CD146 Endothelial Cells Facilitate Renal Interstitial Fibrosis Through Endothelial-to-Mesenchymal Transition

Introduction

Renal fibrosis is the common end point of all progressive kidney diseases [1]. Renal interstitial fibrosis (RIF) stands out as a prominent type of renal fibrosis, ultimately culminating in renal failure. The hallmark pathological feature of RIF is the proliferation of abundant myofibroblasts in the renal interstitium, leading to the excessive secretion of extracellular matrix proteins, such as collagen and fibronectin. However, there is no effective therapy for alleviating renal fibrosis, since the underlying mechanism of RIF has not been fully investigated.

Endothelial cells play a crucial role in the pathogenesis of RIF [2,3]. CD31 serves as the classical biomarker for endothelial cells; hence, endothelial cells are commonly identified as CD31 positive. However, within the CD31+ endothelial cell population, several subsets can exert distinct effects on RIF. CD146 is an immunoglobulin protein predominantly expressed in vascular endothelial cells. Originally identified as a melanoma cell adhesion molecule, CD146 is primarily expressed in endothelial cells and serves multiple functions, acting as both an adhesion molecule and a signaling receptor [4]. It plays a crucial role in maintaining vascular integrity [5] and contributes to neovascularization [6,7]. CD146 serves as a marker indicating endothelial cell activation or injury in chronic renal failure [8,9]. Notably, both the plasma and kidney biopsy expression of CD146 were found to be upregulated in patients with diabetic nephropathy, and these elevations were correlated with adverse renal outcomes [10]. Additionally, the expression of CD146 in renal tubules holds significance in assessing RIF during chronic renal failure [11]. However, while the study focused on tubular CD146, the relationship between endothelial CD146 and RIF remains unclear.

In the present study, we systematically elucidated the relationship between CD146+ endothelial cells and RIF. Initially, we elucidated the role of CD146+ endothelial cells in RIF within patients with chronic kidney disease (CKD) and in unilateral ureteral obstruction (UUO) mice. Subsequently, we investigated the potential involvement of CD146+ endothelial cells undergoing endothelial-mesenchymal transition in RIF. Moreover, platelet-derived growth factor receptor beta (PDGFRβ) plays a crucial role in kidney disease, particularly in renal fibrosis. It is a receptor that activates when it binds to platelet-derived growth factors (PDGFs), triggering cell growth, movement, and change. Mainly found on cells such as pericytes and smooth muscle cells in the kidney, PDGFRβ is linked to the transformation of these cells into myofibroblasts, which are key in renal fibrosis [12]. During fibrosis, the activation of PDGFRβ is associated with the buildup of myofibroblasts, which secrete proteins, such as collagen and fibronectin, contributing to kidney scarring [13]. While the role of PDGFRβ in fibrosis has been studied, how it might work with specific cell types, such as CD146+ cells, and its role in a process called endothelial-mesenchymal transition (EndMT) is not fully understood. Therefore, we delved into the potential impact of PDGFRβ on CD146+ endothelial cells undergoing endothelial-mesenchymal transition (CD146+ EndMT) in mice subjected to UUO.

Material and Methods

HUMAN KIDNEY TISSUE:

Patients with CKD who underwent kidney biopsy were randomly selected for inclusion in this study. The inclusion criteria with patients with CKD were as follows: (1) age range from 18 to 75 years old; (2) chronic kidney structural and functional impairments caused by various reasons; (3) renal biopsy performed at Beijing Friendship Hospital, Capital Medical University between January 1, 2017, and December 31, 2021; and (4) randomly selected patients with minimal glomerular abnormalities as the control group, along with patients exhibiting clearly defined tubulointerstitial damage pathologies, such as chronic tubulointerstitial disease, diabetic nephropathy, and IgA nephropathy, among others.

The exclusion criteria for patients with CKD were as follows: (1) history of acute kidney injury or current dialysis treatment; (2) current use of hormone or immunosuppressant therapies; (3) insignificant tubulointerstitial damage, such as membranous nephropathy; (4) comorbid severe systemic diseases (eg, malignancy, uncontrolled severe infection, heart failure); and (5) incomplete medical records or missing critical information.

Human kidney tissue stained with hematoxylin and eosin (HE) and Masson trichrome staining and kidney paraffin specimens were obtained with ethical approval and consent from Beijing Friendship Hospital, Capital Medical University. The kidney tissue stained with HE and Masson were analyzed for interstitial fibrosis and tubular atrophy (IFTA) scores. Kidney paraffin specimens were used for double immunofluorescence labeling.

MICE:

Experiments involving animals were approved by the Institutional Animal Care Committees of Beijing Friendship Hospital, Capital Medical University. Male C57BL/6 mice (8 weeks old, 20–25 g) were obtained from China Three Gorges University Laboratory Animal Center, and were adaptively housed for 7 days. Mice were divided into 3 groups: sham group, UUO day 3 group, and UUO day 7 group, which were groups of mice 3 and 7 days after surgery of UUO, respectively. Mice in the sham group received the same operation, but without ligation. UUO mice were killed on days 3 and 7, after modeling. Renal tissues from the ligated side and serum samples were collected.

HE AND MASSON STAINING:

The kidney tissue was fixed in 4% paraformaldehyde, at room temperature of 25°C for 24 h, and dehydrated as per the routine laboratory protocol. The mouse kidneys were fixed and embedded, and the kidney tissue was sliced into 4-μm thick sections. For HE staining, the sections were stained and then observed under an optical microscope. Sections were processed as previously described and stained with Masson trichrome to evaluate the degree of renal fibrosis. HE and Masson staining experiments were performed according to the manufacturer’s instructions. After sealing, images were captured under a microscope (Olympus, Japan). The average percentage of fibrotic area was calculated for each experimental group for statistical analysis.

The IFTA scores were as follows:

IFTA score of 0: no damage. No visible interstitial fibrosis or tubular atrophy is observed under the microscope. The renal architecture appears normal, with no expansion of the interstitial space and no shrinkage or distortion of the tubules.

IFTA score of 1: <25% damage. Mild interstitial fibrosis is noted, affecting less than 25% of the cortical area. There may be a slight increase in the interstitial matrix, with some individual tubules showing early signs of atrophy. The overall renal structure remains largely intact.

IFTA score of 2: 25–50% damage. Moderate interstitial fibrosis is present, affecting between 25% and 50% of the cortical area. The interstitial space is more prominently expanded with increased collagen deposition. Tubular atrophy is evident, with a significant number of tubules showing a reduction in size and brush border loss.

IFTA score of 3: >50% damage. Severe interstitial fibrosis is observed, with more than 50% of the cortical area involved. There is extensive collagen deposition, and the interstitial space is markedly expanded. Tubular atrophy is widespread, with the majority of tubules affected, showing significant shrinkage and loss of epithelial cells.

IMMUNOFLUORESCENCE:

Human kidney paraffin specimens were cut into 4-μm-thick sections for immunofluorescence, and the following antibodies were used: CD31 (ab182981, Abcam), CD146 (ab75769, Abcam), and alpha-smooth muscle actin (α-SMA; 14395-1-AP, Proteintech). Nuclei were stained with DAPI (YG-0001, Wanbang Biopharma). CD146⁺ CD31⁺ cells and CD146⁺ α-SMA⁺ cells were counted.

MULTIPLEX IMMUNOFLUORESCENCE STAINING:

As previously reported in the literature [14], multiplex immunofluorescence staining is a powerful technique that enables simultaneous in situ detection of CD31, CD146, α-SMA, and PDGFRβ in the renal interstitium of UUO mice. This technique was used in this study to identify vascular endothelial cell subsets. The target vascular endothelial cell subsets in this study included 3 types as follows: CD146+ CD31+ cells, CD146+ CD31+ α-SMA+ cells, and CD146+ CD31+ α-SMA+ PDGFRβ+ cells.

Multiplex immunofluorescence staining was obtained using a PANO 7-plex IHC kit (0004100100; Panovue, Beijing, China). The following primary antibodies were applied sequentially: CD31 (CST77699; Cell Signaling Technology, Boston, MA, USA), CD146 (ab75769; Abcam), α-SMA (ab7817; Abcam), and PDGFR-β (ab32570; Abcam), followed by horseradish peroxidase-conjugated secondary antibody incubation and tyramide signal amplification. The slides were microwave-heat-treated after each application of trichostatin A. Nuclei were stained with 4′-6′-diamidino-2-phenylindole (D9542; Sigma-Aldrich, Saint Louis, MI, USA) after all the antigens had been labeled. The stained slides were scanned using an Olympus scanner VS200 (Panovue) and analyzed with OlyVIA software. Whole-slide scanning of multispectral fluorescent images was performed using an Olympus VS200 MTL (Olympus), in conjunction with an Olympus UPLXAPO20X objective lens. Whole slide bright field and epifluorescence images were analyzed using QuPath software (Queen’s University of Belfast, Northern Ireland, UK).

STATISTICAL ANALYSIS:

Biological microscopy was used to collect HE staining, Masson staining, and immunofluorescence images; Image J/Image Pro Plus 6.0 was used to analyze the images, and a semi-quantitative analysis of the collagen volume fraction in Masson staining was performed. GraphPad Prism 8.0 was used for graph plotting. Experimental data were analyzed using SPSS 26.0 statistical software, with quantitative data represented as the mean±standard deviation. Comparisons among multiple groups were made using one-way ANOVA, and pairwise comparisons between groups were conducted using the least significant difference method. A P value of less than 0.05 was considered to indicate a statistically significant difference. The correlation between the transdifferentiation of CD146⁺ endothelial cells and RIF was analyzed using the Spearman correlation.

Results

:

Ten patients with CKD who underwent kidney biopsy at Beijing Friendship Hospital were enrolled in this study. Kidney tissues stained with HE and Masson trichrome were obtained from the Department of Nephrology at Beijing Friendship Hospital, and the tissues were analyzed for IFTA scores. Among them, 3 cases had a score of 1, while 2 cases had a score of 2, and 3 cases had a score of 3 (Table 1). Representative images are shown in Figure 1A and 1B.

Paraffin specimens of kidney tissue were examined for CD31 and CD146, and CD146 and α-SMA by double-labeling immunofluorescence. The results suggested that CD146 was co-expressed with CD31, a classical biomarker of vascular endothelial cells in blood vessels of renal interstitium (Figure 1C). Furthermore, CD146+ endothelial cells (CD146+ CD31+ cells) increased in renal interstitium with higher IFTA scores (Figure 1E). In addition, CD146 was co-expressed with α-SMA in vascular endothelial cells, and CD146+ α-SMA+ cell population increased in renal interstitium with higher IFTA scores (Figure 1D, 1F).

To investigate the potential impact of different etiologies on the research results, we conducted a detailed analysis of these subgroups. The results indicated that the increase in the number of CD146+endothelial cells and its impact on the IFTA score are mainly related to the degree of renal interstitial fibrosis, rather than to the specific etiology of CKD. This discovery suggested that CD146+endothelial cells may play a role independent of the cause in the process of renal fibrosis.

SEVERITY OF RIF PROGRESSIVELY WORSENED WITH THE DURATION OF UUO:

The kidney tubules showed mild congestion and swelling in the UUO group (Figure 2A). Compared with the sham group, in the UUO groups, fibrotic signs were evident on postoperative day 3, and more pronounced interstitial fibrosis was observed on postoperative day 7 (Figure 2B, 2C). The progressive worsening of kidney fibrotic changes with increasing UUO duration indicated successful model establishment.

:

We observed a significant decrease in the number and percentage of endothelial cells (CD31 positive) in the renal interstitium of UUO mice, compared with the sham group (Figure 3A–3C). Despite the reduction in endothelial cells, there was an increase in the number and percentage of CD146+ endothelial cells (CD146+ CD31+ double-positive) in UUO mice (Figure 3A, 3D, 3E). This suggested the potential involvement of CD146+ endothelial cells in the fibrotic lesions in UUO mice.

:

Given the increased presence of CD146+ endothelial cells in the renal interstitium of UUO mice, we investigated whether CD146+ EndMT played a role in UUO-induced renal fibrosis. α-SMA was considered as a biomarker for mesenchymal cells. We initially observed a significant increase in the α-SMA H-score and number of α-SMA+ cells in UUO-induced renal fibrosis (Figures 4, 5A, 5B). The results indicated an increase in mesenchymal cells in UUO-induced renal fibrosis. Subsequently, we conducted multiplex immunofluorescence staining to examine 3 biomarkers, CD146, CD31, and α-SMA, in situ to determine the potential involvement of CD146+ EndMT during UUO-induced renal fibrosis. The results indicated a significant increase in the number and percentage of CD146+ endothelial cells associated with EndMT, referred to as CD146+ EndMT, in UUO-induced renal fibrosis (Figures 4, 5C, 5D). Moreover, the elevated CD146+ EndMT was positively correlated with the areas of renal fibrosis in UUO mice (Figure 5E).

:

To explore the potential mechanism of CD146+ EndMT in UUO-induced renal fibrosis, we performed multiplex immunofluorescence staining for PDGFRβ, a co-receptor of CD146, specifically examining its expression in CD146+ EndMT. The staining was able to determine 5-color 4-label components, including DAPI, CD31, CD146, α-SMA, and PDGFRβ, in situ. The results indicated an upregulation of PDGFRβ in CD146+ EndMT in UUO-induced renal fibrosis (Figure 6A). The number and percentage of CD146+CD31+α-SMA+PDGFRβ+ cells increased in the renal interstitium of UUO mouse (Figure 6B, 6C).

Discussion

In this study, we found a positive correlation between renal fibrosis in CKD patients and UUO mice and the number of CD146+ endothelial cells. Furthermore, in CKD patients and UUO mice with severe renal fibrosis, there was a significantly higher number of CD146+ EndMT. Moreover, PDGRβ, a co-receptor of CD146, significantly increased in CD146+ EndMT in renal interstitium of UUO mice, and PDGFRβ+ CD146+ EndMT was significantly higher in UUO mice with severe renal fibrosis. Overall, CD146+ endothelial cells facilitated in RIF via EndMT, with PDGFRβ as the potential mechanism.

Renal endothelial cells present an impressive remodeling potential during vessel repair following kidney damage. However, if the damage persists in CKD, the cells will continue to be compromised and eventually undergo cell death [15]. In line with previous studies, our findings demonstrated a significant reduction in endothelial cells in the renal interstitium. This decrease suggests the involvement of vascular rarefaction in the progression of RIF.

An intriguing observation emerged as, despite an overall decline in the total number of endothelial cells in RIF, the subset of CD146+ endothelial cells exhibited an increase. This elevation was particularly notable in cases of severe RIF in CKD patients and UUO mice. This indicates that CD146+ endothelial cells are implicated in the pathogenesis of RIF. Previous studies have suggested that CD146 has been implicated in organ fibrosis, such as hepatic fibrosis [16] and systemic sclerosis [17]. In a mouse model of nephrotoxic serum-induced glomerulonephritis, the absence of CD146, as observed in knockout mice, resulted in a significant reduction in interstitial renal fibrosis [18]. Our findings align with this result, suggesting that CD146 is indeed implicated in RIF. However, it is noteworthy that CD146 expression exhibited a notable increase primarily in the injured endothelium of the glomerular tuft rather than in the renal interstitium in the above study. In contrast, our study revealed CD146 expression in tubular and glomerular endothelium. The exact cause for this disparity remains uncertain. It is plausible that the variation is attributed to the use of different mouse models; the aforementioned study used nephrotoxic serum-induced glomerulonephritis mice, whereas our study used UUO mice. A recent study has indicated elevated levels of serum and urinary CD146 in UUO mice. Additionally, certain drugs have demonstrated the capacity to diminish the severity of RIF, potentially by reducing serum CD146 [19]. However, the above-mentioned study did not report on CD146 expression in the renal interstitium, a factor that may play a more pivotal role in RIF. Overall, our study, along with others, indicate the involvement of CD146 in RIF. Furthermore, our findings specifically highlight the role of CD146+ endothelial cells in the renal interstitium in contributing to RIF.

RIF is distinguished by the uncontrolled accumulation of extracellular matrix, with myofibroblasts serving as the primary cell type responsible for its production. These myofibroblasts can arise from diverse differentiated cell types in response to injury. EndMT was an important pathogenesis of RIF via contributing to the pool of myofibroblasts [20,21]. Given the association between CD146 and RIF, we aimed to delve deeper into the potential mechanism involving CD146+ endothelial cells in tissue fibrosis by investigating CD146+ EndMT in RIF. The findings indicated a positive correlation between CD146+ α-SMA+ endothelial cells and IFTA in CKD patients. Moreover, there was a substantial increase in CD146+ CD31+ α-SMA+ endothelial cells in UUO-induced renal fibrosis. These results suggest the involvement of CD146+ endothelial cells in EndMT in RIF. Notably, as of now, there has been no identified study investigating the relationship between CD146 and EndMT. CD146 overexpression can promote epithelial-mesenchymal transition in both mouse embryonic fibroblasts and ovarian cancer cells [22]. This phenomenon is intricately linked to TGF-β signaling [23], a key growth factor in the process of EndMT. Additional investigations have indicated that TGF-β can upregulate the expression of CD146. In addition, some researchers have reported that PDGFRβ can play a role in EndMT in atherosclerotic lesion and heart disease [24,25]. We also found PDFGRβ was significantly higher in CD146+ EndMT cells of UUO mice, by multiplex immunofluorescence staining, a powerful technique enabling simultaneous in situ detection of CD31, CD146, α-SMA, and PDGFRβ in the renal interstitium of UUO mice. This result suggests that CD146+ EndMT has been implicated in RIF, with PDGFRβ identified as a potential underlying mechanism.

Certain limitations were present in this study. First, kidney paraffin specimens from CKD patients, not fresh kidney tissues, were utilized. Performing multiplex immunofluorescence on kidney paraffin specimens proved challenging, leading to the use of double-label immunofluorescence for these specimens. Consequently, CD146 was determined with α-SMA, and CD146 with CD31 were assessed on separate tissue slides. Simultaneous determination of CD146, CD31, and α-SMA on a single tissue slide was not feasible. Despite this limitation, given the observed coexpression of CD146 mainly with CD31 in endothelial cells, we maintain that CD146+ α-SMA+ cells are indicative of EndMT. Second, CD146 knockout mice were not used in this study due to restrictions of objective conditions; therefore, further work is required to confirm the role of CD146+ endothelial cells and explore the potential mechanism in RIF.

Conclusions

In the process of RIF, CD146 is mainly expressed in renal interstitial vascular endothelial cells and participates in endothelial-to-mesenchymal transition, which may be related to the PDGF-B/PDGFR-β signaling pathway.