Recombinant thrombomodulin domain 1 rescues pathological angiogenesis by inhibition of HIF-1α-VEGF pathway

Abstract

Pathological angiogenesis (PA) contributes to various ocular diseases, including age-related macular degeneration, diabetic retinopathy, and retinopathy of prematurity, which are major causes of blindness over the world. Current treatments focus on anti-vascular endothelial growth factor (VEGF) therapy, but persistent avascular retina, recurrent intravitreal neovascularization, and general adverse effects are reported. We have previously found that recombinant thrombomodulin domain 1 (rTMD1) can suppress vascular inflammation. However, the function of rTMD1 in VEGF-induced PA remains unknown. In this study, we found that rTMD1 inhibited VEGF-induced angiogenesis in vitro. In an oxygen induced retinopathy (OIR) animal model, rTMD1 treatment significantly decreased retinal neovascularization but spared normal physiological vessel growth. Furthermore, loss of TMD1 significantly promoted PA in OIR. Meanwhile, hypoxia-inducible factor-1α, the transcription factor that upregulates VEGF, was suppressed after rTMD1 treatment. The levels of interleukin-6, and intercellular adhesion molecule-1 were also significantly suppressed. In conclusion, our results indicate that rTMD1 not only has dual effects to suppress PA and inflammation in OIR, but also can be a potential HIF-1α inhibitor for clinical use. These data bring forth the possibility of rTMD1 as a novel therapeutic agent for PA.

Introduction

Pathological angiogenesis (PA) plays an important role in ocular neovascular diseases, such as age-related macular degeneration, diabetic retinopathy, and retinopathy of prematurity in newborns. Unlike normal physiological vessel development, uncontrolled PA is driven by the need for energy substrates and oxygen [1]. Up-regulation of vascular endothelial growth factor (VEGF), which is secreted by tissues in response to hypoxia or inflammation, is regarded as the major cause of PA and targeted in current antiangiogenic therapies for neovascular eye diseases [2]. Clinically, anti-VEGF therapies have been proven to be effective in treating PA [3–6]. However, since VEGF is also essential for normal physiological vessel growth [7], the inhibition of VEGF can lead to significant adverse outcomes [8–10], especially in developing retinas in newborns [11–14], Long-term anti-VEGF injections also increase the release of inflammatory cytokines which may promote the development of geographic atrophy (GA) [15] and a change from angiogenesis to fibrosis [16] that further results in tractional retinal detachments (TRD) in patients with severe diabetic retinopathy [17]. For patients who are refractory to anti-VEGF agents, targeting inflammatory mediators has been implicated in disease progression [18, 19]. Intravitreal implantation of corticosteroids in conjunction with anti-VEGF treatment may be beneficial for those with anti-VEGF resistant disease [20, 21] due to its anti-inflammatory effect; however, ocular side effects are not uncommon [22]. Therefore, effective therapeutic agents for targeting PA without affecting normal retinal vascularization and for those who are refractory to anti-VEGF therapy, are critically needed.

Thrombomodulin (TM) is a cell membrane glycoprotein composed of five structural domains which was first discovered on the vascular endothelium [23]. TM comprises an NH₂-terminal C-type lectin-like domain (TMD1), six consecutively repeated epidermal growth factor (EGF)-like domain (TMD2), a serine and threonine-rich domain (TMD3), a transmembrane domain (TMD4), and a cytoplasmic domain (TMD5) [24]. TM is well known as an important regulator of coagulation and fibrinolysis, an essential cofactor in thrombin-mediated activation of protein C and of TAFI [25]. Further studies revealed that different domains of TM play a variety roles [26]. We have previously demonstrated that recombinant TMD1 (rTMD1) suppressed inflammation [27] and inhibited EGF-induced angiogenesis via the interaction with Lewis Y Ag [28]. However, instead of EGF, VEGF is the main therapeutic target in pathologic ocular angiogenesis (POA) [2]. Clinically, it had been reported that EGF levels were either very low or below the detection limit in POA patients [29], which also indicated that EGF didn’t play a major role in POA. Therefore, we aimed to investigate the biologic significance of the interaction between rTMD1 and VEGF-induced pathological neovascularization in this study. In particular, we assessed the anti-angiogenic and anti-inflammatory effects of rTMD1 in pathological and physiological angiogenesis. We also investigated the role of rTMD1 in hypoxia-inducible factor-1α (HIF-1α)-VEGF pathway. Finally, we compared the effect of rTMD1 with a currently used anti-VEGF agent to determine the potential of rTMD1 as a novel therapeutic agent for PA.

Materials and methods

Ethics statement

All animal studies and procedures adhered to the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Institutional Animal Care and Use Committee at National Cheng Kung University. C57BL/6 mice were purchased from National Laboratory Animal Center (Taiwan). The N-terminal Lectin-like Domain deleted (TM^LeD/LeD) mice were a gift from Dr. Conway [30]. Both male and female mouse pups were used.

Preparation of rTMD1 proteins

rTMD1 expression and purification in the Pichia pastoris expression system was conducted as previously described [27]. The purified rTMD1 proteins were examined by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting.

Cell cultures

Human retina microvascular endothelial cells (HRMECs) were purchased from the Applied Cell Biology Research Institute (Kirkland) and were grown in medium 199 (M199, Gibco BRL) supplemented with 20% Fetal Bovine Serum (FBS, Gibco BRL) and endothelial cell growth supplement (Millipore, Bedford). The experiments were performed with cells between passages 2 and 4.

Tube formation assay

The tube formation assay was conducted in 15-well µ-slides (ibidi GmbH, Germany) coated with Matrigel. To analyze the effect of rTMD1 on VEGF-mediated endothelial tube formation in vitro, HRMECs were starved for 12 h in M199 containing 1% FBS, then preincubated with rTMD1 (10 nM) and VEGF (20 ng/ml; R&D Systems) for 24 h. Total tube length was measured using MetaMorph software (Molecular Device) and compared with the control at the indicated times.

Cell migration assay

The assay was performed in a 48-multiwell Boyden chamber (Neuro Probe) with 8-µm pore-size polycarbonate filters (Neuro Probe) coated with 0.1% gelatin (Sigma-Aldrich). In the analysis of inhibition by rTMD1, HRMECs were starved for 12 h in M199 containing 1% FBS, then preincubated with rTMD1 (10 nM) and VEGF (20 ng/ml) for 24 h before being added to the upper compartment, M199 containing 5% FBS was added to the bottom compartment as chemoattractant, the migrated cells were stained with Liu stain and counted.

Cell viability assay

To evaluate the effect of rTMD1 on viability in HRMECs, WST-1 (Water-Soluble Tetrazolium salt-1) assay (Takara Bio, Shiga, Japan) was used as previously described [31]. Absorbance was measured at 450 nm using a plate reader, and absorption at 630 nm was measured as a background.

Western blot analysis

10 μg Of total protein was separated via 10% SDS-PAGE and transferred onto a polyvinylidene difluoride membrane. The membranes were blocked with 5% nonfat milk powder for 1 h at room temperature. After being probed with specific antibodies against VEGF-R1, VEGF-R2, P-ERK, ERK, p-AKT, AKT, P-JNK, JNK, P-p38, p38, VEGF, hypoxia-inducible factor-1α (HIF-1α), and β-actin at 4 °C overnight, the membranes were incubated with peroxidase-conjugated specific secondary antibodies (Calbiochem EMD Biosciences) for 1 h at room temperature. The signal was detected by enhanced chemiluminescence reagent (Millipore) and a Fujifilm LAS-3000 imager (Fujifilm Life Science).

Experimental oxygen-induced retinopathy (OIR) animal model

OIR was generated as previously described [32]. In brief, neonatal mice were exposed to 75% oxygen from postnatal day 7 (P7) until P12 and returned to room air (21% oxygen) from P12 to P17. During the first phase of hyperoxic exposure (P7–P12), retinal vessels constrict to regulate retinal PO₂ levels, and immature capillaries in the central retina regress leading to a central zone of vaso-obliteration (VO). According to the protocol, after being moved from 75% oxygen chamber to room air at P12, the formation of neo-vascularization (NV) in the mice retina was maximum at P17. Thus, the mice were sacrificed at P12 and P17, respectively. Mice received intraperitoneal administration of rTMD1 or PBS or bevacizumab once a day between P13 and P16. We addressed the quantification method of outlining VO and NV structures in Adobe Photoshop and comparing total VO and NV area with the total retinal area [33].

Histology and immunofluorescent staining

Eyes of mice were enucleated and fixed in 4% paraformaldehyde for 1 h. The corneas and muscles were removed, and the eyecups were placed in 4% paraformaldehyde for another 1 h. Then, the entire retina was dissected from the eyecup, and after rinsing and blocking, the retinas were incubated with the primary antibodies or conjugated antibodies overnight at 48 °C. After rinsing with phosphate buffered saline with Tween (PBST), the retinas were flat-mounted on cover slides with an aqueous mounting medium (Thermo Scientific, Fremont). In other retinas, 20-µm sections were cut with a cryostat (Leica CM 1800, Bannockburn). After washing and blocking, the flat-mounted retinas or cryostat sections were incubated with the primary antibodies overnight at 48 °C. The secondary antibodies were added for 1 h at room temperature, and the nuclei were counterstained with DAPI (Hoechst 33,342; Molecular Probes). The cryostat sections and flat-mounts were incubated with the following antibodies: fluorescein-labeled isolectin B4 (1:150 dilution; Vector Laboratories), DyLight 594 labeled isolectin B4 (1:150 dilution; Vector Laboratories), HIF-1α (1:100 dilution; Novus Biologicals), VEGF (1:100 dilution; Proteintech), BAX (1:100 dilution; Taiclone), and BCL-2 (1:100 dilution; Santa Cruz Biotechnology).

Terminal deoxynucleotidyl transferase (TdT) dUTP nick-end labeling (TUNEL) assay

rTMD1 or PBS was intraperitoneally injected to OIR and normoxia mice and were sacrificed at P17. Enucleated globes were fixed in 4% paraformaldehyde and embedded in paraffin. TUNEL staining was performed with a kit (Millipore) according to the manufacturer’s instructions. TUNEL-positive cells were evaluated in randomly selected fields at a 200 × magnification via light microscope (Carl Zeiss, Chester, VA, USA). There were at least six animals in each group.

In vitro cobalt binding assay

HRMECs were cultured in serum-free M199 and got starvation overnight. The HRMECs were then treated with 400 µM CoCl₂ and 20 nM rTMD1 for 6 h. The cell lysate was collected and stored at − 80 °C for Western Blot analysis to confirm the expression of HIF-1α and VEGF.

Quantitative real-time PCR (qRT-PCR)

Total RNA of mouse retinas under group-specified experimental conditions was extracted and qRT-PCR was performed using the SYBR Green RT-PCR Master mix (Biotool, Houston). Expression of target genes was normalized to β-actin, as an internal control, and measured in triplicate. The 2 − ΔΔCq method was used to calculate target gene expression. The PCR primers were designed based on the NCBI GeneBank database.

Statistical analysis

Comparisons between two groups were analyzed by two-tail Student’s t tests. For comparisons of more than two groups, 1-way ANOVA followed by a Bonferroni multiple comparison test was used. Probability values less than 0.05 were considered statistically significant. All statistical analyses were performed using GraphPad Prism 6.0 software.

Results

Inhibitory effects of rTMD1 on VEGF-mediated human retina microvascular endothelial cells migration and tube formation

To explore the effects of rTMD1 on VEGF-mediated PA, HRMECs were co-treated with rTMD1 in the presence of VEGF. The chemotaxis of HRMECs toward VEGF was significantly inhibited by rTMD1 and showed a significant reduction in tube formation (P < 0.05, Fig. 1A) and cell migration (P < 0.05, Fig. 1B).

Fig. 1.

rTMD1 inhibited VEFG-induced angiogenesis in vitro and in vivo. A rTMD1 inhibited VEGF-induced tube formation on Matrigel. HRMECs were starved for 12 h in M199 containing 1% FBS, then preincubated with rTMD1 (10 nM) and VEGF (20 ng/ml; R&D Systems) for 24 h. Total tube length was measured using MetaMorph software. The data are the mean ± SD (n = 3). *P < 0.01. Similar results were obtained in at least three different experiments. B Chemotaxis was assessed by Boyden chamber migration assay. HRMECs were starved for 12 h in M199 containing 1% FBS, then preincubated with rTMD1 (10 nM) and VEGF (20 ng/ml) for 24 h before being added to the upper compartment, 5% M199 was added to the bottom compartment as chemoattractant, the migrated cells were stained with Liu stain and counted. Values are the mean ± SD (n = 5), and similar results were obtained in at least three different experiments. The magnification bar is 200 µm. NT: non-treated (0 ng/ml of VEGF). C OIR procedure. Mice were exposed to hyperoxia (75%) from P7 to P12. After the oxygen exposure, mice were placed back at room air until P17. Mice received intraperitoneal administration of PBS or rTMD1 (0.8 mg/kg) once a day between P12 and P16. At P17, mice were sacrificed, and eyes were enucleated. Retinal wholemounts were stained with isolectin. D Vehicle or rTMD1 (0.8 mg/kg) were administrated intraperitoneally. rTMD1 significantly suppressed vaso-obliteration (VO) and neo-vascularization (NV) tufts in murine OIR model. Values are expressed as the mean ± SD (n = 6–8 retinas per group). **P < 0.01 and ***P < 0.001

rTMD1 suppressed PA in oxygen-induced retinopathy

To evaluate the effect of rTMD1 in OIR, we analyzed retinal wholemount staining. rTMD1 (0.8 mg/kg) or vehicle were intraperitoneally injected daily as scheduled (Fig. 1C). Administration of rTMD1 significantly suppressed VO (P < 0.01) and NV (P < 0.001) compared to vehicle injection alone (Fig. 1D).

rTMD1 treatment did not induce cell toxicity and suppressed apoptosis in OIR

Using the WST-1 assay to evaluate the cytotoxicity of rTMD1 on HRMECs co-treated with VEGF, we confirmed that rTMD1 did not affect cell viability (Fig. 2A). Retina toxicity of rTMD1 was evaluated by H&E staining. The ratios of the retinal thickness from the inner limiting membrane to the inner nuclear layer and outer nuclear layer were measured. Retinal thickness was significantly decreased in the OIR (P < 0.001) while rTMD1 does not affect the retinal thickness (Fig. 2B). We then determined whether rTMD1 could provide additional anti-apoptotic effects on OIR mice retina. The number of TUNEL-positive cells decreased significantly in all retinal layers after rTMD1 injection (Fig. 2C). Moreover, BAX expression was significantly increased in OIR retina while it was decreased after treatment with rTMD1. In parallel, BCL-2 expression increased after rTMD1 treatment (Fig. 2D). Taken together, these data indicate that rTMD1 did not induce direct cytotoxicity to retinal cells, whereas it elicited a protective effect.

Fig. 2.

rTMD1 treatment did not induce cell toxicity but suppressed apoptosis in OIR. A Effect of rTMD1 on the viability of HRMECs. Cytotoxicity was assessed by the WST-1 proliferation assay after rTMD1 treatment at different concentrations or PBS as control. B For the evaluation of retinal structure change, haematoxylin and eosin staining was performed. Retinal thickness A (internal limiting membrane to the inner nuclear layer) to B (internal limiting membrane to the outer nuclear layer) ratio was measured at P17 under different conditions, including normoxia, OIR, and OIR + rTMD1 (1.6 mg/kg). The A/B ratio was significantly decreased in the OIR while rTMD1 does not affect the retinal thickness. ***P < 0.001. C TUNEL staining was performed to count apoptotic cells at three randomly selected fields (200 ×) per section. Arrows indicate TUNEL-positive cells. D Immunofluorescent stains for BAX and BCL-2 were performed to determine the anti-apoptotic effect of rTMD1 in OIR. BAX expression significantly increased in OIR retina while it decreased after treating with rTMD1. BCL-2 expression increased after rTMD1 treatment. Figures are representative of three independent experiments. INL inner nuclear layer, ONL outer nuclear layer

TMD1 deficiency aggravated PA in OIR

To further confirm the role of rTMD1 in PA, we examined the impact of TMD1 deficiency by comparing the retinal vasculature in C57/BL6 and TMD1 knockout (TM^LeD/LeD) mice. There was no detectable developmental difference in vascular growth between C57/BL6 and TM^LeD/LeD mice at P7, P12, and P25 (Fig. 3A), which indicates that that TMD1 is likely dispensable for physiological developmental retinal angiogenesis. However, VO (P < 0.05) and NV (P < 0.01) were significantly increased in the TM^LeD/LeD mice in OIR (Fig. 3B).

Fig. 3.

TMD1 deficiency increased PA in OIR. A Physiological retinal vascular growth was not affected in TM^LeD/LeD mice. Retinal flat mounts from P7, P12 and P17 TM^LeD/LeD mice revealed normal retinal vasculature with similar levels of vascularized retinal areas compared to WT mice. At P17, retinal vascular growth was not affected in TM^LeD/LeD mice. Values are the mean ± SD (n = 4 per group). B Retinas of WT andTM^LeD/LeD mice. Retinal flat mounts were stained with isolectin (green) to visualize vessels, vaso-obliteration areas pseudo-colored red and tufs pseudo-colored yellow. Representative images are shown. Data are presented as mean ± SD. Fold change compared with the control group was calculated (n = 3–10 retinas per group). *P < 0.05 and **P < 0.01

rTMD1 regulates VEGF expression via hypoxia-inducible factor-1α in vitro and in vivo

We found that rTMD1 did not directly change the expression levels of VEGF receptor-1 (VEGFR-1), VEGFR-2 or their downstream signaling pathways (Fig. 4A). We therefore focused on other factors that regulate VEGF expression. It is well documented that the HIF-1α-VEGF pathway plays an important role in retinal neovascularization [34–36], and that HIF-1α is significantly upregulated at P17 in OIR [37]. Therefore, we investigated whether rTMD1 could regulate HIF-1α-VEGF pathway. We used an in vitro cobalt binding assay to mimic hypoxia. After incubating HRMECs with CoCl₂ and rTMD1 for 6 h, the protein levels of HIF-1α and VEGF were significantly upregulated (P < 0.05) in the CoCl₂ group and significantly reduced (P < 0.05) after rTMD1 treatment (Fig. 4B). In OIR, immunofluorescent staining of TM^LeD/LeD mice revealed higher HIF-1α and VEGF expression in the vehicle treatment group compared to WT mice. rTMD1 treatment decreased HIF-1α and VEGF expression in both TM^LeD/LeD and WT mice (Fig. 4C).

Fig. 4.

rTMD1 inhibits HIF-1α in vitro and in vivo. A rTMD1 did not directly change the expression level of VEGF receptor or its downstream factors, including JNK, AKT, ERK, and p38. B rTMD1 significantly inhibited HIF-1α activity induced by cobalt chloride. VEGF activity is also significantly inhibited by rTMD1. Values are the mean ± SD, and similar results were obtained in at least three different experiments. *P < 0.05, **P < 0.01 and ***P < 0.001. C Immunofluorescence staining of OIR WT and TM^LeD/LeD mice at P17. HIF-1α and VEGF expression significantly increased in TM^LeD/LeD mice while rTMD1 decreased HIF-1α and VEGF expression in both WT and TM^LeD/LeD mice

rTMD1 is comparable to bevacizumab and exhibits additional anti-inflammatory effects

To evaluate the therapeutic potential of rTMD1 in PA, we compared the effect of rTMD1 to the current widely used anti-VEGF agent, bevacizumab. We found that both rTMD1 and bevacizumab can effectively reduce NV at P17 (P < 0.001, Fig. 5A). The administration of rTMD1 or anti-VEGF agent didn’t reduce body weight. The body weight data of PBS-treated, rTMD1-treated and anti-VEGF-treated OIR mice were 6.16 ± 0.53, 6.08 ± 0.67, and 6.33 ± 0.59 g, respectively. There is no significant difference (P = 0.712) in body weight among the groups. Since rTMD1 is already well-known for its anti-inflammatory properties [30], we also investigated the effects of rTMD1 on the expression of inflammatory markers in OIR mice retina. The results of qRT-PCR showed that mRNA levels of IL-6 and ICAM-1 were upregulated in the OIR and bevacizumab treatment groups (P < 0.05) but downregulated after administration of rTMD1 (P < 0.05, Fig. 5B). Accordingly, our data indicated that rTMD1 not only suppressed VEGF-induced PA but also exhibited additional anti-inflammatory effects compared to bevacizumab.

Fig. 5.

rTMD1 is comparable to bevacizumab and exhibits anti-inflammatory effects. A rTMD1 and bevacizumab reduce NV and VO at P17. rTMD1 exhibits comparable therapeutic effects to bevacizumab. ***P < 0.001 compared to PBS. B mRNA levels of HIF-1α and VEGF were suppressed after rTMD1 treatment. IL-6 and ICAM-1 were upregulated in OIR after PBS and bevacizumab injection but downregulated when treating with rTMD1.Values are the mean ± SD (n = 3–6 per group), and similar results were obtained in at least 3 different experiments. *P < 0.05 and **P < 0.01

Discussion

In the present study, we found that rTMD1 inhibited PA without affecting physiological retinal vessel growth. Our data suggested that rTMD1 played an inhibitory role by reducing VEGF induction in vitro and in vivo, which occurred at least partly via downregulation of HIF-1α [38]. The effect of rTMD1 was not only comparable to bevacizumab in terms of its antiangiogenic effect, but it also provided anti-inflammatory effects by attenuating inflammatory cytokine expression, especially IL-6 and ICAM-1, in the retina of the OIR mice. To the best of our knowledge, this is the first study to describe the dual anti-angiogenic and anti-inflammatory effects of rTMD1 in PA.

Anti-VEGF therapy is the current conventional therapy for PA and repeated injections are common. Previous clinical studies have shown that long-term anti-VEGF therapy may lead to retinal pigment epithelium cell and photoreceptor atrophy [8, 15]. Animal studies also revealed that administration of VEGF inhibitors may significantly increase VO area [39]. These findings indicated that long-term VEGF therapy may impede physiological vascular development and lead to neurodegeneration. However, different from anti-VEGF therapy, our data revealed that although rTMD1 significantly inhibited VEGF-induced cell migration and tube formation, the cell proliferation function was preserved. Furthermore, rTMD1 increased BCL-2 and decreased BAX and TUNEL expression in OIR mice retina, providing mechanistic insights into the protective properties of rTMD1 in PA. It has been reported that soluble TM (sTM) has anti-apoptotic effects [40] and that high levels of circulating sTM shed from endothelial cells are associated with diminished risk of cardiovascular events [41]. Thus, in OIR, which is also a stressful condition, we found that rTMD1 has similar anti-apoptotic effects and lack of TMD1 aggravates PA. Taken together, rTMD1 inhibited PA without affecting physiological vessel growth and had anti-apoptotic effects that may be safe for long-term use.

Unlike traditional anti-VEGF therapies, we found that rTMD1 did not directly bind to VEGF or VEGFR-1 or VEGFR-2. To maintain VEGF at sufficient levels for physiological angiogenesis, it is reasonably to consider the upstream regulatory factors instead of inhibiting VEGF function directly [42]. Among these upstream factors, HIF-1α is the most important oxygen dependent regulator that contributes to PA [43–46]. Our results revealed that rTMD1 suppressed HIF-1α and VEGF expression in the CoCl₂ assay and OIR model, which indicates that rTMD1 exhibits its anti-angiogenic effect by regulating the HIF-1α-VEGF pathway. Moreover, according to our result (Fig. 5B) and previous reports, rTMD1 can significantly suppress IL-6 expression in OIR retina and other tissues [47, 48]. Since it demonstrated that activation of IL-6 could upregulate HIF-1α expression [49, 50], we proposed that rTMD1 inhibits HIF-1α by reducing IL-6 expression in the OIR retina in our study. The advantage of targeting HIF-1α is that inhibition of the HIF-1α-VEGF pathway could suppress pathological angiogenesis without affecting physiological angiogenesis [37, 43, 51]. Notably, HIF-1α inhibitors have not entered the clinic due to adverse reactions locally and systemically [37]. In this study, we first found that rTMD1 could inhibit VEGF-induced PA via HIF-1α without affecting physiological angiogenesis. Thus, rTMD1 may be a potential HIF-1α inhibitor for clinical use.

rTMD1 also has potent anti-inflammatory properties [27, 30]. Previous studies have showed that angiogenesis, under both physiological and pathological situations, is usually associated with inflammation [52]. sTM levels have been reported to be related to the inflammation status which reflects endothelial damage [28], and serves as a marker for clinical severity of cardiovascular disease [53, 54], sepsis [55], and diabetes [56]. Administration of different forms of rTM exhibit different properties, depending on which domain(s) were used. For example, rTMD123 can dampen formation of experimental abdominal aortic aneurysm [57], while rTMD23 can promote corneal epithelial wound healing [58]. As for rTMD1, Conway et al. first reported that TM^LeD/LeD mice have more severe inflammatory responses [30]. Subsequent studies showed that this property was at least partially achieved through sequestration of high mobility group box B1 (HMGB1) [59] and inhibition of complement activation [60]. The possibility that other TMD1 interacting molecules may participate in anti-inflammation is worth investigating.

During the progression of GA, TRD, and diabetic macula edema, the serum or vitreous level of inflammatory cytokines, especially IL-6 and ICAM-1, are closely related to the severity of diseases [61–64]. Thus, for patients who are refractory to anti-VEGF agents, IL-6 and ICAM-1 may be additional therapeutic target molecules. In this study, we found that rTMD1 significantly reduced IL-6 and ICAM-1 expression in PA, which may be beneficial to the patients who had poor responses to anti-VEGF. Indeed, IL-6 levels were significantly elevated after bevacizumab injection. Our findings are consistent with previous reports that IL-6 was up-regulated after bevacizumab treatment [65], which may promote the fibrotic process and further lead to TRD that has been found after multiple anti-VEGF injections [66]. Together, our data showed that compared to bevacizumab, rTMD1 has an additional anti-inflammatory effect which may be beneficial for PA treatment and decrease the possibility of fibrosis.

In conclusion, we demonstrated for the first time that rTMD1 suppressed PA via inhibition of the HIF-1α-VEGF pathway, and reduced inflammation in the retina of OIR mice. rTMD1 inhibits PA but spares normal physiological vessel growth; thus, rTMD1 may be considered as a potential therapeutic strategy for PA.

Author contributions

YHH, ICP and HLW involved in design and conduct of study. Collection and management of the data were done by YHH, CHK and ICP. YHH, CHK, ICP, YSC, SHT, EMC and HLW participated in management and interpretation of the data. YHH, CHK, EMC and HLW participated in preparation, review, and approval of the manuscript.

Funding

Supported by Grants Ministry of Science and Technology 109-2628-B-006-032 to YHH and 110-2320-B-006-051 to HLW, Taiwan.

Availability of data and material

All data are available in the main text.

Code availability

Not applicable.

Declarations

Conflict of interest

The authors have no relevant financial or non-financial interests to disclose.

Ethics approval

All efforts were made to reduce both animal suffering and the number of animals used. All animal studies and procedures adhered to the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Institutional Animal Care and Use Committee at National Cheng Kung University.