[PPT] - Modeling in Systems Biology: Progress, Problems and Applications to PowerPoint Presentation

SLIDE 1

Modeling in Systems Biology: Progress, Problems and Applications to Biotechnology and Biomedicine

Oleg Demin

Moscow State University, Institute for Systems Biology SPb

MCCMB, 2007

SLIDE 2

Development of quantitative description of biological processes and their application to biotechnology and biomedicine

Cellular metabolism (mitochondria, chloroplast, bacteria, hepatocyte, platelet…)
Cell signaling (NFkB, EGFR, Ca2+, cAMP, caspases, MAP …)
Gene regulatory networks (PurR, IclR, FruR …)
RNA metabolism (yeast)
Pathway reconstruction (PTH, osteoblasts, iflammation…)
Methods and software for control of industrial biotechnology processes

(PACS, strain improvement ….)

Methods and software for kinetic modeling (DBSolve, Model Creator, …)
Databases and Information Systems (Stem Cell Encyclopedia, Inflammation …)
Protein-protein interactions (electrostatic interactions, prediction
f binding energy and 3D-structure of complexes)

Goals Areas of Expertise

SLIDE 3

Biosystems Informatics Institute (UK, NewCastle) Edinburgh University (UK, Edinburgh) Vytautas Magnus University (Lithuania) Centre National de la Recherche Scientifique Universite Montpellier (France) Amsterdam Free University (The Netherlands) University of Barcelona (Spain) Moscow State University (Russia) Institute of Cancerogenesis, N.N. Blokhin Cancer Research Center (Russia) RiboSys Consortium: EC FP 6 EUROCOLI: European alliance

Partners and Collaborations

SLIDE 4

Challenges in Biomedicine and Biotechnology

How to discover new drugs in fastest, cheapest and most optimal way?
How to minimize drug side effect or/and to find new drugs without it?
How to optimize production of new drugs and their components?

AIM To present model based strategy how to cope with the problems Biosimulations and Informatics

Drug production Drug discovery Drug Safety

SLIDE 5

Mathematical Model

Structural Data OMICs Experimental Data High Throughput Data Biochemical Data Molecular Biology Data Physiology Data Clinical Data New knowledge about functioning and regulatory mechanisms of biological systems

Model Based Strategy

SLIDE 6

Model Based Strategy

liver: hepatocyte, mitocondria Bacteria: E.coli,

M. Tuberculosis

yeast

What organisms, organs, tissues and cells have been processed?

Cardiomiocyte: mitochondria

Blood system: platelets, endothelium cells

Plant: chloroplasts

steoblast

Stem cells

SLIDE 7

Model Based Strategy

hepatotoxicity, polluants effect

Tuberculosis

Osteoporosis,

steoarthrit

What diseases and drug side effects have been processed?

Heard attack Side effects of NSAIDs inflammation Breast cancer

SLIDE 8

Model Based Strategy

Kinetic modeling

Pathway reconstruction

Flux balance analysis

What modeling techniques do we use?

Analysis of protein docking

Enzyme kinetics

Structural modeling

SLIDE 9

Systems Biology Biosimulation/ data integration and interpretation Kinetic modeling Pathway reconstruction Data measurement Statistical models Proteomics Metabolomics Genomics

… …

Biosimulation: what we are talking about?

SLIDE 10

Topics discussed in this presentation

What kinetic model is?
Stages of kinetic model development and validation
What types of experimental data can be integrated by kinetic

model?

How kinetic model can integrate different experimental data?
Applications to biotechnology and biomedicine

Examples:

strain improvement
drug safety assessment of NSAIDs

SLIDE 11

KM should:

Take into account key properties of the biosystem

(stoichiometry, dynamics and regulation)

Clearly describe the key properties of the biosystem in terms of easily

understandable and measurable parameters (Vmax, Km, Kd, Ki, etc)

Reproduce correctly all known responses of the biosystem to external

and internal perturbations

Key KM requirements

Kinetic Modeling Approach

M. Noble, A. Kolupaev, O. Demin, F. Tobin and I. Goryanin, Biotechnology & Bioengineering, 2006, accepted

Metelkin, Е, Goryanin, I. and Demin, О., Biophys. J., 2006, 90, 423-432

What kinetic model is?

Kinetic model (KM) is system of ordinary differential equations describing dynamics and regulations of the corresponding biological system

SLIDE 12

Main stages of kinetic model development

Stoichiometry of metabolic pathway and elucidation of the key

enzymatic and genetic regulations: Kinetic scheme and N - matrix of stoichiometric coefficients

System of differential equations describing dynamics of the

pathway: dx/dt=N·v(x;e,K) Here, x=[x1,…xm] is vector of metabolite concentrations and v=[v1,…vn] is vector of rate laws

Description of individual enzymes:
catalytic cycle;
derivation of the rate laws for enzymatic reactions;
estimation of kinetic parameters of enzymatic reactions from

in vitro data, available from literature

Validation of the whole model using in vivo data

Goryanin I., Lebedeva G., Mogilevskaya E., Metelkin E. and Demin O. Methods Biochem Anal. 2006;49:437-88

SLIDE 13

in vivo in vitro

Metabolic networks Genetic networks

Properties

f Purified

enzymes Pertrubation Experiments

n Crude

extract Pertrubation Experiments

n cell culture

Measurement Of steady State fluxes Properties

f Purified

Operators And Regulatory proteins mRNA Time series Enzyme Expression profiles Protein structure

What types of experimental data can be integrated by kinetic model?

SLIDE 14

How KM integrates different types

f experimental data

1) Takes into account in vitro data measured for individual enzymes 2) Uses in vitro data measured on crude extract 3) Is validated against in vivo experimental data

Examples:

Signaling networks in platelets and endothelium cells initiated by

prostaglandins

Metabolic pathways of prostaglandins’ biosynthesis
E.coli central catabolic pathways
E.coli purine biosynthesis pathway
E.coli histidine biosynthesis pathway

SLIDE 15

In vitro experimental data

Properties of purified enzymes
kinetics (Histidinol dehydrogenase)
pH and temperature dependence (Histidinol dehydrogenase)
Perturbation experiments on crude extract
Adenylate degradation of burines biosynthesis pathway of E.coli

SLIDE 16

Minimal kinetic model construction.

Enzyme catalytic cycle construction

3D structure analysis
pH adjustment
temperature adjustment
rate equation derivation
identification of the unique set of parameters to fit all available kinetic

experimental data

minimal model from all available selection

EXAMPLE:

Histidinol dehydrogenase catalyzes two consecutive reactions in histidine biosynthesis pathway: 1) Hol + NAD = Hal + NADH (histidinol oxidation) 2) Hal + NAD = His + NADH (histidinal oxidation) Hol – histidinol, Hal – histidinal, His - histidine

SLIDE 17

PRPP PRATP PRAMP ProFAR PRFAR IGP IAP HolP Hol Hal His ADP AMP ppGpp ATP HisG HisI HisI PPi HisA HisHF HisB HisC HisB HisD HisD

+

+ Gln Glu AICAR His His1 HisC Glu KG Pi NAD NADH NAD NADH HisHF

Minimal kinetic model construction. Pathway of histidine biosynthesis

Purine biosynthesis

Ammonia assimilation Respiratory chain

PPi

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Reactions catalyzed with histidinol dehydrogenase

O.V.Demin, I. I. Goryanin, S.Dronov, G.V.Lebedeva Kinetic model of imidazole glycerol phosphate synthase from Escherichia coli , Biochemistry (Russian), 2004, 69(12): 1625-1638

SLIDE 18

Minimal kinetic model construction. Data on histidinol dehydrogenase available from literature

1) Structural data on catalytic site organization 2) Order of substrate’s binding and product’s release 3) pH dependence of maximal activity of histidinol oxidation (Hol and NAD as substrate) 4) pH dependence of maximal activity of histidinal oxidation (Hal and NAD as substrate) 5) Dependencies of initial rate of histidinol and histidinal oxidation at pH=7.5, pH=7.7 and pH=9.3 6) Time dependence of histidinol oxidation at pH=8.9

SLIDE 19

E E°NAD Hol°E Hol°E°NAD Hal°E°NADH E°NADH Hal°E His°E°NADH E His°E Hal°E°NAD

NAD Hol Hol NAD Hal Hal NADH NAD His NADH NADH His Hal

k1 k-1 k2

Minimal kinetic model construction. CATALYTIC CYCLE OF HISTIDINOL DEHYDROGENASE We constructed “minimal” catalytic cycle using

(1) structural data on catalytic site organization and (2) experimentally proved order of substrate’s binding and product’s release

SLIDE 20

Hal°EH2°NAD

EH EH°NAD Hol°EH Hol°EH°NAD Hal°EH°NADH EH°NADH Hal°EH His°EH°NADH EH His°EH Hal°EH°NAD

Hol Hol NAD Hal Hal NADH NAD His NADH NADH His Hal

k1 k-1 k2

EH EH2 Hol°E Hol°EH2 E°NAD EH2°NAD Hol°EH2°NAD Hol°E°NAD Hal°EH2°NADH Hal°E°NADH Hal°E°NAD E°NADH EH2°NADH Hal°EH Hal°EH2 E EH2 His°E°NADH His°EH2°NADH His°EH2 His°E

H H H H H H H H H H H H H H H H H H H H H H

NAD

“Minimal” assumptions on pH dependence enabling rate equation to fit experimental data: 1)

nly once protonated states of enzyme are active

2) There are 3 groups of enzyme states which differ each other in proton binding affinity

Minimal kinetic model construction. pH DEPENDENCE OF CATALYTIC CYCLE OF HISTIDINOL DEHYDROGENASE (1)

SLIDE 21

pH

11 10 9 8 7 6 5

Relative activity of enzyme (%)

100 90 80 70 60 50 40 30 20 10

Using experimentally measured data on (3) pH dependence of maximal activity of histidinol oxidation (Hol and NAD as substrate) and (4) pH dependence of maximal activity of histidinal oxidation (Hal and NAD as substrate) We’ve found values of parameters responsible for pH dependence

Histidinol

xidation

Histidinal

xidation

Minimal kinetic model construction. pH DEPENDENCE OF CATALYTIC CYCLE OF HISTIDINOL DEHYDROGENASE (2)

SLIDE 22

NAD (mM)

10 9 8 7 6 5 4 3 2 1

Reaction rate

0.25 0.2 0.15 0.1 0.05

Concentration (mM)

0.12 0.11 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01

Reaction rate

0.3 0.25 0.2 0.15 0.1 0.05 Histidinol oxidation at pH=7.7 Histidinol oxidation at pH=9.3 Histidinal oxidation at pH=7.7 Histidinol oxidation at pH=7.5 Histidinal oxidation at pH=7.5

Time (min)

6 5 4 3 2 1

NADH (mM)

0.04 0.035 0.03 0.025 0.02 0.015 0.01 0.005

Hol Hal His NAD NADH NADH NAD vHol vHal

pH=8.9 NAD(0)=0.5 mM Hol(0)=20 mM HDtotal=2.7 mM

Using experimentally measured (5) Dependencies of initial rate of histidinol and histidinal oxidation at pH=7.5, pH=7.7 and pH=9.3 and (6) Time dependence of histidinol oxidation at pH=8.9 We’ve found values of other kinetic parameters

Minimal kinetic model construction. Other kinetic parameters

SLIDE 23

In vitro experimental data

What types of in vitro experimental data for

individual enzymes can we integrate into KM?

structural features of enzymes/proteins
dependencies of initial velocity on substrates/products/effectors concentration
time series
pH and temperature dependencies
dependencies on Mg and other metal concentration
dependence on electric potential difference across the membrane
Problems encountered
we do not take into account how ionic strength influences reaction rate
we are not sure that estimated set of parameters is unique
there are contradictory experimental data in literature
in vitro experimental literature data measured for individual enzymes is not sufficient to

characterize their kinetic properties

Preliminary conclusions:

What experimental data can be used to identify these unknown parameters?

SLIDE 24

In vitro experimental data

Properties of purified enzymes
kinetics (Histidinol dehydrogenase)
pH and temperature dependence (Histidinol dehydrogenase)
Perturbation experiments on crude extract
Adenylate degradation of burines biosynthesis pathway of E.coli

SLIDE 25

R5P PRPP PRA FGAM NCAIR AIR

ATP AMP

Glycine gln glu

PPi ADP Pi ATP

GAR FGAR

formate

ADP ATP Pi f-THF THF ADP Pi ATP Gln Glu ATP ADP Pi ATP ADP Pi HCO3

AICAR SAICAR CAIR

Asp ADP ATP Pi

fumarate

NADPH NADP

prs purF purD purT purN purL purM purK 1 2 3 4 5 6 7 8 9 11 10 57 59 58 61

GR G GdR X A Hx AR HxR HxdR AdR FAICAR

f-THF THF

IMP XMP GMP GDP GTP ADP AMP sAMP

NAD NADH

Glu

Gln ATP PPi AMP NH3 ATP AMP PPi ATP ADP ATP ADP ATP ADP Pi R1P Pi dR1P PRPP PPi PRPP PPi ATP ADP GTP GDP Asp Pi

fumarate

R5P R1P Pi Pi dR1P Pi dR1P NH3 H2O NH3 H2O R1P Pi PRPP PPi

ATP

ADP ATP Pi R1P

XR

purE purC purB purH purH guaB guaC guaA purA purB adk apt amn deoD deoD deoD deoD deoD gsk gpt gpt hpt gpt, hpt deoD gsk gmk ndk xapA add add

dADP dGDP

nrd nrd

ATP ADP

ndk

dATP dGTP

DNA DNA RNA RNA

dAMP dGMP

ushA ushA ushA ushA ushA ushA ygfP dgt

Pi Pi Pi Pi Pi Pi

12 13 14 15 16 17 18 19 21 22 23 24 25 26 27 28 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 50 49 51 52 53 54 60 55 56

ADP ATP ADP ATP

E.Coli purine biosynthesis pathway

SLIDE 26

R5P PRPP PRA FGAM NCAIR AIR

ATP AMP

Glycine gln glu

PPi ADP Pi ATP

GAR FGAR

formate

ADP ATP Pi f-THF THF ADP Pi ATP Gln Glu ATP ADP Pi ATP ADP Pi HCO3

AICAR SAICAR CAIR

Asp ADP ATP Pi

fumarate

NADPH NADP

prs purF purD purT purN purL purM purK 1 2 3 4 5 6 7 8 9 11 10 57 59 58 61

GR G GdR X A Hx AR HxR HxdR AdR FAICAR

f-THF THF

IMP XMP GMP GDP GTP ADP AMP sAMP

NAD NADH

Glu

Gln ATP PPi AMP NH3 ATP AMP PPi ATP ADP ATP ADP ATP ADP Pi R1P Pi dR1P PRPP PPi PRPP PPi ATP ADP GTP GDP Asp Pi

fumarate

R5P R1P Pi Pi dR1P Pi dR1P NH3 H2O NH3 H2O R1P Pi PRPP PPi

ATP

ADP ATP Pi R1P

XR

purE purC purB purH purH guaB guaC guaA purA purB adk apt amn deoD deoD deoD deoD deoD gsk gpt gpt hpt gpt, hpt deoD gsk gmk ndk xapA add add

dADP dGDP

nrd nrd

ATP ADP

ndk

dATP dGTP

DNA DNA RNA RNA

dAMP dGMP

ushA ushA ushA ushA ushA ushA ygfP dgt

Pi Pi Pi Pi Pi Pi

12 13 14 15 16 17 18 19 21 22 23 24 25 26 27 28 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 50 49 51 52 53 54 60 55 56

ADP ATP ADP ATP

E.Coli purine biosynthesis pathway: adenylate degradation

SLIDE 27

20 18 16 14 12 10 8 6 4 2

concentration, mM

2 1

ATP AMP ADP Strain SO 003 – normal enzymes of adenylate degradation.

Experiments were initiated by addition

f 2.5 mM ATP. Extracts catalise rapid

conversion of ATP to ADP and Pi. Equilibration of ATP and ADP through adenylate kinase (adk) reaction provides AMP for adenylate degradation.

concentration. mM

time, min

20 18 16 14 12 10 8 6 4 2 0,3 0,2 0,1

A AR

time, min

20 18 16 14 12 10 8 6 4 2 concentration, mM 0,6 0,5 0,4 0,3 0,2 0,1

Hx HxR

Rapid production of AMP is followed by increase of adenine (A) and adenosine (AR) Inosine (HxR) and hypoxanthine (Hx) are formed in approximately equal amounts as the major end- products of AMP catabolism

SLIDE 28

time, min

20 18 16 14 12 10 8 6 4 2

concentration, mM

1

AMP A HxR AR Strain SO 105 – purine nucleoside phosphorylase (deoD) is missing

When deoD is missing, adenine is a major early product. After 20 min incubation Inosine becomes a major product. Formation of both AR and Hx depend

n the presence of deoD.

Strain SO 433 – both purine nucleoside phosphorylase (deoD) and adenosine deaminase (add) are missing

concentration, mM

time, min

20 18 16 14 12 10 8 6 4 2 1

A AR

Adenine is the major degradation product Adenosine is the only product that appears in significant amounts. No inosine was detected. Hypoxanthine concentration remained near zero. Adenine can be formed from adenylate by a pathway which does not depend on deoD and add. This activity is due to AMP nucleosidase (amn)

SLIDE 29

concentration, mM

time, min

20 18 16 14 12 10 8 6 4 2 1

A AR HxR Hx

Verification of parameter values found using independently measured experimental data

Experiment: Reaction is initiated by adding 1 мМ of adenine Satisfactory coincidence experimental data and model curves has been found Strain SO 003 – normal enzymes of adenylate degradation.

Parameters were determined which could not be estimated from in vitro data.

A unique set of parameters (maximal velocities, equilibrium constants, etc) was determined which allows to describe all experimental data on adenylate degradation both in the strains both with normal and altered pathway of adenylate metabolism

SLIDE 30

In vitro experimental data

What model parameters can be identified on the basis of in

vitro experimental data measured on crude extract?

kinetic parameters of individual enzymes
intracellular enzyme concentrations
But model, verified on the data cannot be applied to predict

response to extracellular perturbation.

Preliminary conclusions:

What experimental data can be used to fill this gap? In vivo experimental data quantifying coupled functioning of gene regulation and metabolism are needed

SLIDE 31

Experiment used to verify the model: dynamics of nucleotides in response to guanosine addition (30 mg/ml) to bacterial cultures of different E.coli strains (Petersen C. // J. Biol. Chem. v.279, 1999, 5348)

Parent strain CN1524 gsk-3 mutant CN1932 Gsk-3, guaC mutant CN2133

E.coli purine model validation, WT and mutants

SLIDE 32

In vivo experiments on growing bacterial cultures (E.coli parent strain, gsk-3, and gsk-3/guaC mutant).

All regulatory network (both metabolic and gene) is involved in pathway response to GR addition

GTP

RNA

Metabolic regulation: feedback inhibition feedback activation

GR G A Hx AR HxR IMP XMP GMP GDP ADP AMP sAMP

NAD NADH

Glu

Gln ATP PPi AMP NH3 ATP AMP PPi ATP ADP ATP ADP ATP ADP Pi R1P PRPP PPi PRPP PPi ATP ADP GTP GDP Asp Pi

fumarate

R5P R1P Pi NH3 H2O R1P Pi PRPP PPi

ATP

ADP ATP NADPH NADP

guaB guaC guaA purA purB adk apt amn deoD deoD gsk hpt hpt deoD gsk gmk ndk add

ATP RNA

14 15 16 21 22 23 24 25 26 30 31 32 33 34 35 37 39 40 41 60 59

R5P PRPP PRA

ATP AMP

gln glu

PPi ATP ADP

prs purF ATPasa purD purM purN purK purT purE purL purC purB purH

Biosynthesis: Pyrimidines Histidine Tryptophan NAD

Gene regulation: Repression by PurR G and Hx are corepressors

gsk-3

SLIDE 33

X[0] 120 110 100 90 80 70 60 50 40 30 20 10 GTP 1.4 1.3 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

ATP GTP dATP

X[0] 120 110 100 90 80 70 60 50 40 30 20 10 PRPP 1.6 1.5 1.4 1.3 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3

PRPP

Wild type strain E.coli. Response to guanosine addition.

metabolic regulation Gene regulation In the initial phase – metabolic regulation; In the late phase – gene regulation Kinetic response results from the complex interplay of metabolic and gene regulations

E.coli purine model validation, WT and mutants

SLIDE 34

ATP and GTP pools increase. Pool of PRPP is depleted

X[0] 120 110 100 90 80 70 60 50 40 30 20 10 ATP 9 8.5 8 7.5 7 6.5 6 5.5 5 4.5 4 3.5 3 2.5 2 1.5 1

ATP

X[0] 120 110 100 90 80 70 60 50 40 30 20 10 GTP 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1

GTP

X[0] 120 110 100 90 80 70 60 50 40 30 20 10 PRPP 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1

PRPP

Mutant strain E.coli (gsk-3). Response to guanosine addition.

E.coli purine model validation, WT and mutants

SLIDE 35

Strategy to integrate different types of experimental data into KM

has been developed

KM can be considered as depository of all available structural and

functional information which has predictive power

Preliminary conclusions

Availability of raw experimental data is prerequisite to model

development and validation

From kinetic modelers point of view, does not matter what

experimental technique is used and what quantitative experimental data is measured. To be integrated in KM the data

should be true
is accompanied with clear description of experimental

procedure and conditions

SLIDE 36

Applications of modeling to resolve problems arising in Biotechnology and Biomedicine

Strain improvement: maximization of

thymidine production

Safety assessment of NSAIDs

SLIDE 37

Modeling and optimization

f pyrimidine intermediate

(drug precursor) production in an industrial E. coli strain

SLIDE 38

PRPP

rotic acid

dihydro-

rotic acid

carbamoyl-aspartat HCO3

2ATP 2ADP+Pi glutamine glutamate

carAB

aspartate Pi

pyrBI pyrC pyrD pyrE carbamoyl-phosphate

H2O AH2 A

upp U C CR UR codA

NH3 H2O

cdd udp udk CTP UTP

R1P Pi H2O R5P

udk

ATP ADP ADP ATP ATP ADP ADP ATP glutamine glutamate

pyrG

ATP ADP

ndk ndk UdR CdR

H2O NH3

cdd deoA

dR1P Pi

R5P

ATP AMP RNA RNA RNA

OMP

Py Pyri rimi midine dine bios biosynthes ynthesis is pa pathw thway ay in in wil wild d ty type pe E.

E. coli

coli K12 K12

dCDP dUDP dUTP UMP UDP CMP CDP nrd nrd dUMP dTMP dTDP TdR dTTP T dut thyA dCTP

ATP ADP ADP ATP

ndk ndk

NH3

dcd

PPi CH2-THFA DHFA H2O

tmk

ADP ATP ADP ATP

ndk

ADP ATP

tdk pyrF pyrH cmk

RNA

tdk

dR1P Pi

deoA deoB, deoC nupC, nupG

Mod

del of

el of wil wild d type type pyri yrimidine midine biosynthe biosynthesi sis was de as develop eloped ed

SLIDE 39

PRPP

rotic acid

dihydro-

rotic acid

carbamoyl-aspartate HCO3

2ATP 2ADP+Pi gln glu

carAB

asp Pi

pyrBI pyrC pyrD pyrE carbamoyl-phosphate

H2O AH2 A

upp U C CR UR codA

NH3 H2O

cdd udp udk CTP UTP

R1P Pi H2O R5P

udk

ATP ADP

ADP ATP

ATP ADP

ADP ATP

gln glu

pyrG

ATP ADP

ndk ndk UdR CdR

H2O NH3

cdd deoA

dR1P Pi

R5P

ATP AMP RNA RNA RNA

OMP

Pyrimidine Biosynthesis Modified.

dCDP dUDP dUTP UMP UDP CMP CDP T4 nrdAB T4 nrdC dUMP dTMP dTDP

TdR

dTTP T

T5 dut T4 td thyA

dCTP

ATP ADP ADP ATP

ndk ndk

NH3

dcd

PPi CH2-THF DHF H2O

tmk

ADP ATP ADP ATP

ndk

ADP ATP tdk

pyrF pyrH cmk

RNA

tdk

dR1P Pi

deoA deoB, deoC nupC, nupG

ATP

PPi PPi

CO2

ATP ADP

THF T4 frd Pi

udp

Pi

PBS2

dTMPase

PBS2 dTMPase

PPi

T5 dut L-serine glycine serA

FBR

prs

FBR

cpa (B. subtilis) pyrB (B. subtilis)

glyA

CO2

gcvTHP

3-P-glycerate

dCMP

DNA

SP8 deaminase

T4 dCTPase

yeiK yeiK

Pi + Ribose

DNA yeiK Pi + ribose

pyruvate formate pflBA

CoA Acetyl-CoA

ADE3 yeast

ppk ppk ppk ppk

dUDPase

rpsA ppk

WT WT mo model l sh should b ld be adjuste justed to to current t ind industri strial l str strain ain

SLIDE 40

Second plasmid:PL-T5 dut serA cmk rpsA E. coli pyrG (n+) 40.14 2632/609/927 Second plasmid:PL-T5 dut serA

E. coli pyrG (n+)

38.59 2632/609/895 Second plasmid:PL-T5 dut

E. coli pyrG (n+)

38.18 2632/609/890 Second plasmid: E. coli pyrG (n+)2 supE, thi, hsdR2, deoA75, tdk-1, udp-1, ndk::kan, chr::Tn5::dTMPase thyA, cdd::kan, pyrG 36.68 2632/609/882 Added E. coli thyA with Plac to 596 (for plasmid stabilization) supE, thi, hsdR2, deoA75, tdk-1, udp-1, ndk::kan, chr::Tn5::dTMPase  -, thyA::Tn10 31.81 2576/609 Added  cI857 gene to 590 and it was removed supE, thi, hsdR2, deoA75, tdk-1, udp-1, ndk::kan, chr::Tn5::dTMPase  -, Nic+, Bio+ 29.74 2560/596 Added E. coli dcd,udk operon and T4 frd to 532 25.78 2549/590 2451 ndk::kan 17.8 2549/532 T4 nrdCAa-iB T4 td 9.77 2451/532 T4 nrdCAa-iB1 7.987 2451/366 PL-T4 nrdCAB supE, thi, hsdR2, deoA75, tdk-1, udp-1, nadA50, chr::Tn5::dTMPase (chlD-pgl),  cI857BAMH1 6.74 2451/343 Plasmid Genes Host Genotype

S. P.(mg / g DCW /

L / hr.) Strain/plasmid

2451/ 343 2451/ 366 2451/ 532 2549/ 532 2549/ 590 2560/ 596 2576/ 609 2632/ 609/ 882 2632/ 609/ 890 2632/ 609/ 895 2632/ 609/ 927 10 20 30 40 50

m g T d R /g DCW / L i te r / Ho u r (4 -6 h o u rs p o s t i n

Shake Flask Specific Productivity

Progress on Main Development Path

Model was adjusted to current industrial strain with the use of shake flask data on TM production of different strains, derived from

riginal parent strain

Target Metabolite production

SLIDE 41

11.79 0.1 0.1 0.04 0.1 10.0 0.01 0.1 0.1 (0.06) 0.02 0.8 2549/532 22.50 22.38 22.38 22.20 16.27 15.97 15.60 5.44 4.70 4.25 ? TdR, microM/min 0.0 0.0 0.0 0.0 0.0 10.0 10.0 10.0 10.0 10.0 12.74 Cdd 0.1 5.0 0.8 2.0 1.0 5.0 0.5 (0.06) 0.02 0.8 2632/609/927 1.0 5.0 0.8 2.0 1.0 5.0 0.5 (0.06) 0.02 0.8 2632/609/957 0.1 0.1 0.8 2.0 1.0 5.0 0.5 (0.06) 0.02 0.8 2632/609/895 0.1 0.1 0.8 2.0 0.1 5.0 0.5 (0.06) 0.02 0.8 2632/609/890 0.1 0.1 0.04 2.0 0.1 5.0 0.5 (0.06) 0.02 0.8 2632/609/882 0.1 0.1 0.04 0.1 0.1 5.0 0.5 (0.06) 0.50 0.8 2576/609 0.1 0.1 0.04 0.1 0.1 5.0 0.5 (0.06) 0.02 0.8 2549/590 0.1 0.1 0.04 0.1 0.01 0.1 0.1 15 0.02 0.8 2451/532 0.1 0.1 0.04 0.1 0.01 0.1 0.1 15 0.001 0.8 2451/366 0.1 0.1 0.04 0.1 0.01 0.1 0.1 15 0.001 0.04 2451/343 0.1-1 ? ? <9 ? ? 0.55 16.4 ? ? Vm,estimated from literature (mM/min) PRPP Cmk Dut (T5dut) PyrG THFA-syn Udk Dcd Ndk (ppk) thyA (T4 td) Nrd

0,00 5,00 10,00 15,00 20,00 25,00 2451/343 2451/366 2451/532 2549/532 2549/590 2576/609 2632/609/882 2632/609/890 2632/609/895 2632/609/927 2632/609/957

TdR production rate, microM/min

Assumptions: 1) Recycling of CH2THF is considered 2) Vm are decreased to fit experimental data,

Changes of the colour in a column (from white to yellow and orange) indicate changes in corresponding enzyme Vm; Red colour marks the strains which provide significant increase in TdR production rate.

TM production rate

Model allowed to reproduce historical data on TM production increase in industrial strains

SLIDE 42

Predictions of Kinetic model

1. Overexpression of upper pathway genes resulted in about 3-fold increase in

TM production rate.

2. Overexpression of other genes of pyrimidine biosynthesis pathway does not

change TM production

Upper pathway limitation

Changes in TM production resulted from 20-fold variation in enzime activity under condition of ATP saturation

10 20 30 40 50 60 70 80 90 100

normal VcarAB VpyrB VpyrC VpyrD VpyrE VpyrF VpyrH Vnrd Vppk Vdut VpyrG Vdcd VdTMPase21 Vyeik22 Vupp Vyeik24 Vcmk Vudk26 Vudk27 V_28 VcodA e_thyA VdTMPase37 Vyeik38 e_tmk (decrease) k_dTTP_DNA k_ATP_syn k_Q_syn k_CH2THF_syn k_TdR_out

TdR production rate, microM/min

SLIDE 43

PRPP

rotic acid

dihydro-

rotic acid

carbamoyl-aspartate HCO3

2ATP 2ADP+Pi gln glu

carAB

asp Pi

pyrBI pyrC pyrD pyrE carbamoyl-phosphate

H2O AH2 A

upp U C CR UR codA

NH3 H2O

cdd udp udk CTP UTP

R1P Pi H2O R5P

udk

ATP ADP

ADP ATP

ATP ADP

ADP ATP

gln glu

pyrG

ATP ADP

ndk ndk UdR CdR

H2O NH3

cdd deoA

dR1P Pi

R5P

ATP AMP RNA RNA RNA

OMP dCDP dUDP dUTP UMP UDP CMP CDP T4 nrdAB T4 nrdC dUMP dTMP dTDP

TdR

dTTP T

T5 dut T4 td thyA

dCTP

ATP ADP ADP ATP

ndk ndk

NH3

dcd

PPi CH2-THF DHF H2O

tmk

ADP ATP ADP ATP

ndk

ADP ATP tdk

pyrF pyrH cmk

RNA

tdk

dR1P Pi

deoA deoB, deoC nupC, nupG

ATP

PPi PPi

CO2

ATP ADP

THF T4 frd Pi

udp

Pi

PBS2

dTMPase

PBS2 dTMPase

PPi

T5 dut L-serine glycine serA

FBR

prs

FBR

cpa (B. subtilis) pyrB (B. subtilis)

glyA

CO2

gcvTHP

3-P-glycerate

dCMP

DNA

SP8 deaminase

T4 dCTPase

yeiK yeiK

Pi + Ribose

DNA yeiK Pi + ribose

pyruvate formate pflBA

CoA Acetyl-CoA

ADE3 yeast

ppk ppk ppk ppk

dUDPase

rpsA ppk

SLIDE 44

957 994 957 994 957 994

Strain

10 20 30 40 50

mg TM/L/hr./g DCW

1 2 3 4 5 6 7 8

mg Orotic Acid/L/hr./g DCW or % UdR

TM SF# 227- 6 hr.

2632/609/957 (contol); 2632/609/994 (p yrE

in 957)

TM

Orotic Acid

% UdR

957 957 998 957 957 998 957 957 998 Strain 20 25 30 35 40 45 TM (mg/L/hr./g DCW) 5 10 15 20 Orotic Acid ( mg/L/hr./g DCW) or % UdR

TM SF# 230 - 6 hr. -

2632/609/957 vs 2632/609/957/998 ( Bacillus pyr BCD cpa in Tc R

pSC101) TM Orotic Acid UdR

Experimental confirmation of model prediction

Strain 998 carries Bacillus Sub. pyrBCD , cpa additionally to strain

957. This results in about 10%

increase of TM production and about 2-3 fold increase in orotic acid. This indicates that pyrE limits TM production in 957 and 998. Amplification of pyrE in 957 results in increase in TM production

Upper pathway limitation

pyrBCD cpa

pyrE

SLIDE 45

ATP (mM) 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 Reaction rate (1/min) 5 000 4 500 4 000 3 500 3 000 2 500 2 000 1 500

UMP=0.12 mM 0.09 mM 0.06 mM 0.04 mM

UMP (mM) 0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 Reaction rate (1/min) 6 000 5 500 5 000 4 500 4 000 3 500 3 000 2 500 2 000 1 500

ATP=0.065 mM 0.04 mM 0.025 mM 0.015 mM

How to achieve further improvement of TM production? (in the strain with amplified upper pathway genes) Knockout pyrH and insert UMP kinase from Arabidopsis taliana. This enzyme catalyzes is feedback resistant with respect to UTP inhibition.

Predictions of Kinetic model

pyrH overexpression

Fitting pyrH rate equation to experimental data:

Potential for further improvement – modification of the regulatory properties of the enzymes catalyzing reactions of the pathway

SLIDE 46

Amplification of UMP kinase (-fold) 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 TdR production rate (%) 145 140 135 130 125 120 115 110 105 100

Arabidopsis UMP/CMP kinase Escherichia coli PyrH

Under condition of high energy state (i.e. respiratory rate and rate of oxidative phosphorylation is high enough) and amplified pyrD, pyrE and PRPP supply we found that 1) 20-fold amplification of E. coli pyrH results in about 25% elevation in TM production (blue line) 2) 20-fold amplification of Arabidopsis UMP/CMP kinase in strain with mutated pyrH results in about 45% elevation in TM production (black line)

Predictions of Kinetic model:

UMP kinase overexpression:

SLIDE 47

Experimental confirmation of model prediction

Strain 1010-1 (blue box) results from knockout of pyrH and amplification

f Arabidopsis UMP kinase in strain

957 (red box). These genetic changes are accompanied by increase in TM production.

UMP kinase overexpression:

pyrH, Arabidopsis

957 1006 1010-1 1008-5 957 1006 1010-1 1008-5 957 1006 1010-1 1008-5

Strains

10 20 30 40 50

S.A. (mg TM /L/h/g DCW)

5 10 15 UdR%, O.A. (mg OA/L/h/g DCW)

TM SF #250_6 hrs HPLC Data (LM + 50 mM KHPO4+ 1.0 g/L MgCl2)

2632/609/957; 2632/609/1006; 2632/609/1010-1; 2632/609/957/1008-5

TM % UdR

Orotic Acid

1006 = Aradidopsis UMP kinase in 942 1010-1 = Aradidopsis UMP kinase in 957 1008-5 =

E. coli nrd

HEIF pSC101 derived vector

SLIDE 48

Model predictions and experimental confirmation

N Prediction Result Status 1

Pyr D

verexpression

TM and orotate increase Confirmed

2

Prs overexpression PRPP increase C

nfirmed

3

PyrD/E

verexpression

TM increase Confirmed

4

PyrH

verexpression

TM increase Confirmed

5

Gln deficiency Nitrogen limitation Confirmed

6

ATP increase Energy limitation Confirmed

7

Phosphate optimum Pi inhibits TM production Confirmed

8

Acetate inhibits cellular growth NAD/NADH level decreases acetate Not tested

9

Sorbitol decreases Acetate NA D/NADH level decreases acetate Not tested

Nine predictions has been made for the mutation program and media

ptimisation

Seven of them led to the production increase

SLIDE 49

Conclusions

Kinetic modeling can serve as the powerful tool in strain improvement programs. Thymidine production in industrial strain has been increase by 25%

SLIDE 50

Safety assessment of NSAIDs

SLIDE 51

NSAIDs safety problem

NSAIDs – popular drugs for pain relief and antipyretic, more recently started to be used in cancer and even depression.

Main target – COX1,2
Wide range of adverse effects
Aspirin – risk of gastro-intestinal bleeding
Selective COX-2 inhibitors (Coxibs) - efficient in pain relief

but with unfavourable side effects (heart attacks)

– Vioxx withdrawal from the market – cost Merck $billions, with

ngoing legal costs

– FDA suggests Vioxx has contributed to >20 000 heart attacks & sudden cardiac deaths during its stay on market

The exact mechanism of NSAID action, and the origin of

many undesirable adverse effects still remain poorly understood.

SLIDE 52

AIM OF THE PROJECT:

To develop kinetic models of prostaglandin H synthases (COX-1,

COX-2), prostaglandin biosynthesis and signaling pathways in platelet and endothelium cells and apply the models to drug safety assessment including following NSAIDs: aspirin, celecoxib, diclofenac, naproxen, indomethacin, ibuprofen

Development of technology to drug safety assessment of NSAIDs

SLIDE 53

The Cyclooxygenase Reaction

Arachidonic acid

+ 2O2 PGG2 + H2O PGG2 PGH2

The enzyme has two activities: Cyclooxygenase and peroxidase

Cyclooxygenase (COX) is a membrane bound enzyme responsible for the

xidation of arachidonic acid to Prostaglandin G2 (PGG2) and the subsequent

reduction of PGG2 to prostaglandin H2 (PGH2).

SLIDE 54

Main assumptions in our model:

COX is a bifunctional enzyme with two distinct activities:

cyclooxygenase (COX) and peroxidase (POX)

Radical mechanism of COX functioning
Self-inactivation of COX and POX activity
Two isoforms COX-1 and COX-2

23 enzyme states and 55 reactions considered in the model

SLIDE 55

120 110 100 90 80 70 60 50 40 30 20 10

AA consumption., mkM

2,2 2 1,8 1,6 1,4 1,2 1 0,8 0,6 0,4 0,2

Time, s [AA]=20 mkM 2 mkM 1 mkM 0.5 mkM

Time, s 300 250 200 150 100 50 [Adrenochrome], mkM 160 140 120 100 80 60 40 20 1.08 mkM 0.81 mkM 0.54 mkM 0.27 mkM 0.16 mkM [COX-1]=1.61 mkM

Validation of the COX model

The COX model was developed and successfully validated on more than 150 independent studies globally

Rate constant Identified value Literature value k1 40 mM-1 s-1 k_1/k1= 1-3 mM k5 1.1 mM-1 s-1

k6

0.7 mM-1 s-1

k7

18 mM-1 s-1 14 mM-1 s-1 k9 332 s-1 350 s-1 k_in1 0.011 s-1 0.013 s-1

All kinetic parameters (22 in total) of COX catalytic cycle were identified:

Subset of these predictions checked against literature, all were accurate
Some estimates were not previously able to be observed, bordering on impossible / impractical to

measure experimentally...

Validated kinetic model can be used for analysis / prediction of the enzyme interaction with the inhibitors

SLIDE 56

Effects of inhibitors (NSAIDs) introduced to the COX model:

Aspirin

+

1,2
Indomethacin

+ + 1,2

Naproxen

1-,2+ + 1,2

Diclofenac

+ + 1

Ibuprofen
+

1,2

Celecoxib

1-,2+ + 2

Vioxx

1-,2+ + 2

Time dependence Reversibility

f binding

Selectivity to COX1,2

1- COX1; 2 - COX2

SLIDE 57

The model allows for consistent description of experimental data on inhibitory effects of different types of NSAIDs in vitro

1,2

Preincubation time, sec

Aspirin

0.8 mM 2.35 mM 4.36 mM

0,2 0,4 0,6 0,8 1 1,2 1000 2000 3000 4000 5000

Relative COX activity Preincubation time, sec

Indomethacin

2.2 mM 3.8 mM 5.4 mM

0,2 0,4 0,6 0,8 1 500 1000 1500

1.4 mM Relative COX activity

20 40 60 80 100 200 400 600 800 1000 1200

Ibuprofen concentration, mM

Ibuprofen

Relative COX activity Experimental data from: Varfolomeev S.B. Prostaglandins - molecular biological regulators. 1985. Publishing Moscow State

University. in Russian

We need the data on inhibitor binding to COX (rate constants for binding) to describe and predict the inhibitor effects in vitro.

Points – experimental data; Curves – model predictions

1

We also can identify parameters of NSAID binding to COX on the base of experimental curves and further use them in “in vivo” model.

Rel COX-2 activity

Celecoxib

0.5 mM 1 mM 2 mM

0,2 0,4 0,6 0,8 10 20 30 40 50 60 70

Experimental data from: Gierse J. K. et al Kinetic basis for selective inhibition of cyclo-oxygenases.

Biochem. J. (1999)

339, 607-614

Preincubation time, sec

SLIDE 58

Model predicts the discrepancy between in vitro/in vivo estimates of IC50 for Aspirin

0.2 0.4 0.6 0.8 1 1.2 0.1 1 10 100 1000

“in vivo” model IC50=2 uM IC50=300 uM in vitro model ASA, uM

PGH2 production, rel.un.

1. Warner T., Giuliano F., et al PNAS 96, 1999, 7563
2. Quellet M., Riendeau D., Percival M. PNAS 98, 2001, 14583
3. Mitchell J. , Akarasereenont P., et al PNAS 90, 1994, 11693
4. Chulada P., Langenbach R. J. Pharm. Exp. Ther. 280, 1997,606
5. Kargman S., Wong E. et al Bioch. Pharm. 52, 1996, 1113

In vivo COX-1

Whole blood assay 1.3 [1], Platelet 1.3 [2] Endothelial cells 1.5 [3] Fibroblasts 2.6 [4]

In vitro COX-1

Purified enzyme: 30-200 [5]

Experimental estimates of IC50 (mM) for Aspirin:

This discrepancy results from:

Higher concentration of substrates in vitro, preincubation with NSAID;
Accumulation of acetylated COX in running conditions due to additional

COX inhibition by remaining peroxidase activity.

I.e as a consequence of accumulation of acetylated COX under in

vivo conditions

SLIDE 59

Model predicts, that selectivity for Celecoxib in vivo depends on substrate concentration:

0.2 0.4 0.6 0.8 1 1.2 0.01 0.1 1 10 100

IC50=0.6 uM IC50=10 uM S1=0.1uM S1=0.01 uM Celecoxib, mM PGH2 production, rel.un.

0,2 0,4 0,6 0,8 1 1,2 0,001 0,01 0,1 1

S1=0.1 uM S1=0.01 uM IC 50=0.007 uM IC50=0.8 uM

Celecoxib, mM

Selectivity = 12.5 Selectivity = 85.7 COX1 COX2

IC50 in vitro 0.003 uM [1] IC50 in vitro 20 uM [1]

NB: Fast Coxib action allows use of in vitro data as a more reliable indicator than for other NSAIDs

1. Gierse J., Koboldt C., et al Biochem J. 339, 1999, 607
2. Warner T., Giuliano F., et al PNAS 96, 1999, 7563
3. Chan et al J Pharm Exp Ther 1999, 290(2) 551

Experimental estimate: Selectivity = IC50COX1/ IC50COX2 = 1.4 – 7 [2,3]. THEREFORE, experimental results for SELECTIVITY probably reflect higher 0.1 mM range S1 concentrations

Model allows to scale selectivity

SLIDE 60

Selective COX2 inhibitor can block aspirin effect – experimental phenomena

bserved, not previously explained

Our model explains and predicts the phenomena for the first time

50% inhibition plane High dose Aspirin High dose Coxib

For both single drug and drug combinations:
Aspirin, Ibuprofen, Naproxen, Celecoxib, Indomethacin, Diclofenac,…

Prediction of NSAID action on target

SLIDE 61

On the base of in vitro and in vivo experimental data on NSAID

binding to COX (rate constants of binding) model of COX has been developed

Model can be used to:
Predict/scale NSAID IC50 in any given in vivo conditions
Estimate the range of effective NSAID doses
Determine preferred dosing regimens for combined NSAID

use

Conclusion from NSAID studies of COX activity:

SLIDE 62

PGD2

(ext)

TXA2

(ext)

AA PLA2 COX-1 PGH2

TBXAS HHT TXA2

TXB2 inactivation

Ph.Lip.-AA

PGH2

(ext)

TXB2

(ext)

AA

(ext)

cAMP ATP PKA IP3R Ca2+ AC Ca2+ ER PLC

IP3

PIP2 AMP Ca- ATPase degradation PGE2

(ext)

R1 Gq R2 Gs thrombin, ADP, TXA2 (ext) PGI2 (ext) , iloprost (IP); PGD2

(ext) (DP)

PKC degradation

DAG

Platelet model

TXA2 biosynthesis
transmembrane transport
signalling pathways activated by prostanoids
Ca2+ fluxes involved in platelet activation
NSAIDs action on cyclooxygenase

SLIDE 63

Time, min

uM

IP3 cAMP Ca2+ COX-1

thrombin iloprost

Part of “active” IP3R High dose of iloprost effectively inhibits platelets activation by thrombin Stimulation of platelets by iloprost leads to cAMP- dependent PKA activation that phosphorilates IP3- dependent calcium channels (IP3R) on endoplasmic reticulum (ER). This results in in the inhibition of IP3R sensitivity to IP3, and a decrease in Ca2+ outflux from ER

Data from

JBC(2002)277:29321

Platelet model validated on perturbation experimental data with receptor agonists

SLIDE 64

AA PLA2 COX-1,-2 PGH2

TXAS HHT TXA2

TXB2 inactivation…

Ph.Lip.-AA

PGH2

(ext)

TXB2

(ext)

AA

(ext)

cAMP ATP PKA IP3R Ca2+ AC Ca2+ ER PLC

IP3

PIP2 AMP Ca- ATPase degradation PGE2

(ext)

R1

Gq

R2

Gs thrombin, TXA2 (ext)

TXA2

(ext)

PKC degradation

DAG

PGE2

PGES

PGI2 PGI2

(ext)

PGIM

(ext)

PGIS

PGI2 (ext) , iloprost (IP); PGD2

(ext) (DP)

Endothelium cell model

prostanoid biosynthesis (PGI2, PGE2, TxA2)
transmembrane transport
signalling pathways activated by prostanoids
Ca2+ fluxes involved in EC activation
NSAIDs action on cyclooxygenase

SLIDE 65

PGIMext TXB2ext PGE2ext PGH2ext “Resting” ECs (COX1 only) “IL1 - treated” ECs (COX1 + COX2)

In agreement with experimental data, activation of COX2, PGIS and PGES genes expression in IL1-treated ECs leads to increasing of production of PGI2 and PGE2, but not TXA2…

Time, min Increasing of Ca2+ Time, min

EC model validated on experimental data with Ca – ionophore stimulation

SLIDE 66

“Two cell model” takes into account:

1) Three compartments: endothelium cell, platelet and blood 2) Platelets express COX-1 only 3) Endothelium cells express COX-1 at normal conditions and both COX-1 and COX-2 at inflammation 4) Endothelium cells produce prostacyclin but platelets produce TXA2 5) Both endothelium cells and platelets produce PGD2, PGE2 and PGF2 6) Both endothelium cells and platelets can export/import PGH2

Application of the model to predict possibility of clot formation.

The aim of development of the model is to understand how clot formation depends on COX-1 and COX-2 inhibition

SLIDE 67

General scheme

Endothelial cell (HUVEC)

Arachidonic acid (exogenous) PGH2

,PGD2, PGE2,

PGF2a PGI2 (prostacyclin)

BLOOD Platelet

Arachidonic acid

Membrane

Phospholipids AA

transporter

PGH2 PGI2 (prostacyclin) PGH2

transporter

AA

transporter

Phospholipase A2 PGD2, PGE2, PGF2a

Transporter

COX-1 COX-2 Prostacyclin synthase Nonenzymatic Arachidonic acid

Membrane

Phospholipids AA

transporter

PGH2 TXA2 Thromboxane A2 PGH2

transporter

TXA2

transporter

Phospholipase A2 PGD2, PGE2, PGF2a

Transporter

COX-1 Thromboxane synthase Nonenzymatic TXA2 Thromboxane A2 Nonenzymatic

SLIDE 68

“Two cell model” predicts extracellular concentrations of prostanoids

Stimulated HUVECs. No inhibitors

50 100 150 200 250 300 350 400 2 4 6 8 10

Time (min) Concentration (nM)

PGH2_ext TXA_ext PGI_ext PGs_ext

PGD2 + PGE2 + PGF2a

Stimulated HUVEСs. COX-2 inhibited

50 100 150 200 250 300 350 2 4 6 8 10 Time (min) Concentration (nM) PGH2_ext TXA_ext PGI_ext PGs_ext

PGD2 + PGE2 + PGF2a

Inflammatory conditions

(endothelium express both COX-1 and COX-2)

COX1,2 inhibitors

+ COX-2

inhibitors 1) Extracellular prostacyclin (green) and tromboxan (red) does not change their profile with time upon COX-2 inhibition 2) Level of extracellular PGH2 and sum of PGD2+PGE2+PGF2a increases drastically upon inhibition of COX-2

SLIDE 69

allowed us to evaluate probability of clot formation in terms
f intracellular Ca2+ level and “platelet sensitivity” to

thromboxane.

predicted the combined effect of key prostanoids involved

in clotting: PGI2, PGD2, PGH2 and TXA2

allowed us to analyse possible adverse NSAID effects,

resulting from complex interplay of multiple variables controlling clot formation

Coupled platelet-endothelium model

SLIDE 70

CONCLUSIONS

Strategy to construct models of large

scaled biochemical systems and integrate different types of experimental data into KM has been developed

This strategy has been successfully

applied to resolve wide range of problems arising in Biotechnology and Biomedicine

SLIDE 71

Our team:

Modelers

Galina Lebedeva Alexey Goltsov Tatiana Plyusnina Anastasiya Lavrova Ekaterina Zobova Evgeniy Metyolkin, Aleksandr Dorodnov Aleksey Kolupaev Kirill Peskov Sergey Mironov

Bioinformatics

Ekaterina Goryacheva Yuriy Kosinskiy Andrey Dubinsky

Scientific programming

Nail Gizzatkulov Aleksandr Klimov

Acknowledgements: Edinburgh University, Biosystems Informatics Institute, GlaxoSmithKline

SLIDE 72

http://www.insysbio.ru