Changes in Aqueous Cytokine Levels Following Intravitreal Aflibercept in Treatment-Naive Patients with Diabetic Macular Edema

Abstract

Purpose:

To investigate the changes in aqueous humor cytokine levels in response to short-term aflibercept therapy in treatment-naive patients with center-involving diabetic macular edema (DME).

Methods:

This is a prospective cohort study that included patients with treatment-naive DME with central subfield macular thickness ≥310 μm on optical coherence tomography from July 2015 to May 2017. Patients received 3 monthly intravitreal aflibercept injections. Aqueous samples for cytokine analysis were obtained before the first and third injections. Levels of various cytokines were measured using multiplex immunoassay. Main outcome measures were changes in aqueous cytokine levels from baseline to month 2.

Results:

A total of 17 patients were enrolled and 16 completed the study. The mean age was 57.2 ± 8.1 years. The following cytokines were significantly higher at month 2 versus baseline: transforming growth factor-beta (TGF-β)1 (P = 0.004), TGF-β2 (P = 0.017), inducible protein (IP)-10 (P = 0.011), and hepatocyte growth factor (HGF) (P = 0.02). There were significant reductions in the levels of vascular endothelial growth factor (VEGF) (P < 0.001), placental growth factor (PlGF) (P = 0.028), interleukin (IL)-6 (P = 0.011), and platelet-derived growth factor-AA (PDGF-AA) (P = 0.003).

Conclusions:

In treatment-naive patients with DME, short-term aflibercept therapy not only results in VEGF and PlGF suppression, but also leads to reduced levels of IL-6 and PDGF-AA and higher concentrations of TGF-β1, TGF-β2, HGF, and IP-10.

Introduction

Diabetic macular edema (DME) is an important cause of central visual loss caused by accumulation of excess fluid in the extracellular space of the neurosensory retina.¹ The microvascular and structural retinal changes associated with this condition are a consequence of increased inflammation, ischemia, and oxidative stress that alter the local cytokine milieu.² Various angiogenic and inflammatory cytokines, such as vascular endothelial growth factor (VEGF), influence vascular permeability and therefore may contribute to the disease pathogenesis.^2,3 In fact, significantly higher concentrations of multiple cytokines, including VEGF, transforming growth factor-beta (TGF-β), hepatocyte growth factor (HGF), interleukin (IL)-6, and monocyte chemoattractant protein-1 (MCP-1), have been found in patients with DME when compared with healthy controls.⁴ As a consequence, current therapies for DME focus on pharmacological modulation of these factors.

The treatment of DME has been revolutionized with the advent of anti-VEGF therapy. Ranibizumab, a monoclonal antibody fragment, and bevacizumab, its full-length monoclonal antibody, were the first intravitreal anti-VEGF agents to be widely used and both bind to VEGF-A isoforms.^5,6 A third anti-VEGF drug, aflibercept, a recombinant fusion protein that binds to VEGF-A, VEGF-B, and placental growth factor (PlGF), was later released and shown to be superior to ranibizumab and bevacizumab in DME patients with poor vision after 1 year of treatment.^5,7

Despite the well-established efficacy of all anti-VEGF agents, variability in response to anti-VEGF therapy in DME remains a challenge. A recent study showed that 14.6% of patients with DME were considered nonresponders to long-term anti-VEGF therapy.⁸ The reason for this variability is not entirely understood, but is likely related to the complex disease cascade rather than being solely VEGF mediated. Interestingly, our group has demonstrated that baseline aqueous levels of intercellular adhesion molecule (ICAM)-1 were associated with DME severity, whereas VEGF levels were not, despite the proven efficacy of anti-VEGF agents targeting primarily VEGF.⁹ The interplay of various cytokine pathways is supported by the fact that short-term ranibizumab therapy, which has no other affinity other than for VEGF-A, has shown to reduce levels of several aqueous cytokines other than VEGF, including IL-6, PlGF, and ICAM-1.¹⁰ Because aqueous cytokines may respond differently to other anti-VEGF agents in patients with DME, the objective of this study is to investigate the changes in aqueous humor cytokine levels in response to short-term intravitreal aflibercept therapy in treatment-naive patients with center-involving DME.

Methods

This is a prospective cohort study performed at St. Michael's Hospital, Toronto, Canada, from July 2015 to August 2017. This study was approved by St. Michael's Hospital Research Ethics Board and adhered to the tenets of the Declaration of Helsinki. All participants provided written informed consent to participate in this study.

Participants

Adult patients with center-involving DME with central subfield macular thickness (CST) ≥310 μm on spectral-domain optical coherence tomography (SD-OCT) were included in this study. Study eye exclusion criteria were active proliferative diabetic retinopathy (PDR), history of pan-retinal photocoagulation (PRP) within the last 12 months or anticipated need for PRP in the following 6 months, history of intraocular surgery or focal laser photocoagulation in the last 3 months, previous pars plana vitrectomy, coexistent macular pathology, uncontrolled glaucoma, and previous intravitreal injection of any agent to either eye. Patients who were on renal dialysis, systemic immunosuppression, or on any ocular steroids were also excluded.

Study design

At baseline, patients underwent a complete ophthalmological examination, including Snellen best-corrected visual acuity (BCVA) and dilated fundoscopy. Color fundus photograph and fluorescein angiogram were obtained to rule out any occult vascular pathology. Diabetic retinopathy was graded according to the International Clinical Diabetic Retinopathy and Diabetic Macular Edema Disease Severity Scales.¹¹ SD-OCT images (Cirrus HD-OCT; Carl Zeiss Meditec AG, Germany) were captured using the 512 × 128 macular cube protocol. Each image was reviewed by the treating physician to confirm diagnosis, accurate alignment, foveal centration, and segmentation and to exclude coexisting macular pathology. CST, which corresponds to the average retinal thickness within the central early treatment diabetic retinopathy study (ETDRS) subfield (1 mm diameter), and the macular volume (MV), which represents the total retinal volume within the entire ETDRS circle (6 mm diameter), were automatically calculated by the SD-OCT.

Each participant received a monthly intravitreal aflibercept 2.0 mg (0.05 mL) injection for 3 months (baseline and months 1 and 2). In addition, patients were seen at month 3 for assessment of visual and anatomic response following their final aflibercept injection. At each follow-up visit, patients had Snellen BCVA and SD-OCT repeated.

Patients were classified into responders and nonresponders to short-term aflibercept therapy according to SD-OCT CST and MV parameters. A MV responder was defined as a patient with ≥10% reduction in MV from baseline to month 3. A patient was considered a CST responder when there was a reduction of ≥50% in excess CST (≥310 μm) from baseline to month 3.

Sample collection and analysis

Aqueous humor fluid (0.2 mL) was collected immediately before each aflibercept injection at baseline and months 1 and 2 through an anterior chamber paracentesis using a 30-gauge needle inserted into the anterior chamber through the limbus under a sterile technique. Samples were stored at −80°C within 2 h of collection and not thawed until analysis. Aqueous samples were analyzed for cytokines that have been previously described to be associated with DME: fibroblast growth factor-2 (FGF-2), IL-6, IL-8, IL-10, inducible protein (IP)-10, MCP-1, VEGF, platelet-derived growth factor-AA (PDGF-AA), HGF, PlGF, ICAM-1, vascular cell adhesion molecule-1 (VCAM-1), TGF-β1, TGF-β2, and TGF-β3. Multiplex immunoassays of samples (HCYTOMAG-60K-08, HCVD2MAG-67K-02, TGFBMAG-64K-03; MilliporeSigma, Burlington, MA) were carried out in duplicate using the Luminex^® platform according to the manufacturer's protocol.

Statistical analysis

The primary outcome of this study was the change in concentrations of aqueous humor cytokines from baseline to month 2. IL-10 was excluded from the analysis since its measurements clustered at the lower level of sensitivity of the assay. Snellen BCVA was converted to logMAR visual acuity for statistical analysis. Descriptive statistics were used to summarize data. A Shapiro–Wilk's test was performed to examine whether the samples were normally distributed. Wilcoxon signed-rank test was performed to compare continuous variables, including aqueous cytokine levels at baseline versus month 1 and month 2. Median percent changes for cytokine levels were calculated using the formula: $100 * (\frac{m e d i a n m o n t h 2 v a l u e - m e d i a n b a s e l i n e v a l u e}{m e d i a n b a s e l i n e v a l u e}) .$ Data analysis was performed using SPSS software (version 26; IBM Corp., Armonk, NY). A P value <0.05 was considered statistically significant.

Results

A total of 17 patients were initially enrolled in this study, but 1 patient withdrew at month 1. Therefore, data of 16 eyes belonging to 16 patients who completed the study are presented. The mean age was 57.2 ± 8.1 years and 10 (62.5%) were male. At baseline, the mean logMAR BCVA was 0.39 ± 0.16 (Snellen 20/50), mean CST was 430.9 ± 85.5 μm, and mean MV was 11.9 ± 1.2 mm³. Additional baseline data are described in Table 1.

Table 1.

Baseline Characteristics of the Study Population (n = 16)

Characteristics	Value
Age, years, mean ± SD	57.2 ± 8.1
Gender, n (%)
Male	10 (62.5)
Female	6 (37.5)
Lens status, n (%)
Phakic	13 (81.2)
Pseudophakic	3 (18.8)
Severity of retinopathy, n (%)
No diabetic retinopathy	1 (6.25)
Mild NPDR	4 (25.0)
Moderate NPDR	2 (12.5)
Severe NPDR	8 (50.0)
Inactive PDR	1 (6.25)
LogMAR BCVA, mean ± SD	0.39 ± 0.16
Central subfield thickness, μm, mean ± SD	430.9 ± 85.5
Macular volume, mm³, mean ± SD	11.9 ± 1.2

BCVA, best-corrected visual acuity; NPDR, nonproliferative diabetic retinopathy; PDR, proliferative diabetic retinopathy; SD, standard deviation.

According to the pre-established criteria of clinical response based on SD-OCT parameters, 14 (87.5%) patients were considered CST responders and 7 (43.8%) were MV responders. The median changes in CST and MV from baseline to month 3 were −92.5 μm (IQR −141.0 to −61.8 μm) and −1.00 mm³ (IQR −1.6 to −0.7 mm³). Details of clinical response to aflibercept therapy are summarized in Table 2.

Table 2.

Changes in Clinical Features from Baseline to Month 3 (n = 16)

Variable	Change between baseline and month 3, mean (SD) [range]	P
LogMAR BCVA	0.02 (0.09) [−0.17 to 0.18]	0.42
Central subfield thickness, μm	−94.6 (68.3) [−202.0 to 87.0]	0.002^*
Macular volume, mm³	−1.1 (0.6) [−2.1 to −0.2]	<0.001^*

Statistically significant (P < 0.05).

Overall, the following cytokines were significantly higher at month 2 versus baseline: TGF-β1 (median change: 61.2%, P = 0.004), TGF-β2 (median change: 7.6%, P = 0.017), IP-10 (median change: 66.1%, P = 0.011), and HGF (median change: 34.8%, P = 0.02). Other cytokines demonstrated significantly reduced levels at month 2 when compared with baseline: VEGF (median change = −100%, P < 0.001), PlGF (median change = −100%, P = 0.012), IL-6 (median change = −43.4%, P = 0.011), and PDGF-AA (median change = −10.9%, P = 0.003). There were no significant changes in the levels of IL-8, MCP-1, ICAM-1, VCAM-1, FGF-2, and TGF-β3 (Table 3).

Table 3.

Aqueous Humor Cytokine Concentrations at Baseline and at Months 1 and 2 Following Intravitreal Aflibercept (n = 16)

Cytokine	Cytokine levels, pg/mL, mean (SD) [range]				P^a
Cytokine	Baseline	Month 1	Month 2	Change between baseline and month 2, mean (SD)	P^a
VEGF	192.1 (212.2) [57.6–918.5]	3.2 (13.1) [0–52.4]	3.9 (15.9) [0–63.8]	−188.1 (213.6)	<0.001^*
IL-6	37.1 (71.6) [0–282.0]	18 (53.2) [0–214.9]	13.4 (24.4) [0–92.8]	−23.7 (51.9)	0.011^*
IL-8	13.9 (12.2) [0–37.3]	18.3 (18.8) [0–58.0]	21.1 (23.1) [0–71.7]	7.2 (16.6)	0.104
IP-10	130.0 (87.2) [40.3–354.7]	193.7 (134.6) [19.6–500.9]	207.9 (154.6) [0–547.9]	77.9 (114.5)	0.011^*
MCP-1	1829.1 (659.4) [866.7–3081.3]	1631.3 (823.6) [360.4–3614.5]	1684.3 (953.8) [0–3914.3]	−144.8 (614.6)	0.361
PDGF-AA	38.4 (10.1) [18.4–54.7]	31.4 (12.9) [0–48.8]	30.4 (12.1) [0–45.8]	−8.0 (8.8)	0.003^*
ICAM-1	22427.6 (24082.2) [147.4–66384.2]	15387.7 (17770.6) [258.8–70220.2]	26393.7 (40931.6) [0–139774.5]	3966.1 (36093.5)	0.836
VCAM-1	6247.9 (10464.6) [1722.8–44883.0]	5101.7 (5590.4) [1780.2–24183.2]	5776.1 (7797.8) [0–32508.8]	−471.8 (3620.4)	0.959
FGF-2	50.9 (86.2) [0–257.8]	69.8 (122.1) [0–384.5]	76.0 (127.1) [0–462.9]	25.1 (150.1)	0.878
HGF	949.7 (428.8) [363.2–1634.2]	1158.7 (660.4) [261.2–2636.6]	1340.1 (787.3) [0–3172.9]	390.4 (569.4)	0.02^*
PlGF	4.4 (6.1) [0–20.6]	0.6 (2.6) [0–10.2]	0 (0) [0–0]	−4.4 (6.1)	0.012^*
TGF-β1	88.6 (44.7) [31.7–170.2]	151.1 (92.3) [38.8–344.6]	160.5 (110.3) [0–429.3]	71.9 (78.6)	0.004^*
TGF-β2	17394.3 (4402.7) [9321.9–22466.5]	18922.5 (5057.6) [9970.2–27420.5]	18511.7 (7178.8) [0–27558.4]	1117.4 (5707.9)	0.017^*
TGF-β3	44.1 (12.1) [19.4–57.8]	48.8 (15.6) [24.1–80.4]	48.0 (18.5) [0–72.3]	3.9 (19.5)	0.427

Comparative analysis between baseline and month 2.

Statistically significant (P < 0.05).

FGF-2, fibroblast growth factor-2, HGF, hepatocyte growth factor; ICAM-1, intercellular adhesion molecule-1; IL, interleukin; IP-10, inducible protein-10; MCP-1, monocyte chemoattractant protein-1; PDGF-AA, platelet-derived growth factor-AA; PlGF, placental growth factor; TGF-β, transforming growth factor-beta.; VCAM-1, vascular cell adhesion molecule-1; VEGF, vascular endothelial growth factor.

Discussion

This study emphasizes the complex interplay that exists between various cytokines in the pathogenesis of DME and in response to aflibercept therapy. Aflibercept is a recombinant fusion protein that binds to VEGF-A, VEGF-B, and PlGF.⁷ Therefore, our findings that aqueous VEGF and PlGF levels were both significantly reduced in response to aflibercept therapy are consistent with the mechanism of action of the drug. Surprisingly, other cytokines demonstrated changes in their concentrations throughout the course of short-term treatment with aflibercept, supporting the idea that along with the suppression of VEGF and PlGF, there is also an upstream and downstream modulation of various other cytokines. Interestingly, despite having been previously identified as a biomarker for disease severity in DME and to be significantly downregulated following ranibizumab therapy, ICAM-1 did not show the same behavior in response to aflibercept therapy in our study.^9,10 One hypothesis is that that the expression of ICAM-1 is likely still being influenced by other non-VEGF mediated pathways that are perhaps downregulated in response to intravitreal therapy with ranibizumab but not with aflibercept. It is also possible that the smaller sample size in this study could be responsible for the conflicting results.

Aqueous levels of PDGF-AA and IL-6 are known to be higher in patients with DME.^2,3,12 Interestingly, both cytokines had their concentrations reduced following 2 intravitreal aflibercept injections. Reduction in aqueous IL-6 levels in patients with DME following anti-VEGF injections has also been previously reported with both aflibercept and ranibizumab.^10,13 IL-6 is a significant proinflammatory cytokine essential for the induction of acute inflammatory reactions, regulation of immune processes, and also plays a role in angiogenesis and vascular permeability.¹⁴ Because its expression is upregulated by the binding of VEGF to VEGF receptor-2, it would be expected that the suppression of VEGF levels by anti-VEGF agents could lead to downregulation of IL-6, which may be an indirect favorable effect of these drugs.¹⁴ PDGF-AA plays an important role in the angiogenesis cascade by promoting migration and proliferation of endothelial cells along with an increase in the recruitment of pericytes.¹⁵ Targeting PDGF signaling has been the basis for various studies looking at the development of other antiangiogenic medications that could act in combination with anti-VEGF therapy.^15,16 Thus, it is interesting to observe that aflibercept alone resulted in reduction of PDGF-AA levels, which has been previously observed in DME patients treated with intravitreal triamcinolone, but not in eyes treated with ranibizumab or bevacizumab alone.^17,18

The most surprising finding in this study was the increasing levels of various cytokines involved in inflammation following aflibercept injection, which had not been described in other studies that looked at cytokine variations after intravitreal ranibizumab or bevacizumab therapy in patients with DME.^4,10,18 One hypothesis is that the upregulation of certain inflammatory cytokines could be associated with the pathogenesis of aflibercept-related sterile intraocular inflammation, a poorly understood complication that has been described with this medication.^19,20 We observed that aqueous concentrations of IP-10, HGF, TGF-β1, and TGF-β2 were significantly higher after 2 injections of aflibercept when compared with baseline. The upregulation of granulocyte-macrophage colony-stimulating factor and fractalkine following aflibercept therapy in patients with DME has been recently observed in a study that included 20 patients with DME who were not necessarily treatment-naive.¹³ Our study adds to their findings, as we observed an upregulation of cytokines (HGF, TGF-β1, and TGF-β2) that had not been tested.¹³ In contrast, the increasing IP-10 levels seen in our study differ from the significantly reduced levels shown by Mastropasqua et al.¹³ However, since our study exclusively included treatment-naive patients, a plausible hypothesis is that there could be an initial upregulation of IP-10 in response to aflibercept therapy in patients who have not been previously exposed to anti-VEGF therapy. IP-10 is a cytokine that enhances immune reactivity but is also a potent inhibitor of both IL-8- and FGF-induced angiogenic activity.²¹ Therefore, an upregulation of IP-10 following aflibercept therapy could be an advantageous effect.²¹ IP-10 has been reported to be significantly higher in DME patients compared with controls, and Shiraya et al. have demonstrated that the treatment for DME can be more effective in eyes that present with higher baseline levels of IP-10.^13,17

HGF is an endothelium-specific growth factor that regulates cell growth, morphogenesis of various types of cells, and has also been implicated in retinal angiogenesis and vascular permeability.²² Its expression is exacerbated by hypoxic conditions, and HGF concentrations are significantly higher in patients with PDR when compared with healthy controls.²³ Most importantly, in vivo analysis has shown that HGF-induced retinal permeability is of similar magnitude to that induced by VEGF, suggesting that HGF may play an important role not currently targeted by current anti-VEGF agents. A significant correlation between variations in HGF levels and changes in the amount of macular edema has been shown in patients with DME treated with intravitreal dexamethasone implant, which supports the hypothesis of HGF being a contributor to macular edema.²⁴ In contrast, our study showed a significant median increase of 34.8% in the levels of HGF following aflibercept injections, which may be a potential factor for suboptimal response to anti-VEGF therapy. Therefore, intravitreal steroids may be helpful in patients who are nonresponders to anti-VEGF therapy, given its broader spectrum of action on other inflammatory molecules.

TGF-β is a multifunctional growth factor and modulator of vascular cells. Its inhibition may lead to abnormalities in the retinal microcirculation, including reduced retinal perfusion, impaired peripheral vascular autoregulation, and reduced endothelial barrier function.²⁵ However, the upregulation of TGF-β can result in exacerbated synthesis of extracellular matrix molecules, which is known to be involved in epiretinal membrane formation in patients with PDR and in the development of diabetic nephropathy.^26–28 Therefore, it is unclear at this time how the upregulation in TGF-β levels following aflibercept therapy could impact the response to treatment.

This study has some limitations. The small sample size limited the statistical power to determine the associations between anatomic response and aqueous cytokine concentrations. However, the purpose of this study was to simply assess the changes in aqueous cytokines following intravitreal aflibercept therapy. Another potential limitation is that cytokines were measured from aqueous fluid samples although DME is a disease involving the posterior segment, and therefore, analysis of vitreous fluid would have been ideal. Due to safety concerns involving vitreous sampling along with previous evidence validating aqueous sampling for posterior segment diseases, we felt that obtaining aqueous samples would be the preferred approach.²⁹

In conclusion, intravitreal aflibercept therapy results in an upregulation and downregulation of various aqueous cytokines other than those directly related to its mechanism of action in treatment-naive patients with DME. This study demonstrated significantly reduced aqueous levels of VEGF, PlGF, IL-6, and PDGF-AA and an increase in the cytokines related to inflammation, including TGF-β1, TGF-β2, IP-10, and HGF. Further studies are needed to determine the association of baseline and follow-up aqueous cytokine levels with anatomic response to treatment with intravitreal aflibercept in patients with DME.

Footnotes

Acknowledgment

The authors would like to thank Gurbaksh Basi from the Princess Margaret Genomics Centre, Toronto, Canada, for assistance with cytokine analysis.

Author Disclosure Statement

R.H.M. receives grants from Novartis and Bayer, personal fees from Novartis, Bayer, Allergan, Roche, and Bausch & Lomb, but none of them related to the submitted work. V.R.J., M.Y.K.M., and M.B. have no competing financial interests.

Funding Information

No funding was received for this article.

References

Antcliff

R.J.

, and Marshall

The pathogenesis of edema in diabetic maculopathy. Semin. Ophthalmol. 14:223–232, 1999.

Funatsu

, Noma

, Mimura

, Eguchi

, and Hori

Association of vitreous inflammatory factors with diabetic macular edema. Ophthalmology. 116:73–79, 2009.

Jonas

J.B.

, Jonas

, Neumaier

, and Findeisen

Cytokine concentration in aqueous humor of eyes with diabetic macular edema. Retina. 32:2150–2157, 2012.

Wen

, Jiang

, Zheng

, and Zhou

Six-month changes in cytokine levels after intravitreal bevacizumab injection for diabetic macular oedema and macular oedema due to central retinal vein occlusion. Br. J. Ophthalmol. 99:1334–1340, 2015.

Wells

J.A.

, Glassman

A.R.

, Ayala

A.R.

, et al. Aflibercept, bevacizumab, or ranibizumab for diabetic macular edema. N. Engl. J. Med. 372:1193–1203, 2015.

Lucentis (Ranibizumab injection)—Product Monograph. Published 2015. https://www.novartis.ca/sites/www.novartis.ca/files/lucentis_scrip_e.pdf accessed January 15, 2020.

Aflibercept—National Drug Monograph. Published 2012. https://webcache.googleusercontent.com/search?q=cache:RPMWZaFnczQJ:https://www.pbm.va.gov/PBM/clinicalguidance/drugmonographs/AfliberceptDrugMonograph.doc+&cd=8&hl=pt-BR&ct=clnk&gl=ca&client=safari accessed January 15, 2020.

Felfeli

, Juncal

V.R.

, Hillier

R.J.

, et al. Aqueous humor cytokines and long-term response to anti-vascular endothelial growth factor therapy in diabetic macular edema. Am. J. Ophthalmol. 206:176–183, 2019.

Hillier

R.J.

, Ojaimi

, Wong

D.T.

, et al. Aqueous humor cytokine levels as biomarkers of disease severity in diabetic macular edema. Retina. 37:761–769, 2017.

10.

Hillier

R.J.

, Ojaimi

, Wong

D.T.

, et al. Aqueous humor cytokine levels and anatomic response to intravitreal ranibizumab in diabetic macular edema. JAMA Ophthalmol. 136:382–388, 2018.

11.

Wilkinson

, Ferris

F.L.

, Klein

R.E.

, et al. Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales. Ophthalmology. 110:1677–1682, 2003.

12.

Lee

W.J.

, Kang

M.H.

, Seong

, and Cho

H.Y.

Comparison of aqueous concentrations of angiogenic and inflammatory cytokines in diabetic macular oedema and macular oedema due to branch retinal vein occlusion. Br. J. Ophthalmol. 96:1426–1430, 2012.

13.

Mastropasqua

, D'Aloisio

, Di Nicola

, et al. Relationship between aqueous humor cytokine level changes and retinal vascular changes after intravitreal aflibercept for diabetic macular edema. Sci. Rep. 8:1–9, 2018.

14.

Kotake

, Noma

, Yasuda

, Motohashi

, Goto

, and Shimura

Comparing cytokine kinetics between ranibizumab and aflibercept in central retinal vein occlusion with macular edema. Ophthalmic Res. 61:210–217, 2019.

15.

Sadiq

M.A.

, Hanout

, Sarwar

, et al. Platelet derived growth factor inhibitors: a potential therapeutic approach for ocular neovascularization. Saudi J. Ophthalmol. 29:287–291, 2015.

16.

Siedlecki

, Wertheimer

, Wolf

, et al. Combined VEGF and PDGF inhibition for neovascular AMD: anti-angiogenic properties of axitinib on human endothelial cells and pericytes in vitro. Graefes Arch. Clin. Exp. Ophthalmol. 255:963–972, 2017.

17.

Shiraya

, Kato

, Araki

, Ueta

, Miyaji

, and Yamaguchi

Aqueous cytokine levels are associated with reduced macular thickness after intravitreal ranibizumab for diabetic macular edema. PLoS One. 12:e0174340, 2017.

18.

Sohn

H.J.

, Han

D.H.

, Kim

I.T.

, et al. Changes in aqueous concentrations of various cytokines after intravitreal triamcinolone versus bevacizumab for diabetic macular edema. Am. J. Ophthalmol. 152:686–694, 2011.

19.

Hahn

, Kim

J.E.

, Stinnett

, et al. Aflibercept-related sterile inflammation. Ophthalmology. 120:1100–1101, 2013.

20.

Hahn

, Chung

M.M.

, Flynn

Jr.

, H.W., et al. Postmarketing analysis of aflibercept-related sterile intraocular inflammation. JAMA Ophthalmol. 4:421–426, 2015.

21.

Strieter

, Kunkel

, Arenberg

, et al. Interferon Y-inducible protein 10 (IP-10), a member of the C-X-X chemokine family, is an inhibitor of angiogenesis. Biochem. Biophys. Res. Commun. 210:51–57, 1995.

22.

Canton

, Burgos

, Hernández

, et al. Hepatocyte growth factor in vitreous and serum from patients with proliferative diabetic retinopathy. Br. J. Ophthalmol. 84:732–735, 2000.

23.

Clermont

A.C.

, Cahill

, Salti

, et al. Hepatocyte growth factor induces retinal vascular permeability via MAP-kinase and PI-3 kinase without altering retinal hemodynamics. Invest. Ophthalmol. Vis. Sci. 47:2701–2708, 2006.

24.

Campochiaro

P.A.

, Hafiz

, Mir

T.A.

, et al. Pro-permeability factors in diabetic macular edema; the diabetic macular edema treated with ozurdex trial. Am. J. Ophthalmol. 168:13–23, 2016.

25.

Walshe

T.E.

, Saint-Geniez

, Maharaj

A.S.R.

, et al. TGF-β is required for vascular barrier function, endothelial survival and homeostasis of the adult microvasculature. PLoS One. 4:e5149, 2009.

26.

Shi

G.J.

, Shi

G.R.

, Zhou

J.Y.

, et al. Involvement of growth factors in diabetes mellitus and its complications: a general review. Biomed. Pharmacother. 101:510–527, 2018.

27.

Harada

, Mitamura

, and Harada

The role of cytokines and trophic factors in epiretinal membranes: involvement of signal transduction in glial cells. Prog. Retin. Eye Res. 25:149–164, 2006.

28.

Goldfarb

, and Ziyadeh

TGF-B. A crucial component of the pathogenesis of diabetic nephropathy. Trans. Am. Clin. Climatol. Assoc. 112:27–33, 2001.

29.

Noma

, Funatsu

, Yamasaki

, et al. Aqueous humour levels of cytokines are correlated to vitreous levels and severity of macular oedema in branch retinal vein occlusion. Eye (Lond.), 22:42–48, 2008.