Aqueous humour proteins and treatment outcomes of anti-VEGF therapy in neovascular age-related macular degeneration.

We aimed to construct a better model for predicting treatment outcomes of anti-vascular endothelial growth factor therapy for neovascular age-related macular degeneration (nAMD) using the concentrations of aqueous humour proteins at baseline and during treatment. From the data of 48 treatment-naïve nAMD eyes that received intravitreal ranibizumab pro re nata for up to 12 months, we used the aqueous humour concentrations of C-X-C motif chemokine ligand 1 (CXCL1), CXCL12, CXCL13, interferon-γ-induced protein 10, monocyte chemoattractant protein 1 (MCP-1), C-C motif chemokine ligand 11, interleukin 6 (IL-6), IL-10, and matrix metalloproteinase 9 (MMP-9). After stepwise regression, multivariate analysis was performed to identify which predictors were significantly associated with best-corrected visual acuity (BCVA) changes and the number of injections. The results demonstrated that besides male sex (β coefficient = -0.088, P = 0.040) and central retinal thickness (β coefficient = 0.00051 per μm, P = 0.027), MCP-1 (β coefficient = 0.44, P < 0.001) and IL-10 (β coefficient = -0.16, P = 0.033) were significantly correlated with baseline BCVA. Additionally, high MCP-1 at baseline (β coefficient = -0.20, P = 0.015) and low CXCL13 at baseline (β coefficient = 0.10, P = 0.0054) were independently associated with better BCVA change at 12 months. High MMP-9 at the first injection (β coefficient = 0.56, P = 0.01), CXCL12 at the third injection (β coefficient = 0.10, P = 0.0002), and IL-10 at the third injection (β coefficient = 1.3, P = 0.001) were predictor variables associated with the increased number of injections. In conclusion, aqueous humour protein concentrations may have predictive abilities of BCVA change over 12 months and the number of injections in pro re nata treatment of exudative nAMD.

Introduction

Many biomarker studies have been conducted to investigate the associations between the intraocular concentrations of anti/pro-angiogenic factors and treatment outcomes in neovascular age-related macular degeneration (nAMD), with a hope to improve prognostic accuracy. For example, several studies have demonstrated that aqueous humour protein concentrations may help differentiate good responders from poor responders in the setting of anti-vascular endothelial growth factor (VEGF) treatment [1-3]; however, the prognostic value of such aqueous humour protein concentrations in clinical care remains under debate. Besides the information obtained at presentation, additional information acquired after/during initial dosing and maintenance therapy may help predict long-term treatment outcomes; therefore, the importance of assessing disease activity after the loading dose has been highlighted in recent studies [4-11]. For example, visual and anatomical outcomes up to 12 weeks are associated with long-term treatment outcomes such as visual acuity and treatment frequency [4-11]. Some studies [12,13], including ours [13], measured aqueous humour proteins during treatment and demonstrated their change after treatment. As such, the investigation of aqueous humour proteins at baseline and after the loading phase may be useful to better predict treatment outcome. Furthermore, information is also lacking on whether cytokine concentrations in the aqueous humour are correlated with visual acuity (VA). Thus far, no studies have used aqueous humour samples at baseline and later visit(s) to investigate the relationship between these concentrations and VA at baseline and treatment outcome.

We previously investigated changes in the concentrations of multiple aqueous humour proteins during the induction phase, i.e. before and after the initial 3 monthly consecutive injection phase, of ranibizumab treatment in patients with nAMD, using initial samples [13]. Among C-X-C motif chemokine ligand 1 (CXCL1), CXCL12, CXCL13, interferon-γ-induced protein 10 (IP-10), monocyte chemoattractant protein 1 (MCP-1), C-C motif chemokine ligand 11 (CCL11), interleukin 6 (IL-6), IL-10, and matrix metalloproteinase 9 (MMP-9), selected based on previous clinical and animal studies [12-14], we found that pro-inflammatory cytokines are elevated in the aqueous humour of nAMD patients, while MMP-9 levels are decreased [13]. We also found that after the induction phase, inflammatory cytokine levels in the aqueous humour were strongly suppressed, while MMP-9 levels are increased. No cytokine was significantly different between type 1 and 2 choroidal neovascularization (CNV).

In the present study, we built on our previous work and aimed to clarify whether we could construct a better model for predicting treatment outcomes of anti-VEGF therapy for nAMD by using the concentrations of aqueous humour proteins at baseline and during treatment. We used stepwise regression followed by multiple regression analysis to investigate the relationship between the concentrations of various cytokines and VA at baseline and at 2 and 12 months after treatment, as well as the number of injections during the first year of anti-VEGF monotherapy with a standardised pro re nata treatment regimen.

Materials and methods

Study design and approval

We used data from our previous prospective study [13]. The present study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all patients. Institutional Review Board approval was obtained from the Japan Community Health Care Organization, Tokyo Shinjuku Medical Center.

Data collection

The study procedure has been described previously [13]. In brief, we analysed 48 eyes of 48 treatment-naïve patients with nAMD (typical nAMD, 21 eyes; polypoidal choroidal vasculopathy [PCV], 27 eyes) with a mean age of 72.9 years. The patients received three consecutive intravitreal injections of ranibizumab monthly from November 2010 to August 2012 at the Japan Community Health Care Organization Tokyo Shinjuku Medical Center. Patients with other ocular diseases, including glaucoma and anterior uveitis, were excluded. The comparison of cytokine concentrations before the first and third intravitreal ranibizumab injections was published elsewhere [13].

Samples of aqueous humour (approximately 0.2 mL) were collected by manual aspiration into disposable syringes through a clear corneal paracentesis under an operating microscope immediately before each injection. The samples were transferred immediately to sterile tubes, snap-frozen, and stored at −80°C until required.

The concentrations of the following aqueous humour proteins were determined using a multiplex cytokine assay (Procarta® Cytokine Assay Kit; Affymetrix, Inc., Santa Clara, CA) according to the manufacturer’s instructions: CXCL1, IP-10, CXCL12, CXCL13, MCP-1, CCL11, IL-6, IL-10, and MMP-9. We imputed the concentration of half the minimal detection limit when the concentration was lower than the detection limit. Because aqueous protein concentrations have been reported to not reliably reflect vitreous cytokine concentrations [14], 5 vitreous humour samples were examined preliminarily for the presence of the 9 proteins. The presence of all studied proteins in the vitreous humour was confirmed, and the mean (95% confidence interval) concentrations were 5.7 (2.3–13.5), 87 (47–159), 370 (240–570), 12 (5–28), 567 (436–737), 4.4 (3.0–6.4), 4.6 (1.7–12.5), 0.20 (0.10–0.36), and 1.6 (0.2–11.2) pg/mL for CXCL1, IP-10, CXCL12, CXCL13, MCP-1, CCL11, IL-6, IL-10, and MMP-9, respectively. We confirmed that dye angiography did not affect the measurements (data not shown). The concentration of VEGF, as measured using an enzyme-linked immunosorbent assay (Quantikine® Human VEGF Immunoassay; R&D Systems, Inc., Minneapolis, MN), was available in 29 samples, as has been reported elsewhere [13]. We first subjected the samples to multiplex analysis and then the leftover was subjected to ELISA. Therefore, we were able to measure VEGF concentrations in 29 (60%) cases prior to anti-VEGF therapy.

Clinical examinations

To diagnose typical nAMD, PCV [15], and retinal angiomatous proliferation, both fluorescein and indocyanine green angiography were performed routinely for nAMD cases at our institution, with some exceptions (e.g. contraindications due to drug allergy, liver dysfunction, or recent cerebrovascular event) [13]. For the measurement of axial length, A-mode ultrasonography (UD-6000; Tomey Corp., Aichi, Japan) was used. B-mode ultrasonography (UD-6000) was used To examine posterior vitreous detachment (PVD), as previously described [16-21], and the eyes were either categorised into the PVD group or the without PVD group.

All patients received 3 monthly ranibizumab (0.5 mg/0.05 mL) injections followed by pro re nata injections [9]. Each patient underwent a best-corrected VA (BCVA) test with a 5 M Landolt C VA chart (i.e. decimal VA chart) and spectral domain optical coherence tomography (OCT; Cirrus HD-OCT Model 4000; Carl Zeiss Meditec AG, Oberkochen, Germany) at the initial and each follow-up visit. The patients were followed up on a monthly basis, and retreatment was given based on the qualitative assessment of OCT; if any persistent or recurrent intra-retinal or subretinal fluid and/or clinically detectable retinal haemorrhage was observed, an injection was recommended. We did not include BCVA loss in the retreatment criteria.

BCVA was converted to a logarithm of minimum angle of resolution (logMAR) values for statistical analysis. Greatest linear dimension (GLD) was determined based on fluorescein angiography. Disease duration was based on the patient’s symptoms. Central retinal thickness (CRT) was determined from the distance between the inner limiting membrane and Bruch’s membrane. Central choroidal thickness (CCT) was defined as the distance from Bruch’s membrane to the choroid-scleral border. The values for thickness were determined manually at the foveal centre using the OCT calliper function.

Statistical analysis

Statistical analysis was performed using JMP Pro software version 11.2.0 (SAS Institute, Cary, NC). A two-sided paired t-test was performed to compare the changes of BCVA and CRT at baseline and at 12 months. Associations between baseline factors, cytokine concentrations, and visual and anatomical outcomes were examined with Pearson’s correlation coefficient after a normal distribution was confirmed with Shapiro-Wilk’s W-test or Spearman’s rho for categorical variables. Stepwise variable selection (using minimum Bayesian information criterion, increasing the number of variables) was performed using baseline BCVA, BCVA change at 12 months, number of ranibizumab injections, and CRT and CCT (at baseline and change at 2 months) as response variables, and disease type, disease duration (months), sex, age, axial length (mm), PVD, GLD (μm), CRT (μm), CCT (μm), BCVA, and each log concentration of aqueous humour proteins as predictor variables. The concentrations of aqueous humour proteins were log-transformed because of the lognormal distribution of these variables. To select the predictor variables in multiple regression analysis, we used the non-fixed stepwise method due to its objectivity and to avoid the risk of multicollinearity [22]. We then used multiple regression analysis after stepwise variable selection. P < 0.05 was considered to be significant.

Results

BCVA and central retinal thickness over 1 year of treatment

After receiving 3 monthly intravitreal injections of ranibizumab, all 48 eyes received pro re nata intravitreal injections of ranibizumab. The changes in BCVA and CRT over 12 months are shown in Fig 1. BCVA and CRT improved by 0.12 logMAR units and 180 μm after treatment, respectively (both P < 0.01).

Fig 1

Best-corrected visual acuity (BCVA) and central retinal thickness (CRT) over 1 year of treatment in neovascular age-related macular degeneration (nAMD).

Forty-eight eyes with nAMD received 3 monthly intravitreal injections of ranibizumab followed by pro re nata intravitreal injections of ranibizumab. a: Last mean BCVA (0.29) was better than initial mean BCVA (0.41) (paired t test. P = 0.0028). b: Last mean CRT (300 μm) was better than initial mean CRT (480 μm) (paired t test. P < 0.0001).

Factors associated with baseline VA

At baseline, worse BCVA was associated with increased CRT (r = 0.43, P = 0.020), IP-10 (r = 0.44, P = 0.016), MCP-1 (r = 0.54, P = 0.0025), and IL-6 (r = 0.44, P = 0.018) in univariate analysis (Table 1). Multiple regression analysis after stepwise variable selection demonstrated that besides male sex (β coefficient = −0.088, i.e. regression coefficient was −0.088 logMAR for men [as 1] vs women [as 0], P = 0.040) and CRT (β coefficient = 0.00051 logMAR per μm, P = 0.027), logMCP-1 (β coefficient = 0.44, i.e. baseline BCVA worsened by 0.44 logMAR per logMCP-1, P < 0.001) and logIL-10 (β coefficient = −0.16, P = 0.033) were independently associated with baseline BCVA (Table 1).

Table 1

Factors associated with baseline visual acuity.

	BCVA (logMAR)
	Univariate	Multivariate
Age (years)	−0.070 (0.71)
Sex (male)	−0.20 (0.058)	−0.088 (−0.170 to −0.007, 0.040)*
BCVA (logMAR)
Disease type (PCV)	0.028 (0.79)
Duration of disease (months)	−0.043 (0.82)
GLD (μm)	0.25 (0.19)
PVD (+)	0.024 (0.82)
Axial length (mm)	−0.063 (0.74)
CRT (μm)	0.43 (0.020)*	0.00051 (0.00007 to 0.00094, 0.027)*
CCT (μm)	0.15 (0.42)	0.00083 (−0.00016 to 0.00181, 0.11)
VEGF (pg/mL)	0.13(0.50)
CXCL1 (pg/mL)	0.14 (0.46)
IP-10 (pg/mL)	0.44 (0.016)*
CXCL12 (pg/mL)	0.20 (0.30)
CXCL13 (pg/mL)	0.27 (0.16)
MCP-1 (pg/mL)	0.54 (0.0025)*	0.44 (0.20 to 0.68, <0.001)*
CCL11 (pg/mL)	0.18 (0.35)
IL-6 (pg/mL)	0.44 (0.018)*
IL-10 (pg/mL)	−0.13 (0.50)	−0.16 (−0.31 to −0.02, 0.033)*
MMP-9 (pg/mL)	0.044 (0.82)
R2		0.49

Univariate: Pearson’s correlation (for continuous variables) and Spearman’s rho (for categorical variables). R (P value).

Multivariate analysis was performed after stepwise variable selection (BIC, forward method). β Coefficient (95% confidence interval, P value).

*P < 0.05.

BCVA: best-corrected visual acuity; CCL11: C-C motif chemokine ligand 11; CCT: central choroidal thickness; CRT: central retinal thickness; CXCL1: C-X-C motif chemokine ligand 1; CXCL12: C-X-C motif chemokine ligand 12; CXCL13: C-X-C motif chemokine ligand 13; GLD: greatest linear dimension; IL-6: interleukin 6; IL-10: interleukin 10; IP-10: interferon-γ-induced protein 10; MCP-1: monocyte chemoattractant protein 1; MMP-9: matrix metalloproteinase 9; PCV: polypoidal choroidal vasculopathy; PVD: posterior vitreous detachment; VEGF: anti-vascular endothelial growth factor.

Factors associated with BCVA at 12 months

To investigate the factors associated with treatment outcome at 12 months, we used demographic characteristics, OCT parameters at baseline and at 2 months, and aqueous humour biomarker concentrations at baseline and at 2 months. An improvement of BCVA was associated with worse BCVA at baseline (r = -0.31, P = 0.031) and high MCP-1 at baseline (r = −0.29, P = 0.042) in univariate analysis (Table 2). Stepwise selection demonstrated that the parameters at baseline and 2 months were independently associated with the change in BCVA. Among the baseline demographic factors, the presence of PCV (β coefficient = -0.095, P = 0.00076) and smaller GLD (β coefficient = 4.3e−5, P = 0.00068) were associated with better BCVA change at 12 months (Table 2). Among the cytokine concentrations at baseline and at 2 months, high MCP-1 at baseline (β coefficient = −0.20, P = 0.015) and low CXCL13 at baseline (β coefficient = 0.10, P = 0.0054) were also independently associated with better BCVA change at 12 months. Overall, 69% of the BCVA change over 12 months could be explained by the combination of these factors (R2 = 0.69) (Table 2).

Table 2

Factors associated with best-corrected visual acuity change at 12 months.

	BCVA (logMAR)
	Univariate	Multivariate
Age (years)	−0.13 (0.49)
Sex (male)	0.14 (0.10)
BCVA (logMAR)	−0.31 (0.031)
BCVA change at 2 months (logMAR)	0.67 (<0.0001)	0.86 (0.63 to –1.08, <0.0001)*
Disease type (PCV)	−0.088 (0.30)	-0.095 (0.044 to 0.146, 0.0008)*
Duration of disease (months)	−0.017 (0.93)
GLD (μm)	0.28 (0.14)	4.3e-5 (2.0e-5 to 6.5e-5, 0.0007)*
PVD (+)	−0.11 (0.20)
Axial length (mm)	−0.14 (0.48)
CRT (μm)	0.040 (0.84)
CRT change at 2 months (μm)	0.34 (<0.019)
CCT (μm)	0.18 (0.36)
CCT change at 2 months (μm)	−0.13 (0.36)
VEGF (pg/mL)	−0.18 (0.36)
CXCL1 (pg/mL)	−0.25 (0.18)
CXCL12 (pg/mL)	−0.15 (0.43)
CXCL13 (pg/mL)	−0.018 (0.93)	0.10 (0.03 to 0.17, 0.0054)*
IP-10 (pg/mL)	−0.048 (0.80)
MCP-1 (pg/mL)	−0.29 (0.042)	−0.20 (−0.34 to −0.05, 0.015)*
CCL11 (pg/mL)	−0.14 (0.45)
IL-6 (pg/mL)	−0.11 (0.58)
IL-10 (pg/mL)	0.12 (0.53)
MMP-9 (pg/mL)	−0.024 (0.90)
R2		0.69

Univariate analysis: Pearson’s correlation (for continuous variables) and Spearman’s rho (for categorical variables). R (P value).

Multivariate analysis was performed after stepwise variable selection (BIC, forward method). β Coefficient (95% confidence interval, P value).

*P < 0.05.

BCVA: best-corrected visual acuity; CCL11: C-C motif chemokine ligand 11; CCT: central choroidal thickness; CRT: central retinal thickness; CXCL1: C-X-C motif chemokine ligand 1; CXCL12: C-X-C motif chemokine ligand 12; CXCL13: C-X-C motif chemokine ligand 13; GLD: greatest linear dimension; IL-6: interleukin 6; IL-10: interleukin 10; IP-10: interferon-γ-induced protein 10; MCP-1: monocyte chemoattractant protein 1; MMP-9: matrix metalloproteinase 9; PCV: polypoidal choroidal vasculopathy; PVD: posterior vitreous detachment; VEGF: anti-vascular endothelial growth factor.

Factors associated with the number of injections within 12 months

Factors associated with the number of ranibizumab injections were also investigated in univariate analysis, which demonstrated that only CXCL12 at the third injection (r = -0.40, P = 0.0045) was associated with the number of injections among the baseline characteristics, baseline OCT parameters, baseline cytokines, and 3-month cytokine concentrations. However, stepwise selection and multiple regression analysis demonstrated that the predictor variables initial BCVA (β coefficient = −1.5, P = 0.04), high MMP-9 at the first injection (β coefficient = 0.56, P = 0.01), CXCL12 at the third injection (β coefficient = 0.10, P = 0.0002), and IL-10 at the third injection (β coefficient = 1.3, P = 0.001) were associated with the increased number of injections (R2 = 0.38). Importantly, clinical and demographical parameters excluding aqueous humour cytokines were not associated with the number of injections (Fig 2).

Fig 2

Correlation between the number of injections and cytokine concentrations.

a: CXCL1, b: IP-10, d: CXCL13, e: MCP-1, f: CCL11, g: IL-6, and j: VEGF concentrations did not significantly correlate with the number of injections. c: CXCL12 concentration at the third injection was correlated with the number of injections in univariate analysis (P = 0.0045) and predicted the number of injections in multivariate analysis (P = 0.0002). h: IL-10 concentration at the third injection was a predictor variable that was positively associated the number of injections in multivariate analysis (P = 0.001). i: MMP-9 concentration at the first injection predicted the number of injections in multivariate analysis (P = 0.01).

Discussion

Anti-VEGF therapy has achieved unprecedented success in the treatment of nAMD [23]. Angiogenesis and vascular hyperpermeability in nAMD are caused mainly by VEGF; however, the long-term management of exudative nAMD with anti-VEGF therapy remains challenging [24,25], partly because of significant contributions from other factors [26]. As we and others have shown, eyes with nAMD have elevated concentrations of several cytokines/chemokines (e.g. MCP-1 [27] and C-X-C motif chemokines such as IP-10 [28]) in the aqueous humour. Such molecules partly modulate the activity of CNV directly through pro- and anti-angiogenic activity or by recruiting monocytes, which positively and negatively regulate inflammation by inducing production of other angiogenic/inflammatory cytokines, such as VEGF, IL-6 [29], and IL-10 [30]. Importantly, in addition to cytokines and chemokines, tissue proteases such as MMP-9 [31], and intracellular adhesion molecules such as intercellular adhesion molecule 1 and vascular cell adhesion molecule 1 [32-34], are also overexpressed in CNV and accelerate CNV growth [35]. Such molecules are also detected in the aqueous humour.

In this study, we investigated the associations between the concentrations of 9 aqueous humour proteins (CXCL12, CXCL13, IL-10, IP-10, IL-6, MCP-1, CCL11, CXCL1, and MMP-9) from 48 treatment-naïve eyes with nAMD at baseline and at 2 months after the first ranibizumab injection and BCVA, CRT, CCT, and the number of injections.

Overall, regarding ocular imaging biomarkers, our results confirmed the findings of previous studies. Here, eyes with a thick choroid were relatively resistant to anti-VEGF therapy, similar to previous reports [36]. As expected, the change inBCVA at 12 months was better in eyes with a small GLD and bad initial BCVA.

VA and cytokines

A previous study revealed that MCP-1 was not associated with initial BCVA, but IL-10 was positively correlated with the change of VA [37]. The results of the present study demonstrated that initial BCVA was better in eyes with a lower concentration of MCP-1 and higher IL-10 at baseline.

MCP-1 is a major chemoattractant for inflammatory monocytes and aggravates inflammation [38]. IL-10 is generally considered an anti-inflammatory cytokine that inhibits inflammation and the synthesis of pro-inflammatory cytokines [39]. Considering these biological links, it is rational to surmise that the balance between inflammatory and anti-inflammatory factors such as MCP-1 and IL-10 is an important indicator of inflammation in the setting of nAMD [40,41]. Importantly, we found that these two cytokines were independently associated with BCVA besides CRT. It is generally believed that the increase in MCP-1 attracts ocular macrophages. leading to the digestion of the retinal pigment epithelium and Bruch’s membrane [42]. Histologic examinations of surgical specimens showed macrophages located in regions with atrophy of the retinal pigment epithelium, Bruch’s membrane breakdown, and CNV [43-47]. In contrast, IL-10 is believed to play a significant role in suppressing the formation of subretinal fibrosis [48]. IL-10 is involved in regulating the early and later movement of macrophages in and out of the injured nerve, lowers the expression of proinflammatory chemokines and cytokines, and plays a role in the inducing macrophages to shift from proinflammatory to anti-inflammatory via myelin-phagocytosis [49]. The results of the present study support the idea that MCP-1 and IL-10 control such different aspects of inflammation occurring in eyes with nAMD.

We also demonstrated that higher MCP-1 and lower CXCL13 levels at baseline were associated with a better change of BCVA at 12 months. Considering the aforementioned association showing that eyes with a higher MCP-1 concentration had a worse baseline BCVA, these eyes with higher MCP-1 concentrations at baseline had a better chance to improve BCVA, considering the ceiling effects of VA, which at least in part explains why higher MCP-1 concentrations at baseline were associated with a greater improvement of BCVA. CXCR5, the receptor for CXCL13, may be involved in protecting the retinal pigment epithelium and retinal cells during ageing, and its loss may lead to AMD-like pathological changes in aged mice [50]. However, the role of CXCL13 is less clear in the context of nAMD compared with other cytokines.

Number of injections and cytokines

We assumed that the number of injections was an important parameter that reflects disease activity more than VA, because VA is also affected by numerous factors such as anatomical changes including the integrity of the external limiting membrane and the ellipsoid zone. Here, we showed that cytokine concentrations were strongly associated with the number of injections required for the first year, while clinical and demographic factors were not. Importantly, as shown in this study, other than MMP-9, aqueous humour proteins at baseline might not be associated with the number of injections; however, CXCL12 and IL-10 at 2 months may be useful to predict the number of ranibizumab injections after the initial loading dose. MMP-9 has a role in the control of the extracellular matrix. In the retinal pigment epithelium/choroid, MMP-9 is involved in the remodelling of Bruch’s membrane and controls the permeability of the outer retinal barrier [51]. MMP-9 is considered to be involved in large pore formation in Bruch’s membrane, which is critical for CNV to grow into the subretinal space [52]. Further studies including the balance of extracellular matrix remodelling factors such as TIMP/MMP [53] are required to understand how low concentrations of MMP-9 are associated with the number of injections. CXCL12 is a chemotactic factor for lymphocytes and promotes angiogenesis [54,55]. Interestingly, bone marrow-derived cells distributed in a mouse laser-induced CNV model are CXCL12 receptor-positive [56]. Therefore, the increased concentration of CXCL12 at 2 months may indicate the presence of residual subclinical inflammation/inflammatory cell accumulation after the loading dose. Notably, CXCL12 may be another important molecule involved in CNV activity, as suggested previously [57]. Interestingly, those patients with high IL-10 concentrations at 2 months required more frequent injections. This may be in line with a recent hypothesis that a subpopulation of macrophages, referred to as alternatively activated macrophages, may be involved in nAMD [58]. Inflammation occurring in eyes with nAMD is mediated by several types of immune cell, among which myeloid cells (macrophages/monocytes) are considered to be of the utmost importance. Myeloid cells can be classified into at least two subpopulations, i.e. classically activated and alternatively activated monocytes/macrophages, which control two different aspects of inflammation; in general, the former exacerbates inflammation in the acute phase of inflammation, while the latter promotes angiogenesis in the chronic phase of inflammation, although this depends on the context in which inflammation occurs. IL-10 is produced by alternatively activated macrophages. Therefore, it is tempting to consider that the high levels of IL-10 are associated with an increased number of alternatively activated macrophages and aberrant angiogenesis, and because of such increased angiogenic drive, those patients with high levels of IL-10 at 2 months required an increased number of injections. In this aspect, MCP-1 is indispensable for classically activated macrophages, which accelerate inflammation, but not angiogenesis, and possibly leading to exudative changes. Although further studies are needed to validate our findings, our results underscore the importance of focusing on aqueous humour protein levels after loading injections for prognostication.

Limitations

The current study has several limitations. The sample size was relatively small, only Japanese individuals were involved, and PCV and typical nAMD were mixed. The high heterogeneity of the studied samples as per AMD traits limits the study’s conclusions. This study was also limited by a lack of a standardized and masked treatment protocol. The current results should be validated in other cohorts. However, we believe our results are still relevant considering that most of the patients adhered to the protocol presented and the pro re nata treatment regimen is one of the most commonly employed therapeutic approaches. Further studies are warranted to confirm our results.

Conclusion

Higher MCP-1 and lower CXCL13 concentrations at first injection were predictor variables that were significantly associated better BCVA prognosis at 12 months. Higher concentrations of CXCL12 at the third injection were associated with a smaller number of ranibizumab injections. Additionally, IL-10 concentrations at 2 months was associated with the number of injections. Treatment outcomes can be predicted by investigating the concentrations of these aqueous humour proteins, which are potential biomarkers for precision medicine.

(TIF)

Click here for additional data file.

Factors associated with baseline characteristics.

(DOCX)

Click here for additional data file.

Factors associated with CRT and CCT changes at 2 months.

(DOCX)

Click here for additional data file.

Supplementary information of central retinal thickness and central choroidal thickness.

(DOCX)

Click here for additional data file.

Influence of fluorescein and indocyanine green to measurements using multiplex cytokine assay.

(DOCX)

Click here for additional data file.

17 Sep 2019

PONE-D-19-23157

Inflammatory cytokines and treatment outcomes of anti-VEGF therapy in neovascular age-related macular degeneration

PLOS ONE

Dear Dr. Takahashi,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Both expert reviewers indicated that your submission was interesting but that the results do not fully support your conclusions.  Reviewer 1 indicated flaws in your methodology, particularly statistical analysis and reviewer 2 indicated that your present paper overlaps with your earlier paper, Sakamoto S, Takahashi H, Tan X, et al.

Br J Ophthalmol 2018;102:448–454.  I share this reviewers concerns that you may be presenting the same results in both papers, for example Table 2 in the earlier paper seems to present the same results as Table 1 in the current paper. In rewriting your paper for PLoS One,  please be clear what distinguishes the new paper from the older one.

We would appreciate receiving your revised manuscript by Nov 01 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Alfred S Lewin, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

1. We noticed you have some minor occurrence of overlapping text with the following previous publication(s), which needs to be addressed:

https://bjo.bmj.com/content/102/4/448

https://jamanetwork.com/journals/jamaophthalmology/fullarticle/426210

https://www.ncbi.nlm.nih.gov/pubmed/26674868

In your revision ensure you cite all your sources (including your own works), and quote or rephrase any duplicated text outside the methods section. Further consideration is dependent on these concerns being addressed.

2. Thank you for including your competing interests statement; "Dr Takahashi received lecturer’s fees from Kowa Pharmaceutical, Novartis Pharmaceuticals, Bayer Yakuhin, and Santen Pharmaceuticals, and grants from Bayer Yakuhin and Novartis Pharma, outside this work. Dr Inoue received lecturer’s fees from Kowa Pharmaceuticals, Novartis Pharmaceuticals, Bayer Yakuhin, and Santen Pharmaceuticals outside this work. Dr Kawashima received lecturer’s fees from Kowa Pharmaceutical, Novartis Pharmaceuticals, and Santen Pharmaceuticals outside this work. Dr Yanagi received lecturer’s fees and grants from Santen Pharmaceuticals outside this work. He is an advisory board member for Bayer Pharmaceuticals and a consultant for Santen Pharmaceuticals. Dr Arai, Dr Tan, Dr Inoda, Dr Sakamoto, and Dr Fujino declare no potential conflict of interest."

Please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials, by including the following statement: "This does not alter our adherence to  PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests).  If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include your updated Competing Interests statement in your cover letter; we will change the online submission form on your behalf.

Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This a very interesting study on aqueous humour cytokines as biomarkers of nAMD treatment response as per VA and number of injections at 12 months. Nonetheless, shows a strong rationale and well-intended methodology. However, there are significant flaws in the Methods and Results section that, by instance, should not represent a major problem to solve. In addition, some information is lacking and some, by contrast, does not even need to be presented as it is later forgotten in the discussion. If most points are indeed faced and cleared it could then be ready to undergo a more detailed review.

Point by point considerations:

Abstract: consider a section-based writing instead such as Purpose-Methods-Results-Conclusion. Also, correlation coefficients mean association whereas beta coefficients mean prediction, which is not the same and therefore the conclusion should be rewritten.

Introduction: as a whole, we miss some more referenciated information on the local and systemic importance of inflammation for nAMD which, after all, is the rationale of the present study.

49: "AMD" use should be stated in the very first part of the text, not afterwards

54-55 and 57-58: consider softening such strong statements that don't even show a reference

60-64: explanation of the present study should not be present in the Introduction section

73-74: a reference is lacking in this statement

85: was this informed consent written or oral? Please specify

92: were non-retinal diseases such as glaucoma or anterior uveitis excluded?

95: how was this obtained? Through a clear corneal paracenthesis? On the slit lamp? Please specify

101-102: how where 0 values handled?

103-104: reference that statement

105: how were these vitreous samples obtained? on which basis and time? same subjects? on what inclusion criteria and randomisation?

106-107: need dispersion statistics such as standard deviation or IQR when a mean/median is showed

109: how were these 29 subjects chosen? on which basis?

114-115: does this mean ICG and FA was undertaken in all cases?

125: specify which ranibizumab concentration is used

126: "essentially as described in the CATT", please specify which deviation from CATT protocol this means

127: treatment decisions are therefore relied on the physician, with no protocol adjusted or blind assessment. Since this, conclusions as per visual acuity prognosis at 12 months are unfair as treatment is not defined nor controlled in the group.

129: specify which chart of vision is used

142: Pearson correlation demands a normality assumption, which is not clearly stated in the text, please update accordingly, also referring to used normality tests.

145: log transformation of which variables? cytokines? please specify

157: add dispersion statistics to AL mean

158-159: definition criteria for PCV and type of nAMD is not stated in the manuscript, please update

167-168: paired tests have not been claimed in the methods section, please update

170: are these means? medians? please state so; which statistical test is used?

182: where is this 2.7 coming from?

Table 2, 3, S1 and S2: it is not clear to me whether these are correlation coefficients, beta or standardized beta. Please update accordingly. Also a 95% CI should be informed for all data and a total R square in all tables.

201 and on: "better" is an unclear term, use "improvement" instead.

231: a Table 4 on this topic is also mandatory as previously done

252-255: even though S1/S2 tables are supplied, consider stating the main found coefficients in the manuscript in order to give a clear vision of their magnitudes

301-302: such bivariant statement is not further explained in the manuscript and could be forsaken.

303: a clear discussion on the reason why lower MCP-1 is associated to better BCVA at baseline and, at the same time, higher levels are linked to better BCVA at 12 months is mandatory. This is point is forgotten in the discussion.

314: baseline MMP-9 significant association is not discussed and even goes against this statement.

324: IL-10 results are not discussed even though is an important agent with also results regarding VA at 12 months.

325-327: these interesting results should be discussed. If the authors find CRT and CCT to fall beyond the manuscript agenda is should be better to ban this topic from the start of the presented investigation.

331: PCV/AMD categorization is a strong limitation of the study that tampers its conclusions; authors could consider, given such proportion, to also present all results with subgroup analysis between PCV+ and PCV- cases that could end up with interesting findings that could however still keep statistical significance.

332: authors should also refer to the high heterogeneity of the studied samples as per AMD traits as well as the lack of a standardized and blinded treatment protocol that limit the study conclusion, especially regarding VA and number of injections.

335: IL-10 and MMP-9 findings regarding number of injections are missing without reason.

338: "imaging parameters" have been lightly discussed in the manuscript, this could be skipped in the conclusion or otherwise enrich their previous discussion

Conclusion: I would recommend a complete rewriting focusing on how these results could influence present investigations and settle up future ones.

Reviewer #2: This study by Arai et al. reports intravitreal concentrations of CXCL-12, CXCL-13, IL-10, IP-10, IL-6, MCP-1, CCL-11, CXCL-1 and MMP-9 from 48 eyes of 48 patients. The study seems well conducted, although the paper could be more precise.

Overall comments:

- Some of the data is previously published—as I can tell, at least the entire table 1 has been previously published? I think the text should be very clear on what has already been published, preferably leave those results out of this study.

- The study describes inflammatory cytokines, although some are anti-inflammatory, and some are not cytokines at all?

- The list of measured markers are in different order in the abstract, in the text, and in the tables. It is a minor thing, but might help the reader to use the same order.

- Please state your reason for selecting these specific assays in the introduction.

Specific comments:

Page 3, line 62: Observations are made at baseline, after 2 months, and after 12 months. Later in the paper (e.g. page 10, line 204) you report results after 3 months? Later again (page 14, l. 254) reports measures at 2 months.

Page 3, line 66: define induction phase.

Page 4, line 77-80: Is already part of the method section.

Page 4, line 90: Patients received treatment – not eyes.

Page 4, line 91: refers to aforementioned centres, but these are not mentioned?

Regarding cytokine assays: Lacks information on how many measures were above/below detection range? And what where the CV values?

Page 5 line 103-106: Please rephrase/clarify sentence.

Page 5-6 line 116-118: “To identify PCV, A-Mode ultrasonography was used to measure axial length” This line is confusing, please rephrase.

Regarding clinical examinations: There is no clear definition of the diagnostic criterias.

Page 6 line 133-139: More suitable in the clinical examination-section.

Page 7, line 144-146: What models in what variable?

Page 7, line 158-160: refers to “typical AMD”, with no clear definition of this.

Table 2: Heading should probably read: Factors associated with baseline VA.

Table 2, and 3: It is unclear what selection of variables for the multivariate analysis is based on? In the subheading it states that P<0.05, but then why select sex (P=0.058) and CCT (P=0.42), but not IL-6 (P=0.018) in table 2?

Page 11, line 207-209: This line is confusing: BCVA change is associated with BCVA change?

Discussion: I am having much difficulties in distinguishing what results are from the previously published study, and what is new. Please only include the new findings to discussion and conclusion. A better use of space would be to discuss the potential impact of the relevant chemokines on CNV sustainability.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Marie Krogh Nielsen

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

1 Dec 2019

Dear editor and reviewers,

Thank you for reviewing our manuscript for its submission to PLOS ONE. We hereby submit a revised version of the manuscript that addresses the point-by-point response raised during the review process.

Both expert reviewers indicated that your submission was interesting but that the results do not fully support your conclusions. Reviewer 1 indicated flaws in your methodology, particularly statistical analysis and reviewer 2 indicated that your present paper overlaps with your earlier paper, Sakamoto S, Takahashi H, Tan X, et al.Br J Ophthalmol 2018;102:448–454. I share these reviewers concerns that you may be presenting the same results in both papers, for example Table 2 in the earlier paper seems to present the same results as Table 1 in the current paper. In rewriting your paper for PLoS One, please be clear what distinguishes the new paper from the older one.

Thank you for pointing out the duplication. We have deleted all duplicated sections.

1. We noticed you have some minor occurrence of overlapping text with the following previous publication(s), which needs to be addressed:

https://bjo.bmj.com/content/102/4/448

https://jamanetwork.com/journals/jamaophthalmology/fullarticle/426210

https://www.ncbi.nlm.nih.gov/pubmed/26674868

In your revision ensure you cite all your sources (including your own works), and quote or rephrase any duplicated text outside the methods section. Further consideration is dependent on these concerns being addressed.

We have checked this and updated in this revision.

2. Thank you for including your competing interests statement; "Dr Takahashi received lecturer’s fees from Kowa Pharmaceutical, Novartis Pharmaceuticals, Bayer Yakuhin, and Santen Pharmaceuticals, and grants from Bayer Yakuhin and Novartis Pharma, outside this work. Dr Inoue received lecturer’s fees from Kowa Pharmaceuticals, Novartis Pharmaceuticals, Bayer Yakuhin, and Santen Pharmaceuticals outside this work. Dr Kawashima received lecturer’s fees from Kowa Pharmaceutical, Novartis Pharmaceuticals, and Santen Pharmaceuticals outside this work. Dr Yanagi received lecturer’s fees and grants from Santen Pharmaceuticals outside this work. He is an advisory board member for Bayer Pharmaceuticals and a consultant for Santen Pharmaceuticals. Dr Arai, Dr Tan, Dr Inoda, Dr Sakamoto, and Dr Fujino declare no potential conflict of interest."

Please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials, by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests). If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include your updated Competing Interests statement in your cover letter; we will change the online submission form on your behalf.

We added the statement of sharing policies to the end of the COI statement and to the cover letter.

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

We have updated the Methods section in accordance with your comment.

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

We have uploaded all data to the following site: https://doi.org/10.6084/m9.figshare.7126922.v1.

Reviewer #1: This a very interesting study on aqueous humour cytokines as biomarkers of nAMD treatment response as per VA and number of injections at 12 months. Nonetheless, shows a strong rationale and well-intended methodology. However, there are significant flaws in the Methods and Results section that, by instance, should not represent a major problem to solve. In addition, some information is lacking and some, by contrast, does not even need to be presented as it is later forgotten in the discussion. If most points are indeed faced and cleared it could then be ready to undergo a more detailed review.

Thank you for your comments. Overall, we toned down the strong statements in the Introduction. Additionally, the Methods and Results sections have been re-written to include additional details, as described below. We believe that we have carefully responded to all of your comments.

Abstract: consider a section-based writing instead such as Purpose-Methods-Results-Conclusion.

We simply followed the authors’ guidelines of PLOS ONE; the sample Abstract in the authors’ guidelines does not include subheadings. However, we did structure the Abstract so that Purpose was first, followed by Methods and Results, and put the Conclusions last. However, we are not sure if we are allowed to write subheadings as you suggested and left it as it is. If EBM and AE allow, we are happy to accept your proposal to use subheadings in the Abstract.

Also, correlation coefficients mean association whereas beta coefficients mean prediction, which is not the same and therefore the conclusion should be rewritten.

We have corrected the points as you suggested. Overall, we used the term ‘predictive variables’ to describe what we have done correctly (page 18, lines 356–361)

Introduction: as a whole, we miss some more referenciated information on the local and systemic importance of inflammation for nAMD which, after all, is the rationale of the present study.

We have added 9 references related to the importance of local and systemic inflammation in nAMD to strengthen our hypothesis (refs 4 to 12).

49: "AMD" use should be stated in the very first part of the text, not afterwards

We have carefully reviewed the manuscript and corrected all abbreviations as appropriate.

54-55 and 57-58: consider softening such strong statements that don't even show a reference

We have toned down these statements and added references as appropriate (page 3, lines 55–58).

60-64: explanation of the present study should not be present in the Introduction section

We have moved this explanation to the last part of the Introduction -(page 4, line 80-86).

73-74: a reference is lacking in this statement

We have added references (page 4, line 76).

85: was this informed consent written or oral? Please specify

We have indicated that we received written informed consent from the patients (page 4, line 91).

92: were non-retinal diseases such as glaucoma or anterior uveitis excluded?

We excluded other ocular diseases such as glaucoma or anterior uveitis (page 5, lines 100).

95: how was this obtained? Through a clear corneal paracenthesis? On the slit lamp? Please specify

We obtained aqueous humour through a clear corneal paracentesis under an operating microscope. This has been clarified in the revised manuscript (page 5, lines 103–105).

101-102: how where 0 values handled?

We imputed half the lowest detection value for values lower than the detection limit (page 5, lines 110–111).

103-104: reference that statement

We have added a reference. (page 5, line 113).

105: how were these vitreous samples obtained? on which basis and time? same subjects? on what inclusion criteria and randomisation?

After the study protocol was approved by the Institutional Review Board, we initially collected vitreous humour samples from 5 cases. However, we decided not to proceed with further vitreous sampling thereafter because of safety concerns; we noted that the vitreous was stuck in the needle, and the procedure could potentially increase the risk of developing proliferation of vitreous. Moreover, the volume of vitreous humour that was collected by vitreous tap was very low. The 5 vitreous humour samples were all from nAMD eyes.

106-107: need dispersion statistics such as standard deviation or IQR when a mean/median is showed

We have added standard deviation as you suggested. (page 5, lines 115–117).

109: how were these 29 subjects chosen? on which basis?

We were not able to determine VEGF concentrations in all patients because the sample volume was insufficient for measuring the concentrations of cytokines (multiplex analysis) and VEGF (enzyme-linked immunosorbent assay [ELISA]). We first subjected the samples to multiplex analysis and then the remaining samples were subjected to ELISA. Therefore, we were able to measure VEGF concentrations in only 29 (60%) cases prior to anti-VEGF therapy. Additionally, this was stated in our previous publication, and therefore, we cited our previous publication here.

114-115: does this mean ICG and FA was undertaken in all cases?

You are correct; all cases underwent fluorescein and indocyanine green angiography. This is our routine work-up for nAMD in our institution, because of the relatively high prevalence of PCV among nAMD cases in the region where the study was conducted (Japan) (page 5, lines 122–125).

125: specify which ranibizumab concentration is used

The ranibizumab concentration was 0.5 mg in 0.05 mL of solution (page 6, line 130). We believe this is the only available dosage outside the US.

126: "essentially as described in the CATT", please specify which deviation from CATT protocol this means

We have changed this line and detailed the PRN regimen we employed to describe the protocol clearly.

127: treatment decisions are therefore relied on the physician, with no protocol adjusted or blind assessment. Since this, conclusions as per visual acuity prognosis at 12 months are unfair as treatment is not defined nor controlled in the group.

This is a limitation of the present study.

129: specify which chart of vision is used

We measured best corrected visual acuity (BCVA) using a 5 M Landolt C VA chart.

142: Pearson correlation demands a normality assumption, which is not clearly stated in the text, please update accordingly, also referring to used normality tests.

A normal distribution was confirmed with Shapiro-Wilk’s W-test (page 7, lines 151–153).

145: log transformation of which variables? cytokines? please specify

We added “of aqueous humour proteins” (page 7, line 159).

157: add dispersion statistics to AL mean

We have rewritten and updated the Results accordingly.

158-159: definition criteria for PCV and type of nAMD is not stated in the manuscript, please update

We have rewritten and updated the Results accordingly.

167-168: paired tests have not been claimed in the methods section, please update

We have rewritten and updated the Methods for statistical analysis and have described all tests performed in this paper.

170: are these means? medians? please state so; which statistical test is used?

Since the concentrations had a lognormal distribution, they are geometric means. We originally included Table 1 to clarify the demographics of the patients, but this was the same as in Table 2 of our previous study (Sakamoto et al., BJO 2018) and we have deleted the entire table and put the necessary information in the Methods.

182: where is this 2.7 coming from?

By 2.7 (=e), we meant Napier’s constant, because we used natural logarithms. However, we have changed this to “per logMCP-1” to avoid any confusion.

Table 2, 3, S1 and S2: it is not clear to me whether these are correlation coefficients, beta or standardized beta. Please update accordingly. Also a 95% CI should be informed for all data and a total R square in all tables.

We have updated the tables according to your comments. However, adding 95% confidence intervals (CIs) made the tables too complex. Most guidelines recommend including at least a P value, and not a 95% CI, because it provides sufficient information to assess the significance of the value. R2 values were stated for all multivariate analysis.

201 and on: "better" is an unclear term, use "improvement" instead.

We have corrected this.

231: a Table 4 on this topic is also mandatory as previously done

Sorry, but we do not have a Table 4; we carefully reviewed and corrected all tables.

252-255: even though S1/S2 tables are supplied, consider stating the main found coefficients in the manuscript in order to give a clear vision of their magnitudes

We chose not to include this topic in the revised manuscript and all supplementary tables have been deleted.

301-302: such bivariant statement is not further explained in the manuscript and could be forsaken.

We have deleted this sentence.

303: a clear discussion on the reason why lower MCP-1 is associated to better BCVA at baseline and, at the same time, higher levels are linked to better BCVA at 12 months is mandatory. This is point is forgotten in the discussion.

Lower MCP-1 concentrations were associated with better BCVA at baseline and, at the same time, higher levels were linked to a better BCVA ‘gain’ at 12 months. We have added ‘The current results demonstrated that eyes with higher MCP-1 concentrations had worse baseline BCVA compared to those with lower MCP-1 concentrations. Those with higher MCP-1 concentrations at baseline had a better chance of BCVA gain compared to those with lower MCP-1 concentrations, which is probably because of the low baseline BCVA of this group, at least in part’.

314: baseline MMP-9 significant association is not discussed and even goes against this statement.

We have added ‘Moreover, the loss of structural recuperative power will necessitate repeated injections’.

324: IL-10 results are not discussed even though is an important agent with also results regarding VA at 12 months.

The Discussion section has been re-written and we have added a discussion about the results of IL-10. Inflammation occurring in eyes with nAMD is mediated by several types of immune cell, among which myeloid cells (macrophages/monocytes) are considered to be of the utmost important. Laboratory studies suggest that myeloid cells can be classified into at least two subpopulations, namely, classically activated and alternatively activated monocytes/macrophages, which control two different aspects of inflammation; in general, the former exacerbates inflammation in the acute phase of inflammation, while the latter exacerbates angiogenesis in the chronic phase of inflammation, although there is some debate on this issue. IL-10 is produced by alternatively activated macrophages, which contribute to angiogenesis. It is tempting to consider that in addition to exudative changes, low levels of IL-10 are associated with the aberrant control of angiogenesis, which may affect the number of injections.

325-327: these interesting results should be discussed. If the authors find CRT and CCT to fall beyond the manuscript agenda is should be better to ban this topic from the start of the presented investigation.

We have deleted these results and focused instead on the factors associated with treatment outcome in the current manuscript. We decided that CRT and CCT are two important topics that will be analysed and discussed in a separate paper in the future.

331: PCV/AMD categorization is a strong limitation of the study that tampers its conclusions; authors could consider, given such proportion, to also present all results with subgroup analysis between PCV+ and PCV- cases that could end up with interesting findings that could however still keep statistical significance.

We agree with your opinion; therefore, in our current analysis, we tried to elucidate the effects of disease type (CNV with or without PCV), and included this as a covariant in stepwise variant selection. However, disease type was selected only as a predictor variable of BCVA change and in no other situations.

332: authors should also refer to the high heterogeneity of the studied samples as per AMD traits as well as the lack of a standardized and blinded treatment protocol that limit the study conclusion, especially regarding VA and number of injections.

We have revised the Limitations section accordingly. However, we believe our results are still relevant considering that most of the patients adhered to the protocol presented and the treatment regimen was based on one of the most commonly employed therapeutic approaches.

335: IL-10 and MMP-9 findings regarding number of injections are missing without reason.

We have now discussed this in the revised manuscript.

338: "imaging parameters" have been lightly discussed in the manuscript, this could be skipped in the conclusion or otherwise enrich their previous discussion

We have deleted this from the manuscript.

Conclusion: I would recommend a complete rewriting focusing on how these results could influence present investigations and settle up future ones.

Thank you for your constructive suggestions. We have added a section summarizing how the current investigation builds on previous studies in this field and how it impacts on future directions. ‘Our results clearly demonstrated that different aqueous humour proteins, including cytokines associated with classically and alternatively activated macrophages, may be involved in AMD in a different way. Part of this could be ascribed to the fact that each factor is involved in different aspects of inflammation, i.e. exudation and angiogenesis. Further studies focusing on different subclasses of immune cells and their markers are warranted for better clinical care.

Reviewer #2: This study by Arai et al. reports intravitreal concentrations of CXCL-12, CXCL-13, IL-10, IP-10, IL-6, MCP-1, CCL-11, CXCL-1 and MMP-9 from 48 eyes of 48 patients. The study seems well conducted, although the paper could be more precise.

Overall comments:

- Some of the data is previously published—as I can tell, at least the entire table 1 has been previously published? I think the text should be very clear on what has already been published, preferably leave those results out of this study.

We have deleted the replicated data from the revised manuscript.

- The study describes inflammatory cytokines, although some are anti-inflammatory, and some are not cytokines at all?

You are correct. We measured the aqueous humour concentrations of several proteins, not only cytokines; therefore, we changed the terminology to more accurately describe what was measured.

- The list of measured markers are in different order in the abstract, in the text, and in the tables. It is a minor thing, but might help the reader to use the same order.

We have re-ordered this and rewritten the Abstract, text, and tables.

- Please state your reason for selecting these specific assays in the introduction.

We have added the reason underlying the selection of these assays.

Specific comments:

Page 3, line 62: Observations are made at baseline, after 2 months, and after 12 months. Later in the paper (e.g. page 10, line 204) you report results after 3 months? Later again (page 14, l. 254) reports measures at 2 months.

We measured at 2 months (at the 3rd injection). We apologise for the confusion and have corrected this mistake.

Page 3, line 66: define induction phase.

We have added ‘initial 3 monthly consecutive injection phase’. (Page 3, lines 69–70).

Page 4, line 77-80: Is already part of the method section.

We have deleted the duplicated section.

Page 4, line 90: Patients received treatment – not eyes.

We have corrected this.

Page 4, line 91: refers to aforementioned centres, but these are not mentioned?

We recruited patients from two institutions. This sentence did not make any sense at all and we have corrected the mistake.

Regarding cytokine assays: Lacks information on how many measures were above/below detection range? And what where the CV values?

No proteins were above the detection range. The number of samples below the detection range, number of samples below the detection limit, and coefficient of variation (CV) of CXCL1, IP-10, CXCL12, CXCL13, MCP-1, CCL11, IL-6, IL-10, and MMP-9 were 1, 0, 3.6%, 0, 0, 3.8%, 0, 0, 6.7%, 0, 0, 4.6%, 0, 0, 2.1%, 0, 0, 4.7%, 0, 0, 2.0%, 68, 37, 4.5%, and 17, 1, 3.4%, respectively. Most of the samples were within the detection range, except for MMP-9.

Page 5 line 103-106: Please rephrase/clarify sentence.

We rephrased and clarified the sentence on page 5, lines 111–113.

Page 5-6 line 116-118: “To identify PCV, A-Mode ultrasonography was used to measure axial length” This line is confusing, please rephrase.

We apologise for our mistake. We have corrected the sentences. ‘To identify PCV…’ was an infinitive clause for the previous sentence.

Regarding clinical examinations: There is no clear definition of the diagnostic criterias.

We apologise for the lack of diagnostic criteria. We added references for the diagnosis of PCV and RAP to page 6, line 122.

Page 6 line 133-139: More suitable in the clinical examination-section.

We moved this paragraph to the clinical examination section on page 6, lines 139–145.

Page 7, line 144-146: What models in what variable?

We have revised the statistical analysis section.

Page 7, line 158-160: refers to “typical AMD”, with no clear definition of this.

We have revised the Results section.

Table 2: Heading should probably read: Factors associated with baseline VA.

We have changed the heading of Table 1 as you suggested (page 8 line 189).

Table 2, and 3: It is unclear what selection of variables for the multivariate analysis is based on? In the subheading it states that P<0.05, but then why select sex (P=0.058) and CCT (P=0.42), but not IL-6 (P=0.018) in table 2?

We apologise for the confusion. To clarify this, we have rewritten the Abstract and Methods to describe the methods used. We used stepwise variable selection, described on page 8, lines 158–159, to build the best prediction model that minimizes the Bayesian Information Criterion. This is an objective method that automatically determines which variables should be included in the model. We used the forward (increasing) method.

Page 11, line 207-209: This line is confusing: BCVA change is associated with BCVA change?

We apologize for the confusion. The larger change in BCVA at 2 months was associated with an increased change of BCVA of 12 months. We deleted the sentence previously on page 12, lines 222–224.

Discussion: I am having much difficulties in distinguishing what results are from the previously published study, and what is new. Please only include the new findings to discussion and conclusion. A better use of space would be to discuss the potential impact of the relevant chemokines on CNV sustainability.

Although we initially thought that including such information was useful for the readers of the current manuscript, we have revised the text to delete the duplicated results (Table 1) and other previous findings to avoid concerns of possible duplicate publication. The Discussion was also re-written to focus more on the new findings of the current study (deleted lines from the Discussion, previously on page 17, lines 286–290).

Submitted filename: Response to Reviewers .docx

Click here for additional data file.

19 Dec 2019

PONE-D-19-23157R1

Aqueous humour proteins and treatment outcomes of anti-VEGF therapy in neovascular age-related macular degeneration

PLOS ONE

Dear Dr. Takahashi,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Two reviewers evaluated this revised manuscript and both agreed that you have addressed the issues raised previously. Thank you for your efforts.  Both have made some suggestions for additional clarification of your findings. Please address these additional comments.

We would appreciate receiving your revised manuscript by Feb 02 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Alfred S Lewin, Ph.D.

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #3: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #3: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #3: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #3: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I do congratulate the authors for their job improving the manuscript, which is now technically sound and expressively strong. I can only point out two minor questions to undertake.

Lines 118-120: clarify whether VEGF in these 29 samples was vitreous or aqueous humour. Moreover, since not all samples were analyzed for VEGF search (as stated in your response), it ends up behaving as a huge bias for the study. Some could even think that the authors choose which samples they finally analyzed. I feel it is not enough to state a reference to back up such reasoning and I would rather include a detailed explanation, like the one that has been provided in the response, in the manuscript.

Conclusion: lacks a statement on IL-10 at 2 months associated with the number of injections, which has been extensively tackled in the Discussion, and should be at least briefly commented here.

Reviewer #3: In this paper Arai et al. investigated the relationship between Aqueous humour proteins and anti-VEGF therapy outcomes in nAMD. The authors have addressed previous comments. I have two additional minor comments:

1. Abstract: The authors state in the final sentence that "In conclusion, aqueous humour protein concentrations may be good predictors of BCVA change over 12 months and the number of injections in pro re nata treatment of exudative nAMD.". This is not my impression from this study - this study can only conclude that aqueous humor proteins may have predictive abilities - to determine whether they are good predictors, further biomarker analyses need to be performed.

2. Regarding the limitations of this study. This is perhaps in the lines of my comment #1. A limitation of this study is the lack of another independent sample, wherein the potentially predictive abilities of aqueous humor proteins can be tested. This approach allows concluding whether these proteins truly possess predictive abilities and how good they are in doing so.

3. Regarding methods: To my knowledge, the Procarta® Cytokine Assay Kit is a multiplex cytokine assay that is read using a fluorescence-based method. Do you have any insight from a pilot study or any investigation into whether fluorescein or indocyanine green (which is used in the retinal angiography in relation to diagnosis) influences measurements obtained from the Procarta® Cytokine Assay Kit? If not, then I think it would be worth mentioning that this is a potential source of bias and thus a limitation that must be kept in mind when interpreting the results of this study. Also, it is unclear in the manuscript when you did the sampling. Immediately after the diagnosis? Immediately prior to the injection? Immediately after the injection? Was it prior to the injection in baseline, but after at the first follow-up? These factors are important details with potential influence to the results. These factors should be outlined in a clear fashion.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

2 Feb 2020

Reviewer #1: I do congratulate the authors for their job improving the manuscript, which is now technically sound and expressively strong. I can only point out two minor questions to undertake.

Lines 118-120: clarify whether VEGF in these 29 samples was vitreous or aqueous humour. Moreover, since not all samples were analyzed for VEGF search (as stated in your response), it ends up behaving as a huge bias for the study. Some could even think that the authors choose which samples they finally analyzed. I feel it is not enough to state a reference to back up such reasoning and I would rather include a detailed explanation, like the one that has been provided in the response, in the manuscript.

First, these 29 samples were aqueous humour. Second, we have added explanations as you suggested. (lines 127-129)

Conclusion: lacks a statement on IL-10 at 2 months associated with the number of injections, which has been extensively tackled in the Discussion, and should be at least briefly commented here.

We have updated the conclusion section in accordance with your comment. (lines 380-381)

Reviewer #3: In this paper Arai et al. investigated the relationship between Aqueous humour proteins and anti-VEGF therapy outcomes in nAMD. The authors have addressed previous comments. I have two additional minor comments:

1. Abstract: The authors state in the final sentence that "In conclusion, aqueous humour protein concentrations may be good predictors of BCVA change over 12 months and the number of injections in pro re nata treatment of exudative nAMD.". This is not my impression from this study - this study can only conclude that aqueous humor proteins may have predictive abilities - to determine whether they are good predictors, further biomarker analyses need to be performed.

We have updated the abstract section in accordance with your comment. (line 45)

2. Regarding the limitations of this study. This is perhaps in the lines of my comment #1. A limitation of this study is the lack of another independent sample, wherein the potentially predictive abilities of aqueous humor proteins can be tested. This approach allows concluding whether these proteins truly possess predictive abilities and how good they are in doing so.

We agree with this. We have revised the limitation section. (lines 369-370)

3. Regarding methods: To my knowledge, the Procarta® Cytokine Assay Kit is a multiplex cytokine assay that is read using a fluorescence-based method. Do you have any insight from a pilot study or any investigation into whether fluorescein or indocyanine green (which is used in the retinal angiography in relation to diagnosis) influences measurements obtained from the Procarta® Cytokine Assay Kit? If not, then I think it would be worth mentioning that this is a potential source of bias and thus a limitation that must be kept in mind when interpreting the results of this study. Also, it is unclear in the manuscript when you did the sampling. Immediately after the diagnosis? Immediately prior to the injection? Immediately after the injection? Was it prior to the injection in baseline, but after at the first follow-up? These factors are important details with potential influence to the results. These factors should be outlined in a clear fashion.

We performed a preliminary test, and confirmed that the fluorescein and indocyanine green did not affect the measurement of cytokine levels in this Cytokine Assay Kit. We have added the results as S2-text for your review. In the text we mentioned “dye angiography did not affect the measurements (data not shown).” (lines 123 to 124)

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

5 Feb 2020

Aqueous humour proteins and treatment outcomes of anti-VEGF therapy in neovascular age-related macular degeneration

PONE-D-19-23157R2

Dear Dr. Takahashi,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Alfred S Lewin, Ph.D.

Section Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

14 Feb 2020

PONE-D-19-23157R2

Aqueous humour proteins and treatment outcomes of anti-VEGF therapy in neovascular age-related macular degeneration

Dear Dr. Takahashi:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Alfred S Lewin

Section Editor

PLOS ONE