Interplay of HIF-1α, SMAD2, and VEGF signaling in hypoxic renal environments: impact on macrophage polarization and renoprotection

Abstract

This study investigates the roles of hypoxia-inducible factor (HIF-1α), SMAD2, and vascular endothelial growth factor (VEGF) in renal repair under hypoxic conditions, focusing on their impact on macrophage phenotype transformation. Bioinformatics analysis identified SMAD2 as a key gene in renal injury, correlating with HIF-1α and VEGF levels. In a cohort of 60 pediatric patients, non-AKI individuals exhibited higher VEGF and SMAD2 levels but lower HIF-1α compared to AKI patients. Using a chronic hypoxia rat model, bevacizumab treatment exacerbated renal damage, as indicated by elevated serum creatinine (79.4 ± 61.7 μmol/L), increased inflammatory markers, and heightened HIF-1α expression. Bevacizumab’s inhibition of VEGF also impaired macrophage phenotype modulation. In vitro experiments revealed that hypoxia alone had minimal direct effects on macrophage polarization but enhanced IL-4-induced M2 polarization, further amplified by SMAD2 and VEGF overexpression. These findings underscore the distinct yet interconnected roles of HIF-1α, SMAD2, and VEGF in shaping the hypoxic renal microenvironment and influencing macrophage polarization. The study highlights VEGF’s critical role in renal repair and its interaction with hypoxia and inflammatory pathways, which may modulate macrophage polarization and ultimately impact renal outcomes. These insights suggest a promising therapeutic strategy for kidney diseases.

Introduction

Oxygen homeostasis is vital for cellular energy production, and the kidneys are highly susceptible to hypoxic conditions. Hypoxia triggers kidney injury through multiple pathophysiological mechanisms accompanied by inflammatory responses [1,2]. Cellular adaptation to hypoxia is primarily mediated by hypoxia-inducible factor (HIF), a heterodimer comprising an inducible α subunit (HIF-α; the main hypoxia-sensitive component) and a constitutively expressed β subunit (HIF-β). Three distinct α subunit isoforms have been identified: HIF-1α, HIF-2α, and HIF-3α, each with tissue-specific expression patterns and functions [3]. In the kidney, HIF-1α predominantly localizes to tubular epithelial cells, which are the primary sites of injury in kidney injury. It responds rapidly to hypoxia, with transcriptional upregulation occurring during the early phases of injury. In contrast, HIF-2α is mainly expressed in endothelial cells, interstitial fibroblasts, and peritubular cells, but is notably absent from tubular epithelia, while HIF-3α functions primarily as a negative regulator through competitive inhibition of HIF-1α and HIF-2α transcriptional activity [3,4].

Our previous studies demonstrate that under moderate hypoxia (10% O₂), HIF-1α upregulates vascular endothelial growth factor (VEGF) to promote angiogenesis and renal protection. In severe hypoxia (7% O₂), however, HIF-1α accumulation correlates with amplified inflammation and renal damage, while VEGF expression declines. This reveals a hypoxia-dependent duality in HIF-1α/VEGF signaling, where VEGF-mediated angiogenesis serves as a pivotal protective mechanism under moderate hypoxia [5,6].

Macrophages, one of the major inflammatory cell types, exhibit a dual role in kidney injury. During the early phase of renal damage, macrophages can polarize toward a pro-inflammatory phenotype (MΦ-1), while during the repair phase, a transition to an anti-inflammatory phenotype (MΦ-2) is observed, which mitigates inflammation and supports tubular cell regeneration [7]. Previous studies indicate that macrophages can stimulate endothelial proliferation and angiogenesis, stabilizing tip cell fusion [8,9]. The potential relationship between macrophage plasticity and the hypoxia-dependent duality in HIF-1α/VEGF signaling remains an area of active investigation.

Emerging evidence indicates that inflammatory signaling pathways interact with the HIF-1α/VEGF axis, potentially influencing macrophage polarization [10–12]. Apart from hypoxia-induced pathways, SMAD indirectly modulates VEGF expression, thereby facilitating angiogenesis [11–14]. The duration and severity of hypoxia also affect these interactions, with prolonged hypoxia altering SMAD2 expression and potentially impacting the protective role of SMAD2 under chronic hypoxic conditions [15,16]. The dynamic interplay between HIF-1α/VEGF signaling, inflammatory mediators, and macrophage phenotypic switching remains unclear and requires further investigation.

A description of the study profile is shown in Figure 1. In this study, we investigate the crosstalk between inflammatory pathways and the HIF-1α/VEGF axis in early kidney injury, focusing on their dual regulation of macrophage phenotypic switching. Through integrated bioinformatics analysis of blood samples from acute kidney injury (AKI) patients, validation in chronic hypoxia rat models, and in vitro macrophage polarization assay. The identification of a link connecting inflammatory-hypoxic crosstalk and macrophage polarization dynamics suggests potential therapeutic implications for mitigating kidney injury progression.

Figure 1.

Flow chart of study design. (A) Clinical validation phase: Identification of SMAD2 as a key target gene associated with kidney injury through bioinformatic analysis, followed by comparative analysis of HIF-1α, SMAD2, and VEGF expression levels in plasma samples between AKI and Non-AKI groups. (B) In vivo validation: Establishment of a chronic hypoxia rat model to investigate the correlation between kidney injury biomarkers and expression levels of HIF-1α, SMAD2, and VEGF under hypoxic conditions. (C) In vitro interaction analysis: Investigation of the association between HIF-1α, SMAD2, and VEGF expression dynamics and macrophage polarization (M1/M2 phenotypes) in RAW 264.7 cells, exploring their potential interplay under hypoxic conditions. VEGF: vascular endothelial growth factor; HIF-1α: hypoxia-inducible factors; AKI: acute kidney injury.

Materials and methods

Target gene analysis in acute kidney injury patients

The target gene screening process is shown in Figure S1. We searched PubMed, Embase, and Web of Science for microRNAs (miRNAs) related to kidney injury using terms like ‘acute kidney injury’ and ‘AKI’ with ‘microRNA’. Inclusion criteria were: (a) use of plasma or serum samples, (b) at least ten cases in any group, and (c) miRNA validation by qRT-PCR. Exclusions included protocols, reviews, abstracts, and editorials. miRNA sequences were explored via miRBase. Target gene prediction used miRTarBase, DIANA TOOLS, TargetScan, and miRDB, analyzing genes found in at least three databases. Gene ontology (GO) enrichment was done using DAVID, and KEGG classified protein pathways. Cytoscape visualized the network between AKI pathways and genes. To validate hub genes’ diagnostic value in AKI, we analyzed kidney samples from NCBI GEO (GSE30718). Data were processed with GEOquery, removing probes for multiple genes to retain the highest expression. Principal component analyses (PCAs) used FactoMineR, differential analysis used limma, and ggplot2 was for visualization.

Inclusion and data collection for patients

Children admitted to the pediatric intensive care unit (PICU) of Xinhua Hospital Shanghai Jiao Tong University School of Medicine, from January 2022 to June 2023, were included. The study was approved by the ethics committee (Approval No. XHEC-C-2022-012-1) and registered (NCT06197828).

Inclusion criteria: (a) age 28 days to 16 years. (b) PICU stay >48 h. (c) Indication for urinary catheter placement and ability to monitor renal ultrasound [17]. (d) at least two serum creatinine (SCr) measurements post-admission, with hourly urine output recorded for Kidney Disease: Improving Global Outcomes AKI assessment [18]. Exclusion criteria: (a) history of chronic renal insufficiency or pre-admission dialysis. (b) previous renal surgery. (c) Long-term nephrotoxic drug use [19]. (d) family history of kidney disease or conditions like lupus. (e) refusal to consent. (f) incomplete records.

Patients meeting inclusion criteria and passing exclusion screening constituted the study cohort. Baseline SCr was defined as the most recent measurement before PICU admission or, when unavailable, age- and sex-specific normal reference values [20]. Patients were monitored for seven days following PICU admission and classified into AKI and Non-AKI groups based on serum creatinine measurements, in accordance with the 2012 KDIGO guidelines [18]. Baseline data were collected upon PICU admission, including demographics, such as sex, age, body mass index (BMI), pediatric critical illness score (PCIS), and pediatric risk of mortality III scores (PRISM III). Renal function indicators, cytokines, renal perfusion metrics, and urinary oxygen indicators were collected at PICU admission, 24 h, and 48 h. Renal function: SCr, estimate glomerular filtration rate (eGFR), neutrophil gelatinase-associated lipocalin (NGAL), and renal resistive index (RRI). Cytokines: interleukin-6 (IL-6), interleukin-10 (IL-10), and tumor necrosis factor (TNF-α). Urinary oxygen indicators: urinary lactate, urinary partial pressure of oxygen (PuO₂), and urinary partial pressure of carbon dioxide (PuCO₂). Interventions throughout the treatment period were analyzed via retrospective chart review, including continuous renal replacement therapy (CRRT) and mechanical ventilation.

Construction of a chronic hypoxia rat model with renal impairment

Using the ProOx-100 system (Shanghai TOW Intelligent Technology Company, China), we established a chronic hypoxia rat model by adjusting inhaled oxygen levels in a normobaric environment, as described previously [5,6,21]. In evaluating the rat model of hypoxia-induced renal injury, we referenced: (i) the 2012 KDIGO guidelines for elevated SCr levels [18], (ii) Wang et al.’s criteria indicating a 40% increase in NGAL from baseline as suggestive of kidney injury [22], and (iii) Renal histopathology scoring was assessed using the scoring system described by Choi et al. [23].

In preliminary experiments, rats were exposed to oxygen concentrations ranging from 21 to 5% for 14 days. No significant kidney injury occurred at 10–21% oxygen, while 5% oxygen resulted in mortality. We identified 7% oxygen as the threshold concentration that induced significant kidney injury without mortality. Experimental periods shorter than 14 days yielded insufficient pathological changes. Therefore, the final protocol utilized 21% oxygen for normoxia and 7% for hypoxia over a 14-day exposure period.

Based on our previous findings highlighting VEGF’s critical role in renal repair, we employed bevacizumab (MedChemExpress LLC, Monmouth Junction, NJ, USA), a recombinant humanized IgG1 monoclonal antibody that specifically targets VEGF, inhibiting angiogenesis by preventing VEGF from binding to its tyrosine kinase receptors [24]. Despite being a humanized antibody, bevacizumab effectively recognizes and binds to rat VEGF due to the highly conserved structure of VEGF across species, making it suitable for experimental animal models [25]. Research has also shown that bevacizumab can influence podocyte VEGF secretion, thereby altering renal VEGF expression levels and potentially increasing the risk of glomerular filtration barrier injury [26]. Our previous research has demonstrated that bevacizumab administration affects both VEGF expression and its downstream effects, potentially exacerbating renal damage [6]. In this study, bevacizumab was administered intravenously at 5 mg/kg, twice weekly [26,27]. It was diluted in saline, with an equivalent volume of saline as a placebo.

Rats were divided into four groups: (a) 21%O₂ + normal saline (21%O₂ + NS); (b) 21%O₂ + bevacizumab (21%O₂ + Bev); (c) 7%O₂ + normal saline (7%O₂ + NS); (d) 7%O₂ + bevacizumab (7%O₂ + Bev). This study was conducted in strict accordance with the recommendations of the guidelines of the Institutional Animal Care and Use Committee of Xinhua Hospital Affiliated with Shanghai Jiao Tong University School of Medicine (Shanghai, China), and the experimental protocol was approved by the Ethics Committee of the same institution (XHEC-F-2021-001).

Paraffin embedding and sectioning

Rat kidney specimens were fixed in 4% formaldehyde for 24 h, dehydrated in graded ethanol (70–100%), and embedded in paraffin. Sections (2–3 μm) were cut, mounted on a 42 °C water bath, dried at 65 °C overnight, labeled, and stored.

Hematoxylin-eosin staining and renal injury scoring

Kidney sections were deparaffinized in xylene, rehydrated through graded ethanol, and subjected to hematoxylin-eosin (H&E), Masson’s trichrome (MT), and periodic acid-Schiff (PAS) staining according to previously described study [5]. After dehydration and clearing, sections were mounted and staining quality was verified microscopically. Renal injury was scored on H&E-stained sections at ×20 magnification by evaluating five random fields for tubular injury, such as brush border loss and epithelial degeneration. Scores ranged from 0 (normal) to 5 (>90% damage) [23].

Culture and treatment of RAW264.7 macrophages

The RAW264.7 mouse macrophage line was obtained from Procell Life Science & Technology (China) and cultured in a sterile environment using specific media.

A preliminary experiment was conducted to assess the oxygen tolerance of RAW264.7 cells. Cells were cultured in a tri-gas incubator (Ox-100C, Shanghai TOW Intelligent Technology Company, China) at varying oxygen concentrations: 21, 15, 10, and 5% O₂, with temperature maintained at 37 °C and humidity at 5–10%. At 10 and 5% O₂, macrophages ceased adherent growth after 48 h, exhibiting significant cell detachment and death (Figure S2(A)). CCK-8 assay indicated a marked decrease in cell viability at these lower oxygen levels (Figure S2(B)). Consequently, subsequent experiments were conducted at 15% O₂. Cells were incubated at 37 °C with 5% CO₂ until 90% confluence, then washed, detached, and passaged. They were seeded at 1 × 10⁵ cells per well in 6-well plates, incubated overnight, and treated according to experimental groups.

Cells were exposed to normoxia (21%O₂), hypoxia (15%O₂), normoxia with IL-4-induced polarization (21%O₂ + IL-4), hypoxia with IL-4-induced polarization (15%O₂ + IL-4). The 15%O₂ + IL-4 group can be further subdivided as: overexpression control (15%O₂ + IL-4 + OC), SMAD2 overexpression (15%O₂ + IL-4 + SMAD2-OE), VEGF overexpression (15%O₂ + IL-4 + VEGF-OE), knockdown control (15%O₂ + IL-4 + KC), SMAD2 knockdown (15%O₂ + IL-4 + SMAD-KD), and VEGF knockdown (15%O₂ + IL-4 + VEGF-KD).

Cells were cultured under hypoxia for 48 h. Groups receiving IL-4 (20 ng/mL, Peprotech, USA) were pre-treated for 24 h before hypoxic exposure. For SMAD2 and VEGF overexpression (15%O₂ + IL-4 + SMAD-OE, 15%O₂ + IL-4 + VEGF-OE) and knockdown groups (15%O₂ + IL-4 + SMAD-KD, 15%O₂ + IL-4 + VEGF-KD), cells were transfected with specific plasmids and cultured for 48 h before IL-4 treatment and hypoxia exposure. Refer to Figure S3 for grouping details.

Co-culture of RAW 264.7 cells with glomerular endothelial cells

One day before co-culture, primary mouse glomerular endothelial cells (CP-M063, Procell Life Science & Technology, China) were seeded in the lower chamber of a Transwell system to achieve 30% confluence. Treated RAW 264.7 cells were counted using a cell counter and seeded in the upper chamber at 2 × 10⁶ cells per well. Co-cultures were maintained for 72 h before proceeding with further experiments.

Enzyme-linked immunosorbent assay procedure and reagents

The ELISA was performed according to the kit instructions. Kits used included HIF-1α (YX-080906H), SMAD2 (YX-191301H), and VEGF (YX-220507H) from Shanghai Yuanxin Biotechnology (China), as well as IL-10 (SEA056Ra), TNF-α (SEA133Ra), IL-6 (SEA079Ra), and NGAL (SEB388Ra) from Wuhan USCN Business Co., Ltd (China).

Immunofluorescence staining

Tissue sections were deparaffinized, rehydrated, and underwent antigen retrieval using microwave heating with pH 8.0 EDTA. Blocking of non-specific sites was achieved with 3% BSA. Primary antibodies, including CD68 (GB13430, Servicebio, China), CD206 (GB13438, Servicebio, China) and CD86 (P80609, MedChemExpress, USA) were incubated overnight at 4 °C. Secondary antibodies were applied for 50 min, followed by FITC application and autofluorescence quenching. Nuclei were counterstained with DAPI.

Western blot analysis

Protein concentrations were determined using the bicinchoninic acid assay. Membranes were blocked and incubated with primary antibodies, including HIF-1α (H6536, Sigma-Aldrich, USA), VEGFA (GB13034, Servicebio, China), SMAD2 (12570-1-AP, Proteintech, USA), VE-Cadherin (23195, Cell signaling, USA), and Vimentin (5741, Cell signaling, USA) overnight at 4 °C. Secondary antibodies were applied, and detection was performed using enhanced Chemiluminescence.

Real-time quantitative polymerase chain reaction

Macrophages were trypsinized, centrifuged, and total RNA was extracted with Trizol. RNA integrity was confirmed, and cDNA was synthesized. qPCR was performed using SYBR Green, following the manufacturer’s instructions, with specificity confirmed by melt curve analysis. mRNA expression was calculated using the 2^−ΔΔCt method [28]. Primers used for qPCR validation are listed in Table S1.

Flow cytometry analysis of macrophage phenotype

Hypoxia-treated macrophages were rinsed with phosphate buffer saline, trypsinized, and centrifuged. Cells were incubated with PE-CD86 (11-0862-82, Thermo Fisher Scientific, USA) and Alexa Fluor^® 488-CD206 antibodies (12-2061-82, Thermo Fisher Scientific, USA) at 4 °C. After washing, cells were analyzed by flow cytometry for CD86 and CD206 expression.

Statistical analysis

The Shapiro-Wilk test assessed data normality. Normally distributed variables are shown as mean ± SD and compared using t-tests or ANOVA with Scheffe’s test. Non-normal variables are shown as median [M (P25, P75)] and analyzed with Mann-Whitney U or Kruskal-Wallis tests. Categorical variables are ratios, evaluated with chi-square or Fisher’s exact test. Spearman correlation assessed continuous variable relationships. Analyses used SPSS 25.0; significance at p < 0.05. Image processing used Adobe Illustrator (v.26.4.1), Image-Pro Plus 6.0, and FlowJo Version 10 (FlowJo LLC). Mean optical density was calculated as integrated optical density divided by the positive staining area, using three randomly selected regions.

Results

Identification of SMAD2 as a key target in acute kidney injury and its interaction with the HIF-1α/VEGF pathway

Six articles were analyzed (Figure 2(A)) [29–34]. We identified six differentially expressed miRNAs: two up-regulated and four down-regulated. Study and miRNA details are in Table 1. Among 681 target genes, miR-195-5p had the most targets (n = 332) (Figure 2(B)). GO analysis showed differentially expressed genes (DEGs) were enriched in negative regulation of cell differentiation (BP), transcription regulator complex (CC), and transcription factor binding (MF) (Figure 2(C)). KEGG analysis identified 16 pathways linked to kidney injury (p < 0.05), with the top five shown in Figure 2(D). The top five KEGG pathways’ networks were constructed using Cytoscape, identifying hub genes like SMAD2, Smad3, and AKT3 (Figure 2(E)). SMAD2 is regulated by miR-204, miR-668, and miR-27a-3p; SMAD3 by miR-204 and miR-195-5p; AKT3 by miR-204 and miR-195-5p.

Figure 2.

Analysis of key target genes identifying kidney injury. (A) Flowchart illustrating the study selection process from database search to inclusion in the analysis. (B) Number of target genes corresponding to kidney injury-related microRNAs. (C) Gene Ontology (GO) analysis of differentially expressed genes, showing enrichment in biological processes, cellular components, and molecular functions. (D) Top five significantly enriched pathways associated with kidney injury risk. (E) Network depicting the relationship between kidney injury risk pathways and enriched genes from KEGG analysis. Orange nodes represent the five risk pathways: A (hsa04550), B (hsa04068), C (hsa05166), D (hsa05161), and E (hsa04933). (F) Box plot showing sample distribution (Group 1: control; Group 2: kidney injury). (G) Principal component analysis (PCA) plot illustrating sample clustering (Group 1: control; Group 2: kidney injury). (H) Volcano plot of differential gene expression.

Table 1.

Characteristics of microRNAs associated with kidney injury.

Study (PMID)	Country	Sample size (case/control)	Tissue	miRNA	Expression	Location on human chromosome
34240756	China	129/100	Serum	miR-204	Down	chr9: 70809975–70810084 [−]
33231561	China	100/100	Serum	miR-532–3p	Up	chrX: 50003148–50003238 [+]
30325740	China	20/22	Serum	miR-668	Up	chr14: 101055258–101055323 [+]
32492657	China	80/80	Serum	miR-195-5p	Down	chr17: 7017615–7017701 [−]
33175458	China	25/25	Serum	miR-129-5p	Down	chr7: 128207872–128207943 [+] chr11: 43581394–43581483 [+]
31090110	China	18/18	Serum	miR-27a-3p	Down	chr19: 13836440–13836517 [−]

Gene expression data (GSE30718) from GEO verified three hub genes associated with AKI. SMAD2 was significantly up-regulated (logFC = 1.55, p < 0.01) (Figures 2(F–H)). HIF-1α binds to the SMAD2 promoter, facilitating its transcription, and SMAD2 signaling may activate VEGF [35–37]. Our analysis showed interactions among HIF-1α, SMAD2, and VEGF in the hsa05200 pathway (Pathways in cancer) (Figures S4). Therefore, we aim to investigate the roles and interrelationships of HIF-1α, SMAD2, and VEGF in the progression of renal injury.

Association of renal injury severity with HIF-1α, SMAD2, and VEGF expression in pediatric AKI patients

In this study, 60 pediatric patients (median age 6.6 months, 60% male) were included. No fatalities occurred. Sepsis affected 80%, with higher prevalence in the AKI group (27 vs. 21, p = 0.03). The AKI group had higher PRISM III scores, lower PCIS scores, more mechanical ventilation and CRRT, and longer PICU stays (Table S2).

The AKI group showed higher serum creatinine for 48 h and elevated IL-6, TNF-α, and IL-10. No significant differences in RRI or urinary oxygenation indicators were found (Table 2). Within 48 h of PICU admission, HIF-1α levels were higher in the AKI group, while VEGF and SMAD2 were higher in the Non-AKI group (Figures 3(A–C)). At the time of admission, HIF-1α levels were inversely correlated with VEGF (R = −0.581, p < 0.001) and SMAD2 (R = −0.546, p < 0.001), while VEGF and SMAD2 exhibited a strong positive correlation (R = 0.751, p < 0.001) (Figures 3(D–F)).

Table 2.

Comparison of clinical indicators within 48 h of PICU admission between the AKI group and the non-AKI group.

	Patients (n = 60)	Non-AKI (n = 30)	AKI (n = 30)	p-Value
Kidney function
SCr (μmol/L)
Admission	35.0 (20.8, 69.9)	25.5 (18.6, 34.3)	69.9 (41.9, 112.2)	0.019
24 h	31.6 (20.3, 53.9)	20.8 (18.9, 24.3)	53.9 (43.3, 60.6)	0.003
48 h	28.6 (22.1, 48.5)	24.3 (18.3, 28)	48.5 (42.1, 52.3)	0.005
eGFR (mL/min/1.73 m2)
Admission	109.0 (61.6, 155.0)	135.5 (120, 177.9)	61.6 (35.8, 97.2)	0.007
24 h	122.5 (76.1, 163.7)	163.7 (124.9, 188.7)	76.1 (59.2, 88.9)	<0.001
48 h	121.3 (85.4, 161.3)	149.4 (129, 178.2)	85.4 (58.6, 103.0)	0.002
NGAL (ng/mL)
Admission	43.8 (13.8, 248.2)	13.8 (8, 41.2)	248.2 (108.1, 1,044.9)	<0.001
24 h	27.2 (7.8, 265.1)	11.3 (3.7, 14.7)	265.1 (106.0, 545.0)	<0.001
48 h	24.7 (6.4, 105.5)	8 (3.5, 10.8)	105.5 (37.9, 220.2)	0.002
Cytokines
IL-6 (pg/mL)
Admission	48.7 (6.9, 218.0)	20.4 (5.8, 241)	71.8 (7.9, 206.0)	0.853
24 h	12.8 (6.5, 30.9)	11 (4, 20.9)	18.9 (9.5, 47.9)	0.218
48 h	10.2 (4.8, 25.6)	7.7 (2.9, 13.4)	16.3 (8.0, 46.0)	0.089
TNF-α (pg/mL)
Admission	17.2 (10.3, 49.7)	14.3 (7.4, 17.7)	35.3 (10.4, 69.8)	0.165
24 h	13.6 (7.8, 29.0)	17 (10.9, 21.8)	9.6 (6.1, 38.9)	0.684
48 h	17.3 (8.6, 36.0)	15.6 (9, 18.5)	23.0 (8.3, 43.4)	0.481
IL-10 (pg/mL)
Admission	35.4 (8.2, 100.2)	10.4 (5, 34)	61.1 (36.7, 398.0)	0.011
24 h	13.0 (5.2, 45.9)	5.7 (5, 40.2)	14.8 (8.5, 51.6)	0.190
48 h	8.2 (5.0, 23.2)	5.5 (4, 9.6)	10.5 (7.3, 50.0)	0.023
Renal perfusion index
RRI (left kidney)
Admission	0.6 (0.6, 0.7)	0.6 (0.6, 0.7)	0.6 (0.5, 0.7)	0.684
24 h	0.6 (0.6, 0.7)	0.6 (0.6, 0.7)	0.6 (0.6, 0.7)	0.684
48 h	0.6 (0.5, 0.7)	0.6 (0.6, 0.7)	0.6 (0.5, 0.7)	0.123
RRI (right kidney)
Admission	0.6 (0.6, 0.7)	0.6 (0.6, 0.7)	0.7 (0.6, 0.8)	0.529
24 h	0.6 (0.6, 0.7)	0.7 (0.6, 0.7)	0.6 (0.5, 0.7)	0.315
48 h	0.6 (0.6, 0.7)	0.6 (0.6, 0.7)	0.7 (0.6, 0.7)	0.481
Urinary oxygen index
Urinary lactate (mmol/L)
Admission	0.4 (0.2, 0.7)	0.4 (0.1, 0.7)	0.4 (0.4, 0.7)	0.436
24 h	0.3 (0.1, 0.5)	0.3 (0.1, 0.4)	0.4 (0.1, 0.5)	0.481
48 h	0.2 (0.2, 0.5)	0.2 (0.1, 0.3)	0.3 (0.2, 1.0)	0.105
PuO2 (mmHg)
Admission	148.0 (134.0, 160.5)	143.5 (110, 156)	150.5 (141.0, 167.0)	0.190
24 h	154.0 (147.5, 160.5)	150 (146, 154)	159.0 (148.0, 161.0)	0.089
48 h	141.5 (125.5, 153.5)	150.5 (135, 154)	129.5 (91.7, 142.0)	0.143
PuCO2 (mmHg)
Admission	24.7 (15.3, 51.8)	32.4 (14.7, 57.9)	19.8 (17.0, 36.4)	0.796
24 h	29.7 (21.6, 37.4)	29.7 (21.8, 38.7)	29.7 (16.4, 35.5)	0.579
48 h	32.3 (22.3, 58.3)	32.3 (23.1, 46.2)	32.6 (20.0, 70.7)	0.853

PICU: pediatric intensive care unit; AKI: acute kidney injury; SCr: serum creatinine; eGFR: estimate glomerular filtration rate; NGAL: neutrophil gelatinase-associated lipocalin; IL-6: interleukin-6; TNF-α: tumor necrosis factor; IL-10: interleukin-10; RRI: renal resistive index; PuO₂: urinary partial pressure of oxygen; PuCO₂: urinary partial pressure of carbon dioxide.

Figure 3.

Dynamic interplay of HIF-1α, VEGF, and SMAD2 in pediatric AKI. Comparison of expression levels of HIF-1α (A), VEGF (B), and SMAD2 (C) between the AKI and Non-AKI groups at different time points. *p < 0.05. Correlation analysis of HIF-1α with SMAD2 (D), SMAD2 with VEGF (E), and HIF-1α with VEGF (F) within 48 h. (G) Correlation analysis of HIF-1α, VEGF, and SMAD2 with clinical indicators in children admitted to the PICU. Positive correlations are depicted in red, while negative correlations are shown in blue, with color intensity proportional to the Spearman correlation coefficient. Statistical significance, *p < 0.05. SCr: serum creatinine; eGFR: estimated glomerular filtration rate; NGAL: neutrophil gelatinase-associated lipocalin; IL-6: interleukin-6; TNF-α: tumor necrosis factor-alpha; IL-10: interleukin-10; RRI: renal resistive index; PuO₂: urinary partial pressure of oxygen; PuCO₂: urinary partial pressure of carbon dioxide; VEGF: vascular endothelial growth factor; HIF-1α: hypoxia-inducible factors; AKI: acute kidney injury.

HIF-1α correlated positively with kidney injury markers (SCr, NGAL) and negatively with eGFR. SMAD2 and VEGF showed opposite trends. HIF-1α negatively correlated with IL-10, an anti-inflammatory cytokine, while SMAD2 was positively correlated (Figure 3(G)). Sustained HIF-1α activation is associated with markers of renal injury, while elevated VEGF expression appears to be linked to repair processes. SMAD2 exhibits context-dependent effects on renal pathophysiology, requiring nuanced analysis across injury phases.

Association of renal HIF-1α, SMAD2, and VEGF expression in a chronic hypoxia rat model

Given that circulating HIF-1α and SMAD2 are typically undetectable except during cellular injury, we employed tissue-specific analysis in animal models to assess their expression levels in renal tissue. All rats survived to the end of the study (n = 5 per group). On day 14, kidney weight analysis revealed a reduction with decreasing oxygen levels, which was further accentuated by bevacizumab administration (Figures 4(A,B)). In Sprague-Dawley rats, arterial blood gas parameters were assessed under varying oxygen levels. As oxygen decreased, PaO₂ levels significantly declined in both the 7%O₂ + NS and 7%O₂ + Bev groups compared to the 21%O₂ + NS group (Table 3).

Figure 4.

Renal injury assessment in a chronic hypoxia rat model with bevacizumab treatment. (A) Right kidney weight. (B) Left kidney weight. (C) BUN levels. (D) NGAL levels. (E) SCr levels. (F) Macroscopic kidney morphology. (G) H&E staining (×63). (H) Renal injury scores. (I) PAS staining (×40). (J) Masson staining (×63). Compared to 21%O₂ + NS, ^ap < 0.05; compared to 21%O₂ + Bev, ^bp < 0.05. BUN: blood urea nitrogen; SCr: serum creatinine; NGAL: neutrophil gelatinase-associated lipocalin.

Table 3.

Comparison of arterial blood gases in rats across different hypoxic environments.

	21%O2 + NS group	21%O2 + Bev group	7%O2 + NS group	7%O2 + Bev group
PH	7.3 (7.3, 7.4)	7.3 (7.3, 7.4)	7.3 (7.2, 7.3)	7.3 (7.0, 7.3)
PaCO2 (mmHg)	30.3 (24.4, 36.7)	30.8 (26.1, 40.4)	31.9 (30.4, 38.4)	28.0 (25.1, 68.9)
PaO2 (mmHg)	135.0 (117.0, 144.0)	131.0 (130.0, 131.0)	78.0 (51.0, 84.0)ab	77.0 (69.0, 97.0)ab
BEecf (mmol/L)	−11.0 (−12.0, −10.0)	−7.0 (−13.0, −7.0)	−10.0 (−14.0, −6.0)	−13.0 (−16.0, −11.0)
HCO3 (mmol/L)	15.6 (13.6, 17.1)	17.9 (13.3, 19.5)	16.1 (13.7, 19.7)	14.9 (12.9, 15.5)

PaO₂: arterial partial pressure of oxygen; PaCO₂: arterial partial pressure of carbon dioxide; BEecf: base excess of extracellular fluid.

Compared with 21%O₂ + NS group, ^ap < 0.05; compared with 21%O₂ + Bev group, ^bp < 0.05.

Serum analysis revealed that the 7%O₂ + Bev group had the highest levels of BUN, NGAL, and SCr among all groups (Figures 4(C–E)). Macroscopic examination revealed well-formed kidneys in the 21%O₂ + NS group, while 7%O₂ + NS and 7%O₂ + Bev kidneys appeared dark red, with subcapsular hemorrhages observed on day 14 (Figure 4(F)). The 7%O₂ + Bev group exhibited severe renal injury on H&E, PAS, and Masson staining, with higher injury scores, demonstrating that hypoxia exacerbates renal damage and bevacizumab further aggravates it (Figures 4(G–J)). Western blot analysis showed increased HIF-1α and SMAD2 levels in hypoxia and bevacizumab-treated groups, with the highest in 7%O₂ + Bev (Figures 5(A–F)). Serum TNF-α and IL-6 were elevated in hypoxia groups (7%O₂ + NS, 7%O₂ + Bev) and higher in bevacizumab administration groups (Table 4). Renal tissue analysis confirmed that elevated HIF-1α levels correlated with renal injury, while VEGF appeared renoprotective, as its reduction exacerbated injury. SMAD2 expression exhibited a similar trend to HIF-1α.

Figure 5.

Expression of HIF-1α, SMAD2, VEGF, and M2 macrophage markers in rat kidneys under hypoxia and bevacizumab treatment. (A) Western blot images of HIF-1α in rat kidneys. (B) Quantitative analysis of HIF-1α expression. (C) Western blot images of SMAD2 in rat kidney tissue. (D) Quantitative analysis of SMAD2 expression. (E) Western blot images of VEGF in rat kidney tissue. (F) Quantitative analysis of VEGF expression. (G) Comparison of MΦ-2 expression (CD68⁺/CD206⁺) in rat kidneys at ×40 magnification. Nuclei are labeled with blue fluorescence, CD68 with green, and CD206 with red. (H) Comparison of average optical density values of renal MΦ-2 (CD68⁺/CD206⁺). Compared with 21%O₂ + NS group, ^ap < 0.05; compared with 21%O₂ + Bev group, ^bp < 0.05. HIF-1α: hypoxia-inducible factor; VEGF: vascular endothelial growth factor; mOD: mean optical density, calculated as integrated optical density/area.

Table 4.

Comparison of cytokines levels among different groups of rats.

	IL-10 (pg/mL)	TNF-α (pg/mL)	IL-6 (pg/mL)
21%O2 + NS group	16.1 ± 4.6	5.8 ± 1.4	8.3 ± 2.5
21%O2 + Bev group	13.6 ± 3.8	8.0 ± 1.8	12.4 ± 3.7
7%O2 + NS group	33.0 ± 6.2b	12.0 ± 3.8	15.6 ± 5.2
7%O2 + Bev group	32.0 ± 3.2b	16.0 ± 4.5a	22.2 ± 6.5a

IL-6: interleukin-6; TNF-α: tumor necrosis factor; IL-10: interleukin-10.

Compared with 21%O₂ + NS group, ^ap < 0.05; compared with 21%O₂ + Bev group, ^bp < 0.05.

Immunofluorescence double staining was used to assess the polarization of M1 macrophages (CD86⁺/CD68⁺) and M2 macrophages (CD206⁺/CD68⁺) (Figure S5, Figures 5(G,H)). Results indicated that hypoxia increased the expression of both M1 (CD86⁺/CD68⁺) and M2 (CD206⁺/CD68⁺) macrophages. However, when combined with bevacizumab, hypoxia further increased M1 expression while reducing M2 expression.

Induction of SMAD2 and VEGF expression promotes macrophage M2 polarization

Considering that the chronic hypoxia rat model exerts complex effects on the systemic immune system, we conducted in vitro macrophage experiments under hypoxic conditions to specifically investigate the direct regulatory effects of hypoxia on macrophage polarization. Based on preliminary findings (Figure S2), we set the oxygen concentration for cell culture at 15%. As hypoxia alone failed to induce macrophage polarization in pilot experiments, IL-4 was employed to prime polarization [38], followed by evaluation of hypoxia’s effects on macrophages.

Macrophage polarization was assessed by qPCR and flow cytometry. Hypoxia alone (15% O₂) did not induce polarization (Figures 6(A–C)), and preliminary experiments showed no polarization in macrophages cultured under varying oxygen levels (21, 15, 10, and 5%) for 48 h (Figure S6). In further experiments, we found IL-4 induced M2 macrophage polarization, which was further enhanced under hypoxic conditions (Figures 6(A–C)). We also evaluated the correlation between iNOS and Arg-1 mRNA levels and those of IL-6, TNF-α, and IL-10 by qPCR (Figure S7).

Figure 6.

Analysis of macrophage polarization and HIF-1α, VEGF, and SMAD2 expression. (A) iNOS expression by qPCR. (B) Arg-1 expression by qPCR. (C) Flow cytometry of macrophage phenotype conversion. qPCR analysis of HIF-1α (D), SMAD2 (E), and VEGF (F) expression under various conditions. Compared with 21%O₂ + NS group, ^ap < 0.05; compared with 21%O₂ + Bev group, ^bp < 0.05. Compared with 15%O₂ + IL-4 group, ^dp < 0.05; compared with 15%O₂ + IL-4 + OC group, ^ep < 0.05. HIF-1α: hypoxia-inducible factor; VEGF: vascular endothelial growth factor; iNOS: inducible nitric oxide synthase.

Hypoxia alone (15% O₂) or IL-4 treatment increased HIF-1α and VEGF qPCR levels, with VEGF showing a more pronounced rise, while SMAD2 levels remained unchanged. Combined hypoxia and IL-4 appeared to elevate SMAD2 expression (Figures 6(D–F)). To further explore the roles of VEGF and SMAD2 in macrophage polarization, we overexpressed and knocked down both factors, revealing that high VEGF levels strongly correlated with M2 polarization, while SMAD2 exhibited a weaker pro-M2 effect (Figures 6(A–C)). Co-culture of RAW 264.7 cells with mouse glomerular endothelial cells revealed that hypoxia promoted VE-cadherin expression, further enhanced by IL-4 and VEGF overexpression. Conversely, SMAD2 exhibited a modest negative correlation with vimentin, indicating a potential role in mesenchymal transition (Figure S8).

Discussion

The mechanisms of kidney injury and repair involve complex interactions among various pathways and effector cells, prominently featuring hypoxia and inflammatory responses. Our findings suggest that hypoxia-induced dysregulation of HIF-1α, SMAD2, and VEGF expression may collectively contribute to macrophage phenotype transitions, as evidenced by both in vivo and in vitro experimental models.

This study identified SMAD2 as a key target in kidney injury repair through bioinformatics and literature review. SMAD2 influences VEGF expression directly or via the TGF-β1 pathway, promoting angiogenesis and macrophage polarization from M1 to M2, which aids in tissue remodeling [10,13,14]. In our study, serum analysis in AKI patients within 48 h of PICU admission revealed an inverse correlation was identified between HIF-1α and both VEGF and SMAD2, while a direct positive correlation was established between VEGF and SMAD2. Hypoxia is recognized for its role in modulating HIF-1α expression, with a direct correlation to the intensity and duration of the hypoxic condition [39,40]. In contrast, VEGF expression is not directly regulated by hypoxia. The divergent expression profiles of HIF-1α and VEGF under conditions of severe hypoxia may be attributed to multiple factors. In our study, HIF-1α was elevated in AKI, supporting the hypoxia hypothesis. However, VEGF did not initially increase with HIF-1α; they were inversely correlated on days 0 and 1, but positively correlated by day 2. This aligns with Becherirat et al.’s findings of a complex HIF-1α-VEGF relationship, where hypoxia increases HIF-1α, with delayed VEGF response promoting angiogenesis and reducing hypoxia [41].

In pediatric AKI patients, HIF-1α and VEGF expression levels varied over time. To investigate this, we established a chronic hypoxia rat model, chosen over ischemia-reperfusion injury (IRI) due to the prevalence of pulmonary disease (50%) and mechanical ventilation (35%) in our cohort, suggesting chronic hypoxia’s role in disease progression. Additionally, IRI may introduce confounding inflammation [42]. However, renal fibrosis observed in the chronic hypoxia model may not fully reflect acute kidney injury, potentially linked to inflammatory responses. Our study showed that elevated HIF-1α expression correlates with increased proinflammatory cytokines (IL-6, TNF-α), consistent with its established role in binding the IL-6 promoter to enhance transcriptional activity [43]. Notably, HIF-1α also coordinates the release of other mediators, including IL-1β, suggesting its central role in inflammatory amplification [44]. However, it is important to note that in the early stages of renal injury, tubular epithelial cells are primarily affected, and HIF-1α is widely expressed in these cells. With prolonged hypoxia, HIF-2α directly binds to the VEGF promoter to regulate its expression, particularly under chronic hypoxia. While this study did not focus on HIF-2α, its role warrants further investigation [3,4]. In our study, SMAD2 and VEGF showed positive associations with the anti-inflammatory cytokine IL-10. Experimental evidence indicates IL-10 maintains immune homeostasis by activating SMAD2 to suppress TLR2-mediated NF-κB signaling, as evidenced by phospho-SMAD2 reduction in IL-10-deficient intestinal epithelial cells [45]. While direct IL-10 regulation of VEGF remains unproven, clinical analyses revealed significant IL-10/VEGF correlations (r = 0.579), potentially mediated through plasma cell proliferation and indirect angiogenic modulation.

In the chronic hypoxia rat model, we determined the minimum tolerable oxygen concentration to be 7% through preliminary experiments. In contrast, RAW 264.7 cells could not survive 7% O₂ for 48 h, consistent with Gao et al.’s report of impaired mitochondrial respiration in macrophages below 10% O₂ [46]. The discrepancy likely stems from the complex in vivo oxygen delivery system, involving blood flow and adaptive responses, which is absent in simplified in vitro models that rely solely on passive diffusion [47]. Based on preliminary experiments, we set oxygen concentrations at 7% for animal experiments and 15% for cell experiments. Despite the difference, both models demonstrated hypoxia-induced changes in HIF-1α, VEGF, and SMAD2 expression, as well as macrophage function, suggesting comparable biological effects. In the chronic hypoxia rat model, we observed that hypoxia stimulated macrophage activation in the kidney, aligning with Mazzali et al.’s observation of macrophage infiltration in rats exposed to 10% O₂, detectable within 6 h and increasing over time [48]. However, systemic hypoxia likely impacts the entire immune system, including circulating immune mediators, bone marrow-derived immune cells, and interactions within the renal microenvironment, beyond renal macrophages alone [49,50]. To address this limitation and isolate the direct effects of hypoxia on macrophages, we conducted in vitro experiments focusing specifically on macrophage polarization under controlled hypoxic conditions.

Macrophages exhibit remarkable plasticity in diverse microenvironments, with MΦ-1 associated with renal injury and MΦ-2 linked to tissue repair. Our study revealed that hypoxia alone did not induce significant macrophage polarization, but enhanced MΦ-2 polarization with IL-4 induction, aligning with previous findings [51]. In our study, we observed that the overexpression of SMAD2 and VEGF may facilitate the polarization of MΦ-2 macrophages. This is consistent with previous findings showing that VEGF can promote M2 polarization [52]. Existing research has also demonstrated that the knockdown of VEGFR-1 can suppresses the expression of monocyte chemoattractant protein-1, thereby diminishing macrophage infiltration [53]. However, it’s important to note that the VEGF family’s effects on macrophage polarization and kidney injury are context-dependent [54]. The interplay between VEGF signaling and other pathways, such as STAT3, further complicates the picture. STAT3 has been shown to influence macrophage polarization through various signaling pathways, including JAK/STAT, NF-κB, and PI3K/AKT [55]. These pathways can interact with VEGF signaling to modulate macrophage phenotype and function.

This study has several limitations. First, the small sample size of 60 pediatric patients may limit the generalizability of our findings, necessitating validation in larger cohorts. Second, during the hypoxic exposure of rats, necessary procedures conducted in normoxic conditions (lasting ∼5 min) could have affected oxygen control and experimental outcomes. Third, in vitro studies involved interventions, such as hypoxia, IL-4, and SMAD2, VEGF plasmid transfection. These interventions may interact synergistically or antagonistically, possibly activating other stress response pathways impacting cell proliferation, migration, and other functions. Further comprehensive evaluations of macrophages are required to elucidate these interactions.

In conclusion, this study elucidates the intricate interplay between HIF-1α/VEGF signaling and SMAD2 in kidney injury, demonstrating that inflammation and hypoxia synergistically contribute to renal damage. Hypoxia alone does not induce MΦ-2 polarization but enhances IL-4 effects, further amplified by SMAD2 and VEGF overexpression. These findings highlight the intricate regulatory roles of macrophages in renal pathology under hypoxic conditions, offering valuable insights into the treatment and management.

Funding Statement

This work was supported by the Youth Project of the National Natural Science Foundation of China (82200754) and the Xinhua Hospital Development Fund. The author acknowledges with gratitude the financial support from the K. C. Wong Medical Fellowship Fund at Shanghai Jiao Tong University.

Ethical approval

This study strictly adhered to ethical standards, encompassing both human and animal research components. The human study was approved by the ethics committee (Approval No. XHEC-C-2022-012-1) and registered with a clinical trial registry (Registration No. NCT06197828). The animal experiments were conducted in strict accordance with the guidelines of the Institutional Animal Care and Use Committee of Xinhua Hospital Affiliated with Shanghai Jiao Tong University School of Medicine (Shanghai, China), and the experimental protocol was approved by the Ethics Committee of the same institution (Approval No. XHEC-F-2021-001).

Author contributions

Study design: Yaya Xu and Jiayue Xu. Data collection: Jiru Li and Xiangmei Kong. Methodology development and verification: Haoyun Mao and Zhiyi Du. Data visualization and statistical analysis: Jiahui Zan and Lili Xu. Data interpretation: Wen Qian and Zhushengying Ma. Manuscript preparation: Yaya Xu, Wen Qian, Zhushengying Ma, and Lili Xu. Supervision, final approval, and senior authorship: Yueniu Zhu and Xiaodong Zhu.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Additional data not included in the repository can be obtained from the corresponding author upon reasonable request.