Live-Cell

Analysis

HandbookA

Guide

to

Real-Time

Live-Cell

Imaging

and

AnalysisFifth

EditionTable

of

ContentsTable

of

ContentsNext

ChapterNext

Section1.

Introducing

Real-Time

Live-Cell

Analysis2.

From

Images

to

Answers6.

Kinetic

Cell

Migration

and

Invasion

Assays6a.

Kinetic

Scratch

Wound

Assays

6b.Kinetic

Chemotaxis

Assays5.

Kinetic

Assays

for

Studying

Immune

Cell

Models5a.

Kinetic

Assays

for

Immune

Cell

Activation

and

Proliferation5b.Kinetic

Assays

for

Immune

Cell

Killing5c.

KineticAssays

for

NETosis5d.

Kinetic

Assays

for

Immune

Cell

Differentiation,Phagocytosis

and

Efferocytosis8.

Kinetic

Assays

for

Utilizing

Complex

Models8a.

Kinetic

Multi-Spheroid

Assays8b.

Kinetic

Single

Spheroid

Assays8c.

Single

Spheroid

Invasion

Assay

8d.

Embedded

Organoid

Assay9.

Kinetic

Assays

for

Studying

Neuronal

Models9a.

Kinetic

NeuriteAnalysis

Assays9b.

Kinetic

Neuronal

ActivityAssays9c.

Kinetic

NeuroimmuneAssays3.

Cell

Culture

Quality

Control

Assays3a.

2D

Cell

Culture

Quality

Control

3b.

OrganoidCulture

Quality

Control4.

Kinetic

Cell

Health,

Proliferation

and

Viability

Assays4a.

Kinetic

Proliferation

Assays4b.Kinetic

Apoptosis

Assays4c.

Kinetic

Cytotoxicity

Assays4d.

Kinetic

Assays

forStudying

Cell

Cycle4e.

Kinetic

Assays

for

Monitoring

ATPand

Mitochondrial

Integrity7.

KineticAssays

for

Quantifying

Protein

Dynamics7a.

Kinetic

Antibody

Internalization

Assays7b.

Kinetic

Live-Cell

Immunocytochemistry

Assays6c.

Kinetic

Transendothelial

Migration

Assays11.

Appendix:

Protocols

and

Product

Guides10.

Label-Free

Advanced

Cell

Analysis10a.Kinetic

Cell-by-CellAnalysis

Assays10b.Kinetic

Advanced

Label-Free

Classification

Assays(Back

Cover)Contact

Information

|About

the

CoverChapter

1Introducing

Real-Time

Live-Cell

AnalysisThe

biomedical

world

has

come

a

long

way

since

Anton

van

Leeuwenhoek

first

observed

living

cells

with

a

basic

microscope

in

1674.

Using

fluorescent

probes

and

modern

high

resolutionimaging

techniquesitisnow

possible

to

view

labeled

sub-cellular

structures

at

the10-50

nanometer

scale.

For

researchers

working

with

fixed

(dead)

cells,

organelles

can

be

studied

at

even

higher

resolution

using

electron

microscopy.

These

methodsprovide

tremendousinsight

into

the

structure

and

function

of

cells

down

to

the

molecular

and

atomic

level.The

further

development

ofcell

imagingtechniques

has

largely

focused

on

resolving

greater

spatial

detail

within

cells.Examples

include

higher

magnification,

three

dimensional

viewing

and

enhanced

penetration

into

deep

structures.Significantattentionhasalso

beenpaidto

temporal

resolution

–

time-lapse

imaging

has

evolved

for

high-frame

rate

image

capture

from

living

cells

to

address

“fast”biology

such

as

synaptic

transmission

and

muscle

contractility.

Any

consideration

for

technology

advances

at

lower

spatial

ortemporal

detail

mayinitiallyseem

mundane,

or

even

unnecessary.

However,this

wouldfail

to

recognizesome

key

unmet

user

needs.First,

there

is

an

increasing

realization

that

many

important

biological

changesoccur

over

far

longer

time

periods

than

current

imaging

solutions

enable.

For

example,

maturation

and

differentiation

of

stem

cells

can

take

hours,

days

andsometime

weeks,

which

is

hard

to

track

using

existing

methods.

Second,

imagingtechniques

are

not

readily

accessibleto

all

researchers

nor

on

an

everyday

basis.

This

lack

of

accessibility

is

either

due

to

virtue

of

instrumentation

that

is

expensive

and

use-saturated

or

bycomplex

software

that

renders

imageacquisition

and

analysis

the

sole

domain

of

the

expert

user.

Third,

and

particularly

with

regard

to

time-lapse

measurement,the

throughput

of

current

solutions

is

typically

too

low

for

frontline

use

inindustrial

applications.

Finally

and

most

importantly,

researchers

are

increasinglyaware

that

any

perturbance

of

the

cells

in

the

process

of

imaging

(e.g.

fixing,

loss

of

environmental

control)

can

introduce

unwanted

and

misleading

experimentalartifacts.

Together,

these

factors

frame

up

the

requirement

for

solutions

that

enablelonger-term,

non-perturbing

analysesof

cells

at

a

throughput

and

ease

of

usecommensurate

with

non-specialist

users,

and

at

industrial

scale.A

new

generation

of

specializedcompactmicroscopes

and

live-cell

imaging

devices,

are

nowemerging

to

meet

this

need.Designed

to

reside

within

the

controlled,

stable

environment

of

a

cell

incubator,

these

systems

gather

cell

images

(phase

contrast,

bright-field

and/or

fluorescence)

from

assay

microplates

automatically,

repeatedly

and

around

the

clock.

Imageacquisition

is

completely

non-invasiveTable

of

Contents

Previous

ChapterNext

ChapterPrevious

SectionNext

Sectionand

non-perturbing

to

cells,

opening

up

the

opportunity

to

capture

the

full,

and

as

needed,

long-term

time

course

of

the

biology.

Acquisition

scheduling,

analysisand

data

viewing

can

be

conducted

easily

and

remotely,

without

in-depth

knowledge

of

image

processing.

Data

is

analyzed

on

the

fly,

image

by

image,

to

provide

real-time

insight

into

cell

behavior.We

refer

to

this

paradigm,

which

is

differentiated

from

straight

live-cell

imaging

by

the

provision

of

analysed

data

at

scale

as

opposed

to

simply

images,

as

‘real-time

live-cell

analysis’.In

an

ideal

world,

the

images

acquired

from

a

live-cell

imaging

device

would

be

collected

only

from

photons

produced

by

the

sample

of

interest,

and

in

perfect

focus.

However,

this

is

not

the

usual

case.

There

are

multiplesources

of

confounding

signal

present

in

an

image,

each

needing

correction,

removal,

or

cleaning

in

order

to

reveal

information

which

has

been

generated

by

the

sample

elements

of

interest.

Corrections

are

needed

dueto

systematic

aberrations

in

animaging

system

stemming

from

multiple

sources.

For

example,

detector

anomalies(e.g.

detector

bias,

dark

currentvariability,

field

flatness

and

thermal

or

gamma-ray

noise),

opticalissues

(non-flat

optical

components

and

illumination

imperfections)

orundesired

signal

introduced

by

the

sample

are

common

issues.

Autofluorescence

from

cellular

components

or

media,

or

non-

biological

signal

sources

such

as

shading,

or

patterns

arising

from

sample

matrices

or

non-uniform

illumination

due

to

meniscus

effects

in

microwells

must

be

removedbefore

usable,

replicable

information

can

be

extracted.In

order

to

perform

these

corrections,

one

must

be

aware

of

the

effects

of

eachprocess,

and

manipulations

on

the

raw

images

must

be

repeatable,

to

ensurefaithful

capture

of

the

measured

biological

signal

across

images,

experiments,and

devices.

There

are

many

tutorials

and

software

toolkits

available

to

process

images,

however

systems

that

perform

these

corrections

as

a

matterof

course

provide

consistencyand

ease

of

use,

particularly

when

coupledwith

standardized

assays,

reagents

and

consumables

which

normalize

theexperimental

process

(e.g.

the

Incucyte?

Live-Cell

Analysis

System,

and

the

assays

and

reagents

available

fromSartorius).The

consistency

with

which

images

are

acquired

and

processed

strongly

influences

the

ability

to

analyze

thecollecteddata.

Thiscanbeatime-consuming

task,

and

purpose-built

software

that

presents

only

the

toolsnecessary

for

aspecific

scientificquestion

can

remove

what

can

be

a

significant

hurdle

in

the

image

analysis

workflow.While

traditional

compact

microscopes

typically

only

image

from

a

single

micro-

plate

or

flask

at

a

time,

new

live-cell

analysis

devices

such

as

Incucyte?

can

automatically

capture

and

analyze

images

from

multiple

microplates

in

parallel,

thereby

significantly

increasingthroughput

(e.g.

Incucyte

=

6

x

384

well

plates).

With

the

Incucyte?

Live-Cell

Analysis

System,

a

unique

moving

optical

path

design

means

that

the

cells

and

cellplates

remain

stationary

throughout

the

entire

experiment.

This

further

minimizescell

perturbance

and

enables

imaging

and

analyses

of

both

adherent

and

non-adherent

cell

types.This

combination

of

functionality,throughput

and

ease

of

use

revolutionizesNext

SectionNext

ChapterTable

of

Contents

Previous

ChapterPrevious

Sectionthe

way

researchers

can

think

about

imaging

assays

in

living

cells.

Real-time

live-cell

analysis

has

now

been

appliedto

a

wide

rangeof

phenotypic

cellularassaysincludingcellproliferation,

celldeath

and

apoptosis,

immune-cell

killing,

migration,

chemotaxis,

angiogenesis,neurite

outgrowth

and

phagocytosis.

Ineach

case,

the

full

time-course

data

and

‘mini-movies’

of

the

assay

provide

greater

biological

insight

than

end

point

assays.

Novel

analyses

such

as

area

under

curve,

time

to

signal

onset

or

threshold,

and

rate

parameters

(dx/dt)

are

at

times

highly

value

adding.

Simplycalculating

the

assay

signal

at

its

peak

time-point

and/or

atthe

optimal

signal/background

all

helpsinassembling

robust

and

reproducible

assays.

Of

course,

transient

effects

of

treatments

can

be

detected

by

kineticimaging

that

may

otherwise

be

missed

with

end-point

reads.Due

to

its

non-invasive

nature,measurementsfrom

cells

canbe

made

not

only

during

the

assay

itself

but

also

during

the

cell

preparation

and

‘pre-assay’

stage.

For

example,

the

morphologyand

proliferation

rates

of

cells

can

bemonitored

throughout

the

cell

culture

period

and

immediately

post-seeding

on

the

micro-titer

assay

plate.

The

parameter/phenotype

ofinterest

canbe

measured

prior

to

the

addition

of

treatments

to

provide

a

within

well

baseline

measure.Quality

control

of

cells

and

assay

plates

in

this

way

helps

improve

assayperformance

and

consistency

by

ensuring

that

experiments

are

only

conducted

on

healthy,

evenly

plated

cultures

with

theexpected

cell

morphology.The

real-time

live-cell

analysis

approach

also

provides

the

opportunity

to

make

data

driven

decisions

while

the

experiment

is

in

progress.

A

researcher

studying

the

biology

of

vascular

or

neuronal

networks,

for

example,

may

wish

to

first

establisha

stable

network

before

assessing

the

effects

of

compound

treatments

orgenetic

manipulations

(e.g.

siRNAs).With

continuous

live-cell

analysis,

it

is

straightforward

to

temporally

tracknetwork

parameters

and

use

the

real

time

data

to

judge

when

best

to

initiate

the

treatment

regimes.

The

timing

of

adjunct

studies

such

as

analysis

of

metabolitesor

secreted

proteins

in

supernatants

canalsobe

guided.

Drug

washout

studies

may

be

performed

using

the

real-time

data

to

identify

when

an

equilibrium

response

occurs

and

to

trigger

the

timing

of

the

washout

regime.

If

for

any

reasonit

transpires

that

the

experiment

is

notperforming

as

expected,

then

treatments

could

be

withheld

to

save

expensive

reagents

and

follow-on

experiments

canbe

initiated

more

quickly

to

make

up

time.Real-time

live-cell

analysis

is

extremelyhelpful

when

developing,

validatingand

troubleshooting

phenotypic

assays.

Within

a

small

number

of

assay

plates

it’s

usually

possible

to

obtain

a

clearunderstanding

oftherelationshipovertime

between

assay

signal

and

treatments,cell

plating

densities,

plate

coatings

and

other

protocol

parameters.

Scrutiny

ofthe

kinetic

data

and

‘mini-movies’

fromeach

well

help

to

rapidly

pinpoint

sources

of

within-

and

across-plate

variance

and

to

validate

the

biology

of

interest.

This

is

particularly

true

formore

advanced

cell

systems

such

as

co-cultures

where

farmore

permutations

and

combinations

of

protocol

parameters

exist

(e.g.

cell

plating

ratios)

and

the

biology

is

more

complex.Table

of

ContentsNext

SectionNext

ChapterPrevious

ChapterPrevious

SectionIn

summary,

real-time

live-cell

analysis

is

re-

defining

the

possibilities

and

workflows

of

cell

biology.

The

combination

of

ease

of

use,

throughput,

long

term

stability

and

non-

invasive

measurement

enables

researchers

to

monitor

and

measure

cell

behaviors

at

a

scale

and

in

waysthat

were

previously

not

possible,

or

at

the

least,

highly

impractical.

Inthe

following

chapters

of

this

handbook,

we

illustrate

this

with

a

range

of

different

application

examples.Next

SectionNext

ChapterTable

of

Contents

Previous

ChapterPrevious

SectionChapter

2From

Images

to

AnswersIntroductionThe

nature

of

cell

biology

research

typically

requires

that

image-based

methods

are

used

to

capturemoments

in

time

to

enable

comparisons

between

treatment

groups

and

across

imaging

modalities.Sample

information

is

typically

acquired

using

amicroscopeandadigital

camera,andthosemomentsin

time

are

processedand

analyzed.

Imagescaptured

with

a

typical

microscope

camera

are

digital

representations

of

the

analog

information

contained

in

the

sample,

providing

a

means

to

automatically

analyze

the

information

in

the

sample.

Once

thesedigital

snapshots

are

acquired,

image

processing

is

used

to

clean

up

the

data,

and

image

analysis

is

used

to

extract

usable

information

for

analysis.Atthe

core

of

all

of

these

manipulations

are

numbers

–images

are

comprisedof

pixels

(pictureelements),

andeach

pixel

in

an

image

has

a

digital

value

representingthe

brightness

or

intensity

of

that

portion

of

the

sample,

at

a

specific

moment

in

time.By

operating

on

these

values,

either

in

isolation,

or

while

considering

nearby

values,

the

information

in

the

images

can

be

cleaned

of

aberrant

information,

and

data

relevant

tothe

imaged

sample

can

be

extracted

and

measured.Figure

1.

Image

processing

and

analysis

is

accomplished

using

a

number

oftechniques,

guided

by

expert

knowledge

and

software

guidance.

To

ensureprocessing

consistency

across

static

and

kinetic

data,

it

is

important

to

establish

a

set

of

image

processing

parameters

which

enable

operation

on

all

imagesin

an

identical

manner.This

contextually

derived

data

processing

workflowwill

seamlessly

and

automatically

perform

all

of

the

necessary

pre-

and

post-

image

processingsteps,

up

to

and

including

object

analysis

and

graphical

representa-

tion

of

the

experimental

result.

Properly

designed

image

analysis

workflows

are

intended

torequireno

humaninterventionand

processes

imagearchives,gen-

erating

consistent

and

actionable

results

either

in

real-time,

or

post-acquisition.The

ImageProcessingWorkflowUser

Driven

or

AutomatedImageSampleVisualization

and

Preprocessing---Assess

data.Correct

image

defects.

Bleaching.Restoration

and

Reconstruction--Restore

useful

information.Kernal

filtering.Specify

Features--Suppress

noise.Identify

regions

of

interest.Analysis-Extract

parameterslike

area,overlap,

object

number.Classification--Group

objects

into

different

classes.Display

population

data.Next

SectionNext

ChapterTable

of

ContentsPrevious

ChapterPrevious

SectionPerforming

these

steps

on

individual

images

to

generate

sufficient

statistical

power

to

support

a

hypothesis

canbe

a

tedious

process.

However,

when

operating

on

large

numbers

ofimages

which

have

been

collected

in

a

substantially

similar

manner,

the

series

of

operations

performed

to

clean

up

the

data,

extract

desired

information,

and

compare

images

may

be

recorded

and

automatically

applied

to

many

images

in

a

single

experiment.

Once

this

data

has

been

extracted,

treatment

groups

maybecomparedto

assess

differences,

andhypotheses

evaluated.

Scaling

this

tothe

analysis

of

live-cell

experiments

allows

for

the

evaluation

of

temporal

data,

andextendingthis

to

microplate

microscopy

means

that

population

data

may

be

studied

with

ease.

This

basic

workflowis

the

subject

of

countless

tutorials

and

books,

and

the

domain

of

numerous

software

packages

that

offer

a

cornucopia

of

tools

intended

to

answer

a

broad

range

of

scientific

questions.Image

Processing

toRemoveSystematic

or

Sample-InducedArtifactsThe

image

data

we

have

described

above

is

typically

captured

by

detectors

that

convert

analog

information,

specifically

photons,

into

digital

signals.

This

analog

information

is

collected

in

a

matrix

fashion,

spatially

rendered

according

to

location

in

the

sample.

Ideally,

the

signal

undergoing

analog

to

digital

conversionwould

comeonly

from

photons

produced

by

the

sample

ofinterest,andin

perfectfocus.However,this

is

not

the

usual

case.

There

are

multiplesources

of

confounding

signal

present

in

an

image,

each

needing

correction,

removal,

or

cleaning

in

order

to

reveal

information

which

has

been

generated

by

the

sample

elements

of

interest.

Corrections

are

needed

due

to

systematic

aberrationsin

an

imaging

system

stemming

from

multiple

sources.

For

example,

detector

anomalies

(e.g.

detector

bias,

dark

current

variability,

field

flatness

and

thermal

or

gamma-ray

noise),

optical

issues

(non-

flat

optical

components

and

illumination

imperfections)

or

undesired

signalintroduced

by

the

sample

are

common

issues.

Autofluorescence

from

cellularcomponents

or

media,

or

non-biological

signal

sources

(i.e.

shading

or

patterns

arising

from

sample

matrices,

micro-fluidic

channels,

or

non-uniform

illumination

effects

in

microwells)

must

be

removed

before

usable,

replicable

information

can

be

extracted.In

order

to

perform

these

corrections,

one

must

be

aware

of

the

effects

of

eachprocess,

and

manipulations

on

the

raw

images

must

be

repeatable

to

ensure

faithful

capture

of

the

true

biological

signal

across

images.

There

are

many

tutorials

and

software

toolkits

available

to

process

images,

however

systems

that

perform

these

corrections

as

a

matterof

course

provide

consistencyand

ease

of

use,

particularly

when

coupledwith

standardized

assays,

reagents

and

consumables

which

normalize

theexperimental

process

(e.g.

the

Incucyte?

Live-Cell

Analysis

System,

and

the

assays

and

reagents

available

from

Sartorius).

The

consistency

with

which

images

are

acquired

and

processed

will

influence

the

abilityto

analyze

the

collected

data.Previous

SectionNext

SectionNext

ChapterTable

of

ContentsPrevious

ChapterIdentifying

Biologyof

Interest

via

Image

Maskingor

“Segmentation”O(jiān)nce

an

image

has

been

appropriatelyprocessed

to

remove

aberrant

signal,

the

next

step

is

to

identify

the

biology

of

interest.Image

segmentation

is

a

binary

process,

meaning

pixels

are

classifiedas

either

“in”

and

are

included

in

any

enumeration

process,

or

“out”

and

notconsidered

as

part

of

the

sample.

Thesimplest

method

for

determining

which

pixels

are

in

or

out

is

by

thresholding,

or

setting

a

boundary

above

which

allpixels

are

“in”,

and

below

which,

all

pixels

are

“out”.

More

complex

tools

do

exist,

and

more

complexinteractions

can

be

performed

with

multiple

masks,

and

Boolean

operations

(e.g.,

AND,OR,

NOT)in

order

to

hone

in

on

the

exact

pixelsof

scientific

interest.Again,

this

canbe

atime-consuming

task,

and

purpose-built

software

that

presents

only

the

toolsnecessary

for

a

specific

scientificquestion

can

remove

what

can

be

a

significant

hurdle

in

the

image

analysis

workflow.Generating

Actionable

DataAfter

the

pixels

which

satisfy

all

of

themeasurementcriteria

are

identified

in

an

image,

it

is

possible

to

operate

on

this

binary

mask

of

pixels.

The

mask

maybe

analyzed

whole

(for

total

area,

orconfluence

measurements)

or

broken

into

multiple

subparts,

for

example

when

defining

or

counting

objects

in

the

image.

Depending

upon

the

labeling

ofthe

sample,

e.g.

label-free

or

tagged

with

a

specific

marker

such

as

a

fluorescent

reagent

labeling

a

specific

organelle

or

structure,

a

widevarietyof

statistics

maybe

generated.

In

the

case

of

fluorescent

reagent-labeled

images,

these

statistics

may

include

the

mean

intensity

valueof

all

the

pixels

in

the

mask,

the

total

additive

intensity,

the

minimum,

maximum,

or

standard

deviation

of

the

collective

intensity,

or

thefluorescencemask

maybe

used

to

count

numbers

of

objects.

Statistics

maybe

global

for

the

image

asjust

described

(e.g.

total

size

of

the

mask,

or

mean

intensity

of

the

mask)

or

per

object

(e.g.,

area

occupied

byindividualcells).Once

again,

the

appropriate

choiceof

labels,

image

processing,

and

object

identification

can

require

deep

technical

expertise,

as

the

number

of

options

available

to

differentiate

objects

is

very

broad.

For

example,

if

you

are

looking

for

all

red-labeled

nuclei

that

are

also

labeledwith

agreen

reagent(e.g.apoptotic

cells

labeled

with

Incucyte?

Caspase

3/7

GreenDye),

it

is

possible

to

identify

individual

cells

first

using

a

transmitted

light

image

[mask

1],

breaking

that

mask

into

objectsrepresenting

cells

using

image

processing

tools

like

watershed

split,

and

then

classifying

those

objects/cells

based

on

the

included

red

and

green

mean

intensityof

the

includednuclei.

This

task

is

more

easily

performed

when

the

scientific

question

is

well-defined,

the

appropriate

tools

are

utilized,

and

the

images

processed

automatically,

and

without

bias.Previous

SectionNext

SectionNext

ChapterTable

of

ContentsPrevious

ChapterAnalyzing

Image

Data

at

ThroughputNow

that

a

specific

set

of

operationshas

been

constructed

to

process

andanalyze

a

representative

image,

this

same

set

of

operations

may

be

applied

to

all

images

in

an

experiment

in

exactly

the

same

manner.

If

this

set

of

operationsinadequatelyprocesses

the

population

ofimages

included

in

an

experiment,

it

maybe

necessary

to

make

adjustments

to

theset

of

processing

operations

based

uponthe

population

of

images

collected

for

the

task.

In

a

live-cell

imaging

experiment

performed

in

a

96-well

plate,

a

dataset

containing

thousands

of

images

is

perfectly

reasonable.

Many

data

sets

willbe

considerably