Kidney Transplantation–Induced Reduction in Myocardial Fibrosis and Modulation of Plasma Proteome: Results of a Pilot Study

Abstract

Rationale & Objective

Kidney transplantation (KT) improves cardiovascular outcomes in patients with kidney failure, yet the biological mechanisms underlying post-KT myocardial recovery remain unclear.

Study Design

Prospective observational cohort study with matched controls.

Setting & Participants

Fifty participants were enrolled at 2 transplant centers: 28 received a living KT, and a cardiac magnetic resonance imaging (CMRI) scan assessment of interstitial myocardial fibrosis with T1 maps before and nine months post-KT; 22 matched waitlisted controls had baseline and 9-month CMRI scans. The groups were matched using baseline characteristics and T1 values.

Exposure

Receipt of kidney transplantation.

Outcome

Change in native T1 mapping CMRI scan, a surrogate marker of interstitial myocardial fibrosis.

Analytical Approach

Regression models were used to estimate the effect of KT versus no KT (the waitlisted control group) on interstitial myocardial fibrosis, adjusted for demographic characteristics, comorbidities, blood pressure, and medications. Plasma collected at baseline was analyzed using SomaScan v.4.1 (6,472 unique proteins). Regression models were used to identify proteins associated with baseline and longitudinal changes in T1, adjusted for clinical covariates and corrected for multiple testing.

Results

KT recipients had reduced interstitial myocardial fibrosis compared with waitlisted controls (−47.49 ± 81.63 vs +13.77 ± 62.13 ms, P = 0.01). Forty-three plasma proteins were significantly associated with baseline T1 values; macrophage colony-stimulating factor displayed the greatest retest reproducibility (80%) and positive association with T1 (P = 3.1 × 10^-4). Pathway enrichment analyses identified 17 interconnected fibrosis-related proteins with biological relevance to cardiac remodeling.

Limitations

A modest sample size and a single time-point proteomic assessment. External validation is needed.

Conclusions

KT leads to regression of interstitial myocardial fibrosis within 9 months, supporting a structural cardiac benefit of transplantation. Pretransplant proteomic signatures, particularly macrophage colony-stimulating factor, reflect underlying fibrotic burden and may serve as a predictive biomarker to monitor post-KT cardiac recovery.

Plain-language summary

Kidney transplantation results in improved cardiovascular health, but how the 2 are connected is not fully understood. In this study, we followed people with kidney failure who received a kidney transplant and compared them to similar patients who remained on the waitlist. We examined heart imaging to measure changes in diffuse fibrosis (scarring) over 9 months and association with circulating blood proteins. We found that transplants were linked to reduced heart fibrosis, and 1 protein in particular—macrophage colony-stimulating factor—correlated with reduced fibrosis. These findings suggest that kidney transplants may help improve heart tissue structure, and identifying key proteins, such as macrophage colony-stimulating factor, could represent a tool to monitor heart recovery in the future.

Introduction

Cardiovascular diseases (CVD) account for almost half of the deaths in people with kidney failure.¹ Traditional CVD risk factors, derived from the general population, perform poorly in people with kidney failure, suggesting unique pathophysiologic mechanisms. Myocardial remodeling and interstitial myocardial fibrosis are prevalent, progressive, and associated with CVD mortality as chronic kidney disease (CKD) advances to kidney failure.2, 3, 4, 5

Kidney transplantation (KT) reduces CVD death rates by as much as 4-fold.⁶ KT has been associated with improved cardiac function, including echocardiographic measurements of left ventricular (LV) systolic and diastolic function, LV end diastolic dimension, LV mass, and right ventricular systolic pressure.⁷ KT is also associated with a reduction in noncontrast cardiac magnetic resonance imaging (CMRI) scan T1 times, a measure of interstitial myocardial fibrosis.^8,9 However, limitations of the prior studies include the lack of a control group,⁸ or variable times for follow-up CMRI scan.⁹

Though the regression in interstitial myocardial fibrosis may account for lower rates of CVD deaths in KT recipients, the mechanisms are unclear. It is speculated that circulating proteins and metabolites that accumulate in CKD create a uremic milieu that mediates changes in the myocardium.³ Proteomic profiling may offer insights into molecular pathways modulated by KT that contribute to myocardial recovery. These insights could potentially lead to the identification of therapeutic targets for interstitial myocardial fibrosis.

We conducted a prospective observational multicenter study with matched controls to (1) test the hypothesis that KT is associated with improvement in interstitial myocardial fibrosis, as assessed using noncontrast CMRI scans, and (2) to explore plasma proteins associated with baseline and change in myocardial fibrosis post-KT.

Methods

Study Population and Intervention

The Halting Myocardial Fibrosis with Kidney Transplantation (HYFi Kidney) study was conducted between November 2019 and January 2021 at 2 academic centers in Cleveland, OH (University Hospitals Cleveland Medical Center and Cleveland Clinic). Inclusion criteria were adults aged >18 years who met all the requirements to become actively listed for a KT, without infiltrative diseases (eg, amyloid, sarcoidosis, etc.), with no contraindication for CMRI, and with the ability to give informed consent. Participants scheduled for a living KT had concomitant CMRI and blood collection the week before KT and at 9 months post-KT. An age-, sex-, and dialysis vintage-matched 1:1 control group, actively listed for KT but with no living donors, had CMRI and blood collection at baseline and 9 months later (Fig 1). Two participants originally enrolled as controls received a deceased KT after enrollment and had a follow-up CMRI scan and blood collection at 9 months post-KT.

Figure 1.

Study flow diagram. A total of 50 participants were enrolled, and 43 participants completed the study.

Untargeted proteomics assessment using the SomaScanv4.1 assay (SomaLogic Inc) was performed on baseline plasma samples. For participants who were on hemodialysis, plasma was collected immediately after the first dialysis session of the week.

All study procedures were conducted in accordance with the Declaration of Helsinki, and all participants gave written informed consent. The study was approved by the Institutional Review Board at each institution.

Study Measurements

Demographic data were collected at study entry. Clinical, imaging, and laboratory data were obtained at baseline and 9-month visits (Table 1). The estimated glomerular filtration rate (eGFR) was based on the serum creatinine levels.¹⁰ For patients with kidney failure with urine production < 200 mL/d, the eGFR was defined as 0 mL/min/1.73 m².

Table 1.

Baseline Characteristics of HYFi Kidney Study Participants

Characteristic	All Participants (N=48)	Study Group
		Transplant (N=27)	Control (N=21)
Age, y	54 ± 15	53.3 ± 16	55 ± 13
Women	20 (41.7)	12 (44.4)	8 (38.1)
Race
White	37 (77)	22 (81.5)	15 (71)
Black	7 (15)	2 (7.4)	5 (23.8)
Hispanic	3 (6)	3 (11)	0 (0)
Asian	1 (2)	0 (0)	1 (5)
Education level
High school or less, some college, no degree, or other	21 (44)	12 (44)	9 (43)
College or above	27 (56)	15 (56)	12 (57)
Systolic blood pressure, mmHg	134 ± 16	135 ± 16	133 ± 15
Diastolic blood pressure, mmHg	76 ± 8	76 ± 7	75 ± 11
Body mass index, kg/m2	30.7 ± 7	30.9 ± 7	30.5 ± 7
Diabetes	14 (29.2)	5 (18.5)	9 (42.9)
History of coronary artery disease	5 (10.4%)	2 (7.4)	3 (14.3)
CKD stage 4 (eGFR 17 [16-19] mL/min/1.73 m2)	5 (10.4)	2 (7.4)	3 (14.3)
CKD 5 not on dialysis (eGFR 11 [9-14] mL/min/1.73 m2)	21 (43.8)	13 (48.1)	8 (38.1)
CKD 5-dialysis	22 (45.8)	12 (44.4)	10 (47.6)
Hemodialysis	17 (35.4)	8 (29.6)	9 (42.8)
Peritoneal dialysis	5 (10.4)	4 (14.8)	1 (4.8)
Months on dialysis, median [IQR]	28.8 [7.4-42.2]	27.7 [10.8-51.3]	31.4 [3.7-36.6]
Months of transplant listing,	13 ± 11	14 ± 13	12 ± 10
Medications
Statins, n (%)	24 (50)	13 (48)	11 (52)
ACEi/ARBs	24 (50)	15 (55.6)	9 (43)
Laboratory Data
eGFR, mL/min/1.73 m2	10 ± 5	10 ± 5	10 ± 4
Bicarbonate level, mEq/L	24 ± 3.5	24 ± 4	24 ± 3
Hemoglobin level, g/dL	11 ± 2	11 ± 2	12 ± 2
Hemoglobin A1C, %	5.8 ± 1.2	5.7 ± 1.4	5.9 ± 0.96
Serum total cholesterol level, mg/dL	162 ± 44	181 ± 39	137 ± 37
HDL-cholesterol level, mg/dL	41 ± 14	39 ± 10	43 ± 19
LDL-cholesterol level, mg/dL	94 ± 33	106 ± 33	80 ± 27
P3NP level, μg/L∗	14.9 ± 10.6	14.5 ± 9.7	15.5 ± 11.9

Values represent mean ± standard deviation (SD); median [interquartile range (IQR): 25th percentile, 75th percentile], or number (%).

Abbreviations: ACEi, angiotensin-converting enzyme inhibitors; ADMA, asymmetric dimethylarginine; ARB, angiotensin II receptor blocker; eGFR, estimated glomerular filtration rate; HDL, high density lipoprotein; LDL, low density lipoprotein; TMAO, trimethylamine N-oxide; SDMA, symmetric dimethylarginine; P3NP, N-terminal propeptide of procollagen type III.

^∗Normal levels of P3NP are 1.2-4.2 μg/L.

CMRI Scans of Interstitial Myocardial Fibrosis

CMRI scans were performed using a 1.5T Avanto scanner (Siemens Medical Solutions) as previously described.¹¹ Noncontrast (native) myocardial T1 mapping was acquired with a modified Look-Locker inversion recovery sequence in a mid-cavity short-axis slice.^12,13 A detailed description of the standardized noncontrast image acquisition protocol is included in the Supplementary Material. Myocardial T1 is a surrogate measure of interstitial myocardial fibrosis that can be successfully and reliably performed in patients with kidney failure.14, 15, 16 Native T1 values are not influenced by alterations in fluid status, typically observed in patients receiving dialysis.¹⁶ However, to standardize measures with fluid status, a CMRI scan was performed on nondialysis days (18-24 hours after the most recent dialysis treatment). All images were analyzed by investigators blinded to group assignment.

N-terminal Propeptide of Procollagen Type III Assay

N-terminal propeptide of procollagen type III (P3NP) is a collagen-derived serum peptide that is associated with histologically proven myocardial fibrosis.¹⁷ P3NP ELISA kits (Cisbio) were used to measure plasma P3NP in baseline and follow-up stored samples. All samples were run as 1 batch, and the mean interassay coefficient of variation (CV) from replicate controls was 5.5%. Plasma P3NP concentration is expressed as micrograms per liter. A significant correlation was observed between myocardial native T1 values and P3NP levels (Pearson coefficient r = 0.6, P = 0.008) (Fig S1).

Untargeted Proteomic Profiling Using Modified Aptamers

Relative concentrations of 6,472 unique proteins were measured at baseline using the Slow Off-rate Modified Aptamer (SOMAmer) with the SomaScan v.4.1 platform (SomaLogic Inc) and normalized using standard protocols.18, 19, 20, 21 Quality control data, standardization, and formatting for patients receiving hemodialysis were previously described.²² Protein levels were log₂-transformed and standardized in all analyses.

Macrophage Colony-stimulating Factor Assay

Plasma macrophage colony-stimulating factor (M-CSF) was measured at baseline and 9 months using the human M-CSF Quantikine ELISA (DMC00B; R&D Systems) according to the manufacturer’s instructions. All samples were assayed in duplicate, and the mean intra-assay CV score was 5.4%. Plasma M-CSF values are expressed in picograms per milliliter.

Data Analysis

With a target enrollment of 50 participants and an anticipated loss to follow-up of 20% per year, we estimated that the study would have 80% power to detect a mean difference in T1 times of 50 ± 52 ms, at an α error of 0.05, with 36 participants needed to complete the study (18 participants in each group).

Primary analyses for the KT intervention group included combined living and deceased donor KT recipients (N=25). Sensitivity analyses were performed excluding the 2 participants with deceased donor KT (N=23). Participants who remained active on the waiting list and completed the study (N=18) served as controls in all analyses.

Paired t tests were used to compare baseline and 9-month CMRI scan T1 times in each group. The Mann-Whitney U-test was used to compare the 2 groups when the data were not normally distributed. Repeated measures analysis of variance was performed to compare the changes in imaging and laboratory parameters between the groups over time. Results are expressed as mean ± standard deviation, or if nonnormally distributed, as median (25^th and 75th percentile). The Χ² test was used to compare baseline categorical variables between the 2 groups. Multivariable adjusted regression models were used to estimate the effect of KT versus no KT (the waitlisted group) on CMRI scan-derived interstitial myocardial fibrosis. We addressed potential confounding by including adjustments for relevant baseline characteristics, including demographic characteristics (age, sex, and race), body mass index, smoking, diabetes, history of coronary artery disease and heart failure, dialysis status, current renin-angiotensin-aldosterone system inhibitors, systolic blood pressure, and eGFR.

Proteomic Analysis

Linear regression models were run individually for each of the 6,472 unique proteins to identify proteins associated with CMRI scan T1 times. Confounding variables described above were built into the linear regression models after selection of the optimal set of predictors. A Benjamini-Hochberg false discovery rate-adjusted cut-off level of 0.05 was used in all proteomics analyses. DESeq2²³ and Firth’s bias-reduced logistic regression models were used to identify proteins associated with baseline and change in interstitial myocardial fibrosis from baseline to 9 months.

To establish confidence that relevant proteins were implicated in the mechanisms of interstitial myocardial fibrosis, we tested whether protein combinations reside in the same biological pathway using the Kyoto Encyclopedia of Genes and Genomes database.^24,25 To further investigate protein-protein interactions, we utilized k-means clustering on proteins associated with enriched terms within STRING-db v.11.5.²⁶

A significance level of α = 0.05 was used for statistical tests. Analyses were performed using SAS Software version 9.4 of the SAS System for Windows (2023 SAS Institute Inc), and R Statistical Software (version 4.2.2; R Core Team, 2023).

Results

Of the 50 participants enrolled, 2 (4%) were excluded because of poor CMRI scan image quality, resulting in the inclusion of 48 participants with complete data at baseline. Five (10%) were lost to follow-up, resulting in 43 participants who completed the 9-month follow-up visit (Fig 1). There were no differences in baseline characteristics between the excluded participants and the study cohort (Table S1).

The mean age was 54 ± 15 years, 20 (42%) participants were women, and 22 (46%) were receiving dialysis. There were no significant differences in baseline demographic or clinical characteristics between the KT and control groups (Table 1). At follow-up, blood pressure was lower in the transplanted group but not significantly different than the waitlist control group (mean ± SD systolic/diastolic blood pressure 131 ± 10/80 ± 13.8 mmHg vs 139 ± 16/79 ± 10 mmHg, P = 0.06). There were no differences between the groups in the treatment with angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, either at baseline (56% vs 43%, P = 0.21) or at 9 months (16% vs 27%, P = 0.29), in KT versus waitlist participants, respectively. Median (interquartile range) eGFR at 9 months post-KT was 53 (44-58) mL/min/1.73 m².

Using published thresholds for LV mass,²⁷ 5.0% of women and 14.3% of men had LV hypertrophy using CMRI scan images at baseline. There were no baseline differences between CMRI scan images (myocardial native T1) or serologic (P3NP level) markers of myocardial fibrosis between groups (Tables 1 and 2). Other baseline measures of LV structure and function were also not significantly different between groups (Table 2).

Table 2.

Baseline Cardiac Magnetic Resonance Imaging Scan Characteristics of HYFi Kidney Study Participants

Characteristic	All Participants N=48	Transplant N=27	Control N=21	P-value
Primary Outcome Measure of Myocardial Fibrosis
T1a (ms)	1,069.85 ± 58.85	1,076.71 ± 64.13	1,061.04 ± 51.47	0.37
LV Measures
LV mass indexb (g/m2)	62.53 ± 25.43	66.64 ± 30.35	57.24 ± 16.44	0.25
LV end diastolic volume (mL)c	147.21 ± 42.79	150.19 ± 40.52	143.48 ± 46.26	0.59
LV mass/LV end diastolic volume	0.43 ± 0.12	0.44 ± 0.14	0.42 ± 0.10	0.84
LV end systolic volume (mL)d	66.63 ± 32.05	66.93 ± 31.24	66.24 ± 33.84	0.94
Stroke volume (mL)e	80.56 ± 22.98	83.15 ± 21.26	77.24 ± 25.15	0.38
Cardiac output (L/min)f	5.64 ± 1.70	5.55 ± 1.43	5.74 ± 2.03	0.71
LV ejection fraction (%)g	56.17 ± 10.55	56.78 ± 10.82	55.38 ± 10.39	0.65

Values represent mean ± standard deviation. P-values are obtained from an independent t test.

Abbreviations: LV, left ventricular.

Normal values:

^aT1 values (on 1.5T cardiac magnetic resonance imaging scanner): 900-1100 ms.

^bLV mass index: 70 ± 9 g/m² in men and 61 ± 8 g/m² in women.

^cLV end diastolic volume 62-120 mL for males and 58-103 mL for females.

^dLeft ventricle end systolic volume (ESV) [mL]: Women 15-64 mL and ESVI [mL/m²]:10-38 mL/m²; Men 19-88 mL; ESVI [mL/m²]:11-47 mL/m².

^eStroke Volume: 50-100 mL.

^fCardiac output: ∼4-6 L/min (varies by sex, age, and activity level).

^gLV ejection fraction: 50%-75%.

Reduction in Interstitial Myocardial Fibrosis Post-kidney Transplant

There was a significant reduction in myocardial native T1 at nine months in KT versus waitlisted participants (−47.5 ± 81.6 vs +13.8 ± 62.1 ms, P = 0.01, Fig 2). Compared with 39% of controls, 80% of KT recipients experienced a reduction in T1 times at 9 months post-KT. After exclusion of the 2 participants who received a deceased donor KT, the results remained essentially unchanged Table S2). In sensitivity analyses using baseline CKD stage and dialysis modality, a similar trend toward a reduction in T1 times in the KT group was observed; however, statistical significance was achieved only for the larger group with CKD 5 at baseline (mean T1 reduction at 9 months −55.7 ± 75.3 vs 48.3 ± 36.7 ms in the KT vs the waitlisted group, respectively, P = .007) (Fig S2).

Figure 2.

Change in myocardial native T1 times at 9 months. Changes in primary end-point of native T1 values (ms) between kidney transplant (N=25) and waitlist groups (N=18). Whisker box plots show a line at the median, the interquartile range within the blue boxes, and SD/95% CI, −45.0 (−117.3 to −3.3) versus 26.2 (−18.1 to 58.3), P = 0.01 between the 2 groups using the Mann-Whitney U-test. Within each group, compared with baseline T1 values, the 9-month follow-up T1 values were significantly decreased (P = 0.008) in the kidney transplant group and nonsignificantly increased (P = 0.36) in the waitlist group. Abbreviations: Ci, confidence interval; SD, standard deviation.

Left Ventricular Mass and Volumes Post-kidney Transplant

Several measures of myocardial structure and function were improved at 9 months after KT (LV mass index, LV end diastolic and systolic volume, and LV ejection fraction), whereas no significant changes occurred over 9 months in the waitlisted group (Table 3). Of note, the decline in LV end diastolic volume in the KT group was highly significant compared with the control group (Table 3).

Table 3.

Changes in Cardiac Magnetic Resonance Imaging Characteristics in HYFi Kidney Study Participants

	Transplant Group (n=25)		Waitlisted Control Group (n=18)
	Baseline	9-Mo	Baseline	9-Mo
Left Ventricle Measures
LVM (g)	137.08 ± 78.71	105.08 ± 42.09	114.11 ± 37.88	97.94 ± 38.43
ΔLVM	−31.54 ± 55.33		−15.17 ± 35.13
LVMI	66.64 ± 30.35	51.98 ± 17.15a	57.24 ± 16.44	49.19 ± 17.78
ΔLVMI	−14.25 ± 24.47		−8.00 ± 17.62
LVEDV	152.20 ± 40.96	130.32 ± 28.27a	146.0 ± 47.30	151.50 ± 43.01
ΔLVEDV	−21.73 ± 31.21b		5.50 ± 32.33
LVESV	68.08 ± 32.11	49.16 ± 13.71c	66.89 ± 35.72	61.94 ± 20.58
ΔLVESV	−18.77 ± 26.09		−4.94 ± 26.95
SV	83.15 ± 21.26	79.81 ± 19.63	79.22 ± 25.79	89.44 ± 28.29
ΔSV	−3.38 ± 18.28d		10.22 ± 20.81
LVEF	56.68 ± 11.26	62.00 ± 6.70a	56.00 ± 11.09	58.78 ± 6.99
ΔLVEF	5.32 ± 9.19		2.78 ± 10.00

Values represent mean ± standard deviation, or median (interquartile range); P-values obtained from paired sample t test.

Abbreviations: LVEDV, left ventricular end diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end systolic volume; LVM, left ventricular mass; LVMi, left ventricular mass index; SV, stroke volume.

^aP-value < 0.05 or.

^cP-value < 0.01 for the difference between the baseline and 9-month values.

^dP-value < 0.05 or.

^bP-value < 0.01 for the change in LV parameter over 9 months between the transplant and control group.

Plasma Protein Association With Baseline Interstitial Myocardial Fibrosis

At baseline, 126 proteins were significantly (false discovery rate < 0.05) associated with myocardial native T1 times, with most showing positive associations (Fig 3). After adjusting for clinical covariates, 43 proteins remained significant. Among the top hits (Table S3), M-CSF, zinc finger protein 174, S100-A8/A9 heterodimer, integrin β5, thyrotropin receptor, and deoxyribonuclease-2β were positively associated with interstitial myocardial fibrosis.

Figure 3.

Volcano plot of protein levels associated with myocardial fibrosis at study baseline (N=48). Volcano plot showing differentially expressed protein levels associated with myocardial fibrosis as measured by native T1 values. Each data point indicates a specific protein. Positive or negative magnitude fold changes are shown along the X-axis, and statistical significance (−log10 adjusted P values) is depicted along the Y-axis. Dashed horizontal lines show the P-value thresholds. The most significant proteins, with associated gene name labels, are shown in red font. P-values were obtained from linear regression models adjusted for the Benjamini and Hochberg false discovery rate.

Twelve proteins showed > 60% stability, with 10 significantly associated with baseline T1, indicating true biological signals. Overlap between adjusted linear regression analyses and Bayesian variable selection confirmed M-CSF as the top protein (Tables S3-S5). Similarly, the SuSiE²⁸ selection method also prioritized M-CSF, followed by S100-A8/A9 heterodimer, lamin-B2, probable RNA-binding protein 23, and transmembrane and immunoglobulin domain-containing protein 2.

The Kyoto Encyclopedia of Genes and Genomes database analysis identified 17 proteins linked to 22 enriched pathways implicated in the pathogenesis of cardiac remodeling and fibrosis, including arrhythmogenic right ventricular cardiomyopathy, hypertrophic cardiomyopathy, and dilated cardiomyopathy (Fig S3). Protein-protein interaction analyses revealed 3 main clusters: histones and transcription factors involved in gene regulation; an inflammation-related pathway including M-CSF; and proteins involved in cell signaling and adhesion (Fig S4).

Plasma Protein Association with Changes in Interstitial Myocardial Fibrosis Post-kidney Transplant

In the KT group, 24 participants had decreased T1 times at 9 months, reflecting interstitial myocardial fibrosis regression. Logistic regression analyses revealed baseline proteins associated with interstitial myocardial fibrosis improvement post-KT (Table S6) or worsening in the waitlisted group (Table S7), but none of these protein associations remained significant after FDR correction.

Because M-CSF level was highly associated with interstitial myocardial fibrosis at baseline (Fig 3, Table S3), we assessed the change in baseline versus 9-month plasma M-CSF concentrations using ELISA assays. For the entire population, the change in T1 times correlated with the change in M-CSF levels over 9 months overall (r = 0.32, P = 0.05, Fig 4) and in the dialysis subset (Fig S5). Furthermore, longitudinal M-CSF levels declined significantly in KT recipients compared with controls (P = 0.001).

Figure 4.

Change in T1 times versus change in M-CSF during study follow-up in all participants (N=43). The change was defined as the value at 9 months minus the value at baseline. The X-axis shows the change in plasma M-CSF value; the Y-axis shows the change in CMR T1 value. The reduction in T1 times correlated with the reduction in M-CSF post-KT. Pearson correlation (r) = 0.32, P = 0.05. Abbreviations: CMRI, cardiac magnetic resonance imaging; KT, kidney transplantation; M-CSF, macrophage colony-stimulating factor.

Discussion

Cardiac structural and functional abnormalities are common in CKD and are associated with cardiovascular mortality.2, 3, 4, 5 Though CKD is recognized as a major contributor to myocardial remodeling and fibrosis, the pathogenesis of interstitial myocardial fibrosis in advanced CKD is poorly understood. Our major findings are that interstitial myocardial fibrosis is reduced following KT, and a panel of plasma proteins derived from proteomic analysis is associated with interstitial myocardial fibrosis. M-CSF was the top significant protein associated with interstitial myocardial fibrosis and was reduced post-KT. These findings contribute to the limited but growing body of evidence on the cardiac benefits of KT and potentially enhance the understanding of interstitial myocardial fibrosis pathophysiology in the context of CKD by implicating M-CSF-regulated pathways.

Our study aligns with previous research showing that improved, albeit not completely normalized, renal function post-KT is associated with cardiovascular benefits.6, 7, 8 Multiple mechanisms of cardiac remodeling have been proposed, including correction of uremia-related biochemical imbalances, amelioration of chronic extracellular fluid overload, improvement in blood pressure and anemia, and reduction of inflammation and oxidative stress.^7,29,30 We showed that the improvement in interstitial myocardial fibrosis post-KT was independent of baseline or follow-up systolic or diastolic blood pressure, implying a structural rather than a hemodynamic basis for myocardial fibrosis in people with kidney failure. The T1 values at follow-up in both groups overlap with values observed in normal populations. However, longitudinal follow-up of T1 values is a significant predictor of cardiovascular events in various populations.31, 32, 33 Evidence suggests that as little as a 10 ms increase in native T1 times from a baseline value that overlaps with the normal range is associated with a 10% increase in mortality and a 7% increase in heart failure events.³⁴ It is therefore likely that the ∼50 ms difference in pre- versus post-KT times is clinically significant. The detection of small T1 differences over a short time period is also important because it allows for early interventions aimed at mitigating the progression of diffuse fibrosis, potentially preventing the development of heart failure or arrhythmias.31, 32, 33 This clinical implication is particularly pertinent given that CVD remains the leading cause of mortality post-KT,³⁵ owing to transplant-specific factors, including immunosuppression regimen, alterations in lipid and glucose metabolism, allograft dysfunction, rejection, chronic inflammation, etc.

Native CMRI scan T1 maps are widely recognized as the standard noninvasive test for diagnosing interstitial myocardial fibrosis, including in patients with kidney failure,^14,16,36 with good correlation with interstitial myocardial fibrosis assessed using late gadolinium enhancement CMRI scan.³⁷ In the absence of extracellular myocardial edema or conditions such as amyloid deposition, variation in extracellular volume fraction (ECV) is determined mostly by myocardial fibrosis, which has excellent agreement with human histologic measures of the collagen volume fraction.³⁷ We demonstrated that T1 times were reduced in the KT group, whereas T1 continued to increase in the waitlisted control group (Fig 2). These data are consistent with Contti et al.⁸ and suggest that the impact of KT goes beyond merely arresting the decline in cardiac function and may promote structural cardiac repair.

To identify potential mechanistic circulating proteins, we used a proteomics approach. We consistently observed an association between M-CSF levels and interstitial myocardial fibrosis; M-CSF level was positively associated with T1 maps at baseline, reflecting its potential as a mechanistic biomarker for cardiac fibrosis. Furthermore, M-CSF level and T1 were both significantly reduced in KT compared with waitlist controls at 9 months, suggesting that M-CSF might also be a predictive biomarker, which could be monitored for the resolution of cardiac fibrosis posttransplant. Future studies with larger populations would be required to validate M-CSF biomarker status for clinical use.

M-CSF is a cytokine produced by endothelial cells and regulates monocyte to macrophage differentiation, as well as monocyte/macrophage proliferation and survival.³⁸ The role of M-CSF is complicated, though, since it can drive macrophage polarization to become either pro (M1)- or anti (M2)-inflammatory macrophages.³⁸ In the heart, resident macrophages (M2-like) prevent the early stages of cardiac hypertrophy and fibrosis in animal models.³⁹ However, the numbers of resident macrophages decline with aging, and is replaced with monocyte-derived macrophages,^40,41 and pathologic fibrosis, and remodeling depend on the relative abundance of resident and monocyte-derived macrophages.^39,42 Taken together, M-CSF is a master regulator of monocytes/macrophages, with potentially important roles in cardiac pathophysiology that could mediate interstitial myocardial fibrosis of kidney failure.

Plasma M-CSF levels parallel interstitial myocardial fibrosis and are inversely related to kidney function. Bioactive M-CSF is a dimer of alternatively spliced polypeptides, with a molecular weight ranging from 50-130 kDa. These proteins are too large to be excreted by glomerular filtration, and renal handling of M-CSF by secretion or reabsorption is unknown. Therefore, whether KT-associated renal clearance of M-CSF is enhanced is speculative. It is also conceivable that immunosuppressive medications administered with KT could influence plasma M-CSF levels by altering monocyte-macrophage function.

Strengths of our study include proteomics analysis to identify plasma protein levels associated with improvement in interstitial myocardial fibrosis post-KT, validation of the top candidate using ELISA (M-CSF), and the cohort study design with matched controls, which permitted a more thorough assessment of the KT role on interstitial myocardial fibrosis compared with prior studies. However, some limitations deserve mention. First, the relatively small sample size may lack sufficient statistical power to detect all protein levels associated with interstitial myocardial fibrosis or a stronger association between M-CSF and T1 values to generalize the findings. We used various approaches to protein selection analyses to compensate for this and identify a set of highly influential proteins. The proteomics platform identified > 6,000 proteins, which is not comprehensive, and some viable candidates may have been omitted. Because the study included participants with advanced CKD, administration of gadolinium contrast agents available at the time was not feasible because of the risk of nephrogenic systemic fibrosis. This may have limited our ability to measure ECV. However, myocardial native T1 times have been shown to directly correlate with histologic measures of myocardial fibrosis,⁴³ have high interstudy reproducibility in patients receiving hemodialysis,¹⁶ and are comparable to ECV.⁴⁴ Our study also does not account for post-KT immunosuppressive therapy, which may have influenced the findings. The use of angiotensin-converting enzyme inhibitors/angiotensin receptor blockers and blood pressure was not statistically different between the groups; however, a clinically meaningful difference in systolic blood pressure (131 vs 139 mmHg) may account for the observed differences. Additionally, objective measurement of volume status was not collected in the study. Because of the observational nature of our study, causality cannot be inferred; a certain degree of residual confounding (eg, serum cholesterol and C-reactive protein levels were not measured) and selection bias cannot be excluded. Finally, we noted cardiac improvements at 9 months post-KT, but sustained reduction in T1 values and clinical benefits, such as decreased cardiac morbidity and mortality, were not assessed.

Future studies should build on these findings by incorporating longitudinal designs with serial cardiac imaging and repeated proteomic profiling to better characterize the temporal relationship between kidney transplantation and myocardial recovery. Integrating functional cardiovascular endpoints, such as cardiopulmonary exercise capacity, arrhythmia monitoring, or heart failure symptoms, will be essential to determine whether observed changes in imaging biomarkers (eg, T1 times) translate into meaningful clinical outcomes. Such studies will help validate potential biomarkers like M-CSF and clarify the biological pathways through which kidney transplantation may reverse cardiac remodeling.

In summary, our study (1) demonstrates that interstitial myocardial fibrosis is reduced following KT, and (2) contributes insight about M-CSF as a potential prognostic and mechanistic plasma biomarker for interstitial myocardial fibrosis associated with kidney failure. The validity of M-CSF and other candidate proteins as predictive biomarkers for reduction in interstitial myocardial fibrosis post-KT will require further validation. Future, larger studies with longer follow-up are warranted, with the ultimate goal of reducing or preventing the high CVD burden in kidney failure patients.

Supplementary Materials

Supplementary File (PDF)

Figure S1-S5; Table S1-S7