U.S. patent number 3,710,933 [Application Number 05/211,473] was granted by the patent office on 1973-01-16 for multisensor particle sorter.
Invention is credited to James R. Coulter, Mack J. Fulwyler, John A. Steinkamp.
United States Patent |
3,710,933 |
Fulwyler , et al. |
January 16, 1973 |
MULTISENSOR PARTICLE SORTER
Abstract
An apparatus for rapidly and automatically analyzing and sorting
minute particles on the basis of certain preselected
characteristics. Particles flow in suspension through a flow
chamber having multiple sensing means to detect preselected
physical or chemical characteristics of each particle and then are
jetted between charging electrodes and deflection plates. Signals
from the sensors for each particle are compared with preset
standards, and those droplets containing particles having
characteristics not meeting those standards are automatically
charged by the charging electrodes. The deflection plates provide a
constant electric field which deflects charged droplets away from
uncharged droplets, thus sorting particles on the basis of their
conformance or nonconformance to standards set for the preselected
characteristics. This apparatus is particularly applicable to the
rapid and automatic sorting of biological cells.
Inventors: |
Fulwyler; Mack J. (Los Alamos,
NM), Steinkamp; John A. (Los Alamos, NM), Coulter; James
R. (Los Alamos, NM) |
Assignee: |
|
Family
ID: |
22787055 |
Appl.
No.: |
05/211,473 |
Filed: |
December 23, 1971 |
Current U.S.
Class: |
209/3.1;
435/6.12; 209/4; 209/579; 356/73; 377/10; 422/73; 435/40.51;
209/577; 356/39; 356/341; 347/1 |
Current CPC
Class: |
G01N
15/12 (20130101); G01N 15/1459 (20130101); G01N
2015/149 (20130101); G01N 2015/1486 (20130101); G01N
2015/1406 (20130101); G01N 2015/1093 (20130101); B01J
2219/005 (20130101); G01N 2015/1037 (20130101); G01N
2015/1062 (20130101); G01N 2015/1477 (20130101); G01N
2015/1081 (20130101) |
Current International
Class: |
G01N
15/10 (20060101); G01N 15/12 (20060101); G01N
15/14 (20060101); B03b 001/00 () |
Field of
Search: |
;209/3,4,111.7,111.8,75,111.5 ;210/65 ;324/71PC,34,61 ;356/39 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Schacher; Richard A.
Assistant Examiner: Church; Gene A.
Claims
What we claim is:
1. An apparatus for rapidly analyzing and sorting minute particles
on the basis of preselected characteristics or combinations of
preselected characteristics which comprises
a. a flow chamber,
b. a high-intensity light source,
c. means for introducing particles in suspension in a fluid into
said flow chamber,
d. multiple sensing means for detecting preselected physical and
chemical characteristics of said particles in said flow chamber and
producing analog electrical quantities related to said
characteristics,
e. means for comparing said analog electrical quantities with
preselected standards for said characteristics or combination of
said characteristics and producing an electrical sort signal when
said quantities or combinations of said quantities are outside said
preselected standards,
f. means for jetting said fluid from said flow chamber,
g. means for periodically disturbing the jet or produce uniformly
sized droplets sufficiently small that substantially each particle
is isolated in a single droplet,
h. electrical charging means adjacent to the jet path at the
droplet separation zone,
i. electrical delay means whereby said sorting signal activates
said electrical charging means at a time when a particle having
characteristics outside said preselected standards is in said
droplet separation zone, said charging means remaining inactivated
unless said sorting signal is received, and
j. electrical deflecting means whereby charged droplets are
deflected to a separate receptacle from that for uncharged
droplets.
2. The apparatus of claim 1 wherein said preselected chemical and
physical characteristics comprise small angle light scatter,
fluorescence and volume.
3. The apparatus of claim 1 wherein said flow chamber contains
a. means whereby said particles are made to pass along a narrow
stream of fluid, said stream passing through a first region having
a Coulter volume sensing orifice wherein the change of impedance
produced by the passage of each particle is measured and a second
region, termed a viewing region, wherein said stream intersects a
beam of light from said high-intensity light source,
b. an access port whereby said beam of light enters said viewing
region, and
c. multiple viewing ports whereby optical properties of said
particles may be viewed and measured, and said means for jetting
said fluid is a nozzle at the base of said flow chamber.
4. The apparatus of claim 3 wherein said high-intensity light
source is a laser.
5. The apparatus of claim 3 having means for delaying the analog
electrical signal produced by the passage of each particle through
said volume sensing orifice and correlating it with the analog
electrical signals produced for that same particle by the optical
sensing means.
6. The apparatus of claim 5 wherein the optical sensing means
consist of a photodiode and a photomultiplier tube.
7. The apparatus of claim 3 wherein said means for causing said
particles to pass along a narrow stream of fluid consists of
a. a sample inlet tube for introducing said particles in suspension
in said fluid into said flow chamber, said sample inlet tube
extending substantially into said flow chamber, and
b. a first sheath liquid inlet tube concentrically surrounding said
sample inlet tube and extending somewhat beyond it into said flow
chamber, said first sheath liquid inlet tube having a nozzle at its
lower end in which is located said volume sensing orifice, whereby
said particles on leaving said sample inlet tube are surrounded by
a coaxial laminar flow of sheath liquid, said sheath liquid having
a velocity sufficiently high to narrow the flow of particles in
suspension to a stream of desired diameter and surround said stream
substantially coaxially as it passes through said volume sensing
orifice.
8. The apparatus of claim 7 wherein said viewing region is
surrounded by a reservoir of sheath liquid, said reservoir
extending substantially above said viewing region.
9. The apparatus of claim 8 wherein said reservoir is fed by a
second sheath liquid inlet tube.
10. The apparatus of claim 9 wherein said second sheath liquid
inlet tube extends to near the base of said reservoir, and said
reservoir has a flushing outlet tube as its top whereby gases
produced in said reservoir may be periodically or continuously
flushed therefrom.
11. The apparatus of claim 10 wherein said sample inlet tube serves
as one electrode for said volume sensing orifice and said second
sheath liquid inlet tube serves as the second electrode for said
orifice.
12. The apparatus of claim 9 wherein means are provided for
controlling the pressures of liquids entering said flow chamber
through said sample inlet tube, said first sheath liquid inlet
tube, and second sheath liquid inlet tube.
13. The apparatus of claim 12 wherein said pressures are controlled
differentially.
14. The apparatus of claim 9 wherein said means for periodically
disturbing said jet consists of a piezoelectric crystal coupled to
said flow chamber and oscillated at a desired frequency.
15. The apparatus of claim 9 wherein said nozzle in the base of
said flow chamber extends into said reservoir to near the plane of
said viewing region, and means are provided for aligning said
volume sensing orifice with the orifice in said nozzle.
16. The apparatus of claim 15 wherein said alignment means consists
of a plurality of adjusting screws uniformly spaced around said
first sheath liquid inlet outside said flow chamber whereby said
first sheath inlet tube is rotated about an axis located partially
within said reservoir.
17. An apparatus for rapidly sorting biological cells failing to
meet preset standards of normality from cells meeting such
standards by imparting an electrical charge to fluid droplets
containing the abnormal cells and passing the charged droplets
through a static electrical field whereby the charged droplets are
deflected into a separate receptacle from that of uncharged
droplets containing normal cells, which comprises in
combination
a. means for introducing a suspension of cells in fluid into a flow
chamber,
b. a flow chamber containing (1) means for causing said cells to
pass substantially singly and in spaced relationship along a narrow
stream, said stream having a first region wherein cell volume is
measured and a second region, termed a viewing region, wherein a
beam of laser light intersects the flow of cells in suspension, (2)
an access port whereby said beam of laser light enters said viewing
region, (3) a first viewing port whereby the laser light, after
intersecting said flow of cells may be viewed and measured, (4) a
second viewing port at a substantial angle to said first viewing
port whereby light given off by the excitation of said laser light
of fluorescent dyes or stains attached to a cell or constituents of
a cell may be viewed and measured, and (5) a nozzle whereby said
cells in suspension are jetted into air,
c. means for generating a first electrical signal proportional to
the volume of each cell as it passes through said first region of
said channel in said flow chamber,
d. laser means for illuminating each cell as it passes through said
second region of said channel,
e. means for measuring the scatter in the laser light as it emerges
from said first viewing port and generating a second electrical
signal proportional to the amount of scatter,
f. means for collecting and measuring the fluorescent light emitted
through said second viewing port and generating a third electrical
signal proportional to the amount of fluorescent light,
g. means for delaying said first electrical signal and correlating
said signal for each cell with said second and third signals
produced by that same cell,
h. means for comparing said signals or combinations of said signals
with predetermined value ranges for such signals for cells
considered to be normal, and producing a sort electrical signal if
any of said first, second, and third signals or combinations of
said signals are outside said predetermined value ranges,
i. means for periodically disturbing the jet of fluid emerging from
said flow chamber to produce uniformly sized droplets sufficiently
small that substantially each cell is isolated in a single
droplet,
j. electrical charging means adjacent to the jet path at the
droplet separation zone,
k. electrical delay means whereby said sort signal is used to
activate the electrical charging means adjacent to the jet path at
a time when the cell determined to be abnormal is in the droplet
separation zone, said charging means remaining inactivated unless
said sort signal is received, and
l. electrical deflecting means whereby charged droplets are
deflected to a separate receptacle from that for uncharged
droplets.
Description
BACKGROUND OF THE INVENTION
The invention described herein was made in the course of, or under,
a contract with the U.S. ATOMIC ENERGY COMMISSION. It relates to an
apparatus for automatic minute particle analysis and sorting and
more particularly to an apparatus wherein the volume, shape, and
fluorescence of biological cells in suspension in a continuously
flowing fluid are rapidly and automatically measured and analyzed
to determine if the cells appear to be normal or abnormal, and
cells indicated to be abnormal are physically separated from their
normal counterparts.
In cytology there is an increasing demand for automated cell
analysis and differentiation. Presently, the screening of
cytological material, e.g., for the detection of cancerous or
malignant cells, is typically done by a hierarchy of two or more
levels of screening. Initially, cell samples are prescreened
visually by an observer to search out those that appear to contain
abnormal cells. These are then set aside for later examination by a
trained cytotechnologist or pathologist who makes the final
judgment as to whether the cells are indeed cancerous. Although
this method presently works well, it has a number of disadvantages.
It is slow, requires considerable technician time, thus making it
costly, and is nonquantitative in that the criteria of abnormality
used are largely subjective. Because of the time and cost, it is
difficult to apply it to very large populations. Moreover, many,
perhaps most, of the cellular specimens submitted to the medical
laboratory are normal. For example, in cytologic examination for
uterine cervical carcinoma, 98 percent of the women examined do not
have cancer. The net result of this-- as larger populations are
examined for cancer--is to lower the level of alertness and
interest of those that must do the prescreening. This, in time,
results in a test that becomes less quantitative and more costly as
personnel turnover increases.
The art reveals that many of these disadvantages could be overcome
by application of flow systems methods of cell analysis to the
prescreening process. Flow systems analysis allows observation of
individual cells as they flow in suspension sequentially through a
small detection volume. Large numbers of cells can be observed in
short time periods and rapid automatic prescreening procedures
developed. Common parameters used are light absorption,
fluorescence, or scatter, or volume of the observed particles.
While the literature reveals various claims that these parameters
have been observed quantitatively, a primary difficulty is that a
single parameter is frequently insufficient to differentiate
quantitatively between normal and abnormal cells. Multiparameter
analysis increases the ability to distinguish among different types
of cells. Additionally, because the majority of the cells observed
are normal, it is highly desirably that means be provided to sort
abnormal from normal cells so that the sample provided for later
screening consists of a preponderance of cells believed to be
abnormal. These various considerations and the present state of the
art are set forth in considerable detail in Part A of "Automated
Cytology: A Symposium by Correspondence," Acta Cytologica, Vol. 15,
Nos. 1-3 (1971).
In U.S. Pat. No. 3,380,584, one of the present inventors (Fulwyler)
discloses an apparatus for sorting minute particles suspended in a
fluid. Sorting is accomplished in accordance with a selected
parameter which may be size, volume, presence of radioactivity,
color, fluorescence, light absorption, or any quantity capable of
being translated into an electrical quantity. The particle
separator disclosed in that patent, however, is based on single
rather than multiparameter measurement.
Only one apparatus for sorting abnormal cells from large
populations of normal cells on the basis of multiparameter analysis
is known in the art. Kamentsky and Melamed, in Science, Vol. 156,
p. 1364 (1967) reveal a spectrophotometric cell sorter which
physically separates cells of predetermined optical properties from
large populations of cells in suspension. The sorting is done on
the basis of multiple optical measurements, and the separation
system depends on fluid switching principles popular about 1964 for
computer design. This spectrophotometric cell sorter has the
disadvantages of being relatively slow and of being unable to
provide a sample consisting primarily of the cells sought to be
further screened. For example, the best efforts with this cell
sorter produce final concentrations of the selected (i.e., abnormal
as opposed to certain preset standards of normality) cells of about
1:5 from initial concentrations in the range of 1:10,000.
The art teaches that performance of cell volume sensing instruments
employing the principle of the Coulter counter in which a cell
changes the impedance of a narrow orifice as it passes through an
orifice can be improved if the cell suspension is surrounded by a
coaxial flow of cell-free liquid as it passes through the orifice.
Thus, for example, Merrill et al., in Rev. Sci. Instru. Vol. 42, p.
1157 (1971) reveal an improved cell volume analyzer with a coaxial
flow of the cell suspension inside a sheath of cell-free solution
through the sensing orifice. This apparatus, however, is not a cell
sorter and operates as a single parameter analyzer. Although
Merrill et al. suggest that it may be used for multiparameter
analysis, the art does not reveal that it has been so used.
SUMMARY OF THE INVENTION
Using a high-speed flow system and electronic and optical sensing,
we have developed an apparatus for rapidly and automatically
analyzing and sorting minute particles on the basis of certain
preselected characteristics or combinations of these
characteristics. The apparatus is an outgrowth of that disclosed in
U.S. Pat. No. 3,380,584 and allows particle separation on the basis
of multiparameter analysis. It is particularly applicable to the
analysis and sorting of biological cells.
In one embodiment of the apparatus useful for sorting abnormal
(malignant) cells from normal cells, cellular volume, small-angle
light scatter, and fluorescence are measured for each cell and
compared with preset standards, and cells failing to meet these
standards are separated from cells conforming to the standards.
Cell samples stained with an appropriate fluorescent dye are
diluted and suspended in physiological saline solution and
introduced into a flow chamber on the axis of a moving stream of
saline solution which acts as a sheath to confine the cell stream
to the central axis of the system. Within the chamber, cells flow
sequentially through an orifice which serves as a Coulter volume
sensor wherein cell volume is electronically measured. The cells
flowing in suspension in the saline solution next intersect an
argon-ion laser beam. The individual cells scatter light and the
dye bound to the cell is excited to fluoresce. The scattered light
provides quantitative information on cell size and shape, and the
fluorescence is a quantitative measure of any cell constituents to
which a fluorescent dye is bound, e.g., DNA content. Small-angle
light scatter is measured in the forward direction and fluorescence
perpendicular to the cell stream and the laser beam. After passing
through the laser beam the cell suspension jets out into air
through a coaxially aligned nozzle at the exit end of the flow
chamber. A piezoelectric crystal mechanically coupled to the flow
chamber is used to produce uniform droplets by regularly disturbing
the emerging liquid jet. Most cells are effectively isolated into
single droplets although not all droplets contain cells and certain
droplets may contain two or more cells. Droplets containing
selected cells are electrically charged and then deflected into a
separate receptacle by a static electric field. An oscilloscope
monitors individual signal pulses while a multichannel pulse-height
analyzer, printer, and plotter provide and record pulse amplitude
distributions representative of cell volume, light scatter, and
fluorescence or combinations of these characteristics. A variable
delay pulse generator triggered by a single-channel pulse-height
analyzer produces droplet charging pulses which are delayed to
allow the cell being sorted to travel from the sensing region to
the point of droplet formation and charging.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a block diagram showing the manner in which the apparatus
of this invention may be used in a cancer screening program.
FIG. 2 is a simplified view of the apparatus showing the flow
chamber and the charging and deflection plates used to achieve
particle sorting.
FIG. 3 is an enlarged simplified cut-a-way view of the sensing
portion of the flow chamber.
FIG. 4 is a detailed cross-sectional view of the flow chamber
useful with a preferred embodiment of the invention.
FIG. 5 is a block diagram of the optical and electrical elements of
a preferred embodiment of the invention.
FIG. 6 is a portion of a logic and switching block diagram for the
multiparameter signal processing unit indicated in FIG. 5.
FIG. 7 is a continuation of the diagram of FIG. 6.
GENERAL DESCRIPTION
The apparatus of this invention may readily be used for rapid and
automatic multiparameter analysis and sorting of various types of
particles. The size of the particles analyzed is limited by the
size of the Coulter volume sensing orifice. It will be apparent
that a limitation on the type of particles that can be analyzed and
sorted by this apparatus is that the particles be capable of
analysis on the basis of their physical and chemical
properties.
The figures within this specification are directed toward an
embodiment of this invention useful in the analysis and sorting of
abnormal from normal cells in a cytological screening program for
the determination of cervical cancer. The scheme is outlined in
FIG. 1. Cell samples are prepared for flow system analysis by
appropriate dilution, treatment to avoid clumping, staining with
fluorescent dyes, etc., as required for the particular form of
automated analysis to be used. In this particular scheme, the
cellular parameters measured are cell volume, small-angle light
scatter, and fluorescence. The fluorescence measurements depend on
the use of biochemically specific stains. Sensors to make these
particular measurements are compatible with each other and with
electronic sorting of cells. The electronic sorting technique is
similar to that described in U.S. Pat. No. 3,380,584. As each cell
is analyzed, a signal from each sensor is transmitted to a
multiparameter signal processing unit, processed, and compared with
predetermined criteria of abnormality. Thus, while the cell is
still in the vicinity of the sensing region, the signals obtained
from the sensors and representing measured cell characteristics are
processed to yield ratios, overlapping ranges, etc., which most
effectively describe abnormal cells. The processed signals are
electronically compared with specified standards, and the
corresponding cell is designated as normal, abnormal, or ambiguous.
Once the signals have been obtained, the time required for signal
processing and the sorting decision is on the order of 25 .mu.sec.
Classification of a cell as abnormal or ambiguous produces a signal
causing a droplet containing that cell to be deflected away from
the droplets containing normal cells. Results of analysis of this
cell may be stored separately from data for normal cells of the
sample. Sorted abnormal or ambiguous cells are counterstained and
held for visual examination by a cytologist. To aid his evaluation
of the sorted cells, distributions of the various measured cellular
characteristics or combinations of the characteristics of the
entire sample or only the abnormal cells under examination are
available from processed data storage. The apparatus of this
invention provides both printouts and an oscilloscope display of
the data.
Although the specific embodiment disclosed herein is based on the
multiparameter analysis of the volume, fluorescence, and
small-angle light scatter of individual cells, it will be readily
apparent to one of reasonable skill in the art that the analytical
and sorting techniques embodied in this invention are readily
applicable to other forms of high-speed sensing, and that the
electronic and mechanical components of the embodiment described
may readily be altered to allow for the measurement of other
parameters. For example, using the flow chamber described herein,
the small-angle light scattering sensing may be replaced with
sensors capable of detecting light absorption or fluorescence at an
additional wavelength.
DESCRIPTION OF THE PREFERRED EMBODIMENT
FIGS. 2 and 3 illustrate the basic flow system and sensing region
of the apparatus of this invention. An appropriately prepared cell
sample is introduced as a continuously flowing suspension into flow
chamber 3 through sample entry tube 1 from pressurized reservoir
72. Within chamber 3 tube 1 is centered and extends partially
through a larger tube 19 which tapers to a nozzle 22 at its lower
end. A continuous flow of cell-free liquid, known as a sheath
liquid, is introduced into tube 19 through sheath input tube 18
from pressurized reservoir 70 and flows coaxially 20 around tube 1.
As the cell stream exits from tube 1 it is reduced in diameter 17
as it obtains the velocity of the sheath liquid. Relative
velocities and flow rates are determined by a differential pressure
regulator system. It is necessary that the coaxial flow of sheath
liquid and cells in suspension be essentially laminar in nature to
avoid turbulence effects as the sheath liquid and cell suspension
pass through volume sensing orifice 21 in nozzle 22. The pressure
differential between sheath liquid 20 and the cell stream is
adjusted to provide a cell stream 28 through orifice 21 having a
diameter such that most cells pass through one at a time. Orifice
21 serves as a Coulter volume sensing orifice in which the
impedance is changed in accordance with the volume of the cell
passing through. The laminar flowing sheath liquid acts not only to
control the size of the cell stream passing through orifice 21 but
also to center it within the orifice, thus substantially reducing
electric field edge effects affecting the volume sensing.
After leaving orifice 21, cell stream 28 intersects a beam 4 of
intense light. Flow chamber 3 is provided with an entry port 35
through which this beam may be focused on the cell stream. Ports 36
(one not shown in FIG. 2) serve as exits for fluorescence 9 and
light scatter 11 produced as the result of the interaction 16 of
the light with a single cell as the cells flow sequentially in cell
stream 28. Light beam 4 intersects cell stream 28 at an angle of
approximately 90.degree.. Fluorescence is given off in all
directions but is only measured in a cone 9 extending at right
angles to the plane of intersection of cell stream 28 and light
beam 4. These 90.degree. angles are critical only insofar as they
serve to simplify the optical measurements involved.
Cell stream 28 jets from flow chamber 3 through exit nozzle 26. It
is essential that orifice 21 be properly aligned with nozzle 26.
Misalignment may result in interference with the optical
measurements because of turbulence and will affect the proper
timing by which measurements associated with a particular cell are
translated into a sorting signal because the cells lose their
sequential and equal spacing within cell stream 28. It is also
necessary that the cell stream and its surrounding sheath liquid
leave flow chamber 3 at a sufficiently high velocity to form a jet
10. Because of the pressure drop associated with orifice 21 an
additional source of sheath liquid is required to provide a
sufficient pressure in region 30 for an adequate jet 10 to form.
This second sheath liquid input 25 occurs through tube 54 which is
connected to sheath input tube 12. Tube 12 is in turn connected to
pressurized sheath liquid reservoir 71. The input 25 is
sufficiently far from cell stream 28 that it does not introduce any
effect on the cell flow aside from increasing the velocity through
nozzle 26. Sheath liquid input 25 has the additional advantage of
allowing the flow chamber to be flushed of accumulated gases.
Because the volume sensing involves the use of electrical current
there is a tendency for electrolytic dissociation to occur in the
liquids present in flow chamber 3. The dimensions of the flow
chamber are sufficiently large that the gas bubbles resulting from
such dissociation rise to the top of the flow chamber and may be
temporarily stored there without disrupting the optical or
electrical sensing. However, it is desirable that some means be
present to periodically flush any collected gases from the system.
This is readily accomplished by means of input tube 12. When
flushing is required, a valve in flushing outlet tube 2 is opened
and additional sheath liquid is added to flow chamber 3 through
inlet tube 12.
A vibration means 31 is coupled to flow chamber 3 by means of
coupling rod 32. Vibrations imparted to flow chamber 3 produce
minute disturbances or bunching 29 on jet 10. By producing these
disturbances at a proper frequency, determined by the jet diameter
and velocity, they are made to grow in amplitude by surface tension
until jet 10 is broken into evenly spaced, uniformly sized droplets
13. In this way cells in suspension are isolated into liquid
droplets. The manner in which vibration means 31 is coupled to flow
chamber 3 is not critical except that the vibration frequency must
be kept relatively constant. Typically, a piezoelectric crystal is
used as the vibration source, but other means may readily be
used.
The droplets 13 thus produced pass between charging electrodes 5
where those droplets determined to contain abnormal cells on the
basis of an analysis of the volume, light scatter, and fluorescence
of each cell are charged. Droplets indicated to contain normal
cells are not charged. This particular sequence is followed on the
assumption that most droplets will contain normal cells and hence
the simplest approach is to charge only the droplets containing
abnormal cells. The reverse procedure can just as readily be
adopted, however. The droplets then pass through an electrostatic
field between deflection plates 8. Under the influence of this
field charged droplets 14 are deflected into separate receptacles 7
than the uncharged particles 15.
FIG. 4 is a detailed cross-sectional view of a flow chamber adapted
for use with this invention. Particles in suspension flow into the
chamber through tube 1. This tube, which also serves as one
electrode for the Coulter volume sensor, may be formed from any
suitably conductive material. In the preferred embodiment this tube
is formed from platinum-rhodium alloy to avoid corrosion problems
resulting from the use of physiological saline solution as both the
carrier liquid for biological cells and the sheath liquid
surrounding the carrier liquid. Tube 1 is aligned coaxially within
a larger tube 19 by means of a guiding star 46 and is provided with
a nozzle 47. This nozzle may be composed of platinum, but there is
no requirement that this be so. A noncorrosive material such as an
appropriate plastic will serve as well. Additionally, guiding star
46 and nozzle 47 may be combined and formed of the same material.
Tube 19 ends in a nozzle 22 in which Coulter volume sensing orifice
21 is emplaced. Tube 19 may be of any nonconductive material such
as glass. In the preferred embodiment, volume sensing orifice 21 is
in a sapphire insert 34 bonded to a glass tube 19. Sapphire is used
because of its ready availability, but insert 34 may consist of any
suitable nonconductive material in which an appropriate orifice can
be produced.
Sheath liquid is introduced into tube 19 through inlet tube 18.
Sheath liquid also enters cuvette 23 through inlet tube 12 and
flushing tube 54. Flushing tube 54 also serves as the second
electrode for the Coulter volume sensor so that both tube 54 and
tube 12 must be of suitably conducting material. In the preferred
embodiment, tube 12 is composed of platinum and tube 54 of
platinum-rhodium alloy. Flushing outlet tube 2 connecting to the
interior region 30 of cuvette 23 is provided for the removal of any
gases that may be produced in region 30. In the preferred
embodiment, outlet tube 2 is provided with a valve; however, if
desired, sheath liquid may be continuously flowed from flushing
tube 54 through cuvette 23 and out outlet tube 2.
Cuvette 23 which is open at the top, may be made of any material
that is an insulator and is transparent to light at the wavelengths
used for light beam 4. In the preferred embodiment, cuvette 23 is
composed of quartz, primarily because a quartz cuvette of the
desired size is readily available commercially. Centered in the
base of cuvette 23 is an opening 58 through which nozzle 26
extends. To minimize wall effects on the velocity of the cells
being jetted from nozzle 26, it extends nearly to plane A-A in
which the optical measurements on the individual cells are made.
This extension also allows volume sensing orifice 21 to be more
easily aligned with orifice 59 in nozzle 26.
Cuvette 23 is surrounded by a body shell 24 of metal. Shell 24
serves to protect cuvette 23 and also to shield the Coulter volume
sensor from outside electronic noise. At the base of shell 24 is
sealing end plate 56 and nozzle retaining end plate 57. End plate
56 has a circular opening 60 centered in it through which the
extended portion of nozzle 26 is passed. Nozzle 26 is secured by
threading it into central opening 61 in end plate 57.
Attached to the upper portion of body shell 24 is draw nut 55.
Threaded into draw nut 55 is enclosure cap 52. Cuvette 23 is sealed
closed within shell 24 and cap 52 by means of O-ring 48 located in
well 62 in end plate 56 and gasket 51 adjacent to cap 52. Enclosure
cap 52 extends partially into cuvette 23 to form well 63. At the
base of well 63 is centered a circular opening 64 through which
tube 19 extends. Angular adjusting seal gland 44 threads into well
63 until it is flush with the top of enclosure cap 52. Gland 44 has
centered within it truncated conical shaped well 65 having lip 66
near its top. Well 65 has a circular opening 67 centered in its
base through which tube 19 passes.
Around the upper end of tube 19 and attached to it is sealing
collar 43. Atop sealing collar 43 and tube 19 is tube connector and
positioner 41. Positioner 41 has a channel 68 through it by which
tube 1 enters tube 19 and is coaxially aligned in the upper portion
of tube 19. Sheath liquid inlet tube 18 also enters positioner 41
and by means of channel 69 passes sheath liquid into tube 49. Inlet
tubes 1 and 18 are surrounded by a shielding tube 40 which serves
to prevent the introduction of electrical noise into the flow
chamber through either tube 1 or tube 18. Shielding tube 40 and
positioner 41 are held in place on top of tube 19 by means of
connector compression cap 42 which threads onto sealing collar 43.
O-ring 50 provides a seal between collar 43 and positioner 41. With
tube 19 inserted, wells 65 and 67 are sealed from cuvette 23 by
O-rings 70. The seal of these O-rings can be adjusted by screwing
seal gland 44 further in or out of enclosure cap 52 thus allowing
tube 19 to be moved in or out of cuvette 23 as desired. Located
equidistant around seal gland 44 are four adjusting lock screws 45
(only one of which is shown in FIG. 4). These lock screws 45
provide a ready means by which orifice 21 in the end of tube 19 may
be aligned with orifice 59 in nozzle 26. As indicated earlier in
this specification, it is essential to the proper functioning of
this apparatus that these two orifices by accurately aligned.
A block diagram of the electrical and optical elements of an
embodiment of this invention useful in the rapid, automatic
analysis and sorting of abnormal from normal cells is given in FIG.
5. A piezoelectric crystal coupled to the flow chamber serves as
the source of vibration for producing droplets at the desired
frequency. The light source used in this embodiment is an argon-ion
laser which is appropriately focused to intersect the cell stream
in the flow chamber. The light beam entering the flow chamber
normally has an elliptical cross section to aid in the analysis of
the cell structure by means of the resultant light scatter and
fluorescence. The laser beam is optically shaped such that it has
an elliptical cross section as the laser beam-cell stream
intersection. The elliptical cross section improves ease in
operation (alignment), increases signal strength improving
resolution, and allows characterization of cell structure
(nuclear-to-cytoplasm ratio) and doublet discrimination. Doublets
are two cells passing through the flow chamber in contact with each
other. To the volume sensor they appear as one abnormally large
cell. To avoid receiving erroneous data from the sensors, they must
therefore somehow be discriminated against.
An argon-ion laser is used as the light source because cancerous
cells usually contain substantially more DNA than do normal cells,
and the fluorescence of an excited Feulgen dye biochemically bound
to the DNA in a cell is a quantitative indication of the DNA
present in the cell. The argon-ion laser emits light at a
wavelength suitable for exciting this dye to fluoresce. Pulse of
fluorescence coming from the flow chamber as the result of the
interaction of the light beam with the cells are focused by a lens
system on a pinhole and then onto a photomultiplier tube. The
signal from this photomultiplier tube is amplified and fed into a
multiparameter signal processing unit.
Theory predicts that small angle light scatter (at angles between
0.5.degree. and 2.0.degree.) by spherical particles of 5 to 20
microns diameter is nearly proportional to volume. Since most
mammalian cells have diameters in this range, small-angle light
scattering is attractive as a means of obtaining size and
structural information for single cells at high speed. Thus in the
preferred embodiment light scattered between 0.5.degree. and
2.0.degree. by the cells is passed through a collecting lens system
and into a photodiode. Light scattered less than 0.5.degree. is
passed to a beam dump. This avoids having the photodiode
overwhelmed by light that has not interacted with the cell stream.
The photodiode signal is amplified and also fed into the
multiparameter signal processing unit.
The signal produced by the passage of individual cells through the
Coulter volume sensing orifice has the advantage of already being
electrical in nature so that all that is required of it is to
amplify it and feed it also into the multiparameter signal
processing unit.
As its name indicates, the multiparameter signal processing unit
processes these input signals, compares them with certain preset
standards, and then provides three types of output signal. One
signal is transmitted to a multichannel pulse height analyzer which
in turn provides digital printouts, a pulse height analyzer
display, and histograms of the data obtained from the
multiparameter signal processing. The signals from the processing
unit may also be directly monitored by means of an oscilloscope
display. Finally, an output from the processing unit is passed
through a single channel analyzer and separation logics and droplet
charging generator to provide droplet charging pulses which act to
separate selected cells from the cell stream. It is apparent that
time delay means are used in conjunction with the multiparameter
signal processing unit to coordinate all sensor signals with a
particular cell.
The signals received by the multiparameter processing unit can be
processed in various ways to modify their dependence on the
measured property. For example, a signal proportional to cell
volume (r.sup.3) can be processed to make it linearly proportional
to cell radius (r) or to area (r.sup.2). Because only one piece of
data is produced by each sensor for each cell, the amount of
information to be processed is small, and the requirements placed
on the electronics are not great. By using a two-dimensional
pulse-height analyzer, a two parameter frequency distribution of
cells can be obtained. Three- or more parameter analysis requires
the data capacity of a small computer. As an alternative to storage
of all information, logical restrictions can be imposed on the
analysis scheme, thus lessening the electronics requirement. For
example, in the fluorescence distribution of all cells within a
certain volume range is desired, this can be obtained by analyzing
the fluorescence of only those cells which produce a volume signal
corresponding to the desired range. Likewise, the volume of cells
within a certain fluorescence range can be obtained. If
biologically useful information is provided, analysis is possible
on the basis of three or more such logical requirements.
Alternatively, the processed signals from several sensors can be
combined as ratios (or sums, differences, etc.) and the frequency
distribution of the combination determined among a population. For
example, by using a RNA-specific fluorescent stain and sensors for
fluorescence and volume, a ratio of the processed signals can be
formed to obtain a distribution of RNA density among a population
of cells. Likewise, a type of nuclear-to-cytoplasmic ratio is given
by using a nucleus-specific fluorescent stain to give a measure of
nuclear volume and total cell volume measurement by scattered light
or by the Coulter sensor.
A logic and switching block diagram for a multiparameter signal
processing unit useful in the analysis and sorting of abnormal from
normal cells is given in FIGS. 6 and 7. The multiparameter signal
processing unit is the central electronic processing unit for
single parameter analysis, ratio computing, serial or sequential
analysis and subsequent cell sorting. Basically, the signal
processing unit serves as a central analog electronics computing
interface between the sensors, the multichannel pulse height
analyzer, and the cell separator control (single channel analyzer,
separation logics, droplet charging generator). The signal
processing unit also provides x and y outputs for a dual parameter
pulse height analyzer. Amplified signal (0.4 to 8.0 V) pulses from
the Coulter or cell volume (CV), light scatter (LS), and
fluorescence (FL) sensors are fed directly to the processing unit.
The unit has separate inputs for the volume, light scatter and
fluorescence signals. If desired, a second wavelength fluorescence
signal, such as red, may be substituted for the light scatter
input, or wide angle light scatter substituted for the fluorescence
input.
The signal processing unit serves as a
gated-signal-peak-sense-and-hold device capable of both single and
dual processing of the sensor signals. The processing unit is
divided into three sections: input conditions, signal processing,
and output routing. The Input Condition section (shown in FIG. 6)
consists of Operation Mode and CV-FL/LS Delay selector switches.
The eight position operation mode switch allows signal parameter
cellular analysis, i.e., CV, LS, and FL, and dual parameter
analysis of cells, i.e., CV and FL, CV and no FL, CV and LS, LS and
FL, and LS and no FL. Since the Coulter volume signal arrives prior
to either the light scatter of fluorescence signals, a variable
(0-190 .mu.sec in steps of 10 .mu.sec) CV-FL/LS Delay switch is
used to set the proper CV to FL/LS signal delay. This delay is on
the order of 170 .mu.sec. The delay need be only used in the dual
parameter analysis mode when Coulter volume is to be analyzed.
The Signal Processing section consists of Ratio and Serial (Input
and Analyze) Analysis selector switches (see FIG. 7). A six
position ratio selector switch allows the following ratios to be
computed: CV/FL, CV/LS, FL/CV, FL/LS, LS/CV, and LS/FL. It is
mandatory that the operation mode and ratio selector switch
coincide, e.g., operation mode selector in the LS and FL position
and ratio selector in either the LS/FL or FL/LS ratio position. It
is also important that the CV to FL/LS signal delay be used if
ratios containing CV are to be computed. The serial or sequential
analysis section consists of input and analyze selector switches
(four positions each). The serial analysis input selector switch
selects either CV, FL, LS, or Ratio signals to be inputed to an
external single channel pulse height analyzer (Serial SCA). If the
signal amplitude falls within a variable width (0.4 to 8.0 V)
preset SCA window, and SCA trigger pulse is produced and is
returned to the signal processing unit gating on the serial
analysis analyze linear gate, thus allowing either the CV, FL, LS,
or Ratio signal to be analyzed, as determined by the Serial
analysis analyze selector switch. Both the Serial analysis input
and analyze selector positions (CV, FL, LS, and Ratio) must
correspond to the operation mode, CV-FL/LS delay, and ratio
selector switches whenever required, e.g., Serial analysis input --
CV, Serial analysis analyze -- FL, Operation mode -- CV and FL,
CV-FL/LS delay .apprxeq. 180 .mu.sec, and ratio -- off.
The Output Routing section (see FIG. 7) consists of pulse height
analyzer (PHA) input, separator input, and dual parameter analyzer
input selector switches. The PHA input can select single parameters
(CV, LS, or FL), ratios, serial analysis input and analyze
parameters to be routed externally to a multichannel pulse height
analyzer, whereas, the separator input selector can route single
parameters, ratios, or serial analysis analyze signals to the
separator control. The dual parameter pulse height analyzer
selector switch provides outputs of CV.sub.x -FL.sub.y, CV.sub.x
-LS.sub.y, LS.sub.x -FL.sub.y, CV.sub.x -Ratio.sub.y, FL.sub.x
-Ratio.sub.y, and LS.sub.x -Ratio.sub. y, where the x and y
subscripts refer to the x and y axes of the dual parameter PHA. By
interchanging the x and y axis inputs the above can be
inverted.
OPERATIONAL SEQUENCE
Cell populations to be tested are first stained (Fluorescent
Feulgen, etc.) are placed in aqueous suspension, such as normal
saline. Fixed or unfixed cells can be measured. Prior to placing
the cell suspension in the cell reservoir, it is filtered through a
60-70 micron nylon mesh screen to remove large debris and clumps.
The electronics are in a standby condition, system pressurized, the
laser turned on. The system is aligned and adjusted prior to cell
measurements. If cell sorting is desired, the droplet generator
oscillator-amplifier which electrically drives the piezoelectric
crystal (or equivalent) transducer must be turned on. Droplet
formation is checked by illuminating the emerging liquid jet near
the flow chamber with a strobe light or equivalent light source.
The strobe light is synchroflashed with respect to the oscillator
frequency. Droplet formation can then be viewed using a microscope.
For a given exit nozzle diameter and flow rate, droplet formation
can be adjusted by varying the voltage and frequency applied to the
piezoelectric crystal. Typical values are 15 volts RMS (sinusoidal)
at 40 to 50 kHz. The droplet charging electrode is placed astride
the point of droplet formation (separation) about 5/16 inch below
the flow chamber to ensure maximum droplet charging. Typical
charging pulses are 50 volts for 100-200 .mu.sec. The electrostatic
deflection plates are located 2 to 3 inches below the flow chamber
and spaced about 1/2 to 3/4 inches apart. A differential of 15 kV
d.c. is normally applied to the deflection plates. A sample
collection beaker or appropriate collection system is placed 8 to 9
inches from the flow chamber exit side and is slightly offset from
the main jet (uncharged) so as to only collect the deflected
droplets (charged). If it is not desired to sort out cells, then
the above procedure can be omitted.
Suspended cells are placed in a pressurized (23.4 psi) cell
reservoir. Pressurized sheath fluid No. 1 (24.0 psi) and sheath
fluid No. 2 (20.0 psi) are turned on and proper droplet formation
achieved if sorting is desired. Sheath fluid No. 1 with no cell
stream has a flow rate of 0.3 ml/min. Sheath fluid No. 2 flow rate
is approximately 3.9 ml/min. The total flow rate exiting the 86
.mu. diameter exit nozzle is thus 4.2 ml/min. For a typical cell
stream diameter of about 20 .mu., the cell reservoir pressure of
23.4 psi corresponds to a cell stream flow rate of about 0.08
ml/min. The cell stream flow rate can easily be adjusted from 0 to
0.3 ml/min (100 percent) by adjusting the cell reservoir pressure
relative to sheath No. 2 reservoir pressure (.+-. 0.2 psi), holding
sheath No. 1 reservoir pressure fixed. Sheath No. 1 pressure
relative to sheath No. 2 pressure normally remains fixed, but can
be varied if desired. Increasing sheath No. 1 pressure relative to
sheath No. 2 decreases the transit time of cells through the flow
chamber. As the sample on/off valve is turned on, cells pass from
the cell reservoir into the flow chamber via the sample inlet tube.
The inlet tube serves as the Coulter volume signal electrode. From
the inlet tube cells pass through the volume sensing orifice (75
.mu. diameter aperture) wherein cell volume is sensed. Orifices of
other sizes may readily be substituted. A particle free sheath
(sheath No. 1) flows coaxially around the sample inlet tube and
serves to centrally align the cell stream as it passes through the
volume sensing orifice, thus improving the resolution of cell
volume and fluorescence/light scatter measurements. Typical d.c.
aperture currents flowing from the volume signal electrode through
the orifice can be adjusted from 0.05 to 1.0 mA. The aperture
current and amplifier gain can be adjusted to give volume signal
pulses (0.4 to 8.0 V) which in turn are fed to the CV input of the
multiparameter signal processing unit. Typical volume signal
risetime is about 20 .mu.sec with pulse widths of 40 .mu.sec.
Upon exiting the volume sensing orifice the cells next intersect
the laser beam thereby scattering light and fluorescing. Typical
time delays between initiation of the Coulter volume and
fluorescence/light scatter pulse are in the order of 160-180
.mu.sec. Fluorescence and light scatter electro-optical pulses are
amplified (0.4 to 8.0 V) and fed to their respective inputs on the
signal processing unit. Typical risetimes are in the order of 1-2
.mu.sec with pulse widths of about 5 .mu.sec. A second particle
free sheath liquid (sheath No. 2) of normal saline serves to reduce
the effect of the pressure drop created by the Coulter sensing
orifice.
Once the properties of the cell have been measured, it exits the
flow chamber via the exit nozzle contained in a liquid droplet
which can subsequently be separated. The approximate time delay
between cell sensing and droplet formation is in the order of 1400
.mu.sec.
Signals from volume, light scatter, and fluorescence sensors and
amplifiers are thus fed to the multiparameter signal processing
unit for subsequent analysis and routing. The processing unit
serves as an interface between the multichannel pulse height
analyzer, cell separation logics and droplet charging generator,
and dual parameter pulse height analyzer (not shown in FIG. 5). The
signal processing unit must be properly set up as previously
discussed, depending upon the requirements for each experimental
run.
In a typical experiment where it is desired to analyze and sort out
abnormal cells from a given population mixture, a number of
different approaches might be meaningful. The multichannel pulse
height analyzer would be used first to display frequency
distribution histograms of single parameters (volume, light
scatter, and fluorescence) ratios of parameters, or possibly to
serially analyze parameters, e.g., analyze the fluorescence for a
given cell volume range, etc. The dual parameter analyzer could
also be used to analyze various dual parameter frequency
distribution histograms that might be needed. From either or both
the multichannel PHA and Dual Parameter PHA displays it is possible
to pick out abnormalities from various histograms, e.g., an
abnormally large nuclear-to-cell volume ratio. From this type of
information the cell separation logics-droplet charging generator
can be set up to physically sort out those cells exhibiting
questionable properties for microscopic examination and
identification. The lower and upper threshold level of the single
channel pulse height analyzer (SCA) is set to accept pulse
amplitudes (ratios, etc.) from cells exhibiting abnormal
characteristics. The SCA then triggers the droplet charging
generator which produces a delayed (1400 .mu.sec) droplet charging
pulse (50 V peak for 100-200 .mu.sec). Once the characteristics of
abnormality have been obtained it may not be necessary to sort out
the abnormal cells for screening under the microscope, but only
further automate the analysis procedure to aid in rapid disease
diagnosis.
It should be noted that the optimum operating condition would be
that in which all cells pass singly through the volume sensing
orifice and light beam and only one cell is caught in each droplet.
As a practical matter, this is most difficult to achieve. The cells
are frequently widely spaced in the cell stream such that numerous
droplets contain no cells. This presents no particular problem;
however, when two cells are in actual contact (thus forming a
doublet) or so closely spaced that the sensors cannot discriminate
between them, then sensing data are received which indicate
abnormal cells and a sort signal goes out. If, as is likely in the
usual population of cells, the doublet is composed merely of two
normal cells, this serves to dilute the purity of the sorted
sample. While very careful attention to sample preparation may
substantially reduce the presence of doublets, and the use of
discrimination techniques aid in reducing erroneous sort signals
caused by the presence of such doublets or very closely spaced
cells, this results in an increase in the time and cost of analysis
and sorting. Pragmatically, it is therefore frequently desirable to
allow a certain small percentage of doublets and closely spaced
normal cells to be sorted with abnormal cells. Although this
reduces the purity of the sorted sample, it does not greatly hinder
analysis of the sample. Typically, 90 percent or more of the cells
are isolated singly into droplets. That is to say, 90 percent or
more of the droplets containing cells have only a single cell
within them.
* * * * *