Plasma Galectin-9 Is a Useful Biomarker for Predicting Renal Function in Patients Undergoing Native Kidney Biopsy Open Access

The study protocol was approved by the institutional review board, and all participants provided written informed consent in accordance with the Declaration of Helsinki. Between October 2018 and March 2021, a total of 259 adult patients undergoing clinically indicated percutaneous kidney biopsy in our department were recruited. The exclusion criteria included an age under 20 years (n = 2), refusal to sign the informed consent (n = 4), participation in other research studies (n = 2), and having undergone kidney transplant (n = 8). The study ultimately included 243 patients. Renal biopsy specimens were routinely processed for histologic examination, and an expert nephropathologist interpreted the test results. Urine and plasma were collected the day before kidney biopsy.

Plasma samples were stored at −80°C until analysis. Quantification of human plasma galectin-9, interleukin 6 (IL-6), and monocyte chemoattractant protein-1 (MCP-1) was performed with an in-house, multiplex, bead-based immunoassay as previously described.^24,25  In brief, a set of fluorescent beads coated with capture antibody targeted to the biomarker of interest was used. Samples were then incubated with the beads at room temperature. After washing away unbound substances, the beads were incubated with a mixture of biotinylated secondary antibodies and a streptavidin-phycoerythrin conjugate. Plasma levels of human galectin-9, IL-6, and MCP-1 were measured and quantified with a Bio-Plex 200 analyzer (Bio-Rad, Hercules, California). All samples were tested in duplicate to ensure accuracy.

The blood and urine samples were laboratory tested to assess renal function, proteinuria, and other baseline parameters. Renal function was estimated by using the Modification of Diet in Renal Disease Study equation,²⁶  and proteinuria was corrected for urine creatinine concentration in spot urine.

The histopathologic abnormalities were assessed with digital images obtained from paraffin-embedded biopsy specimens stained with hematoxylin-eosin, periodic acid–Schiff, Jones methenamine silver, and Masson trichrome staining. Whole slide images were acquired with the Hamamatsu NanoZoomer S210 Digital Slide Scanner (Hamamatsu Photonics, Bridgewater, New Jersey). Images at ×100 magnification were viewed with NDP.view2 software (Hamamatsu, Japan), and the rates of renal cortical involvement by interstitial fibrosis and tubular atrophy (IFTA) were evaluated. The IFTA degree was determined semiquantitatively for each individual stain and averaged to obtain an overall mean, as described previously.²⁷  We adopted the semiquantitative scoring system of Srivastava et al²⁸  for assessing the severity of IFTA (none, ≤10%; mild, 11%–25%; moderate, 26%–50%; and severe, >50%). They found that moderate and severe versus minimal IFTA were independently associated with kidney disease progression across a diverse group of kidney diseases. All quantitative image analyses were performed blinded to avoid biases.

To determine the intrarenal galectin-9 expression, paraffin-embedded renal biopsy tissues (2-μm thick) from healthy individuals (n = 7) and patients (n = 35) were processed for immunostaining as described previously.²⁹  The renal function was comparable in these patients and the remaining individuals in this cohort. The primary antibody was rabbit polyclonal anti–galectin-9 antibody, which detected amino acid residues within the aa 271–340 region (Atlas Antibodies; HPA047218). An isotype-matched irrelevant antibody served as a negative control. After incubation with horseradish peroxidase–conjugated secondary antibody, the color reaction was developed with diaminobenzidine. Sections were then counterstained with hematoxylin. The intensity and abundance of glomerular, tubular, and interstitial staining were estimated semiquantitatively from 0 to 3 (absent, weak, moderate, or strong). A mean score was calculated for each section from at least 10 high-power fields.

Immunofluorescence staining was performed to identify cell types that produce galectin-9 in cryostat-cut sections of kidneys from the patients. A rabbit polyclonal antibody against galectin-9 (1:50; HPA047218, Atlas Antibodies) was used for detecting galectin-9 by indirect immunofluorescence staining with fluorescein isothiocyanate–conjugated secondary antibodies (1:100; 111-095-003, anti-rabbit immunoglobulin G [IgG], Jackson ImmunoResearch). Antibodies for CD68 (1:200; ab201340, Abcam), CD20 (1:200; ab215866, Abcam), and CD3 (1:10, ab17143, Abcam) were used for detecting macrophages, B cells, and T cells, respectively, by indirect immunofluorescence staining with rhodamine-conjugated goat anti-mouse IgG antibody (1:100; 115-025-146, Jackson ImmunoResearch). Nuclei were then counterstained with DAPI (4′,6-diamidino-2-phenylindole, Sigma-Aldrich). Images were viewed under a fluorescence microscope (Nikon ECLIPSE TS100, Nikon, Japan).

Murine macrophage J774A.1 cells (Bioresource Collection and Research Center, Hsinchu, Taiwan) were cultured in Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum at 37°C in a humidified atmosphere containing 5% CO₂. To characterize the expression pattern of endogenous galectin-9 in J774A.1 cells under varying pathologic conditions, J774A.1 cells were seeded in 6-well plates at a 5 × 10⁶ cells/well density. The cells were then polarized into the M1 phenotype by treating with lipopolysaccharide (LPS, 100 ng/mL) or interferon-γ (20 ng/mL) or into the M2 phenotype by treating with IL-4 (20 ng/mL) + IL-13 (20 ng/mL), respectively. Cells were stimulated for the indicated periods before harvesting and stored at −80°C until further analysis.

The kidney tissue for the analysis of gene expression was immediately stored in the RNA stabilization reagent RNAlater (Thermo Fisher Scientific) according to the manufacturer's recommendations. The tissue samples were manually microdissected into glomerular and tubulointerstitial compartments. Total RNA of the dissected tubulointerstitial compartments and J774 cells was extracted and then reverse transcribed into cDNA as previously reported.³⁰  Quantitative real-time polymerase chain reaction (qPCR) assays were performed by using a 2× SYBR Green qPCR Master Mix (BioTools, New Taipei City, Taiwan) with a real-time PCR instrument (Applied Biosystems QuantStudio 3 Real-Time PCR System, Thermo Fisher Scientific). All reactions were run in triplicate. The relative mRNA expression levels of each sample's target genes were determined and normalized to the β-actin (ACTB) expression. Supplemental Table 1 (see supplemental digital content containing 3 tables and 5 figures at https://meridian.allenpress.com/aplm in the February 2023 table of contents) lists the primer sequences for the qPCR analysis.

The participants' baseline characteristics are expressed as means ± standard deviation, medians and interquartile range (IQR), or numbers and percentages. The study population was divided into 2 subgroups on the basis of median plasma galectin-9 values. The groups were compared with Student t and Mann-Whitney U tests for the continuous variables (for normal and nonnormal distributions, respectively) and with the χ² test for the categorical variables. To compare 3 or more groups, we performed a 1-way analysis of variance followed by the Bonferroni post hoc test or the Kruskal-Wallis test followed by the Dunn test, based on the data distribution. Correlations between variables were determined by using the Spearman rank correlation coefficient. The multivariate logistic regression revealed the independent risk factors for high plasma galectin-9 levels and advanced CKD. Parameters with P values < .10 in the univariate analysis and those considered relevant were entered as covariates. A receiver operating characteristic (ROC) curve analysis was performed to assess the performance of plasma galectin-9 in diagnosing advanced CKD. We also evaluated the impact of plasma levels of galectin-9 on adverse kidney outcomes in this patient group. The adverse renal outcome was defined as a composite of an eGFR decline of 50% or more, end-stage renal disease, or mortality due to cardiovascular or renal etiologies. The cumulative risk of renal end points was calculated by using the Kaplan-Meier method and compared with the log-rank test. Additionally, Cox regression analysis was used to determine the prognostic role of galectin-9 in patients with CKD. An analysis with P < .05 was considered statistically significant. Data were analyzed with SPSS version 23.0 software (SPSS Inc, Chicago, Illinois) and GraphPad Prism 9 (GraphPad Software, San Diego, California).

Table 1 presents the clinical characteristics of the 243 patients who underwent kidney biopsy (mean age, 56.4 ± 16.4 years; 39.1% women; median eGFR, 36 mL/min/1.73 m² [IQR, 17–65 mL/min/1.73 m²]; proteinuria, 2.93 mg/mg creatinine [IQR, 1.05–7.68 mg/mg]). Hypertension, diabetes, and cardiovascular disease were present in 99 (40.7%), 62 (25.5%), and 27 (11.1%) participants, respectively. The patients were classified into 6 diagnostic categories on the basis of their clinical and histopathologic findings. The most common kidney disease categories were nonproliferative glomerulopathy, proliferative glomerulonephritis, and diabetic kidney disease. Supplemental Table 2 presents the histologic diagnoses of the renal biopsies included in each category. The median levels of plasma galectin-9, IL-6, and MCP-1 were 750.1 pg/mL, 8.2 pg/mL, and 145.9 pg/mL, respectively.

To further assess the relationship between galectin-9 and clinical features, the patients undergoing native renal biopsy were divided into 2 groups on the basis of median plasma galectin-9 values (Table 1). The high galectin-9 group had a larger percentage of elderly patients and patients with hypertension and diabetes mellitus than the low galectin-9 group. The high galectin-9 group tended to have more severe anemia, lower eGFR values, a larger percentage of patients with proteinuria, and increased plasma IL-6 and MCP-1 levels. Compared with the low galectin-9 group, the high galectin-9 group was more likely to have diabetic kidney disease and less likely to have nonproliferative glomerulopathies. The high galectin-9 group had a greater degree of IFTA than the patients in the low galectin-9 group (Table 1).

Spearman correlation coefficient analysis was performed to determine the correlation between circulating galectin-9, renal function, proteinuria severity, degrees of IFTA, and systemic inflammatory markers in these patients. The correlation analysis revealed that galectin-9 levels were positively correlated with proteinuria, IFTA, IL-6, and MCP-1 and inversely correlated with eGFR, as presented in Figure 1, A and B. Our results also suggested that level of eGFR was significantly related to most of the variables listed here; however, there was no correlation between eGFR and plasma concentrations of MCP-1. These findings led to speculation that plasma galectin-9 levels may add value to assess and monitor the inflammatory status in patients with CKD.

Logistic regression was used to determine the factors influencing plasma galectin-9 levels (Table 2). In the unadjusted analyses, older age, hypertension, diabetes mellitus, lower eGFR and hemoglobin levels, higher proteinuria, higher plasma IL-6 and MCP-1 levels, absence of nonproliferative glomerulopathy, and greater IFTA severity were associated with high circulating galectin-9 levels. After adjusting for age, sex, and variables with P values < .10 in the univariate analyses, only lower eGFR and higher plasma MCP-1 concentrations remained associated with increased plasma galectin-9 levels.

There is an increased risk of CKD-related complications as eGFR falls below 45 mL/min/1.73 m².³¹  We therefore used a logistic regression analysis to determine the factors that predicted the presence of CKD stage 3b or higher (Table 3). In the unadjusted analyses, plasma galectin-9 levels were significantly associated with more advanced CKD stage, and this association persisted after adjusting for potential confounders.

We used an ROC analysis to evaluate the ability of plasma galectin-9 to identify advanced CKD (Supplemental Figure 1, A). The optimal cutoff point for detecting advanced CKD was 689.6 pg/mL (sensitivity, 74.0%; specificity, 74.2%; AUC [area under the curve], 0.797; 95% CI, 0.742–0.852). Our analyses also demonstrated that plasma galectin-9 had a high accuracy for diagnosing the earliest stages of CKD (Supplemental Figure 1, B and C).

The prognostic ability of plasma galectin-9 to predict the risk of adverse renal outcome was evaluated. Among 193 patients who were followed up for at least 1 year, 52 composite renal outcomes occurred: 45 in the high galectin-9 group and 7 in the low galectin-9 group (hazard ratio, 6.441; 95% CI, 2.902–14.294; see Supplemental Figure 2 and Supplemental Table 3). After multivariate adjustment, a trend for increased risk of renal adverse events could be observed in patients with elevated plasma galectin-9 levels.

Periodic acid–Schiff and Masson trichrome staining (Figure 2, A through D) showed morphologic changes in the renal cortex, including global glomerulosclerosis, interstitial inflammation, and IFTA in patients with CKD, compared with those without CKD. To identify the cell types producing galectin-9 during CKD progression, we analyzed the presence of galectin-9 protein from paraffin-embedded kidney biopsy samples, using immunohistochemistry (Figure 2, E through H). Galectin-9 was not detectable in the kidneys of healthy controls (Figure 2, E and F). For those with advanced CKD, galectin-9 expression was significantly upregulated, mainly in the interstitial cells, with very little expression in tubules or glomeruli (Figure 2, G and H; Supplemental Figure 3). We found no difference between the expression of galectin-9 in both the glomerular and tubular compartments of either the fibrotic or nonfibrotic kidneys.

Recently, a publicly available single-nucleus RNA sequencing data set on human diabetic kidney samples showed enhanced expression of LGALS9 mRNA, which encodes human galectin-9, in the leukocyte subsets and endothelial cells when compared with those of healthy controls (Supplemental Figure 4).³²  Using double-immunofluorescence staining, we found that most galectin-9–positive leukocytes in the interstitial space of patients with CKD were macrophages (Figure 2, I through L). Galectin-9 expression was also detected in a small number of T and B cells (Supplemental Figure 5).

From these results, we quantified LGALS9 mRNA expression in the renal tubulointerstitial compartment in the patients with paired plasma and kidney biopsy samples and evaluated its associations with the clinical and laboratory parameters. Intrarenal LGALS9 RNA expression correlated significantly with the plasma galectin-9 concentrations in 58 patients who underwent percutaneous kidney biopsy (r_s = 0.55; P < .001) (Figure 3, A). We found that LGALS9 expression was positively correlated with gene expression levels of various extracellular matrix molecules and was inversely correlated with eGFR (Figure 3, B through F), suggesting the intimate relationship between galectin-9 and renal fibrosis in CKD progression.

To determine which factors might upregulate galectin-9 expression by monocytes/macrophages, murine J774 cells were treated with various stimulants, including LPS, interferon-γ, and IL-4 + IL-13. Cells were then confirmed to be polarized into corresponding states 24 hours after stimulation by using qPCR to assess the expression of M1- and M2-related genes (Figure 4, A through D). Galectin-9 expression from macrophages was significantly increased by interferon-γ but not by LPS or IL-4 + IL-13, suggesting that interferon-γ–activated macrophages might be the main source of galectin-9 production under pathophysiologic conditions (Figure 4, E).

In this study, we observed a significant increase in plasma galectin-9 concentrations in the patients with renal impairment who underwent kidney biopsy, which correlated inversely with renal function and directly with proteinuria and IFTA extent. Lower eGFRs and higher plasma MCP-1 concentrations appeared to be linked to elevated circulating galectin-9 levels. We showed that increased plasma galectin-9 levels were associated with the presence and severity of renal dysfunction. The immunohistochemical and qPCR analysis of microdissected tubulointerstitial components of human renal biopsy samples showed increased galectin-9 expression in the interstitial space of fibrotic kidneys compared with that of histologically normal controls, and the major source of galectin-9 within the injured kidney appeared to be macrophages. The in vitro studies demonstrated that galectin-9 production in monocytes/macrophages was significantly enhanced by interferon-γ. Our findings therefore suggest that plasma galectin-9 can be considered a good biomarker for diagnosing CKD and that it might play a key role in the pathogenesis of kidney disease progression.

There are numerous possible mechanisms that might explain the higher plasma galectin-9 levels in the patients with CKD than in those with normal renal function. Galectin-9 has a molecular weight of approximately 36 kDa, and a decreased GFR might result in its accumulation in the systemic circulation. The current study also showed that plasma galectin-9 levels correlated significantly with the intrarenal mRNA expression of the LGALS9 gene in the patients with paired plasma and kidney biopsy samples. This finding suggests that injured kidneys could be a potential source of circulating galectin-9. Plasma galectin-9 concentrations might also reflect a delicate balance between the immune response and host. Studies have noted that galectin-9 is involved in the macrophage polarization process, thereby influencing the pathogenesis of numerous diseases.^33–35  Our study also demonstrated that MCP-1, an important mediator regulating migration and infiltration of monocytes/macrophages into inflammation sites, is an independent predictor for plasma galectin-9 levels after accounting for other confounding variables. Given that activation of kidney-resident macrophages and circulating monocytes is common in CKD, these cells might be an important contributor to elevated plasma galectin-9 levels in this population.^36–38 

Histopathologic lesions on native kidney biopsies in kidney disease of diverse etiology have recently been shown to provide prognostic information independent of clinical parameters, such as comorbid conditions, proteinuria, and eGFR.^28,39  Regardless of the underlying cause, tubulointerstitial injury is always present as kidney disease progresses. Several studies have suggested that, even among patients with chronic glomerulonephritis, the extent of tubulointerstitial damage has more impact on the subsequent decline in renal function than the degree of glomerular injury.^40,41  Tubulointerstitial injury causes infiltration of inflammatory cells in the interstitium, leading to IFTA and CKD progression. This study found that increasing IFTA severity was associated with lower eGFR after a multivariable adjustment. Our data confirm the importance of IFTA, which is consistent with previous findings.

Galectin-9 exerts pleiotropic effects on numerous tissues by regulating cellular processes such as proliferation, differentiation, apoptosis, and immunomodulation.^42,43  Beyond its role in normal physiologic processes, several studies have found that the circulating galectin-9 was elevated in certain medical conditions, such as infectious disease, autoimmune disease, malignancy, and renal disease.¹⁹  Although galectin-9 affects glomerular injury in certain immune-mediated glomerular diseases,^20,21  less is known about the relationship between this protein and renal tubulointerstitial damage during CKD progression. Researchers have recently demonstrated the importance of galectin-9 in fibrotic remodeling in various organ systems, including the liver, skin, and lungs.^44–46  Galectin-9 is highly expressed in the liver and systemic circulation in chronic hepatitis C, and an excess of galectin-9 can enhance the production of proinflammatory and profibrotic cytokines from Kupffer cells, which might result in liver damage and fibrosis.⁴⁴  Galectin-9 expression has also been reported as upregulated in dermal fibroblasts of skin sections in systemic sclerosis.⁴⁶  Increased galectin-9 expression led to a dysregulated balance of T helper 1/T helper 2 (T_H1/T_H2) immune responses that promoted skin fibrosis in these patients. Our results indicate that human galectin-9 is strongly expressed in interstitial macrophages at sites of tubulointerstitial injury and fibrosis. Galectin-9 expression correlates with the extent of tubulointerstitial fibrosis and tubular cell damage. In vitro experiments have shown that galectin-9 expression in macrophages is significantly upregulated by interferon-γ, which participates in the pathogenesis of renal fibrosis.⁴⁷  Collectively, these findings suggest that galectin-9 might be involved in CKD progression, partly via immune system modulation.

Several limitations to this study need to be acknowledged. First, our preliminary investigation has revealed that elevated plasma galectin-9 is linked to poor outcome in patients with CKD. Further studies using a larger population are still needed to validate these findings. Second, although appropriate adjustments were performed in the multivariable model, unmeasured and residual confounding factors might still exist. A third limitation is that the study population had only a small proportion of patients with diabetic kidney disease, who constitute most cases of CKD worldwide. Given that the patients exhibited a broad mix of renal pathologies, the results are more likely to be applicable to patients with kidney disease of diverse etiology. Fourth, although only a few patients with a pathologic diagnosis of acute kidney injury were included in the cohort, this may influence the association between galectin-9 expression and investigated parameters. Lastly, we acknowledge the limitations of this work that failed to elucidate the role of galectin-9 in renal immune homeostasis and tubulointerstitial fibrosis. Well-designed studies using genetically modified animals are needed to clarify these questions.

In summary, this study showed that galectin-9 expression in kidneys and peripheral blood was significantly increased in the patients with CKD when compared with healthy controls. Plasma galectin-9 levels were negatively correlated with renal function and positively correlated with proteinuria and IFTA severity. The in vitro study showed that interferon-γ–stimulated macrophages might contribute to galectin-9 production within damaged kidneys. These findings provide valuable information for further investigation of galectin-9 as a prognostic biomarker and therapeutic target in patients with CKD.

We thank the Academia Sinica Inflammation Core Facility, IBMS for technical support.