Intraocular Concentration of Stem Cell Factor/c-KIT and Galectin-1 in Retinal Diseases

Abstract

Purpose: To investigate the intraocular concentration profiles of stem cell factor (SCF)/c-KIT, galectin-1 (GAL-1), and vascular endothelial growth factor (VEGF)-A with regard to retinal disease and treatment response.

Methods: The study group included 13 patients with dry age-related macular degeneration (AMD), 196 with neovascular AMD (nAMD), 21 with diabetic macular edema (DME), 10 with retinal vein occlusion (RVO), and 34 normal subjects with cataracts. Aqueous humor levels of SCF, c-KIT, GAL-1, and VEGF-A were analyzed by immunoassay according to disease group and treatment response.

Results: Increased aqueous levels of SCF, c-KIT, and GAL-1 were observed in eyes with nAMD (2.67 ± 3.66, 296.84 ± 359.56, and 3945.61 ± 5976.2 pg/mL, respectively), DME (1.64 ± 0.89, 238.80 ± 265.54, and 3701.23 ± 4340.54 pg/mL, respectively), and RVO (4.62 ± 8.76, 509.63 ± 647.58, and 9079.60 ± 11909.20 pg/mL, respectively) compared with controls (1.13 ± 0.24, 60.00 ± 0.00, and 613.27 ± 1595.12 pg/mL, respectively). In the eyes of nAMD, the levels of all three cytokines correlated positively with VEGF-A levels. After intravitreal injections of anti-VEGF agents, the levels of GAL-1 and VEGF-A decreased significantly, whereas those of SCF and c-Kit showed no significant change. Eyes of nAMD patients with improved vision after treatment had significantly lower levels of c-KIT, GAL-1, and VEGF-A at baseline.

Conclusions: The intraocular levels of cytokines were significantly elevated in eyes with nAMD, DME, and RVO compared to the controls and they showed different response to anti-VEGF treatment. With this result and their known association with angiogenesis, these cytokines may be potential therapeutic targets for future research.

Various diseases that are associated with ischemic changes in the retina pose a threat to vision. Age-related macular degeneration (AMD) is a leading cause of vision loss in the developed world.¹ Late AMD denotes geographic atrophy and neovascular AMD (nAMD). In nAMD, vascular endothelial growth factor (VEGF) is a key mediator of exudative disease activity, the overexpression of which by hypoxic retinal pigment epithelium causes loss of the blood–retinal barrier of the choriocapillaris.² Intravitreal injection of anti-VEGF is currently the mainstream of treatment. It necessitates repeated treatment at intervals determined by the half-lives of the therapeutics to maintain therapeutic concentrations,³ which creates a significant socioeconomic burden. Similar to nAMD, diabetic retinopathy (DR) and retinal vein occlusion (RVO) are ischemic retinal diseases that can greatly reduce vision and quality of life. Accompanying macular edema (ME) is the leading cause of vision loss in these diseases. Hypoxia causes inflammation of the retina mediated by various cytokines, leading to ME.^4–7 VEGF is also an important therapeutic target for these conditions, as for nAMD. 

However, a significant proportion of patients respond poorly to anti-VEGF agents,⁸ resulting in a need for the development of new treatments. Many ongoing studies are looking at cytokines, not just VEGF, as potential therapeutic targets. The expression profiles of several cytokines in serum, plasma, and aqueous humor of patients with nAMD include interleukin (IL)-1 and IL-2,^9–11 IL-4 and IL-10,^10,11 IL-6,^10,12–15 IL-12,^13,14 IL-8,^10,13–16 interferons (IFNs),^13,14,17 transforming growth factor (TGF),^18–20 and tumor necrosis factor alpha (TNF-α).^{13,14,16,19–22} Recently, other lesser known cytokines, such as stem cell factor (SCF) and galectins, have been implicated in the pathogenesis of retinal diseases. 

SCF is a multifunctional cytokine that binds to c-KIT as a ligand.²³ It plays key roles in regulating cellular functions, such as survival, proliferation, differentiation, and migration.²⁴ Some studies have shown that the hypoxia-induced increases in SCF/c-KIT signaling are associated with retinal diseases and pathological ocular neovascularization.^6,25 These reports have demonstrated upregulation of SCF and c-KIT in mural models of ischemic retinal diseases, including DR. Galectins are powerful stimulators of angiogenesis by engaging a different set of endothelial cell surface receptors, activating unique signaling pathways regulating angiogenic cascade.²⁶ In particular, galectin-1 (GAL-1) is an important cytokine involved in retinal vascular disease.²⁷ 

However, little clinical evidence is available regarding the involvement of SCF/c-KIT and GAL-1 in different retinal diseases and their responses to anti-VEGF treatment. Thus, this study aimed to identify the mechanisms associated with these cytokines in multiple retinal diseases by measuring aqueous humor concentrations and assessing their potential as therapeutic targets. To this end, we analyzed the concentrations of SCF/c-KIT, GAL-1, and VEGF-A in the aqueous humor of patients with RVO with non-exudative AMD, nAMD, diabetic ME (DME), or ME. 

Materials and Methods

This study was approved by the Institutional Review Board of Seoul National University Bundang Hospital (IRB no. B-1907/550-306) and adhered to the tenets of the Declaration of Helsinki. Informed consent was obtained from all participants before their participation in the study. 

Our study enrolled 274 subjects, including 240 patients who were diagnosed with various retinal diseases and had visited Seoul National University Bundang Hospital for decreased visual acuity. Patients were recruited between October 30, 2009, and March 4, 2022, referring to the date of specimen collection. Retinal diseases included dry AMD (n = 13), neovascular AMD (n = 196), DME (n = 21), and RVO (n = 10). The patients included in the study were treatment naïve. In addition, we established a control group of patients who underwent cataract surgery and had no known retinal disease (n = 34). The exclusion criterion was patients not consenting to the sampling of aqueous humor at baseline. 

Every patient underwent thorough ophthalmic examinations, including best-corrected visual acuity (BCVA) checks with autorefraction and spectral-domain optical coherence tomography (SD-OCT; SPECTRALIS OCT; Heidelberg Engineering, Heidelberg, Germany). BCVA and central macular thickness (CMT) were measured using SD-OCT at each visit. After establishing the initial diagnosis, aqueous fluid was obtained via anterior chamber paracentesis during intravitreal anti-VEGF injections. 

The treatment protocol was applied to all patients in a controlled manner as follows. Treatment was based on the T&E (treat and extend) method for nAMD patients and the PRN (pro re nata) method for patients with DME or RVO ME. For patients with nAMD, three loading injections were given monthly, followed by injections with a repeat interval based on the recurrence of disease activity as determined by the presence of fluid on OCT. For DME and RVO patients, the clinician determined the need for PRN dosing based on the presence of ME on OCT. In patients with nAMD, aqueous humor was collected at the first and third visits of the initial monthly loading injections. This allowed us to compare the concentrations of cytokines and VEGF-A before and after the treatment with anti-VEGF agents. 

A Luminex multiplex immunoassay system (Luminex Bio-Plex 200; Bio-Rad, Hercules, CA, USA) was used for simultaneous quantification of multiple cytokines. Concentrations of SCF, CD117 (soluble c-KIT), GAL-1, and VEGF-A were analyzed using the ProcartaPlex Human custom-developed 4-plex kit (cat. no PPX-04-MX04KOR; Invitrogen, Waltham, MA, USA). Standard concentrations for the four targets included in the kit were diluted to 1/2.44 based on the concentrations provided by the manufacturer. The multiplex assay kit can quantitatively measure multiple cytokines from as little as 50 µL of aqueous humor. Each sample was run as a single measurement for a limited amount of collected aqueous humor. 

Statistical Analyses

We compared the demographic and clinical data of patients according to disease entity. BCVAs were analyzed using converted logarithmic minimum angle of resolution (logMAR) values. Very low visual acuity was replaced with the appropriate logMAR value, as suggested by Lange et al.²⁸: finger count, 2.0; hand motion, 2.3; light perception, 2.6; and no light perception, 2.9. BCVA and CMT were expressed as mean ± SD. 

Wilcoxon's signed-rank test was used to compare the concentrations of cytokines before and after the use of anti-VEGF agents in patients with nAMD. For further analysis, we categorized the subgroups based on visual acuity improvement and CMT reduction of 10% or more. Baseline cytokine concentrations were compared between groups using the Mann–Whitney U test. The relationships between each disease and the cytokine levels were analyzed using Spearman's rank correlation analysis. In addition, the correlation between the baseline concentrations of VEGF and other cytokines was determined. When interpreting the results of statistical tests, P < 0.05 was considered statistically significant. When performing multiple comparisons, we used the Bonferroni correction. All statistical analyses were performed using SPSS Statistics 26.0 for Windows (IBM, Chicago, IL, USA). 

Results

Baseline Characteristics of All Subjects

Demographics and Clinical Characteristics of Subjects

The specific demographics, baseline BCVA, laterality, and mean concentrations of the three cytokines and VEGF-A are summarized in Table 1. The mean age of all patients was 73.2 ± 8.7 years, and men accounted for 49.2% of the total. The prevalence rates of diabetes and hypertension were 40.4% and 25.0%, respectively. Of 34 patients of the control group, 17 patients (50%) had diabetes and seven patients (21%) had hypertension. Two patients (6%) had both. 

Table 1.

 

Demographics, Clinical Characteristics, and Mean Cytokine and VEGF-A Concentrations of Subjects According to Disease Entities

 

Cytokine and VEGF-A Concentration Profile in Various Retinal Diseases

The concentrations of cytokines and VEGF-A at baseline were summarized by disease entity. The exact values and relative ratios of concentration levels are shown in Table 1 and Figure 1, respectively. The ratios in Figure 1 were derived by normalizing the concentration levels of each cytokine in each disease entity to those of the control. Concentrations of the cytokines and VEGF-A were higher in the nAMD, RVO, and DME groups than in the control group. In dry AMD, only c-KIT showed elevated levels. Overall, the cytokine concentration ratios were highest in RVO, followed by nAMD/DME, and they were lowest in dry AMD subjects (P < 0.05). 

Figure 1.

 

Levels of cytokines and VEGF-A in the aqueous humor according to disease entities. This study enrolled patients with one of four categories of retinal diseases: dry AMD (n = 13), neovascular AMD (n = 196), DME (n = 21), or RVO (n = 10). Relative ratios as compared to control were calculated for adjustment. The error bars represent standard errors of the mean rather than standard deviations to allow better visualization of the data, because most concentration levels were skewed toward out-of-range values. Relative ratio values are shown beneath the error bars. Asterisks indicate statistically significant differences compared to controls. In the order of dry AMD, neovascular AMD, DME, and RVO, all of the cytokines and VEGF-A showed significant differences among the retinal disease entities: C-KIT (1.48, 4.94, 3.98, and 8.49, respectively; P < 0.001), SCF (0.94, 2.36, 1.45, and 4.08, respectively; P = 0.016), GAL-1 (0.57, 6.43, 6.03, and 14.81, respectively; P < 0.001), and VEGF-A (1.18, 4.83, 5.76, and 25.97, respectively; P < 0.001).

A statistically significant positive correlation was found between the baseline concentrations of VEGF-A and the three cytokines in patients with nAMD (Table 2, Supplementary Fig. S1). In the other disease groups, the sample size was smaller than required; hence, the correlations were not statistically significant, but the three cytokines correlated positively with VEGF-A in all subjects, including the controls. The correlation coefficients were highest for RVO, followed by DME, dry AMD, and controls, in that order. 

Table 2.

 

Spearman Correlation Analysis Between Cytokines and VEGF-A According to Disease Entities

 

Cytokine Concentrations Before and After anti-VEGF Treatment in nAMD Patients

Pre- and post-injection aqueous cytokines were measured in 69 patients with nAMD who received intravitreal injections of anti-VEGF agents. Of these, 45 received aflibercept injections and 24 received ranibizumab injections. The changes in the concentration of each cytokine before and 35.6 ± 7.2 days after injection are shown in Table 3 and Figure 2. Statistically significant differences in the concentrations of GAL-1 and VEGF-A were found between pre- and post-injection measurements (P = 0.046 and P < 0.001, respectively), whereas no statistically significant differences in c-KIT and SCF were noted (P = 0.948 and P = 0.857, respectively). 

Table 3.

 

Cytokine and VEGF-A Concentrations for Patients with nAMD Who Received Anti-VEGF Treatment

 

Figure 2.

 

Comparison of cytokine and VEGF-A concentration levels in patients with nAMD before and after anti-VEGF injections. Concentration values are presented on a logarithmic scale. Error bars indicate standard errors of the mean. Wilcoxon's signed-rank test was performed to compare the changes before and after injections of each cytokine and VEGF-A. GAL-1 (6183.24 pg/mL vs. 5227.06 pg/mL; P = 0.046) and VEGF-A (1420.17 pg/mL vs. 913.97 pg/mL; P < 0.001) showed significant differences before and after injection, respectively.

Subgroup Analysis of nAMD Patients According to the Change in VA and CMT

Table 4 and Figure 3 show the details related to visual improvements in terms of BCVA and CMT in 196 patients with nAMD at a 1-year follow-up. Of these, 66 patients showed an improvement in vision, whereas the remaining 130 did not. In contrast, 87 patients showed CMT improvement, but 109 showed no CMT improvement. Statistically significant differences in baseline c-KIT (P = 0.001), GAL-1 (P = 0.001), and VEGF-A (P = 0.006) levels were seen between the groups with or without VA improvement. The baseline concentrations of these cytokines were lower in the VA-improvement group than those in the control group; however, no significant differences in baseline levels of SCF were noted (P = 0.199). When comparing groups with or without CMT improvement, no differences in any of the cytokines were observed. Figure 3 shows the relative ratios of cytokine concentrations as compared to the controls, with the exact values documented below the bar graph. 

Table 4.

 

Therapeutic Responses and Their Association With Visual Prognosis and Cytokine Concentrations at Baseline

 

Figure 3.

 

Comparison of baseline levels of aqueous cytokines according to anti-VEGF treatment response in nAMD (n = 196). The relative concentration ratio of each cytokine to the control are indicated as bars. Error bars represent the standard errors of the mean. (A) Sixty-six patients showed a good response as indicated by BCVA improvement, but 130 patients showed a poor response. Relative ratios compared to controls are documented beneath the bar graphs with error bars. The results showed statistically significant differences in c-KIT (3.38 vs. 5.73; P = 0.001), GAL-1 (4.69 vs. 7.32; P = 0.001), and VEGF-A (3.32 vs. 5.6; P = 0.006). Concentrations of these cytokines were lower in the VA improvement group as compared to the VA non-improvement group, whereas SCF (1.72 vs. 2.69; P = 0.199) showed no significant difference. (B) Eighty-seven patients showed improvement in CMT, but 109 patients did not. No significant differences in the concentrations of each cytokine were seen between these two groups.

Table 5 shows cytokine and VEGF concentrations measured after 1 month of treatment in 69 patients with nAMD. Statistically significant differences were observed in SCF (P = 0.033), GAL-1 (P = 0.043), and VEGF-A (P = 0.023) between the groups with or without VA improvement. Concentrations of these cytokines were lower in the VA improvement group. When comparing groups with or without CMT improvement, there were no differences in all cytokines except VEGF-A. 

Table 5.

 

Therapeutic Responses and Their Association With Visual Prognosis and Cytokine Concentrations After 1 Month of Treatment

 

Discussion

To our knowledge, this is the first study to describe and compare the concentration profiles of c-Kit, SCF, and Gal-1 in various retinal diseases, including DME, RVO, dry AMD, and nAMD. Previously, several studies have reported that SCF/c-KIT and GAL-1 are involved in RVO, DME, and AMD.^4–6 However, these studies were limited to a single disease entity or to a specific cytokine. For SCF and soluble c-KIT, the mean concentration levels in proliferative DR patients were significantly upregulated as compared to those in patients with quiescent DR or without diabetes.²⁴ Additionally, a fully monoclonal antibody targeting c-KIT showed potential efficacy in the treatment of pathological choroidal neovascularization.²⁴ GAL-1 is commonly upregulated in patients with DME^4,5 and nAMD.²⁹ Wu et al.²⁹ reported that GAL-1 promoted choroidal neovascularization in a murine model of nAMD. Moreover, intravitreal injection of OTX008, a selective small-molecule inhibitor of GAL-1, inhibited neovascularization in a murine model of oxygen-induced retinopathy.³⁰ 

Concentration Profiles of Inflammatory Cytokines in Various Retinal Disease

Figure 1 shows the relative concentration profiles of SCF, c-Kit, GAL-1, and VEGF-A in RVO with ME, nAMD, or DME. They have in common that they were all elevated compared to the control; however, their concentration profiles were significantly different from each other. RVO had the highest levels of all cytokines. 

We postulate that the different pathophysiology of the retinal insult is responsible for this difference. DR and RVO share a common aspect of the pathophysiology in that the insult to the retinal vasculature causes hypoxia of the retinal layers supplied by these vessels. In nAMD, however, the insult to the choriocapillaris causes hypoxia of the retinal pigment epithelium rather than the retina. VEGF levels in the aqueous humor of RVO patients are known to correlate positively with the extent of the non-perfused area.³¹ We postulate that the higher intraocular VEGF levels in RVO and DME than in nAMD are due to the larger size of the ischemic area in RVO and DME. 

The commonality of high cytokine levels that we observed in three retinal diseases can be explained as follows. Although the pathogenesis and affected tissues of these diseases are not identical, the response to hypoxia shares common features at the molecular level, including pathways related to cytokines such as VEGF. Theoretically, each of these cytokines is associated with the upregulation of VEGF in the retinal response to hypoxia. SCF/c-KIT signaling promotes phosphorylation of glycogen synthase kinase-3 beta (GSK-3β), which results in the translocation of β-catenin, which enhances the transcription of target genes involved in angiogenesis.⁶ In turn, this stimulates the expression of WNT target genes, including VEGF, TNF-α, and hypoxia-inducible factor 1 (HIF-1).³² Aberrant activation of WNT signaling plays a pathogenic role in retinal diseases, including DR and AMD.³³ On the other hand, GAL-1 aberrantly stimulates VEGFR2, the receptor for VEGF-A, which in turn activates the VEGF signaling pathway. Hypoxia facilitates endothelial cell remodeling, which modulates the binding of GAL-1. This leads to the activation of VEGFR2.³⁴ 

Changes of Cytokine Concentration in Response to Anti-VEGF Therapy in nAMD Patients

As shown in Table 3, the concentrations of VEGF and GAL-1 were decreased after anti-VEGF treatment in patients with nAMD, while those of SCF and c-KIT showed no change. This result indicates that intravitreal injections of anti-VEGF agents have no influence on the intraocular concentration of SCF and c-KIT. In addition, we found a significant differences in concentration levels of c-KIT, GAL-1, and VEGF-A between groups of nAMD patients with good and bad treatment responses (Tables 4, 5). Based on these findings, the aqueous concentrations of c-Kit and Gal-1 could serve as good biomarkers of functional cell survival in AMD, even when obtaining morphological success after anti-VEGF therapy. 

We interpreted the results to mean that c-Kit and GAL-1 are induced more in response to more severe disease states that threaten photoreceptor survival. It is known that SCF/c-KIT have protective roles for photoreceptors against harmful environments and inflammation, resulting in improved survival of photoreceptors. In a murine model, c-Kit mutation resulted in exacerbated photoreceptor degeneration, and the overexpression of SCF can protect photoreceptors from light toxicity and genetic disease model of retinitis pigmentosa. The c-Kit⁺ retinal progenitor cells derived from retinal organoids suppress microglial activation, gliosis, and the production of inflammatory mediators, resulting in improved survival of grafted cells in retinal degenerative disease models.³⁵ GAL-1 is closely related to VEGF in that it affects VEGF-receptor segregation, internalization, trafficking, endothelial cell proliferation, migration, and vascular permeability.³⁶ 

We postulate that SCF did not show such a relationship with treatment response because of its characteristic regulation with positive feedback. Binding of SCF to c-KIT upregulates HIF-1α, which induces and upregulates the expression of SCF in response to both hypoxia and growth factor receptor activation.³⁷ Within a positive feedback loop, SCF tends to be easily overexpressed when there is stimulation such as disease-induced hypoxia, which may negate the proportionality of SCF to disease severity.³⁷ 

Limitations

This study had several limitations. First, it was a retrospective registry study and exploratory, as it used the aqueous humors from patients who consented to banking. Therefore, the number of patients in each group and the disease severity were not calculated or controlled before the study. We could not enroll enough patients to ensure statistical power in disease groups other than nAMD. The sample size for the Spearman correlation analysis was 194 with type I and type II error rates of 0.05 and 0.2 when the expected correlation coefficient was assumed to be 0.2. Therefore, the only disease that met this sample size was nAMD. DME, RVO, dry AMD, and the control group did not meet the sample size requirement for Spearman's correlation analysis. 

Second, except for 29 patients, the participants did not undergo secondary aqueous sampling. To include as many patients as possible in the reporting of baseline concentration profiles, patients who only sampled at baseline were included. The main reason for not performing secondary sampling was the difficulty in preparing for the clinical procedure, such as using a cryopreservation device and microscope. This reduced the statistical power of the analysis. We have reported the concentration profiles of these cytokines as a descriptive study to aid future analytical studies. Previously, the three cytokines other than VEGF were known to be related to cellular responses to hypoxia in the retina, but information was lacking in patients. A future follow-up study with a larger number of patients could find that the insignificant differences in our study become significant. 

Third, two anti-VEGF agents (ranibizumab and aflibercept) were used as treatment, which may have resulted in different treatment responses and cytokine changes. Also, measuring the concentration in the aqueous humor is an indirect measure of the concentration in the target tissue and the retina, and there may be discrepancies between the aqueous and retinal tissues. 

Conclusions

In this study, we showed that concentrations of SCF, c-Kit, and GAL-1 were elevated in the aqueous humors of eyes of patients with nAMD, DME, and RVO. These levels also correlated with VEGF-A concentrations. The current mainstay of treatment for retinal vascular diseases is anti-VEGF agents. These drugs target VEGF-A and VEGF-B, and a significant number of patients show a poor response to treatment. As mentioned, the pathogenesis of retinal disease is not limited to VEGF and may involve other cytokines. A better understanding of the intraocular profiles of these cytokines and their role in response and resistance to treatment is essential for the future development of new agents related to these cytokines. 

Acknowledgments

Supported by Novelty Nobility (NN2019); a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute funded by the Ministry of Health & Welfare, Republic of Korea (HI20C0025); and a National Research Foundation of Korea grant funded by the Korea government (MSIT; RS-2023-00248480). 

Disclosure: Y.J. Choi, None; H.M. Kim, None; T.-Y. Na, Novelty Nobility (E); K.H. Park, None; S.G. Park, Novelty Nobility (F); S.J. Woo, Novelty Nobility (C, F)