Principles of Particle Sorting 31 Electrostatic Sorting

Most analytical flow cytometers are enclosed in that cells are aspirated from a reservoir and hydrodynamically focused so that they pass one by one through a light source, generally from one or more lasers. At this point, scattered light and fluorescence signals are generated, detected, and measured. After this, cells are removed under vacuum to a waste reservoir. In general, flow sorters use a principle involving the electrostatic deflection of charged droplets similar to that used in ink-jet printers. To sort particles by this method, the process has to be performed in a more open system where cells are ejected into air in a stream of sheath fluid.

Any fluid stream ejected into air will break up into droplets but this is not a stable process; the distance from the orifice that the stream begins to break up will depend on many factors such as the orifice size, the pressure of the sheath fluid, the ambient temperature, and the viscosity of the fluid. However, if a stationary wave of vibration of known frequency and amplitude is applied to the fluid stream, it is possible to stabilize the break-off, and for a given set of conditions, the size of the droplets and the distance between drops will also stabilize. In a flow sorter, this vibration is produced by a transducer, which is generally a piezo-electric crystal acoustically coupled to the nozzle. As cells are ejected from a nozzle, they pass through one or more laser beams and at this point - the moment of analysis - information is gathered about the cell or particle (Figure 1). The distance, and therefore time, between the point of analysis and the point at which the cell breaks off from the solid stream in a droplet is constant and under given conditions can be calculated. This distance between the laser intercept and the break-off point is measured in drop equivalents and is often referred to as the drop delay. The calculation and monitoring of this is critical and is the factor that makes a sort successful or not. The drop break-off can be observed microscopically under stroboscopic illumination to allow the break-off point to be monitored. The drop delay is calculated by determining how many drops are in the distance between the analysis point and the break-off point; drops will start to form as soon as the stream emerges from the nozzle but will be coalesced until the break-off point. There are several ways of measuring the drop delay: by counting the number of drops formed in a known distance, by sorting beads onto slides at varying drop delays and checking microscopically, or by viewing fluorescent beads in sorted side streams. The precise way of calculating drop delay will vary with the type of sorter used.

Fig. 1. A diagram of a generalized cell sorter (A). Particles are introduced into a column of pressurized sheath fluid, and as they emerge from the nozzle, they pass through one or more laser beams. At this point, the moment of analysis, the cytometer gathers information about the fluorescence characteristics of the particle. After passing through the stream for the break-off distance, the stream is charged when the cell breaks off into a drop (moment of charging). There will be a variable number of satellite drops that are formed from the fluid connecting the drops as they form. These satellites should be "fast-merging" (i.e., quickly become coalesced with the preceding drops). Charged drops then pass through two high-voltage deflection plates and are deflected into collection vessels or aspirated to waste. The break-off point is seen in real time under stroboscopic illumination (B).

Fig. 1. A diagram of a generalized cell sorter (A). Particles are introduced into a column of pressurized sheath fluid, and as they emerge from the nozzle, they pass through one or more laser beams. At this point, the moment of analysis, the cytometer gathers information about the fluorescence characteristics of the particle. After passing through the stream for the break-off distance, the stream is charged when the cell breaks off into a drop (moment of charging). There will be a variable number of satellite drops that are formed from the fluid connecting the drops as they form. These satellites should be "fast-merging" (i.e., quickly become coalesced with the preceding drops). Charged drops then pass through two high-voltage deflection plates and are deflected into collection vessels or aspirated to waste. The break-off point is seen in real time under stroboscopic illumination (B).

Once the drop delay is calculated, it is possible to charge through the stream at the precise moment that the first drop is forming. Therefore, individual drops, as they break away from the solid stream, can be independently charged and will carry a positive charge, a negative charge, or will remain uncharged. The individual drops then pass through a static electrical field created by two charged plates. The voltage between the plates will be in the range of 2000 to 6000 V depending on the flow sorter used and the number of populations required from the sort. Charged drops are attracted to the plate of opposite polarity and will be deflected into collection vessels, which may be, for example, Eppendorf tubes, 6-mL centrifuge tubes, 15- or 50-mL conical tubes, or multiwell plates (anything up to 384-well plates). Initially, flow sorters were able to sort only two populations, one to the right and one to the left. However, the most recently introduced sorters (MoFlo [Dako, Fort Collins, Co.], FACS Aria [BD Biosciences, San Jose, CA], and Influx [Cytopeia Incorporated, Seattle, WA]) have the ability to sort four populations and they do this by using a variable charge so two populations may be sorted to the right and two to the left of the unsorted center stream.

The formation of drops and the determination of the drop delay are the factors that enable the flow sorter to be able to sort cells. However, to sort a pure population, it is necessary to ensure that the particles of interest be contained in a drop that does not also include an unwanted event. To ensure purity, normally only the drop containing - or likely to contain - the required event is charged. Although droplet formation in the absence of a sample is stable, stream dynamics and minor disturbances in the flow mean that the position of the event within the drop is not certain and that the charging pulse may not always be in phase with the drop production. This is especially true if an event is near the edge of a drop, and it may appear in the preceding or the following drop, and so to avoid cells loss more than one drop may be charged. Whether this is the preceding or following drop depends on the position of the event within the drop.

Sorting will never be a 100% efficient process, and some particles of interest will be missed and lost. Particles may arrive at the analysis point too close together to be analyzed separately and these are often termed hardware aborts; the flow sorter will ignore these events, as it cannot be sure whether they are wanted or unwanted. The number of hardware aborts will be influenced by the dead time of the electronics - that is, the time needed to finish processing one event before the next can be analyzed - and this is more important in analog systems than in more modern digital systems. Additional considerations are the sample pressure, the size of the cell, the cell concentration, and "stickiness" of the cells.

Another way that events of interest may be lost is when they can be analyzed separately but are too close together in the fluid stream to be included in separate drops; these are often termed software or coincidence aborts. Rather than compromise purity, the drop would not normally be sorted if it contained an unwanted event and these events would go to waste. However, it is possible to override this if the investigator is interested in a particular population even at the expense of purity; this is often termed enrich mode.

In fact, all sorters give the user the ability to vary how the sort is performed depending on the user's needs. Each cytometer does this in a slightly different way, but broadly speaking there are three sort modes: one that is optimized for purity, which would be the general default mode; there will be some cell loss where there are high coincidence aborts. The second mode is the enrich mode, where all events of interest are sorted at the expense of some reduction of purity. The third mode is often termed counter or single-cell mode and this is used when high count accuracy is required (e.g., when sorting single cells into multiwell plates for cloning or for sorting known numbers of cells for functional assays or transplantation). Here, the cytometer electronics will sort a drop only when it contains only one event of interest, when that event is in the center of the drop, and there are no possible coincident events in adjacent drops.

The speed of sorting depends on the time taken to generate a droplet, which depends on the frequency of the transducer. The frequency of droplet production in a stream-in-air sorter is determined by the jet velocity and the jet diameter and is defined as f = v/4.5 d, where v is the velocity of the fluid and d the diameter of the orifice (37). A typical sorter of the mid-1980s, sorted using a sheath pressure of approx 12.5 psi, will generate a sheath velocity of approx 10 m/s. This allows about 27,000 drops per second to be produced. Typically, flow rates for cells would be approx 5000 so we would expect only one drop in five to contain an event. To sort more quickly at a given orifice size, flow sorters have had to be developed that run at higher pressure and therefore greater sheath velocity. Increasing the sheath pressure to 60 psi allows approx 100,000 drops to be produced every second, giving an approx fourfold increase in the speed of sorting. The optimal drop drive frequency will vary with the nozzle orifice diameter. Table 1 shows typical sort pressures, drop drive frequencies, and sort times for a range of common nozzle sizes. The maximum speed of sorting is limited not by the number of drops that may be produced but by the sheath pressure needed to produce them; a drop frequency of 250,000 per second is possible but only if a pressure of 500 psi is used and this would be contraindicated for live cells although these pressures have been used in sorting chromosomes (38).

Was this article helpful?

0 0

Post a comment