Logical Modeling Peripheral T Cell Differen5a5on Jim Faeder - - PowerPoint PPT Presentation

▶
logical modeling peripheral t cell differen5a5on
SMART_READER_LITE
LIVE PREVIEW

Logical Modeling Peripheral T Cell Differen5a5on Jim Faeder - - PowerPoint PPT Presentation

Nov 13, 2023 41 likes •716 views

Logical Modeling Peripheral T Cell Differen5a5on Jim Faeder Department of Computa.onal and Systems Biology CMACS PI Mee5ng New York University October 29, 2010 Acknowledgements Faeder Lab Department of Computa5onal and Systems Biology

slide-1
SLIDE 1

Logical Modeling Peripheral T Cell Differen5a5on

Jim Faeder

Department of Computa.onal and Systems Biology

CMACS PI Mee5ng New York University October 29, 2010

slide-2
SLIDE 2

Acknowledgements

  • Faeder Lab

Department of Computa5onal and Systems Biology

– Natasa Miskov‐Zivanov – John Sekar – Leonard Harris – Jus5n Hogg – Jintao Liu – Arshi Arora – Jose Tapia

  • Morel and Kane Labs

Department of Immunology

– Michael Turner – Lawrence Kane – Penelope Morel

  • Funding:

– NSF (Expedi5ons in Compu5ng) – NIH (P01, Dendri5c Cell Vaccines)

slide-3
SLIDE 3

Peripheral T cell differen5a5on

  • T cell subpopula5on ra5os are cri5cal for numerous

immune and auto‐immune pathologies

Source: Ochs et al., J Allergy Clin Immunol, 2009.

slide-4
SLIDE 4

Peripheral T cell differen5a5on

  • T cell subpopula5on ra5os are cri5cal for numerous

immune and auto‐immune pathologies

  • Key target for immunomodula5on therapy in cancer*

* Whiteside, T.L. “Inhibi5ng the Inhibitors...”, Expert Opin. Biol.

  • Ther. (2010), 10, 1019.

B. Tumor cell VEGFR Co-stim MHC class II Immune activation co-stimulation adjuvants EGFR TLR Treg Teffectors Proliferation/ differentiation Adoptive T-cell transfers PD-L1 Co-stim MHC class I CD137 MDSC CTL Proliferation/ differentiation Apoptosis Cytokine/ chemokine receptors Cytokine/ chemokine receptors TCR T CD8+ DC DC TLR TLR TLR TLR TLR TLR

slide-5
SLIDE 5

Dominant Role of Antigen Dose in CD4Foxp3 Regulatory T Cell Induction and Expansion1

Michael S. Turner, Lawrence P. Kane, and Penelope A. Morel2

The Journal of Immunology

Naïve T cells s5mulated with low Ag doses produce a high percentage of regulatory cells, which falls off as dose is increased.

slide-6
SLIDE 6

Dominant Role of Antigen Dose in CD4Foxp3 Regulatory T Cell Induction and Expansion1

Michael S. Turner, Lawrence P. Kane, and Penelope A. Morel2

The Journal of Immunology

Inverse correla5on between Foxp3+ Treg expansion and TCR signaling via Akt/mTOR/pS6.

slide-7
SLIDE 7

Key Findings

  • Treg induc5on is determined by Ag dose
  • Mechanism is T cell intrinsic

– Observed with both iDC and mDC – Observed with plate‐bound an5‐CD3/CD28

  • Inverse correla5on between mTOR ac5va5on

at 18h and Foxp3+ Treg at 7 days

  • No exogenous TGF‐β
slide-8
SLIDE 8

Modeling Goals

  • Determine whether known mechanisms are

sufficient to explain experimental

  • bserva5ons.
  • Suggest addi.onal experiments to iden5fy

missing mechanisms and clarifying areas of uncertainty.

  • Iden5fy other early markers of the response.
  • Incorporate signals through other receptors

predic.ve model.

slide-9
SLIDE 9

Rule‐Based Modeling of Signal Transduc5on

TCR/CD3

  • ITAM
  • ITAM

CD28

PRS

CD28

PRS ITAM2

  • ITAM2

Fc Fab

  • CD28

Fab

  • CD3

Fc Fab Fab

ZAP-70 Lck

SH3 PTK SH2 pY192 pY394 pY505

Vav1

DH SH2 pY267 pY280 pY826

Gads

SH3-C SH2

SLP-76

SH2 pY113 pY128 pY145 PRS Ubn-Lys PRS SH3 PTK

Itk

pY512 SH2 pY273 pY237

  • IgG

Fab Fab

Cbp/PAG

PRS

SLAP-130

pY571

Fyn

SH3 SH2 PTK pY531

ZAP-70

SH2-N SH2-C PTK pY493 pY319 pY315 pY292 RxxK

CD45

PTP

Csk

SH2

LAT

pY191 pY132

PLC1

SH2-C pY783 pY775 pY771 SH2-N

SHP-1

PTP SH2-N pY61

WASp

PBD PRS WH2 pY291

Cbl

PRS TKB pY731 RING UbcH7 pY240

Nck1

SH2 SH3-3 pY111 pY123 pY199

Grb2

SH3-N SH2

Sos1

PRS pY317 C E E E M M M C C C C C M C C C C C C C M C C GDP

Cdc42

RHO GTP M C

  • 21. PLCγ1

Gene names: PLCG1, PLC1 Uniprot accession number: P19174 Molecule type definiton: PLCG1(SH2 N,SH2 C,Y771∼u∼p,Y775∼u∼p,Y783∼u∼p) Domain structure: In the map of molecular interactions, PLCγ1 is represented with the following graph:

PLC1

SH2-C pY783 pY775 pY771 SH2-N

Phospholipase Cγ1 is an enzyme essential for T cell activation (127). It cleaves phos- phatidylinositol 4,5-bisphosphate, generating the second messengers diacyl glycerol (DAG) and inositol 1,4,5-trisphosphate (IP3) (128). IP3 binds to receptors on the endoplasmic reticulum, leading to release of Ca 2+ (129). Itk phosphorylates PLCγ1 on Y783, which is important for activation (51,130,131). PLCγ1 binds to phosphorylated LAT (111). The

Wiring diagram Object‐oriented model of protein Hu, Chylek, and Hlavacek, in prepara5on.

slide-10
SLIDE 10

Rule‐Based Modeling of Signal Transduc5on

TCR/CD3

  • ITAM
  • ITAM

CD28

PRS

CD28

PRS ITAM2

  • ITAM2

Fc Fab

  • CD28

Fab

  • CD3

Fc Fab Fab

ZAP-70 Lck

SH3 PTK SH2 pY192 pY394 pY505

Vav1

DH SH2 pY267 pY280 pY826

Gads

SH3-C SH2

SLP-76

SH2 pY113 pY128 pY145 PRS Ubn-Lys PRS SH3 PTK

Itk

pY512 SH2 pY273 pY237

  • IgG

Fab Fab

Cbp/PAG

PRS

SLAP-130

pY571

Fyn

SH3 SH2 PTK pY531

ZAP-70

SH2-N SH2-C PTK pY493 pY319 pY315 pY292 RxxK

CD45

PTP

Csk

SH2

LAT

pY191 pY132

PLC1

SH2-C pY783 pY775 pY771 SH2-N

SHP-1

PTP SH2-N pY61

WASp

PBD PRS WH2 pY291

Cbl

PRS TKB pY731 RING UbcH7 pY240

Nck1

SH2 SH3-3 pY111 pY123 pY199

Grb2

SH3-N SH2

Sos1

PRS pY317 C E E E M M M C C C C C M C C C C C C C M C C GDP

Cdc42

RHO GTP M C

Wiring diagram Hu, Chylek, and Hlavacek, in prepara5on.

BIONETGEN / NFSIM

Reaction Volume Reaction Volume Molecule Types:

A b

c

B

a

C

a

b a

Reaction Rules

A-C binding

A Reactants C Reactants

A-B binding

A Reactants B Reactants

A-B unbinding

A-B Reactants

A b

c

C

a

A b

c

B

a

A b

c

B

a

A b

c

B

a

A b

c

B

a

C

a

C

a

C

a

slide-11
SLIDE 11

Rule‐Based Modeling of Signal Transduc5on

TCR/CD3

  • ITAM
  • ITAM

CD28

PRS

CD28

PRS ITAM2

  • ITAM2

Fc Fab

  • CD28

Fab

  • CD3

Fc Fab Fab

ZAP-70 Lck

SH3 PTK SH2 pY192 pY394 pY505

Vav1

DH SH2 pY267 pY280 pY826

Gads

SH3-C SH2

SLP-76

SH2 pY113 pY128 pY145 PRS Ubn-Lys PRS SH3 PTK

Itk

pY512 SH2 pY273 pY237

  • IgG

Fab Fab

Cbp/PAG

PRS

SLAP-130

pY571

Fyn

SH3 SH2 PTK pY531

ZAP-70

SH2-N SH2-C PTK pY493 pY319 pY315 pY292 RxxK

CD45

PTP

Csk

SH2

LAT

pY191 pY132

PLC1

SH2-C pY783 pY775 pY771 SH2-N

SHP-1

PTP SH2-N pY61

WASp

PBD PRS WH2 pY291

Cbl

PRS TKB pY731 RING UbcH7 pY240

Nck1

SH2 SH3-3 pY111 pY123 pY199

Grb2

SH3-N SH2

Sos1

PRS pY317 C E E E M M M C C C C C M C C C C C C C M C C GDP

Cdc42

RHO GTP M C

Wiring diagram Hu, Chylek, and Hlavacek, in prepara5on.

BIONETGEN / NFSIM

Reaction Volume Reaction Volume Molecule Types:

A b

c

B

a

C

a

b a

Reaction Rules

A-C binding

A Reactants C Reactants

A-B binding

A Reactants B Reactants

A-B unbinding

A-B Reactants

A b

c

C

a

A b

c

B

a

A b

c

B

a

A b

c

B

a

A b

c

B

a

C

a

C

a

C

a

Issues

  • Models are very Kme‐consuming to construct.
  • Limited knowledge about wiring.
  • Lack of high‐resoluKon data.
  • Lack of measured parameters.
slide-12
SLIDE 12

Rule‐Based Modeling of Signal Transduc5on

TCR/CD3

  • ITAM
  • ITAM

CD28

PRS

CD28

PRS ITAM2

  • ITAM2

Fc Fab

  • CD28

Fab

  • CD3

Fc Fab Fab

ZAP-70 Lck

SH3 PTK SH2 pY192 pY394 pY505

Vav1

DH SH2 pY267 pY280 pY826

Gads

SH3-C SH2

SLP-76

SH2 pY113 pY128 pY145 PRS Ubn-Lys PRS SH3 PTK

Itk

pY512 SH2 pY273 pY237

  • IgG

Fab Fab

Cbp/PAG

PRS

SLAP-130

pY571

Fyn

SH3 SH2 PTK pY531

ZAP-70

SH2-N SH2-C PTK pY493 pY319 pY315 pY292 RxxK

CD45

PTP

Csk

SH2

LAT

pY191 pY132

PLC1

SH2-C pY783 pY775 pY771 SH2-N

SHP-1

PTP SH2-N pY61

WASp

PBD PRS WH2 pY291

Cbl

PRS TKB pY731 RING UbcH7 pY240

Nck1

SH2 SH3-3 pY111 pY123 pY199

Grb2

SH3-N SH2

Sos1

PRS pY317 C E E E M M M C C C C C M C C C C C C C M C C GDP

Cdc42

RHO GTP M C

Wiring diagram Hu, Chylek, and Hlavacek, in prepara5on.

BIONETGEN / NFSIM

Reaction Volume Reaction Volume Molecule Types:

A b

c

B

a

C

a

b a

Reaction Rules

A-C binding

A Reactants C Reactants

A-B binding

A Reactants B Reactants

A-B unbinding

A-B Reactants

A b

c

C

a

A b

c

B

a

A b

c

B

a

A b

c

B

a

A b

c

B

a

C

a

C

a

C

a

Issues

  • Models are very Kme‐consuming to construct.
  • Limited knowledge about wiring.
  • Lack of high‐resoluKon data.
  • Lack of measured parameters.

We did not “stand and fight” this Kme. Wisdom or cowardice?

slide-13
SLIDE 13

A Simpler Approach Boolean Networks

  • The state of an element in the signaling network

can be described by a Boolean variable, expressing that it is:

– Ac5ve or present (on or ‘1’) – Inac5ve or absent (off or ‘0’)

  • Boolean funcKons:

– Represent interac5ons between elements – The state of an element is calculated from states of

  • ther elements
  • The resul5ng network is a Boolean network
  • Long history of applica5ons to biology.
slide-14
SLIDE 14

Logical Modeling Approach

  • Generaliza5on of Boolean – variables may have more than 2

values.

  • Systema5c study of the dynamics of large systems:

– Depends largely on the interconnec5on structure

  • Does not require numerical parameters.
  • Discrete networks provide informa5on about:

– Mul5‐sta5onarity – Stability – Oscillatory behavior

  • Highly relevant for obtaining qualitaKve measures

– Perturba5ons – Environment – Alterna5ve wiring of the network

slide-15
SLIDE 15

Boolean Network Modeling Example

Biological network

p1 p2 p3 Proteins: p1, p2, p3

slide-16
SLIDE 16

Boolean Network Modeling Example

p1*= p2 OR p3 p2*= NOT p1 AND p3 p3*= p1 AND NOT p3

Biological network Boolean network

p1 p3 p2 p1 p2 p3 Proteins: p1, p2, p3

slide-17
SLIDE 17

Biochemical Examples

PDK1 mTORC2 Akt

Akt’ = PDK1 AND mTORC2

PI3K PTEN PIP3

PIP3’ = PI3K AND NOT PTEN

Note that PTEN overrides PI3K here.

slide-18
SLIDE 18

Boolean Models Are Logic Circuits

X1 X3 X2

x1(t+1) = x2(t) or x3(t) x2(t+1) = not x1(t) and x3(t) x3(t+1) = x1 (t) and not x3(t)

x1 x2 x3

S1 S2 S6 S8 S7 S3 S5 S4 State transiKon diagram Boolean network Logic circuit network

slide-19
SLIDE 19

Dynamics of a Boolean Model

S1 S2 S6 S8 S7 S3 S5 S4 AUractors Point aUractor Dynamic aUractor state x1x2x3 s1

000

s2

001

s3

010

s4

011

s5

100

s6

101

s7

110

s8

111

p1 p2 p3

slide-20
SLIDE 20

Different Methods for Simula5ng Network Dynamics

000 001 101 111 110 010 100 011 000 001 101 011 111 010 100 110 state x1x2x3 s1

000

s2

001

s3

010

s4

011

s5

100

s6

101

s7

110

s8

111

Synchronous updates Asynchronous updates

p1 p2 p3

x1(t+1) = x2(t) or x3(t) x2(t+1) = not x1(t) and x3(t) x3(t+1) = x1 (t) and not x3(t)

slide-21
SLIDE 21

Different Methods for Simula5ng Network Dynamics

000 001 101 111 110 010 100 011 000 001 101 011 111 010 100 110 state x1x2x3 s1

000

s2

001

s3

010

s4

011

s5

100

s6

101

s7

110

s8

111

Synchronous updates Asynchronous updates

p1 p2 p3

x1(t+1) = x2(t) or x3(t) x2(t+1) = not x1(t) and x3(t) x3(t+1) = x1 (t) and not x3(t)

Determinis5c Non‐Determinis5c

slide-22
SLIDE 22

Model Construc5on Process

!"#$%&'$()*+ ,$-.&(/+-(.+ 0&*12**&3(+ 4-)56-7+83.$9+ :3%'29-;3(+-(.+ <((3)-;3(+ =3/&1-9+83.$9+ 0$%&>-;3(+ ?2)*&.$+ $"#$%)*+ =3/&1-9+83.$9+ @&'29-;3(+ 83.$9+A-9&.-;3(+ 4%$.&1;3(*+

slide-23
SLIDE 23

Model Construc5on Process

!"#$%&'$()*+ ,$-.&(/+-(.+ 0&*12**&3(+ 4-)56-7+83.$9+ :3%'29-;3(+-(.+ <((3)-;3(+ =3/&1-9+83.$9+ 0$%&>-;3(+ ?2)*&.$+ $"#$%)*+ =3/&1-9+83.$9+ @&'29-;3(+ 83.$9+A-9&.-;3(+ 4%$.&1;3(*+

Almost 1 year!

slide-24
SLIDE 24

The Model

!"#$%&'$()*+ ,$-.&(/+-(.+ 0&*12**&3(+ 4-)56-7+83.$9+ :3%'29-;3(+-(.+ <((3)-;3(+ =3/&1-9+83.$9+ 0$%&>-;3(+ ?2)*&.$+ $"#$%)*+ =3/&1-9+83.$9+ @&'29-;3(+ 83.$9+A-9&.-;3(+ 4%$.&1;3(*+

~25 variables / 50 edges

!"#

$%# &'(! &'(!) #*+, #*-+ &#) " ./0- ! # ".12 3,24 "&56 3$!*1 73, !89 "*24 &3#40&3#- )%5: ;&<)#4 ;&<)#- "."1 ./0-) =!21 9>?@ABCD?9EC@ ?D9EF?9EC@ E@GEHE9EC@ 6(0!& 6(0$: !"04 )?A 5)2 (CA =I@ =62 3&!&J "2#0% K#?-L

!"04 6(0!& 6(0$:

!"#$

3$!*1 6(0!& 3&!&J !"04 6(0!& 3&!&J

%&'()

(CM71 !"04 6(0!& 3&!&J

!"#$*!+ +

6(0$:

(CM71 6(0!& 6(0$: 3$!*1 3&!&J

slide-25
SLIDE 25

The Model

!"#$%&'$()*+ ,$-.&(/+-(.+ 0&*12**&3(+ 4-)56-7+83.$9+ :3%'29-;3(+-(.+ <((3)-;3(+ =3/&1-9+83.$9+ 0$%&>-;3(+ ?2)*&.$+ $"#$%)*+ =3/&1-9+83.$9+ @&'29-;3(+ 83.$9+A-9&.-;3(+ 4%$.&1;3(*+

~25 variables / 50 edges

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

Model rules

slide-26
SLIDE 26

Receptors:

T cell receptor (TCR) Co‐s5mula5on through CD28 IL‐2 receptor (IL‐2R) TGFβ receptor (TGFβR)

Transcrip5on factors:

AP‐1, NFAT, NFκB, SMAD3, STAT5

Genes:

IL‐2, CD25, Foxp3

Other important elements:

PTEN, PI3K, PIP3, PDK1, Akt, mTORC1, mTORC2, TSC1‐TSC2, Rheb, S6K1, pS6

!"#

$%# &'(! &'(!) #*+, #*-+ &#) " ./0- ! # ".12 3,24 "&56 3$!*1 73, !89 "*24 &3#40&3#- )%5: ;&<)#4 ;&<)#- "."1 ./0-) =!21 9>?@ABCD?9EC@ ?D9EF?9EC@ E@GEHE9EC@ 6(0!& 6(0$: !"04 )?A 5)2 (CA =I@ =62 3&!&J "2#0% K#?-L

!"04 6(0!& 6(0$:

!"#$

3$!*1 6(0!& 3&!&J !"04 6(0!& 3&!&J

%&'()

(CM71 !"04 6(0!& 3&!&J

!"#$*!+ +

6(0$:

(CM71 6(0!& 6(0$: 3$!*1 3&!&J

slide-27
SLIDE 27

Influence sets

Element Influence set Element Influence set

PI3K TCR, CD28, IL‐2, IL‐2R AP‐1 Fos, Jun Akt PDK1, mTORC2 ERK Ras mTORC1 Rheb, PKC‐θ JNK Ras mTORC2 PI3K, S6K1 Fos ERK Foxp3 NFAT, AP‐1, STAT5, Smad3 Jun JNK IL‐2 NFAT, AP‐1, NFκB, Foxp3 NFAT Ca CD25 NFAT, AP‐1, NFκB, STAT5, Foxp3 Ca TCR STAT5 IL‐2, IL‐2R PDK1 PIP3 NFκB PKC‐θ, Akt TSC1‐TSC2 Akt Smad3 TGFβ, Akt, mTORC1 Rheb TSC1‐TSC2 PIP3 PI3K, PTEN S6K1 mTORC1 Ras TCR, CD28, IL‐2, IL‐2R pS6 S6K1

slide-28
SLIDE 28

Influence sets

Element Influence set Element Influence set

PI3K TCR, CD28, IL‐2, IL‐2R AP‐1 Fos, Jun Akt PDK1, mTORC2 ERK Ras mTORC1 Rheb, PKC‐θ JNK Ras mTORC2 PI3K, S6K1 Fos ERK Foxp3 NFAT, AP‐1, STAT5, Smad3 Jun JNK IL‐2 NFAT, AP‐1, NFκB, Foxp3 NFAT Ca CD25 NFAT, AP‐1, NFκB, STAT5, Foxp3 Ca TCR STAT5 IL‐2, IL‐2R PDK1 PIP3 NFκB PKC‐θ, Akt TSC1‐TSC2 Akt Smad3 TGFβ, Akt, mTORC1 Rheb TSC1‐TSC2 PIP3 PI3K, PTEN S6K1 mTORC1 Ras TCR, CD28, IL‐2, IL‐2R pS6 S6K1

slide-29
SLIDE 29

Logical modeling approach

Akt’ = PDK1 and mTORC2

slide-30
SLIDE 30

Influence sets

Element Influence set Element Influence set

PI3K TCR, CD28, IL‐2, IL‐2R AP‐1 Fos, Jun Akt PDK1, mTORC2 ERK Ras mTORC1 Rheb, PKC‐θ JNK Ras mTORC2 PI3K, S6K1 Fos ERK Foxp3 NFAT, AP‐1, STAT5, Smad3 Jun JNK IL‐2 NFAT, AP‐1, NFκB, Foxp3 NFAT Ca CD25 NFAT, AP‐1, NFκB, STAT5, Foxp3 Ca TCR STAT5 IL‐2, IL‐2R PDK1 PIP3 NFκB PKC‐θ, Akt TSC1‐TSC2 Akt Smad3 TGFβ, Akt, mTORC1 Rheb TSC1‐TSC2 PIP3 PI3K, PTEN S6K1 mTORC1 Ras TCR, CD28, IL‐2, IL‐2R pS6 S6K1

slide-31
SLIDE 31

Logical modeling approach

PIP3’ = PI3K and not PTEN

slide-32
SLIDE 32

Logical modeling decisions

  • Number of levels for element values

– TCR variable represents level of an5gen s5m.

  • No an5gen (TCR_LOW = 0, TCR_HIGH = 0)
  • Low an5gen dose (TCR_LOW = 1, TCR_HIGH = 0)
  • High an5gen dose (TCR_LOW = 0, TCR_HIGH = 1)
slide-33
SLIDE 33

TCR_LOW vs. TCR_HIGH

x

TCR_LOW not strong enough to overcome inhibi.on by PTEN.

slide-34
SLIDE 34

Logical modeling decisions

  • Choice between OR and AND:

– Example:

mTORC1’ = Rheb and (or?) PKC‐θ

slide-35
SLIDE 35

Logical modeling decisions

  • Choice between AND and OR:

PKC‐θ Rheb 1 1

slide-36
SLIDE 36

Logical modeling decisions

  • Choice between AND and OR:

PKC‐θ Rheb 1 1 1

mTORC1’ = Rheb and PKC‐θ

‘and’ rule means both are necessary for ac5va5on

slide-37
SLIDE 37

Logical modeling decisions

  • Choice between AND and OR:

PKC‐θ Rheb 1 1 1 1

mTORC1’ = Rheb

slide-38
SLIDE 38

Logical modeling decisions

  • Choice between AND and OR:

PKC‐θ Rheb 1 1 1 1 1

mTORC1’ = Rheb or PKC‐θ

‘or’ rule means either one is sufficient for ac5va5on

slide-39
SLIDE 39

Simula5on setup

  • Simula5on:

– For given ini5al condi5ons, computes system trajectory – Usually 20‐40 steps to reach steady state

  • Scenarios (ini5al condi5ons and rules)

– Simulated 300 5mes – Results show the percentage of being equal ‘1’ across all runs

slide-40
SLIDE 40

Model Valida5on

  • Three main scenarios:
  • 1. High vs. Low an5gen dose
  • 2. High an5gen dose, then removed
  • 3. High an5gen dose, then Akt or mTOR inhibitors added

Results are s.ll preliminary.

slide-41
SLIDE 41

An5gen Dose Dependence

Logical model results Experimental data

Source: Turner et al., The Journal of Immunology, 2009, 183, 4895‐4903.

slide-42
SLIDE 42

An5gen Dose Dependence

Experimental data

80nM 8nM 0.8nM Foxp3 CFSE Total Naïve

Source: Turner et al., The Journal of Immunology, 2009, 183, 4895‐4903.

Logical model results

slide-43
SLIDE 43

Foxp3 vs. pS6

!" #!" $!" %!" &!" '!" (!" )!" *!" +!" #!!" ,-./012" #*3-4" %5" )5"

!"#$

%&'($)*!'#*$+,-#$

6789%":;<" 90(":;<"

!" #!" $!" %!" &!" '!" (!" )!" *!" +!" #!!" ,-./012" #*3-4" %5" )5"

!"#$

%&'$()!*#)$+&,#$

6789%":;<" 90(":;<"

10 20 30 40 50 60 70 80 90 100 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Update round

Foxp3 [%] pS6 [%] 10 20 30 40 50 60 70 80 90 100 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Update round

Foxp3 [%] pS6 [%]

High AnKgen Dose Low AnKgen Dose

Experiment Experiment Model Model

slide-44
SLIDE 44

An5gen Removal

Remove TCR a|er 18 hrs

Foxp3

Experimental data

Source: Sauer et al., PNAS 105:7797, 2008.

slide-45
SLIDE 45

An5gen Removal

Remove TCR a|er 18 hrs

Foxp3

Experimental data

Source: Sauer et al., PNAS 105:7797, 2008.

Logical model results

10 20 30 40 50 60 70 80 90 100 TIME3 TIME4 TIME5 TIME6 TIME7 TIME8 TIME9 TIME10 no removal

% Th % Treg

slide-46
SLIDE 46

Akt and mTOR inhibitors

Experimental data

Source: Sauer et al., PNAS 105:7797, 2008.

!" #!" $!" %!" &!" '!" (!" )!" *!" +!" #!!" !" #" $" %" &" '" (" )" *" +" #!" ##" #$" #%" #&" #'"

!"#$%&" '()*+%"$,-.)" ,-./0#"123141567" 895"123141567" 4653" 671:12;<"

slide-47
SLIDE 47

Akt and mTOR inhibitors

Akt inhibitor mTORC1 inhibitor both together

Foxp3

Experimental data

Source: Sauer et al., PNAS 105:7797, 2008.

!" #!" $!" %!" &!" '!" (!" )!" *!" +!" #!!" !" #" $" %" &" '" (" )" *" +" #!" ##" #$" #%" #&" #'"

!"#$%&" '()*+%"$,-.)" ,-./0#"123141567" 895"123141567" 4653" 671:12;<"

slide-48
SLIDE 48

Low An5gen Trajectory

slide-49
SLIDE 49

Low An5gen Trajectory

slide-50
SLIDE 50

Low An5gen Trajectory

slide-51
SLIDE 51

Low An5gen Trajectory

Low dose steady state

slide-52
SLIDE 52

High An5gen Trajectory

slide-53
SLIDE 53

High An5gen Trajectory

Suppression of PTEN allows signal to reach Akt/mTOR axis. Could PIP3 level be a good early predictor

  • f cell fate?
slide-54
SLIDE 54

High An5gen Trajectory

No.ce that mTORC1 is ac.vated at same .me as STAT5. If STAT5 ac.va.on happens first, Foxp3 expression can happen transiently before mTOR suppression occurs.

slide-55
SLIDE 55

STAT5 vs. mTOR

!"#

$%# #&'( #&)' *#+ ! ,-.) " # ",/0 !12 "&03 *4#3.*4#) +%56 7*8+#3 ","/ ,-.)+ 9!0/ 2:;<=>?@;2A?< ;@2AB;2A?< A<CADA2A?< 4*!*E

!".3 FG.!* FG.$6

!"#$

4$!&/ FG.!* 4*!*E !".3 FG.!* 4*!*E

%&'()

G?HI/ !".3 FG.!* 4*!*E

!"#$*!+ +

FG.$6

G?HI/ 4*!*E FG.$6 "0#.% J#;)K FG.!* FG.$6 FG.!* +;= $5&,L$

  • 8FM

+;N $50) $5003 $50O G?= 9P< 5+0 9F0 4%8+* !".3 !"# !"$ %&''( %&') %&'* &!' +,'

  • .#

+/0 1")2 3'14! ,-56 ,-"7 894) 894)!# +5': ;656< 38:' 383: 3='( 5>? 6;1(46;1) !@AB C6D!1(

  • .EF:

534( !"# "$%& '()% *+,-. 5,= 8,G&!6&! 8,G&!6&=H G59I& 7I--&! %&=8I%HJ"#AH ;KD!6HJ"#A

Network Diagram Circuit Diagram

slide-56
SLIDE 56

STAT5 vs. mTOR

!"#

$%# #&'( #&)' *#+ ! ,-.) " # ",/0 !12 "&03 *4#3.*4#) +%56 7*8+#3 ","/ ,-.)+ 9!0/ 2:;<=>?@;2A?< ;@2AB;2A?< A<CADA2A?< 4*!*E

!".3 FG.!* FG.$6

!"#$

4$!&/ FG.!* 4*!*E !".3 FG.!* 4*!*E

%&'()

G?HI/ !".3 FG.!* 4*!*E

!"#$*!+ +

FG.$6

G?HI/ 4*!*E FG.$6 "0#.% J#;)K FG.!* FG.$6 FG.!* +;= $5&,L$

  • 8FM

+;N $50) $5003 $50O G?= 9P< 5+0 9F0 4%8+* !".3 !"# !"$ %&''( %&') %&'* &!' +,'

  • .#

+/0 1")2 3'14! ,-56 ,-"7 894) 894)!# +5': ;656< 38:' 383: 3='( 5>? 6;1(46;1) !@AB C6D!1(

  • .EF:

534( !"# "$%& '()% *+,-. 5,= 8,G&!6&! 8,G&!6&=H G59I& 7I--&! %&=8I%HJ"#AH ;KD!6HJ"#A

Network Diagram Circuit Diagram Intermediate events may be very fast. Test effect of varying the “buffer” length.

slide-57
SLIDE 57

STAT5 vs. mTOR

!"# !"$ %&''( %&') %&'* &!' +,'

  • .#

+/0 1")2 3'14! ,-56 ,-"7 894) 894)!# +5': ;656< 38:' 383: 3='( 5>? 6;1(46;1) !@AB C6D!1(

  • .EF:

534( !"# "$%& '()% *+,-. 5,= 8,G&!6&! 8,G&!6&=H G59I& 7I--&! %&=8I%HJ"#AH ;KD!6HJ"#A

Circuit Diagram Longer buffer means STAT5 wins race less

  • |en.

!" #!" $!" %!" &!" '!" (!" )!" *!" +!" #!!" #" %" '" )" +" ##" #%"

!"#$%&'()& *+,&-.,$& ,-./" 012340" 56-78"

slide-58
SLIDE 58

Role of CD25‐>STAT5‐>Foxp3

  • This pathway drives transient Foxp3

expression at high Ag dose and sustained expression at low dose (in the model).

  • Experiments suggest that both CD25

expression and pSTAT5 remain low in Foxp3‐ cells.

slide-59
SLIDE 59

Role of CD25‐>STAT5‐>Foxp3

  • This pathway drives transient Foxp3

expression at high Ag dose and sustained expression at low dose (in the model).

  • Experiments suggest that both CD25

expression and pSTAT5 remain low in Foxp3‐ cells.

  • Implies weak TCR s5mula5on may not be

enough to drive CD25. Could Foxp3 be driving CD25 instead?

slide-60
SLIDE 60

PTEN regula5on

  • PTEN blocks mTOR ac5va5on

at low dose resul5ng in 100% Treg – not observed.

  • Kine5cs of PTEN / PIP3 could

be very informa5ve.

  • Interplay with kine5cs of

CD25 / Foxp3 expression.

  • PI3K ac5vity increased by IL2

signaling and may par5ally

  • vercome PTEN block.
slide-61
SLIDE 61

Complex Interac5on between mTORC1 and mTORC2

  • mTORC2 ac5va5on s5ll

unclear:

– Possible ac5va5on by PI3K

  • r PIP3

– Nega5ve feedback from mTORC1 through S6K1

  • OscillaKons for high

an5gen dose

slide-62
SLIDE 62

Complex Interac5on between mTORC1 and mTORC2

  • mTORC2 ac5va5on s5ll

unclear:

– Possible ac5va5on by PI3K

  • r PIP3

– Nega5ve feedback from mTORC1 through S6K1

  • OscillaKons for high

an5gen dose

TCR PI3K PTEN PIP3 PDK1 Akt TSC1-TSC2 Rheb mTORC1 S6K1 mTORC2 IL-2 Foxp3

Step

slide-63
SLIDE 63

Complex Interac5on between mTORC1 and mTORC2

  • mTORC2 ac5va5on s5ll

unclear:

– Possible ac5va5on by PI3K or PIP3 – Nega5ve feedback from mTORC1 through S6K1

  • OscillaKons for high

an5gen dose

  • Solved by using three

levels for PI3K.

slide-64
SLIDE 64

mTOR role in Foxp3 expression

  • Links between mTORC1 and mTORC2 and the Foxp3

expression are not well understood

– Early mTORC1 signaling helps increase Foxp3 expression (through chroma5n remodeling) – Prolonged mTORC1 signaling inhibits Foxp3 – mTORC2 ac5va5on takes longer than mTORC1 ac5va5on – pS6 as a readout of mTORC1 ac5vity decreases a|er 18 hours – Both mTORC1 and mTORC2 are necessary for Foxp3 inhibi5on

slide-65
SLIDE 65

mTOR role in Foxp3 expression

  • Links between mTORC1 and mTORC2 and the Foxp3

expression are not well understood

– Early mTORC1 signaling helps increase Foxp3 expression (through chroma5n remodeling) – Prolonged mTORC1 signaling inhibits Foxp3 – mTORC2 ac5va5on takes longer than mTORC1 ac5va5on – pS6 as a readout of mTORC1 ac5vity decreases a|er 18 hours – Both mTORC1 and mTORC2 are necessary for Foxp3 inhibi5on

  • Further Experiments: correla5on between levels of mTORC1

and mTORC2 and the level of Foxp3 expression

slide-66
SLIDE 66

Conclusions

  • Logical modeling approach allows collabora5ve

model development.

  • Preliminary model reproduces dependence of
  • utcome on an5gen dose and dura5on.
  • Model focuses a~en5on on several key elements

– Rela5ve kine5cs of CD25 / Foxp3 expression – Role of differen5al PTEN regula5on – Possible role of Smad3 – Nega5ve feedback between mTORC1 and mTORC2 – mTORC1/2 regula5on of Foxp3

slide-67
SLIDE 67

Future modeling steps

  • Experimen5ng with three instead of two levels

– Increase in number of variables is not significant in terms of simula5on run5me

  • Modeling of popula5on of cells
  • Explora5on of the system’s sensi5vity