Plasma neuronal exosomes serve as biomarkers of cognitive impairment in HIV infection and Alzheimer's disease

Abstract

Fluid biomarkers for cognitive impairment have the advantage of being relatively noninvasive and capable of monitoring neuronal and other brain cell health in real time. Biomarkers can predict the onset of dementing illness, but also correlate with cognition in a dynamic way allowing us to follow treatment responses and determine brain recovery. Chronic HIV infection causes cognitive impairment in a subset of individuals suggesting "premature aging." Exosomes are small extracellular vesicles that are shed from all cells. They are important in normal cell-to-cell communication as they contain cellular proteins, mRNA transcripts, and miRNAs. Exosome cargo varies depending on the health of the cell and pathological state; specific proteins/mRNAs and/or miRNAs are present and may serve as biomarkers. Exosomes of variable cellular origin can be isolated from peripheral blood by various methods. Neuron-derived exosomes (NDEs) can be isolated using a precipitation/immunoaffinity approach using antibodies against neuronal cell adhesion molecule L1CAM and the contents queried for central nervous system (CNS) disorders including HIV-associated neurological disorders (HAND) and Alzheimer's disease (AD). As these studies are recent, numerous questions arise including which neuronal proteins are in NDEs and whether their contents differ in different CNS pathologies or with age. In addition, can the NDE cargo predict as well as diagnose cognitive impairment and could exosomal contents be used as therapeutic biomarkers, or theramarkers, of neuronal recovery from effective treatment? This mini-review will show some new data and review recent studies on NDE from individuals with HIV infection and AD. HIV-associated neurocognitive disorders (HAND) are pathologies seen in a subset of individuals with chronic HIV infection. They belong to the spectrum of neurodegenerative diseases that result in death or dysfunction of neurons with similarities to Alzheimer disease (AD) but also distinctive differences (reviewed (Canet et al., Front Cell Neurosci 12: 307, 2018)). Both disorders are difficult to diagnose without neuropsychological testing and both need new biomarkers to judge progression as well as recovery with treatment. Both disorders involve neuroinflammation and several common targets. AD is associated with aging and HIV is thought to initiate premature aging. In HIV infection, amyloid beta (Aβ), which is deposited in "plaques" in AD, is soluble and its relevance to HIV-associated cognitive impairment is controversial (Achim et al., J Neuroimmune Pharmacol 4: 190-199, 2009; Rempel and Pulliam, AIDS 19: 127-135, 2005). Aβ deposition is required for AD pathological diagnosis, but is not necessarily causative (Barage and Sonawane, Neuropeptides 52: 1-18, 2015; Hardy and Selkoe, Science 297: 353-356, 2002; Morris et al., Acta Neuropathol Commun 2: 135, 2014). Neurofilament light (NF-L) is a surrogate marker in plasma and cerebrospinal fluid (CSF) for neurodegeneration (Abu-Rumeileh et al., Alzheimers Res Ther 10: 3, 2018; Mattsson et al., JAMA Neurol 74: 557-566, 2017) but continues to be a controversial biomarker for both HAND and AD (Gisslen et al., EBioMedicine 3: 135-140, 2016; Kovacs et al., Eur J Neurol 24:1326-e77, 2017; Norgren et al., Brain Res 987: 25-31, 2003; Rolstad et al., J Alzheimers Dis 45: 873-881, 2015; Yilmaz et al., Expert Rev Mol Diagn 17: 761-770, 2017). Blood biomarkers are needed to advance both HAND and AD fields, as blood draws are less costly than neuroimaging and are minimally invasive compared to lumbar punctures required for CSF acquisition. Extracellular vesicles (EVs) are nanoscale membranous vesicles shed from all cells including those of the central nervous system (CNS) and found in all biofluids; they are divided into exosomes (30-150 nm) originating from late endosomes/multivesicular bodies and microvesicles (150-1000 nm) produced through budding of the plasma membrane. Both types of vesicles are implicated in the pathogenesis of neurodegenerative diseases and may provide biomarkers (Bellingham et al., Front Physiol 3: 124, 2012). In this report, we call the vesicles exosomes, since they are the predominant vesicles in our preparations. They are involved in cell-to-cell communication in normal homeostasis and can be carriers of toxic proteins (Aβ, tau) (Sardar Sinha et al., Acta Neuropathol 136: 41-56, 2018) shed by cells as waste or actively secreted in a degenerative process (review Gupta and Pulliam, J Neuroinflammation 11: 68, 2014). The idea that exosomes originating from a specific cell can be recovered in the plasma using cellular surface markers of interest is intriguing. Neuron derived exosomes (NDEs) were first described in 2015 and isolated using antibodies against neural cell adhesion molecules NCAM or L1CAM, after total plasma exosome isolation (Fiandaca et al., Alzheimers Dement 11: 600-607 e1, 2015). Characterization of NDEs follows guidelines endorsed by the International Society for Extracellular Vesicles and includes Nanoparticle Tracking Analysis (NTA) to determine EV concentration and average diameter; Western Blots for EV markers; ELISAs for neuronal proteins and transmission EM for visualization (Sun et al., AIDS 31: F9-F17, 2017; Tang et al., FASEB J 30: 3097-106, 2016). This innovative isolation of an exosome sub-population has generated interest in using NDE as biomarkers for neurodegenerative diseases like AD, HAND, traumatic brain injury, posttraumatic stress disorder and more (reviews Agoston et al., Brain Inj 31: 1195-1203, 2017; Gupta and Pulliam, J Neuroinflammation 11: 68, 2014; Hu et al., Cell Death Dis 7: e2481, 2016; Karnati et al., J Neurotrauma, 2018; Osier et al., Mol Neurobiol, 2018). Several biomarkers from plasma NDEs were recently reported by the Pulliam lab to be elevated in general cognitive impairment (Sun et al., AIDS 31: F9-F17, 2017). We review our collective data here on HAND and AD and add to the characterization of plasma NDEs as exciting biomarkers of neurodegeneration.

Keywords: Alzheimer’s; Biomarker; Dementia; Exosomes; HIV; Neurons.

Figure 1:. Protein expression and correlation of NDE biomarkers

from control (N=12) and HIV-infected individuals (N=23). (A) Tetraspanins CD63, CD81 and ALIX measured by ELISA in NDE. NDE from controls (C) and HIV-infected subjects (HIV) significantly differed in CD63 and CD81 levels. ALIX levels were not statistically different from NDE from controls and HIV infection. (B) Levels of proteins associated with neuronal dysfunction were assayed by ELISA from control subjects who were neuropsychologically normal (NPN, N=21) and neuropsychologically impaired (NPI, N=14) subjects that included 11 HIV-infected and 3 uninfected control subjects (Sun et al, 2017). HMGB1, high mobility group box1; NF-L, neurofilament-light; Aβ, amyloid β. (Adapted from (Sun et al, 2017)) (C) Aging and a decrease in NDE targets of HIV-infected subjects. Aging correlated with significant decreases in HMGB1 and NF-L in NDE from HIV-infected individuals. No effect was observed in NDE from controls. HIV-infected subjects had a wider age range (35–67 years) than the controls (51–65). Spearman correlation was used. (Adapted from (Sun et al, 2017))

Figure 2:. Protein targets in NDEs as biomarkers for AD.

Potential protein biomarkers are grouped by their relevance to different aspects of AD pathogenesis: Aβ Pathology and Synaptic changes (since Aβ directly induces synaptic pathology); Glucose Hypometabolism / Insulin Resistance (with a focus in insulin and downstream signaling cascades); and Tau Pathology / Neurodegeneration (including deficient neurotrophic factors). These targets were identified in referenced discovery studies by Kapogiannis, Goetzl, et al. and may be increased or decreased in AD patients compared to controls.

Figure 3:. Heatmap of differential protein expression from EVs.

Heatmap showing log-transformed expression results (NPX values) for proteins measured in three distinct exosome populations (total, immunoprecipitated by anti-L1CAM and immunoprecipitated by anti-IgG2a) with the Olink Biomarker Discovery Neurology panel. Only proteins above the Lowest Limit of Detection are included. To help interpret results the two columns on the right indicate whether these proteins have previously been shown to be present or enriched in neurons (red - previously shown to be enriched in neurons compared to other tissues; green – previously shown to be expressed in neurons, but also in other cells; and blue – not previously shown to be expressed in neurons) and brain (red - previously shown to be enriched in brain tissue; blue - not previously shown to be expressed in brain tissue) according to Human Protein Atlas (https://www.proteinatlas.org/)and PubMed search results.