Article Text

PDF
Aqueous cytokine and growth factor levels indicate response to ranibizumab for diabetic macular oedema
  1. Masahiko Shimura,
  2. Kanako Yasuda,
  3. Ryosuke Motohashi,
  4. Osamu Kotake,
  5. Hidetaka Noma
  1. Department of Ophthalmology, Tokyo Medical University Hachioji Medical Center, Hachioji, Tokyo, Japan
  1. Correspondence to Dr Masahiko Shimura, Department of Ophthalmology, Tokyo Medical University Hachioji Medical Center, 1163 Tate-machi, Hachioji, Tokyo 193-0998, Japan; masahiko{at}v101.vaio.ne.jp

Abstract

Background/Aims To investigate the relations between aqueous humour levels of cytokines/growth factors and treatment response to intravitreal ranibizumab (IVR) for diabetic macular oedema (DME)

Methods Sixty-eight eyes of 68 patients with treatment-naïve centre-involved DME, central macular thickness (CMT) greater than 400 μm and visual acuity (VA) worse than logMAR 0.3 were recruited. Each patient received monthly IVR injection (0.5 mg/0.05 mL) until CMT was reduced to below 300 μm. Additional IVR was given to maintain CMT below 300 μm during the clinical course of 6 months with monthly follow-up. Aqueous concentrations of cytokines/chemokines and growth factors were measured using samples obtained just before first IVR injection. CMT and VA were monitored monthly for up to 6 months. The number of monthly IVR injections given during the 6-month study period was also recorded.

Results Twenty-four eyes showed CMT <300 μm soon after the first IVR injection (good responders), while 12 eyes did not reach the goal after six consecutive injections (poor responders). Baseline CMT and VA were not significantly different between the two groups. However, the good responders showed significant increases in baseline aqueous concentrations of vascular endothelial growth factor (VEGF), placenta growth factor, soluble VEGF receptor-1 (sVEGFR1), monocyte chemoattractant protein-1, intercellular adhesion molecule-1, interleukin 6 and inducible protein-10, but not of sVEGFR2, compared with poor responders.

Conclusions Response to ranibizumab treatment for DME appears to be associated with aqueous concentrations of VEGFR1 family and certain inflammatory cytokines, but not with clinical parameters.

Statistics from Altmetric.com

Diabetic retinopathy is the leading cause of blindness and visual disability in the working age group in economically developed societies.1 Diabetic macular oedema (DME) is a sight-threatening complication and the most common cause of visual loss in diabetic patients, with a prevalence of 6.81% among diabetic patients.2

Based on the recent scientific evidence implicating vascular endothelial growth factor (VEGF) in the pathogenesis of DME,3 anti-VEGF treatment for DME has been established.4 ,5 Studies have shown that anti-VEGF therapy is effective in reducing macular oedema and improving visual function, and has additional advantages of safety (not increasing intraocular pressure) and rapid response.6 ,7 Conventionally, three consecutive monthly injections of intravitreal ranibizumab (IVR) are given in the initial loading phase, and additional injections are decided according to the clinician's opinion based on functional and/or morphological assessments.8 In the clinical setting, while some patients achieve complete regression of DME shortly after the first IVR injection, others do not respond even after the six consecutive injections. Thus, treatment response to ranibizumab is case dependent, and possibly related to intraocular concentrations of some cytokines.

In this prospective study, we measured baseline aqueous concentrations of cytokines that were reported to be related to DME activity.9 ,10 We analysed the relationship between these concentrations and the number of IVR injections required to achieve and maintain anatomical resolution of DME.

Methods

This prospective study was conducted in accordance with the Institutional Guidelines of Tokyo Medical University Hachioji Medical Center Clinical Research Ethics Committee and was approved by the Institutional Review Board prior to the study (#H-49). The procedures conformed to the tenets of the World Medical Association's Declaration of Helsinki. Informed consent was obtained from each of the patients after they were provided sufficient information on the procedures to be used.

Patients eligibility

Type II diabetic patients with treatment-naïve, centre-involved diffuse DME were recruited in this study from January 2013 to December 2015 in our outpatient clinic. Centre-involved diffuse DME was defined as central macular thickness (CMT) greater than 400 μm obtained from optical coherence tomography (OCT) mapping program data, best-corrected logMAR visual acuity (VA) between 1.0 and 0.3, and lower leakage originated from leaking microaneurysms.8 Eyes with severe proliferative retinopathy or severe cataract were excluded. If both eyes in the same patient were eligible, the more severe eye was selected. Patients who had systemic disorders except hypertension and hypercholesterolaemia were excluded.

Clinical parameters

All patients received a comprehensive ocular examination before and after the treatment. VA was measured using the logMAR chart (5 m) (NEITZ LVC-10, Tokyo, Japan) and CMT was measured using Spectralis OCT (Heidelberg Engineering, Heidelberg, Germany) during follow-up examinations. The Spectralis OCT mapping images were generated using the currently available Spectralis software. The CMT was obtained from the thickness of the 1 mm central field on the OCT mapping image.

Intravitral injection

At the time of intravitreal injection, topical anaesthesia was induced by applying 0.4% oxybuprocaine (Benoxil; Santen Pharmaceutical Co., Osaka, Japan) at least three times. A topical antimicrobial drug, gatifloxacin hydrate 0.3% ophthalmic solution (Gatiflo; Senju Pharmaceutical Co., Osaka, Japan), was administered four times/day in both eyes for at least 1 week after each drug injection. Aqueous humour was sampled just before the first IVR injection. Following disinfection and draping, a mean volume of 0.1 mL of aqueous humour was collected carefully by anterior chamber limbal paracentesis using a 30 G needle attached to an insulin syringe. Subsequently, a 0.05 mL volume of 0.5 mg of ranibizumab (Lucentis, Novartis Pharma KK, Tokyo, Japan) was injected into the vitreous cavity using a sharp 30 G needle at a distance of 3.5 mm from the limbus. Immediately after collection, aqueous humour samples were transferred to sterile plastic tubes and stored at −80°C until analysis.

Study design and definition of treatment response and recurrence

All eligible eyes received the initial IVR injection. The clinical course was monitored by measuring CMT and VA every 4 weeks up to 24 weeks of the study period. IVR was given monthly until CMT decreased to below 300 μm. Thereafter, additional IVR was given as appropriate to maintain CMT below 300 μm during the 6-month period.

The number of IVR injections required to reduce CMT to below 300 μm (initial loading injections) and the total number of IVR injections (loading and maintenance injections) given during the study period of 24 months were recorded.

Measurement of cytokines and growth factors

The methods of measurement of cytokines and growth factors used in this study were described previously.11 ,12 Briefly, a suspension array technology (xMAP; Luminex. Austin, Texas, USA) was used. Capture bead kits (Beadlyte; Upstate Biotechnology, Lake Placid, New York, USA) were employed for the detection of including VEGF, placenta growth factor (PlGF), soluble VEGF receptor (sVEGFR)-1, sVEGFR-2, monocyte chemoattractant protein 1 (MCP-1), intercellular adhesion molecule-1 (ICAM-1), platelet-derived growth factor (PDGF)-AA, interleukin (IL)-6, IL-8 and interferon gamma-inducible protein (IP)-10. Undiluted aqueous humour samples (25 μL) were incubated overnight (16–18 hours) for PlGF and ICAM-1 assays or for 2 hours for assays of other factors. Kits were used according to the manufacturer's instructions. Standard curve for each factor was generated (in duplicate) using the reference standards supplied in each kit. All incubation steps were performed at room temperature and in the dark. Samples were read using a suspension array system. To avoid between-run imprecision, we measured all samples from a single patient collected before and after intervention in the same run. Control samples were included in all runs. The levels of these factors in the aqueous humour samples were within the detection ranges of the assays, with the minimum detectable concentrations being 1.59 pg/mL for sVEGFR-1, 44.81 pg/mL for sVEGFR-2, 0.64 pg/mL for VEGF, 0.37 pg/mL for PlGF, 0.03 ng/mL for soluble ICAM-1, 1.2 pg/mL for MCP-1, 0.64 pg/mL for PDGF-AA, 0.29 pg/mL for IL-6, 0.14 pg/mL for IL-8 and 0.12 pg/mL for IP-10.

Statistical analyses

The data are presented as mean±SD. Analyses were performed with SAS System 9.3 software (SAS Institute, Cary, North Carolina, USA). The minimum required sample size from the results of this study (probability level=0.05, desired statistical power level=0.8, mean change to be detected between the groups=1.0, expected SD=1.5, the number of groups=6) was 59 eligible eyes in this study. Statistical differences between predrug and postdrug application clinical data were assessed by Student's paired t-test. To compare baseline aqueous concentrations or clinical parameters among six groups classified by the number of initial loading injections, Kruskal-Wallis one-way analysis of variance (ANOVA) on ranks test was performed. Two-tailed p values of <0.05 were considered to indicate statistical significance.

Results

Sixty-eight patients with DME were recruited and completed the study. The age of the patients averaged 67.2±5.5 (range 57–79) years old. All patients had type 2 diabetes, and the mean duration of diabetes averaged 9.6±2.9 (range 4–16) years. Forty-nine of 68 (72.1%) patients had a history of hypertension (table 1). During the clinical course, HbA1c and tChL did not exceed 8.0% and 250 mg/dL, respectively, in all patients, and blood pressure was well controlled by the patient's internists.

Table 1

Baseline patient characteristics

Number of IVR injections required to resolve DME

In each patient, after comprehensive ocular examination, monthly IVR injections were given until achieving CMT below 300 μm (initial loading injections). The numbers of injections ranged from 1 to 6 in all eyes, and averaged 2.84±1.87. Eyes were classified into six groups according to the number of initial loading injections required to resolve DME (figure 1A). Twenty-four of 68 (35.3%) eyes showed prompt regression of DME after the first IVR injection, while 12 (17.6%) eyes had persistent DME even after the sixth injection.

Figure 1

Distributions of the number of intravitreal ranibizumab (IVR) injections required initially to resolve diabetic macular oedema (A) and total number of IVR injections received during the 6-month study period (B). (A) Sixty-eight eyes were classified by the number of IVR injections required to reduce central macular thickness (CMT) to below 300 μm during 6 months (initial loading injections). White bar indicates eyes that achieved CMT <300 μm after the first IVR injection. Black bar indicates eyes that did not achieve the goal even after 6 consecutive IVR injections. Grey (light to dark) bars indicate eyes that required two to five IVR injections. Circle chart shows proportions of six groups. (B) Graph shows the distribution of total number of IVR injections (loading+maintenance injections) received during 6 months when eyes were classified by the number of initial loading injections as in (A). Bars are coded as in (A).

During the clinical course of 6 months, the total number of IVR injections required to achieve and maintain CMT below 300 μm (one to six times) in all eyes averaged 3.23±1.32. Figure 1B shows the total number of IVR injections given in the 6-month study period when the eyes were classified by the number of initial loading injections required to resolve DME. Of 24 eyes showing prompt regression after the first injection (figure 1A), seven eyes maintained DME resolution during the clinical course without further injection, while seven eye required one more (total 2 times), six eyes required two more (total 3 times) and four eyes required three more (total 4 times) injections (figure 1B). Fifty-six (82.4%) eyes achieved CMT below 300 μm during the clinical course of 24 weeks.

Clinical outcome after 6 months of IVR therapy

Just before the first IVR injection, mean VA (logMAR) and CMT were 0.58±0.24 and 557.2±147.9 μm, respectively, in all eyes. After 6 months of IVR therapy, both VA and CMT improved significantly to 0.44±0.32 and 305.5±101.5 μm, respectively (both p<0.001). The mean improvement in VA was −0.14±0.19, with 19 of 68 eyes (27.9%) showing improvement≤−0.3 while 12 eyes (17.6%) showing >0. The mean improvement in CMT was 42.9±19.4% (p<0.001).

Treatment response and initial clinical parameters

The baseline characteristics of eligible patients were classified by the number of initial loading IVR injections required to resolve DME (table 1), but found no significant differences of each systemic factor among six groups.

While, according to this classification, there was no significant difference in baseline VA among the six groups (p=0.954) (figure 2A). In contrast, significant differences in VA at 6 months were observed among groups, and the group requiring six consecutive injections showed significantly worse final VA than the group that required a single injection (p=0.033) (figure 2B). Significant differences in VA improvement were also observed among groups, and the group that required a single injection was significant better than the groups that required four and six injections (p<0.001) (figure 2C).

Figure 2

Changes in visual acuity (VA) and central macular thickness (CMT) from baseline to after 6 months of intravitreal ranibizumab (IVR) therapy. (A) Box and whiskers plots showing median and 25th and 75th percentiles of VA at presentation in six group classified by the number of initial loading injections required to resolve diabetic macular oedema. (B) Box and whiskers plots of VA at the final visit after 6-month IVR therapy in six groups. (C) Change in VA (ΔVA) in six groups. Line connecting two groups indicates significant difference. (D) Box and whiskers plots showing median and 25th and 75th percentiles of CMT at presentation in six group classified by the number of initial loading injections required to resolve diabetic macular oedema. (E) Box and whiskers plots of CMT at the final visit after 6-month IVR therapy in six groups. (F) Change in CMT (%) in six groups. Line connecting two groups indicates significant difference.

Also, there was no significant difference in baseline CMT among the 6 groups (p=0.786) (figure 2D). In contrast, there were significant differences in CMT at 6 months among groups, and the group that required 6 consecutive injections were significant worse than the groups that required 1–3 injections (all p<0.001) (figure 2E). Significant differences in CMT improvement were also observed among groups, and the group that required six consecutive injections group was significantly worse than the other five groups (all p<0.001) (figure 2F).

Treatment response and initial biological parameters

Figures 3 and 4 show the relations between the number of initial loading IVR injections required to resolve DME and baseline aqueous humour levels of some cytokines. Multivariate analysis revealed significant differences between the group that required a single injection and the group that required six consecutive injections in aqueous VEGF (p=0.023), PlGF (p=0.030), sVEGFR1 (p=0.019), MCP-1 (p=0.030), ICAM-1 (p=0.020), IL-6 (p=0.021) and IP-10 (p=0.036), but not in sVEGFR2 (p=0.87), IL-8 (p=0.70) or PDGF-AA (p=0.827).

Figure 3

Aqueous humour levels of vascular endothelial growth factor (VEGF) family. Aqueous levels (mean±SD) of VEGF (A), placenta growth factor (PlGF) (B), soluble VEGF receptor-1 (VEGFR1) (C) and VEGFR2 (D) in six groups classified by the number of initial loading injections required to resolve diabetic macular oedema. Line connecting two groups indicates significant difference.

Figure 4

Aqueous humour levels of inflammatory cytokines. Aqueous levels (mean±SD) of monocyte chemoattractant protein (MCP)-1 (A), intercellular adhesion molecule (ICAM)-1 (B), interleukin (IL)-6 (C), IL-8 (D), platelet-derived growth factor (PDGF)-AA (E) and inducible protein (IP)-10 (F) in six groups classified by the number of initial loading injections required to resolve diabetic macular oedema. Line connecting two groups indicates significant difference.

Discussion

In this study, we showed some interesting findings about anti-VEGF treatment for DME. Response to IVR for DME, defined by the number of initial loading IVR injections required to regress DME, was diverse and case sensitive, and was associated with some cytokines and growth factors related to inflammation.

Compared with previous studies,8 ,13 ,14 the treatment protocol in this study was unique. We used a morphology-based reinjection protocol because VA is a subjective index which is sometimes affected by psychological factors. In our study protocol, VA gain of more than logMAR 0.3 (equivalent to gain of >15 letters) was achieved in 27.9% of DME eyes and the mean change in VA was −0.14 (equivalent to seven-letter gain) after 6 months of IVR therapy. Compared with previous studies in which 22.6%–30% eyes achieved VA gain of more than 15 letters and mean VA improvement of 5.9–9.0 letters after 12 months,8 ,15 our protocol appears to have the same efficacy. It is noteworthy that 3.23±1.32 IVR injections were given during 6 months. According to previous studies of IVR for DME,8 ,15 all patients received three initial consecutive monthly IVR injections, and the total numbers of IVR injections during 6 months were 4.8 to 5.0.14 Although the significance of continuous IVR injection for ‘stabilisation’ in the loading phase cannot be denied, our treatment protocol without routine loading injections may have the advantage of reducing the number of IVR injections.

In our study, 24 of 68 (35.3%) eyes showed prompt regression of DME after the first IVR injection, and 7 of the 24 eyes did not require further IVR injection during the 6-month study period. For these good responders, initial routine IVR injections could be an overtreatment. On the other hand, 12 of 68 (17.6%) eyes had persisted DME even after six consecutive injections, and these eyes were significantly worse in VA and CMT changes compared with the other eyes. For these poor responders, other treatments for DME such as triamcinolone, grid photocoagulation and vitrectomy should be considered.

In this study, although it is hard to calculate the relationships between response to ranibizumab and aqueous levels of biological factors because the number of initial loading IVR injections was qualitative parameter, there were significant differences in baseline aqueous levels of some biological factors between good and poor responders. In good responders, aqueous levels of VEGF, PlGF and sVEGFR1 were higher than those in poor responders. It is not surprising because high level of VEGF causes vascular hyperpermeability,16 and thus suppression of VEGF by IVR results in better resolution of oedema. It is interesting that only aqueous sVEGFR2 level did not show a significant difference. Previous study indicated that breakdown of the intercellular junction of blood–retinal barrier correlated closely with VEGFR1 activity but not with VEGFR2.17 Thus, IVR-responsive DME may reflect VEGFR1-mediated cell signalling which is activated by VEGF and PlGF.18 The pathogenesis of DME is multifactorial, involving vascular hyperpermeability and leakage from neovascularisation or microaneurysms, chronic inflammation and vitreomacular traction.19 Our results suggest that the main cause of DME in IVR-responsive DME may be VEGFR1-mediated vascular hyperpermeability.

In addition, we also found significantly higher levels of some inflammatory cytokines in IVR-responsive cases. In this study, eyes with a history of photocoagulation and inflammatory disease were excluded. Thus, increases in these cytokines were induced intrinsically. Although the reason why some but not all the inflammatory cytokines were related to response to IVR remains unclear, previous studies indicated that vitreous levels of inflammatory cytokines including MCP-1, ICAM-1 and IL-6 correlated with the level of VEGF in DME.9 ,10 Thus, higher levels of VEGF together with inflammation-related cytokines are novel characteristics in the pathogenesis of IVR-responsive DME.

According to previous studies,20 ,21 aqueous levels of some cytokines correlate with disease severity in DME, and hypoxia caused by fall in perfusion pressure in the retinal capillaries may be playing an important role in the development of DME.22 Recent study indicated that the size of hypoxic retina correlated significantly with aqueous VEGF levels.23 In this study, although the size of hypoxic retina including non-perfusion area or retinal neovascularisation was not analysed, severity of DME may correlate with sensitivity to anti-VEGF drugs. Further study in future will reveal this hypothesis.

According to this study results, our morphology-based individualised IVR injection protocol without initial routine IVR injections for DME was comparable in efficacy to the conventional regimen, while response to IVR was case dependent and was associated with aqueous levels of VEGF family but not clinical parameters. Although this study is an exploratory research with a short study period of 6 months, the results suggest the usefulness of individualised treatment for DME. Larger scale studies with longer follow-up are required to confirm our hypothesis.

Acknowledgments

The authors would like to thank Teresa Nakatani for her helpful comments and assistance in language editing.

References

View Abstract

Footnotes

  • Contributors MS: the conception and design of the work. KY, RM and OK: the acquisition of data. HN and MS: the analysis of data. All authors: the interpretation of data and discussion, final approval of the version to be published. MS: writing manuscript.

  • Funding This work is supported by JSPS KAKENHI Grant Number JP25462737 and 2013 Scientific Grant in aid for Clinical Research about age-related ocular disease.

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval Tokyo Medical University Hachioji Medical Center Clinical Research Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.