Retinal vein occlusion (RVO) occurs when a blockage in one of the major retinal veins obstructs blood outflow from the retina, causing accumulation of fluid, blood, and inflammatory cells. This condition affects between 1 and 2% of persons over the age of 40 (Ho et al., 2016), and can cause variable degrees of vision impairment, depending on the location (macular vs. peripheral), duration, and severity of ischemic injury.

Current treatments for RVO target retinal edema and neuroinflammation via anti-VEGF (vascular endothelial growth factor) treatments, or reduce broad-spectrum inflammation through corticosteroids (Ho et al., 2016). VEGF levels increase in response to hypoxia, and act on endothelial cells to promote vasodilation and increase vascular permeability (Apte et al., 2019). VEGF-neutralizing therapies can be highly effective at resolving retinal swelling and improving visual function, however individual treatment response is variable. Refractory edema persists in over 50% of eyes treated for RVO (Campochiaro et al., 2014), and over 20% of eyes experience visual acuity loss of >15 ETDRS [Early Treatment Diabetic Retinopathy Study] letters over 5  years (Wecker et al., 2017). RVO can damage retinal microvasculature, resulting in poor retinal perfusion and capillary ischemia. Ischemic RVO, typically defined as cases where the ischemic index (percent area of non-perfused retina) is ≥30%, is associated with worse visual outcomes, and poor treatment response (Khayat et al., 2018).

The laser-induced murine model of RVO has been used extensively to investigate mechanisms underlying RVO pathology (Khayat et al., 2017). In this model, an image-guided laser system is used to create a localized occlusion to one or more branch retinal veins. Optimized RVO induction and evaluation protocols enable highly reproducible measurement of retinal edema, inflammation, and neuronal injury (Colón Ortiz et al., 2021; Chen et al., 2022). Consistent with clinical findings, inhibiting VEGF signaling attenuates retinal edema and improves retinal nonperfusion following experimental RVO (Fuma et al., 2017). One novel mediator of retinal injury in RVO is caspase-9, which acts as a multimodal instigator of neurovascular and astroglial dysfunction (Avrutsky et al., 2020; Colon Ortiz et al., 2022). Selective in vivo inhibition of caspase-9 can be achieved through administration of Pen1-XBir3 (Akpan et al., 2011; Avrutsky et al., 2020), a formulation containing the caspase-9-inhibitory domain of XIAP protein (BIR3) (Denault et al., 2007) crosslinked to a cell penetrating peptide, Penetratin-1 (Dupont et al., 2011). Typically known for its role as an initiator of the intrinsic apoptosis pathway, caspase-9 mediates diverse inflammatory and degenerative pathologies through both apoptotic and nonapoptotic mechanisms (Avrutsky and Troy, 2021). In RVO, both VEGF and caspase-9 signaling modulate ischemic injury by regulating vascular endothelial cells.

Here, we compared the efficacy of VEGF-neutralization against a cell permeant topical caspase-9 inhibitor (Pen1-XBir3), following induction of RVO. Using an integrated panel of ophthalmic imaging readouts, we evaluated the progression of retinal ischemia, edema, and neurodegeneration in wild-type mice treated with either a VEGF-neutralizing antibody or a topical caspase-9 inhibitor.

The standard of care for RVO is therapy with injections of anti-VEGF. We have shown that targeting caspase-9 in a mouse model of RVO provides substantial morphologic, cellular and functional protection (Avrutsky et al., 2020; Colon Ortiz et al., 2022). To compare the efficacy of inhibiting caspase-9, vs. blocking VEGF, mice were followed for 8 days following induction of RVO and treatment with either caspase-9 inhibitor eyedrops (Pen1-XBir3), or intravitreal injection of anti-VEGF antibody. We also evaluated a group that received combination therapy to determine if there were additive or potentiating effects of the treatment. Anti-VEGF/PBS administration was performed 16 h prior to induction of RVO to allow for screening out of animals with retinal morphological changes due to the intravitreal injection procedure. Caspase-9 inhibition was achieved by administration of 10 μg Pen1-XBir3 eyedrops immediately following RVO induction, followed by a second dose at 24 h as described in (Avrutsky et al., 2020). RVO-induced retinal pathology was measured by OCT imaging at 1, 2, and 8 days. ERG imaging was performed on dark-adapted animals at 7 days (Figure 1A). Occlusions induced by the laser RVO model typically last through 48 h, and resolve within 1 week. Consequently, the RVO model induces transient retinal detachment and intraretinal swelling characterized by thickening of GCL, IPL, INL and OPL layers at 24–48 h post RVO. By 8 days post-RVO, retinal edema resolves, revealing atrophy of retinal neuronal layers (Figure 1B).

This study represents a framework for comparing in vivo efficacy of experimental treatments against a clinically-established mechanism of action (VEGF-neutralization) in a mouse model of RVO. Our prior work has shown that non-apoptotic activation of caspase-9 regulates edema, gliosis and neuronal dysfunction in a well-defined mouse model of RVO (Avrutsky et al., 2020; Colon Ortiz et al., 2022). Here, we compare the functional efficacy of Pen1-XBir3, a caspase-9 inhibitor, with an established VEGF neutralizing treatment protocol. The data show that topical treatment with Pen1-XBir3 is equal or superior to intravitreal delivery of VEGF neutralizing antibodies across multiple in vivo measures.

Orthogonal measures capture different dimensions of RVO pathology in response to caspase-9 inhibition, VEGF-neutralization, or a combination treatment. Notably, our study was powered to capture therapeutic differences between Pen1-XBir3 treatment versus control, and was not powered to evaluate differences between the combination treatment and either monotherapy. Since neither therapy offers complete retinal protection by itself, a combination treatment was attempted to test the feasibility of using caspase-9 inhibition as an adjunct to VEGF neutralization. To maximize potential efficacy, both treatments were applied at maximal possible dose, based on limitations of intravitreal injection and eye-drop volumes. Stronger protection in eyes treated with the caspase-9 inhibitor suggest that caspase-9 signaling is a more critical target for treating retinal edema and capillary ischemia in RVO. Meanwhile the combination of the two therapies was more effective than anti-VEGF monotherapy, but did not show any improvement over the efficacy of Pen1-XBir3 alone. One limitation of this study is that this was the first attempt to co-administer caspase-9 and VEGF antagonists in a retinal injury model. The dosing levels and methods of delivering the respective treatments were not optimized for a combination therapy approach.

This study captured functional and morphological endpoints, and did not enable molecular profiling of signaling changes in response to the treatment arms. Future studies will be required to determine whether and how caspase-9 and VEGF signaling may interact in RVO. Both caspase-9 and VEGF are multimodal regulators of tissue injury and hypoxia response, with several intersecting mechanisms of action. Both proteins can regulate endothelial cell survival through modulation of autophagy and mitochondrial function (Domigan et al., 2015; An et al., 2019; Spengler et al., 2020). Additionally, both VEGF and caspase-9 have multiple immunomodulatory activities (Reinders et al., 2003; Avrutsky and Troy, 2021).

Using in vivo measures over time increases the translational relevance of preclinical models by utilizing techniques which are commonly used in the clinic to diagnose and follow the evolution of disease in persons receiving treatment for RVO. OCT measures retinal thickness, ischemia (using DRIL) and inflammation (tracking HRFs). Fundoscopic imaging follows resolution of retinal occlusions. ERGs reflect integrity of the retinal neuronal network. Efficacy of each treatment arm was equivalent with regard to HRFs but for all other measures Pen1-XBir3 provided superior efficacy compared to VEGF neutralization. Pen1-XBir3 protected neuronal function, while VEGF neutralization did not affect ERG response. These findings are consistent with clinical data noting extensive retinal degeneration after successful resolution of edema with anti-VEGF treatment (Hasegawa et al., 2017), and lack of functional improvements in retinal regions of severe nonperfusion at baseline (Rachima et al., 2020).

Comparison of VEGF-neutralization and caspase-9 inhibition strategies in a murine model of RVO demonstrate vascular protection by anti-VEGF, and both vascular and neuronal protection with inhibition of caspase-9. These data demonstrate comparison of an experimental therapy against a clinically-validated treatment modality, and support developing therapies to target pathways that are not VEGF-driven for the treatment of RVO.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

CT and MA contributed to the conception and design of the study. MA, CC, and JL performed the experimental procedures and data analysis. SS and GS generated XBir3 protein used in the study. CT, MA, and CC contributed to the writing, revision, and figures. CT acquired the funding. All authors contributed to the article and approved the submitted version.

This work was supported by a Sponsored Research Agreement with Opera Therapeutics to CT, by the National Institute of Neurological Disorders and Stroke (RO1 NS081333 to CT) and the Department of Defense Army/Air Force (DURIP to CT) and by the National Eye Institute (T32 EY013933 to MA).

CT and MA are listed as inventors on patent applications filed by the Trustees of Columbia University in the City of New York related to the therapeutic use of caspase-9 inhibitors. MA received consulting income from Opera Therapeutics.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.