm. Data are expressed as the mean intensity per cell
ratio ± SD. Statistical analysis was performed using ANOVA followed by
the Newman-Keuls test; all tests were two-sided. E, F Cells were
cultured under the same conditions to collect total RNA from each
treatment group and assess the mRNA expression of α-smooth muscle actin
(ACTA2), fibronectin (FN1), collagen I (COL1a1), and transforming
growth factor β1 (TGFB1). n = 5 in (E) and n = 6 in (F). Normally
distributed data are presented as mean ± SD, while skewed data are
expressed as the median with interquartile range. Statistical
significance was assessed using ANOVA, followed by the Newman-Keuls or
Dunn’s test; all tests were two-sided. The source data are available in
the Source Data file.
Single-cell RNA sequencing analysis reveals distinct transcriptional profiles
in corisin-treated cells
Single-cell RNA sequencing analysis was performed to evaluate the
transcriptional impact of corisin on human renal primary proximal
tubular epithelial cells, comparing corisin-treated cells with those
treated with a scrambled peptide control. The violin plots show
significantly higher expression levels, while the Uniform Manifold
Approximation and Projection (UMAP) plots confirm higher expression
intensities across the cell population for various senescence-related
and SASP markers, including cyclin-dependent kinase inhibitor 1 A
(CDKN1A), serpin family E member 1 (SERPINE1), C-X-C motif chemokine
ligand 8 (CXCL8), insulin-like growth factor binding protein 5
(IGFBP5), heparin-binding EGF-like growth factor (HBEGF), and EMT
markers such as tropomyosin 2 (TPM2) and transgelin (TAGLN)
(Fig. [227]8A, B). The expression of other EMT markers, including
myosin light chain 9 (MYL9) and calponin 1 (CNN1), was also
significantly upregulated in corisin-treated cells compared to
scrambled peptide controls (Fig. [228]8C). Additionally, several other
SASP markers were significantly elevated in the presence of corisin
compared to the scrambled peptide, including growth arrest and
DNA-damage-inducible alpha (GADD45A), myelocytomatosis oncogene (MYC),
fibroblast growth factor 1 (FGF1), tumor necrosis factor (TNF), C-C
motif chemokine ligand 20 (CCL20), amphiregulin (AREG), inhibin subunit
beta A (INHBA), C-X-C motif chemokine ligand 5 (CXCL5), C-X-C motif
chemokine ligand 3 (CXCL3), and nerve growth factor (NGF)
(Fig. [229]8D).
Fig. 8. Corisin induces the expression of senescence-associated factors in
primary renal proximal tubular epithelial cells.
[230]Fig. 8
[231]Open in a new tab
A Renal proximal tubular epithelial cells (RPTEC) were cultured for
24 h in the presence of corisin or scrambled peptides and subsequently
subjected to single-cell RNA sequencing analysis. B Violin plots and
Uniform Manifold Approximation and Projection (UMAP) plots depict the
expression of senescence markers and senescence-associated secretory
phenotype (SASP) components in corisin- and scrambled peptide-treated
cells. Statistical analysis was conducted using a two-sided
Mann–Whitney U test. C Violin plots and UMAP plots illustrating the
expression of epithelial-mesenchymal transition (EMT) markers in
corisin- and scrambled peptide-treated cells. Statistical analysis was
conducted using a two-tailed Mann-Whitney U test. D Additional
senescence, SASP, and EMT markers in corisin- and scrambled
peptide-treated cells. Statistical analysis was conducted using a
two-sided Mann-Whitney U test. For panels B, C, and D, single-cell RNA
sequencing was performed using three independent RPTEC cultures per
stimulant, yielding 9327 high-quality cells for the corisin-treated
group and 8186 cells for the scrambled peptide (control) group. Violin
plots in (B–D), represent the distribution of gene expression values.
The width of each violin corresponds to the kernel density estimation.
The thick vertical bar within each violin denotes the interquartile
range (25th-5th percentile), and the thin line (whisker) extends to the
minimum and maximum values within 1.5 times the interquartile range.
The median is not explicitly displayed.
The relationship between cellular senescence and EMT after corisin
stimulation was evaluated using single-cell analysis. Specifically,
expression levels of mesenchymal markers, including α-smooth muscle
actin (ACTA2), CNN1, TAGLN, and TPM2, were compared between cells with
high and low expression of the senescence marker p21, using contingency
table analysis. Chi-square testing indicated no significant increase in
mesenchymal marker expression in cells with elevated p21 levels,
suggesting that senescence does not directly drive EMT at the
single-cell level during corisin stimulation (Supplementary
Fig. [232]23A–D). Notably, CXCL8 (IL-8) and growth differentiation
factor 15 (GDF15), key components of the senescence-associated
secretory phenotype (SASP), were significantly upregulated in p21-high
cells (Supplementary Fig. [233]24A, B). These findings imply that
senescence and EMT are unlikely to co-occur within individual
corisin-stimulated cells. Instead, EMT may result from SASP factors
secreted during corisin-induced senescence rather than being a direct
consequence of cellular senescence. Similarly, no significant increase
in the proapoptotic markers p53, BAX and caspase 3 was observed in
cells exhibiting elevated p21 levels, suggesting that senescence and
apoptosis likely occur in distinct cell populations during corisin
stimulation (Supplementary Fig. [234]25A–C).
A heatmap demonstrates the distinct transcriptional profiles of key
upregulated and downregulated genes in corisin-treated cells compared
to scrambled peptide controls (Fig. [235]9A). The volcano plot analysis
highlights the most significantly upregulated genes in corisin-treated
cells, including CNN1, CXCL5, IGFBP5, INHBA, p53-induced protein with a
death domain 1 (PIDD1), and wingless-type MMTV integration site family,
member 7 A (WNT7A), with fold changes exceeding 2 and P-values below
0.05. Conversely, genes such as coiled-coil domain containing 160
(CCDC160) and krüppel-like factor 15 (KLF15) were among those
significantly downregulated (Fig. [236]9B).
Fig. 9. Corisin stimulation in primary renal proximal tubular epithelial
cells promotes pathways associated with the DNA damage response, cellular
senescence, the senescence-associated secretory phenotype (SASP), and
myofibroblast-like differentiation.
[237]Fig. 9
[238]Open in a new tab
A Heatmap showing gene expression in cells stimulated with corisin or a
scrambled peptide control. B Volcano plots showing differentially
expressed genes between corisin- and scrambled peptide-treated renal
proximal tubular epithelial cells (RPTEC), based on single-cell RNA
sequencing. Each point represents a gene, with the x-axis showing the
log[2] fold change and the y-axis showing the -log[10] adjusted
p-value. Differential expression analysis was performed using a
two-tailed Mann–Whitney U test. Since gene expression was assessed
across thousands of genes, p-values were adjusted for multiple
comparisons using the Benjamini-Hochberg method to control for the
false discovery rate. C Kyoto Encyclopedia of Genes and Genomes (KEGG)
pathway enrichment analysis reveals significant overrepresentation of
pathways related to cellular senescence in corisin-treated renal
proximal tubular epithelial cells. Enrichment analysis was performed
using the hypergeometric test, and p-values were adjusted for multiple
comparisons using the Benjamini-Hochberg method to control the false
discovery rate. D Categorization of differentially expressed genes by
their biological roles, demonstrating their involvement in various
cellular processes.
Pathway enrichment analysis using Kyoto Encyclopedia of Genes and
Genomes (KEGG) terms revealed significant enrichment of pathways
associated with p53 signaling, cellular senescence, and interleukin-17
(IL-17) signaling, among others. Notably, pathways involved in cell
cycle regulation, cytokine-cytokine receptor interaction, and
cancer-related pathways were also highlighted, indicating a broad
impact of corisin on processes linked to inflammation, senescence, and
cell proliferation (Fig. [239]9C).
Categorization of the differentially expressed genes by their
biological roles revealed increased expression of markers associated
with the DNA damage response (blue cluster), cellular senescence (red
cluster), myofibroblast-like factors (green cluster), and
senescence-associated secretory phenotype (yellow cluster) in
corisin-treated cells compared to scrambled peptide controls
(Fig. [240]9D).
These results collectively suggest that corisin stimulation in renal
tubular epithelial cells promotes pathways associated with the DNA
damage response, induces cellular senescence, contributes to the
development of a senescence-associated secretory phenotype, and
enhances myofibroblast-like differentiation. This further indicates
that corisin may play a significant role in driving renal fibrosis by
promoting both cellular senescence and a fibrotic phenotype.
Discussion
This study demonstrates that increased corisin release by the
microbiome in patients with diabetic CKD and mice with kidney fibrosis
correlates with disease severity and renal dysfunction, contributes to
the progression of kidney fibrosis, and interacts with human serum
albumin to facilitate its transport to the kidneys, where it
accelerates cellular senescence and induces epithelial-mesenchymal
transition.
The involvement of microbial dysbiosis in the pathogenesis of kidney
fibrosis in diabetic CKD has garnered increasing attention due to its
significant impact on inflammation, metabolic disturbances, and overall
kidney function. Disruption of the gut microbiota is associated with
heightened systemic inflammation and altered immune responses, which
are crucial in the progression of kidney fibrosis^[241]66. Dysbiosis
has also been linked to the activation of the intrarenal
renin-angiotensin system, further contributing to kidney injuries in
diabetic nephropathy^[242]67. Additionally, microbial dysbiosis can
participate in overproduction and accumulation of bacterial
metabolites, including albumin-bound uremic toxins such as indoxyl
sulfate, p-cresyl sulfate, phenyl sulfate, and 4-ethylphenyl
sulfate^[243]19–[244]22. These metabolites contribute to oxidative
stress, kidney cell apoptosis, inflammation, and profibrotic activity,
exacerbating kidney damage in diabetic nephropathy^[245]68,[246]69.
However, the potential of specific microbiome-derived metabolites as
reliable biomarkers for dysbiosis or therapeutic targets remains
uncertain^[247]66,[248]70. Our previous research found elevated levels
of corisin, a microbiome-derived peptide, in the blood of patients with
various severe conditions, including idiopathic pulmonary fibrosis with
acute exacerbation, COVID-19-associated acute lung injury, and acute
cholangitis^[249]27,[250]30,[251]71. Corisin induces apoptosis of
parenchymal cells, including those in the lungs and kidneys, and
promotes inflammation by stimulating the secretion of inflammatory
cytokines and chemokines and activating the coagulation
system^[252]28,[253]29. However, its role in diabetic nephropathy
remained unclear. In this study, we recapitulated the apoptotic effects
of corisin on kidney cells using primary human podocytes and renal
tubular epithelial cells and demonstrated that an anti-corisin mAb
significantly attenuates corisin-induced cellular damage in vitro while
mitigating kidney fibrosis and dysfunction in diabetic mice.
Furthermore, we observed that elevated circulating corisin levels in
diabetic patients with CKD and in mice with diabetic kidney fibrosis
correlate with renal dysfunction and fibrosis progression.
Interestingly, serum corisin levels were also significantly elevated in
non-diabetic patients with CKD compared to healthy controls, suggesting
that dysregulation of corisin may also occur in CKD of non-diabetic
etiology. These observations underscore the role of corisin in the
progression of kidney fibrosis and highlight its potential as a
therapeutic target in this debilitating condition. However, the extent
of corisin’s interaction with other uremic nephrotoxins that accumulate
during diabetic nephropathy, as well as the magnitude of their additive
effects, remains unclear and warrants further investigation.
Dysregulated apoptosis is intricately linked to cellular senescence.
With aging, cells accrue damage and oxidative stress, heightening
apoptotic activity that contributes to tissue degeneration and
fibrosis, ultimately precipitating a decline in organ
function^[254]57,[255]58. Senescent cells amass in tissues over time,
exacerbating fibrosis by releasing an array of pro-inflammatory and
profibrotic factors, termed the Senescence-Associated Secretory
Phenotype (SASP)^[256]72. This phenotype includes cytokines, growth
factors, and proteases, which perpetuate a chronic inflammatory state,
fostering ongoing fibrosis^[257]73. Such mechanisms underscore
senescence as a pivotal factor in the pathogenesis of organ fibrosis,
particularly in diabetic CKD. A critical adaptive response enabling
cells to circumvent senescence is EMT^[258]60. This process involves
the transformation of epithelial cells into myofibroblasts, which are
central to fibrotic tissue remodeling and are prolific producers of
collagen^[259]74. These myofibroblasts may originate from resident
epithelial cells, fibroblasts, or circulating fibrocytes, leading to
excessive extracellular matrix deposition and scar
formation^[260]75,[261]76. In diabetic CKD, these fibrogenic processes
are intensified by hyperglycemia-induced oxidative stress and the
activation of TGFβ signaling pathways, further driving
fibrosis^[262]77.
While EMT is well-documented in vitro and has been observed in certain
in vivo models, its significance in human CKD remains a subject of
considerable controversy^[263]78. Recent evidence suggests that rather
than undergoing a complete transition into myofibroblasts, renal
tubular epithelial cells may acquire mesenchymal characteristics and
the ability to secrete profibrotic cytokines while remaining attached
to the basement membrane, a phenomenon termed partial
EMT^[264]79,[265]80. This concept offers a more nuanced perspective on
how epithelial cells contribute to kidney fibrosis while retaining
aspects of their epithelial identity. In our current study, we observed
that exposure to corisin significantly enhanced markers of cellular
senescence, including SAβGal activity and the expression of p16, p21,
p27, and p53, while concurrently decreasing the proliferation marker
Ki-67. The SASP protein osteopontin levels were notably elevated in
human podocytes and primary renal tubular epithelial cells, suggesting
that corisin induces senescence in renal cells. Regarding EMT, corisin
promoted mesenchymal transformation in vitro in both podocytes and
renal tubular epithelial cells, potentially via the secretion of SASP
factors. These findings suggest that corisin-associated senescence may
contribute to the expansion of collagen-producing cells, although the
in vivo significance of this EMT-promoting effect remains to be
elucidated. Importantly, most senescence and EMT markers were
attenuated by co-treatment with an anti-corisin monoclonal antibody.
Additionally, single-cell RNA sequencing analysis of renal proximal
tubular epithelial cells reinforced these observations, demonstrating
that corisin induces cellular senescence, contributes to the
development of a SASP, and enhances myofibroblast-like differentiation.
Collectively, these findings suggest that corisin plays a significant
role in driving renal fibrosis by promoting cellular senescence and a
fibrotic phenotype.
Cellular senescence is a dynamic, heterogeneous, and multistage process
regulated by multiple factors, including cell type, the extent of
damage, duration of stress exposure, and the nature of intracellular
signaling pathways activated^[266]81,[267]82. Although it is well
established that senescent cells often upregulate anti-apoptotic
proteins (e.g., members of the BCL-2 family) as a survival mechanism
under chronic stress, this is not universally observed across all cell
types or stages of senescence^[268]81,[269]83. For instance, there is
evidence showing that senescent fibroblasts are resistant to apoptotic
stimuli than their younger counterparts, despite expressing low levels
of BCL-2 proteins^[270]84. Conversely, vascular endothelial cells
exhibit increased susceptibility to apoptosis during senescence^[271]85
associated with increased activation of ROS-associated signaling
pathways^[272]86, and with reduced Bcl-2 expression alongside elevated
levels of the pro-apoptotic protein Bax^[273]87. In our study,
stimulation with corisin elicited hallmarks of both senescence and
apoptosis. Specifically, human podocytes and renal tubular epithelial
cells displayed elevated expression of senescence markers, including
p21 and SAβGal. Concurrently, corisin exposure for 24 and 48 h led to a
marked reduction in Bcl-2 and survivin levels, accompanied by increased
apoptotic cell death. Flow cytometric analysis supported this dual
response, revealing both G0/G1 phase arrest, characteristic of
senescence, and an increased proportion of cells in the S and subG1
phases, the latter often reflecting DNA fragmentation associated with
apoptosis. These findings suggest that corisin induces a hybrid stress
phenotype that engages both senescence- and apoptosis-related signaling
pathways, mirroring cellular responses previously described in
endothelial cells under prolonged stress. Depending on the nature and
intensity of the insult, cells may activate both programs in parallel,
with the ultimate fate determined by the balance of pro-survival and
pro-death signals^[274]81. Furthermore, our single-cell transcriptomic
analysis revealed that cells with high p21 expression did not uniformly
express pro-apoptotic markers, indicating the existence of distinct
subpopulations, some undergoing apoptosis and others entering
senescence. This observation raises the possibility that corisin
induces sufficient stress to trigger apoptosis in one subset of cells
while priming another for senescence. Future time-course studies and
deeper single-cell analyses will be essential to delineate the temporal
dynamics and molecular regulators governing these divergent cell fates.
Albumin is essential for the transport and distribution of various
substances, including microbial metabolites, within the human body. It
binds to and carries protein-bound substances, ensuring their movement
between different body compartments, which is vital for maintaining
physiological homeostasis^[275]88,[276]89. Albumin facilitates the
transport of hormones, metals, and fatty acids by binding to specific
sites, making it a key player in the body’s regulatory
mechanisms^[277]89. Additionally, it is involved in the transport of
drugs and toxic substances, significantly influencing their
pharmacokinetics and toxicokinetics^[278]90. Albumin may also carry
substances into cells through its various receptors^[279]91. In our
present study, utilizing molecular dynamics simulations and validation
experiments, we found that corisin interacts with human albumin,
suggesting serum albumin as a potential carrier of corisin. This
interaction improves the structural conformation of corisin and
enhances its pro-apoptotic activity against podocytes and renal tubular
epithelial cells. Moreover, we detected corisin-albumin complexes in
the urine of diabetic patients with CKD, supporting the role of albumin
in transporting corisin to the intratubular spaces of the kidneys under
diabetic conditions. Albuminuria is a well-recognized marker of CKD
progression, particularly in patients with diabetic
nephropathy^[280]92–[281]94. Once albuminuria develops, kidney function
declines more rapidly^[282]92–[283]94. While the underlying mechanism
remains unclear, evidence suggests that albuminuria induces cytotoxic
effects on tubular epithelial cells and podocytes, leading to
apoptosis, inflammation, fibrosis, and subsequent nephron damage,
ultimately exacerbating kidney dysfunction^[284]51–[285]53. Our
findings suggest that corisin, when bound to albumin, may partly
contribute to the toxic effects associated with albuminuria in diabetic
patients with CKD. Albumin receptors may facilitate the entry of
albumin-bound corisin into podocytes and proximal renal tubular
epithelial cells, further exacerbating cellular injury (Fig. [286]10).
Specifically, we identified cubilin, an albumin receptor highly
expressed in kidney cells and crucial for albumin reabsorption in
tubular epithelial cells, as a mediator of corisin penetration into
kidney cells^[287]95–[288]97. This suggests that cubilin may be a
critical contributor to cell injury in diabetic nephropathy.
Collectively, our results indicate that albumin and its receptor
cubilin play pivotal roles in facilitating corisin’s entry into kidney
cells and promoting intracellular damage, thereby accelerating kidney
injury and fibrosis in diabetic nephropathy. Further research into the
interactions between albumin, cubilin, and corisin may offer novel
therapeutic avenues for preventing or slowing kidney disease
progression in diabetic patients.
Fig. 10. Corisin-induced kidney cell damage and fibrosis in diabetic chronic
kidney disease and the protective potential of anticorisin monoclonal
antibody.
[289]Fig. 10
[290]Open in a new tab
Diabetes-associated dysbiosis increases corisin release from the
microbiome into systemic circulation, where it binds to serum albumin.
The corisin-albumin complex reaches the glomeruli and proximal tubular
epithelial cells, binding to cubilin, an albumin receptor, and thereby
facilitating corisin entry into podocytes and proximal tubular
epithelial cells. Within these cells, corisin induces a
senescence-associated secretory phenotype, resulting in the elevated
secretion of inflammatory cytokines, chemokines, matrix
metalloproteinases, and growth factors that promote inflammation,
epithelial-mesenchymal transition, apoptosis of podocytes and tubular
epithelial cells, myofibroblast recruitment, and extracellular matrix
deposition (e.g., collagen I). Areas of the glomeruli and tubules
affected by increased apoptosis are subsequently replaced by fibrotic
tissue, accelerating disease progression and leading ultimately to a
fatal outcome. The anti-corisin monoclonal antibody binds to corisin
peptides, blocking their pro-senescence activity and mitigating disease
progression. mAb, monoclonal antibody.
Upon cellular entry, the precise intracellular molecular mode of action
of corisin remains incompletely understood. However, our previous
investigations demonstrated that corisin induces apoptosis through
mitochondrial dysfunction and activation of the intrinsic apoptotic
pathway. Specifically, corisin treatment leads to increased
accumulation of reactive oxygen species (ROS), mitochondrial membrane
depolarization, and activation of proapoptotic factors. Moreover, our
earlier research revealed that corisin promotes p53 phosphorylation, a
key regulator that governs both apoptosis and senescence. Building on
these findings, the present study establishes that corisin accelerates
cellular senescence. Given that mitochondrial dysfunction and oxidative
stress are well-recognized drivers of cellular senescence and that p53
plays a dual regulatory role in apoptosis and senescence via p21
modulation, our current findings, demonstrating elevated expression of
senescence-associated markers, including p21 and p53, strongly suggest
that corisin-induced cellular stress and mitochondrial dysfunction
contribute to senescence-related pathways. Nonetheless, further
investigations are essential to conclusively determine whether
corisin-mediated mitochondrial dysfunction and oxidative stress
actively drive senescence and influence long-term cellular outcomes.
The present study has several limitations. First, the clinical
component was based on a relatively small patient cohort from a single
center, which may limit the generalizability of the findings. Second,
the specific or additive contribution of corisin could not be clearly
distinguished from that of other uremic retention solutes and toxins
known to contribute to fibrogenic processes. Third, we did not
investigate the potential involvement of corisin-producing bacteria at
other mucosal sites, such as the lungs or skin, which may also
influence systemic corisin levels. Finally, the mechanistic
understanding of the downstream signaling pathways through which
corisin promotes fibrosis remains incomplete. Nonetheless, the findings
presented in this study provide a strong foundation for future
investigations to define the precise role of corisin in kidney fibrosis
and assess its potential as a novel therapeutic target.
In conclusion, this study highlights the critical role of corisin in
the progression of diabetes-associated kidney fibrosis. Our findings
demonstrate that corisin is significantly upregulated in patients with
diabetic CKD and in animal models of kidney fibrosis, correlating
strongly with disease severity and renal dysfunction. Corisin enhances
kidney fibrosis through its interaction with human serum albumin, which
facilitates its transport to the kidneys. This interaction promotes
cellular senescence and induces EMT, mechanisms central to the
progression of kidney fibrosis. Importantly, our therapeutic
intervention using an anti-corisin monoclonal antibody significantly
improved renal function and halted fibrosis in diabetic mice,
underscoring corisin’s potential as a novel therapeutic target
(Fig. [291]10). These novel insights could lead to innovative
treatments that mitigate renal deterioration and improve patient
outcomes in diabetic nephropathy, a significant worldwide health
burden.
Methods
Patients
This cross-sectional study included 35 patients with diabetic chronic
kidney disease (CKD) who presented to our institutions between June and
December 2023. Inclusion criteria consisted of a diagnosis of DM and
CKD. Patients undergoing anticancer treatment and those with incomplete
data or insufficient sample quality were excluded from the study
(Fig. [292]1A). Supplementary Table [293]1 summarizes the patients’
demographics, clinical characteristics, underlying medical conditions,
treatments, and routine laboratory data. The diagnosis and management
of DM followed the criteria recommended by the American Diabetes
Association (ADA)^[294]98. The diagnosis and classification of CKD
adhered to the Kidney Disease: Improving Global Outcomes (KDIGO)
guidelines^[295]99,[296]100. Blood and urine samples were obtained from
all patients and centrifuged, and the serum and urine supernatants were
collected and stored at -80 °C until analysis. Five patients (2 males
and 3 females; mean age ± SD: 51.4 ± 9.4 years) with non-diabetic
chronic kidney disease (CKD) were also included as disease controls.
The underlying etiologies of CKD were arterial hypertension (3 cases),
hyperuricemia with obesity (1 case), and hemolytic uremic syndrome (1
case). Of these patients, four were classified as stage G3 and one as
stage G5, with a mean eGFR of 37.1 ± 18.5 mL/min/1.73 m². In addition,
serum samples from 20 healthy volunteers and urine samples from 5
healthy volunteers were included as healthy controls.
Human kidney sections
Formalin-fixed, paraffin-embedded human kidney sections were obtained
from OriGene, including samples from patients with chronic kidney
disease and those with normal histology. These sections were used for
immunostaining of p21, corisin, and assessment of DNA fragmentation.
Animals
Male wild-type (WT) C57BL/6 J mice (Nihon SLC, Hamamatsu, Japan) and
TGFβ1 transgenic (TG) mice with a C57BL/6 J background, which naturally
develop progressive and fatal kidney fibrosis, were utilized in the
experiment^[297]31. The mice, aged 8–9 weeks, weighed between 20 and
26 g. The TGFβ1 transgenic (TG) mouse model, on a C57BL/6 J background,
was specifically engineered to overexpress the full-length human TGFβ1
gene in glomerular podocytes under the control of the mouse podocin
promoter^[298]31. The TG mouse was generated by preparing a chimeric
podocin-TGFβ1 bacterial artificial chromosome (BAC) transgenic
construct, which contained the full-length coding exons and intervening
introns of the human TGFβ1 gene, replacing the mouse podocin gene locus
via BAC-mediated recombination and genetic engineering^[299]31. The
transgenic mouse line was established by pronuclear injection of the
chimeric construct into C57BL/6 J mouse embryos, and transgenic
founders were verified by Southern blot analysis^[300]31. All
experimental animals were bred and housed in a specific pathogen-free
environment, maintained at a temperature of 21 °C with a 12-hour
light/dark cycle, within the experimental animal facility of Mie
University. Each mouse cage was equipped with wood-wool nesting
material, and the animals had free access to water and food. Genotyping
of the TG mice was performed using standard PCR analysis with DNA
extracted from tail samples and specific primer pairs^[301]31.
Ethical statement
All subjects participating in the clinical investigation provided
written informed consent, and the study protocol was approved by the
Ethical Committees for Clinical Investigation of Mie University
(approval No: H2021-029, approval date: 2021/02/09) and conducted
following the Principles of the Declaration of Helsinki. Written
informed consent was obtained from all participants. This study
complies with ethical standards, with efforts to ensure an inclusive
design across sex and age, equitable contributions across geographical
backgrounds, and data analysis that considers relevant biological and
social factors. The Recombinant DNA Experiment Safety Committee
(approval No: I-744, approval date: 2022/11/26; approval No: I-629,
approval date: 2023/08/09) and the Committee for Animal Investigation
of Mie University approved the experimental protocols (approval No:
2023-17, approval date: 2024/01/09; approval No: 30-14, approval date:
2023/09/04; approval No: 29-23-sai3-h1, approval date: 2023/09/04). We
performed all experimental procedures following internationally
approved laboratory animal care principles published by the National
Institute of Health ([302]https://olaw.nih.gov/). The research followed
the ARRIVE Guidelines for animal investigation, and variables were
measured blindly of the treatment groups.
Reagents
The clear cell renal carcinoma (Caki-2) cell line, derived from the
epithelium of the proximal tubules, and human primary podocytes were
obtained from the American Type Culture Collection (Manassas, VA),
while human renal primary tubular epithelial cells (RPTEC) were sourced
from LONZA (Houston, TX). RPMI 1640 medium (RPMI) was acquired from
Nacalai Tesque (Kyoto, Japan), while fetal bovine serum (FBS) was
obtained from Bio Whittaker (Walkersville, MD). L-glutamine,
penicillin, and streptomycin were procured from Invitrogen (Carlsbad,
CA). Synthetic corisin and its corresponding scrambled peptide were
synthesized and supplied by Peptide Institute Inc. (Osaka, Japan) and
Thermo Fisher Scientific (Waltham, MA).The siRNAs against each albumin
receptor were purchased from Origene (Rockville, MD, USA). Fluorescein
isothiocyanate (FITC)-labeled corisin and FITC-labeled scrambled
peptide were also from Peptide Institute Inc. (Osaka, Japan), and
MitoTracker™ Red CMXRos was from Thermo Fisher Scientific (Waltham,
MA).
Evaluation of corisin interaction with recombinant human albumin
Mixtures of 20 μg recombinant human albumin with several amounts of
synthetic corisin or synthetic scrambled peptide were prepared in
saline and incubated at 37 °C. The mixture was thoroughly mixed by
pipetting, centrifuged, and incubated at 94 °C for 5 min. The samples
were then loaded onto a 10-20% SDS-polyacrylamide gel electrophoresis
(SDS-PAGE) gel. Subsequently, Western blotting was performed using
anticorisin mAb 21 A or anti-albumin antibody.
Evaluation of the interaction between corisin and human serum albumin
Serum samples from patients with diabetic nephropathy and healthy
controls were diluted at a ratio of 1:20 and incubated for 10 min with
corisin (4 µg) diluted in saline. As a control, a mixture containing
corisin (4 µg) and recombinant human albumin (rhAlb, 5 µg) was also
prepared. Each mixture was subsequently subjected to electrophoresis,
followed by Western blot analysis using monoclonal antibodies specific
to corisin or albumin.
Evaluation of the pH-dependent interaction between corisin and human serum
albumin
Serum samples from patients with diabetic nephropathy and healthy
controls were diluted 1:20 and incubated for 10 min with or without
10 µg of corisin in saline solutions adjusted to various pH levels
using hydrochloric acid. In a separate experiment, 2.5 µg of
recombinant human albumin (rhAlb) was incubated with 3 µg of corisin in
saline under different pH conditions. Each mixture was then subjected
to electrophoresis and analyzed by Western blotting using monoclonal
antibodies specific for corisin or human albumin.
Gel band excision for mass spectrometry analysis
A mixture of corisin and rhAlb in a saline solution and a control
sample containing rhAlb alone were prepared. The mixtures were
thoroughly mixed by pipetting, centrifuged, and incubated at 94 °C for
5 min. Subsequently, the samples were subjected to 10–20% SDS-PAGE.
Upon completion of electrophoresis, the gel was carefully removed,
briefly rinsed with distilled water, and then immersed in a fixative
solution with agitation for 30 min. Following fixation, the gel was
transferred to a Coomassie blue staining solution and agitated for
approximately 2 h to visualize the protein bands. The gel was then
destained, and images were captured. Bands of interest were excised
from the gel using a sterile scalpel for subsequent mass spectrometry
analysis.
Mass spectrometry to determine co-localization of corisin with recombinant
human albumin
Proteomic analyses were performed at the Proteomics Core Facility of
the Roy J. Carver Biotechnology Center at the University of Illinois
Urbana-Champaign. Excised gel bands were cut into 1 mm cubes and
de-stained thrice with 50% acetonitrile in 50 mM triethylammonium
bicarbonate. The gel pieces were then dehydrated with acetonitrile and
swelled with 150 µL of 50 mM triethylammonium bicarbonate containing
1 µg of mass spectrometry-grade trypsin (Pierce). The samples were
digested overnight at 37 °C, and on the following day, the digested
peptides were extracted three times with 50% acetonitrile in 5% formic
acid. The samples were dried in a speed vac, desalted with StageTips,1,
and then dried again.
The peptides were resuspended in 15 μL of 5% acetonitrile in 0.1%
formic acid, and 1 μL of each sample was injected into an UltiMate 3000
RSLCnano system coupled to an Orbitrap Fusion mass spectrometer (Thermo
Fisher Scientific). Peptide separation was performed using a 25 cm
Acclaim PepMap 100 C18 column (Thermo Fisher Scientific) maintained at
50 °C, with mobile phases consisting of 0.1% formic acid in water (A)
and 0.1% formic acid in acetonitrile (B). The gradient elution was
carried out over 45 min from 2% to 36% B at a flow rate of 300 nL/min,
followed by column washing and equilibration. The mass spectrometer
operated in positive ion mode. MS1 (first-stage mass analysis) scans
were acquired over a range of 300–2000 m/z at a resolution of 120,000.
The most abundant ions detected in MS1 scans were subjected to CID
(collision-induced dissociation) fragmentation with a normalized
collision energy of 35%, and MS2 (second-stage mass analysis) scans
were acquired in the ion trap, with a total cycle time of 3 s. The
isolation window was set at 1.6 m/z, and dynamic exclusion was applied
with a duration of 60 s.
Raw LC-MS data were analyzed using Mascot Distiller v.2.8.4.0 and an
in-house Mascot server v.2.8.2 (Matrix Science). Database searches were
performed against the Homo sapiens reference proteome from UniProt
(104,451 entries) and a custom FASTA file containing the sequences of
corisin and the scrambled peptide. Search parameters included a peptide
mass tolerance of 10 ppm, a fragment mass tolerance of 0.6 Da, and
allowance for up to two missed tryptic cleavages. Protein
identifications were filtered in Mascot using the default significance
threshold of p < 0.05.
Co-immunoprecipitation of corisin with human albumin
A preparation containing 5 μg of recombinant human albumin and 50 ng of
corisin was constituted in 20 μL of Tris-buffered saline (TBS) and
subsequently incubated at 37 °C for 30 min. Following this, 2 μg of
anti-corisin monoclonal antibody (mAb) at a concentration of 1 mg/mL
was introduced to the solution and incubated at 4 °C for an additional
30 min. Protein G agarose beads (10 μL) were added, and the mixture was
further incubated for 30 min at 4 °C. After the incubation, the beads
were meticulously washed three times with TBS to remove any nonspecific
interactions. Thereafter, 10 μL of 2x sodium dodecyl sulfate (SDS)
loading buffer was added to the beads to elute the proteins, followed
by centrifugation to separate the supernatant. The eluted proteins’
supernatant was then applied to a 10-20% SDS-PAGE.
Post-electrophoresis, Coomassie blue staining was employed to analyze
the protein complexes resultant from the co-immunoprecipitation
procedure.
Cell culture
All cells were cultured in RPMI 1640 medium, enriched with 10% fetal
calf serum, 0.03% (w/v) L-glutamine, 100 IU/ml penicillin, and
100 μg/ml streptomycin. The cultures were maintained in a controlled
environment featuring a 5% CO[2] atmosphere at a constant temperature
of 37 °C.
Cell stimulation with corisin
Given corisin’s interaction with albumin and its stability in the
presence of albumin, unless otherwise specified, all in vitro cell
assays were performed using corisin diluted in a solution containing
recombinant human albumin, unless otherwise specified. While the serum
albumin concentration in human plasma is typically 35-50 g/L ( ~3.5–5%
w/v), we used 0.5% HSA as a physiologically relevant lower-bound
concentration to approximate albumin levels in the renal tubular
microenvironment, particularly under pathological conditions such as
diabetic nephropathy where proteinuria leads to albumin leakage into
the tubular lumen. In our experimental setup, the apical side of the
tubular epithelial cells faced the medium, simulating luminal (urinary)
exposure, where corisin could be present in proteinuric states.
Apoptosis assay
To assess apoptosis, cells (4 × 105 cells/well) were cultured in
12-well plates until they reached subconfluency. Subsequently, they
were serum-starved for 12 h and then exposed to stimulants for 48 h.
Apoptosis was quantified by flow cytometry (FACScan, BD Biosciences,
Oxford, UK) using fluorescein-labeled annexin V and propidium iodide
(FITC Annexin V Apoptosis Detection Kit with PI, Biolegend, San Diego,
CA). The gating strategy is shown in Supplementary Fig. [303]26.
Apoptosis was also evaluated using the terminal deoxynucleotidyl
transferase dUTP nick end labeling (TUNEL) assay, which was conducted
at MorphoTechnology Corporation in Sapporo, Hokkaido, Japan, following
standard protocols.
Flow cytometric analysis of cell cycle
Cells were harvested using trypsin/EDTA and washed once with
phosphate-buffered saline (PBS). Following centrifugation, the cell
pellet was vortexed thoroughly. To fix the cells, cold 70% ethanol was
added dropwise to the pellet while vortexing to prevent clumping. The
cells were then incubated in ethanol at 4 °C for 60 min. After
fixation, the cells were washed twice with PBS by centrifugation at 850
× g, and the supernatant was discarded. The resulting pellet was
resuspended, and cells were treated with ribonuclease by adding 50 μl
of RNase A (from a 100 μg/ml stock solution). Subsequently, 200 μl of
propidium iodide (PI) was added from a 50 μg/ml stock solution for DNA
staining. For flow cytometric analysis, forward scatter (FSC) and side
scatter (SSC) were measured to identify single-cell populations. To
exclude cell doublets, pulse processing was performed by analyzing
PI-Area versus PI-Width.
Cell proliferation assay
Cells were seeded at a density of 1 × 10⁴ cells per well in 100 μL of
culture medium in a 96-well plate. The cells were then stimulated with
0, 20, or 40 μg/mL of corisin, with or without the addition of
200 μg/mL of anti-corisin mAb (21 A), and incubated for 24 h at 37 °C
in a humidified atmosphere containing 5% CO₂. Following stimulation,
20 μL of CellTiter 96® AQueous One Solution Reagent (Promega
Corporation, Madison, MI, USA) was added to each well, and the plate
was incubated for an additional 2 h at 37 °C. Cell proliferation was
assessed by measuring absorbance at 490 nm using a 96-well plate
reader.
Evaluation of nuclear morphology
Primary human podocytes and primary normal human renal tubular
epithelial cells were cultured for 48 h in the presence of corisin at
concentrations of 20 µg/mL and 40 µg/mL or an equivalent concentration
of scrambled peptide as a control. Following incubation, the cells were
fixed and stained with 4’,6-diamidino-2-phenylindole (DAPI) to
visualize nuclear morphology. Nuclear parameters, including nuclear
area, perimeter, and circularity, were quantitatively assessed using
ImageJ, an open-source image analysis software. Automated image
segmentation and thresholding were applied to ensure consistent
measurement of nuclear features.
Evaluation of corisin targeting mitochondria in human renal cells
Human renal tubular Caki-2 cell lines and primary human podocytes were
cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum
(FBS) at a seeding density of 1 × 104 cells per well. The cells were
grown on collagen I-coated 4-well culture slides (Corning 354557,
Biocoat) and incubated at 37 °C in a humidified atmosphere with 5%
CO[2]overnight. The following day, the culture medium was replaced with
RPMI-1640 containing 1% FBS, and incubation continued at 37 °C
overnight. Subsequently, the medium was replaced with a solution
containing 20 µg/mL of fluorescein isothiocyanate (FITC)-labeled
corisin in 0.5% rhAlb or an equivalent dose of FITC-labeled scrambled
peptide, and the cells were incubated for 4 h at 37 °C. After
incubation, the medium was aspirated, and the cells were washed thrice
with phosphate-buffered saline (PBS) devoid of calcium and magnesium to
remove residual dye or peptide. Cells were then treated with a 20 nM
MitoTracker staining solution to label mitochondria for 30 min at
37 °C, followed by fixation in 4% paraformaldehyde in PBS (without
calcium and magnesium) for 10 min at room temperature, and subsequently
washed three times with PBS. The nuclei were stained with 1 µg/mL DAPI
(4’,6-diamidino-2-phenylindole) in PBS for 5 min at room temperature,
followed by additional PBS washes. Visualization was performed using a
fluorescence microscope (Olympus Corporation, Tokyo, Japan), and images
were captured and analyzed using image processing software.
Evaluation of corisin penetration into renal tubular cells in the absence of
albumin
RPTECs were seeded at a density of 2 × 10⁴ cells per well in a 24-well
culture plate and allowed to adhere. The culture medium was then
replaced with RPMI, either without albumin or supplemented with 0.5%
rhAlb. FITC-labeled corisin was added at a final concentration of
40 μg/mL in either saline or 0.5% rhAlb, and the cells were incubated
for 4 h at 37 °C. After incubation, the medium was removed, and the
cells were stained with 50 nM Mitotracker Red CMXRos, prepared from a
1 mM stock solution in DMSO, followed by a 30-minute incubation at
37 °C. The cells were then washed with PBS (-Ca, -Mg) and fixed with 4%
paraformaldehyde (PFA) in PBS (-Ca, -Mg) for 10 min at room
temperature. After fixation, the cells were washed twice with PBS (-Ca,
-Mg)) and stained with 1 μg/mL DAPI in PBS (-Ca, -Mg) to visualize
nuclei, followed by two additional PBS washes. Finally, fluorescence
microscopy analysis was performed to assess corisin penetration and
mitochondrial localization.
Western blotting
The cells were washed twice with cold phosphate-buffered saline and
subsequently lysed in RIPA buffer. This buffer consisted of 10 mM
Tris-Cl (pH 8.0), 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate,
0.1% SDS, 140 mM NaCl, and 1 mM phenylmethylsulfonyl fluoride.
Protease/phosphatase inhibitors, including 1 mM orthovanadate, 50 mM
β-glycerophosphate, 10 mM sodium pyrophosphate, 5 μg/mL leupeptin,
2 μg/mL aprotinin, and 5 mM sodium fluoride, were also added to the
buffer. The resulting mixture was centrifuged at 17,000 x g for 10 min
at 4 °C, and the protein concentration was determined using the Pierce
BCA protein assay kit from Thermo Fisher Scientific Incorporation
(Waltham, MA). Samples with equal protein amounts were mixed with
Laemmli sample buffer and then separated using SDS-PAGE. For Western
blotting, nitrocellulose membranes and anti-corisin mAb were used. The
intensity of the bands on the blots was quantified with the NIH
ImageJ1.54 m program developed by Wayne Rasband at the NIH Research
Service Branch (wayne@codon.nih.gov).
Molecular Dynamics simulation of the conformational ensemble of corisin
Molecular Dynamics simulation (MDS) set-up
Two separate simulation systems were built to observe the
conformational ensemble of peptide: the unfolded/extended structure of
the peptide, which was constructed in PyMol 2.1
([304]https://pymol.org/2/), and the folded structure of the peptide,
which was built by SWISS-MODEL ([305]https://swissmodel.expasy.org/).
The N-terminal of the peptide was capped using the neutral acetyl group
(ACE), and the C-terminal was capped using a neutral methylamine group
(NME). The peptide was solvated in an orthogonal TIP3P water box, and
the systems were neutralized by 150 mM NaCl using Packmol
v18.169^[306]101.
All MDS were performed using the Amber18 software package employing
Amber ff14SB forcefield. Each MDS system was first minimized for 50000
cycles. The minimization method switched from the steepest descent for
the first 5000 cycles to a conjugate gradient for the remaining 45000
cycles. The systems were heated from 0 K to 300 K under the NVT
ensemble. The heating step was conducted for 3 ns using a Langevin
thermostat with a collision frequency of 2 ps-1. The systems were
equilibrated in an NPT ensemble for 2 ns with a pressure of 1 bar using
a Monte Carlo barostat. The systems were further equilibrated in an NPT
ensemble (300 K and 1 bar) for 50 ns and underwent production
runs^[307]102. The SHAKE algorithm was used to constrain
hydrogen-containing bonds^[308]103. All systems underwent hydrogen mass
repartitioning (HMR)^[309]104, which redistributes mass between
hydrogen atoms and their covalently bonded heavy atoms within the
peptide. This adjustment permits an increase in the simulation timestep
to 4 fs.
For two systems, ~20 μs simulation was obtained by applying the
adaptive sampling method^[310]105, that was used to efficiently sample
the conformational space of the peptide. In this study, the least
count-based adaptive sampling was used to find the new conformational
states quickly. Adaptive sampling was performed as follows: (1) Run a
series of short MDS from a collection of starting structures. (2)
Cluster all collected simulation data using the K-means algorithm. The
distances of all pairs of residues separated by two or more residues
were used as the reaction coordinates for generating 100 clusters. (3)
Randomly pick one state from each of 25 clusters with the least
population as seeds to start a new simulation. (4) Repeat steps 1-3
until the sampling reaches convergence.
In adaptive sampling, a bias was introduced as simulations initiated
from the least populated states after each round, potentially resulting
in a population distribution that deviates from the real equilibrium
population of states. To eliminate this sampling bias, a Markov State
Model (MSM) was constructed to interconnect all clusters by estimating
the transition probability matrix across all conformational
states^[311]106–[312]108.
Markov State Model (MSM) Construction
MSM was built by the python package PyEMMA 2.5.6^[313]109. The
distances of all pairs of residues separated by two or more residues
were used for featurizing the simulation data, and 136 residue-residue
distances were included in the feature matrix. The optimal number of
microstates for MSM was selected by maximizing VAMP score^[314]110. The
lag time of 20 ns was chosen from the implied time scale plot and
Chapman-Kolmogorov test^[315]111 was performed to validate the
Markovian behavior of the MSM.
Transition path theory (TPT)
TPT was applied to calculate the transition timescale between different
conformational states^[316]112. The transition time from state A to
state B was estimated from mean free passage time (MFPT), which is the
inverse of the rate of the reaction (1/
[MATH:
kAB
:MATH]
) calculate by the following equation (1):
[MATH:
1/kAB<
/mi>=τ∑i=1
mπi(1−qi+<
/msubsup>)/F :MATH]
where τ is lag time of MSM, π_i is the stationary probability of
microstate i, F is the total flux between state A and state B, q_i^+ is
the forward committor probability that when the system is at microstate
i, it will reach the state B instead of state A. TPT was performed by
the python package PyEMMA 2.5.6^[317]109.
Trajectory analysis
CPPTRAJ module in AMBER (Assisted Model Building with Energy
Refinement) 22^[318]42 and the python package MDTraj^[319]113 were used
to analyze trajectory data, and VMD 1.9.3^[320]114 and PyMol 2.1
([321]https://pymol.org/2/) were used to visualize MDS snapshots. The
Python package Matplotlib^[322]115 was used to generate the 2-D plot of
the free energy landscape.
Definition of Fraction of Native Contacts
Native contacts are interactions between residues that are in contact
within the native structure^[323]116. The fraction of native contacts,
denoted as Q, is calculated by the following equation (2):
[MATH: QX=1<
/mn>S∑(i,
j)∈S11
+exp[β(rij<
mfenced close=")"
open="(">X−λ
rij0)] :MATH]
where X is a conformation, r[ij](X) is the distance between atom i and
j in conformation
[MATH: X :MATH]
,
[MATH:
rij
0 :MATH]
is the distance between heavy atoms i and j in the native conformation,
S is the set of all pairs of heavy atoms (i, j) belonging to the
residue θ^i and θ^j such that |θ^i – θ^j| >3 and
[MATH:
rij
0 :MATH]
< 4.5 Å, β = 5 Å^−1 is a smoothing parameter, λ = 1.8 for the all-atom
model.
Molecular dynamics simulation of corisin interaction with human and bovine
albumin
Molecular dynamics simulation set-up
We prepared two systems for MDS: (i) HSA-corisin (Human Serum
Albumin-corisin) and (ii) BSA-corisin (Bovine Serum Albumin-corisin).
The complex structures of HSA-corisin and BSA-corisin were predicted by
AlphaFold-Multimer^[324]40. The systems were solvated in a TIP3P water
box and neutralized by 150 mM NaCl. Amber ff14SB force field was used
for protein parameterization. All systems underwent minimization and
equilibration first by using Amber 18 software^[325]117. Each MDS
system was minimized for 5000 cycles with the steepest descent
algorithm and then further minimized for 45000 cycles with the
conjugate gradient algorithm. The temperature of minimized systems was
increased from 0 K to 300 K under the NVT (Number of particles, Volume,
Temperature) ensemble. Heating steps were performed for 3 ns by using a
Langevin thermostat with a collision frequency of 2 ps-1. The heated
systems were pressurized to a constant pressure of 1 bar under the NPT
ensemble. The pressure of the systems was controlled by using a Monte
Carlo barostat. During heating and pressurized steps, the backbone
atoms of proteins were restrained with a weight of 5 kcal/mol/Å2. Then,
restraints were removed, and systems were further equilibrated for
50 ns in an NPT ensemble (300 K and 1 bar). The production runs were
performed using OpenMM^[326]41. Shake algorithm was used to constrain
the hydrogen-containing bonds^[327]103. Hydrogen mass repartitioning
(HMR) was applied to all systems, which redistributes mass between
hydrogen atoms and heavy atoms of protein/peptide to allow the time
step of the simulation to be increased to 4 fs^[328]104.
Trajectory analysis
CPPTRAJ in Amber 18 was used to process MDS trajectories^[329]42.
Feature calculation and data analysis were performed by the Python
package MDTraj^[330]113. VMD 1.9.3^[331]114 was used to visualize the
MDS trajectories.
Linear Interaction analysis
Linear Interaction Energy (LIE)^[332]118 function in CPPTRAJ was used
to estimate non-bonded interactions between all-atoms in the ligand and
all-atoms in the surroundings. 10 Å was used as a cutoff to determine
the surrounding atoms of ligands. The binding free energy was
calculated as a linear combination of electrostatics and van der Waals
interactions for all-atom pairs based on the following formula (3)
(default values:
[MATH: α=0.16,β=0.5,γ=0.0 :MATH]
):
[MATH: ΔELIE=αEboundvdW<
/mi>−Efreevd
W+βEboundele<
/mi>−Efreeel
e+γ :MATH]
Induction of exacerbation of kidney fibrosis in TGFβ1 transgenic mice
This experiment was conducted to investigate whether the systemic
administration of corisin exacerbates kidney fibrosis. Female TGFβ1
transgenic (TG) mice, aged 9 weeks, were randomly assigned to two
treatment groups. A group received intraperitoneal injections of
synthetic corisin (n = 5) at a dose of 5 mg/kg body weight,
administered three times per week for two weeks. Another group (n = 6)
was administered a synthetic scrambled peptide at the exact dosage and
via the same route for two weeks. In a previous study, we demonstrated
that administering corisin at a dose of 5 mg/kg every two days for two
weeks exacerbated the progression of lung fibrosis. Based on these
findings, we adopted a similar dosing regimen in the current
experiment, administering six doses over two weeks to ensure sufficient
and sustained exposure to corisin while assessing its effects on the
worsening of kidney fibrosis. Plasma and urine samples were collected
on Days 1, 8, and 15 from both groups to assess kidney function. On Day
16, euthanasia was performed, followed by the final collection of
blood, urine, and kidney tissues.
Induction of DM in TGFβ1 TG mice with kidney fibrosis
DM was induced in WT and TGFβ1 TG mice through intraperitoneal
administration of streptozotocin (STZ) (Sigma, St. Louis, MO) at a dose
of 40 mg/kg body weight for five consecutive days. Mice that received
an equivalent volume of sterile physiological saline (SAL) via
intraperitoneal injection served as controls. Blood glucose levels were
assessed in the fourth week post-STZ administration. DM was confirmed
through a glucose tolerance test. Mice with blood glucose levels
exceeding 200 mg/dL were classified as diabetic. Diabetic and
non-diabetic TGFβ1 TG mice were euthanized at week nine following STZ
administration, and blood, urine, and kidney tissue samples were
collected for further analysis.
Induction of diabetic nephropathy in WT mice
To expedite the onset of DM-associated nephropathy, a group of
wild-type mice underwent unilateral nephrectomy followed by STZ
administration. The procedure was performed under deep anesthesia in
sterile conditions. A surgical incision was made along the right
dorso-lumbar region, allowing visualization of the ureter and renal
artery. Both structures were carefully ligated, and the kidney was
excised at the proximal region near the renal hilum. The incision was
then sutured, and the mice were allowed to recover for four weeks.
After complete recovery, the mice received intraperitoneal injections
of STZ at a dose of 40 mg/kg body weight for five consecutive days to
induce DM. Control mice were administered an equivalent volume of
sterile physiological saline under identical conditions. Blood samples
were collected in the fourth-week post-STZ injection to measure glucose
levels. Mice with blood glucose levels exceeding 200 mg/dL were
classified as diabetic and included in the study cohort. Four weeks
after DM diagnosis, the mice were humanely sacrificed under deep
anesthesia, and blood and kidney tissue samples were collected for
further analysis. To assess the degree of fibrosis in each mouse group,
Masson’s trichrome staining was performed on formalin-fixed,
paraffin-embedded tissue sections, and the extent of fibrosis was
quantified using WinROOF imaging software.
Evaluation of the therapeutic effect of a neutralizing anticorisin mAb in
diabetic TGFβ1 TG mice with kidney fibrosis
This experiment aimed to investigate the effects of anticorisin
treatment in diabetic TGFβ1 TG mice with kidney fibrosis. TGFβ1 TG mice
typically develop renal dysfunction as early as 8 weeks of age,
compared to their WT counterparts^[333]31. However, this renal
dysfunction remains relatively stable for several weeks before
progressing to a fatal stage. To minimize early mortality during the
experimental period, human TGFβ1 TG mice with relatively mild renal
dysfunction were selected for DM induction, enabling the observation of
disease exacerbation while reducing the risk of premature death. DM was
induced in the TGFβ1 TG mice by intraperitoneal injections of
streptozotocin (STZ) (Sigma, St. Louis, MO). STZ was administered at a
dosage of 40 mg/kg body weight for five consecutive days. A negative
control group of WT mice was administered an equivalent volume of
saline (SAL) intraperitoneally. DM induction was confirmed by a
non-fasting blood glucose level of 200 mg/dL or higher and by an
intraperitoneal glucose tolerance test. Following DM induction, the
mice were assigned to three groups: (1) TGFβ1 TG mice treated with
anticorisin mAb (n = 5), (2) TGFβ1 TG mice treated with control
immunoglobulin G (IgG) (n = 5), and (3) wild-type (WT) mice as a
negative control (n = 4). The treatment phase began in the fifth week,
where the mice received intraperitoneal injections of either
anticorisin mAtb or control IgG at a dose of 10 mg/kg three times per
week for eight weeks. The reported half-life of the anti-corisin mAb is
approximately three days. To maintain a steady antibody level, minimize
significant troughs in its concentration, and evaluate its efficacy, we
administered three doses per week for eight weeks^[334]29. At the end
of the 13-week experimental period, the mice were sacrificed, and
samples of blood, urine, and kidneys were collected for further
analysis to evaluate the impact of anticorisin treatment on kidney
fibrosis.
Specimen collection and handling procedures
The animals were euthanized by intraperitoneal injection of an overdose
of 5% isoflurane. Following euthanasia, samples were collected for
biochemical analysis and histological staining. Blood was obtained via
closed-chest cardiac puncture and collected into tubes containing
10 U/mL heparin as an anticoagulant. Urine samples were gathered using
metabolic cages. After systemic circulation was flushed with
physiological saline, each kidney was dissected, isolated, and weighed.
The left kidney was stored at -80 °C for subsequent analysis, while the
right kidney was fixed in 10% paraformaldehyde, dehydrated, embedded in
paraffin, and sectioned into 3-μm thick slices for staining with
hematoxylin and eosin (H&E), periodic acid-Schiff (PAS), or Masson’s
trichrome. Images of the resected kidneys were captured using an
Olympus BX53 microscope equipped with a digital camera (Olympus DP73,
Tokyo, Japan). Glomerular sclerosis was evaluated based on PAS
staining. For each mouse, 25 glomeruli were randomly selected and
scored as follows: 0 for normal glomeruli, 1 for mild mesangial
thickening (PAS-positive area <25%), 2 for moderate segmental sclerosis
(PAS-positive area 25–50%), 3 for severe segmental sclerosis
(PAS-positive area 50–75%), and 4 for global sclerosis (PAS-positive
area ≥75%). The mean score from the 25 glomeruli was defined as the
glomerular sclerosis score. Six investigators, blinded to the treatment
groups, performed the scoring. Renal fibrosis was assessed using
Masson’s trichrome staining. Ten images of the kidney cortex per mouse
were randomly acquired, and the ratio of Masson’s trichrome-positive
area to the total kidney cortex area was calculated using WinROOF image
processing software (Mitani Corp., Fukui, Japan).
Evaluation of diabetes status and kidney functional parameters
Non-fasting blood glucose levels were periodically monitored to assess
the diabetic state. The glucose tolerance test was conducted by
intraperitoneal injection of glucose at a dose of 1 g/kg following
overnight fasting, with blood glucose levels measured at 0, 15, 30, 60,
and 120 min. Blood glucose levels were determined using the glucose
oxidase method with a glucose assay kit (Dojindo, Mashiki, Japan).
Blood albumin levels were measured using the A/G B test (Wako, Osaka,
Japan), and urinary albumin levels were assessed using the Mouse Urine
ELISA Kit (Exocel, San Diego, USA). Blood creatinine levels were
measured using the Mouse Creatinine Kit (Crystal Chem, Elk Grove
Village, USA), urinary creatinine levels were determined using the
Creatinine Companion Kit (Exocel, San Diego, USA), and blood urea
nitrogen levels were measured using the Urea Nitrogen Detection Kit
(Arbor Assays, Ann Arbor, USA).
Immunohistochemical staining
Staining for corisin was performed using an in-house anti-corisin
monoclonal antibody (clone A21; dilution 1:300), and commercially
available antibodies against p21 (cat. no. sc-6246; dilution 1:50;
Santa Cruz Biotechnology, Dallas, TX, USA), F4/80 (cat. no. 28463-1-AP;
dilution 1:1200; Proteintech, Rosemont, IL, USA), and p53 (cat. no.
ab131442; dilution 1:400; Abcam, Cambridge, MA, USA). This procedure
was conducted at MorphoTechnology Corporation in Sapporo, Hokkaido,
Japan, following standard protocols. Evaluation of SAβGal activity was
performed using X-Gal (Sigma-Aldrich) and a commercial kit (Cellular
Senescence Detection Kit Spider BGAL, Dojindo, Osaka, Japan). Olympus
BX50 microscope with an OlympusDP70 digital camera (Olympus
Corporation, Tokyo, Japan) was used for data collection. WinROOF2018
software (Mitani Corporation, Tokyo, Japan) was used for data
collection, and the public domain NIH ImageJ1.54 m program was used for
image analysis.
Biochemical analysis
The concentrations of interleukin-1β (IL-1β), transforming growth
factor-β1 (TGFβ1) (R&D Systems, Minneapolis, MN), tumor necrosis
factor-α (TNFα), monocyte chemoattractant protein-1 (MCP-1) (BD
Bioscience, BD OptEIA kits, San Diego, CA), connective tissue growth
factor (CTGF) (Abcepta, San Diego, CA), thrombin-antithrombin (TAT;
Affinity Biologicals, Ontario, Canada), and D-dimer (Bioss Antibodies,
Woburn, MA) were quantified using commercially available enzyme-linked
immunoassays (ELISAs) following the manufacturer’s instructions.
Collagen type I levels were determined using an ELISA that employed an
anti-collagen type I antibody and a biotin-conjugated anti-collagen
type I antibody from Rockland Immunochemicals Inc. (Limerick, PA).
Creatinine levels were measured using an enzymatic method, while blood
urea nitrogen was assessed via a colorimetric method (NCal™
NIST-Calibrated Kit; Arbor Assays, Ann Arbor, MI) following the
manufacturer’s instructions. Liver-type fatty acid-binding protein
(L-FABP) and kidney injury molecule 1 (KIM-1) were quantified using
commercial enzyme immunoassay kits (R&D Systems). Corisin levels were
measured using an in-house ELISA as previously described^[335]29.
Briefly, a polyclonal anti-transglycosylase 351 antibody was used to
coat a 96-well plate at 2 µg/ml in phosphate-buffered saline and
incubated overnight at 4 °C. Following blocking with 1% bovine serum
albumin in phosphate-buffered saline and thorough washing with
phosphate-buffered saline containing Tween, the wells were incubated
with varying concentrations of corisin standards and plasma samples at
4 °C overnight. Subsequent washing steps were followed by the addition
of horseradish peroxidase-conjugated streptavidin (R&D Systems). After
further washing and incubation, a substrate solution was applied for
color development, and absorbance was measured at 450 nm using a
BIOD-RAD iMark™ microplate reader. Corisin concentrations were
calculated from a standard curve, with both inter- and intra-assay
variability maintained below 10%.
Determination of albumin-bound corisin and free corisin in serum from DM
patients
To assess the presence of free corisin in the serum of patients with
diabetes mellitus (DM), 100 µL of tenfold diluted serum was applied to
a Nanosep 10 K Omega ultrafiltration device (Pall Corporation, Port
Washington, NY, USA). The sample was then centrifuged according to the
manufacturer’s instructions to separate the serum into two fractions: a
flow-through fraction (<10 kDa) and a retained fraction (>10 kDa).
Corisin levels were subsequently quantified using enzyme-linked
immunosorbent assay (ELISA), as described above, in the >10 kDa
retained fraction, <10 kDa flowthrough fraction, and unfractionated
serum, the latter representing the total corisin concentration prior to
filtration.
Amplification of DNA fragments encoding corisin/corisin-like peptides in
healthy subjects and diabetic CKD patients
A large alignment of the polypeptide sequences of the IsaA-like
transglycosylases from diverse bacteria was created to develop PCR
primers that specifically amplify the DNA encoding the proapoptotic
peptides in each urine sample. Two sequences flanking the proapoptotic
peptides (SVKAQF and WGTGSV) are highly conserved, and the conserved
codon usage in the genes allowed the design of two primers, Corisin-F
5’ATCAGTTAAAGCTCAATTC and Corisin-R 5’GCTACTGAACCAGTACCCCATG, as the
forward and reverse primers, respectively. The primer pair amplifies
~150 bp DNA fragments containing the coding sequences of the
proapoptotic peptides. The primers were validated by extracting
community genomic DNA from the urine of healthy controls (n = 4) and
patients with diabetic CKD (n = 18). The right size fragment was
amplified from all extracted genomic DNA, except for the negative
control (same PCR mixture, without DNA), and the corisin/corisin-like
peptides in the urine samples were obtained by using the DNA sequencing
and bioinformatics analyses described below.
Shotgun DNA library construction and sequencing
In this process, each PCR product was resolved in agarose gel and
purified (QIAquick PCR purification kit, Qiagen). Importantly, each PCR
product was considered a library of the kidneys and urinary tract of a
particular CKD patient or healthy control. The DNA was next used in
preparing unique libraries by using the NEBNext DNA Library Prep Master
Mix Set for Illumina with Unique Dual Indexes to prevent index
switching. Library preparations was done using an EpMotion 5075 liquid
handler (Eppendorf). The individually barcoded libraries were
amplified, quantitated with Qubit and resolved on a Fragment Analyzer
to confirm the absence of free primers and primer dimers and the
presence of DNA of the expected size. Libraries were pooled in
equimolar concentration and further quantitated by qPCR on a BioRad CFX
Connect Real-Time System (Bio-Rad Lab Inc., CA). The pooled barcoded
shotgun libraries were loaded on a MiSeq flowcell for sequencing
(Illumina).
Bioinformatic analyses
Sequences were processed using the TADA Nextflow-based
workflow^[336]119,[337]120, which implements protocols for denoising
and retaining single-nucleotide resolution amplicon sequence variants
(ASVs)^[338]121 from known and custom amplicon sequences. Resulting
ASVs were tabulated and quantified per sample, assessed for potential
chimeric sequences, and finally compared with known reference sequences
using BLASTN^[339]122. Multiple sequence alignment was performed using
DECIPHER^[340]123, with final protein translations performed in R using
the Bioconductor Biostrings library^[341]124.
Gene expression analysis
RNA was isolated from kidney cells or tissue using the Sepasol RNA-I
Super G reagent from Nacalai Tesque Inc., Kyoto, Japan. Complementary
DNA (cDNA) was synthesized from 2 μg of total RNA with an oligo-dT
primer and ReverTra Ace Reverse Transcriptase (Toyobo Life Science
Department, Osaka, Japan). Standard PCR was then conducted using
primers listed in Supplementary Table [342]3. Depending on the target
gene, PCR was performed for 26 to 35 cycles, with denaturation at 94 °C
for 30 s, annealing at 65 °C for 30 s, and elongation at 72 °C for
1 min, followed by a final extension at 72 °C for 5 min. mRNA
expression levels were normalized to glyceraldehyde 3-phosphate
dehydrogenase (GAPDH) mRNA expression.
Single-cell RNA sequencing analysis
Cell culture and processing
Human Renal Proximal Tubule Epithelial Cells (RPTEC) were obtained from
LONZA (Houston, TX) and cultured in Renal Epithelial Cell Growth Medium
(LONZA) under standard conditions (37 °C, 5% CO₂). At 70% confluency,
the culture medium was replaced with fresh medium containing either
20 μg/mL corisin (prepared as a stock solution at 2 mg/mL in 0.5%
recombinant human albumin) or 20 μg/mL scrambled peptide in 0.5%
recombinant human albumin. Cells were incubated under these conditions
for 24 h. After treatment, cells were harvested using trypsinization,
washed in cold PBS, pelleted by centrifugation (300 g, 5 min, 4 °C),
and cryopreserved in freezing media containing BAMBANKER (GC LYMPHOTEC
Inc., Tokyo, Japan). Frozen cells were stored in liquid nitrogen until
further processing. Single-cell RNA sequencing was conducted by Takara
Bio Inc. (Shiga, Japan) ([343]https://www.takarabio.com/). Cell
viability was assessed prior to library preparation using a Countess II
Automated Cell Counter (Thermo Fisher Scientific), ensuring >80%
viability. Approximately 10,000 cells per sample were loaded into the
Chromium Controller (10x Genomics) for droplet encapsulation. The cDNA
library was prepared using the Chromium Next GEM Single Cell Fixed RNA
Sample Preparation Kit, Chromium Next GEM Chip Q Single Cell Kit, and
Dual Index Kit TS Set A (10x Genomics) following the manufacturer’s
protocols. Sequencing was performed using the Novaseq X Plus system
(Illumina) to achieve a target sequencing depth of 50,000–100,000 reads
per cell.
Data preprocessing and quality control
Raw sequencing data were demultiplexed, aligned, and quantified into
UMI-filtered counts using Cell Ranger (v.4.0.0; 10x Genomics) against
the hg38 reference genome (refdata-gcs-GRCh38-2020-A). Quality control
was performed to exclude low-quality cells and artifacts: cells with
fewer than 200 detected genes, fewer than 500 UMIs, or greater than 10%
mitochondrial gene expression were excluded. After filtering, 9327
cells from corisin-treated samples and 8,186 cells from scrambled
peptide-treated samples were retained for downstream analysis.
Data normalization, clustering, and visualization
Normalized gene expression data were generated using the LogNormalize
method in the Seurat R package (v.4.0.0), scaling each gene’s
expression by total UMI counts and multiplying by a scale factor of
10,000. Data were further scaled using the ScaleData function, and
highly variable genes were identified for dimensionality reduction.
Clustering was performed using the Louvain algorithm implemented in the
FindClusters function with a resolution parameter set to 0.3. Uniform
Manifold Approximation and Projection (UMAP) was used for
dimensionality reduction and visualization with the Seurat RunUMAP
function using default parameters.
Trajectory and Differential Gene Expression Analysis
Trajectory analysis was performed using Monocle2 (v.2.20.0), with input
gene expression matrices processed through size factor and dispersion
estimation steps. Cells were ordered along pseudotime based on the most
variable genes, as determined by the “dpFeature” method in Monocle2.
Differentially expressed genes (DEGs) were identified using the Seurat
FindMarkers function with a Wilcoxon rank-sum test. DEGs were defined
by a fold change >2 and P-value < 0.05.
Pathway and Network Analysis
Pathway enrichment analysis was conducted using Enrichr
([344]https://maayanlab.cloud/Enrichr/) with the KEGG and Reactome
libraries. Pathways with false discovery rates (FDR) <0.05 were
considered significant. Highly upregulated genes in corisin-treated
cells were analyzed for protein-protein interactions (PPI) using the
STRING database (v.11.5; [345]https://string-db.org/) with a minimum
interaction confidence score of 0.7. The resulting PPI networks were
visualized and analyzed using Cytoscape (v.3.9.1).
Statistical analysis
Data with a normal distribution are presented as mean ± standard
deviation (SD) or standard error of the mean (SEM), while skewed data
are expressed as median (interquartile range). Data distribution was
assessed using the Shapiro-Wilk or Kolmogorov-Smirnov test. Statistical
differences between two normally distributed variables were evaluated
using a two-sided unpaired Student’s t-test, whereas differences among
three or more normally distributed variables were analyzed using
one-way analysis of variance (ANOVA) followed by the Newman-Keuls post
hoc test. For skewed data, the two-sided Mann-Whitney U test was
applied for comparisons between two groups, while the Kruskal-Wallis
ANOVA followed by Dunn’s post hoc test was used for multiple-group
comparisons. The associations between eGFR and other variables were
evaluated using univariate and multivariate linear regression analyses.
Variables with p < 0.05 in the univariate analysis were included in the
multivariate model. The non-parametric Spearman’s rank correlation
coefficient was used to assess correlations between corisin and other
variables. Corrections for multiple comparisons were applied where
appropriate. A P-value of less than 0.05 was considered statistically
significant. Statistical analyses were conducted using GraphPad Prism
version 10.5 (GraphPad Software, Inc., San Diego, CA).
Reporting summary
Further information on research design is available in the [346]Nature
Portfolio Reporting Summary linked to this article.
Supplementary information
[347]Supplementary Information^ (38.4MB, pdf)
[348]41467_2025_61847_MOESM2_ESM.docx^ (15.7KB, docx)
Description of Additional Supplementary Files
[349]Supplementary dataset 1^ (24.9KB, xlsx)
[350]Supplementary dataset 2^ (26.7KB, xlsx)
[351]Reporting Summary^ (123KB, pdf)
[352]Transparent Peer Review file^ (990.5KB, pdf)
Source data
[353]Source Data^ (7.8MB, xlsx)
Acknowledgements