SEARCH

SEARCH BY CITATION

Keywords:

  • flow cytometry;
  • phytoplankton;
  • T-cells;
  • sheath fluid

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. Results
  6. DISCUSSION
  7. Acknowledgements
  8. Literature Cited

Traditional flow cytometers use a sheath fluid to position particles or cells for cytometric measurements, but the need for sheath fluid greatly complicates flow cytometric instrumentation. A cytometric detector that is free of the requirements of sheath fluid can simplify the design of flow cytometers and can extend their use into a number of areas. We designed a flow cytometer that uses a combination of three photodetectors to sense the position of a particle in sample stream. The position-sensitive detectors create a virtual core in the sample stream that eliminates the need for sheath fluid. In this article, we demonstrate the efficacy of a virtual-core flow cytometer (VCFC) using test particles, immunofluorescently labeled thymocytes, and raw seawater. The VCFC performs accurate measurements that can be used for a number of uses including environmental monitoring or simple immunology tests. © 2009 International Society for Advancement of Cytometry


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. Results
  6. DISCUSSION
  7. Acknowledgements
  8. Literature Cited

Although the main application of flow cytometry is in the detection and enumeration of different cell types for immunology, the technique is used in disparate research fields, such as oceanography and stem cell biology (1–6). In nearly all flow cytometers, particles or cells pass through the center of a tightly focused laser beam. Scattered and fluorescent light generated by the interaction of the particles with the laser light is analyzed to determine characteristics such as the presence and quantity of fluorescent molecules. To accurately and precisely measure these characteristics, the cytometer must consistently position the particles with respect to the focused laser beam. There are many methods for positioning particles in a stream including hydrodynamic focusing (7), dielectrophoretic focusing (8, 9), capillaries (10, 11), and acoustic focusing (12–14). The majority of commercially available flow cytometers use a hydrodynamically focused sample core surrounded by a clear carrier fluid. In settings other than a well-equipped laboratory, the availability of sheath fluid has limited the utility of flow cytometers. For environmental monitoring, several attempts have been made to use either filtered, readily available fluids, or recycled sheath fluid (15–20) to obviate the need for several liters of sheath fluid each day. Unfortunately, the dependence on processing large amounts of sheath fluid greatly increases the complexity of these instruments.

To circumvent the necessity for sheath fluid, we designed a flow cytometer that senses the position of particles that pass near the center of a tightly focused laser beam. Using the position-sensitive information, the data acquisition circuitry can construct a virtual-core flow cytometer (VCFC) that is free of the shackles of sheath fluid. A flow cytometer that is free from the constraints of a special sheath fluid would find applications in new settings where the technology could provide crucial information such as long-term environmental monitoring, and medical diagnoses in resource-poor settings. In this report, we present a design for a VCFC as well as initial results of its efficacy.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. Results
  6. DISCUSSION
  7. Acknowledgements
  8. Literature Cited

Flow Cytometry

Flow cytometric measurements were performed using an Influx high-speed cell sorter (BD Biosciences, San Jose, CA) whose forward scatter detector was modified to include the position sensitive detector (PSD) (see Fig. 1). The machine and its electronics are described in more detail elsewhere (21–24). The position-sensitive detector consists of a 50× long-working-distance microscope objective (Mitutoyo Plan Apo, Edmund Optics, Barrington, NJ), mirrored aperture (HTA Photomask, San Jose, CA), and three photomultiplier tubes (PMT; R3896, Hamamatsu, Bridgewater, NJ). The front of the mirror and the image of the stream are monitored using a relay mirror and a CCD camera (V-1208-C/CS, Marshall Electronics, El Segundo, CA). Data were obtained using custom software (Spigot, BD Biosciences, San Jose, CA) and analyzed using Flowjo (Tree Star, Ashland, OR). Fluorescent microspheres (1- and 6-micron YG, Polysciences, Warrington, PA) were used for alignment and for initial tests of the PSD. A 457-nm 300 mW laser (Melles Griot) was used as the light source for measurements on raw seawater. A 200 mW Sapphire 488-nm laser (Coherent, Santa Clara, CA) was used for measurements of T cells. For most of the measurements shown, the samples were run at relatively low pressures, between 5 and 12 psi. At these pressures, the velocity of stream was between 8 and 12 m/s. For seawater measurements (where the sample density was relatively low), ∼5 mL was analyzed in the instrument at a rate of 250 events/s. For an accurate assessment of the murine thymocytes, ∼0.5 mL of sample was analyzed at a rate of 1800 events/s.

thumbnail image

Figure 1. The position sensitive detector can be constructed as the central portion of a virtual-core flow cytometer. A 50× long working distance microscope objective collects scattered and fluorescent light from a stream and focuses the light onto the surface of a pinhole mirror. Light that passes through the pinhole is analyzed by a set of detectors that first determine whether the particle is properly positioned for a fluorescence measurement. Photomultiplier tubes, dichroic mirrors, and bandpass filters are used to make a bivariate measurement of two fluorescence intensities, from the fluorochromes PE and PE-Cy5. The instrument was evaluated using murine thymocytes stained for CD4 and CD8 expression.

Download figure to PowerPoint

Murine Thymocytes

Murine thymocytes were stained for CD4 and CD8 expression and fixed following a protocol described elsewhere (25). Cells were suspended at approximately 1 × 106/mL and analyzed using custom software (Spigot, Becton Dickinson Biosciences, San Jose, CA and Flowjo, Treestar, Ashland, OR).

Results

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. Results
  6. DISCUSSION
  7. Acknowledgements
  8. Literature Cited

Sheath fluid management and precise sample injection can be eliminated when one uses a combination of photodetectors that also sense the position of the particles. To achieve this goal, we have developed an optical system that borrows from the imaging principles of confocal microscopy as shown in Figures 1 and 2. This sensing system is based on the principle that light from a properly focused object will pass through a pinhole located in the image plane, whereas improperly focused images will cast light onto multiple detectors located at the edge of the pinhole. At its heart, the detector system has a mirrored surface with an aperture to act as a pinhole. Right-angle prisms that send light toward two PMTs located orthogonal to the mirror plane are glued onto the mirror's surface over the edges of the aperture as shown in Figure 2d. This sensor is placed in the image plane of the lens as shown in Figure 2a. This configuration can be used to evaluate the sharpness of the projected light spot. Scattered or fluorescent light originating from the virtual core will be sharply focused and will pass through the center of the pinhole mirror aperture undeviated and into the central (fsc) detector as in Figure 2a. Out of focus light originating from outside the virtual core will produce a blurred image on the pinhole mirror, and will send light to one or both of the orthogonal detectors as in Figures 2b and 2c. The signals from the three PSDs can be analyzed to ignore measurements of particles that do not follow a proper trajectory through the focal region of the optical system. Thus, data can then be recorded for signals from the particles that are properly positioned. We choose to use a highly reflective (>99.5%) surface for the aperture rather than an opaque surface so one can adjust the focus of the 50× objective onto the surface of the mirror using an inexpensive CCD camera. This camera allows one to accurately align the laser, PSD detectors, and the sample stream with one another.

thumbnail image

Figure 2. Position sensitive detector. The position sensitive detector uses three photomultipliers (labeled fsc, d1, and d2) to monitor a stream in air sorter, combining these measurements to determine which events come from the center of the stream and are suitable for a cytometric measurement. Scattered light from the center of the stream passes through the center of the pinhole mirror undeviated as in (a). Light from outside the virtual core will either enter only one of the detectors, or be detected in all three as in (b) and (c). (d) The aperture in the pinhole mirror and prisms glued to the distal surface of the mirror to direct light to the orthogonal detectors. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

A challenge of this type of detector is to analyze the data in a fashion that discriminates between particles that are in the virtual core and particles that are outside the core, but positioned in a manner analogous to Figure 2c. In this position, the scattered laser light will be detected equally by both PSD detectors. If the gains are chosen judiciously, bivariate data from detectors d1 and d2 will lie along the diagonal axis of a bivariate plot of d1 versus d2. The effect of larger particles outside of the virtual core appearing analogous to small particles in the core but on the PSD axis can be mitigated, but not completely eliminated, by normalizing each of the PSD signals to the FSC signal. The FSC signal for a particle outside the core is small because the 50× objective has a relatively shallow depth of field and because the laser light less intense outside the virtual core as well. The measured virtual core is ∼25 μm from front to back, and 15–20 μm from side to side.

As constructed, the PSD consists of a 50× long-working-distance microscope objective, the pinhole mirror, and three PMTs. This objective was chosen because it offered a reasonable compromise between allowing one to reliably discriminate particles in the virtual core, and was not too difficult to align with respect to the lasers and other optical components. To facilitate alignment, a CCD camera monitors the image on the surface of the mirror to assist in the initial setup of the system and is imbedded in the Influx cell sorter. For a comparison of the PSD's efficacy, we tested the sensitivity of the device versus a traditional sample and sheath fluid cytometer. Once the stream is aligned to the PSD and the laser using the CCD camera, 1-micron and 6-micron diameter fluorescent beads were injected into the center of the stream and the stream was translated from side to side to test the response of the detector. Results of these tests are shown schematically in Figure 2. Side-to-side motion of the beads of ∼25 μm for the two sizes of beads results in the curves shown in Figure 3. Also shown in the figure are the positions of 1 μm polystyrene bead at positions corresponding to Figure 2a through 2c. These tests show that the PSD detector can be used as the heart of a VCFC that senses the position of particles whose sizes are suitable for traditional sheath-based cytometry. The PSD may be further refined by using different shapes for the pinhole aperture rather than the rectangular aperture shown in the PSD (as shown in Fig. 2d). The aperture's shape may be modified to selectively send more or less light to the orthogonal detectors (d1 and d2 in Fig. 2). By gluing the prisms that send light to the orthogonal detectors over the longer sides of an H-shaped aperture, one can increase the contribution from the orthogonal position-sensitive detectors. This refinement can help increase the discrimination between particles that are in the virtual core and those that lie outside the core.

thumbnail image

Figure 3. (a) The virtual core is determined by combining measurements from the three detectors. The virtual core is shown as an ellipse along the diagonal axis of the bivariate plot of d1 versus d2 where both have been normalized by the intensity of the fsc detector. Normalization helps further discriminate between small particles in the virtual core and larger ones that are not optimally positioned. The two curves shown represent a side to side motion of 1- and 6-micron beads. The dots represent where data from a 1-micron bead would fall for the corresponding panels of Figure 2. (b) Bivariate plot of d1/fsc versus d2/fsc where an ellipse has been drawn to schematically indicate the position of the virtual core. A side-to-side sweep of ∼25 μm of 1-micron polystyrene beads results in the curve shown. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

The VCFC's utility was evaluated further by testing its ability to perform simple immunophenotyping and to measure phytoplankton in raw seawater. The position-sensitive detector was modified to add two fluorescence channels that measured fluorescent light in the forward direction. For measurements of murine T cells, the forward scatter detector samples a portion of the light collected by the microscope objective; the majority of the collected light is directed toward the fluorescence detectors as in Figure 1. The murine T cells were labeled for CD4 and CD8 expression using fluorescent monoclonal antibodies. To directly compare the VCFC with a traditional sheath-based cytometer, we equipped the system with a fluidic valve that allowed us to switch from normal operation, where the sample is introduced into the center of the sheath core, and the virtual core operation, where the sample is injected directly into the sheath fluid using a syringe pump. Because the detector arrangement is identical for both systems, the test allows one to directly compare the efficiency and efficacy of the virtual core cytometer without having to consider the effect of switching fluorescence detectors.

Figure 4 shows bivariate measurements of murine T cells based on CD4 (x-axis) versus CD8 (y-axis) expression. Figure 4a shows the labeled sample measured with a cytometer using sheath fluid to constrain the sample to the center of the stream. In addition to cells that are negative for both functional groups, one can distinguish three labeled populations: a CD4+ population, a CD8+ population, and a small percentage of doubly labeled cells. The ratio of the CD4+ cells to CD8+ cells for this sample is 2.11. Figures 4b and 4c show the same sample measured using the virtual-core cytometer, shown with and without the PSD gate active. The gate discriminates between the cells distributed throughout the stream and the ∼1.4% that are correctly positioned for fluorescence measurements. By considering only cells properly positioned for a measurement, the VCFC recovers the cytometric information from a sample whose cells distributed throughout the 200-μm stream. The virtual-core cytometer not only recovers the general characteristics of the bivariate plot, but is also quantitative; CD4/CD8 ratio from the VCFC is 2.14, and differs from the ratio determined traditionally by less than 2%.

thumbnail image

Figure 4. Bivariate plots of CD4 and CD8 expression for murine thymocytes. (a) A bivariate measurement using sheath fluid to localize the sample to the center of the stream. (b) The same sample measured using the VCFC without applying a gate that determines the virtual core. (c) The same data as in (b) with a gate applied to isolate the events in the virtual core. Approximately 1.4% of the events shown in (b) pass through the virtual core. This technique not only recovers the general morphology of the plot, but can be used quantitatively. CD4/CD8 ratios determined using the data in (a) and (c) differ by less than 2%. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

Similar results may also be obtained using the virtual core cytometer to measure more complex, environmental samples. Raw seawater from Puget Sound was measured using both the traditional and the virtual core cytometer. Particles in the seawater were classified based on their phycoerythrin (PE) and chlorophyll (Chla) content using a 527-27 bandpass filter for PE and a 692-40 bandpass filter for the chlorophyll. Results for the traditional cytometer are shown in a bivariate plot in Figure 5a, and for the VCFC in Figure 5b. Also noticeable in the plots are 1-μm and 6-μm fluorescent beads that are used to set the gains of the fluorescence detectors and to help measure concentrations of the phytoplankton that are present in the sample. Again, the virtual core cytometer reproduces the both the general shape of the plot and can be used for ratiometric concentration measurements of the constituent groups.

thumbnail image

Figure 5. The VCFC was further evaluated using raw seawater from Puget Sound. Phytoplankton may be enumerated and classified by their expression of chlorophyll and PE. The figure shows data measured in a traditional flow cytometer (a) as well as data measured using the VCFC (b). The populations evident in the plot belong to phytoplankton as well as 1-micron and 6-micron YG beads used for calibration. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. Results
  6. DISCUSSION
  7. Acknowledgements
  8. Literature Cited

This work demonstrates a cytometer that does not rely on clear sheath fluid and localization of the sample at the center of the sheath fluid to make cytometric measurements. A cytometer that is based on a PSD that creates a virtual core lacks many of the subsystems required to ensure sample localization at the center of the fluid stream and allows the construction of a rugged, simple instrument. Not only is this instrument less complex, its sheath fluid requirements are greatly reduced. Rather than needing large quantities of sterile sheath fluid (usually several liters at a time), samples for the VCFC can be diluted with sheath solution to a suitable concentration and run directly.

The relative lack of complexity and less reliance on reagents makes a cytometer built around the PSD suitable for use in field laboratories and settings where a traditional laboratory cytometer is impractical. As an example, this system could be used to monitor phytoplankton populations in the ballast water of ships. A system with the PSD at its core can be automated to periodically sample the ballast water to enumerate phytoplankton functional groups using ratios of endogenous fluorophores. Although the system is, at this point, relatively immature, one could also foresee its use for more complex quantitation of marine groups through use of image-based analysis or internal standards, such as fluorescent microspheres. These particles can be injected into a known quantity of an environmental sample. The sample can then be characterized using the VCFC and environmental concentrations can be inferred by comparison with the known concentration of the microsphere standard.

A further advantage to removing the sheath fluid from the cytometer is the ability to completely characterize samples. In sheath fluid-based flow cytometry, one could attempt to reanalyze relatively precious samples to better characterize constituent populations. Reanalysis is impractical, however, because the sample is diluted ∼100-fold by sheath fluid with each successive analysis. In the VCFC, the sample is diluted only to the extent that is necessary to make the first measurement. At this point, one has the opportunity to measure the sample several more times without additional dilution until either photobleaching or other physiological effects compromise the sample's fluorescent characteristics.

The system described herein is the simplest implementation of the VCFC. As with traditional cytometers, the VCFC can be incorporated into a more complex system. For example, one could add additional scatter detectors and fluorescence detectors in directions orthogonal to the stream and the PSD to make the measurement more akin to commercially available flow cytometers. This system can be engineered to any degree of complexity, including the addition of extra lasers to excite other fluorophores. Although the data presented here have addressed only two applications, CD4/CD8 ratios and oceanography, there are many realms in which the VCFC could make accurate cytometric measurements of particles and cells.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. Results
  6. DISCUSSION
  7. Acknowledgements
  8. Literature Cited

The authors gratefully acknowledge the help of Eric O'Connor of the MRC Clinical Sciences Centre who supplied the stained murine T-cell sample used in this work.

Literature Cited

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. Results
  6. DISCUSSION
  7. Acknowledgements
  8. Literature Cited