The effect of systemic levels of TNF-alpha and complement pathway activity.

Updated: Nov 11


Systemic levels of pro-inflammatory cytokines and activated complement components affect the risk and/or progression of neovascular age-related macular degeneration (AMD). This study investigated the effect of serum pro-inflammatory cytokine levels and complement pathway activity on the clinical response to vascular endothelial growth factor (VEGF) inhibition in neovascular AMD.

Methods Sixty-five patients with a new diagnosis of neovascular AMD were observed over a six-month period in a single-centre, longitudinal cohort study. At each visit, the visual acuity score (VAS), central macular thickness (CMT), serum levels of CRP, pro-inflammatory cytokines (TNF-α, IL-1β, IL-2, IL-6 and IL-8), and complement pathway activity were measured. Participant DNA samples were sequenced for six complement pathway single nucleotide polymorphisms (SNPs) associated with AMD. Results A statistically significant difference in VAS was observed for serum levels of TNF-α only: there was a gain in VAS (from baseline) of 1.37 for participants below the 1st quartile of mean concentration compared to a reduction of 2.71 for those above the 3rd quartile. Statistical significance was maintained after Bonferroni correction (P value set at <0.006). No significant differences in CMT were observed. In addition, statistically significant differences, maintained after Bonferroni correction, were observed in serum complement activity for participants with the following SNPs: CFH region (rs1061170), SERPING1 (rs2511989) and CFB (rs641153). Serum complement pathway components did not significantly affect VAS. Conclusions Lower serum TNF-α levels were associated with an increase in visual acuity after anti-VEGF therapy. This suggests that targeting pro-inflammatory cytokines may augment treatment for neovascular AMD.

Introduction Age-related macular degeneration (AMD), a progressive retinal disease that results in the loss of central vision, is predicted to affect 288 million people worldwide by 2040 [1]. Neovascular AMD (nAMD) is a result of choroidal neovascularisation (CNV) and leads to rapid vision loss. The mainstay of current treatment is inhibition of vascular endothelial growth factor (VEGF) [2]. The evidence base for a genetic component in AMD is significant, and numerous single nucleotide polymorphisms (SNPs) have been associated with a patient’s risk of developing AMD [3]. SNPs in genes of the complement pathway, including the complement factor B (CFB) gene region [4, 5], the C2 [4, 5] and C3 [6] genes have been reported to affect the risk of developing AMD. Uncontrolled activation of the complement pathway is limited by a set of complement regulatory proteins: Factor H and Factor I (encoded by the CFH and CFI genes, respectively), regulate the alternative complement pathway [7], whereas the C1 inhibitor is a regulator of the classical pathway [8]. Genetic variants at the Regulators of Complement Activation (RCA) locus on chromosome 1, which contains the CFH gene, contributes to AMD risk [9,10,11], in addition to the CFI gene region on chromosome 4 [12,13,14], and the SERPING1 gene that encodes the C1 inhibitor [15, 16]. Studies have shown elevated levels of complement activation fragments to be independently associated with AMD [17,18,19]. Furthermore, complement activation has been demonstrated to be associated with stage of AMD [20]. In addition, systemic activation of the alternative complement pathway and complement components is associated with AMD genotypes [21], including the CFH SNP rs1061170 (Y402H) [19] and the CFI region SNP rs10033900 [17, 21]. A meta-analysis by Hong et al. reported that treatment-naïve patients carrying the CFH SNP, rs1061170 (Y402H), were more likely to achieve an improved outcome to anti-VEGF treatment [22]. Furthermore, visual outcome was improved after anti-VEGF treatment for patients carrying a low-risk CFH genotype and low CFH risk score [23]. Expression of acute phase proteins and pro-inflammatory cytokines can also affect the risk of AMD development and/or progression: CRP is an acute phase protein and marker of systemic inflammation that is an independent risk factor for AMD [24]. IL-6 is a known cytokine stimulus of CRP release by the liver [25], and both have been associated with AMD progression [26]. CRP has been demonstrated to induce IL-8 expression by human retinal pigment epithelium (RPE) cell lines [27], and both IL-6 and IL-8 are expressed by RPE cells on complement activation [28], by degenerating RPE cells [29], and are associated with drusen formation [30]. Systemic levels of IL-6 have been found to be associated with the progression rate of geographic atrophy secondary to AMD [31]. In addition, patients with AMD have been shown to express higher levels of circulating IL-1β than age-matched controls [32]. IL-2 has been implicated in the pathogenesis of AMD as activation of IL-2 signalling pathways has been observed [33] and IL-2 contributes to extracellular matrix formation and the development of fibrosis in AMD [34]. TNF-α, a pro-inflammatory cytokine that is known to mediate CNV formation in experimental models by upregulating VEGF expression by RPE cells [35], has also been demonstrated to promote the angiogenic drive of active CNV lesions [36]. Patients with elevated levels of serum TNF-α have been shown to respond favourably to VEGF inhibition [32]. Although the studies mentioned above have investigated the role of complement pathway SNPs, complement pathway activity and systemic concentrations of pro-inflammatory cytokines on AMD pathogenesis, relatively few studies have investigated their functional effect on outcomes of VEGF inhibition. The primary aim of this study was to investigate the effect of serum levels of pro-inflammatory cytokines (TNF-α, IL-1β, IL-2, IL-6 or IL-8) and complement pathway activity on the clinical response to VEGF inhibition in neovascular AMD. A secondary aim was to investigate the effect of complement pathway SNPs, associated with AMD, on serum complement activity in the same cohort of patients.

Materials and methods Study approval, registration and regulation This study was conducted in accordance with the Research Governance Framework for Health and Social Care (2005) and Good Clinical Practice. Ethical approval was obtained from the National Research Ethics Committee (NRES) South Central- Southampton A. This study adhered to the tenets of the Declaration of Helsinki. The University Hospital Southampton NHS Foundation Trust was the sponsor of this study, and The University of Southampton undertook the research study. All patient samples and data were anonymised for the purpose of this study. Patient DNA and serum samples were stored for future studies. Procedures for handling, processing and storage of patient data were in compliance with the UK Data Protection Act (1998). Patient recruitment, consent, and investigation Patients were recruited to the study after informed consent by the ophthalmology department of University Hospital Southampton NHS Foundation Trust. Patients were invited to take part if they met the principle inclusion criteria for the study: (1) over the age of 50; (2) a new diagnosis of neovascular AMD in one eye, treated with an initial loading dose of three, monthly Ranibizumab intravitreal injections; (3) White ethnicity (to limit any effects of ethnic variation on outcomes of VEGF inhibition in neovascular AMD). The exclusion criteria were: (1) bilateral diagnosis of neovascular AMD (one of the exploratory endpoints of the study was the development of nAMD in the second eye); (2) a macular co-pathology; (3) poor venous access that prevents a peripheral blood samples being taken. All patients recruited to this study had a diagnosis of neovascular AMD, confirmed on fundus fluorescein angiography, that was made by a consultant ophthalmologist specialising in medical retina diseases. Indocyanine green angiography was carried out for patients to rule out polypoidal choroidal vasculopathy (PCV)- patients with PCV were not invited to take part in the study. Patients were eligible to enrol for the study after their third intravitreal Ranibizumab injection and subsequently invited to a baseline visit (Fig. 1). Informed consent was taken from participants at this visit, and their demographic details, medical history and baseline LogMAR visual acuity score (VAS) was recorded (number of letters on an ETDRS chart). A baseline central macular thickness (CMT) was also measured using optical coherence tomography (OCT) (Topcon, Berkshire, UK). A blood sample was taken at the baseline visit for serum cytokine and genetic analysis. Participants were reviewed by a study investigator and received treatment with an intravitreal ranibizumab injection if they had active neovascular AMD. Following the baseline visit, participants attended for six, monthly follow-up visits. At each visit, the VAS and CMT was recorded, a blood sample was taken, and the patient reviewed by a study investigator before any treatment for active disease.

Detection of serum cytokine levels and activated end components of complement pathways Serum was isolated from participant blood samples using standard density-gradient ultracentrifugation at 1355 × g for 10 min at 21 °C (Eppendorf, Stevenage, UK). Patient serum cytokine levels were measured using semi-quantitative assays by Meso Scale Discovery (Rockville, Maryland, USA) as per the manufacturer’s instructions. All cytokine measurements were undertaken in triplicate using the assay, and cytokine measurements were within the reading range of the kit. Functional assessment of classical and alternative pathway complement activity in patient serum samples was undertaken using Wieslab semi-quantitative ELISA Assays (SVAR Life Sciences, Malmo, Sweden) as per the manufacturer’s instructions. Measurement of activated end components of classical and alternative complement pathways was expressed as a percentage relative to the fluorescence intensity of the positive control, derived from human serum components, supplied with the testing kit. Genetic analysis DNA was extracted from peripheral blood mononuclear cells of patient blood samples using erythrocyte lysis buffer (Fisher Scientific, Loughborough, UK) as previously described [37]. DNA concentrations were measured using the Nanodrop ND1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA). Sequence analysis of participant DNA samples was undertaken by LGC Genomics (Hoddesdon, UK) on the following six SNPs associated with the complement pathway and AMD risk: CFH region: rs1061170; CFI region: rs10033900; SERPING1/C1-INH: rs2511989; CFB: rs641153; C2: rs9332739; C3: rs2230199. Statistical analyses The GraphPad Prism software version 8.2 (GraphPad Software, Lo Jolla, Ca, USA) was used for statistical analyses and graphical representation of the data obtained in this study. Assessment of normality of continuous variables was determined by quantile–quantile plots of the residuals using GraphPad Prism. The unpaired t test with Welch’s correction was used to determine statistically significant differences in changes of visual acuity scores, central macular thickness and percentage activity of activated end components of complement pathways compared to positive controls. Statistical significance was set at the P < 0.05 value. As this is a preliminary/pilot study, the patient sample size was determined using a rationale laid out by S.A. Julious where a sample size of at least 12 is recommended [38]. Our patient cohort was stratified into quartiles of ~16 in line with this recommendation.

Results Serum classical or alternative complement pathway activity and functional response to anti-VEGF intravitreal injections A total of 65 patients with a new diagnosis of neovascular AMD were recruited to participate in this study (Fig. 1). Participant demographics are summarised in Table 1. Study participants were stratified into quartiles according to average serum concentration of an inflammatory protein over seven study visits, in order to amplify the functional effects of small changes in serum concentration (Table 1). The study first investigated any significant differences in the visual acuity score (VAS) or central macular thickness (CMT) change from baseline at each visit between participants who had a mean serum concentration of classical pathway (Supplementary Fig. 1A, B) or alternative pathway (Fig. 2A, B) complement components below the first quartile and above the third quartile. There was a statistically significant difference in the VAS change from baseline, −2.78 (SD = 7.01) vs. −0.34 (SD = 8.51) for mean serum alternative pathway components (P = 0.048), using an unpaired t test with Welch’s correction (Fig. 2A), but significance was not maintained after a Bonferroni correction was applied (P value set at <0.006).