About

Derek S. Welsbie is an Associate Professor in the Department of Ophthalmology at UC San Diego. His research activities and funding focus on kinase multitargeting for glaucoma neuroprotection, retinal ganglion cell survival, and optic neuropathy. He has been involved as a Principal Investigator in projects supported by NIH grants, including studies on mitochondrial complex I deficiency models and retinal neuron survival. His work encompasses the development of novel therapeutic strategies for optic nerve and retinal diseases, including neuroprotection, axon regeneration, and retinal cell reprogramming. Welsbie's research integrates high-content high-throughput functional genomics, deep learning-based clinical decision support tools, and multi-omic analyses to advance understanding and treatment of neurodegenerative eye conditions.

Research topics

• Biology
• Medicine
• Machine Learning
• Computer Science
• Surgery
• Artificial Intelligence
• Chemistry
• Neuroscience
• Cell biology
• Genetics
• Pathology
• Materials science
• Biophysics
• Optics
• Immunology
• Bioinformatics
• Ophthalmology
• Virology
• Physics
• Nanotechnology
• Biochemistry
• Computational biology

Selected publications

• Serial analysis of macular and circumpapillary structures in glaucomatous eyes with peripapillary retinoschisis

  Canadian Journal of Ophthalmology · 2026-04-01

  article
  PublisherDOI

• Author response: Afadin sorts different retinal neuron types into accurate cellular layers

  2026-01-14

  peer-reviewOpen access
  PublisherDOI

• Optineurin-facilitated axonal mitochondria delivery promotes neuroprotection and axon regeneration

  Nature Communications · 2025-02-20 · 19 citations

  articleOpen access

  Optineurin (OPTN) mutations are linked to amyotrophic lateral sclerosis (ALS) and normal tension glaucoma (NTG), but a relevant animal model is lacking, and the molecular mechanisms underlying neurodegeneration are unknown. We find that OPTN C-terminus truncation (OPTN∆C) causes late-onset neurodegeneration of retinal ganglion cells (RGCs), optic nerve (ON), and spinal cord motor neurons, preceded by a decrease of axonal mitochondria in mice. We discover that OPTN directly interacts with both microtubules and the mitochondrial transport complex TRAK1/KIF5B, stabilizing them for proper anterograde axonal mitochondrial transport, in a C-terminus dependent manner. Furthermore, overexpressing OPTN/TRAK1/KIF5B prevents not only OPTN truncation-induced, but also ocular hypertension-induced neurodegeneration, and promotes robust ON regeneration. Therefore, in addition to generating animal models for NTG and ALS, our results establish OPTN as a facilitator of the microtubule-dependent mitochondrial transport necessary for adequate axonal mitochondria delivery, and its loss as the likely molecular mechanism of neurodegeneration. Optineurin is linked to ALS and glaucoma, but the mechanism is unknown. Here, the authors show that optineurin mutation causes deficits in axonal mitochondria transport and neurodegeneration, enhancing of which achieves axon protection and regeneration.

  PublisherOA PDFDOI

• Incidence of glaucoma in patients with systemic lupus erythematosus: a nationwide cohort study in South Korea

  British Journal of Ophthalmology · 2025-11-23

  article

  BACKGROUND/AIMS: To assess the incidence of glaucoma in patients with systemic lupus erythematosus (SLE) and to identify associated risk factors using a nationwide population-based cohort. METHODS: This retrospective cohort study analysed data from the Korean National Health Insurance Service (2008-2022). A total of 9682 patients with SLE were identified using ICD-10 code M32 and rare intractable disease code V136 and matched 1:1 to non-SLE controls by age and sex. The incidence of glaucoma (ICD-10 codes H40 and H42) and glaucoma suspect (H40.0) was assessed. Multivariate logistic regression was used to identify risk factors for glaucoma, including long-term corticosteroid use (≥180 days). RESULTS: The incidence of glaucoma was significantly higher in the SLE group than in controls (11.34% vs 3.74%; p<0.0001), with a greater prevalence in younger patients (<40 years) and females. Glaucoma suspect cases were also more common in the SLE group (35.56% vs 30.25%; p<0.0001). SLE was independently associated with glaucoma (adjusted OR: 2.00, 95% CI 1.69 to 2.38), and prolonged corticosteroid use further increased the risk (OR: 1.75, 95% CI 1.51 to 2.02). Annual incidence trends showed a rising pattern over time, especially among SLE patients. CONCLUSIONS: SLE is associated with an increased risk of glaucoma, particularly among younger individuals and females. Prolonged corticosteroid therapy significantly contributes to this risk. These findings support the need for regular ophthalmic screening and judicious corticosteroid management in patients with SLE.

  PublisherDOI

• Author response: Afadin Sorts Different Retinal Neuron Types into Accurate Cellular Layers

  2025-02-21

  peer-reviewOpen access

  Neurons use cell-adhesion molecules (CAMs) to interact with other neurons and the extracellular environment: the combination of CAMs specifies migration patterns, neuronal morphologies, and synaptic connections across diverse neuron types. Yet little is known regarding the intracellular signaling cascade mediating the CAM recognitions at the cell surface across different neuron types. In this study, we investigated the neural developmental role of Afadin–, a cytosolic adapter protein that connects multiple CAM families to intracellular F-actin. We introduced the conditional Afadin mutant to an embryonic retinal Cre, Six3-Cre–. We reported that the mutants lead to the scrambled retinal neuron distribution, including Bipolar Cells (BCs), Amacrine Cells (ACs), and retinal ganglion cells (RGCs), across three cellular layers of the retina. This scrambled pattern was first reported here at neuron-type resolution. Importantly, the mutants do not display deficits for BCs, ACs, or RGCs in terms of neural fate specifications or survival. Additionally, the displayed RGC types still maintain synaptic partners with putative AC types, indicating that other molecular determinants instruct synaptic choices independent of Afadin. Lastly, there is a significant decline in visual function and mis-targeting of RGC axons to incorrect zones of the superior colliculus, one of the major retinorecipient areas. Collectively, our study uncovers a unique cellular role of Afadin in sorting retinal neuron types into proper cellular layers as the structural basis for orderly visual processing.

  PublisherDOI

• Afadin sorts different retinal neuron types into accurate cellular layers

  eLife · 2025-02-21 · 1 citations

  preprintOpen access

  Neurons use cell-adhesion molecules (CAMs) to interact with other neurons and the extracellular environment: the combination of CAMs specifies migration patterns, neuronal morphologies, and synaptic connections across diverse neuron types. Yet little is known regarding the intracellular signaling cascade mediating the CAM recognitions at the cell surface across different neuron types. Using mouse genetics and viral labeling, we investigated the neural developmental role of Afadin (Mandai et al., 1997; Takai and Nakanishi, 2003; Takahashi et al., 1999), a cytosolic adapter protein that connects multiple CAM families to intracellular F-actin. We introduced the conditional Afadin mouse mutant (Beaudoin et al., 2012) to an embryonic retinal Cre, Six3 Cre (Oliver et al., 1995; Liu and Cvekl, 2017; Diacou et al., 2018). We reported that the mouse mutants lead to the scrambled retinal neuron distribution, including bipolar cells (BCs), amacrine cells (ACs), and retinal ganglion cells (RGCs), across three cellular layers of the retina. This scrambled pattern was first reported here at neuron-type resolution. Importantly, the mutants do not display deficits for BCs, ACs, or RGCs in terms of neural fate specifications or survival. Additionally, the displayed RGC types still maintain synaptic partners with putative AC types, indicating that other molecular determinants instruct synaptic choices independent of Afadin. Lastly, there is a significant decline in visual function and mis-targeting of RGC axons to incorrect zones of the superior colliculus, one of the major retinorecipient areas. Collectively, our study uncovers a unique cellular role of Afadin in sorting retinal neuron types into proper cellular layers as the structural basis for orderly visual processing.

  PublisherOA PDFDOI

• Cell-Intrinsic Vulnerability and Immune Activation Cooperate to Drive Degeneration in a Mitochondrial Complex I Deficiency Model of Optic Neuropathy

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-10-09

  preprintOpen access

  Abstract Mitochondrial dysfunction is a central hallmark of many optic neuropathies, yet the mechanisms linking intrinsic metabolic stress to retinal ganglion cell (RGC) degeneration remain unclear. To bridge this gap, we developed conditional transgenic models targeting the mitochondrial complex I subunit Ndufs4 in the retina. Broad deletion of Ndufs4 in the retina resulted in vision loss, progressive RGC degeneration, and pronounced immune activation before overt RGC death. Strikingly, depletion of myeloid cells significantly preserved RGCs, demonstrating that inflammation is not simply a downstream consequence but a participant in the degeneration process. To further distinguish between intrinsic and extrinsic mechanisms, we generated a mosaic model in which only subsets of retinal cells lacked Ndufs4 . In this paradigm, the degeneration first appeared selectively in mutant regions, suggesting that mitochondrial impairment within RGCs is necessary to initiate vulnerability. At later stages, however, the degeneration extended beyond mutant territories, highly suggestive of a propagation through non-cell autonomous processes. Together, these findings support a model in which mitochondrial dysfunction creates the conditions for neuronal vulnerability, while immune responses govern the timing and extent of cell loss. This framework explains the consistent co-occurrence of metabolic deficits and neuroinflammation in optic neuropathies and highlights the importance of their interactions in disease progression. By clarifying the intersection of intrinsic and extrinsic mechanisms, this work advances our understanding of RGC degeneration and provides a conceptual basis for deciphering pathogenic processes across diverse optic neuropathies.

  PublisherOA PDFDOI

• Modeling neurodegeneration in the retina and strategies for developing pan-neurodegenerative therapies

  Molecular Neurodegeneration · 2025-10-14 · 3 citations

  reviewOpen access

  BACKGROUND: Glaucoma Research Foundation's third Catalyst for a Cure team (CFC3) was established in 2019 to uncover new therapies for glaucoma, a leading cause of blindness. In the 2021 meeting "Solving Neurodegeneration," (detailed in Mol Neurodegeneration 17(1), 2022) the team examined the failures of investigational monotherapies, issues with translatability, and other significant challenges faced when working with neurodegenerative disease models. They emphasized the need for novel, humanized models and proposed identifying commonalities across neurodegenerative diseases to support the creation of pan-neurodegenerative disease therapies. Since then, the fourth Catalyst for a Cure team (CFC4) was formed to explore commonalities between glaucoma and other neurodegenerative diseases. This review summarizes outcomes from the 2023 "Solving Neurodegeneration 2" meeting, a forum for CFC3 and CFC4 to share updates, problem solve, plan future research collaborations, and identify areas of unmet need or opportunity in glaucoma and the broader field of neurodegenerative disease research. MAIN BODY: We summarize the recent progress in the field of neurodegenerative disease research and present the newest challenges and opportunities moving forward. While translatability and disease complexity continue to pose major challenges, important progress has been made in identifying neuroprotective targets and understanding neuron-glia-vascular cell interactions. New challenges involve improving our understanding of the disease microenvironment and timeline, identifying the optimal approach(es) to neuronal replacement, and finding the best drug combinations and synergies for neuroprotection. We propose solutions to common research questions, provide prescriptive recommendations for future studies, and detail methodologies, strategies, and approaches for addressing major challenges at the forefront of neurodegenerative disease research. CONCLUSIONS: This review is intended to serve as a research framework, offering recommendations and approaches to validating neuroprotective targets, investigating rare cell types, performing cell-specific functional characterizations, leveraging novel adaptations of scRNAseq, and performing single-cell sorting and sequencing across neurodegenerative diseases and disease models. We focus on modeling neurodegeneration using glaucoma and other neurodegenerative pathologies to investigate the temporal and spatial dynamics of neurodegenerative disease pathogenesis, suggesting researchers aim to identify pan-neurodegenerative drug targets and drug combinations leverageable across neurodegenerative diseases.

  PublisherOA PDFDOI

• Recovery of retinal terminal fields after traumatic brain injury: evidence of collateral sprouting and sexual dimorphism

  bioRxiv (Cold Spring Harbor Laboratory) · 2025-06-01

  preprintOpen access

  Abstract The central nervous system is characterized by its limited regenerative potential, yet striking examples of functional recovery after injury in animal models and humans highlight its capacity for repair. Little is known about repair of pathways/circuits after traumatic brain injury (TBI), which results in disruption of connectivity. Here we utilize a mouse model of diffuse traumatic axonal injury (Impact-acceleration TBI) in order to explore, for the first time, the evolution of structural and functional changes in the terminal fields of the injured visual system. Retinal ganglion cell (RGC) axons and synapses were genetically labeled via AAV transduction, while anterograde and transsynaptic tracers were used to mark terminals and postsynaptic cells. Functional connectivity and visual integrity were assessed by monitoring c-Fos expression following light stimulation and pattern-reversal visual evoked potentials (pVEPs). Our findings demonstrate that, although TAI results in approximately a 50% loss of RGC axons and terminals, surviving RGCs undergo collateral sprouting, a form of compensatory branching of surviving axons, that restores terminal density to pre-injury levels. Transsynaptic tracing and c-Fos mapping confirmed the reestablishment of connectivity, which was also associated with significant improvements in visual function as measured by pVEPs. Interestingly, the recovery process exhibited sexual dimorphism, with female mice showing delayed or incomplete repair. Moreover, collateral sprouting proceeded normally in Sarm1 knockout mice, evidence of some independence from Wallerian degeneration. Our findings show that collateral sprouting may be an important mechanism of circuit repair in TAI and may represent a promising target for therapeutic interventions. Significance Homotypic collateral sprouting -the process by which uninjured axons from the same neuronal source extend new branches to reinnervate targets deprived of their original connections- is a fundamental yet understudied mechanism for CNS repair following injury. Unlike heterotypic sprouting, involving sprouting from unrelated pathways, homotypic sprouting offers potential to restore circuit architecture after partial lesions. Here, we employed a model of diffuse axonal injury in the mouse visual system to examine this mechanism. Our research demonstrates surviving retinal ganglion cell axons can re-establish terminal fields, achieving structural and functional connectivity. Importantly, we discovered significant sex differences: female mice showed delayed/incomplete recovery compared to males. These findings provide evidence of repair of brain circuits perturbed by TBI and the role of homotypic sprouting.

  PublisherOA PDFDOI

• Recovery of Retinal Terminal Fields after Traumatic Brain Injury: Evidence of Collateral Sprouting and Sexual Dimorphism

  Journal of Neuroscience · 2025-12-15

  articleOpen access

  The central nervous system is characterized by its limited regenerative potential, yet striking examples of functional recovery after injury in animal models and humans highlight its capacity for repair. Little is known about repair of pathways/circuits after traumatic brain injury (TBI), which results in disruption of connectivity. Here we utilize a mouse model of diffuse traumatic axonal injury (impact-acceleration TBI) in order to explore, for the first time, the evolution of structural and functional changes in the terminal fields of the injured visual system. Retinal ganglion cell (RGC) axons and synapses were genetically labeled via AAV transduction, while anterograde and transsynaptic tracers were used to mark terminals and postsynaptic cells. Functional connectivity and visual integrity were assessed by monitoring c-Fos expression following light stimulation and pattern-reversal visual evoked potentials (pVEPs). Our findings demonstrate that, although TAI results in an ∼50% loss of RGC axons and terminals, surviving RGCs undergo collateral sprouting, a form of compensatory branching of surviving axons, that restores terminal density to preinjury levels. Transsynaptic tracing and c-Fos mapping confirmed the reestablishment of connectivity, which was also associated with significant improvements in visual function as measured by pVEPs. Interestingly, the recovery process exhibited sexual dimorphism, with female mice showing delayed or incomplete repair. Moreover, collateral sprouting proceeded normally in Sarm1 knock-out mice, evidence of some independence from Wallerian degeneration. Our findings show that collateral sprouting may be an important mechanism of circuit repair in TAI and may represent a promising target for therapeutic interventions.

  PublisherOA PDFDOI

Recent grants

• Kinase Multitargeting for Glaucoma Neuroprotection

  NIH · $3.2M · 2024–2028

• Vision Biostatistics

  NIH · $17.7M · 2023–2028

Frequent coauthors

• Donald J. Zack

  Johns Hopkins University

  75 shared

• Katherine L. Mitchell

  Johns Hopkins University

  46 shared

• Amit K. Patel

  University of California, San Diego

  44 shared

• Cynthia Berlinicke

  Johns Hopkins University

  30 shared

• Zhiyong Yang

  30 shared

• Robert N. Weinreb

  University of California, San Diego

  26 shared

• Valentin M. Sluch

  Novartis (United States)

  23 shared

• Xitiz Chamling

  Johns Hopkins Medicine

  23 shared

Labs

• Derek Welsbie | UCSD ProfilesPI

Education

• Ph.D., Ophthalmology

  University of California, San Diego

  2009

• M.D., Medicine

  University of California, San Diego

  2005

• B.A., Biology

  University of California, San Diego

  2001