Characterization of Brain Metabolic State under Injury using Two-Photon Microscopy

Background

Traumatic brain injury is the leading cause of death among individuals under the age of 45 in the US, with an incidence of 1.5 million each year. Beyond fatality, TBI results in severe long-term disabilities, both mentally and physically. Pathophysiologically, Traumatic brain injury can be divided into 2 phases: a primary mechanical impact on the brain followed by secondary biochemical and inflammatory cascades of different types of brain cells. The two major biochemical cascades that we plan to characterize are oxidative stress and glutamate excitotoxicity. Following the injury, an influx of excess calcium ions into the mitochondria triggers the production of reactive oxygen species (ROS) and free radicals. These molecules depolarize the mitochondrial inner membrane, disrupting the electron transport chain and inhibiting the oxidative phosphorylation process. This deprives the nerve cells of ATP and facilitates apoptosis. In junction with oxidative stress, glutamate and aspartate neurotransmitters accumulate at the synapses as the impaired glutamate transporters fail to recycle excess glutamate from injured neurons. These molecules bind to NMDA and AMPA receptors that promote calcium, potassium, and sodium uptake. Cell depolarization triggers downstream cascades that prolong the effect of oxidative stress. From these observations, it can be said that the biochemical pathways involved in secondary injury are highly complex and intertwined. Despite ongoing research, the understanding of its mechanisms and consequences remains incomplete. Notably, secondary injury can develop over a long period of time, increasing the severity of the original injury. In other words, patients with mild TBI can suffer neurological problems and long-term disability months after the injury. For this reason, patients with mild TBI have no initial symptoms and, therefore, are often undiagnosed, preventing early treatment.

The long-term objective is to develop  a technique that can diagnose TBI on a molecular level, which is the biochemical cascade of secondary injury.

Shortcomings of Previous Work

To examine the long-term cellular effects of mild TBI, the 3D-engineered brain tissues (EBT) of neurons and glial cells (astrocytes and microglia) are injured using the controlled cortical impactor (CCI), mimicking a mild blast TBI. The EBT model, while a simplified human brain, still undergoes most of the complex secondary response following the impact and, thus, can be used to develop diagnostic and treatment frameworks for TBI. However, with CCI, it is challenging to completely characterize this model due to the evolving complex cellular environment and unpredictable changes arising from the interactions between multiple cell types.

As a solution, we propose to examine TBI via its constituents by introducing a specific secondary injury to 2D brain cell cultures and study the cellular metabolic interactions and environments in a controlled manner.

About the Imaging Modality and Relevant Analysis

Two-photon excited fluorescence (TPEF) can be used to assess the functional and morphological changes of the injured brain cells by obtaining the metrics of cellular metabolic function. TPEF detects autofluorescent signals from several key biomolecules, namely FAD, NADH, LipDH, and lipofuscin. The former three are metabolic co-enzymes which are implicated in most metabolic perturbations, and lipofuscin is a complex of fluorescent proteins and lipids that accumulates under cellular stress conditions. These endogenous fluorophores can be analyzed using computational techniques that reveal underlying concentration-based and metabolic shifts in the samples: redox ratio, mitochondrial clustering, phasor analysis, and spectral deconvolution.

• Redox ratio is the relative ratio of glycolytic to oxidative metabolism. In brief, it is computed by obtaining a “NADH image” (755ex/460em) and a “FAD image” (860ex/525em) and dividing them according to the formula (NADH/(NADH+FAD)).
• Mitochondrial clustering is the extent of mitochondrial fractionation, which occurs in response to ROS accumulation. It is computed by segmenting and cloning mitochondrial regions in an image. Then, the power spectral density of the cloned image is computed, which determines the image frequency. Highly fractionated mitochondria will have a high frequency, and vice versa.
• Phasor analysis is a technique to obtain fit-free visualizations of FLIM images with overlapping concentrations of lifetimes over different pixels. In brief, time-series fluorescence lifetime data is sine and cosine transformed, giving two coordinates g and s that correspond to the lifetime, tau, of the fluorescent decay. Any one tau localizes on a circular plot (see Fig. 2). The localization of the (g, s) coordinate pair for any given pixel is determined by the linear combination of different tau values constituent in the pixel. A fluorophore’s binding environment affects its lifetime, but its concentration does not (i.e. higher concentrations of a single fluorophore simply cause a shift in the phasor distribution towards that fluorophore’s lifetime). As such, by assessing the overall phasor distribution, conditions such as shifts in relative concentrations of fluorophores and shifts in fluorophore binding configuration can be observed.
• Spectral constituents are obtained from the overall spectral intensity curve via non-negative matrix factorization. In this method of spectral deconvolution, the user specifies the number of total constituents and the model computes optimal concentrations of non-negative vectors and weights that minimize the error (residual) from the overall spectrum. In this way, concentrations and emission spectra of constituent fluorophores are determined.

Compared to state-of-the-art diagnosis procedures like MRI, TPEF is more sensitive to cellular-level metabolic shifts.

However, TPEF fails to detect non-fluorescence metabolites such as lactate. Additionally, while it is known that an increase in redox ratio correlates to an increase in glycolytic metabolism and vice versa, conducting redox ratio studies in conjunction with exact biochemical measurements will allow us to quantify how shifts in oxidative and glycolytic metabolism affect our optical readouts. This lack of specificity is a critical roadblock to using TPEF to study injured brain metabolism.

The specific goal is to characterize and map optical metrics to specific altered metabolic pathways predicted by a metabolic computational model.

The central hypothesis is that, to identify the pathways, relevant biochemical metrics from assays and mass spectrometry will be incorporated into the computational model. 

In general, TPEF is a commonly used neuroimaging technique due to its high-depth penetration and potential for metabolic sensitivity, so many groups researching the impact of TBI or other neurodegenerative diseases choose to use two-photon imaging. Additionally, biological assay and mass spectrometric methods are well-investigated in the context of TBI.

The novelty of this project lies in correlating a non-invasive, label-free method (TPEF) with these invasive methods for the eventual use of optical methods alone for a diagnostic TBI model.

A non-invasive, label-free platform for the assessment of TBI does not exist to our knowledge. As such, the novelty of this study depends on the identification of TBI biomarkers, not just the development of a two-photon platform to study brain injury. In general, we propose to formulate a relationship between output molecular concentrations from assays and optical readouts via a computational model. We hypothesize that, under different perturbed or injured conditions, the trend in optical readouts and molecular concentration will be different since different metabolic mechanisms are involved. Therefore, we can say that a specific trend in optical readouts will be characteristic of a set of output concentrations and from the computational mode, specific altered metabolic pathways. This will allow us to characterize the optical readouts for a specific injury condition.

Tasks

Experiments