Combining In Vivo Corneal Confocal Microscopy With Deep Learning-Based Analysis Reveals Sensory Nerve Fiber Loss in Acute Simian Immunodeficiency Virus Infection

PURPOSE

To characterize corneal subbasal nerve plexus features of normal and simian immunodeficiency virus (SIV)-infected macaques by combining in vivo corneal confocal microscopy (IVCM) with automated assessments using deep learning-based methods customized for macaques.

METHODS

IVCM images were collected from both male and female age-matched rhesus and pigtailed macaques housed at the Johns Hopkins University breeding colony using the Heidelberg HRTIII with Rostock Corneal Module. We also obtained repeat IVCM images of 12 SIV-infected animals including preinfection and 10-day post-SIV infection time points. All IVCM images were analyzed using a deep convolutional neural network architecture developed specifically for macaque studies.

RESULTS

Deep learning-based segmentation of subbasal nerves in IVCM images from macaques demonstrated that corneal nerve fiber length and fractal dimension measurements did not differ between species, but pigtailed macaques had significantly higher baseline corneal nerve fiber tortuosity than rhesus macaques (P = 0.005). Neither sex nor age of macaques was associated with differences in any of the assessed corneal subbasal nerve parameters. In the SIV/macaque model of human immunodeficiency virus, acute SIV infection induced significant decreases in both corneal nerve fiber length and fractal dimension (P = 0.01 and P = 0.008, respectively).

CONCLUSIONS

The combination of IVCM and robust objective deep learning analysis is a powerful tool to track sensory nerve damage, enabling early detection of neuropathy. Adapting deep learning analyses to clinical corneal nerve assessments will improve monitoring of small sensory nerve fiber damage in numerous clinical settings including human immunodeficiency virus.

Differential regulation of TREM2 and CSF1R in CNS macrophages in an SIV/macaque model of HIV CNS disease

HIV-associated neuroinflammation is primarily driven by CNS macrophages including microglia. Regulation of these immune responses, however, remains to be characterized in detail. Using the SIV/macaque model of HIV, we evaluated CNS expression of triggering receptor expressed on myeloid cells 2 (TREM2) which is constitutively expressed by microglia and contributes to cell survival, proliferation, and differentiation. Loss-of-function mutations in TREM2 are recognized risk factors for neurodegenerative diseases including Alzheimer's disease (AD), amyotrophic lateral sclerosis (ALS), and Nasu-Hakola disease (NHD); recent reports have also indicated a role for TREM2 in HIV-associated neuroinflammation. Using in situ hybridization (ISH) and qRT-PCR, TREM2 mRNA levels were found to be significantly elevated in frontal cortex of macaques with SIV encephalitis compared with uninfected controls (P = 0.02). TREM2 protein levels were also elevated as measured by ELISA of frontal cortex tissue homogenates in these animals. Previously, we characterized the expression of CSF1R (colony-stimulating factor 1 receptor) in this model; the TREM2 and CSF1R promoters both contain a PU.1 binding site. While TREM2 and CSF1R mRNA levels in the frontal cortex were highly correlated (Spearman R = 0.79, P < 0.001), protein levels were not well correlated. In SIV-infected macaques released from ART to study viral rebound, neither TREM2 nor CSF1R mRNA increased with rebound viremia. However, CSF1R protein levels remained significantly elevated unlike TREM2 (P = 0.02). This differential expression suggests that TREM2 and CSF1R play unique, distinct roles in the pathogenesis of HIV CNS disease.

Deep learning-based analysis of macaque corneal sub-basal nerve fibers in confocal microscopy images

BACKGROUND

To develop and validate a deep learning-based approach to the fully-automated analysis of macaque corneal sub-basal nerves using in vivo confocal microscopy (IVCM).

METHODS

IVCM was used to collect 108 images from 35 macaques. 58 of the images from 22 macaques were used to evaluate different deep convolutional neural network (CNN) architectures for the automatic analysis of sub-basal nerves relative to manual tracings. The remaining images were used to independently assess correlations and inter-observer performance relative to three readers.

RESULTS

Correlation scores using the coefficient of determination between readers and the best CNN averaged 0.80. For inter-observer comparison, inter-correlation coefficients (ICCs) between the three expert readers and the automated approach were 0.75, 0.85 and 0.92. The ICC between all four observers was 0.84, the same as the average between the CNN and individual readers.

CONCLUSIONS

Deep learning-based segmentation of sub-basal nerves in IVCM images shows high to very high correlation to manual segmentations in macaque data and is indistinguishable across readers. As quantitative measurements of corneal sub-basal nerves are important biomarkers for disease screening and management, the reported work offers utility to a variety of research and clinical studies using IVCM.

Comparative Anatomy of the Mammalian Corneal Subbasal Nerve Plexus

Purpose

The subbasal nerve plexus (SNP) is the densest and most recognizable component of the mammalian corneal innervation; however, the anatomical configuration of the SNP in most animal models remains incompletely described. The purpose of the current study is to describe in detail the SNP architecture in eight different mammals, including several popular animal models used in cornea research.

Methods

Corneal nerves in mouse, rat, guinea pig, rabbit, dog, macaque, domestic pig, and cow eyes were stained immunohistochemically with antiserum directed against neurotubulin. SNP architecture was documented by digital photomicrography and large-scale reconstructions, that is, corneal nerve maps, using a drawing tube attached to a light microscope.

Results

Subbasal nerve fibers (SNFs) in mice, rats, guinea pigs, dogs, and macaques radiated centrally from the corneoscleral limbus toward the corneal apex in a whorl-like or spiraling pattern. SNFs in rabbit and bovine corneas swept horizontally across the ocular surface in a temporal-to-nasal direction and converged on the inferonasal limbus without forming a spiral. SNFs in the pig cornea radiated centrifugally in all directions, like a starburst, from a focal point located equidistant between the corneal apex and the superior pole.

Conclusions

The results of the present study have demonstrated for the first time substantial interspecies differences in the architectural organization of the mammalian SNP. The physiological significance of these different patterns and the mechanisms that regulate SNP pattern formation in the mammalian cornea remain incompletely understood and warrant additional investigation.