Distinguishing Modes of Cell Death Using ImageStream®
Multispectral Imaging Cytometry
Thaddeus George, Brian Hall, Cathy Zimmerman, Keith Frost, Mike Seo, David Perry,
William Ortyn, David Basiji, and Philip Morrissey
Amnis® Corporation, 2505 Third Avenue, Suite 210, Seattle, WA 98121-1480, www.amnis.com
Abstract
ImageStream technology allows for multi-spectral imaging of cells in flow. The ImageStream
100 platform can produce six simultaneous high resolution images of each cell at rates
exceeding 100 cells per second. We have exploited the unique ability of the technology to
simultaneously generate brightfield, darkfield, and fluorescence cell imagery to classify cells
undergoing apoptotic or necrotic cell death. Cells dying by apoptosis activate caspase-3 and
lose plasma membrane asymmetry and expose phosphatidylserine on their external leaflet,
which can be detected by fluorochrome-labeled annexin V. Necrotic and late apoptotic cells
become permeant to both annexin V and to 7-AAD, a cationic nucleotide-binding dye. Despite
these similarities, apoptotic and necrotic cells are morphologically distinguishable: necrotic cells
. cells appear shrunken and have fragmented
appear swollen with intact nuclei, while apoptotic
nuclei with condensed chromatin. We analyzed data acquired from a mixture of untreated
(annexin V-, 7-AAD- live), camptothecin-treated (annexin V+, 7-AAD- early apoptotic and annexin
V+, 7-AAD+ late apoptotic), and H2O2-treated (annexin V+, 7-AAD+ necrotic) tumor cells. The
annexin V+, 7-AAD+ late apoptotic and necrotic cells were readily distinguishable based on the
morphological criteria of brightfield area and darkfield texture derived from the associated
image set. These data demonstrate the powerful utility of the ImageStream technology in
settings that require associated morphologic analysis.
a.
b.
0 .3
8
6
0.2 8
4
0.2 6
75
a.
0.2 4
70
65
2
Necrotic
0
0.2 2
60
55
0.1
1
10
HLA-PE Total Int ensity * 10000
100
0.1
100
0 .2
50
0.1 8
45
40
0.1 6
35
30
0.1 4
Cells
25
c.
Late Apoptotic
1.5
0 .1
15
1
10
Development of ImageStream technology was partially supported by NIH grants 9 R44
CA01798-02, 1 R43 GM58956-01 and 1 R 43CA94590-01.
0.0 8
5
0
0.05
2.5
2
0.1 2
20
0.1
0.15
0.2
0.25
0.3
0.35
Brightfield Area * 10000
b.
0.4
0.45
0.5
0.5
0.0 6
0.55
0.1
1
488 Scatter Frequency * 10
Double Positive
Scatter Annexin V
HLA
7-AAD Brightfield
0
10
Annexin V
7-AAD
Composite
Scatter Annexin V
HLA
1
10
HLA-PE Total Intensity * 1000
7-AAD Brightfield
Annexin V
7-AAD
Composite
100
Introduction
Death by apoptosis is a complex, tightly regulated process in which a cell orchestrates its own
destruction in response to specific internal or external triggers. Physiologic apoptosis is critical
to the maintenance of tissue and immune cell development and homeostasis. Disruption of the
physiologic control of apoptosis results in pathology: inappropriately low apoptotic rates can
result in cancer or autoimmune disease, while high rates can result in neurodegenerative
disease or immunodeficiency. Because apoptosis is an actively regulated process, the
opportunity exists to modulate cell death therapeutically. As a result, intense research aimed at
cell death regulation has increasingly relied on and produced methods to identify and quantify
apoptotic cells.
d.
Early Apoptotic
1
By providing accurate, rapid, and multiparametric measurements of cellular phenotype and
function, flow cytometry has become a common tool used to study a wide variety of cell
processes. Numerous flow cytometric methods are currently available for the identification and
quantification of apoptosis. However, the probed changes are not always specific or unique to
apoptotic cells. Furthermore, flow cytometric techniques cannot directly measure the specific
morphologic changes associated with apoptosis, and it is often necessary to utilize standard
manual and subjective microscopy to confirm flow cytometric results.
Amnis has developed the ImageStream 100 flow imaging platform that simultaneously captures
brightfield, darkfield, and four channels of fluorescence imagery of cells in flow with high
sensitivity. The ability to discriminate cell types in a heterogeneous population using
morphological features is greatly enhanced with multi-mode imagery. In this study, we
demonstrate the use of the ImageStream 100 platform to discriminate apoptotic, live, and
necrotic cells using both fluorescent and morphometric parameters.
a.
b.
R3
R1
R2
e.
Live
10
1
10
Annexin V Total Intensity * 1000
100
Figure 4: Morphometric Resolution of Necrotic from Late Apoptotic Cells Using Multimode Imagery
Scatter Annexin V
HLA
7-AAD Brightfield
Annexin V
7-AAD
Composite
By plotting the brightfield area against 488 nm scatter texture (frequency) of cells within the double positive (annexin V+, 7-AAD+) gate (a), two
populations were distinguishable: cells with low area and high texture and cells with high area and low texture. Analysis of HLA-PE staining
revealed that low texture/high area cells were derived from the necrotic sample (b) while high texture/low area cells were derived from the
apoptotic sample (c). Morphologic examination of cells within the image gallery confirmed that the high area/low texture cells were necrotic
(HLA-PE+, uniform 7-AAD staining) cells (d), while low area/high texture cells were apoptotic (HLA-PE-, fragmented 7-AAD staining) cells (e).
c.
c.
R3
R1
R2
Advanced apoptotic and necrotic cells differ morphologically but cannot be distinguished based solely on annexin V and 7-AAD fluorescence
using standard flow cytometry. The morphology provided by ImageStream’s multispectral image data enables resolution of the two populations
using the IDEAS package. IDEAS automatically calculates all the standard intensity-based parameters and statistics employed in flow
cytometry, while also quantitating numerous morphological features such as cell area, perimeter, aspect ratio, texture, spot counts, and others.
Late Apoptotic
Early Apoptotic
Necrotic
Live
R3
R1
R2
d.
Figure 1: Analysis of Cell Death Using Standard Flow Cytometry
Phosphatidylserine exposure on the cell surface is an early event in the apoptotic process that can
be identified by staining with annexin V. Permeability to and nuclear staining with 7-AAD occurs
later in the apoptotic process. Necrotic cells are permeable to and stain with both reagents.
Untreated Jurkat cells (a), Jurkat cells treated with 1 µM camptothecin for 18 hrs (b) or Jurkat cells
treated with 0.3% H202 for 1 hour (c) were stained with Alexa Fluor® 488 conjugated annexin V
(Molecular Probes, Eugene, Oregon) and 7-AAD for 10 minutes at room temperature in Ca2+
containing annexin V binding buffer (Becton Dickinson Biosciences, Pharmingen, San Diego, CA).
Cells were washed twice, resuspended in 2% paraformaldehyde and analyzed for annexin V and 7AAD staining (top row) or forward and side laser light scatter (bottom row) using a FACSort flow
cytometer (Becton Dickinson). The majority of untreated cells were annexin V-, 7-AAD- (R1).
Camptothecin treatment resulted in the appearance of early (annexin V+, 7-AAD-, R2) and late
(annexin V+, 7-AAD+, R3) apoptotic cell populations. Like late apoptotic cells, the majority of
peroxide-treated necrotic cells stained with annexin V and 7-AAD (R3).
e.
Figure 3: ImageStream Recapitulation of Flow
Cytometric Analysis of Cell Death
Brightfield
Annexin V
7-AAD
Composite
Darkfield
a.
b.
In order to demonstrate the concordance of ImageStream data with
flow cytometric and microscopic data, untreated, apoptotic, and
necrotic Jurkat cells prepared as described in Figure 1 were run on an
ImageStream 100 system. In this experiment, all cells were stained
with Alexa Fluor® 488 conjugated-annexin V and 7-AAD. Necrotic cells
were stained separately with PE-conjugated anti-HLA-A,B,C before
mixing to distinguish them from advanced apoptotic cells. Each cell
was simultaneously imaged in 488 nm laser light scatter, green (500550 nm, annexin V channel), orange (550-600 nm, HLA channel), red
(600-650 nm, 7-AAD channel) fluorescence, and brightfield. In (a),
object size as measured by brightfield area (in pixels) was plotted
against object complexity as measured by laser light scatter total
intensity (in pixels) and shows the singlet gate. (B) Cells were grouped
into live, early apoptotic, or double positive populations based on total
intensities of annexin V and 7-AAD staining. With Amnis’ Image Data
Exploration and Analysis Software (IDEAS™), each data point on the
bivariate plot can be clicked to observe cell imagery and each
population gate generates a corresponding image gallery.
Representative 5-channel images along with annexin V / 7-AAD
composite images are shown for each gated population. Compared to
live cells (c), early apoptotic cells (d) appear shrunken with complex
brightfield morphology. The double positive population (e) contains
both HLA-PE- late apoptotic (with condensed, fragmented 7-AAD
nuclear staining pattern) and HLA-PE+ necrotic cells (larger cells with
uniform 7-AAD nuclear staining).
c.
Calculation of a complex morphometric feature using the IDEAS software to estimate cytoplasmic area of a cell was performed by
subtracting the nuclear area (as defined by 7-AAD staining) from the total cellular area (as defined by Brightfield area). A bivariate
plot of cells contained in the “All Cells” gate (which contained live cells, early and late apoptotic cells and necrotic cells) revealed
four distinguishable subpopulations of cells. Visual inspection of cells contained in these subpopulations confirmed that live cells to
lie in the lower right quadrant (i.e. cytoplasmic areahi, scatterlo), early apoptotic cells in the upper right quadrant (i.e. cytoplasmic
areahi, scatterhi), late apoptotic cells in the upper left quadrant (i.e. cytoplasmic arealo, scatterhi) and necrotic cells to lie in the lower
left quadrant (i.e. cytoplasmic arealo, scatterlo). Backgating of the four cell populations that had been identified using alternative
criteria confirmed their identity with live cells and early apoptotic cells shown in Figure 3B (shown in blue and green, respectively),
and with the necrotic and late apoptotic cells with cells shown in Figure 4a (shown in yellow and red, respectively).
Conclusion
Figure 2: Analysis of Cell Death Using Microscopy
Untreated (a), apoptotic (b), or necrotic (c) Jurkat cells prepared as described in Figure 1 were mixed
with an equal volume of antifade and visualized on slides using a Nikon Eclipse E600 fluorescence
microscope equipped with bandpass filters appropriate for FITC (535/40 nm) and 7-AAD (630/60 nm)
fluorescence. 7-AAD staining distinguished the fragmented nucleus of the advanced apoptotic cell
from the intact nuclei of necrotic cells (annexin V / 7-AAD composite images, center column).
Necrotic cells also exhibit markedly less darkfield intensity than apoptotic cells (right column).
Figure 5: Resolution of Live, Early and Late Apoptotic, and Necrotic Cells Using
Morphometric Features Based on Scatter Variation, Brightfield Area, and Nuclear Area
®
This study demonstrated the ability of the ImageStream 100 multispectral imaging cytometer to correctly identify distinct modes of cell
death based on fluorescent and morphologic measurements. Similar to a standard flow cytometer, the ImageStream could identify live,
early apoptotic, and late apoptotic or necrotic cells with 7-AAD and annexin V fluorescent staining. Brightfield, fluorescence and laser
scatter (or darkfield) imagery obtained with standard and ImageStream microscopy revealed the shrunken, irregular appearance of late
apoptotic cells, with condensed chromatin and fragmented nuclei. Necrotic cells appeared swollen, with intact, uniformly staining nuclei.
Using morphology-based measurements, the ImageStream platform correctly distinguished necrotic from late apoptotic cells in a mixed
sample: the smaller size (indicated by lower brightfield area) and increased texture (as indicated by high laser scatter frequency)
distinguished 7-AAD+ apoptotic cells from necrotic cells (high brightfield area and low laser scatter frequency). Using scatter and a
complex feature derived from brightfield and nuclear areas, we could distinguish all four types of cells without using the annexin V dead
cell marker. Furthermore, the fidelity of these classifiers could be validated by inspection of image galleries corresponding to the gated
populations. By providing morphology-based information in addition to fluorescence and laser scatter signals, the ImageStream platform
allows for discrimination of cell types not feasible with standard flow cytometry.