(639g) Fluorescence Lifetime Sorting with A Frequency Domain Analysing Flow Cytometer

The realm of chemical engineering is expanding rapidly and has incorporated into it principles of bio-engineering and biological techniques. One such method is flow cytometry and is used in research and diagnostic medicine to deduce information about cells or microparticles based upon their relative size, internal complexity and fluorescence character. This is determined by using hydrodynamic properties to focus a flowing stream of a sample of interest into a single flowline and individually interrogate across a Gaussian profile laser beam. From the scattered and fluoresced light we can determine things about the cells such as its place in the cell cycle, protein expression levels and surface marker characteristics. The “killer app” for flow cytometry came in the 1980’s when CD4+T-cell level determination led to a rapid and accurate determination of when HIV infection became full blown AIDS.

These measurements of fluoresced photons by a photodetector have served the cytometry community well in the past yet suffer from a very critical flaw. Fluorophores emit light over a broad spectral range and experimenters will often encounter spectral overlap from two emitting fluorescent species. When spectral overlap between fluorophores emitting in regions that exhibit cellular autofluorescence occurs there is a compounding issue whereby the exogenous signal cannot be separated from the endogenous fluorescence. Although mathematical compensation methods exist, they are only numerical subtractions and not true measurements.

In previous word, we have shown that the phase shift of each fluorophore can be measured to overcome the issues of spectral overlap by capturing the fluorescent molecule’s excited state lifetime. Our methods digitally capture amplified voltage waveforms from a cytometer’s detector. A Fourier transform and digital processing is done to extract the excited state lifetimes of fluorophores on flowing cells. In this contribution we present results for implementing the digital methods with sorting of cells and fluorescence microspheres

Material & Methods

The testbed for this experiment was a modified commercial sorting flow cytometer. We replaced the stock laser lines with solid state lasers that can either be directly modulated or modulated using an electro-optic modulator to impart periodic wave characteristics into the beam. This allows for single and multiple harmonic information to be imparted into the signal for Fourier analysis. Signals are routed from the cytometer photodetectors through high frequency amplifiers and into a specialized data acquisition system. It is within this system that standard cytometric data is recorded and the Fourier transform on the signals is done to extract out the phase information at signal harmonics of interest. The recorded phase, at a particular frequency, is proportional to the excited state fluorescence lifetime of a fluorophore.

The phase shift is a measure of how much a periodic signal lags to a reference. The degree of lag of an excited state fluorescence process only equates to a measure in the tens of nanoseconds, and a high frequency signal is needed in order to determine a phase shift that can be discerned by sampling equipment.

Our investigation is of a set of two fluorescent microspheres whose lifetimes have previously been individually measured on our system. By using modulation levels in the megahertz range we are able to quantify and distinguish between the excited state lifetimes of the two microspheres and set up gating regions to send back to the sorting cytometer. This instrument is then able to utilize an electro-static deflection system to deflect droplets created by a piezo-electric resonator from the flowing stream of the cytometer, into specific bins. These samples are then tested for sort purity by analysis on a separate dedicated analysis cytometer.

Results

With these approaches we demonstrate that using frequency domain techniques can lead to the ability to distinguish different fluorophores based solely on their excited state fluorescence lifetime. Additionally we show that it is possible to not only distinguish in software, but to redirect signals back to a sorting cytometer and be able to sort microparticles based upon fluorescence lifetime characteristics. Ultimately these techniques can be implemented in commercial cytometry and expand the functionality of instruments by allowing for intensity and temporal characteristics to be used to distinguish between fluorophores in cellular systems to aid in research and diagnostics.