UNIVERSITY UBRARIES - ML or
`CEO
`oy, a
`
`32777 014087121
`
`5
`
`
`
`ie| a»|QBre|ao|Botl
`
`
`
`LOW CYTO
`
`
`
`am *#@/)| @@)0O@|O@+|
`
`
`
`O@
`
`Oo@
`
`ABS Global, Inc. and Genus ple — Ex. 1019, cover1
`
`

`

`PRACTICAL FLOW
`CYTOMETRY
`
`Fourth Eclition
`
`(cid:36)(cid:37)(cid:54)(cid:3)(cid:42)(cid:79)(cid:82)(cid:69)(cid:68)(cid:79)(cid:15)(cid:3)(cid:44)(cid:81)(cid:70)(cid:17)(cid:3)(cid:68)(cid:81)(cid:71)(cid:3)(cid:42)(cid:72)(cid:81)(cid:88)(cid:86)(cid:3)(cid:83)(cid:79)(cid:70)(cid:3)(cid:177)(cid:3)(cid:40)(cid:91)(cid:17)(cid:3)(cid:20)(cid:19)(cid:20)(cid:28)(cid:15)(cid:3)(cid:70)(cid:82)(cid:89)(cid:72)(cid:85)(cid:3)(cid:21)
`ABS Global, Inc. and Genusplc — Ex. 1019, cover 2
`
`

`

`PRACTICAL FLOW
`CYTOMETRY
`
`Fourth Edition
`
`HOWARD M. SHAPIRO
`
`GEORGE MASON UNIVERSITY
`UNIVERSITY LIBRARIES
`
`@)WILEY-LISS
`
`A JOHN WILEY & SONS, INC., PUBLICATION
`
`(cid:36)(cid:37)(cid:54)(cid:3)(cid:42)(cid:79)(cid:82)(cid:69)(cid:68)(cid:79)(cid:15)(cid:3)(cid:44)(cid:81)(cid:70)(cid:17)(cid:3)(cid:68)(cid:81)(cid:71)(cid:3)(cid:42)(cid:72)(cid:81)(cid:88)(cid:86)(cid:3)(cid:83)(cid:79)(cid:70)(cid:3)(cid:177)(cid:3)(cid:40)(cid:91)(cid:17)(cid:3)(cid:20)(cid:19)(cid:20)(cid:28)(cid:15)(cid:3)(cid:70)(cid:82)(cid:89)(cid:72)(cid:85)(cid:3)(cid:22)
`ABS Global, Inc. and Genusplc — Ex. 1019, cover 3
`
`

`

`Copyright © 2003 by John Wiley & Sons, Inc. All rights reserved.
`
`Published by John Wiley & Sons, Inc., Hoboken, New Jersey.
`
`No part ofthis publication may be reproduced, stored in a retrieval system ortransmitted in any form or
`by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except as
`permitted under Section 107 or 108 of the 1976 United States Copyright Act, withouteither the prior
`written permission of the Publisher, or authorization through paymentofthe appropriate per-copyfee to
`the Copyright Clearance Center, inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400,fax
`(978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should
`be addressed to the Permissions Department, John Wiley & Sons, Inc., 11! River Street, Hoboken, NJ
`07030, (201) 748-6011, fax (201) 748-6008, e-mail: permreq@wiley.com.
`
`Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in
`preparing this book, they make no representation or warranties with respect to the accuracy or
`completeness ofthe contents of this book and specifically disclaim any implied warranties of
`merchantability or fitness for a particular purpose. No warranty may be created or extended by sales
`representatives or written sales materials. The advice and strategies contained herein maynot be suitable
`for yoursituation. You should consult with a professional where appropriate. Neither the publisher nor
`author shall be liable for any loss of profit or any other commercial damages, including but not limited
`to special, incidental, consequential, or other damages.
`
`For general information on our other products and services please contact our Customer Care
`Departmentwithin the U.S. at 877-762-2974, outside the U.S. at 317-572-3993 or fax 317-572-4002.
`
`Wiley also publishes its booksin a variety ofelectronic formats. Some contentthat appears in print,
`however, may not be available in electronic format.
`
`Library ofCongress Cataloging-in-Publication Data:
`
`Shapiro, Howard M. (Howard Maurice), 1941—
`Practical flow cytometry / Howard M. Shapiro. — 4th ed.
`Includes bibliographical references and index.
`ISBN 0-471-41 125-6 (alk. paper)
`|. Flow cytometry.
`[DNLM:|. Flow Cytometry. QH 585.5.F56 $529p 2002]_1.Title.
`QHS585.5.F56 S48 2002
`571.6'0287—dc21
`
`2002002969
`
`Printed in the United States ofAmerica.
`
`10987654321
`
`ABSGlobal, Inc. and Genus ple — Ex. 1019, cover 4
`
`

`

`introduced by the lower number ofelectrons it starts with
`and, in fact, there is also a statistical aspect to the PMT’s
`gain mechanism.
`Unfortunately, the high-gain, low noise amplifiers we'd
`need to use photodiodes as sensitive fluorescence detectors
`don’t exist. There are, however, solid-state devices called
`avalanche photodiodes (APDs), which combine high
`quantum efficiency with a mechanism that can produce
`gains as high as a few thousand when a voltage is applied
`across the diode. While APDs are now used for both scatter
`and fluorescence detection in some commercial
`flow
`cytometers, they do not match thesensitivity of PMTs.
`The photodetectors we have been talking about are
`sources of electric current. A preamplifier, which is the
`first stage in the analog signal processing electronics, con-
`verts the current output from its associated detector to a
`voltage. The preamplifier also accomplishes the important
`task of DC baseline restoration.
`An ideal flow cytometer is something like an ideal dark
`field microscope; when there’s no cell in the observation
`region, the detector shouldn't be collecting any light atall.
`In practice, there’s always some small amountoflight com-
`ing in. In the case ofthe scatter detectors, most ofthis light
`is stray scattered light from the illuminating beam; in the
`case of the fluorescence detectors, the light background may
`come from fluorescence excited in various optical elements
`such as the flow chamber, lenses, and filters, from fluores-
`cence due to the presence of fluorescent materials in the
`medium in which cells are flowing, and from Raman scat-
`tering, which produces light at frequencies corresponding to
`the difference between theillumination frequency and the
`frequencies at which absorption changes molecular vibra-
`tional states. In flow cytometry, the majorinterference due
`to Raman scattering results from scattering by water; when
`488 nm illumination is used, this scattering occurs at about
`590 nm, and mayinterfere with detection of signals from
`probes labeled with phycoerythrin, which fluoresces near this
`
`The net result of the presence of all of the abovemen-
`tioned stray light sources is that there are some photons
`reaching the detectors in a flow cytometer even when there
`isn’t a cell at the observation point, producing some current
`at the detector outputs. There may also be some contribu-
`tion from the so-called dark current of the detector, which
`results from the occasional electron breaking loose from the
`cathode due to thermal agitation. There are some situations
`in which performance of photodetectors is improved by re-
`frigerating them to reduce dark current; flow cytometry in
`the contexts we're discussing isn’t one of them. Even with
`the detectors in liquid nitrogen, we'd have to deal with the
`background light, which will contribute a signal with an
`average value above zero to whatever signal we collect from
`thecells.
`The backgroundsignal can be considered as the sum of a
`constant direct current (DC) component and a variable
`alternating current (AC) component,representing the fluc-
`
`Overture / 55
`
`tuations due to photon statistics and to other sources of
`variation in the amountofstray light reaching the detector.
`Oneimportantsource of such variation may be light source
`noise, i.e., fluctuations in the light output of the laser or
`lamp used for illumination; in some circumstances, particu-
`larly scatter measurements of small particles, source noise
`can be the majorfactor limiting sensitivity.
`What we'd like to measure when a cell does pass by the
`observation station is the amountoflight coming from the
`cell, not this amountplus the backgroundlight. We can do
`this, to a first approximation, by incorporating an electronic
`circuit that monitors the output of the detector and uses
`negative feedback to subtract the slowly varying component
`of the output from the input, thereby eliminating most of
`the DC backgroundsignal, and restoring the baseline value
`of the preamplifier output to ground.
`In practice, baseline restorers will keep their voltage out-
`puts within a few millivolts of ground when no cells are
`coming by. When a cell does arrive,
`it will scatter and
`probably emit small amounts of light, which will be col-
`lected and routed to the various detectors, producing tran-
`sient increases, or pulses, in their output currents, which
`will result in voltage pulses at the preamplifier outputs. At
`this point, as was noted on p. 17,all of the information we
`wanted to get from the cell resides in the heights, areas,
`widths, and shapes of those pulses; we will ultimately con-
`vert these to digital values, in which form they can be dealt
`with by the computers that are almost universally used for
`data analysis in flow cytometry. However, before we get into
`the details of how pulse information is processed, we ought
`to consider the only element of Figure 1-21 that has been
`neglected to this point, namely, the cell flowing through the
`apparatus, and how it gets there.
`
`Putting the Flow in Flow Cytometry
`Figure 1-21 describes the cell as being in the center of
`the cuvette, and I have already talked about a core or sample
`stream ofcells that is about 20 ym wide, while mentioning
`thar the internal dimensions of the cuvette are on the order
`of 200 by 200 pm. The space between the core and thein-
`ner walls of the cuvette is occupied by another stream of
`flowing fluid, called the sheath. How the core and sheath
`get where they are can be appreciated from a look at Figure
`1-22.
`Fluid mechanics tells us that, if one smoothly flowing
`stream offluid (i.e., the core stream) is injected into the cen-
`ter of another smoothly flowing stream of fluid (i.e., the
`sheath stream), the two streams will maintain their relative
`positions and not mix much, a condition called laminar
`flow. There are generally differences in fluid flow velocity
`from theinside to the outside of the combined stream, but
`the transitions are even. If the velocities of the two streams
`are initially the same, and the cross-sectional area of the ves-
`sel in which they are flowing is reduced, the cross-sectional
`areas of both streams will, obviously, be reduced, but they
`will maintain the same ratio of cross-sectional areas they had
`
`ABSGlobal, Inc. and Genus ple — Ex. 1019, p. 55
`
`

`

`56 / Practical Flow Cytometry
`
`at the injection point. If the sheath stream is flowing faster
`than thecore stream at the injection point, the sheath stream
`will impinge on the core stream, reducingits cross-sectional
`area. In the flow chamber of a flow cytometer, both mecha-
`nisms of constricting the diameter ofthe core stream may be
`operative.
`The core stream, which contains the cell sample, is in-
`jected into the flowing water or saline sheath stream at the
`top of a conical tapered region that, in the flow chamber
`shown in thefigure, is groundinto the cuvette. The areas of
`both streams are reduced as they flow through the tapered
`region and enter the flat-sided region in which cells are ob-
`served. Core and sheath streams may be driven either by gas
`pressure (air or nitrogen), by vacuum, or by pumps; most
`instruments use air pressure. Constant volume pumps, ¢.g.,
`sytinge pumps, which,
`if properly designed, deliver a
`known volume of sample per unit time with minimum pul-
`sation, provide finer control over the sampleflow rate. Since
`knowing the sample flow rate makes it easy to derive counts
`ofcells per unit volume, flow cytometric hematology analyz-
`ers incorporate constant volume pumps; why fluorescence
`flow cytometers, in some cases made by the same manufac-
`turers, do not remains something ofa mystery.
`The overall velocity of flow through the chamber is gen-
`erally determined by the pressure or pumpsetting used to
`drive the sheath. If the sheath flow rate is increased with no
`change in the core flow rate,
`the core diameter becomes
`smaller and the cells move faster; if the sheath flow rate is
`decreased under the same circumstances, the core diameter
`becomes larger and the cells move more slowly. In some
`circumstances, it is desirable to adjust sheath flow rates; if
`cells move more slowly, they spend more time in the illumi-
`nating beam, receive proportionally more illumination, and
`they therefore scatter and emit proportionally more light. If
`the amountoflight being collected from cells is the limiting
`factor determining sensitivity, slowing the flow rate can im-
`provesensitivity, allowing weaker signals to be measured.
`This aside, it is generally preferable to be able to control
`the core diameter, and therefore the volume of sample and
`number ofcells analyzed per unit time, without changing
`the velocity at which cells flow through the system. This is
`done by leaving the sheath flow rate constant and changing
`the driving pressure or pump speed for the core fluid. More
`drive for the core results in a larger core diameter; more cells
`can be analyzed in a given time,butprecision is likely to be
`decreased because theillumination from a Gaussian beam is
`less uniform over a larger diameter core. Less drive for the
`core gives a smaller core diameter and a slower analysis rate,
`but precision is typically higher. When the cytometer is be-
`ing used to measure DNA content, precision is important;
`when it is being used for immunofluorescence measurement,
`precision is usually of much less concern.
`The use of sheath flow as just described has proven es-
`sential in making flow cytometry practical. Without sheath
`flow, the only way of confining 10 pm cells within a 20 pm
`diameter stream would be to observe them in a 20 pm di-
`
`ameter capillary or in a stream in air produced by ejecting
`the cells through a 20 ym diameter orifice. This would very
`quickly run afoul of Shapiro’s First Law (p. 11). As a matter
`of fact, even with sheath flow, Shapiro’s First Law frequently
`came into play when cell sorters were typically equipped
`with 50 ym orifices. That orifice size was fine for analyzing
`and sorting carefully prepared mouse lymphocytes, but peo-
`ple interested in analyzing things like disaggregated solid
`tumors might encounter mean intervals between clogs of
`two minutes or so. With the larger cross-sectional areas of
`the flow chambers now used in most flow cytometets, clogs
`are notnearly the problem they once were.
`Clogs, however, are not the only things that can disturb
`the laminar flow pattern in the flow chamber. Air bubbles
`perturb flow, as do objects stuck inside the chamber but not
`large enough to completely obstruct it. In the first commer-
`cial cell sorters, the standard method for getting rid of air
`bubbles was to remove the chamber from its mount while
`the apparatus was running, and turn it upside down; the
`bubble would rise to the top and emerge from the nozzle
`along with a stream of sheath and sample fluid that would
`spray all over the lab. This technique became inappropriate
`with the emergence ofAIDSin the 1980's. Now, even drop-
`let sorters incorporate an air outlet (which I have referred to
`elsewhere as a “burp line”) for getting rid of bubbles. In
`some flow cytometers with closed fluidic systems,
`the air
`bubble problem is minimized by having the sample flow in
`
`OUTLET FOR
`PURGING
`BUBBLES, ETC.
`
`‘\
`
`J
`
`OF CUVETTE
`
`CONICAL
`TAPERED
`REGION
`
`FLAT-SiDED
`REGION
`OF CUVETTE
`
`INLET FOR
`SAMPLE OR
`CORE FLUID
`CONTAINING
`
`CELLS
`
`INLET FOR
`SHEATH FLUID
`
`CORE
`INJECTOR
`
`TAPERING
`CORE STREAM
`
`OBSERVATION
`POINT
`
`SHEATH
`
`STREAM —~~__ FLUID
`
`OUTLET
`
`Figure 1-22. A typical flow chamber design.
`
`ABSGlobal, Inc. and Genus ple — Ex. 1019, p. 56
`
`

`

`at the bottom and outat the top, essentially turning Figure
`1-22 upside down; bubbles are more orless naturally carried
`out of the flow chamber.
`Disturbances in laminar flow, whether due to bubbles or
`junk, often result in the core stream deviating from its cen-
`tral position in the flow chamber andin differences in veloc-
`ity between differentcells at different points within the core.
`Turbulent fluid flow is now described mathematically using
`chaos theory; you can recognize turbulent flow in the flow
`chamberbythe chaos in your data.
`For the present, we will assume the flow is laminar, the
`optics are aligned, and the preamplifiers are putting out
`pulses with their baselines restored, and consider the next
`step along the way toward getting results you can put into
`prestigious journals and/or successful grant applications.
`
`Signal Processing Electronics
`We have already mentioned that a cell is going to pass
`through the focused illuminating beam in a flow cytometer
`in something under 10 ps, during which time the detectors
`will produce brief current pulses, which will be converted
`into voltage pulses by the preamplifiers. Using analog peak
`detectors, integrators, and/or pulse width measurementcir-
`cuits, followed by analog-to-digital conversion, or, alterna-
`tively, rapid A-D conversion followed by digital pulse proc-
`essing (p. 21), we will reduce pulse height, area, and width
`to numbers, at least some of which will, in turn, be propor-
`tional to the amounts of material in or on the cell that are
`scattering or emitting light. But which numbers?
`First, let’s tackle the case in which the focal spot, in its
`shorter dimension, along theaxis offlow, is larger than the
`cell, meaning that there is some time during the cell’s transit
`through the beam at which the whole cell is in the beam.
`Because the beam is Gaussian, the whole cell may not be
`uniformly illuminated at any given time, but intuition tells
`us that when the center of the cell goes through the center of
`the beam, we should be getting the most light to the cell and
`the most light out of it. The preamplifier outputsignal, after
`baseline restoration, is going to be roughly at ground before
`the cell starts on its way through the beam, andrise as the
`cell passes through, reaching its peak value or height when
`the center ofthe cell is in the center of the beam, and then
`decreasing as the cell makes its way out of the beam. Since
`the wholecell is in the beam when the pulse reaches its peak
`value, this value should be proportional to the total amount
`ofscattering or fluorescent material in or on thecell.
`Things get a lircle more complicated when the beam is
`the size of the cell, or smaller. In essence, different pieces of
`the cell are illuminated at different times as thecell travels
`through the beam.In order to come up with a value repre-
`senting the signal for the whole cell, we have to take the
`area, or integral, rather than the height of the pulse. There
`are two ways to do this with analogelectronics. Oneis to
`change the frequency response characteristics of the pream-
`Plificr, slowing it down so thar it behaves as an integrator, in
`the sense that the height of the pulse coming out of the
`
`Overture / 57
`
`slowed-down preamplifier is proportional to the area or in-
`tegral of the pulse chat would come outofthe original fast
`preamplifier. Putting the slowed pulse into a peak detector
`then gives us an output proportional to the area or integral
`we're trying to measure. Alternatively, we can keep the fast
`preamplifier, and feed its output into an analog integrator
`instead of a peak detector.
`If we decide to do digital pulse processing, we have to
`digitize the pulse trains from the preamplifier outputs rap-
`idly enough so that we have multiple samples or “slices” of
`each pulse. We can then add the values of a number ofslices
`from the middle ofthe pulse to get an approximation of the
`area, or integral; eight slices will do, but sixteen are better.
`This works pretty well. However, if we're only taking eight
`or sixteen slices of a pulse, we may not get as accurate a peak
`value or a pulse width value as we could using analog elec-
`tronics.
`The peak value we get ftom digital processing is simply
`the largest of our eight or sixteen slices. These provide us
`with only a fairly crude connect-the-dots “cartoon” of the
`pulse, thus, while there is a substantial likelihood thar the
`largest digitized slice is near the peak value, there is a rela-
`tively low probability that the digitization will occur exactly
`whenthe peak value is reached.
`Similarly, if we estimate pulse width from the number of
`contiguous slices above a set threshold value, we will have a
`fairly coarse measurement; if the digitization rate gives us at
`most sixteen slices, our range of pulse widths runs from 1 to
`16, with each increment representing at least a 6 percent
`change over the previous value. If we had fast enough ana-
`log-to-digital converters to be able to take a few hundred
`slices of each pulse, and fast enough DSP chips to process
`the data, we could get rid of analog peak detectors and pulse
`width measurementcircuits, but we're not there yet. The
`digital integrals are already good enough to have been incor-
`porated into commercial instruments.
`
`Is It Bigger than a Breadbox?
`I have been referring to benchtop flow cytometers and
`big sorters, but I haven’t shown you any pictures. Now’s the
`timeto fix that.
`Figure 1-23, on the next page, shows the Becton-
`Dickinson FACScan,
`the first
`really successful benchtop
`flow cytometer, introduced in the mid-1980’s. It uses a sin-
`gle 488 nm illuminating beam from an air-cooled argon ion
`laser, and measures forward and side scatter and fluorescence
`at 530 and 585 and above 650 nm. Thedata analysis system
`is an Apple Macintosh personal computer, shown in front of
`the operator.
`Figure 1-24 (courtesy of Cytomation) shows that com-
`pany’s MoFlo high-speed sorter. The optical components,
`including two water-cooled ion lasers and a large air-cooled
`helium-neon laser, are on an optical table in front of the
`operator. Most of the processing electronics are in the rack
`to the operator’s left; the two monitors to her right display
`data from an Intel/Microsoft type personal computer.
`
`ABSGlobal, Inc. and Genus ple — Ex. 1019, p. 57
`
`

`

`166 / Practical Flow Cytometry
`
`Thegallium arsenide R636 is useful in spectrofluorome-
`ters because it has a relatively flat response curve, but its
`maximum gain is quite low; unless you need to work at 900
`nm, the R1477 and R3896 are bemer choices. For most
`work at or below 500 am,or for scatter measuremens out as
`far as 633/635 nm, the bargain-priced 931R will do a fine
`job, bur, for measuring fluorescence anywhere above 500
`nm, a tube with a higher quancum efficiency is worth its
`price. I found, when I excited propidium with less than 5
`mW at 488 nm, that the extra quantum efficiency of an
`R928 (vs. a 931B) helped lower measurement CVs.
`I have already mentioned the compact Hamamatsu de-
`tector modules that incorporate 1/2" PMTs. The current
`H7710-03 features an R6357 PMT; other, less expensive
`modules in the series are made with less spectacular tubes,
`which will probably be fine at 550 am and shorter wave-
`lengths.
`Hamamatsu has also gone in some other interesting di-
`rections in PMT development. They have made ultraminia-
`ture PMTs thar fit into the 16 mm diameter, 12 mm long
`TO- 8 “can” package normally used for transistors and di-
`odes. The first generation of these aibes had neither high
`gain nor high fluorescence sensitivity, but the newest offer-
`ings, the R7400U series, include at least one tube with high
`red sensitivity; however, while the quantum efficiency ofthis
`tube is competitive, the gain (5 x 10°) is still on the low side.
`These PMTsare also available in modules; I have been told
`that neither the tubes nor the modules are significantly
`cheaper than the larger varieties.
`The other notable Hamamatsuoffering is a multianode
`PMT,with a square or linear array of anodes and fine mesh
`dynodes. The different anodes respond to light impinging
`on different areas of the cathode,at least up to a point. The
`linear array multianode PMT can receive the light dispersed
`from a grating, with the outputs from the different anodes
`then providing spectral
`information. Zeiss has apparently
`used a multianode PMT in a spectral detector for its Meta
`confocal microscope system; I have also heard of one being
`used in an experimental flow cytometer.
`
`Photomultipliers: Inexact Science
`After all this discussion of PMT sensitivity, I am obliged
`to let you in on one of the dirty litle secrets of electro-
`optics; the tabulated values are a rough guide. There is a lot
`ofvariation from device to device in most of the important
`parameters; cathode sensitivity and gain for a given applied
`voltage will vary over at least a 2:1 range, and individual
`variations in photocathode composition make for individual
`deviations from the spectral response curves of Table 4-3.
`The good news is that manufacturers cest the sensitivity of
`individual PMTs and provide the results to the buyer. So,if
`you acquire pwo R3896's, you probably want to use the
`“hotter” one at the longer wavelength. My impressionis that
`plain silicon photodiodes don’t vary nearly as much as
`PMTs, although avalanche diodes may.
`
`Charge Transfer Devices: CCDs, ClDs, Etc,
`
`You are, by now,likely to have encountered the charge
`coupled device, or CCD,either in its low-cost form in your
`camcorder or digital camera, or in its rarer, cooled, more
`esoteric and expensive guise in imaging cyromerers designed
`for low light level measurements. CCDs are one of a class of
`photodetectors described as charge transfer devices; there
`are also, for example, charge injection devices, or CIDs,
`In all of these, exposure to light causes accumulation of
`electric charge in individual elements that are usually ar-
`ranged in a linear or rectangular array; attached dectronic
`circuitry senses the amountof stored charge in each element
`at regular intervals. Charge transfer devices are well suited
`for imaging; because they integrate over time, they are useful
`for measurement of low light intensities, especially when
`cooled. However, they tend to be relatively slow, and, on
`that account, they have not been widely used in flow cy-
`tometry. Newer,faster arrays may be useful in polychromatic
`detection for measurement ofemission spectra in flow’.
`I hear thar there are now ways ofgetting gain out of
`CCDs, bur I don't have either details or confirmation. New
`CMOS image sensors are starting t give CCDs « run for
`their money in the commercial camera markets; whether
`they will make inroads in science remains to be seen,
`Intel, which joined forces with Martel to produce the
`QX3 Computer Microscope, a cute roy that uses a CCD to
`provide 320 by 200 pixel images, decided in late 2001 to
`stop making the gadgets;
`they came on the market at
`$119.95, and I've snapped up a few for $49.95. There is still
`time to introduce your kids or grandkids to microscopy and
`cytometry via this route,
`
`4.6 FLOW SYSTEMS
`
`to this
`Ic has probably not escaped your notice that
`point, in this book on flow cytomerry, I have gone into great
`detail abour light, optics, light sources, lenses, filters, and
`detectors and said very little indeed about flow systems,
`without which flow cytometry wouldn’t be flow cytometry.I
`claim there has been a method to chis madness. All the other
`stuff doesn't change just because you work in a flow system,
`and all the other stuff works when you don't work in a flow
`system. You can use the same light sources, and the same
`lenses, and the same filters and detectors, to illuminate and
`collect and detect light from cells on slides, or in culture
`dishes, or in microtiter (or nanotiter) plates, or in small cap-
`illaries, as you use to do the same jobs for cells in flow sys-
`tems, A few chapters from now, we will discuss parameters
`and probes, virtually all of which can be measured in or ap-
`plied to cellls in static as well as in flow cytometers. There are
`some cytometric tasks for which flow cytometry is prefer-
`able, and some for which it is not, but most of the funda-
`mentals of flow cytometry are the fundamentals of cytome-
`try in general. Among those that are not are the theoretical
`and practical details of fluid mechanics or hydrodynamics
`and flow systems, to which we now turn.
`
`ABSGlobal, Inc. and Genus ple
`
`— Ex. 1019, p. 166
`
`

`

`How Flow Cytometers Work / 167
`
`same time through the narrow and wide portions of the cap-
`illary, tht flow velocities at different points in the system will
`be different, i.c., higher in the narrow portions than in the
`wider ones. In fact, the product of cross-sectional area, A,
`and average flow velocity, v, remains constant and equal to
`the volume flow rare, Q, at any point along the flow system.
`But whyare we talking about “average” velocity?
`Water, which is che major component of both the sheath
`and core fluids and which therefore determines their flow
`characteristics, is not what physicists call an ideal liquid; ic
`exhibits viscosity, which, in physical terms, means that some
`work must be done on a volume of fluid to get ir to change
`its shape, While the everyday definition ofviscosity conjures
`up fluids such as glycerin, which has a viscosity abour 1000
`times that ofwater, the effects of the viscosity ofwater on its
`pattern of flow are noticeable enough. In particular, we ob-
`serve chat the stationary flow ofwater through small tubes is
`laminar. If we look at a cylindrical cube of radius R contain-
`ing flowing water, we find that the velocity of water at dif-
`ferent distances from the axis or center of the tube varies.
`Velocity is highest along the axis; at the walls of the tube,
`there is actually a thin boundary layer of water that is nor
`moving (i.c., it has zero velocity). At any intermediare point
`a distance r along the radius, the velocity is proportional to
`(R - 1)’. This produces a so-called parabolic profile of flow
`velocities, as if the water were broken up into thin cylindri-
`cal layers (laminae in Latin) that were sliding over one an-
`
`CORE INLET —
`
`
`_SHEATH INLET
`
`CORE
`INJECTOR
`
`SHEATH
`CONSTRICTS
`
`CORE
`SLUG FLOW
`
`LAMINAR FLOW
`WITH PARABOLIC
`VELOCITY PROFILE
`
`P
`NECKDOWN
`REGION
`
`
`"| y
`
`In a flow cytometer,it is the task of the flow system to
`transport cells in the sample to and through the measure-
`mentstation(s). In “static” microspectrophotometers or im-
`age analysis systems, the same job is usually delegated to
`precisely made and well-controlled mechanical hardware.
`However, while the mechanical transport system in a static
`cytometer may be inactive while actual measurements are
`being made, the flow system in a flow cytometer is continu-
`ally active, and must move the entire cohort of cells in a
`sample past the measurementstation(s) along almost idenui-
`cal
`trajectories at almost identical velocities if satisfactory
`data are to be obtained from the measurementprocess. This
`requires that a stable flow pattern be achieved and main-
`tained, and both designers and users of flow cyrometers must
`play active roles in this process.
`
`Flow System Basics
`The design of flow systems and the underlying physical
`principles have been discussed at
`length by Pinkel and
`Stovel™ and by Kachel, Fellner-Feldegg, and Menke’. If
`you feel a strong urge to design your own flow cytometer,
`you will probably want to refer to one or both of those pub-
`lications. If you're willing to put up with whar the manufac-
`turers give you, and/or to do things my way, stick with me,
`and I will expand on the briefdiscussion of flow systems thar
`appeared on pp. 55-57, hoping to hir che high points of the
`references just cited,
`Almost all modern optical flow cytometer designs make
`use of sheath flow, or hydrodynamic focusing, to confine
`the sample or core fluid containing the cells to the central
`portion of a flowing stream ofcell-free sheath fluid. Sheath
`flow improves the precision with which the cell sample can
`be positioned in the observation region of the cytometer by
`restricting cells co the central region of the stream, and re-
`duces the likelihood of obstruction of the flow system. Sta-
`ble, unobstructed flow minimizes variations in the position
`and velocity of the core stream; when flow becomes unstable
`or turbulent, due to obstruction or other causes, measure-
`ments are likely to become imprecise and inaccurate.
`Figure 4-37 illustrates some aspects of fluid flow in a
`flow cytometer. Core (sample) and sheath inlet tubes are
`shown near the top of the flow chamber; near where these
`enter, there may also be a third port which can be connected
`to vacuum, allowing easy removal of air bubbles and back
`suction to clear clogs out of the orifice. Application of vac-
`uum is more likely to be successful for the first purpose than
`for the second.
`
`Gently Down the Stream: Laminar Flow
`Flow must be stable from the region of the core injector
`tip downward if core velocity and position are to be main-
`tained well enough vo allow good measurements to be made.
`We want stationary or streamline flow, a condition charac-
`terized by the constancy over time of flow velocity at any
`given point in the system. Since the law of conservation of
`mass dictates that the same volume offluid must pass in the
`
`Figure 4-37, Fluid flow In a flow cytometer.
`
`ABSGlobal, Inc. and Genus ple — Ex. 1019, p. 167
`
`

`

`168 / Practical Flow Cytometry
`
`other. So, we can't assume that the velocity of the fluid will
`be constant acr