1 The role of microglia activation in early glaucomatous optic neuropathy
This study found that when IOP was pathologically elevated, microglial in ONH could sense mechanical changes and be transformed into a harmful phenotype, resulting in the secretion of pro-inflammatory cytokines, which would induce neurotoxic astrocytes, a triggering event at the early stages of glaucoma. Recently, a study using a chronic glaucoma cat model with LTBP2 gene mutation also confirmed that the activation of microglial cells in ONH occurs earlier than the activation of astrocytes, further proving an earlier activation of microglial cells in the early stages of the neuroinflammation [26]. Burgoyne et al. recently conducted a study employing laser-induced damage to trabecular meshwork cells to construct a non-human primate model of glaucoma, which confirmed that the activation of astrocytes may not be the earliest manifestation of glaucomatous damage [27].
Our study showed that microglial cells mainly secreted pro-inflammatory cytokines such as TNF-a, IL-1a, and IL-1β after mechanical intervention. These findings align with previous research indicating that inflammation-induced microglial release of TNF-a, IL-1a, and C1q can induce neurotoxic astrocytes [9]. However, significant upregulation of C1q was not observed in this study (Fig. 2C), suggesting that initial mechanical intervention may not promote C1q secretion and stimulation from inflammatory factors or other harmful factors may be required.
In addition, the inflammatory factors secreted by activated microglia after mechanical intervention not only induce neurotoxic astrocytes but also have inflammatory damage effects. The IL-1 family is the main regulatory factor for triggering innate immunity and inflammation, with IL-1α and IL-1β as the main pro-inflammatory factors [28]. An acute high IOP mouse model showed that IL-1β can promote RGC death [29], thus IL-1β is involved in both inducing neurotoxic astrocytes [9] and the damage of glaucomatous optic nerve. TNF-α can lead to RGC death and inflammatory damage through the TNFR signaling pathway, which has been confirmed in glaucoma animal models and pathological retinas of glaucoma patients [30-32]. Therefore, the pro-inflammatory factors secreted by microglia under mechanical stress could also lead to RGC loss and play a direct role in glaucomatous optic nerve damage.
Studies have suggested that in the progression of glaucoma, monocytes that are homologous with microglial cells infiltrate the ONH, potentially contributing to astrocyte activation and inflammatory responses. In a DBA2J glaucoma mouse model, 9-month-old mice were used (in which IOP elevation started in 6-month-old and optic nerve damage began to manifest after 9-month-old) to simulate the long-term effects of chronic high IOP and early-stage pathological damage. The results showed that the molecular pathway changes in microglia primarily involved metabolic pathways, mitochondrial-related proteins, and oxidative phosphorylation, suggesting that the inflammatory pathway may be primarily driven by infiltrating monocytes in ONH in later stages of glaucoma [33]. However, in a chronic glaucoma animal model using intravenous injection of hypertonic saline, it was observed that both astrocytes and microglia proliferated significantly in ONH in the early stages of the disease, with no evidence of monocyte infiltration [34]. Clinical samples from glaucoma patients also did not show infiltrating monocytes in the retina or optic nerve [5]. Moreover, under normal physiological conditions, astrocyte endfeet form tight connections with vascular endothelial cells and pericytes to form the neurovascular unit and the blood-retinal barrier to maintain the "immune privilege" [35]. This phenotype should prevent immune cell infiltration in the absence of astrocyte activation or damage, indicating that monocyte infiltration is not an early event triggered by mechanical intervention. The animal model used in our study exhibited a more significant and rapid increase in IOP compared with DBA2J glaucoma mouse model, with a relatively short duration of IOP elevation. There was no significant monocyte ( no both PTPRC+ and ITGAX+ spot) infiltration observed in the spatial transcriptome(Additional file 3), suggesting that microglial activation is the primary factor. In addition, we applied mechanical intervention to microglia in vitro and revealed that the conditioned medium after intervention could induce neurotoxic astrocytes, further confirming that the involvement of exogenous monocytes is not necessary.
2 Microglial play a role in early glaucomatous optic nerve damage by sensing mechanical changes mainly through Piezo1.
Microglia are recognized as sensors of the nervous system that detect damage signals including immune and inflammatory factors. In neurodegenerative diseases, they are responsible for sensing various types of damage signals through some specific receptors, activating the neuroimmune system, and playing a protective or detrimental role [36]. For example, Toll-like receptors (TLR) respond to sense damage-associated molecular patterns (DAMP) and Trem2 initiates activation in response to neurodegeneration-associated molecular patterns (NAMP) [37, 38]. However, the specific sensors utilized by microglia to respond to mechanical stimuli remain unclear. Studies have shown that P2X7 on the surface of microglia can detect mechanical changes and activate microglia, yet P2X7 can also sense ATP changes caused by mechanical forces, indicating that it is not a pure mechanical stress sensor [14, 39]. Our study identified the specific role of the ion channel Piezo1 as a pure mechanical sensor in converting mechanical signals into biological signals at the early stages of glaucoma. However, we cannot rule out the possibility of other receptors acting at the same time, because the addition of piezo inhibitors did not fully restore the cytokines to control level. It would be an important question for future studies.
The researchers only started to gain a better understanding of the PIEZO family, a major class of mechanosensors, over the past decade [24, 25]. They play a critical role in a range of physiological functions and diseases related to mechanics, such as touch [40, 41], baroreception [42], red blood cell function [43], and vascular development [44]. Previous studies have demonstrated the expression of Piezo1 in certain astrocytes in the brain, and under inflammatory conditions, it can upregulate or activate the secretion of inflammatory factors that inhibit astrocytes, serving as a negative feedback mechanism for astrocyte activation [45]. Additionally, astrocytes in the brain can sense changes in the mechanical microenvironment by releasing Ca2+ and ATP, which promotes neuroregeneration and behavioral functions such as memory and learning in the hippocampus [46]. In retina, Tatjana C. Jakobs et al. also detected the expression of Piezo family in astrocytes in the ONH using single-cell quantitative PCR [47]. However, both Yang Liu et al. and our team have found that, unlike astrocytes in the brain, Piezo1 on the surface of astrocytes in ONH primarily participated in the regulation of cell cycle and cytoskeletal pathway as a mechanosensor and did not regulate neurotoxic astrocyte activation[18].
Recently,Malko, P et al also demonstrated that Piezo1 channel activation downregulates the pro-inflammatory function of microglial cells that were prior primed or activated with lipopolysaccharide (a widely used to induce inflammatory responses), especially production of TNF-α and IL-6, by initiating intracellular Ca2+ signaling to inhibit the adhesion plaque kinase-dependent NF-κB inflammatory signaling pathway[48]. However, it has also been confirmed that the activation of Piezo1 upregulates the pro-inflammatory function of vascular endothelial cells that response to the pathological blood flow shear forces by initiating intracellular Ca2+ signaling to activate the NF-κB pathway[49]. These results indicate that the biological functions of intracellular Ca2+ are complex multidirectional action. In the case of inflammation, intracellular Ca2+ are already out of balance, piezo1 activation may regulate intracellular Ca2+ homeostasis, thus inhibiting inflammation, as show in Choi’s study[47]. When the cells are normal, the intracellular Ca2+ are in balance, the intracellular Ca2+ flow induced by piezo1 activation caused by abnormal mechanical forces may lead to the imbalance of intracellular Ca2+ homeostasis, thus playing a role in promoting inflammation, as shown in our study and Malko, P’ s results[48].
There is also a study on the association between Piezo1 polymorphisms and glaucoma. Researchers had discovered that over 30% of African populations carry point mutations in Piezo1 [43], and African populations have a higher risk of developing glaucoma and tend to exhibit more severe glaucoma phenotypes [50]. Therefore, researchers specifically selected African populations carrying the most common Piezo1 mutation (e756del) to examine glaucoma phenotypes and conduct follow-up studies. The results showed that the mutation carriers had higher IOP, thinner nerve fiber layer thickness, and more severe disease progression during the follow-up, though without significant difference. When the same mutation was introduced in mice, no significant impact on IOP and RGC morphology was observed [51]. These findings suggest that the association between Piezo1 and glaucoma may not be straightforward and requires further investigation. The researchers also acknowledged certain limitations of the clinical data. First, despite careful consideration, the potential confounding effect of age on the observed phenotypes cannot be entirely ruled out. In addition, dividing participants into glaucoma patients and normal individuals in the study resulted in relatively small sample sizes within each subgroup.
Our study suggested that regulating the mechanical activation of microglia could be a potential intervention strategy for glaucoma. There have been many reports on targeting microglia to regulate optic nerve damage in glaucoma. In a rat model of ischemia-reperfusion, intravitreal injection of CSFR antibodies inhibited microglial proliferation and activation, promoting RGCs survival and partial recovery of visual function [52]. Other studies have shown that intervention with minocycline, which suppresses microglial activation, can counteract axoplasmic flow impairment and neuronal damage caused by high IOP [53]. Recent studies showed that radiation therapy applied to the eyes or head of glaucoma model mice can effectively protect against optic nerve damage, with a significant reduction in microglial numbers playing a role in the therapeutic mechanism [54, 55]. However, there are also conflicting findings. For example, in the optic nerve crush model, depletion of microglia using the compound PLX5622 had no significant effect on RGC and optic nerve damage but influenced astrocyte rearrangement in the optic nerve region [56]. In a mouse model of NMDA-induced optic nerve damage, intervention with PLX5622 eliminated microglia and yet accelerated RGC loss [57]. These findings suggested that the activation of microglia exhibits heterogeneity, and that inhibiting the pathological harmful phenotype of microglia while preserving their beneficial physiological function would be more advantageous than directly eliminating microglia. Our study found that Piezo1 could provide a potential intervention target for inhibiting the harmful phenotype of microglia.
3 First description of spatial transcriptomic characteristics of glaucomatous ocular tissues
Our study provided, for the first time, insights into the changes in RNA transcription level in the retina, ONH, and optic nerve in glaucoma. Previous transcriptome studies of glaucoma animal models have mostly used microarray analysis of total RNA from retinal tissue, such as in the high IOP model of rats injected with hypertonic saline or the DBA2J inherit glaucoma mouse model, and the early changes in those models, mostly consistent with the spatial transcriptome results of our study, mainly involve the pathways of immune reactions and glial activation [58, 59]. There has been a lack of targeted research on ONH, the main location that is affected at the beginning of glaucomatous pathological changes, especially at the in vivo level. The mechanical stress mainly applies to ONH in the process of optic nerve damage in glaucoma [21-23]. This is the location where the RGC axons exit the eyeball and the axons are directly exposed to both intracranial and intraocular pressure and are located at the blind end of cerebrospinal fluid reflux, making them highly susceptible to mechanical factors. Previous finite element analysis showed that the ONH experiences the strongest stress and strain during pressure changes [56]. The spatial transcriptome sequencing results of our study at the in vivo level suggested that the most rapid change at the early stage of glaucoma occurred at the ONH and was mainly initiated by glial activation of astrocytes and microglia.