• Aucun résultat trouvé

An age-and-cyclin-structured cell population model with proliferation and quiescence for healthy and tumoral tissues

N/A
N/A
Protected

Academic year: 2021

Partager "An age-and-cyclin-structured cell population model with proliferation and quiescence for healthy and tumoral tissues"

Copied!
29
0
0

Texte intégral

(1)

HAL Id: inria-00081301

https://hal.inria.fr/inria-00081301v2

Submitted on 29 Jun 2006

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

tissues

Fadia Bekkal Brikci, Jean Clairambault, Benjamin Ribba, Benoît Perthame

To cite this version:

Fadia Bekkal Brikci, Jean Clairambault, Benjamin Ribba, Benoît Perthame. An age-and-cyclin- structured cell population model with proliferation and quiescence for healthy and tumoral tissues.

[Research Report] RR-5941, INRIA. 2006. �inria-00081301v2�

(2)

inria-00081301, version 2 - 29 Jun 2006

a p p o r t

d e r e c h e r c h e

N0249-6399ISRNINRIA/RR--5941--FR+ENG

Thème NUM

An age-and-cyclin-structured cell population model with proliferation and quiescence for healthy and

tumoral tissues

Fadia Bekkal Brikci, Jean Clairambault, Benjamin Ribba and Benoît Perthame

N° 5941

Juin 2006

(3)
(4)

Unité de recherche INRIA Rocquencourt

Domaine de Voluceau, Rocquencourt, BP 105, 78153 Le Chesnay Cedex (France)

Téléphone : +33 1 39 63 55 11 — Télécopie : +33 1 39 63 53 30

healthy and tumoral tissues

Fadia Bekkal Briki

, Jean Clairambault

, Benjamin Ribba

and

Benoît Perthame

§

Thème NUM Systèmes numériques

Projet BANG

Rapportde reherhe 5941 Juin 2006 25pages

Abstrat: We present a nonlinear model of the dynamis of a ell popula-

tiondividedinaproliferativeandaquiesentompartments. Theproliferative

phaserepresentstheompleteelldivisionyle(G1−S−G2−M)ofapopula-

tionommittedtodivideatitsend. Themodelisstruturedbythetimespent

by a ellinthe proliferativephase, and by the amountof ylin D/(CDK4 or

6) omplexes. Cellsan transit fromone ompartment tothe other, following

transitionrules whih dieraording tothe tissue state: healthy or tumoral.

The asymptoti behaviour of solutions of the nonlinear model is analysed in

Projet BANG, UR Roquenourt, Institut de Reherhe en Informatique et en Au-

tomatique,BP105,F78153Roquenourt,Frane,Fadia.Bekkal_Brikiinria.fr

Projet BANG, UR Roquenourt, Institut de Reherhe en Informatique et en Au-

tomatique, BP 105, F78153 Roquenourt and INSERM U 776 Rythmes biologiques et

aner, Hpital Paul-Brousse, 14, Av. Paul-Vaillant-Couturier, F94807 Villejuif, Frane,

Jean.Clairambaultinria.fr

Institut de Médeine Théorique, Département de Pharmaologie Clinique EA3736,

FaultédeMédeineR.-T.-H.Laënne,UniversitéLyon1,RueParadin,F69376LyonCedex

08,Frane,brupl.univ-lyon1.fr

§

DépartementdeMathématiquesAppliquées,ÉoleNormaleSupérieure,45,rued'Ulm,

F75005ParisandProjetBANG,URRoquenourt,InstitutdeReherheenInformatique

etenAutomatique,BP105,F78153Roquenourt,Frane, Benoit.Perthameens.fr

(5)

both ases, exhibitingtissue homeostasis ortumour exponentialgrowth. The

model is simulated by numerial solutions whih onrm its theoretial pre-

ditions.

Key-words: Cellyle, Populationdynamis,PartialDierentialEquations,

Cylins, Tumour growth,Caner.

(6)

quiesene pour des tissus sains et tumoraux

Résumé : Nous présentons un modèle non linéaire de la dynamique d'une

populationellulairediviséeenunompartimentproliférantetunompartiment

quiesent. Laphasede proliférationreprésentel'ensembledu yle de division

ellulaire(G1SG2M) d'unepopulationdeellulesappelées àsediviser

ennde yle. Lemodèleest struturéparletempspasséparuneelluledans

laphasede proliférationetpar laquantité de omplexesylineD/(CDK4 ou

6). Lesellulespeuventpasserd'unompartimentàl'autre,ensuivantdes lois

detransitionquidièrentsuivantqueletissuqu'ellesonstituentestdenature

saine ou tumorale. Le omportement asymptotique des solutions du modèle

nonlinéaireest analysédanslesdeuxas, etmontreunehoméostasietissulaire

danslepremier,etuneroissaneexponentielle(tumorale)dansleseond. Des

simulationsnumériques du modèle onrment ses préditions théoriques.

Mots-lés : Cyle ellulaire, Dynamique des populations, Équations aux

dérivées partielles,Cylines, Croissane tumorale, Caner.

(7)

Contents

1 Introdution 5

2 Moleular mehanisms involved in the G1 phase 6

3 Physiologially strutured model 7

3.1 Unequal division . . . . . . . . . . . . . . . . . . . . . . . . . . 9

3.2 Transitionontrolbetween proliferationand quiesene . . . . . 10

4 Analysis and qualitative behaviour 12

4.1 Linear problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

4.2 Healthy tissue: Non-extintion (a prioriboundfrom below) . . . 14

4.3 Healthy tissue: Limited growth (apriori bound from above) . . 15

4.4 Tumoraltissue: Unlimitedgrowth . . . . . . . . . . . . . . . . . 16

4.5 Steady state forhealthy tissue . . . . . . . . . . . . . . . . . . . 17

5 Numerial simulations 19

6 Disussion and onlusion 21

(8)

1 Introdution

Living tissues, subjet to renewal, are onstituted of two dierent ategories

of ells: the proliferating ells (p) and the quiesent ells (q). Proliferating ells grow and divide, givingbirth at the end of the ellyle tonew ells, or

else transit tothe quiesent ompartment(often referred toas the G0 phase),

whereas quiesent ells donot grownor divide but eithertransit tothe prolif-

erative ompartment orelse stay in G0 and eventually dierentiate aording tothe tissuetype.

Inatumourellpopulationthe numberofproliferatingellsinreases on-

tinuously aslong as itis malignantand ative,whereas in a normal(healthy)

ell population, the size of the proliferative ompartment remains bounded

sine the total number of ells, proliferating and quiesent, remains onstant

(atleast inthe mean,e.g. by averagingover24hours)soastomaintaintissue

homeostasis.

Duringtherstphase(oftenreferredtoasG1)oftheproliferatingellyle,

until the restrition point (R) in late G1 has been reahed, proliferating ells may enter the quiesent G0 phase and stopproliferation. Indeed, experiments

byZetterbergandLarsson[10,34℄showedthattherestritionpoint(R)divides

the G1 phase into two parts: before R, ells may enter the quiesent phase,

but one it has been passed, they are ommitted to proeed along the other

phases(S,G

2,M, whih willnot be onsideredhere as suh) untilelldivision.

The swithing of ells between quiesene and proliferation depends on

extraellularenvironmentalonditions suhasgrowth and antigrowth fators,

and is regulated dierently in normal and tumour ells, due to dierenes in

the expression of the involved genes.

Cellpopulationmodelswithproliferativeandquiesentompartmentshave

beeninvestigatedbyauthorswhostudiedtheasynhronousexponentialgrowth

property [3, 14, 15, 28℄. Our goal here is to design a generi ell population

modelappliable toboth anerand normaltissue growth.

Unlimitedtumourgrowth, byoppositiontohealthytissuehomeostasis,an

beseen inpartiularasaderegulationof transitionsbetweenproliferativeand

quiesentompartments. Furthermore,reentmeasurements[16℄indiatethat

ylins are the most determinant ontrolmoleules forphase transitions.

For these reasons, and sine we are interested in studying in parallel the

behaviourofhealthy and tumourells, westruture ourellpopulationmodel

inage andylin ontent, aproess whihwedesribestep by stepinsetions

2and3. Insetion4,weanalysethetheoretialpropertiesofthe model,whih

we illustrate by numerial simulations in setion 5. Finally, some omments

and future prospets are briey developed insetion 6.

(9)

2 Moleular mehanisms involved in the G1 phase

A variety of proteins are produed during the proliferative ell yle. The

progressionof aellthrough the yle is ontrolled by omplexesomposed of

two proteins: a ylin (strutural protein) and a ylin dependent kinase (or

CDK),anenzymewhihisneededforthe ylintoativate. Eahphaseofthe

ellyle has spei ylin/CDK omplexes. In partiular, ylin D/(CDK4

or 6) and ylin E/CDK2 ativate duringthe G1 phase. Cylin D isthe rst

ylin whihis synthesized at the beginning of the ellyle. The level of y-

linD isontrolledbythe extraellularenvironment. Thus, ylin D synthesis

isindued by spei growth fators (GFs)[5℄, and its level deays when ells

are deprived of GFs. GFs bind to spei reeptors onthe external ytoplas-

mimembrane,stimulatinganintraellularsignallingpathway(Ras/Raf/Map

kinase) by means of whih ylin D is eventually synthesized (see [2, 4, 30℄,

for more details). Experiments reported in [17, 35℄ show the important role

of ylin D as a regulator of the transition between G1 and G0. They show

that a redued exit from G1 to G0 ours when ylin D is overexpressed, whereas non overexpressing ells remain in G0. Progression through the re-

strition point (R) is essentially related to ylin D level inasmuh as when thereisasuient amountofylin D,ells pass therestrition pointand are

ommitted to proeed through the rest of the ell yle. Moreover, ylin D

makesomplexeswitheitherCDK4 orCDK6 kinasesandthese omplexesare

able to phosphorylate other proteins whih are importantfor ellprogression

in the G1 phase through the restrition point and further for the rest of the

elldivision yle: DNA repliation, mitosisand elldivision.

In this paper, we are interested in the moleular interations that are re-

latedto the ativityof the ylin D/(CDK4or 6) omplexes. Several authors

[24,25,26, 32℄havedesribed andsimulated,underspei assumptions,part

of these omplex reations. Here, we give a simple modelto desribe the a-

tivity of ylin D/(CDK4 or 6) indued by growth fators.

Let x be the amount of the omplexes ylin D/(CDK4 or 6) and w an

aggregatedvariablerepresenting the amountofthe various moleulesinvolved

inthesynthesisofylinDsuhasRas/Raf. Weonsiderxandwasregulating

variablesinasimplenonlinearsystemofordinarydierentialequations(ODEs)

with respet to age a in the G1 phase. Synthesis of x ours at a rate c1 and

its degradation at a rate c2; we assume that the synthesis of w is indued by

(10)

growthfators (GFs)ataonstantratec3,itsdegradationourring atarate

c4. The ODE modelan thus be writtenas follows:

dx

da =c1 x

1 +xwc2x, x(0) =x0 >0, dw

da =c3c4w, w(0) =w0 >0.

(1)

Theonlynonlinearityofthe modelisloatedinthe termc1

x

1 +x representing a positively autoregulating oeient with saturation for x under the linear

inuene of the lumped variable w. Substituting the solution of the seond equationof (1), we an redue(1) toone equation inx:

dx da =c1

x 1 +x

c3

c4

+ec4a(w0c3

c4

)

c2x, x(0) =x0. (2)

This holds only for the G1 phase sine we assume that ylin amount x and

agearemainonstantinG0 phase. Anaturalquantityarisesinthequalitative analysis of (2), the x-nullline:

X(a) = c1

c2

c3

c4

+ec4a(w0 c3

c4

)

1.

We assume that w0 c3

c4

and c1c3 > c2c4 whih is a way to express that the

lumped variable w is inreasing from its initial to its asymptoti value, and

that in the early G1 phase the overall synthesis of the hemials involved in

the progression of the G1 phase overomes their degradation. Therefore, a fundamental property of equation (2) is that the ylin onentration x is

limitedby:

xmax= c1c3

c2c4

1>0. (3)

We keep this simple model for our next purpose whih is to desribe a

populationof ells, in proliferativeor quiesent state.

3 Physiologially strutured model

We onsider here only twophases: a quiesent one (physiologiallyG0)and a

proliferativeone (physiologially G1SG2M). The ellpopulations we study are rstly strutured by the time spent inside the proliferative phase.

Thisphaserepresentsherethe ompleteelldivisionyle sineellbirth,and

thistimeinthephasewillhereafterbereferredtoasa,forphysiologialage in the yle. As proposed in[6, 33℄, we alsostruture the model by the amount

of ylin D/(CDK4 or 6) omplexes, denoted by variable x. Indeed, as men-

tioned earlier, this biologial quantity is the most important determinant of

(11)

progression up tothe restrition pointR inthe late G1 phase.

Letp(t, a, x)andq(t, a, x)berespetivelythe densitiesofproliferatingand quiesent ells with age a and ontent x inylin D/(CDK4 or 6) omplexes

attime t.

We also onsider a total weighted population, i.e., an eetive population

density, N dened by:

N(t) =

+

Z

0 +

Z

0

ϕ(a, x)p(t, a, x) +ψ(a, x)q(t, a, x)

dadx. (4)

Heretheweightsϕandψrepresentenvironmentalfatorssuhasgrowthand anti-growth fators ating on the populations of proliferating and quiesent

ells, respetively. N is the density of the fration of the total population

onsisting inthe ells that are sensitivetothese fators and are thusqualied

toinuene,e.g. by amehanism apparented todensity inhibition,the G0/G1

transition. This exludes forinstane apoptotior pre-apoptotiells.

Exitsfromthe quiesent ompartmentare dueeithertoapoptosis (physio-

logialelldeath)ataratedortotransitiontotheproliferativephaseaord- ingtoa reruitment orgetting in funtion G, whih isassumed todepend

onthe total weighted populationN. We alsoassume that ells may leave the

proliferative ompartment for the quiesent one aording to a demobilisa-

tion or leak funtion L(a, x). These funtions L and G, whih represent

the ore mehanism of exhange between proliferation and quiesene in our

model, willbe desribed insetion 3.2. The modelmay thus be writtenas:

∂tp(t, a, x) +

∂a0p(t, a, x)) +

∂x1(a, x)p(t, a, x)) =

(L(a, x) +F(a, x) +d1)p(t, a, x) +G(N(t))q(t, a, x),

∂tq(t, a, x) =L(a, x)p(t, a, x)(G(N(t)) +d2)q(t, a, x).

(5)

Quiesentellsareassumedtobehaltedintheirindividualphysiologialevolu-

tion,inthesensethatoneaellbeomesquiesent,itsageandylin ontent

are xed at their last values as belongingto a proliferative ell. In this way,

quiesent ells donot age and donot hange their ylin ontent.

The parameter Γ0 denotes the evolution speed of physiologial age a with

respettotimet,whihisassumedtobeonstantinthismodel;ifforexample Γ0 = 0.5,itmeansthat physiologialagea evolves twieasslowlyasreal time t. Similarly,thefuntionΓ1 representstheevolutionspeedofylin D/(CDK4

(12)

or 6) with respet to time, i.e., Γ0 times the speed

dx

da of x with respet to

physiologialage a,whih isgiven by equation(2), with w1 =w0 c3

c4 <0: dx

da = Γ1(a, x) Γ0

=c1

x 1 +x

c3 c4

+ec4aw1

c2x.

The parameters d1, d2 are apoptosis rates for proliferatingand quiesent ells respetively, and F(a, x) is the fration of ells whih leave the proliferative populationtodivide aording toa proess whih willbedesribed later.

To ompletethe desription ofthe model(5), wespeify initialonditions:

p(0, a, x) =pi(a, x), a 0, x0, (6)

and

q(0, a, x) =qi(a, x), a0, x0, (7)

where pi and qi are nonnegative funtions.

Inthe followingsetion,wedesribeaonditionforenteringthe proliferat-

ingphase(physiologiallyinG1)atagea= 0,butnotethat nosuhondition

is needed at x= 0, sine ylin level x= 0 is never reahed in the proess

desribed by (2) beause Γ1 vanishes at x= 0.

3.1 Unequal division

The distributionof the ellularmaterialbetween daughter ells isassumed to

beunequal. Duetovariabilityinylinontentbetweenthetwodaughterells

when division ours (see [18℄, and [31℄ for a relationwith ageing), some ells

mayinheritabiggeramountofertainproteinsasylins,whereasothersstart

the yle with a smaller amount of the same proteins. We onsider that the

distributionoftheamountofylin D/(CDK4or 6)betweenthetwodaughter

ellsis given by aonditionaldensity f(a, x, y)suh thatthe probabilityfora daughter ell, born from a mother ell with ontent y in ylin D/(CDK4 or

6) with x1 yx2, tohave itselfontent x inylin D is:

x2

Z

x1

f(a, y, x)dy

+

Z

0

f(a, y, x)dy .

(13)

We also onsider that all newborn ells are at birth in the proliferative om-

partment. Then we have the following ondition atthe boundarya = 0, p(t,0, x) = 2

Γ0 +

Z

0 +

Z

0

f(a, x, y)p(t, a, y)dady. (8)

The following onditions followfromthe above interpretation:

(1)The ylinamountofadaughter ellissmallerthan thatofitsmotherell

atthe timeof division:

f(a, x, y) = 0 if x > y.

(2)Theylin amounty ofthe motherellisexatlyonserved andshared by

the two daughters

f(a, x, y) = f(a, yx, y)

and

Z + 0

f(a, x, y)dx=F (a, y),

whereF(a, y) isthe fration of ellswhihat age a andylin ontent y leave

the proliferatingphaseto undergo elldivision. Theseells disappear and are

replaed by two daughter ells whih immediately restart in the proliferative

phase fortheir own part.

Wehoose for F astandard Hill funtion:

F (a, y) = k1yγ1

k2γ1 +yγ11l[A,+[(a),

where k1 is the maximum eet of ylin D on elldivision , k2 is the ylin

ontent yielding its half-maximum eet, γ1 is the Hill oeient tuning the

steepnessofthe swithat y =k2 between 0andk1 fortheeet, andA isthe

minimalellyle duration;we alsoonsider thatylin repartitionisuniform

afterdivision:

f(a, x, y) = F (a, y)

y 1l[0,y](x).

3.2 Transition ontrol between proliferation and quies-

ene

Lynh [19℄ has studied the eet of a transription fator that inhibits the

proliferationof humanolon anerells by reduingylin D geneexpression

and hene induing an aumulation of ells in G0. Deprivation of growth fators (GFs) inthe early G1 phase alsoleads to a low ylin D level in ells,

when ylin D/CDK4 is the only ylin/CDK omplex present, and the low

level of ylin D is suh that ells exitG1 toenter the G0 phase.

Références

Documents relatifs

Comparison of the 530 genes deregulated in wing imaginal discs expressing CycG ΔP with these 889 genes showed that only 62 genes were both deregulated (39 up- and 23 down-regulated)

Our approach is based on duality arguments (see Proposition 2.7 which is adapted from a result due to Krein-Rutman [14]) and allows us to obtain the existence of eigenvectors in

Le nombre de cornéules n'est vraisemblablement pas un caractère sta- ble chez les Mesogastrura mais dans l'état actuel de la systématique, nous sommes obligés de séparer

The smooth entropy formalism [Ren05,RW04] has been introduced in classical and quantum information theory to study general one-shot scenarios, in which nothing needs to be assumed

The constraint mentioned here is a simple constraint consisting of limiting the drug delivery output at the cell level (p53 or CDK inhibitor concentration) in a given healthy

The TPA-induced gene set is enriched in gene hallmarks (i.e., gene cohorts that change similarly in different contexts; Subramanian et al., 2005 ) associated with the cell

β− spectra from short lived fission products are possible ; in the same irradiation channel, measurements of integral ß-spectrum with the Mini-INCA detectors

Dans le chapitre 3, nous avons introduit un nouvel hydrogel pour études de cellules in vitro, avec incorporation de manière covalente de complexes cyclométallés d’Ir (III) comme