In vivo characterization of a retinal cellular biomarker of inflammation in multiple sclerosis

Neuroinflammation plays a central role in the progression of multiple sclerosis (MS). While neuroinflammation contributes to neuronal damage, it also has regenerative roles, highlighting the complex relationship between immune responses and neurodegeneration. However, capturing the cellular processes involved in inflammation and regeneration in vivo remains a challenge. The retina, a part of the CNS tissue, provides a unique, non-invasive window into these processes. High-resolution retinal imaging techniques, such as adaptive optics scanning laser ophthalmoscopy (AOSLO), enable the observation of cellular dynamics in the retina, advancing our understanding of neuroinflammation in MS. In this prospective cohort study, we used a custom-modified AOSLO to examine the retinal ganglion cell layer in patients with MS (pwMS) (n = 51) with varying phenotypes: (i) relapsing-remitting MS (RMS) with recent optic neuritis (ON) (<6months) (ON-RMS, N = 31), (ii) RMS without recent ON (N = 12), and (iii) progressive MS (PMS, N = 8). Healthy controls without ocular pathology (N = 9) were recruited. Longitudinal imaging was performed using an AOSLO at 3, 6, and 12 months for some patients. We focussed on identifying immune cells in the inner retina, particularly during and after acute ON, and assessed these findings in relation to clinical symptoms and other imaging modalities, including optical coherence tomography and MRI. Cellular infiltrates were detected most frequently in ON-RMS patients (93.6%) compared to the other groups. Cell density was higher (P < 0.05) in the affected eye of pwMS after recent ON and declined over time, though with individual variability. In one case, cells appeared 7 weeks before clinical ON symptoms. Most cells had lymphocyte-like morphology with a round shape, compact cytoplasm, and a nucleus-like structure, and some contained multiple intracellular structures. Their mean size was 12.0 µm (5.8-35.5 µm) and they were primarily located in the ganglion cell layer, but absent from the outer retina. Cells were frequently found near capillaries and larger vessels, often interacting with them, with some partially within vessels, suggesting migration or infiltration. However, movement was minimal, with most remaining stationary for minutes to an hour. Cell velocity was estimated at 3.2 ± 1 µm/h. In ON-RMS patients, cell density was not associated with optic nerve lesion location or length. Our results suggest that AOSLO imaging can detect subtle retinal inflammatory changes not captured by conventional clinical systems, offering a promising tool for monitoring neuroinflammation in MS and other neurodegenerative diseases. These findings support the potential of high-resolution retinal imaging as a non-invasive biomarker for tracking neuroinflammation and therapeutic responses in pwMS.