Glaucoma, affecting over 60 million people globally, is the second leading cause of blindness, with POAG being predominant. Glaucoma is characterized by the progressive loss of retinal ganglion cells (RGCs) through apoptosis (1, 2), which shares molecular similarities with central nervous system (CNS) degenerative disorders like AD, indicating a common pathological mechanism. The retina, an extension of the CNS, and its barrier (inner blood-retinal barrier, iBRB) resemble the blood-brain barrier (BBB) (3). It was reported that circulating immune cell migration through an impaired BBB and glial activation contributes to the progression of AD (4). Similarly, impairment of the iBRB is observed in glaucoma, which has been proven to be pivotal in determining the fate or prognosis of neuroinflammation pathological outcomes for POAG (5). A current theory proposes that when systemic immune and inflammatory components entering the retina or brain through the impaired iBRB or BBB initiate a self-exacerbating cycle of neuroimmune responses, then it would lead to the development of clinical disease phenotypes such as POAG or AD (6). Similar pattern has been stated in previous studies about brain structural changes and neurodegenerative processes in relation to clinical severity and cognitive symptoms in glaucoma (7–10). Although, there is little agreement on the role of chronic systemic diseases in developing of neurodegenerative diseases (11, 12), this novel perspective could enhance our understanding of the etiology, mechanisms, and potential therapies for POAG and AD.

DM, a common systemic disease, is associated with heightened risks of both POAG and AD (13, 14). DR and depression are common comorbidities of DM (15, 16), sharing pathological mechanisms that contribute to the breakdown of iBRB and BBB, potentially playing a causative role in POAG and AD (17). Associations have been demonstrated in previous studies, including several meta-analyses (18, 19), between POAG with AD (20, 21), DM and depression (22). Therefore, we hypothesized that DR and depression, serving as indicators of iBRB and BBB impairment, might act as intermediary processes between DM and POAG and AD, and could possibly accelerate the progression of POAG and AD. Meanwhile, these processes were also interfered with intrinsic factors associated with AD and POAG, such as age, gender and APOE E4 status (23). A coherent pathological picture that explains these complex relationships is likely to describe molecular and mechanistic similarities of these disorders and may pave the way for the development of novel and effective therapies. This study aims to develop a comprehensive model exploring the association between DM and POAG, AD, across different gender groups and APOE E4 genotypes in a large cohort, in which depression and DR serve as intermediate factors, with CRP levels evaluated to gauge inflammatory conditions in different contexts.

DM was found to be associated with heightened risks of both POAG and AD in previous studies, likely due to shared pathological mechanisms such as oxidative stress, inflammation, vascular dysfunction, and impaired insulin signaling pathways (25). Thus, our study supports a current hypothesis that POAG and AD may be thought of as diabetes of the brain (25). Additionally, immune and inflammatory components from the systemic circulation enter the brain and retina through the impaired BBB and BRB, which are also implicated in DM-related DR and depression, and could also initiate a self-exacerbating vicious cycle of neuroimmune responses that lead to the development of POAG and AD (6). In this study, we first explored the associations between DM and POAG, AD, on the conception that DR and depression serve as intermediate factors between these associations. We discovered that patients with DR showed higher prevalence of POAG, together with higher prevalence of AD among DM patients with DR and depression. Moreover, we tried to integrate the effects of APOE E4 allele, gender and CRP into our models to enhance our understanding of the common pathological mechanisms in neurodegeneration and further explain the associations among DM, POAG, and AD (26).

Consistent with our findings, an Indian study identified a positive association between glaucoma and DR in a type 2 diabetes mellitus (T2DM) population (OR=2.62) (27). Additionally, other studies have shown that T2DM patients with POAG have a threefold higher risk of developing DR compared to those without POAG (28). Similarly, research using the Danish Registry of Diabetic Retinopathy found that patients with DM and either glaucoma or ocular hypertension were more likely to develop DR within five years (29). Clinical diagnosis of both POAG and DR are often abrupt, making early detection and monitoring of progression challenging. Whilst these findings do not establish causality, they reinforce the association between POAG and DR. Indeed, we continue to regard the development of DR as a crucial factor in the progression of POAG, particularly concerning the role of the BRB in glaucoma. Normally, the BRB’s integrity limits retinal damage; minor disruptions in the BRB are typically transient and quickly repaired, preventing clinical consequences (5). Thus, age-related BRB breakdown alone may not cause POAG. Histological studies have shown BRB damage in glaucoma models, with abnormalities in retinal pigment epithelium (RPE) cells, increased permeability of retinal vessels, and leukocyte infiltration (2). In animal models of transient high intraocular pressure (IOP) glaucoma, BRB impairment has been linked to T-cell infiltration and progressive ganglion cell death (30, 31). These findings suggested that BRB integrity is crucial in POAG progression and breakdown of BRB was the essential factor in the very early stages of POAG (5, 31). DM, especially DR, can compromise the BRB, leading to neuroinflammation in the retina, which may contribute to glaucoma progression. Early dysfunction of the neurovascular unit (NVU) in diabetes has been observed in animal models and patients, leading to impaired neurovascular coupling, loss of autoregulation, and disruption of the iBRB (32). Factors like hypoxia-ischemia, oxidative stress, and inflammation during DR development contribute to both inner and outer BRB breakdown (18, 33–35). These processes indicate that BRB impairments could even precede the clinical onset of DR. The relationship between POAG severity and a significant decrease in retinal vessel density at ONH and macula level further suggests that microvascular damage in the diabetic retina exacerbates retinal neurodegeneration (36, 37). Persistent microvascular leakage in advanced DR induces chronic neural immune-inflammatory responses, eventually leading to neuronal loss (38). Therefore, the disruption of the iBRB caused by DM might explain the significant association between POAG and DM. The higher prevalence of POAG among DR patients could be due to a longer duration of DM and more severe BRB and microvascular damage. Although the precise timing and relationship between glial activation, BRB impairment, and immune-inflammatory cell infiltration in POAG need further clarification, comprehensive research could shed light on the mechanisms underlying progressive RGC loss in glaucoma associated with age-related systemic diseases like DM and DR,which may also explain the progressive RGC loss in glaucoma patients with well-controlled or normal IOP, where BRB impairment, triggered by high IOP or other systemic factors, is likely a critical factor leading to ongoing neuroinflammation and RGC loss independent of IOP levels.

Meanwhile, our study has corroborated the association between DM and AD. Numerous investigations have shown a higher incidence of cognitive decline in individuals with DM. Longitudinal studies conducted in Japan (39) and a five-year prospective study by Yaffe et al. (40) have robustly established DM as a risk factor for AD and cognitive decline, particularly in the context of metabolic syndrome. Neurovascular changes in the brain, similar to those seen in the BRB in DM patients, are believed to contribute to the higher risk of AD (41). The retina, an extension of the CNS, shares similarities with the BBB, suggesting that DM-related BBB impairment, particularly in patients with DR, could be expected (3, 42, 43). The BBB, composed of endothelial cells, pericytes, and astrocytes, regulates the passage of substances into the brain and protects against harmful signals from the bloodstream (44). Migration of circulating immune cells through an impaired BBB, along with glial activation, contributes to the progression of AD and may explain the increased prevalence observed in DM patients, particularly those with DR. Moreover, depression, which can be induced by DM through the strain of diabetes itself and changes in the hypothalamus–pituitary–adrenal (HPA) axis or BBB structure (45–47), has been identified as a risk factor for AD and is prevalent in DM patients. Chronic social stress has been shown to alter BBB integrity in animal models, promoting behaviors akin to depression (48, 49). This suggests that BBB dysfunction may play a role in the development of depression and could contribute to the association between DM, DR, and depression. Furthermore, BBB damage may exacerbate the impact of depression on AD, similar to the key role of BRB disruption in DR on the progression of POAG. Additionally, the APOE E4 allele, a known risk factor for AD, may contribute to BBB breakdown and neuroinflammation, synergizing with systemic inflammation to promote AD onset (50). Our multivariable GLM analysis, adjusted for age and stratified by sex, demonstrated the complex interplay between DM, depression, APOE E4, and AD, underscoring their roles in the pathogenesis of AD.

Although we did not observe an association between POAG and depression after adjusting for age, sex, and VA, previous studies have indicated a higher prevalence of depression in individuals with POAG (22), and increased risk of glaucoma in patients with major depressive disorder (MDD). In our study, females were more likely to experience depression than males, regardless of DM or DR status. Notably, depression was associated with POAG only in females, possibly due to irregular estrogen levels caused by ovarian hormone fluctuations in patients with depression, since estrogen is essential for maintaining RGCs (51, 52). Additionally, we observed a protective effect of the APOE E4 allele in males against POAG, even with DM, which aligns with previous research indicating a reduced risk of POAG associated with this allele (53). Similar findings were revealed by previous studies which showed a difference of AD prevalence among different gender (54). Interestingly, the APOE E4 allele is a known risk factor for AD, suggesting a mechanistic difference between neurodegenerative diseases affecting the eye and brain. This difference may be due to APOE E4 acting as an example of antagonistic pleiotropy—a gene that provides benefits at one stage of life but later presents disadvantages (55)—since POAG typically manifests at a younger age than AD. The post-menopausal loss of estrogen was key in the increased incidence of AD in women (56). Considering the accelerated aging process post-menopause, particularly in females with DM and depression (57–60), the protective effect of the APOE E4 allele may be diminished in this group. Both clinical researches (61, 62) and transgenic mice model (63, 64) has revealed that combination of APOE4 genotype and female sex often exacerbated outcomes across numerous cognitive loss. Although further studies are needed to clarify these mechanisms, these gender differences highlight the complex interplay of DM, depression, and APOE E4 in the context of POAG. Additionally, the correlation between retinal vascular abnormalities, cognitive impairment, and dementia, as well as the similarity between the BRB and the BBB, suggests that retinal function may serve as a promising biomarker for neurodegenerative diseases (65). Previous clinical studies have already shown that POAG patients are at a higher risk of developing AD compared to controls (66) and vice versa (67). Therefore, considering the earlier and more accessible detection of POAG compared to AD, we propose that POAG could potentially be used as an early screening indicator for AD in patients with DM.

Furthermore, although we attempted to elucidate the connection between DM, POAG, and AD by assessing inflammation levels, represented by blood CRP levels, and hypothesized that proinflammatory cytokine levels might explain breaches in the BRB and BBB, our study did not find any association between blood CRP levels and POAG or AD. Despite observing increased CRP levels in DM, which further increased in patients with DR and depression, particularly among females, no correlation with POAG or AD was identified. We further tried to integrate the effects of APOE E4 allele, gender and CRP into our models to enhance our understanding of the common pathological mechanisms in neurodegeneration and further explain the associations among DM, POAG, and AD (26). CRP is a marker of systemic inflammation and increases with age. Although multiple AD-related genes are associated with the level of CRP, the association between blood CRP levels and risk of AD are not conclusive in the literature, with studies showing both low and high levels of CRP in patients with AD (68). Similarly, peripheral inflammatory markers have been linked to a higher risk of vascular dementia, but not AD (31). A Mendelian randomization study suggested a protective effect of CRP on AD, possibly influenced by the methods used to measure CRP (69). It was reported that in the context influenced by possessing the APOE E4 allele, variations in CRP levels—either transformation or suppression—may lead to reduced blood CRP levels (70, 71). Previous studies have observed an inverse correlation between APOE E4 and blood CRP levels, but only within the lower CRP level range, across both elderly and younger populations (72–74). In a large UK Biobank subsample with elevated CRP levels, it was shown that the effect of genetic variants on CRP diminished as CRP levels increased (75). This suggests that chronic, low-grade inflammation, as indicated by CRP, is strongly influenced by genetic factors such as APOE E4, rather than reflecting CRP’s response to acute infections or other stimuli. This may explain why APOE E4 carriers in our study exhibited lower CRP levels regardless of gender or diabetes status, and why we found no association between CRP levels and AD or POAG. The complex inflammatory profiles involving APOE E4, AD, DM, and POAG may offer insights into the systemic inflammation underlying these distinct but related neurodegenerative processes. Future research investigating other proinflammatory cytokines, such as interleukin-6 (IL-6), and exploring CRP’s role in modulating genetic risk for POAG and AD, may further elucidate these complex interactions (76).

Although the underlying mechanisms still require further investigation, previous clinical studies and our data analysis have come to similar conclusion that POAG might come from the consequence of a primary neurodegenerative disease of the CNS, together with BRB and BBB breakdown (77). This insight opens up potential treatment pathways that focus on neuroprotective molecules rather than IOP-targeted medications for POAG. Currently, Coenzyme Q1 (CoQ10) and citicoline are among the most commonly used neuroprotective agents in POAG treatment. CoQ10, a crucial antioxidant that safeguards proteins and DNA from oxidative stress, has been shown to prevent optic nerve astrocyte activation induced by hydrogen peroxide in vitro while also inhibiting RGC apoptosis and loss in animal models (78). Clinical trials and animal studies further suggest that CoQ10 protects optic nerve head (ONH) astrocytes from oxidative stress primarily by preserving mitochondrial function (79). Similarly, citicoline has demonstrated neuroprotective properties by reducing glutamate-mediated excitotoxicity and oxidative stress through the enhancement of neurotrophin levels and mitochondrial support (78). Preliminary studies indicate that the combination of citicoline and vitamin B12 eye drops can help stabilize neuroretinal degeneration and mitigate microvascular damage in DR patients (80). Given these findings, the combined use of CoQ10 and citicoline presents a promising new strategy for glaucoma treatment. Overall, these emerging therapeutic approaches for POAG further support our findings.

Several limitations were identified in our study, particularly our inability to pinpoint the exact onset times of DM, DR, depression, POAG, and AD. Determining the precise onset of each disease is challenging because symptoms may not manifest in early stages, and recorded onset times often only reflect the age at diagnosis. This limitation has prevented us from fully elucidating the causal relationships and could only presented as possible connection between these conditions. Also, we used ICD-10 codes for ascertaining clinical phenotypes might capture a heterogeneous group of patients. An ideal study sample would include a large, population-based cohort with comprehensive ocular phenotyping and chronic disease phenotypes. However, clinical criteria for diagnosis vary in sensitivity across different diseases, which would still impact the consistency and reliability of our findings. Secondly, in this study, we only included age, sex, CRP, and APOE E4 as variables. However, there are undoubtedly additional variables that could influence the complex interplay among DM, POAG, and AD. We believed our results were still valuable for understanding the shared pathological mechanisms involved in DM, POAG, and AD. Thirdly, although European ancestry represented the predominant ethnic group within this study cohort, ethnic variations in glaucoma prevalence and APOE E4 allele distribution were not addressed, limiting analysis to diverse ethnic groups. Including more diverse populations in future analyses could strengthen the findings and enhance their applicability to global populations.

In summary, our findings suggested that DR and depression, as comorbidities related to BRB and BBB impairment in patients with DM, may play crucial roles in the development of POAG and AD among DM patients. Although the complexities of these interactions require further detailed characterization, they provide valuable insights into the underlying mechanisms and potential shared pathways in the development of POAG, AD, and other neurodegenerative diseases. Enhancing knowledge and awareness of these associations could lead to the development of new avenues in the understanding and management of glaucoma, AD, and DM.