• Save
difuhgdlkhjl
Upcoming SlideShare
Loading in...5
×
 

difuhgdlkhjl

on

  • 194 views

 

Statistics

Views

Total Views
194
Views on SlideShare
194
Embed Views
0

Actions

Likes
0
Downloads
1
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Apple Keynote

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n
  • \n

difuhgdlkhjl difuhgdlkhjl Presentation Transcript

  • MODELLING COMPLEX BIOLOGICAL SYSTEMS IN THE CONTEXT OF GENOMICS Evry, May 21-25, 2012On the cellular and supracellular networks controlling regulatory T cells and autoimmunity Jorge Carneiro Instituto Gulbenkian de Ciência, Portugal http://qobweb.igc.gulbenkian.pt
  • Quantitative Organism Biology Cells of multicellular organism cooperate to ensure bodydevelopment and maintenance. They do this in a collective distributed manner, without a global plan. How cells collectively generate organism’s properties? General principles of biological organisation The design and control of artificial systems
  • http://www.embl-heidelberg.de/digitalembryo/
  • How cells get along without a GPS ?
  • How cells get along without a GPS ? global information (system)
  • autoimmune disease and immunopathology, 2 2 2 2 B Foxp3 represents a more specific marker than 5RBhi alone +Foxp3/MIGR1 +MIGR1 None CD25-CD45RBhi alone cell-surface molecules (such as currently used +Foxp3/MIGR1 +MIGR1 None CD25, CD45RB, CTLA-4, and GITR), which Multiple are unable to completely discriminate between regulatory T cells and activated, effector, or memory T cells. scales Colon Mutations in the Foxp3 gene culminate in the development of a fatal lymphoprolifera- tive disorder associated with multiorgan pa- thology both in mice and humans (12–20). FOXP3 is predominantly expressed in human CD25ϩCD4ϩ T cells as well (32). Further- Stomach more, transduction of a mutant Foxp3 lacking the forkhead domain, similar to the mutated Foxp3 in scurfy mice (17), failed to confer suppressive activity to CD25–CD4ϩ T cells Fig. 4. Prevention(fig.IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid of S7). The present results therefore sug- 105 fresh CD25–CD45RBhighCD4ϩ cells either alone (n ϭ 6, where n is the 4 ϫ 10that mutations of thehighCD4ϩgene may alone (n ϭ 6, where n is the mice received gest 5 fresh CD25–CD45RB Foxp3 cells either derived of mice)cause squares)disorders through1.2develop- GFPϩ sorted cells derived from pen squares) or together with 1.2 ϫ 106 GFPϩ sorted cells number from (open these or together with ϫ 106D4ϩ cells infected with Foxp3/MIGR1 (n ϭ 7) (closed circles) or MIGR1 –(n ϭ 5) highmentalcells infected with Foxp3/MIGR1 (n ϭthe (closed circles) or MIGR1 (n ϭ 5) CD25 CD45RB CD4ϩ or functional abnormality of 7) weight is represented as the percentage of initial weight (mean Ϯ SD). Body weight is represented as the percentage of initial weight (mean Ϯ SD) (open circles). CD25ϩCD4ϩ T population. R gnificant difference, P Ͻ 0.01, Foxp3/MIGR1 versus other two groups indicate significant difference, P of T cells by Foxp3 versus other two groups by Astericks by Potentially, generation Ͻ 0.01, Foxp3/MIGR1 R (B) Histopathology of the colon and stomach in each group and in an test. (B) Histopathology of the colon and stomach in each group and in an Mann-Whitney transduction of naıve T cells may provide a ¨ D mouse (None). (C) Colitis (left) and gastritis (right) were unreconstituted SCID mouse (None). (C) Colitis (left) and gastritis (right) were histologically histologically the group cotransferred with MIGR1-infected cells and one transferredTwo micepreviously unstudied therapeutic mode for cells and one transferred with scored. with in the group cotransferred with MIGR1-infected ϩ CD25 CD45RB treatment of autoimmunedebilitation before histological examination. Results –D4 cells alone died of debilitation before histological examination. Results high CD4ϩ cells alone died of and inflammatory are from a total of three independent experiments. shown in (A) to (C) are from a total of three independent experiments. diseases and in transplantation tolerance. 1060 14 FEBRUARY 2003 VOL 299 SCIENCE www.sciencemag.org 14 FEBRUARY 2003 VOL 299 SCIENCE www.scien
  • autoimmune disease and immunopathology, 2 2 2 2 B Foxp3 represents a more specific marker than 5RBhi alone +Foxp3/MIGR1 +MIGR1 None CD25-CD45RBhi alone cell-surface molecules (such as currently used +Foxp3/MIGR1 +MIGR1 None CD25, CD45RB, CTLA-4, and GITR), which Multiple are unable to completely discriminate between regulatory T cells and activated, effector, or memory T cells. scales Colon Mutations in the Foxp3 gene culminate in the development of a fatal lymphoprolifera- tive disorder associated with multiorgan pa- thology both in mice and humans (12–20). FOXP3 is predominantly expressed in human CD25ϩCD4ϩ T cells as well (32). Further- Stomach more, transduction of a mutant Foxp3 lacking the forkhead domain, similar to the mutated Foxp3 in scurfy mice (17), failed to confer suppressive activity to CD25–CD4ϩ T cells Fig. 4. Prevention(fig.IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid of S7). The present results therefore sug- 105 fresh CD25–CD45RBhighCD4ϩ cells either alone (n ϭ 6, where n is the 4 ϫ 10that mutations of thehighCD4ϩgene may alone (n ϭ 6, where n is the mice received gest 5 fresh CD25–CD45RB Foxp3 cells either derived of mice)cause squares)disorders through1.2develop- GFPϩ sorted cells derived from pen squares) or together with 1.2 ϫ 106 GFPϩ sorted cells number from (open these or together with ϫ 106D4ϩ cells infected with Foxp3/MIGR1 (n ϭ 7) (closed circles) or MIGR1 –(n ϭ 5) highmentalcells infected with Foxp3/MIGR1 (n ϭthe (closed circles) or MIGR1 (n ϭ 5) CD25 CD45RB CD4ϩ or functional abnormality of 7) Astericks by R Cell population dynamics weight is represented as the percentage of initial weight (mean Ϯ SD). Body weight is represented as the percentage of initial weight (mean Ϯ SD) (open circles). CD25ϩCD4ϩ T population. gnificant difference, P Ͻ 0.01, Foxp3/MIGR1 versus other two groups indicate significant difference, P of T cells by Foxp3 versus other two groups by Potentially, generation Ͻ 0.01, Foxp3/MIGR1 (ODE) R (B) Histopathology of the colon and stomach in each group and in an test. (B) Histopathology of the colon and stomach in each group and in an Mann-Whitney transduction of naıve T cells may provide a ¨ D mouse (None). (C) Colitis (left) and gastritis (right) were unreconstituted SCID mouse (None). (C) Colitis (left) and gastritis (right) were histologically histologically the group cotransferred with MIGR1-infected cells and one transferredTwo micepreviously unstudied therapeutic mode for cells and one transferred with scored. with in the group cotransferred with MIGR1-infected ϩ CD25 CD45RB treatment of autoimmunedebilitation before histological examination. Results –D4 cells alone died of debilitation before histological examination. Results high CD4ϩ cells alone died of and inflammatory are from a total of three independent experiments. shown in (A) to (C) are from a total of three independent experiments. diseases and in transplantation tolerance. 1060 14 FEBRUARY 2003 VOL 299 SCIENCE www.sciencemag.org 14 FEBRUARY 2003 VOL 299 SCIENCE www.scien
  • autoimmune disease and immunopathology, 2 2 2 2 B Foxp3 represents a more specific marker than 5RBhi alone +Foxp3/MIGR1 +MIGR1 None CD25-CD45RBhi alone cell-surface molecules (such as currently used +Foxp3/MIGR1 +MIGR1 None CD25, CD45RB, CTLA-4, and GITR), which Multiple are unable to completely discriminate between regulatory T cells and activated, effector, or memory T cells. scales Colon Mutations in the Foxp3 gene culminate in the development of a fatal lymphoprolifera- tive disorder associated with multiorgan pa- thology both in mice and humans (12–20). FOXP3 is predominantly expressed in human CD25ϩCD4ϩ T cells as well (32). Further- Stomach more, transduction of a mutant Foxp3 lacking the forkhead domain, similar to the mutated Foxp3 in scurfy mice (17), failed to confer suppressive activity to CD25–CD4ϩ T cells Fig. 4. Prevention(fig.IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid of S7). The present results therefore sug- 105 fresh CD25–CD45RBhighCD4ϩ cells either alone (n ϭ 6, where n is the 4 ϫ 10that mutations of thehighCD4ϩgene may alone (n ϭ 6, where n is the mice received gest 5 fresh CD25–CD45RB Foxp3 cells either derived of mice)cause squares)disorders through1.2develop- GFPϩ sorted cells derived from pen squares) or together with 1.2 ϫ 106 GFPϩ sorted cells number from (open these or together with ϫ 106D4ϩ cells infected with Foxp3/MIGR1 (n ϭ 7) (closed circles) or MIGR1 –(n ϭ 5) highmentalcells infected with Foxp3/MIGR1 (n ϭthe (closed circles) or MIGR1 (n ϭ 5) CD25 CD45RB CD4ϩ or functional abnormality of 7) Astericks by R Cell population dynamics weight is represented as the percentage of initial weight (mean Ϯ SD). Body weight is represented as the percentage of initial weight (mean Ϯ SD) (open circles). CD25ϩCD4ϩ T population. gnificant difference, P Ͻ 0.01, Foxp3/MIGR1 versus other two groups indicate significant difference, P of T cells by Foxp3 versus other two groups by Potentially, generation Ͻ 0.01, Foxp3/MIGR1 (ODE) R (B) Histopathology of the colon and stomach in each group and in an test. (B) Histopathology of the colon and stomach in each group and in an Mann-Whitney transduction of naıve T cells may provide a ¨ D mouse (None). (C) Colitis (left) and gastritis (right) were unreconstituted SCID mouse (None). (C) Colitis (left) and gastritis (right) were histologically histologically the group cotransferred with MIGR1-infected cells and one transferredTwo micepreviously unstudied therapeutic mode for cells and one transferred with scored. with in the group cotransferred with MIGR1-infected ϩ CD25 CD45RB treatment of autoimmunedebilitation before histological examination. Results –D4 cells alone died of debilitation before histological examination. Results high CD4ϩ cells alone died of and inflammatory are from a total of three independent experiments. shown in (A) to (C) are from a total of three independent experiments. diseases and in transplantation tolerance. 1060 14 FEBRUARY 2003 VOL 299 SCIENCE www.sciencemag.org 14 FEBRUARY 2003 VOL 299 SCIENCE www.scien Gene regulatory networks (Logical network dynamics)
  • overviewTolerance, autoimmunity, and regulatory T cellsRegulatory T cells 101How regulatory T cells mediate tolerance ?Modelling T cell population dynamicsOrdinary differential equations ··························· CRMWhat makes a T cell be a regulatory T cell ?Modelling gene regulatory networks and T cell differentiationLogical network formalism ······················ Th cell plasticityIf regulatory T cells are plastic how can tolerance be robust ?Multiscale modelling of cellular and supracellular networksAgent-based stochastic simulations
  • What does the immune system do in the vertebrate organism?
  • Fighting infection
  • Fighting infectionAssimilatingintestinal flora
  • Fighting infectionAssimilatingintestinal flora Rejecting cancer cells
  • Fighting infectionAssimilatingintestinal flora House keeping Rejecting cancer cells
  • Homeostasis Fighting infection and RegulationAssimilatingintestinal flora House keeping Rejecting cancer cells
  • Homeostasis Fighting infection and RegulationAssimilatingintestinal flora House keeping Rejecting cancer cells http://www.sciencemuseum.org.uk/exhibitions/lifecycle/116.asp http://ww.grc.org/Graphics/ programs/2003/cells3.jpg http://www.leukemia-web.org/images/cells.jpg
  • Failure of homeostasis and regulation unleashes pathologic autoimmunity
  • Failure of homeostasis and regulation unleashes pathologic autoimmunity Rheumatoid arthritis Type I diabetes Multiple sclerosis IPEX sindrome Cortesy: Magda Carneiro-Sampaio, São Paulo
  • From: Jean-François Bach, New England J Med
  • “In general, the management of human systemicautoimmune disease is empirical and unsatisfactory. For themost part, broadly immunosuppressive drugs, such ascorticosteroids, are used in a wide variety of severeautoimmune and inflammatory disorders (…)” – Philipe Cohen In: Fundamental Immunology (Ed. W. Paul)
  • Clonal selection theory Stem CellJerne, 1953Burnet, 1957 G.O.D. Ag A 1 2 3 4 ... 111 112 ... 623 ... 1245 ... n Ag B 1 2 3 4 ... 111 112 ... 623 ... 1245 ... n 111 111 1245 1245 111 111 111 111 1245 1245 1245 1245
  • CTL CTL CTL CTL CTL CTL TH CTLAntigen TH TH TH TH TH TH TH B B B B B B B TH
  • CTL CTL CTL CTL CTL CTL TH CTLAntigen TH TH TH TH TH TH TH B B B B B B B TH
  • CTL CTL CTL CTL CTL CTL TH CTLAntigen TH TH TH TH TH TH TH B B B B B B B TH
  • CTL CTL CTL CTL CTL CTL TH CTLAntigen TH TH TH “Self” tolerance by deletion TH TH TH TH B B B B B B B TH
  • CTL CTL CTL CTL CTL CTL TH CTLAntigen TH TH TH TH TH TH TH B B B B B B B TH
  • Reductionist molecular biology approach to Immunology
  • TOLERANCE = absence of clonal expansion = CLONAL DELETION Therapy of autoimmune diseases should aim at “deleting” autoreactive cells or clonesIn the absence of clonal-specific therapies: kill them all !!!
  • Is there any hope ?
  • Tolerance is mediated by regulatory T cellsMason-Sakaguchi WT
  • Tolerance is mediated by regulatory T cellsMason-Sakaguchi CD25- CD4+ T cells E E E T E T T T T (Effector T cells, TE, E) WT
  • Tolerance is mediated by regulatory T cellsMason-Sakaguchi “empty” rag-/- CD25- CD4+ T cells E E E T E T T T T (Effector T cells, TE, E) WT
  • Tolerance is mediated by regulatory T cellsMason-Sakaguchi “empty” rag-/- AID CD25- CD4+ T cells E E E T E T T T T (Effector T cells, TE, E) WT
  • Tolerance is mediated by regulatory T cellsMason-Sakaguchi “empty” rag-/- AID CD25- CD4+ T cells E E E T E T T T T (Effector T cells, TE, E) WT CD25+CD4+T cells R R R R R R R R (Regulatory T cells, TR, R)
  • Tolerance is mediated by regulatory T cellsMason-Sakaguchi “empty” rag-/- AID CD25- CD4+ T cells E E E T E T T T T (Effector T cells, TE, E) WT CD25+CD4+T cells R R R R R R R R (Regulatory T cells, TR, R)
  • Tolerance is mediated by regulatory T cellsMason-Sakaguchi “empty” rag-/- AID CD25- CD4+ T cells E E E T E T T T T (Effector T cells, TE, E) WT CD25+CD4+T cells R R R R R Healthy R R R (Regulatory T cells, TR, R)
  • Tolerance is mediated by regulatory T cellsMason-Sakaguchi “empty” rag-/- AID CD25- CD4+ T cells E E E T E T T T T (Effector T cells, TE, E) WT CD25+CD4+T cells R R R R R Healthy R R R (Regulatory T cells, TR, R)
  • Tolerance is mediated by regulatory T cellsMason-Sakaguchi “empty” rag-/- AID CD25- CD4+ T cells E E E T E T T T T AID (Effector T cells, TE, E) or Healthy WT CD25+CD4+T cells R R R R R Healthy R R R (Regulatory T cells, TR, R)
  • Tolerance is mediated by regulatory T cellsMason-Sakaguchi “empty” rag-/- AID CD25- CD4+ T cells E E E T E T T T T AID (Effector T cells, TE, E) or Healthy WT CD25+CD4+T cells R R R R R Healthy R R R (Regulatory T cells, TR, R)
  • Tolerance is mediated by regulatory T cells CTL CTL CTL CTL CTL CTL TH CTLSelfAntigen TH TH TH TH TH TH TH TH B B B B B B TH B
  • TOLERANCE = absence of clonal expansion = control by regulatory T cells Therapy of autoimmune diseases should aim at stimulating autoreactive regulatory T cells or clones
  • … BUT …there are many open questions
  • CD25+CD4+T cells R R R R R R R R (Regulatory T cells, TR, R)
  • What makes a T cellbe a regulatory T cell ? CD25+CD4+T cells R R R R R R R R (Regulatory T cells, TR, R)
  • What makes a T cell be a regulatory T cell ? CD25+CD4+T cells R RHow do they interact R R R R R R with other cells ? (Regulatory T cells, TR, R)
  • What makes a T cell be a regulatory T cell ? CD25+CD4+T cells R RHow do they interact R R R R R R with other cells ? (Regulatory T cells, TR, R) How do they prevent autoimmune diseases ?
  • What makes a T cell be a regulatory T cell ? CD25+CD4+T cells R RHow do they interact R R R R R R with other cells ? (Regulatory T cells, TR, R) How do they allow efficient immune responses ? How do they prevent autoimmune diseases ?
  • What makes a T cell be a regulatory T cell ? How many and how CD25+CD4+T cells diverse are they ? R RHow do they interact R R R R R R with other cells ? (Regulatory T cells, TR, R) How do they allow efficient immune responses ? How do they prevent autoimmune diseases ?
  • How does their repertoire What makes a T cell compare to that of other T be a regulatory T cell ? cells ? How many and how CD25+CD4+T cells diverse are they ? R RHow do they interact R R R R R R with other cells ? (Regulatory T cells, TR, R) How do they allow efficient immune responses ? How do they prevent autoimmune diseases ?
  • Balance between regulatory and effector T cell subpopulations
  • Is it that simple ?
  • Is it that simple ?“Everything should be made as simple as possible, But not simpler.” Albert Einstein, XX century
  • http://www.medscape.com/content/2004/00/46/84/468446/468446_fig.html
  • http://www.medscape.com/content/2004/00/46/84/468446/468446_fig.html
  • http://www.sewingmachines.us/sewing-machine-636.jpghttp://www.medscape.com/content/2004/00/46/84/468446/468446_fig.html
  • http://www.sewingmachines.us/sewing-machine-636.jpg http://www.roselabiche.com/blog/index.php?2007/11http://www.medscape.com/content/2004/00/46/84/468446/468446_fig.html
  • http://web.mit.edu/2.972/www/reports/sewing_machine/a-sewing_machine.gif
  • Cell population dynamics
  • J. theor. Biol. (2000) 207, 231}254 doi:10.1006/jtbi.2000.2169, available online at http://www.idealibrary.com on Modelling T-cell-Mediated Suppression Dependent on Interactions in Multicellular Conjugates KALET LEON*-?, ROLANDO PEREZ*, AGUSTIN LAGE* D D AND JORGE CARNEIRO- MODELLING T LYMPHOCYTE LINKED SUPPRESSION 249 *Centro de ImmunolognH a Molecular, P.O. Box 16040, Habana 11600, Cuba and -Instituto Gulbenkian de Ciencia, Apartado 14, 2781-901, Oeiras, Portugal ore than two cell types with the alternative mechanisms of linked sup-mechanisms can be speci- (Received on 7 February 2000, Accepted in revised form onreported 2000) pression. Overall, the modelling results 10 Augustsetting of the interaction here and the whole set of observations that we the consequences of each discussed would strongly favour two candidate Tolerance to peripheral body antigens involves multiple mechanisms, namely T-cell-mediatedable 1). The application of suppression of potentially autoimmuneare Recent in vivo and in vitro evidence indicates that mechanisms. These cells. the ones that are trans- regulatory T cells suppress the response of e!ector T cells by a mechanism that requires the tions other than the cur- lated by the "nal model: regulatory T cells inhibit simultaneous conjugation of regulatory and e!ector T cells with the same antigen-presentingminant tolerance can be (APC). Despite this strong requirement, it is not yet clearwhile theywhile both cells are cell the proliferation of e!ector cells what happens are conjugated. Several hypotheses are discussed in the literature. Suppression may result from For example, it was competition of regulatory and e!ectoracells for activation resources on the APC; simple nevertheless dependent on growth factor that the the interaction between regulatory T cellsproduce; an inhibitory signal tocells inhibit the same conjugate; or latter may deliver or regulatory e!ector T cells in the ex- e!ector T cells may acquire the regulatory phenotype during their interaction with regulatory Ls does not involve thecells. The present article tries to further our understanding of T-cell-mediated suppression, T pansion of the population of e!ector cells because they convertnumber of multicellular conjugates of T propose the "rst generalns between the two cells to narrow-downthe formation of to the regulatoryWe cells and APCs. Using this and formalism describing the them candidate mechanisms. phenotype. formalism we derive three particularor both may be operative in of T-cell- resenting cell, as classi- Either mechanism models, representing alternative mechanisms mediated suppression. For each model, we make phase plane and bifurcation analysis, and son & OMalley, 1987; vivo, maybe even dependent on with the history of identify their pros and cons in terms of the relationshipthe life large body of experimentalhat the APC may act as observations on T-cell-mediated Modigliani et al.that accounting for the quantitative the e!ector cells. suppression. We argue (1996b) actually details of adoptive transfers of tolerance requires models with bistable regimes in which eithernnett et al., 1998; Ridge regulatory cells or e!ectors cells dominate the steady cellsFrom this thymic conclude demonstrated that regulatory state. from analysis, wer et al., 1998). The formal- cells actively inhibit mechanism of T-cell-mediated suppression the maintenance of the that the most plausible T epithelium the growth of can both suppressrequires re- regulatory chimeras e!ector T cells, and that the that
  • Modelling T cell population dynamicsnet growth = influx + interaction-dependent — death growth dTi = i + ↵i (T, A, m, k) · Ti · Ti dt
  • Modelling T cell population dynamicsnet growth = influx + interaction-dependent — death growth dTi = i + ↵i (T, A, m, k) · Ti · Ti dt
  • Modelling T cell population dynamicsnet growth = influx + interaction-dependent — death growth dTi = i + ↵i (T, A, m, k) · Ti · Ti dt
  • Modelling T cell population dynamicsnet growth = influx + interaction-dependent — death growth dTi = i + ↵i (T, A, m, k) · Ti · Ti dt
  • Modelling T cell population dynamicsnet growth = influx + interaction-dependent — death growth dTi = i + ↵i (T, A, m, k) · Ti · Ti dt How to choose appropriate functional forms for the interaction terms ?
  • Modelling T cell population dynamicsA taxonomy of putative mechanisms of cell-to-cell interaction R T IL-X R R [Leon et al. J Theor Biol 2000; Carneiro et al. Immunol Rev 2007]
  • Modelling T cell population dynamicsA taxonomy of putative mechanisms of cell-to-cell interaction R T IL-X R R [Leon et al. J Theor Biol 2000; Carneiro et al. Immunol Rev 2007]
  • Modelling T cell population dynamicsA taxonomy of putative mechanisms of cell-to-cell interaction R T IL-X R R CRM [Leon et al. J Theor Biol 2000; Carneiro et al. Immunol Rev 2007]
  • Modelling T cell population dynamics: the crossregulation model Interactive simulation
  • Quantitative assay for cell proliferation+CFSE Nº Cells Log FL1-H Log Intensity of CFSE staining [Leon et al. J.Theor.Biol. 2004]
  • Putting the CRM to the test + anti-CD3 antibody APCs Cell Cycle R R R R E R R R R R R R RB6 Thy1.2 Analysis CD25+CD4+Thy1.2+ 3 day culture E E E E E E E E E E E E E E E E E EB6 congenic Thy1.1 CD25-CD4+Thy1.1+ [Carneiro et al. Immuno. Rev. 2007]
  • Putting the CRM to the test + anti-CD3 antibody APCs Proportions Effector Cells Regulatory Cells Thy1.1:Thy1.2 Thy1.1 Thy1.2 Cell Cycle R R R E R R R R R Day 0 Day 3 R R R R Analysis CD25+CD4+Thy1.2+ ? ? 3 day culture E E E E E E E E E + + anti-CD3100:0 N.D. E E E E E E antibody E E E NDy1.1 CD25-CD4+Thy1.1+ 80:20 47:53 Cell Cycle R R E R R R R Analysis.2+ + anti-CD3 antibody 50:50 24:76 APCs 3 day culture E E E E E E E + E + anti-CD30:100 N.D. Cycle antibody E Cell R R R R E R R R R R R R R ND Analysis1.1+ CD25+CD4+Thy1.2+ 3 day culture Log FL1-H / CFSE intensity E E E E E E E E E E E E E E Cell Cycle R E R E E E E R R R R CD25-CD4+Thy1.1+ Analysis et al. Immuno. Rev. 2007] [Carneiro.2+
  • Putting the CRM to the test + anti-CD3 antibody APCs Proportions Effector Cells Regulatory Cells Thy1.1:Thy1.2 Thy1.1 Thy1.2 Cell Cycle R R R E R R R R R Day 0 Day 3 R R R R Analysis CD25+CD4+Thy1.2+ ? 3 day culture E E E E E E E E E + + anti-CD3100:0 N.D. E E E E E E antibody E E E NDy1.1 CD25-CD4+Thy1.1+ 80:20 47:53 Cell Cycle R R E R R R R Analysis.2+ + anti-CD3 antibody 50:50 24:76 APCs 3 day culture E E E E E E E + E + anti-CD30:100 N.D. Cycle antibody E Cell R R R R E R R R R R R R R ND Analysis1.1+ CD25+CD4+Thy1.2+ 3 day culture Log FL1-H / CFSE intensity E E E E E E E E E E E E E E Cell Cycle R E R E E E E R R R R CD25-CD4+Thy1.1+ Analysis et al. Immuno. Rev. 2007] [Carneiro.2+
  • Putting the CRM to the test + anti-CD3 antibody APCs Proportions Effector Cells Regulatory Cells Thy1.1:Thy1.2 Thy1.1 Thy1.2 Cell Cycle R R R E R R R R R Day 0 Day 3 R R R R Analysis CD25+CD4+Thy1.2+ 3 day culture E E E E E E E E E + + anti-CD3100:0 N.D. E E E E E E antibody E E E NDy1.1 CD25-CD4+Thy1.1+ 80:20 47:53 Cell Cycle R R E R R R R Analysis.2+ + anti-CD3 antibody 50:50 24:76 APCs 3 day culture E E E E E E E + E + anti-CD30:100 N.D. Cycle antibody E Cell R R R R E R R R R R R R R ND Analysis1.1+ CD25+CD4+Thy1.2+ 3 day culture Log FL1-H / CFSE intensity E E E E E E E E E E E E E E Cell Cycle R E R E E E E R R R R CD25-CD4+Thy1.1+ Analysis et al. Immuno. Rev. 2007] [Carneiro.2+
  • Putting the CRM to the test + anti-CD3 antibody APCs Proportions Effector Cells Regulatory Cells Thy1.1:Thy1.2 Thy1.1 Thy1.2 Cell Cycle R R R E R R R R R Day 0 Day 3 R R R R Analysis CD25+CD4+Thy1.2+ 3 day culture E E E E E E E E E + + anti-CD3100:0 N.D. E E E E E E antibody E E E NDy1.1 CD25-CD4+Thy1.1+ 80:20 47:53 Cell Cycle R R E R R R R Analysis.2+ + anti-CD3 antibody 50:50 24:76 ? ? APCs 3 day culture E E E E E E E + E + anti-CD30:100 N.D. antibody E Cell Cycle R R R R E R R R R R R R R ND Analysis1.1+ CD25+CD4+Thy1.2+ 3 day culture Log FL1-H / CFSE intensity E E E E E E E E E E E E E E Cell Cycle R E R E E E E R R R R CD25-CD4+Thy1.1+ Analysis et al. Immuno. Rev. 2007] [Carneiro.2+
  • Putting the CRM to the test + anti-CD3 antibody APCs Proportions Effector Cells Regulatory Cells Thy1.1:Thy1.2 Thy1.1 Thy1.2 Cell Cycle R R R E R R R R R Day 0 Day 3 R R R R Analysis CD25+CD4+Thy1.2+ 3 day culture E E E E E E E E E + + anti-CD3100:0 N.D. E E E E E E antibody E E E NDy1.1 CD25-CD4+Thy1.1+ 80:20 47:53 Cell Cycle R R E R R R R Analysis.2+ + anti-CD3 antibody 50:50 24:76 APCs 3 day culture E E E E E E E + E + anti-CD30:100 N.D. Cycle antibody E Cell R R R R E R R R R R R R R ND Analysis1.1+ CD25+CD4+Thy1.2+ 3 day culture Log FL1-H / CFSE intensity E E E E E E E E E E E E E E Cell Cycle R E R E E E E R R R R CD25-CD4+Thy1.1+ Analysis et al. Immuno. Rev. 2007] [Carneiro.2+
  • Putting the CRM to the test + anti-CD3 antibody APCs Proportions Effector Cells Regulatory Cells Thy1.1:Thy1.2 Thy1.1 Thy1.2 Cell Cycle R R R E R R R R R Day 0 Day 3 R R R R Analysis CD25+CD4+Thy1.2+ + anti-CD3 antibody 3 day culture APCsE EE EE E E E E + + anti-CD3100:0 N.D. antibody E E E E E E E E E NDy1.1 CD25-CD4+Thy1.1+ Cell Cycle R R R E R R R R R + anti-CD3 antibody ? ? R R R R AnalysisCD25+CD4+Thy1.2+ APCs 80:20 47:53 Cell Cycle R R E 3 day culture R R + + anti-CD3 antibody R ? ? + E E E E R E E E E Analysis Cell Cycle E E E E E E R E E E RE R E R R R R R R R R R.2+CD25-CD4+Thy1.1+Thy1.2 Analysis CD25+CD4+Thy1.2+ 50:50 24:76 3 day culture 3 day culture Cell Cycle R R EE E E E E E E E E E R E ER E E E E E E R E E E E E E E R E E Analysis 0:100 N.D.ongenic Thy1.1 ND CD25-CD4+Thy1.1++1.1+ 3 day culture Log FL1-H / CFSE intensity E E E E E E E E E [Carneiro et al. Immuno. Rev. 2007]+
  • Putting the CRM to the test Proportions Effector Cells Regulatory CellsThy1.1:Thy1.2 Thy1.1 Thy1.2Day 0 Day 3100:0 N.D. ND80:20 47:5350:50 24:760:100 N.D. ND Log FL1-H / CFSE intensity [Carneiro et al. Immuno. Rev. 2007]
  • Modelling T cell population dynamics: the crossregulation model R T IL-X R R CRM [Leon et al. J Theor Biol 2000; Carneiro et al. Immunol Rev 2007] [Carneiro et al. Immuno. Rev. 2007]
  • Modelling T cell population dynamics: the crossregulation model Regulatory T cell populations grow as a function of the effector T cells they suppress R R Leon et al. JTB (2000) R E E E 01 . Imm unol 20 iol 2 000 , J al. J Theor B eon et Thesis L alh o PhD sis , Caram hD The
  • the control of the insulin promoter. We had previously observedModelling T cell population dynamics: the crossregulation model Authorship note: Yenkel Grinberg-Bleyer and David Saadoun contributed equally to this work. Eliane Piaggio and Benoît L. Salomon are co–senior authors. that HA-specific Tregs (HA-Tregs) preferentially proliferated and expanded at days 5–7 after transfer in draining pancreatic LNs Conflict of interest: The authors have declared that no conflict of interest exists. (PLNs) of ins-HA homozygous mice (29). When we repeated the Citation for this article: J Clin Invest. 2010;120(12):4558–4568. doi:10.1172/JCI42945. experiment in ins-HA hemizygous recipients, which express lower The Journal of Clinical Investigation http://www.jci.org Volume 120 Number 12 December 2010 Related Commentary, page 4190 Regulatory T cell populations grow as a function of the effector T cells they suppress Yenkel Grinberg-Bleyer,1,2,3 David Saadoun,1,2,3 Audrey Baeyens,1,2,3 Fabienne Billiard,1,2,3 Jérémie D. Goldstein,1,2,3 Sylvie Grégoire,1,2,3 Gaëlle H. Martin,1,2,3 Rima Elhage,1,2,3 Nicolas Derian,1,2,3 Wassila Carpentier,1,4 Gilles Marodon,1,2,3 David Klatzmann,1,2,3 R R Eliane Piaggio,1,2,3 and Benoît L. Salomon1,2,3 1Université Pierre et Marie Curie — Univ Paris 06, 2CNRS UMR 7211, 3INSERM U959, Paris, France. 4Plate-forme Post-Génomique P3S, Hôpital Pitié-Salpêtrière, Paris, France. Leon et al. JTB (2000) CD4+CD25+Foxp3+ Tregs play a major role in prevention of autoimmune diseases. The suppressive effect of Tregs on effector T cells (Teffs), the cells that can mediate autoimmunity, has been extensively studied. How- R E ever, the in vivo impact of Teff activation on Tregs during autoimmunity has not been explored. In this study, we have shown that CD4 + Teff activation strongly boosts the expansion and suppressive activity of Tregs. This helper function of CD4+ T cells, which we believe to be novel, was observed in the pancreas and draining lymph nodes in mouse recipients of islet-specific Teffs and Tregs. Its physiological impact was assessed in autoimmune diabetes. When islet-specific Teffs were transferred alone, they induced diabetes. Paradoxically, when the same Teffs were cotransferred with islet-specific Tregs, they induced disease protection by boost- E ing Treg expansion and suppressive function. RNA microarray analyses suggested that TNF family members were involved in the Teff-mediated Treg boost. In vivo experiments showed that this Treg boost was partially E dependent on TNF but not on IL-2. This feedback regulatory loop between Teffs and Tregs may be critical to preventing or limiting the development of autoimmune diseases. bers of mature DCs in inflamed tissues may favor the activation of 01 unol 20 The peripheral T cell repertoire of any individual contains autoreac- autoreactive Tregs (21–23), which would then turn on their sup- tive T cells specific for a variety of self antigens (1). Their activation pressive activity and exert bystander suppression (24, 25). Thus, during an autoimmune process, there is a local enrichment 00 , J . Imm could induce an autoimmune process, eventually leading to an auto- Bi ofl 2 0 of both autoreactive Teffs and Tregs. Since heefficacy o Treg- immune disease. Severe and prolonged inflammation in a tissue may J T the orbetween acti- mediated suppression depends onal. equilibrium eon et the factor that could tip this Thesis o PhD lead to the activation of pathological autoreactive T cells by several L Tregs (26), any mechanisms (2, 3). At the site of inflammation or in draining LNs, vated Teffs and activated balance to one side or the other could then determine thes, Ca ramalh tissue damage results in an enhanced presentation of autoantigens hD Th esi outcome
  • How does their repertoire What makes a T cell compare to that of other T be a regulatory T cell ? cells ? How many and how CD25+CD4+T cells diverse are they ? R RHow do they interact R R R R R R with other cells ? (Regulatory T cells, TR, R) How do they allow efficient immune responses ? How do they prevent autoimmune diseases ?
  • How does their repertoire What makes a T cell compare to that of other T be a regulatory T cell ? cells ? How many and how CD25+CD4+T cells diverse are they ? R RHow do they interact R R R R R R with other cells ? (Regulatory T cells, TR, R) How do they allow efficient immune responses ? How do they prevent autoimmune diseases ?
  • autoimmune disease and immunopathology, 2 2 2 2 B Foxp3 represents a more specific marker than 5RBhi alone +Foxp3/MIGR1 +MIGR1 None CD25-CD45RBhi alone cell-surface molecules (such as currently used +Foxp3/MIGR1 +MIGR1 None CD25, CD45RB, CTLA-4, and GITR), which are unable to completely discriminate between regulatory T cells and activated, effector, or Colon memory T cells. Mutations in the Foxp3 gene culminate in the development of a fatal lymphoprolifera- tive disorder associated with multiorgan pa- thology both in mice and humans (12–20). FOXP3 is predominantly expressed in human CD25ϩCD4ϩ T cells as well (32). Further- Stomach more, transduction of a mutant Foxp3 lacking the forkhead domain, similar to the mutated Foxp3 in scurfy mice (17), failed to confer suppressive activity to CD25–CD4ϩ T cells Fig. 4. Prevention(fig.IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid of S7). The present results therefore sug- 105 fresh CD25–CD45RBhighCD4ϩ cells either alone (n ϭ 6, where n is the 4 ϫ 10that mutations of thehighCD4ϩgene may alone (n ϭ 6, where n is the mice received gest 5 fresh CD25–CD45RB Foxp3 cells either derived of mice)cause squares)disorders through1.2develop- GFPϩ sorted cells derived from pen squares) or together with 1.2 ϫ 106 GFPϩ sorted cells number from (open these or together with ϫ 106D4ϩ cells infected with Foxp3/MIGR1 (n ϭ 7) (closed circles) or MIGR1 –(n ϭ 5) highmentalcells infected with Foxp3/MIGR1 (n ϭthe (closed circles) or MIGR1 (n ϭ 5) CD25 CD45RB CD4ϩ or functional abnormality of 7) Astericks by R Cell population dynamics weight is represented as the percentage of initial weight (mean Ϯ SD). Body weight is represented as the percentage of initial weight (mean Ϯ SD) (open circles). CD25ϩCD4ϩ T population. gnificant difference, P Ͻ 0.01, Foxp3/MIGR1 versus other two groups indicate significant difference, P of T cells by Foxp3 versus other two groups by Potentially, generation Ͻ 0.01, Foxp3/MIGR1 (ODE) R (B) Histopathology of the colon and stomach in each group and in an test. (B) Histopathology of the colon and stomach in each group and in an Mann-Whitney transduction of naıve T cells may provide a ¨ D mouse (None). (C) Colitis (left) and gastritis (right) were unreconstituted SCID mouse (None). (C) Colitis (left) and gastritis (right) were histologically histologically the group cotransferred with MIGR1-infected cells and one transferredTwo micepreviously unstudied therapeutic mode for cells and one transferred with scored. with in the group cotransferred with MIGR1-infected ϩ CD25 CD45RB treatment of autoimmunedebilitation before histological examination. Results –D4 cells alone died of debilitation before histological examination. Results high CD4ϩ cells alone died of and inflammatory are from a total of three independent experiments. shown in (A) to (C) are from a total of three independent experiments. diseases and in transplantation tolerance. 1060 14 FEBRUARY 2003 VOL 299 SCIENCE www.sciencemag.org 14 FEBRUARY 2003 VOL 299 SCIENCE www.scien Gene regulatory networks (Logical network dynamics)
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? Regulatory T cells are enriched in CD25+CD4+ T cell pool of healthy animals Do not produce IL-2 Produce TGF-beta Express transcription factor Foxp3
  • Modelling gene regulatory networks and T cell differentiation What makes a T cell be a regulatory T cell ? Regulatory T cells are enriched in R CD25+CD4+ T cell pool of healthy animals R R Do not produce IL-2 E Produce TGF-beta E E Express transcription factor Foxp3n et al. J Theor Biol 2000 , J. Immunol 2001 valho PhD Thesis, Caramalho PhD Thesisvin Nat. Immunol. 2002aille et al J.Immunol. 2002 eida et al. J.Immunol. 2002rnton J.Immunol 2004rneiro et al. Immunol Rev. 2007
  • Modelling gene regulatory networks and T cell differentiation What makes a T cell be a regulatory T cell ? Regulatory T cells are enriched in R CD25+CD4+ T cell pool of healthy animals R R Do not produce IL-2 E Produce TGF-beta E E Express transcription factor Foxp3n et al. J Theor Biol 2000 , J. Immunol 2001 valho PhD Thesis, Caramalho PhD Thesis But how come they have this gene expression profile ?vin Nat. Immunol. 2002aille et al J.Immunol. 2002 Is this gene expression profile stable ? eida et al. J.Immunol. 2002rnton J.Immunol 2004rneiro et al. Immunol Rev. 2007
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? ThP/0 IL- 2, T G IL -12 Fβ GF β IL-4 IL-6, T TRTh1 Th17 Th2
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? ThP/0 TR IL- 2, T L- 12 GF I β β IL-6, TGF IL-4Zhou et al, Nat. Immunol. (2009) 10 , 1000Duarte et al. Eur. J. Immunol. (2009) 39 , 948
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? ? ThP/0 TR IL- 2, T L- 12 GF I β β ? IL-6, TGF IL-4 ? ??Zhou et al, Nat. Immunol. (2009) 10 , 1000Duarte et al. Eur. J. Immunol. (2009) 39 , 948
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? T cell gene regulatory network graph Diversity
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? Logical regulatory graphs A regulatory graph is a labelled directed graph where nodes represent genes or components and arcs represent regulatory interactions labelled with a sign when the interaction corresponds to an activation or an inhibition Let G = {G1 , G2 , G3 , ..., Gn } be the set of nodes or components For each Gi G the current activity level is denoted by xi {0, 1, ...,Maxi } Logical rules governing behavior of the nodes or components Kaufman SA (1969), Thomas R (1978)
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? Logical regulatory graphs 1 g1 1 g2 1 2 1 g3
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? Logical state transition graphs Represent the dynamics of the systems represented by logical regulatory graphs, and are defined as: A set of nodes and arcs connecting the nodes The nodes represent the states of the system, defined by a word in temporal logics representing the concatenation of the level of each component of the regulatory network The arcs represent transitions between the pairs of states An updating method specifying the order of the transitions (synchronous vs asynchronous updating)
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? Logical state transition graphs (Asynchronous non-deterministic updating)
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? Logical state transition graphs (Asynchronous non-deterministic updating) cycle stable states
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? The logics of T-cell regulatory networks    NFAT=1 if CD28==1 ⋀ TCR==1 NFAT=0 otherwise   
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? Building blocks of Th cell regulatory networks                    
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? T cell gene regulatory network graph with 65 components Diversity
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? 65 components leads to over 1019 states !!!
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? 65 components leads to over 1019 states !!! A dedicated computational tool - Ginsim http://gin.univ-mrs.fr/ Automatic finding of steady states based on functional circuits Automatic reduction of the network with preservation of functional circuits see: Naldi (2009)
  • Diversity and PlasticModelling gene regulatory networks and T cell differentiation T cell gene regulatory network graph with 34 components Naldi et al. PLoS Comp Biol (2010) What makes a T cell be a regulatory T cell ? Figure 3. Reduced Th regulatory graph, encompassing 34 components. This graph has been obtained by applying the described in Section ‘‘Model reduction’’ to the full model shown in Figure 2. Indirect interactions resulting from the reduction a dotted lines. Greyed-out components can be further reduced to generate a more compact model, which still keeps the differentiation markers. doi:10.1371/journal.pcbi.1000912.g003 PLoS Computational Biology | www.ploscompbiol.org 9 September 2010 | Volume 6 | I
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? T cell lineages and the stable states of the regulatory network Diversity and Plastic Figure 4. Definition of alternative Th subtypes based on the expression of the master regulators. Each of the considered (TBET, GATA3, RORGT and FOXP3) is positively auto-regulated. The first five rows correspond to the canonical Th cell s
  • Diversity and Plastic Diversity and Plastic T cell lineages and the stable states of the regulatory networkModelling gene regulatory networks and T cell differentiation What makes a T cell be a regulatory T cell ? Figure 3. Reduced Th regulatory graph, encompassing 34 components. This graph has been obtained by applying the Figure 3. Reduced Th regulatory graph, encompassing 34 components. This graph has been obtained by applying the described in Section ‘‘Model reduction’’ to the full model shown in Figure 2. Indirect interactions resulting from the reduction a dotted lines. Greyed-out components can be further reduced to generate a more compact model, which still keeps the differentiation markers. doi:10.1371/journal.pcbi.1000912.g003 PLoS Computational Biology | www.ploscompbiol.org 9 September 2010 | Volume 6 | I
  • raph, encompassing 34 components. This graph has been obtained by applying the reduction’ to the full model shown in Figure 2. Indirect interactions resulting from the reduction are display Modelling gene regulatory networks and T cell differentiation What makes a T cell be a regulatory T cell ? T cell lineages and the stable states of the regulatory network see also: Burda et al. PNAS 2011; Hegazy et al. Immunity 2010
  • the master regulators, i.e. that show hybrid patterns. Additional positive circuits (proliferation and STAT3-related) geneModelling gene regulatory networks and T cell differentiation circuit analysis predicts 48 stable patterns (4 for each of the 12 groups; each pattern corresponds to one cell of the table circuits’’). Only 28 of these patterns (greyed cells) are compatible with at least one of the input combinations considered heWhat makes a T cell be a regulatory T cell ? in the cells indicate how many input combinations are compatible with this stable state. Five patterns are not compatible w (cells with dashes). doi:10.1371/journal.pcbi.1000912.g004 It is likely that components or links were missed. To this adds the This cautionary remark notwithstand problem of the definition of the logical functions driving the recapitulates the differentiation of naive c behaviour of the components, particularly those involving complex and Treg subtypes. Strikingly, our model Immunologically relevant environments and input vectors regulatory mechanisms. states expressing markers characteristic o Figure 5. Environmental conditions used for the simulations. Each row corresponds to one prototypic environ combinations of APC and of seven different cytokine inputs. Presence/absence of the different inputs is denoted by grey/w code defined in the first column is used in Figure 7 to denote environmental conditions.
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? Diversity and Plasticity of Th Cell Types Figure 6. Context-dependent stable states and their component expression patterns. A grey cell denotes the activation of the
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? Diversity and Plasticity of Th Cell Types Figure 7. Stability of Th cell subtypes and environment-dependent transitions. This figure summarises several simulation rounds,
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? a b c d e f
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? The structure of the gene network and the micro-environment make it
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? The structure of the gene network and the micro-environment make it Regulatory T cell phenotype is not stable accross different micro-environments
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? The structure of the gene network and the micro-environment make it Regulatory T cell phenotype is not stable accross different micro-environments Regulatory T cells are not autonomous They depend on IL-2-producing effector T cells to sustain their cellular identity
  • Modelling gene regulatory networks and T cell differentiationWhat makes a T cell be a regulatory T cell ? BUT ... the microenvironment depends on the T cell population dynamics !!!
  • autoimmune disease and immunopathology, 2 2 2 2 B Foxp3 represents a more specific marker than 5RBhi alone +Foxp3/MIGR1 +MIGR1 None CD25-CD45RBhi alone cell-surface molecules (such as currently used +Foxp3/MIGR1 +MIGR1 None CD25, CD45RB, CTLA-4, and GITR), which Multiple are unable to completely discriminate between regulatory T cells and activated, effector, or memory T cells. scales Colon Mutations in the Foxp3 gene culminate in the development of a fatal lymphoprolifera- tive disorder associated with multiorgan pa- thology both in mice and humans (12–20). FOXP3 is predominantly expressed in human CD25ϩCD4ϩ T cells as well (32). Further- Stomach more, transduction of a mutant Foxp3 lacking the forkhead domain, similar to the mutated Foxp3 in scurfy mice (17), failed to confer suppressive activity to CD25–CD4ϩ T cells Fig. 4. Prevention(fig.IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid of S7). The present results therefore sug- 105 fresh CD25–CD45RBhighCD4ϩ cells either alone (n ϭ 6, where n is the 4 ϫ 10that mutations of thehighCD4ϩgene may alone (n ϭ 6, where n is the mice received gest 5 fresh CD25–CD45RB Foxp3 cells either derived of mice)cause squares)disorders through1.2develop- GFPϩ sorted cells derived from pen squares) or together with 1.2 ϫ 106 GFPϩ sorted cells number from (open these or together with ϫ 106D4ϩ cells infected with Foxp3/MIGR1 (n ϭ 7) (closed circles) or MIGR1 –(n ϭ 5) highmentalcells infected with Foxp3/MIGR1 (n ϭthe (closed circles) or MIGR1 (n ϭ 5) CD25 CD45RB CD4ϩ or functional abnormality of 7) Astericks by R Cell population dynamics weight is represented as the percentage of initial weight (mean Ϯ SD). Body weight is represented as the percentage of initial weight (mean Ϯ SD) (open circles). CD25ϩCD4ϩ T population. gnificant difference, P Ͻ 0.01, Foxp3/MIGR1 versus other two groups indicate significant difference, P of T cells by Foxp3 versus other two groups by Potentially, generation Ͻ 0.01, Foxp3/MIGR1 (ODE) R (B) Histopathology of the colon and stomach in each group and in an test. (B) Histopathology of the colon and stomach in each group and in an Mann-Whitney transduction of naıve T cells may provide a ¨ D mouse (None). (C) Colitis (left) and gastritis (right) were unreconstituted SCID mouse (None). (C) Colitis (left) and gastritis (right) were histologically histologically the group cotransferred with MIGR1-infected cells and one transferredTwo micepreviously unstudied therapeutic mode for cells and one transferred with scored. with in the group cotransferred with MIGR1-infected ϩ CD25 CD45RB treatment of autoimmunedebilitation before histological examination. Results –D4 cells alone died of debilitation before histological examination. Results high CD4ϩ cells alone died of and inflammatory are from a total of three independent experiments. shown in (A) to (C) are from a total of three independent experiments. diseases and in transplantation tolerance. 1060 14 FEBRUARY 2003 VOL 299 SCIENCE www.sciencemag.org 14 FEBRUARY 2003 VOL 299 SCIENCE www.scien Gene regulatory networks (Logical network dynamics)
  • autoimmune disease and immunopathology, 2 2 2 2 B Foxp3 represents a more specific marker than 5RBhi alone +Foxp3/MIGR1 +MIGR1 None CD25-CD45RBhi alone cell-surface molecules (such as currently used +Foxp3/MIGR1 +MIGR1 None CD25, CD45RB, CTLA-4, and GITR), which Multiple are unable to completely discriminate between regulatory T cells and activated, effector, or memory T cells. scales Colon Mutations in the Foxp3 gene culminate in the development of a fatal lymphoprolifera- tive disorder associated with multiorgan pa- thology both in mice and humans (12–20). FOXP3 is predominantly expressed in human CD25ϩCD4ϩ T cells as well (32). Further- Stomach more, transduction of a mutant Foxp3 lacking the forkhead domain, similar to the mutated Foxp3 in scurfy mice (17), failed to confer suppressive activity to CD25–CD4ϩ T cells Fig. 4. Prevention(fig.IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid of S7). The present results therefore sug- 105 fresh CD25–CD45RBhighCD4ϩ cells either alone (n ϭ 6, where n is the 4 ϫ 10that mutations of thehighCD4ϩgene may alone (n ϭ 6, where n is the mice received gest 5 fresh CD25–CD45RB Foxp3 cells either derived of mice)cause squares)disorders through1.2develop- GFPϩ sorted cells derived from pen squares) or together with 1.2 ϫ 106 GFPϩ sorted cells number from (open these or together with ϫ 106D4ϩ cells infected with Foxp3/MIGR1 (n ϭ 7) (closed circles) or MIGR1 –(n ϭ 5) highmentalcells infected with Foxp3/MIGR1 (n ϭthe (closed circles) or MIGR1 (n ϭ 5) CD25 CD45RB CD4ϩ or functional abnormality of 7) weight is represented as the percentage of initial weight (mean Ϯ SD). Body weight is represented as the percentage of initial weight (mean Ϯ SD) (open circles). CD25ϩCD4ϩ T population. R gnificant difference, P Ͻ 0.01, Foxp3/MIGR1 versus other two groups indicate significant difference, P of T cells by Foxp3 versus other two groups by Astericks by Potentially, generation Ͻ 0.01, Foxp3/MIGR1 R (B) Histopathology of the colon and stomach in each group and in an test. (B) Histopathology of the colon and stomach in each group and in an Mann-Whitney transduction of naıve T cells may provide a ¨ D mouse (None). (C) Colitis (left) and gastritis (right) were unreconstituted SCID mouse (None). (C) Colitis (left) and gastritis (right) were histologically histologically the group cotransferred with MIGR1-infected cells and one transferredTwo micepreviously unstudied therapeutic mode for cells and one transferred with scored. with in the group cotransferred with MIGR1-infected ϩ CD25 CD45RB treatment of autoimmunedebilitation before histological examination. Results –D4 cells alone died of debilitation before histological examination. Results high CD4ϩ cells alone died of and inflammatory are from a total of three independent experiments. shown in (A) to (C) are from a total of three independent experiments. diseases and in transplantation tolerance. 1060 14 FEBRUARY 2003 VOL 299 SCIENCE www.sciencemag.org 14 FEBRUARY 2003 VOL 299 SCIENCE www.scien Stochastic agent-based multiscale model
  • autoimmune disease and immunopathology, 2 2 2 2 B Foxp3 represents a more specific marker than 5RBhi alone +Foxp3/MIGR1 +MIGR1 None CD25-CD45RBhi alone cell-surface molecules (such as currently used +Foxp3/MIGR1 +MIGR1 None CD25, CD45RB, CTLA-4, and GITR), which Multiple are unable to completely discriminate between regulatory T cells and activated, effector, or memory T cells. scales Colon Mutations in the Foxp3 gene culminate in the development of a fatal lymphoprolifera- tive disorder associated with multiorgan pa- thology both in mice and humans (12–20). FOXP3 is predominantly expressed in human CD25ϩCD4ϩ T cells as well (32). Further- Stomach more, transduction of a mutant Foxp3 lacking the forkhead domain, similar to the mutated Foxp3 in scurfy mice (17), failed to confer suppressive activity to CD25–CD4ϩ T cells Fig. 4. Prevention(fig.IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid IBD and autoimmune gastritis by Foxp3-transduced T cells. (A) C.B-17 scid of S7). The present results therefore sug- 105 fresh CD25–CD45RBhighCD4ϩ cells either alone (n ϭ 6, where n is the 4 ϫ 10that mutations of thehighCD4ϩgene may alone (n ϭ 6, where n is the mice received gest 5 fresh CD25–CD45RB Foxp3 cells either derived of mice)cause squares)disorders through1.2develop- GFPϩ sorted cells derived from pen squares) or together with 1.2 ϫ 106 GFPϩ sorted cells number from (open these or together with ϫ 106 Why to go through suchD4ϩ cells infected with Foxp3/MIGR1 (n ϭ 7) (closed circles) or MIGR1 –(n ϭ 5) highmentalcells infected with Foxp3/MIGR1 (n ϭthe (closed circles) or MIGR1 (n ϭ 5) CD25 CD45RB CD4ϩ or functional abnormality of 7) weight is represented as the percentage of initial weight (mean Ϯ SD). Body weight is represented as the percentage of initial weight (mean Ϯ SD) (open circles). CD25ϩCD4ϩ T population. R gnificant difference, P Ͻ 0.01, Foxp3/MIGR1 versus other two groups indicate significant difference, P of T cells by Foxp3 versus other two groups by Astericks by Potentially, generation Ͻ 0.01, Foxp3/MIGR1 trouble ? R (B) Histopathology of the colon and stomach in each group and in an test. (B) Histopathology of the colon and stomach in each group and in an Mann-Whitney transduction of naıve T cells may provide a ¨ D mouse (None). (C) Colitis (left) and gastritis (right) were unreconstituted SCID mouse (None). (C) Colitis (left) and gastritis (right) were histologically histologically the group cotransferred with MIGR1-infected cells and one transferredTwo micepreviously unstudied therapeutic mode for cells and one transferred with scored. with in the group cotransferred with MIGR1-infected ϩ CD25 CD45RB treatment of autoimmunedebilitation before histological examination. Results –D4 cells alone died of debilitation before histological examination. Results high CD4ϩ cells alone died of and inflammatory are from a total of three independent experiments. shown in (A) to (C) are from a total of three independent experiments. diseases and in transplantation tolerance. 1060 Multiscale models will be necessary to 14 FEBRUARY 2003 VOL 299 SCIENCE www.sciencemag.org 14 FEBRUARY 2003 VOL 299 SCIENCE www.scien rationally design molecular therapies
  • Building a multiscale model To address the question: How can regulatory T cell populations persist stably if the Foxp3 cell differentiation program is unstable ? To try to resolve a controversy:Some authors described plasticity of regulatory T cell and of the Foxp3- dependent differentiation program (Zhou et al Nat Immunol 2009, Duarte et al Eur J Immunol 2009) Other authors described no plasticity or reprogramming (Rubtsov et al. Science 2010 ), Yet others propose that regulatory population is heterogeneous (Miyao et al. Immunity 2012)
  • Building a multiscale model Consider an individual agent representing a T cell The state of the agent is defined as: (g, k, c)g is the state of the gene regulatory logical network (with 65-2 components)k is the conjugation state of the cell (free or conjugated to APC)c is the cell cycle state of the(G0,G1,S,G2/M, apoptotic) cell
  • Building a multiscale model Consider an individual agent representing an APC The state of the agent is defined as: ( , k)is the input vector to the network of the conjugated T cells (e.g. juxtacrine cytokines) k is the conjugation state of the APC
  • Building a multiscale modelConsider now the population dynamics of T cells and APC agents. Assuming that the system is Markovian and the waiting times for transitions are exponentially distributed,that at the transition between G2/M and G1 the T cell agents are cloned/duplicated preserving the remaining state variables, Under these conditions the multiscale model can be simulated using the Gillespie algorithm
  • Building a multiscale modelBinomial heap data structure to accelerate the Gillespie algorithm 8 head[H] 2.7 1.8 0.2 4.7 13.2 18.1 26.2 17.9 16.7 29.2 head[H] 2.7 9.8 head[H] 3.7 21.3 (a) 18.1 26.2 head[H] 2.7 1.8 0.2 29.2 (a) 4.7 13.2 18.1 26.2 head[H] 2.7 3.7 9.8 17.9 16.7 29.2 18.1 26.2 21.3 29.2 (b) (b) head[H] 2.7 1.8 head[H] 26.2 18.1 13.2 head[H] 2.7 9.8 3.7 4.7 29.2 17.9 16.7 18.1 26.2 29.2 21.3 (c) (c) Figure 4: Insertion of new event into binomial heap. (a) A binomial heap H with 5 nodes and the heap H consisting of the new node to be inserted. (b) Merging of the two heaps results in H 2.7 head[H] having two 1.8 binomial trees B0 . (c) The two trees B0 are united as shown, resulting in binomial tree B1 . The heap H now has at most one root of each degree. 26.2 13.2 18.1 4.7 17.9 16.7 29.2 21.3
  • Building a multiscale modelT cell gene regulatory network graph with 65 components Diversity
  • Building a multiscale model G0 G0 G1 G1 G1 NFAT+ NFAT+ IL+ NFAT+ IL+ S S S NFAT+ S NFAT+ IL+ IL+ G2/M G2/M G2/M G1 NFAT+ NFAT+ G2/M IL+ G0 IL+ G1 G1 G1 G1 NFAT+ NFAT+ IL+ IL+death rate is thesame in every state :-(
  • Building a multiscale model How to assess the model and set its parameter values ? Define a few classes for the parameters according to the biology involved: binding rates > signaling rates > transcriptional/translational ratesDistribution of cell cycle phase sojourn times from experimental data (Weber et al. submitted) Minimal calibration of the model: can we reproduce classic quantitative results ?
  • Building a multiscale model Reproducing the classics:antigen-dose dependence of cell cycle differentiation
  • the antigen doses. We thus proceeded to dissect these mech- ent CD4+LECAM-11~ subset contained in the total tha CD Building a multiscale model T cell further. anisms population. This CD4 + T cell subset has previo been shown to beataHigh Antigen significantNot Require oflL- Th2 Development source of Doses Does amounts min dev the Presence of T Cells Having a Memory Phenotype. primary cultures and thereby capable ofThe de- influencing velopment ofa Th2 phenotype at high antigen doses could Reproducing the CD4+LECAM-11~30). This Tdirecting the constitutes N classics:27, subset contained in subset developmen phenotype increased production of endogenous IL-4 by result from development and Th2 cells (16, the cell the total CD4 + antigen-dose dependence of Tcell population.differentiation has previously cell cycle This CD4 + T cell subset been shown to be a source of significant amounts oflL-4 in primary cultures and thereby capable of Cultures Primary Cultures S~ondQry influencing Th phenotype development and directing the development of [OVA323.3391 T Cells Recovered Th2 cells (16, I~ lo4~, q~) 27, 30). This T cell subset constitutes N5% e r 100.00 38.7 elp 50.00 47.4 25.00 Primary Cultures 48,0 S~ondQryCultures H 12.50 51.2 + T ll [OVA323.3391 T Cells Recovered 6.25 58.6 C T D 4 Ce q~) 3.13 I~ lo4~, 49.3 , 1.60 32.3 o rt o n a e ath 100.00 0.80 38.722.2 Rep 50.00 47.4 e s e t in .H 0.40 25.00 48,018.2 itiv Do en ew W 0.20 12.50 04; 0.10 51.213.9 e n p m o d e l Andr 58.6 7.0 efin 6.25 943 ool, 0.053.13l 49.3 1.7 ig nia h hoo ef D nt velo c M 1.60 32.3 ,* ifor ical Sc ity Sc0.03 22.2 0.7 ri uya ra* al d 0.80B A C rs 0,01 0.40 18.2 0.3 t of D e geni hib Gar S l e to, Me ive oAl ~eld on Un 0.20 13.9 40 30 20 10 0 100 200 300 p e ans 0.10 7.0 fe c k o O , Pa f Sh ingt 0.05 1.7 e y azu nne ute un Ef n o t 3-tr K In stit rsityo ,Wash 0.03 I00.00 0,01 0.7 44.2 n 0.3 r m +Te L - - Dendritic he P h e oL[ ,* nd A rch nive ology 50.00 o f i D 440 t h30 20 10 0 41.6 ken y,wa T a s 100 200 300 - ese ogy, U Path 25.00 las o49.3 e c t C C e l l p t o r . Hos ph R A X robiol of 12.50 I00.00 r a t f l c 44.250,9 i r b e e n n41.657.6 d t r-----~ I~1 50.00 o e i- L - - Dendritic N ce y A . Mur y, D al Mic 6.25m to h u e m o49.358.9s n o u l d dr - T- - - ~ 25.00 l o p se a Re anc M log c uno Clini nd w r 3.13 d o 12.50v e n d 50,961.6 es c o 4 + - s - 1,60 h I~1 N th mm d a de t e e 0.80 t i g 6.25 e r ,63.2r 57.6 D i By nne of I tal an dom; ent n a -m 3.13 f a v58.9 ltu e d i y t h 0.40n o w e 61.629.3 n i c C C o n s p - h r y c 22.4eu . lo Ke e l l f o r b ity o lls, g 1,60 . rtm ime Kin 110 pa er d 63 c a il 0.10 0.20 0.80 ce a 63.2n s g ells e v e As m 29.315.2 c De f Exp Unite souri * to h u 0.40 tra in e r n t e d e ab T 0.05n p r i [ 3 -22.414,1 g t h e d . -y) s o f et 0.20 t d he , is h + 0.10 di e t 15.2n8.9 e FN t m t rtmen 2 R X is, M ro pa w h a c c o e s . T C D 4 use0.05 C R - p r e s r e c t e ~/ ( I o u n a n t i - e ly 0.03 14,1 r l F e u c e n ve T e n - 8.9i 5.3 o n d a m o f irn - $10 Lo en ki i ose 0.01 0.03 d l ~.D ffield ne, St. in f l u larg f c y t o n a e n d l 1 . 1 0 a n t i g p,M) 5.3e r f e r e a s e150oses 100e d h i s 0.01 She edici u c 0 50 a n y b e ets o f r o m a n t i g D O t h e - 0 . 6 f i n t i n c r ~M)100 r o d50 d . T lls 50 I00 150 150 d 0 50 I00 150 c a s o g of M e e p r e ce + g e n is m inct s o t y p t t h e n a i v sed a ( 0 . 3 u n t s u c i n . 0 5 b t h a t ( nsg ~ v ( n g ~ )[WN~t](ag/ml)[WN~t](ag/ml) i h e) nt t s [IL-4] b [IL-4] i T 4 a d n a n d t g dis p h e n w t h f r o m e r e u dose a m o s p r o ( < 0 cells as o M - 1 h e C D h 2 - m o J Exp Med 1995 um ose elic ) p h l ig ary f f o r e ted, a u c i n T h 2 e s h o m e n t ells w p t i d e eFa tgeu r e ce1. r yDevelopmenty owaE C Al ort yTh2isphenotype is dependent on i rod or , w elop d B e c pe ri e l l l o w ike dF i gl u r e 1. Development o h a 2 IF o thei k d v e dosehused in primary p T f p T e- - 1 N - fL T h e n oT h lth or Th2 phenotype is dependen g f Flow d o o u s n cytometry-sorted, TCR.- S rt v g m h l - antigen f n f 4 + " cultures. o r e to ) a e n t veCD4 +C D ino r y(2.5mXn cultures. n cell e d s e is ( T h a T r e p o e d e t i v a t e d r a n u c e d o f TtheMantigeno doseoused cellsprimary10S/well)o g e Flow cytometry-sorted, T ls e tf T h on lls of is p c i od t (xl3-transgenic, p, m le T ity e m o p e n nd were cultured with flow h
  • Building a multiscale model Reproducing the classics:antigen-dose dependence of cell cycle differentiation
  • Building a multiscale model Reproducing the classics:NNumber of cells antigen-dose dependence of cell cycle differentiation
  • Building a multiscale modelAre the states and micro-environment dependent state transitions of the individual cells preserved ?
  • Building a multiscale modelAre the states and micro-environment dependent state transitions of the individual cells preserved ? a b c d e f
  • Treg_r Treg Treg_r2_i3 Treg_r2_i2 Treg_r2_i1 Treg_r1_i3 Treg_r1_i2 Treg_r1_i1 Treg_1_i2 Treg_1_i1 Th2_r_i4 Th2_r_i3 Th2_r_i2 Th2_r_i1 Th2_n_i2 Th2_n_i1 Th2_a_i2 Th2_a_i1 Th1_r Th1_n_i4 Th1_n_i3 Th1_n_i2 Th1_n_i1 Th1_a_i3 Th1_a_i2 Th1_a_i1 Th17_n Th17_1_i4 Th17_1_i3 Th17_1_i2 Th17_1_i1 Th0_a_i2 Th0_a_i1 Th0 Th2 Th1Are the states and micro-environment dependent state transitions of the individual cells preserved ? Th0 NFAT-5 Th0 NFAT+ Th2 NFAT- Th1 NFAT- Th2 NFAT+ Th1 NFAT+0 Th2 NFAT- Foxp3+ Th1 NFAT- RORgt+ Th2 NFAT+ Foxp3+ Th1 NFAT+ RORgt+ Treg NFAT+ Th17 NFAT+ Th2 NFAT- RORgt+ Th1 NFAT- Foxp3+ Th2 NFAT+ RORgt+ Treg NFAT- Th1 NFAT+ Foxp3+ Th17 NFAT--5 Treg NFAT- RORgt+ Th2 NFAT- Foxp3+ RORgt+ Th1 NFAT- Foxp3+ RORgt+ Treg NFAT+ RORgt+ Th2 NFAT+ Foxp3+ RORgt+ Th1 NFAT+ Foxp3+ RORgt+-10
  • Building a multiscale modelCan T reg cells persist in this model ?
  • Regulatory T cell persistence in a multiscale model Mean T cell conjugation time = 25 h Effector T reg Mean waiting time of Foxp3+ to Foxp3- transition (min)
  • Regulatory T cell persistence in a multiscale model Mean T cell conjugation time = 25 h Bad news for immunotherapy ? Effector T reg Mean waiting time of Foxp3+ to Foxp3- transition (min)
  • Regulatory T cell persistence in a multiscale model Mean T cell conjugation time = 25 h Effector T reg Mean waiting time of Foxp3+ to Foxp3- transition (min)
  • Regulatory T cell persistence in a multiscale model Mean T cell conjugation time = 25 h Effector 1 day T reg Mean waiting time of Foxp3+ to Foxp3- transition (min)
  • Regulatory T cell persistence in a multiscale model Mean T cell conjugation time = 2 h T reg Effector Mean waiting time of Foxp3+ to Foxp3- transition
  • Regulatory T cell persistence in a multiscale model Mean T cell conjugation time = 2 h 1 week T reg Effector Mean waiting time of Foxp3+ to Foxp3- transition
  • Regulatory T cell persistence in a multiscale model Monoallelic cytokine expression ˜ T Paixao et al Mean waiting time of the Foxp3+ to Foxp3- transition in the range of days to weeks is interpretable as chromatin accessibility transition dynamics Figure 5 Representative time series obtained by simulation of the reversible chromatin modification model of allele expression. For depicts the normalized product concentration (line) and the chromatin competence n/N (gray area). The bottom graph represents t (black) and inactive (white) states. Parameter values: N¼100, a¼0.001 minÀ1, b¼a/0.97 minÀ1, a¼0.1 minÀ1, d¼0.1 min r¼0.0083 minÀ1 and x¼0.1. Combining equation (8) with the recursive relation (7) we can cies of monoallelic and biallelic expressing celPaixão et al. Quantitative insights into stochastic monoallelic expression of cytokine genes calculate the stationary probability distribution for all the N+1 locus protein decay rates, which can be as fast as tra states. on the timescale of the slower remodeling eveImmunol and Cell Biol (2007) We can then calculate the probability of finding the locus in a We conclude that this new model of gene transcriptionally active state in steady state as given by equation (9). cooperative and reversible chromatin remodelin N the transcriptional competence of the two all X an can easily reproduce the experimental obse
  • Regulatory T cell persistence in a multiscale model Mean waiting time of the Foxp3+ to Foxp3- transition in the range of days to weeks is interpretable as chromatin accessibility transition dynamicsPaixão et al. Quantitative insights into stochastic monoallelic expression of cytokine genesImmunol and Cell Biol (2007)
  • Regulatory T cell persistence in a multiscale model Activated TR
  • Regulatory T cell persistence in a multiscale model Activated TR weeks
  • Regulatory T cell persistence in a multiscale model Activated TR weeks we ek
  • Regulatory T cell persistence in a multiscale model Activated TR weeks we ek
  • Regulatory T cell persistence in a multiscale model Activated TR weeks we ek hours le urs yc ho ll c Ce
  • Regulatory T cell persistence in a multiscale model Activated TR weeks we ek hours le urs yc ho ll c Ce
  • Regulatory T cell persistence in a multiscale model Activated TR weeks we ek hours DC le urs yc ho ll c Ce
  • Regulatory T cell persistence in a multiscale model Activated TR weeksDC we ek hours DC le urs yc ho ll c Ce
  • In a multiscale model Is the bistability predicted by the CRM still observed ?Are all the states of the gene regulatory logical network preserved ? If so, how large are their basins of attraction ? Are additional cellular states made possible by feedbacks present at the cell population level ?
  • In a multiscale model Is the bistability predicted by the CRM still observed ? YesAre all the states of the gene regulatory logical network preserved ? If so, how large are their basins of attraction ? Are additional cellular states made possible by feedbacks present at the cell population level ?
  • In a multiscale model Is the bistability predicted by the CRM still observed ? YesAre all the states of the gene regulatory logical network preserved ? Yes If so, how large are their basins of attraction ? Are additional cellular states made possible by feedbacks present at the cell population level ?
  • In a multiscale model Is the bistability predicted by the CRM still observed ? YesAre all the states of the gene regulatory logical network preserved ? Yes If so, how large are their basins of attraction ? ... Are additional cellular states made possible by feedbacks present at the cell population level ?
  • In a multiscale model Is the bistability predicted by the CRM still observed ? YesAre all the states of the gene regulatory logical network preserved ? Yes If so, how large are their basins of attraction ? ... Are additional cellular states made possible by feedbacks present at the cell population level ? Yes, but ...
  • What about the controversy? Th precursor Th precursormethylated foxp3 locus demethylated foxp3 locus foxp3-metil foxp3-demethyl Foxp3+ Foxp3+ Foxp3+ Foxp3+methylated foxp3 locus foxp3-metil foxp3-demethyl demethylated foxp3 locus Foxp3- Foxp3- Foxp3- Foxp3- Interpretation by Miyao et al (Immunity, 2012) Re-interpretation
  • Putting the multiscale CRM to the test Hypotheses-dependent parameter values Initial states: Final states: number of agents/cells number of agents/cells agent/cell GRN state vectors agent/cell GRN state vectors Stochastic multilevel agent-based simulation Foxp3.gfpLentiviral transduction Foxp3GFP+ Treg CMV-Foxp3CFP+ Treg normal naive Th rag-/- in vivo WT cell population dynamics foxp3-/- naive Th foxp3-/-
  • Acknowledgements Íris Caramalho Jocelyne Demengeot Danesh TaraporeNuno Sepúlveda Kalet Leon Vanessa Oliveira Lisa Bergman (IGC, Portugal) Tiago Paixão Carline van den Dool Rui Gardner Iris Vilares Aurelien Naldi, Denis Thieffry Claudine Chaouiya (TAGC, Luminy, France) (IGC, Portugal)
  • http://qobweb.igc.gulbenkian.pt/mailto://jcarneir@igc.gulbenkian.pt