55
High-Throughput Cell Engineering as a Tool in Research and Drug Development Tecan Symposium 2008 Herbert Müller-Hartmann Director R&D amaxa AG, a Lonza company

Tecan Symposium 2008...Workflow considerations Depending on number of samples, cells have to be stored in Nucleofector® solution for hours Available automation friendly OPs If no

  • Upload
    others

  • View
    0

  • Download
    0

Embed Size (px)

Citation preview

High-Throughput Cell Engineering as a Tool in Research and Drug Development

Tecan Symposium 2008

Herbert Müller-HartmannDirector R&Damaxa AG, a Lonza company

slide 2

Cells

Transfection Silencing RNAsGene expression arraysCellular imaging and analysis

Dedicated culture mediaSerum Free MediaCell culture matricesInnovative cell culture support

Normal human cellsAdult and embryonic stem cellsDisease cells « Cloned cells » SFM media ready

Assaysin cell

Ease of use

EnablingTechnology

Cell handling Cell analysisCell based assays

Cell based HTS/HCS assayIn Vitro ADMET assaysCell Health assaysCell function assays

Lonza Strategy Focus on Cell Discovery

slide 3

AgendaHigh-Throughput Applications Based on Cell EngineeringFunctional siRNA Screening

Kinome/Cell Cycle siRNA Screening in Human Umbilical Vein Endothelial Cells

Functional Fas-apoptosis Screen in Jurkat T Lymphocytes

Cell Engineering with Non-Nucleic Acid Molecules

Cell Engineering in High Throughput

slide 4

High-Throughput Transfection and Engineering of Cells by Chemical, Viral and Physical Methods

Müller-Hartmann et al., Expert Opin. Drug Discov. 2:1453-1465, 2007

Leading to more efficientTarget researchLead identification and valuationProtein productionOptimization of vectors and methods

Primary human hepatocytes

slide 5

MethodsElectrofusion

Physical deliveryElectroporation

Magnetofection

Sonoporation

Laser irradiation

Microinjection

Ballistic delivery

Chemical delivery

LipofectionCationic polymers

Calcium phosphate transfection

Viral transductionLentivirus

Adenovirus

Adeno-associated virus

Developed for HT

Müller-Hartmann et al., Expert Opin. Drug Discov. 2:1453-1465, 2007

High-Throughput Cell Engineering

slide 6

Increasing utilization of cell-based-assays, supplementing and even replacing biochemical assays

Recent trends in cell engineeringBiologically relevant cells, in particular primary cells

“Biologically inert” manipulation methods (e.g. “endosome-free” delivery)

“Cells as reagents”

High-Throughput Cell Engineering

slide 7

Interferon reponses

01020

30405060

708090

untre

ated

moc

k

0.02

pmol

0.2p

mol

2pm

ol

20pm

ol

untre

ated

moc

k

0.02

pmol

0.2p

mol

2pm

ol

20pm

ol

Nucleofection Lipofection

OA

S-1

mR

NA

(fol

d in

duct

ion;

24

h)

0

1000

2000

3000

4000

5000

6000

IL-6

rele

ase

(pg/

ml;

48h)

OAS-1 expressionIL-6 release

Data generated in collaboration with

Interferon Responses:Related to the Method of DeliveryHeLaS3 cells transfected with a 29-mer siRNA targeting DBI

slide 8

Cells show good growth and morphology on GEM™ beads

Proof-of-principle transfection of cells growing on alginate beads

“Cells as reagents”

AoSMC

24 h growth 48 h growth 96 h growth

CHO-K1

slide 9

AgendaHigh-Throughput Applications Based on Cell Engineering

Functional siRNA ScreeningKinome/Cell Cycle siRNA Screening in Human Umbilical Vein Endothelial Cells

Functional Fas-apoptosis Screen in Jurkat T Lymphocytes

Cell Engineering with Non-Nucleic Acid Molecules

Cell Engineering in High Throughput

slide 10

Loss/Gain of function screensKD of target by RNAi induces phenotype (e.g. Proliferation / cell death)

Modifier / sensitizer screensKD of target by RNAi induces phenotype only in combination withcell treatment (e.g. Pathway induction, cell stress)

Synthetic lethality screensKD of target by RNAi induces phenotype only in certain genetic background (e.g. Comparison of isogenic cell model with wt-Ras vs. Ras-mutated)

Pathway-specific screens using reporter geneKD of target modulates signaling measured by pathway-specific reporter gene (stable reporter cell lines)

RNAi Screening Strategies

slide 11

Workflow & opportunities for Amaxa® Nucleofection®

Nucleofection®

for assay development and generation of stable cell lines

Sam

ple

num

ber

Large libraries

focused libraries

Selected hits

Top hits

Ass

ay

com

plex

ity

96-well Shuttle®

for transfectablesubstrates

Cell-based Screening Approaches

slide 12

simple

ViabilityProliferationApoptosis

Calcium AssaysELISAFlash-LuminescenceReporter Gene assays

complex

High Content Analysis (HCA)Protein-Protein-interaction (e.g. FRET)Pathway specific assays

Standard assaysCommercially availableEasy to optimizeEasy to automate Standard instrumentation

High optimization effortSpecial Instrumentation

Assay Types

slide 13

Workflow considerationsDepending on number of samples, cells have to be stored in Nucleofector® solution for hours

Available automation friendly OPs

If no automation OP available: check for stability in solution, contact Amaxa

Cell typesJurkatHeLa/HeLa-S3Neuro2AHUVEChuman T-cells new OPhuman T-cells old OPstim. human T-cells

More to come …

Conditionsmodified protocolmodified protocolmodified protocolStandard OPStandard OPStandard OPStandard OP

Automation Solutions

slide 14

AgendaHigh-Throughput Applications Based on Cell Engineering

Functional siRNA ScreeningKinome/Cell Cycle siRNA Screening in Human Umbilical Vein Endothelial CellsFunctional Fas-apoptosis Screen in Jurkat T Lymphocytes

Cell Engineering with Non-Nucleic Acid Molecules

Cell Engineering in High Throughput

slide 15

Human siARRAY® - Protein Kinases (Dharmacon):SMARTpool® siRNA targeting 779 genesHuman siARRAY® - Cell Cycle Regulation (Dharmacon):SMARTpool® siRNA targeting 111 genes

siRNA libraries

Untreated cells (in 96-well Solution only)Negative control: siCONTROL® non-targeting siRNA #1 (Dharmacon)Positive control: SMARTpool® targeting PLK-1 (Dharmacon) and CHEK1 (Dharmacon)

Controls

Cell viability using CellTiter-Blue® assay (Promega) 72 h post transfectionData analysis:Z’ factor of controls (quality of experiment)Robust Z-score (hit identification)

Analysis

siRNA Screen in Primary HUVEC Cells

slide 16

siRNA library: 2 nmolLyophilized (-20°C)

Add 100 µlresusp. buffer

siRNA library stock: 20 µM (-20°C)

Transfer1.2 µl/well

Add 4.8 µl/well

2.4x104 cells/18 µl= 2x104 cells/15 µl

Cultured cells(25 flasks)

Add x ml for n wells = 18 µl/well*

Add18 µl/well

Transfer20 µl/well

Add80 µl/well Culture

medium

CTB reagent

siRNA library stock:24 pmol/6 µl (4µM)

Solution: 24 µlsiRNA: 24 pmolCells: 2.4 x 104

Solution: 20 µlsiRNA: 20 pmolCells: 2 x 104

Solution: 4.8 µl

Empty plate

Transfected cellSuspension 20 µl

Transfected cell Suspension 100 µl

3 x 103 cells/well

Analysis Fluoroscan3 x 103 cells/well

Add20 µl/well

3 h incub.

Transfer15 µl/well

96-well Nucleo-fector Solution

Add 85 µl/well

Culturemedium

96-well PCR plate96-well Nucleocuvette® plate96-well culture plate96-well microtiter plate (round)

3 d incubation

* ≈ 3 µl void volume/well

Screening Workflow

slide 17

||33

1'negativepositive

negativepositive

meanmeanSDSD

Z−

+−=

SeparationBand

Data variabilityBand

Data variabilityBand

μs μc

Freq

uenc

y

Assay Signal

3σs 3σc

Robustness and Screening Window

slide 18

Z‘-factor = 0.2

Z‘-factor = 0.5

0%10%20%30%40%50%60%70%80%90%

100%

0 10 20 30 40 50 60

sample #

rel V

iabi

lity

[% o

f unt

reat

ed]

siControl#1 CHEK#1

0%10%20%30%40%50%60%70%80%90%

100%

0 10 20 30 40 50 60

rel V

iabi

lity

[% o

f unt

reat

ed]

siControl#1 PLK

screening screening windowwindow

screening screening windowwindow

Dat

a ge

nera

ted

in c

olla

bora

tion

with

siRNA Screen in Primary HUVEC CellsAssay setup: Robustness – Screening Window

slide 19

HUVEC cell readout shows position/edge effects!

siControl#1

0%10%20%30%40%50%60%70%80%90%

100%

0 8 16 24 32 40 48 56 64 72 80 88 96

sample #

rel V

iabi

lity

[% o

f unt

reat

ed]

Assay setup: Plate uniformity?

siRNA Screen in Primary HUVEC Cells

Data generated in collaboration with

slide 20

0

50

100

150

1 2 3 4 5 6 7 8 9 10 11 12

Column

RFU

0

50

100

150

1 2 3 4 5 6 7 8 9 10 11 12

Column

RFU

1 Nucleofection® sample was plated onto 2 culture plates in different orientations.

ABCDEFGH

1 2 3 4 5 6 7 8 9 10 11 12

Nucleocuvette® Plate

Data generated in collaboration with

Position / Edge Effects are Nucleofection® Independent

slide 21

A

B

C

D

E

F

G

H

1 2 3 4 5 6 7 8 9 10 11 12

Medium only

Performing the Screen While Omitting the Outer Wells Requires Reformatting of Library

slide 22

Transfer from masterplateinto Nucleocuvette® plate

ABCDEFGH

1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12

ABCDEFGH

Reformat after Nucleofection®

into 2 culture plates, medium in the outer wells is necessary!

I II

8 wells for controls40 wells for targets

ABCDEFGH

1 2 3 4 5 6 7 8 9 10 11 12

Reformatting Strategies for Nucleofection® II

slide 23

020406080

100120140160180

B02 F02 G05 G09 E11 Avg C02 G02 G06 G10 D11 Avg D02 G03 G07 G11 C11 Avg E02 G04 G08 F11 B11 Avg

untreated siControl PLK1 CHEK1

RFU

Z‘ =0.43

plate positions

Z‘ =0.55

siRNA Screen in Primary HUVEC CellsPrimary Screen: Reproducibility of controls example plate

High reproducibility of controls (untreated cells, siCONTROL®, PLK-1, CHEK1) independent of plate position

Data generated in collaboration with

slide 24

020406080

100120140160180

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

plate #

RFU

no program siControl PLK1 CHEK1

Primary Screen: Reproducibility of controls over all plates

Data generated in collaboration with

siRNA Screen in Primary HUVEC Cells

slide 25

20,000

40,000

60,000

80,000

100,000

120,000

140,000

160,000

180,000

0 100 200 300 400 500 600 700 800 900 1000

well #

RFU

No programsiControl

CHEK1

PLK1

Raw Data

Data generated in collaboration with

Data Processing

slide 26

robust Z-score

-16

-14

-12

-10

-8

-6

-4

-2

0

2

4

6

0 100 200 300 400 500 600 700 800 900

PLK-1

CHEK-1CHEK-1

PLK-1

Data generated in collaboration with

Data normalization robust Z-Score

Data Processing

In total > 40 primary hits showing reproducible effects >2x MAD of the plate mean

robust Z-Score

slide 27

-14-12-10

-8-6-4-2024

CO

PB2

CN

KSR1

RAPG

EF3

CD

K4

CH

EK1

PYCS

IRAK3

PFKFB3

NLK

MYC

RPS6KL1

EDN

2

PKM2

MST1R

RIO

K1

MAR

K3

PLK2

PIK3C2A

CEN

PE

PLK1

DYR

K4

CKM

T1

CAR

KL

FGFR

1

MAPK11

IKBKB

FLJ13052

FGFR

3

RPS6KC

1

LMTK3

PRKAG

3

PRKC

B1

STK24

NR

BP

FN3KR

P

NEK6

LIM

TESK1

TGFBR

1

r-Z-S

core

Screen 1Screen 2Screen 3

Data generated in collaboration with

Hits sorted by rZ-Scores

Primary Screen: Comparison of three Screens

siRNA Screen in Primary HUVEC Cells

Good correlation between individual screens

slide 28

robust Z score correlation plot

-15-12

-9-6-303

-15 -12 -9 -6 -3 0 3 6

Screen 2

Scre

en 3

COPB2

CNKSR1

RAPGEFCDK4

MYCPYCS

CHEK1

TGFBR1

Primary Screen: Correlation of two Screens

siRNA Screen in Primary HUVEC Cells

Data generated in collaboration with

Good correlation between individual screens

slide 29

ConfirmationRetest identified Hits with same substrate with higher n’s

Redundancy De-convolution of siGENOME™ SMARTpool® (Dharmacon)

Test of 4 different single siRNAsOTP SMARTpool®

Different substrate, contains siRNA sequences different from thesiGENOME™ SMARTpool®

Dose Response

siRNA Screen in Primary HUVEC Cells Validation Strategy

slide 30

2-4 single siRNAs show same phenotype as the poolOTP pools also confirm phenotype (different sequences!)

0%

20%

40%

60%

80%

100%

siC

CH

EK

1si

G p

ool

siG

#1

siG

#2

siG

#3

siG

#4

OTP

poo

lsi

CC

HE

K1

siG

poo

lsi

G #

1si

G #

2si

G #

3si

G #

4O

TP p

ool

siC

CH

EK

1si

G p

ool

siG

#1

siG

#2

siG

#3

siG

#4

OTP

poo

lsi

CC

HE

K1

siG

poo

lsi

G #

1si

G #

2si

G #

3si

G #

4O

TP p

ool

C CDK4 C COPB2 C MYC C PYCS

rel.

viab

ility

[% o

f siC

ontro

l]

Dat

a ge

nera

ted

in c

olla

bora

tion

with

siRNA Screen in Primary HUVEC Cells Validation of TOP Hits

slide 31

Strong correlation of phenotype and knockdown CDK4

OTP and 4 of 4 single siRNAs show phenotype, as well as, KD COPB2

KD above 80% at 24h required for phenotype

OTP and 2 of 4 single siRNAs show phenotype, as well as, KD

0%

20%

40%

60%

80%

100%

120%si

C

CHEK1

siG

poo

l 1µM

siG

poo

l 0,5

µM

siG

poo

l 0,2

5µM

siG

#1

siG

#2

siG

#3

siG

#4

OTP

poo

l

siC

CHEK1

siG

poo

l

siG

poo

l 0,2

5µM

siG

poo

l 0,1

µM

siG

#1

siG

#2

siG

#3

siG

#4

OTP

poo

l

C CDK4 C COPB2

0%

20%

40%

60%

80%

100%

120%

siC

siG

poo

l 1µM

siG p

ool 0

,5µM

siG p

ool 0

,25µ

M

siG #

1

siG #

2

siG #

3

siG #

4

OTP

poo

l

siC

siG

poo

l 1µM

siG p

ool 0

,25µ

M

siG p

ool 0

,1µM

siG #

1

siG #

2

siG #

3

siG #

4

OTP

poo

l

C CDK4 C COPB2

0%

20%

40%

60%

80%

100%

120%

siC

siG

poo

l 1µM

siG p

ool 0

,5µM

siG

poo

l 0,2

5µM

siG #

1

siG #

2

siG #

3

siG #

4

OTP

poo

l

siC

siG

poo

l 1µM

siG

poo

l 0,2

5µM

siG p

ool 0

,1µM

siG #

1

siG #

2

siG #

3

siG #

4

OTP

poo

l

C CDK4 C COPB2

Cel

l tite

r72

h72

h

mR

NA

24h

Validation of TOP HitsDeconvolution of pools, phenotype vs KD

Data generated in collaboration with

slide 32

Summary

Sam

ple

num

ber

Large libraries

focused libraries

Selected hits

Top hits

Ass

ay

com

plex

ity

siRNA Screen in Primary HUVEC Cells

890 targetsin SMARTpool®

~40 Hitsselected by statistics

16 Hits selected for validation

6 Targets confirmed with independent substrates(incl. controls)

slide 33

AgendaHigh-Throughput Applications Based on Cell Engineering

Functional siRNA ScreeningKinome/Cell Cycle siRNA Screening in Human Umbilical Vein Endothelial Cells

Functional Fas-apoptosis Screen in Jurkat T Lymphocytes

Cell Engineering with Non-Nucleic Acid Molecules

Cell Engineering in High Throughput

slide 34

Human siARRAY® ON-TARGETplus™ Apoptosis (Dharmacon):SMARTpool® siRNA targeting 558 genes

siRNA libraries

Untreated cells (in 96-well Solution only)Negative control: siCONTROL® non-targeting siRNA #1 (Dharmacon)Positive controls: SMARTpool® ON-TARGETplus™targeting FAS or Casp 3 (both Dharmacon)

Controls

Induction of apoptosis with Fas-ligand 48h post transfectionCell viability using CellTiter-Glo® assay (Promega) Casp 3/7 using Apo-One® assay (Promega)Data analysis:Z’ factor of controls (quality of experiment)robust Z-score (hit identification)

Analysis

siRNA Screen in Jurkat T Lymphocytes

slide 35

FAS-L caspase 3

apoptosis

FAS

Contributes to peripheral depletion of lymphoid cells Maintenance of self-toleranceDown-regulation of the immune response and homeostasis of the immune system

FAS-Mediated Apoptosis

slide 36

0

50000

100000

150000

200000

250000

300000

Fas siRNA siCONTROL

1 pmol (50 nM)

RLU

Caspase Activity (Caspase-GloTM Assay, Promega corp.)

020000400006000080000

100000120000140000

Fas siRNA siCONTROL

1pmol (50 nM)

RLU

Cell Viability Assay(CellTiter-Glo® Assay, Promega corp.)

Before inductionAfter induction

Phenotypic Knockdown of FAS-Mediated Apoptosis

Data generated in collaboration with

slide 37

First screen assays established only with FAS as positive control

Only FAS detected in screenResponses from other pathway members too weak for consistent results

Conditions were “tweaked” using different positive controlsFAS apoptosis screen repeated

siRNA Screen in Jurkat T Lymphocytes First Results

slide 38

siRNA Screen in Jurkat T Lymphocytes

Optimization of assay parameterssiRNA concentrationsiRNA type (siGENOME®, ON-TARGETplus®, Dharmacon)Cell number during Nucleofection®

Cell plating densityCulture volumeTime point of apoptosis inductionInducer concentrationInduction time before analysisChoice of assays (Cell titer, caspase activity)

FAS-mediated apoptosis – assay setupcaspase3

apoptosis

FASFAS-L

caspase3

apoptosis

FASFAS

slide 39

Data generated in collaboration with

Jurkat FAS screen, CTG

-10

-5

0

5

10

15

20

25

30

35

40

0 80 160 240 320 400 480 560

robu

st Z

-sco

re

BAK1

APAF1CASP8 CASP3

DFFB

FADD

PARP1

FAS

SPATA3

False positive

n=3

Primary Screen Data, Cell Titer Glo Assay

Hits sorted by mean of rZ-Scores of three screensVery good correlation between the three screensOne apparent false positive (identified by screen repetition)

slide 40

FADDStrong mRNA KD helps protect from cell death (see cell titer)

Casp3Strong mRNA KD reduces caspase activation (ApoOne)

CellTiter-Glo®, Promega corp.

Apo-One®

Promegacorp.

mRNA

Dat

a ge

nera

ted

in c

olla

bora

tion

with

CTGlo

0%

100%

200%

300%

400%

500%

600%

700%

siControl FADD pool FADDsingle 1

FADDsingle 2

FADDsingle 3

FADDsingle 4

Casp3 pool Casp3single 1

Casp3single 2

Casp3single 3

Casp3single 4

ApoOne

0%

20%

40%

60%

80%

100%

120%

140%

160%

180%

siControl FADD pool FADDsingle 1

FADDsingle 2

FADDsingle 3

FADDsingle 4

Casp3 pool Casp3single 1

Casp3single 2

Casp3single 3

Casp3single 4

relative mRNA level

0%

20%

40%

60%

80%

100%

120%

siControl FADD pool FADDsingle 1

FADDsingle 2

FADDsingle 3

FADDsingle 4

Casp3 pool Casp3single 1

Casp3single 2

Casp3single 3

Casp3single 4

FADD Casp3

pool

No.

1

No.

2

pool

No.

1

No.

4

pool

No.

1

No.

4

pool

No.

1

No.

2

siRNA FAS-Apoptosis Screen in Jurkat T Lymphocytes

Deconvolution of pools,phenotype vs. knockdown

slide 41

Hit list indicates that FAS apoptosis in Jurkat might primarily use the mitochondrial route

BAK1

Cell titerFAS

FADDBAK1DFFB

PARP1APAF1CASP3

TNFRSF1ATRIM35CASP8

ApoOne/cell titer

BAK1FAS

CASP3APAF1FADD

ALOX12ITGB2LY86LTBR

TP53INP1

siRNA FAS-Apoptosis Screen in Jurkat T Lymphocytes

Dat

a ge

nera

ted

in c

olla

bora

tion

with

slide 42

AgendaHigh-Throughput Applications Based on Cell Engineering

Functional siRNA ScreeningKinome/Cell Cycle siRNA Screening in Human Umbilical Vein Endothelial Cells

Functional Fas-apoptosis Screen in Jurkat T Lymphocytes

Cell Engineering with Non-Nucleic Acid Molecules

Cell Engineering in High Throughput

slide 43

The Rationale Behind Non-nucleic Acid Delivery

Nucleofector® Technology - Accepted leadership for nucleic acid delivery in primary cells and hard-to-transfect cell lines

Efficient delivery of other molecules extends the range of addressable research applications – including high throughput

Current study coversSmall organic moleculesPeptidesRecombinant proteins and antibodies

slide 44

Small organic molecules:Extension / prioritization of lead candidate pipeline by pharma customers by in-vitro validation of compounds transfected by Nucleofection®

Allow customers to screen / validate compounds not permeating membranesExtending available pool of compounds beyond Ro5 (Lipinski)

PeptidesDelivery of peptides for target and pathway validation Enable systems biology approaches by complementing RNAi analysis

Applications

slide 45

Proteins Use as a tool for functional validation of pathways Functional in-vitro validation of recombinant proteins

AntibodiesUse as specific intracellular labels or functional entities (e.g. inhibitory)Intracellular trafficking & pathway validation

Applications

slide 46

Ctrl. Conc.1 Conc.2 Conc.3

Ctrl.Conc.1Conc.2Conc.3

Small Organic Molecule Nucleofection®

Organic molecules per se not crossing the cell membrane are homogenously delivered by Nucleofection® in a concentration dependent way

Flow cytometry analysis of K562 cells transfected with carboxyfluorescein

slide 47

Effekt der extrazell. Inkubation von Staurosporin (7.5 µM) auf HeLa% Viabilität bezogen auf unbehandelte Kontrolle (=100 %)

100.00

63.51

0

20

40

60

80

100

120

-/-/- + Inh.

o. Puls o. Puls

Effect: - 36 %

Effect of extracellular compound Effekt des Nukleofektions-Pulses auf HeLa% Viabilität bezogen auf unbehandelte Kontrolle (=100 %)

100.00

20.57

0

20

40

60

80

100

120

-/-/- + Inh.

o. Puls Puls 16

Effect: - 79.43 %Effect:

- 79 %

Effect of transfected compound

% v

iabi

lity

% v

iabi

lity

Small Organic Molecule Nucleofection®

Transfected organic molecules exhibit functional in-vitro activity

Viability analysis of Jurkat cells transfected with staurosporin

slide 48

Time course fluorescence intensity analysis of wt MDCK after Nucleofection® with calcein

Efflux kinetics Rhodamine 123Normalized (0 min = 100 %)

0

20

40

60

80

100

120

--- bkg 0 min. 2 min. 5 min. 10 min 30 min. 1h

wt MDCK stable-transfected MDCK

MDCK, Efflux kinetics of Calcein (10 µM)

0

100

200

300

400

500

600

700

800

900

--- bkg 0 min. 30 min. 60 min. 120 min 240 min.

x-m

ean

of c

ells

0

10

20

30

40

50

60

70

RLU

of s

uper

nata

nt

FL1 x-mean (FACS) Supernatant

Efflux kinetic comparison of wt MDCK vs. MDCK over-expressing MDR1 transporter

Small Organic Molecule Nucleofection®

Transfected organic molecules are specifically excreted via drugtransporter proteins

slide 49

SH-SY5Y cells transfected with fluorescently tagged 8 AA-peptide (flow cytometry analysis)

SH-SY5Y cells 2 hours post Nucleofection® with fluorescently tagged 8 AA-peptide (confocal images)

- peptide / - pulse+ peptide / - pulse+ peptide / + pulse

Peptide Nucleofection®

Peptides from 8 to 36 amino acids are efficiently delivered by Nucleofection®

slide 50

NuFe vs. competitor C

0

20

40

60

80

100

120

1 2 3 4 5 6 7

untreated ctrl 25 µM 25 µM 25 µM

inhibitory inhibitory inhibitory

+ pulse - + + competitor + competitor

+ TNF stimulation

Nor

mal

ized

Luc

ifera

se a

ctiv

ity (C

trl =

100

%

CHO-K1 Cells stably transfected with a Luciferase reporter of NF-кBpathway activity. Cells were stimulated using Tumor Necrosis Factor (TNF) and luciferase activity was assessed 30 minutes after peptide delivery.

Peptide Nucleofection®

Transfected peptides exhibit functional activity

slide 51

0

20

40

60

80

100

120

140

160

180

200

Autofluorescence Protein Protein + Program1

Protein + Program2

Jurkat cells transfected with fluorescently tagged ovalbumin

HeLa cells transfected with fluorescently tagged GαMIgG

0

20

40

60

80

100

120

Autofluorescence Antibody Antibody+Pulse

Recombinant proteins are efficiently delivered by Nucleofection®

Full size antibodies are delivered by Nucleofection®

Protein Nucleofection®

slide 52

Summary

Manipulation of eukaryotic cells by introducing nucleic acids and other substrates goes HT

Meaningful screening needs careful optimization and standardization of cell culture workflows and assay conditions

Functional siRNA Nucleofection® screening campaigns in difficult-to-transfect cell types relevant for bio-medical research are feasible while maintaining cellular functions

slide 53

Summary

Reproducible and comfirmable hits were identified for:

Cell proliferation in the HUVEC cell model for angiogenesisFas-mediated apoptosis pathway in the T lymphocyte model Jurkat cells

Upcoming applications for cellular delivery of non-nucleic acids

slide 54

Acknowledgements

Amaxa, a Lonza companyNicole SpottkeSheila OffizierSandra DomzalskiSonja SpickerMarkus ZumbansenLudger Altrogge

Thermo Fisher ScientificDharmacon products, Lafayette, CODevin Leake

Schering AG, BerlinClaudia Merz

Qui VanDaniela LitzenbergerYvonne StrübingMichael Kazinski

Thank you for your attention!

Tecan Symposium 2008

Herbert Müller-HartmannDirector R&Damaxa AG, a Lonza company