Background

Osteoarthritis (OA) is the most common form of arthritis and a leading cause of disability worldwide. This disability is largely due to pain, a major symptom of the condition. Pain contributes to functional limitations and reduces quality of life [1,2,3,4]. Largely because of pain, lower extremity OA is well recognized as the leading cause of mobility impairment in older adults [5]. Pharmacologic treatment options for OA are centered around the relief of pain and support for functional improvement in patients. Nevertheless, the efficacy of treatments such as nonsteroidal anti-inflammatory drugs (NSAIDs) can be limited, and can cause significant adverse effects such as cardiorenal and gastrointestinal toxicity [6, 7]. It is therefore important to establish the mechanisms underlying OA pain to aid in drug development for OA treatment.

Vascular endothelial growth factor (VEGF) is a potent stimulator of angiogenesis, and also a contributor to inflammation. VEGF in OA patients has been found to be elevated in the synovial membrane (SM) [8,9,10,11], subchondral bone [12,13,14], synovial fluid (SF) [15,16,17,18], serum [16,17,18,19], and articular cartilage [20,21,22,23,24,25,26,27,28]. In particular, VEGF is strongly expressed in synovial lining cells in OA patients [29]. Intraarticular injection of anti-VEGF antibody reduced synovial inflammation in a rabbit OA model [30]. Several studies have suggested that SF and plasma VEGF concentrations in OA patients correlate with OA severity [16, 17]. In addition, recent studies have reported that VEGF contributes to pain in a rodent neuropathic pain model [31,32,33,34] and cancer pain model [35]. The exact relationship between VEGF expression in SM and pain in human KOA patients is not fully understood.

Recent studies have suggested that several neuropeptides, such as calcitonin gene-related peptide (CGRP) and nerve growth factor (NGF), in the SM are involved in the OA pain pathway [36,37,38,39]. Apelin is a recognized member of the adipose-secreted cytokine family and is initially secreted as a pre-propeptide of 77 amino acid residues, which is then cleaved into a number of active forms [40]. The apelin signaling pathway was shown to play a major role in the development of the functional vascular network [41] and apelin expression was elevated in endothelial cell culture following VEGF stimulation [42]. Several studies also showed that apelin can regulate peripheral pain sensitivity mediated by apelin receptors (APJ) and GABAA receptors [21, 43]. Hu et al. reported that apelin concentration in SF is increased in OA patients [44]. These observations led us to investigate the role of VEGF in regulating apelin in SM and its contribution to the OA pain pathway.

We investigated the relationship between VEGF expression in SM and pain in knee osteoarthritis (KOA) patients. In addition, we investigated whether VEGF regulates apelin expression in the SM.

Methods

Patients

This study was approved by the Institutional Review Board for Clinical Research and Treatment in Kitasato University (approval No. B13–113). Sample size was determined with a power analysis for an alpha of 0.05 and power of 0.80 using G*POWER3. Patients scored their pain on a 0 to 10 cm visual analog scale (VAS). Power analysis showed that 44 SM samples of patients with VAS < 6 and 58 SM samples of patients with VAS ≥ 6 were required to identify a difference in VEGF expression between the two groups. SM samples were harvested from 102 patients undergoing total knee arthroplasty. The study enrolled 22 men and 80 women (age 46–89 years, mean ± standard error (SE) = 73.2 ± 0.8 years; body mass index (BMI) range 18.4–36.7, mean ± SE = 26.0 ± 0.4 kg/m2) with radiographic evidence of KOA (unilateral Kellgren/Lawrence [K/L] grades 2 (3/102, 2.9%), 3 (36/102, 35.3%) and 4 (63/102, 61.8%)). All patients provided informed consent for participation in this study 1 day before surgery. SM samples were harvested intraoperatively from the suprapatellar pouch of each operated knee and immediately stored frozen in liquid nitrogen at − 80 °C until required for extraction of RNA. SM samples obtained from six patients were used for cell culture. The remaining samples were fixed in 4% paraformaldehyde phosphate-buffered solution (Nacalai Tesque, Kyoto, Japan) for 72 h for use in histological analysis.

Quantitative polymerase chain reaction (qPCR) analysis

Extraction of total RNA from SM and cultured SM cells and cDNA synthesis was conducted as reported previously [45]. PCR primer pair sequences for use in qPCR analysis were: VEGF-Forward (5′- TTGCCTTGCTGCTCTACCTC-3′) and VEGF-Reverse (5′- AGCTGCGCTGATAGACATCC-3′) for VEGF amplification (product size: 117 bp); apelin-Forward (5′- GAATCTGCGGCTCTGCGT-3′) and apelin-Reverse (5′- CATCAGGGACCCTCCACACA-3′) for apelin amplification (product size: 76 bp); and GAPDH-Forward (5′-TGTTGCCATCAATGACCCCTT-3′) and GAPDH-Reverse (5′-CTCCACGACGTACTCAGCG-3′) for GAPDH amplification (product size: 223 bp). Specificity of the qPCR products was evaluated using melting curve analysis. Relative mRNA expression levels of VEGF and apelin were evaluated using qPCR (CFX-96®, Bio-Rad, Richmond CA, USA). Expression levels of VEGF and apelin mRNA were normalized to the expression of the housekeeping gene, GAPDH.

Expression levels of VEGF mRNA were compared between the strong/severe (VAS ≥ 6) and mild/moderate pain (VAS < 6) groups (Table 1), using VAS = 6 as a cutoff based on previous reports [39, 46, 47]. The correlation between VAS levels and VEGF mRNA expression was also determined. Relative VEGF expression was calculated based on the mean of all samples of the VAS < 6 group.

Table 1 Characteristics of patients in VAS < 6 and VAS ≥ 6 groups
Full size table

To investigate the relationship between VEGF and K/L grades, the 102 knee OA patients were divided into three groups based on their K/L grade (2, 3, or 4). The clinical characteristics of patients in each of these groups are shown in Table 2. Relative VEGF expression was calculated based on the mean of all samples of the K/L2 group.

Table 2 Clinical characteristics of patients (K/L 2, 3 and 4)
Full size table

Immunohistochemistry

Following fixation, SM samples were embedded in paraffin, sectioned at 3 μm thickness, then deparaffinized (Clear Plus®, FALMA, Tokyo, Japan) and pretreated with sodium citrate buffer (pH 6.0) containing 0.1% polyoxyethylene sorbitan monolaurate (Nacalai Tesque, Kyoto, Japan) at 98 °C for 20 min for antigen retrieval. The sections were subsequently washed three times with phosphate-buffered saline for 5 min and incubated with rabbit polyclonal anti-VEGF antibody (1:100 dilution; Santa Cruz Biotechnology Inc., Santa Cruz CA, USA) and mouse monoclonal anti-apelin antibody (1:100 dilution; Santa Cruz Biotechnology) for 4 h at 4 °C. The sections were additionally incubated with Alexa 488 Fluor®-conjugated goat anti-rabbit IgG antibody (1:100 dilution; Thermo Fisher Scientific, Waltham MA, USA) and Alexa 594 Fluor®-conjugated goat anti-mouse IgG antibody (1:100 dilution; Thermo Fisher Scientific) for 1 h at room temperature. The distribution of fluorescence in SM sections was analyzed using a fluorescence microscope (Axiovert 200®, Zeiss, Jena, Germany).

Synovial membrane cell culture

Synovial membrane cells (SMCs) were isolated from 500 mg SM using 40 mL of a 1 mg/mL collagenase solution. The SMCs were incubated in α-minimal essential media (α-MEM; Nacalai Tesque) containing 10% fetal bovine serum in six-well plates. After 1 week, the SMCs were incubated with vehicle (serum free α-MEM) or 10 or 100 ng/mL human recombinant VGEF (Biolegend, San Diego CA, USA) for 24 h. Subsequently, total mRNA and protein were extracted and used in western blotting and qPCR analysis. Relative expression was calculated based on the mean of all samples of the vehicle-treated group.

Western blotting for apelin

To investigate the regulation of apelin by VEGF, SMCs harvested from six patients were stimulated with 1 ng/mL or 10 ng/mL VEGF for 24 h. Using methodology described elsewhere [48], SMCs were lysed in radioimmunoprecipitation buffer (Thermo Fisher Scientific) containing a protease inhibitor cocktail (Sigma-Aldrich, St. Louis MO, USA). Protein concentration was determined for each cell extract using a bicinchoninic acid (BCA) assay (Thermo Fisher Scientific). A total of 30 μg of each protein was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and electrophoretically transferred to polyvinyl difluoride membranes. These membranes were then blocked with polyvinylidene fluoride (PVDF) blocking reagent (DS Pharma Biomedical, Suita, Japan) for 1 h and incubated overnight at 4 °C with mouse monoclonal primary antibody against apelin (1:200 dilution; Santa Cruz Biotechnology Inc.). The membranes were washed with Tris-buffered saline containing 0.05% Tween and then incubated with horseradish peroxidase-conjugated anti mouse IgG (1:1000 dilution; GE Healthcare, Piscataway NJ, USA). Apelin proteins were visualized by chemiluminescence using an enhanced chemiluminescence detection system (GE Healthcare) and exposure of the membranes to x-ray film. Bands were quantified by densitometric scanning using ImageJ software (NIH, Bethesda MD, USA). Densitometry levels of apelin proteins were normalized against that of β-actin.

Statistical analysis

Differences in VEGF expression between the VAS < 6 and VAS ≥ 6 groups were compared using the Mann-Whitney U-test. Differences in VEGF expression among K/L2, 3 and 4 subjects were compared using the Kruskal-Wallis test. Tukey’s multiple comparisons test was used to compare vehicle control and VEGF-treated cells. The relationship between VEGF expression and VAS was evaluated using Spearman’s correlation coefficient. All statistical analyses were conducted using SPSS software (v. 19.0; SPSS, Chicago IL, USA), with a P value < 0.05 considered statistically significant for all analyses.

Results

Relationship between VEGF expression and VAS

The VAS ≥ 6 and VAS < 6 groups did not differ with regard to patient age, male/female ratio, BMI, or KL 2/3/4 ratio (Table 1). qPCR analysis showed that VEGF expression in the SM was significantly higher in the VAS ≥ 6 group than the VAS < 6 group (Fig. 1a, P < 0.05). VEGF levels were also positively correlated with VAS (Fig. 1b, ρ = 0.346, P < 0.05).

Fig. 1
figure 1

qPCR analysis of VEGF expression in the synovial membranes (SMs) of knee osteoarthritis (KOA) patients according to their VAS score. a “Box and whisker” plot of VEGF expression in the SMs of KOA patients with mild/moderate (VAS < 6) (n = 44) and strong/severe pain (VAS ≥ 6) (n = 58). The plot shows the median values (bold horizontal line), interquartile range (IQR) (box), and range. Outliers beyond 1.5-fold of the IQR from the box are plotted. * P = 0.008. VAS, visual analog scale. b Scatter plot showing the correlation between VAS scores and VEGF levels. Relative expression was calculated based on the mean of all samples of the VAS < 6 group

Full size image

Relationship between VEGF expression and K/L grades

The three K/L grade groups did not differ with regard to patient age, male/female ratio, BMI, or VAS (Table 2). There was no difference in synovial VEGF expression among the K/L2, 3 and 4 groups (Fig. 2).

Fig. 2
figure 2

Relationship between VEGF mRNA expression level and Kellgren/Lawrence (K/L) grade. To investigate the relationship between VEGF and K/L grades, 102 knee OA patients were divided into three groups according to their K/L grade (2, 3, or 4). The three groups did not differ by patient age, male/female ratio, body mass index, or pain score based on the visual analog scale. Relative expression was calculated based on the mean of all samples of the KL2 group. The plot indicates median (bold horizontal line), interquartile range (IQR) (box), and range. Outliers beyond 1.5-fold of the IQR from the box are plotted

Full size image

Distribution of VEGF and apelin in the SMs of KOA patients

Immunohistochemical analysis was conducted to investigate the distribution of VEGF and apelin in the SMs of KOA patients (Fig. 3). Immunostaining showed that VEGF and apelin protein were both expressed in the synovial lining layers (Fig. 3).

Fig. 3
figure 3

Immunostaining of VEGF and apelin in the synovial membranes of knee OA patients. Synovial membranes stained with a DAPI (nuclei), b VEGF, or c apelin. d The merged images. Region enclosed by dotted lines indicates the lining layer (LL), which is above the sublining layer (SLL) in the synovial membrane. Scale bar = 100 μm

Full size image

Effects of VEGF on apelin expression in SMCs

qPCR analysis showed that the expression of apelin mRNA increased significantly in SMCs following 10 and 100 ng/mL VEGF stimulation (1.73-fold and 1.69-fold, respectively, P < 0.05; Fig. 4). Western blotting analysis showed that the expression of apelin was significantly increased in SMCs in the presence of exogenously added 10 and 100 ng/mL VEGF (1.85-fold and 1.56-fold, respectively, P < 0.05; Fig. 5a and b).

Fig. 4
figure 4

Effect of VEGF on apelin mRNA expression. qPCR analysis for apelin. Synovial membrane cells were stimulated with 10 or 100 ng/mL human recombinant VGEF or vehicle for 24 h preceding the extraction and analysis of apelin mRNA. All data are shown as the mean ± standard error (n = 6). *P < 0.05

Full size image
Fig. 5
figure 5

Effect of VEGF on apelin protein expression. Western blotting analysis for apelin. Synovial membrane cells were stimulated with 10 or 100 ng/mL human recombinant VGEF or vehicle for 24 h preceding protein extraction and analysis of apelin protein. a Western blot of apelin and β-actin. b Densitometric analysis of apelin bands. All data are shown as the mean ± standard error (n = 6). *P < 0.05 compared with the vehicle control

Full size image

Discussion

In the SM of KOA patients, VEGF mRNA expression in the VAS ≥ 6 group was significantly higher than that in the VAS < 6 group. VEGF and apelin were both expressed in the synovial lining layers and VEGF stimulated apelin mRNA and protein expression in SM cell culture. Together, these findings indicate that VEGF expression in SM may be involved in knee pain via apelin in KOA patients.

Several studies have reported that VEGF may contribute to chronic pain conditions [31, 32, 49]. Injection of VEGF in spinal cord-injured rats causes mechanical allodynia [49]. VEGF neutralization in rat chronic constriction injury (CCI) models attenuates chronic pain behavior by reducing the VEGF receptor expression level in dorsal root ganglia to inhibit neuropathic pain signaling [32]. Perineural injection of a VEGF inhibitor inhibited tactile allodynia and thermal hyperalgesia caused by partial sciatic nerve ligation [31]. Here, KOA patients who experienced severe/strong pain showed increased VEGF expression levels in SM. These findings suggest that synovial VEGF seems to play an important role in the pain pathway associated with KOA.

Several studies have reported that VEGF regulates apelin expression in vitro and in vivo [42, 50, 51]. VEGF stimulates apelin mRNA in human umbilical venous endothelial cells in vitro [50]. Local injection of bevacizumab, an anti-VEGF antibody, inhibits apelin expression in monkeys with occlusion of the central retinal vein [51]. Apelin-APJ systems are located in the central and peripheral nervous systems [52, 53]. In the central nervous system, apelin and its receptors have been detected in pain-associated regions. Previous studies have reported that intrathecal injection of apelin-13 (the isoform that binds most strongly to the APJ) induces hyperalgesia [21], and when “intrathecal administration of ML221, an APJ blocker, was used, this transiently reduced CCI-induced pain hypersensitivity” [43]. In addition, higher apelin concentrations in serum and SF were found in OA patients compared to non-OA patients [44]. Here, VEGF stimulated apelin mRNA and protein expression in SMCs, suggesting that further investigation of a direct link between apelin and pain may explain the mechanism underlying VEGF-induced OA pain.

A number of limitations of this study warrant mention. First, the absence of a non-KOA control patient population reduces the certainty of our results. Additional evaluations aimed at confirming whether VEGF levels are raised in the SMs of KOA patients compared to non-KOA patients are needed. Second, whether SMCs extracted from OA knees will behave the same as SMCs from healthy knees when treated with VEGF remains to be determined. Third, although our findings support the idea that altered VEGF levels in SMs are associated with KOA pain, whether a direct causative link exists between VEGF and apelin remains to be clarified. Finally, the relationship between apelin and pain in SMs was not determined.

Conclusions

Elevated VEGF expression in SMs was associated with an increase in pain in KOA patients with severe/strong pain. VEGF may regulate apelin expression in SMCs. The present findings suggest that altering the regulation of VEGF and apelin expression in SMs may represent a promising and suitable pharmaceutical strategy for the management of KOA pain.