Transcript Regulatory T cells

Modeling the role of IL2 on CD4+T cell dynamics:
On the effect of IL-2 modulating therapies (2)
Dr. Kalet León Monzón
Summer school Leeds (2009)
Crosstalk model of dominant tolerance to self
The key role of IL-2
R
R
R
Regulatory T cells inhibit the activation and
proliferation of helper T cells by an interaction
which requires their colocalized activation at the
APCs.
T
Helper T cells produce a growth factor that
Regulatory T cells use to proliferate and survive
T
T
Leon et al J. Theor. Biol. 2000
Leon et al J. Immunol. 2001
Leon et al 2002 pHD thesis
IL-2 is a good candidate growth factor
In vitro
Regulatory cells do not express IL-2
Regulatory cells over-express CD25
Regulatory cells proliferate when stimulated in the presence of IL-2
Regulatory cells inhibit IL-2 mRNA expression in their targets
In co-cultures, regulatory T cells proliferate/survive by using the IL-2 produced by their
targets. Moreover they consume this IL-2 starving the normal T cell
In vivo
Animals deficient in IL-2, CD25, IL-2Rb chains exhibit lympho-proliferative disorders and
autoimmunity
These autoimmune disorders, in receptor deficient mice, can be prevented by transfer of
wild-type Tregs. But in Il-2 deficient mice, they can be prevented by the addition of Il-2 or
transfer of IL-2 producing cells.
In-vivo depletion of IL-2 induces autoinmunity in 10-20 day thymectomized BalB/c mice.
Our Master Modelling Plan
To extend the Cross-Regulation model to include IL2 dynamics
-Check in which conditions the extended models conserve the key dynamical
features of the original cross-regulation model.
-Compare models of different hypotheses about the specific role of IL2 in T
cell mediated suppression
To study the effect of IL2 modulating therapies currently used
in clinical or preclinical studies
-Study the impact of IL2 depletion treatments.
-Study the impact of IL2 administration treatments.
Adding IL-2 in the cross-regulation model
Basic interactions for T cells
-E cells produce IL-2 upon activation at the APCs.
-E cells need both APCs and gc cytokine signal to
proliferate and survive in vivo.
-E cells use in vivo either IL2 or another homeostatic
cytokine (denoted IL-a) to proliferate and/or survive.
IL-2
(+)
IL-a
(+)
(-)
(-)
R
E
APC
-R cells do not produce IL-2
-R cells need both APC and IL-2 signals to proliferate and survive in vivo.
-R cells inhibit E cell activation and production of IL2 by a mechanism requiring their colocalized activation at the APCs
-R cells constitutively over express the Alpha chain of IL2R (CD25), while E cells up regulate
this molecule transiently after activation.
-The main sources of IL2 in vivo are the helper CD4+ FoxP3- T cells (our E cells)
-IL-2 concentration is reduced in vivo by both by renal elimination and cell internalization
upon binding to the IL2 receptor
Postulates: regarding T cells Live Cycle
R cells
Cognate
APCs
Death
RA
Production
of IL-2
Death
RN
Thymus
EN
EA
{IL-2, IL-a}
{IL-2}
2x
2x
RC
EC
R cells live cycle
E cells live cycle
Three functional states for E and R cells during their life cycle:
Naïve: T cells that have not interacted recently with APCs
Activated: T cells which have been activated by interaction with APCs.
Cycling: T cells irreversibly committed to cell division. They are
activated T cells that have obtained enough cytokine-related signals
Postulates: regarding IL2 “life Cycle”
IL2 interactions
EN
Cognate
APCs
Production
binding to IL2r
T
IL 2B
IL 2F
R cells
unbinding of IL2r
EA
Internalization
+
degradation
Degradation/ elimination
Processes regarding IL2:
IL2 is produced by E cells upon activation on the APCs
IL2 is degraded by renal elimination while not bound to receptor.
IL2 is internalized and degraded by T cells, following its binding to receptors.
Postulates: regarding interactions with APCs
EN
EA
RN
RA
(-)
R
EN
EN
RA
T cell get signals from APCs upon direct conjugation. Conjugation to APCs is assumed
to be a fast process and taken to quasi-steady state equilibrium.
Cells in different states (naïve, activated, cycling) have different conjugation constants.
Regulatory cells suppress E cell activation or IL2 production by a mechanism dependent
on co-localized activation on APCs.
Postulates: regarding interactions with IL2
EN
EA
+
{IL2 or ILa}
EC
RA
+
{IL2}
RC
EN
RN
RN
The mean number of Il2 receptors bound per
cells is computed.
The fraction of cells differentiating into
cycling cells is computed as a sigmoid
function of this number
Fraction of cells
differentiating
R and E cells in different functional states express different number of IL2 receptors on
their surface. Binding of IL2 to these receptors is assumed to be a very fast process
taken to quasi-steady state equilibrium
1.0
0.8
0.6
SR
0.4
SE
0.2
0
0
500
1000 1500 2000
Bound receptors/ cell
Model Equations
d EA
= K AE  ENB
dt
d EC
= KSE
dt

(SE )h
E 

+


+2 K EC α KS 
E h
h
h
(SE ) +(Il2B ) +(Ilα)
E
P
A

 EA -KdE ENF


η(s -1 )
B 

R
T

 1 
s
A


- KSE  EA
EA h
h

(Il
2
)
+
(
Ilα
)
B
 
E
h
h
h
(SE ) + (Il 2BA ) + (Ilα)

  EA - KPE  EC

h


(S
)
R
  - KdR  RNF
 
R
h
h  RA
 (SR ) + (IL 2B ) 
Regulatory Cells
d RN
= GR - K AR  RNB + 2 K pR  RC + α  KSR
dt
d RC
R
= KS
dt
IL2
Effectors Cells
η(s-1)
 RTB 
d EN
E
B 

G


= E KA EN 1 
dt
 s A
d IL 2
= Gi + K ip  ENB
dt
A
d RA
= KAR  RNB - KSR  RA
dt
R
h

(Il 2B )
 
R
h
h
(SR ) + (Il 2B )
A
A

  RA - KPR  RC

(s -1 )
B


R
T

 1  s A 
- Kdi  IL 2 - KinE (IL 2B A  EA + IL 2BC EC ) E
E
- KinR (IL 2B N  RN + IL 2B A RA + IL 2BC RC )
R
R
R
Plus algebraic relations to compute the number of cells bound to APCs and of Il2 bound to its receptor
Handling parameters
The models have 30 parameters, thus although we are
interested in generic properties not dependent on specific
parameter values, we try to find a good reference set of
parameters based on existing experimental data.
The idea is to use this set of values as a reasonable reference to
study model properties, but then study robustness of model
behavior to parameter changes always around these reference
values.
Thus we do a partial calibration parameterization of the model
Most parameters values are bounded from existent data
Lymph node volume
VNL=10 -6 L
Personal communication
# of dendritic cells.
APCS= 104 – 105
Personal communication
T cell division rates
Cycling time (E=R)
KPE =KPR = ln2/tc1/2 ; tc1/2=5 h
Stoll, Science, 2002. 296: 1873
(Confocal microsc)
R
KAE ;KA ;K SE ;KSR are taken such that
the mean doubling time is
between 6-20 h
T cell death rate
Kd= ln2/t1/2 ; t1/2=7-30 días
Surh et al, 2006
# of High affinity IL2 receptors
700 – 1000
R up to 10 time more
Smith KA, Cell Res, 2006. 16: 11
Conjugation IL2 to IL2Rs
alpha: Kon= 107 M-1s-1 ;
Koff= 10-1 s-1
Beta-Gamma: Kon= 105 M-1s-1 ;
Koff= 10-4 s-1
Wang H, J Exp Med, 1987,
166:1055
IL2/IL2R internalization rate
KinE =KinR = ln2/t1/2 ; t1/2=15 min
Fallon, J Biol Chem, 2000.
275: 6790
Free IL2 degradation Rate
Kdi = ln2/t
Donohue et al, 1984
1/2
; t1/2=20 min
Most parameters values are bounded from existent data
Equal for R and E cells
Constant for IL2
association and
dissociation in the
cell membrane to
diferent receptor
chains
Castellino F. Nature,
2006. 440: 890
García Z. PNAS 2007.
104(11): p. 4553-8
IL2/a + bg:
Kon=10-4 s-1 ; Koff= 10-4 s-1
IL2/bg + a
Wang H, J Exp Med,
1987, 166:1055
Kon> 10 s-1 ; Koff= 10-1 s-1
IL2 production rate
Kpi> 10 3 molécs / h
Yamazaki S, J Exp Med,
2003. 198: 235
Sensibility to IL2
derived signal
SE=SR=500; h=4
Smith KA, 2004
Med Immunol, 3: 3
Bound IL2 (molec./cell) \
Free IL2 (pM)
Conjugation constant Kon= 10-13 L.s-1.cél-1
T to APC
Koff=10 -3 s-1
Bound IL2 (molec./cell)x10-3
Does the extended model does conserve the
key dynamical features of the original
cross-regulation model ?
Bistability
One stable state interpreted as tolerance and the other as immunity
Adaptive response to the quantity of APCs
Change from tolerance to immunity with abrupt increase of APC numbers
Adaptation the tolerant state upon slow increase of APC numbers.
Bistability is conserved: Two steady states
Autoimmunity and Tolerance
Autoimmunity IS
Self-Tolerance TS
4
4
10
10
8
Number of cells
10
3
10
IS
OS
2
TS
10
1
10
OS, IS
ET
6
3
10
RT
IL2
10
1
RT
4
2
10
10
ET
10
0
10
IL2
TS, IS
-1
2
200
300
500
ILa adimen. Conc.
10
IL2 conc. (pM)
IL2 degradation rate Kdi (h-1)
Parameter Regions
10
0
2
4
6
Time (weeks)
8
0
2
4
6
8
Time (weeks)
The actual values of T cell numbers and IL2 concentration in the Tolerant and AutoImmune
States are quite reasonable despite not direct fitting of the parameter values.
Other two types of steady state are observed: The trivial immune state (OS) where both E
and R cell collapse to very small values; The reinforced tolerance state (TR) containing only
expanded R cells, which could be attained when external source of IL2 is added.
Adaptive response to APC variation is also
conserved
Gradual APC change
Abrupt APC change
10
10
ET
10
2
3
A
1
10
6
10
RT
4
10
10
10
A
2
-1
2
10
1
1
2
4
3
5
time
6
Time (weeks)
8
8
RT
1
10
1
3
5
time
7
1
ET
6
10
10
4
10
-1
10
IL2F
2
10
7
10
0
2
4
6
8
10
Time (weeks)
Bifurcation diagrams, with the number of APCs, show also a similar structure to those
obtained with the original cross-regulation model.
Free IL2 conc. (pM)
Number of cells
10
10
Free IL2 conc. (pM)
IL2F
8
10
0
10
3
Number of cells
10
Three necessary parameter conditions for a
reasonable model behavior
R cells use IL2 more effectively at low concentrations
SE / RI
EA
> SR / RI
RA
E cells have net growth rate faster than R cells in IL2 excess
K E  K AE / KdE > K R  K AR / KdR
E cells have to use a cytokine alternative to IL2 to proliferate
Ilα > 0
The effect of knocking down ILa in the model
IL2 degradation rate Kdi (h-1)
10
10
4
3
IS
OS
10
10
If Ila is forced to go to zero, the
system reaches a parameter regime
where only autoimmune or a trivial
steady state could be attained
2
TS
This result would predict a dramatic
effect on T cells populations of
knocking down this hypothetical
cytokine
1
OS, IS
200
TS, IS
300
500
ILa adimen. Conc.
Interestingly the only gamma chain family cytokine whose
knockout has a profound effect on the size of T cell population is
Il7. Thus the model would predict it to be the best candidate for
our ILa cytokine.
Including IL2 breaks locality and
specificity of interactions in crossregulation model
In the cross-regulation model, all the
interactions are local near the APCs
In the extended models, some interactions
are non-local (sharing of IL2)
IL-2
R
R
(+)
IL-a
(+)
(-)
R
T
T
T
(-)
E
APC
R
Implications for independence of
concomitant responses to different selfantigen in a given lymph node
A1
<- E1,R1
<- E2,R2
A2
In the extended models
there should be strong
interference between the
immune reactions of
different clones
Comparing different hypotheses in the
suppressive effect of R cells over E cells
Extended model variant 1:
R cells drive competition for IL2
Extended model variant 2:
R cells also inhibit E cell activation
IL-2
IL-2
(+)
(+)
IL-a
(+)
IL-a
(-)
(-)
R
E
APC
Full coupling of
responses to different
antigens
(+)
(-)
(-)
R
E
APC
Bi-standard interactions
between clones, but still
independent responses
But, another possible solution to be explored
in the future is further locality on the use of
IL2 inside a given lymph node
A1
<- E1,R1
<- E2,R2
A2
But since the suppression
mechanism relies on
competition of E and R
cells for IL2, the lymph
node will have to be
subdivided in sub-domains
presenting different antigens
The extended model can be used to study
the effect of IL2 modulation therapies. i.e
Therapies that deplete IL2 and Therapies
that inject IL2
Let´s study whether these treatments can
trigger a switch of steady states in the
models: Tolerance to immunity and
viceversa.
IL2 depletion therapy can subvert the preexistent
Autoimmune or Tolerant steady states
We model IL2 depletion
treatment by an incerase in
IL2 degradation rate
IL2 degradation rate Kdi (h-1)
10
10
4
3
IS
OS
10
10
2
TS
1
OS, IS
TS, IS
200
300
500
ILa adimen. Conc.
Treatments that deplete IL2 are predicted to both break preexistent tolerance by taking
the system to autoimmune state or to reset a previously autoimmune system back into
the tolerance state.
IL2 depletion can subvert the preexistent
Autoimmune or Tolerant steady states
4
106
+
8
ET
3
10
IL2 conc. (pM)
Breaking
Tolerance
Number of cells
10
2
6
10
10
IL
2
1
10
4
10
0
RT
10
Depletion Dose
10
0
3
10
TS
2
10
5
10
15
10
20
0
4
+
8
1
2
3
4
4
10
10
-
10
RT
2
6
10
10
1
10
ET
4
10
0
10
IL
2
2
10
0
10
-1
Depletion Dose
3
10
IL2 conc. (pM)
Number of cells
4
10
-1
2
10
Reestablishing
Tolerance
IS
5
10
10
3
TS
IS
10
2
10
20
30
Time (weeks)
40
1
2
3
4
5
6
7
Duration of treatment (w)
This supports applications of IL2 depletion both in the treatment of autoimmune diseases and cancer,
but these applications strongly depend on treatment schedules
IL2 Blocking MAbs have been shown to induce
autoimmunity, but in adult thymectomized mice.
5
10
4
DosS
7.10
ΓT= 10-4
IS
4
5.10
ΓT= 10-6
ΓT= 10-8
4
3.10
TS
4
2.10
0
1
2
3
4
Duration of treatment
(weeks)
The model explains the thymus dependence of this effect by showing a decrease in the minimal dose of
depletion treatment, which is required to change state, as the thymus output is decreased.
Setoguchi et al. JEM 2005, 201(5):723–735
Anti-IL2 MAbs can induce tumor regression
2d
5d
1 mg AcM aIL-2 iv
F10
F10 irradiated
5x105 IU/Kg sc IL-2
Montero et al. 2005 unpublished results
IL2 injection therapy can not subvert the preexistent
Autoimmune or Tolerant steady states
10
8
6
RT, IS
10
10
4
Γi
We model IL2 injection
treatment by an increase in the
external source of IL2 (Gi)
10
2
TS, IS
OS, IS
0.1
1
10
100
Ilα
1000
10000
Therapy with IL2 can only reinforce the preexistent
Steady State
Reinforcement of Tolerance
9
10
10
Reinforcement of Autoimmunity
5
9
10
10
ET
10
2
5
10
10
ET
3
10
10
Number of cells
3
10
7
10
3
10
IL2
5
10
1
IL2
1
-1
10
10
2
4
6
10
RT
3
10
0
2
10
8
10
Time (weeks)
12
14
IL2 conc. (pM)
RT
10
4
10
IL2 conc. (pM)
Number of cells
4
7
5
-1
10
10
0
2
4
6
8
10
12
14
Time (weeks)
This might explain the limited success of IL2 therapy in Cancer. It would be predicted to result
just on those patients with a preexistent but insufficient immune response.
It will predict that, despite the role of IL2 as homeostatic factor for Tregs, it might not be used as
therapy for autoimmune diseases.
IL2 administration can reinforce tolerance promoting
allograft acceptance
2 daily doses of IL-2,
sc tail base
Check growth of allogenic tumor
2*105 F10 sc fd
Incidence of MB16F10 tumors
Incidencia tumoral (%)
100
PBS
IL-2 5*103
IL-2 5*104
IL-2 5*105
IL-2 5*106
80
60
40
U/kg
U/kg
U/kg
U/kg
20
0
0
5
10
15
20
25
30
35
40
45
Días
Montero et al Ann. N.Y. Acad. Sci. 1107: 239–250 (2007).
IL2 administration can reinforce tolerance
preventing EAE induction
Interestingly this tolerizing effect of IL2 or Il2+MAbs is only observed when it is administered
before the onset the disease.
Montero et al Ann. N.Y. Acad. Sci. 1107: 239–250 (2007).
Could the particular interface of the IL2 being
recognized by IL2 blocking MAbs make a difference?
Chain a
Dimerization of the b and g chain is
responsible for signaling
IL-2
MAbs, could block the IL2 binding to the
a chain of the receptor without affecting
the binding to the b and g chains or vice
versa
Chain b
Chain g
MAbs blocking binding to b and/or g
chain will fully inhibit signaling, but
those just blocking binding to a chain
will allow it
Thus we study these classes of MAbs being used alone or in preformed complexes with IL2
Modeling therapies with IL-2 and/or anti-IL2
MAbs in a virtual mouse
Therapies of interest
Vs=2.5 mL
Vn=1.0 mL
Lymph Node
#T=107
B
L
O
O
D
IL-2
(+)
IL-a
#APC=1 105
(+) (-)
(-)
E
R
APC
As many repeated units of Lymph nodes as necessary
to guaranteed a T cell repertoire with 108 cells, 107 of
which might be autoreactive T cells that participate in
the modeled dynamics
Renal and/or fecal elimination of molecules
Model parameters are calibrated from independent publicly available Data
MAbs blocking either a or b interfaces of IL2
similarly break a preexistent tolerant state
4
5 mg
8
Number of cells
KD=10-10 M
t1/2=3 days
3
ET
10
RT
RT
6
10
2
10
IL2
ET
4
RT
10
10
IL2
IL
2
2
1
ET
10
IL2 conc. (pM)
Face b
1 mg
3 mg
10
10
0
10
-1
10
4
2.5 mg
8
1.7 mg
1.0 mg
3
ET
6
10
10
IL2
4
10
RT
10
ET
ET
10
IL
2
RT
2
2
RT
1
0
IL
2
10
-1
10
10
0
5
10
15
Time (weeks)
20
0
5
10
15
Time (weeks)
20
0
5
10
15
20
Time (weeks)
The difference between these MAbs is just quantitative. Affinity for IL2, half life and even the
particular Kon and Koff values strongly influence the stringency of the effect.
IL2 conc. (pM)
Face a
Number of cells
10
10
MAbs blocking b, but not a, interface of IL2 can
reverse autoimmunity, reestablishing tolerance
Face b MAb
Face a MAb
4
+
8
-
8
2
10
10
1
10
4
ET
IL2
10
0
10
-1
2
10
10
0
10
20
30
Time (weeks)
40
Number of cells
RT
6
10
-
ET
3
10
6
2
10
10
IL2
1
10
RT
4
10
0
10
2
-1
10
0
10
20
30
40
10
Time (weeks)
Despite achieving similar reduction of IL2 concentration, these MAbs behave in a
qualitatively different way.
Interface b MAbs reduce the actual IL2 related signal more, triggering a larger
reduction of E cell number that ultimately allows the reestablishment of tolerance.
IL2 conc. (pM)
Dos=20 mg/d
10
+
10
3
IL2 conc. (pM)
t1/2=3 days
Number of cells
10
KD=10-10 M
4
10
What about injecting Immunecomplexes?
MAbs + IL2
Both types of Immunocomplexes reinforce tolerant
state.
Interface a MAb +IL2
4
4
1
10
ET
4
10
0
10
IL
2
2
10
0
-1
10
5
10
15
Time (weeks)
20
RT
6
10
2
10
1
10
ET
4
10
0
IL
2
2
10
0
10
-1
10
5
10
15
Time (weeks)
20
Conc. (pM)
2
10
10
3
10
20 mg/d 1 week
5
IL2 conc. (pM)
RT
6
10
Number of cells
3
10
10
5 mg/d 1 week
8
10
IL2 conc. (pM)
Number of cells
10
20 mg/d 1 week
8
10
Availability of Free
IL-2 signal
Interface b MAb +Il2
Complexes
3
10
1
IL
2
10
-1
10
0
5
10
15
20
Time (weeks)
Both types of immunocomplexes appear to expand regulatory T cell activity in the model,
although interface b derived complexes appear quantitatively more efficient at low doses
However injection of interface a MAbs +IL2 provokes a very large and sustained increase in
the availability of IL2 signal, that could explain its reported effect over MP CD8+ cells.
Immunocomplex with interface b but not interface a
anti- IL2 MAbs can reestablish Tolerant State.
Face b MAb +IL2
Face a MAb +IL2
4
4
10
10
0.3 mg/d 50week
10
2
10
IL2
1
10
4
RT
10
0
10
-1
2
10
10
0
10
20
30
40
50
Time (weeks)
60
70
3
10
ET
2
10
6
10
IL2
1
10
4
RT
10
0
IL2 conc. (pM)
ET
IL2 conc. (pM)
Number of cells
3
10
6
0.3 mg/d 50week
8
10
Number of cells
8
10
10
-1
10
2
10
0
10
20
30
40
50
60
70
Time (weeks)
Despite both types of immunocomplexes appearing to expand regulatory T cell activity in
the model, only interface b derived complexes appear capable of reversing the
autoimmune state.
Conclusions through the Modeling relationship
Postulates, Axioms
Experimental
Observations
Model
d N1
 F1 ( N1 , N 2 , N 3 , N 4 )
dt
d N2
 F2 ( N1 , N 2 , N 3 , N 4 )
dt
d N3
 F3 ( N1 , N 2 , N 3 , N 4 )
dt
Predictions
Model Predictions and explanations
It explains:
•
The Cross-regulation model is compatible with IL2 being a growth factor of both E and R cells.
•
That in the absence of IL7 the immune system could be either in a steady state with a few E and R cells
sustained by thymic output or in an autoimmune steady state.
•
That IL2 depletion can induce autoimmunity and that the absence of thymus could favor this effect.
•
That IL2 depletion treatments could favor tumor rejection or potentiate cancer vaccines.
•
That IL2 injection treatment could be effective in a subset of cancer patients or can be good adjuvants for
some vaccines.
•
That Il2 injection treatment could reinforce tolerance and promote allograft acceptance if applied before graft
implant.
It predicts:
1. That IL7 is the likely gamma chain family cytokine that helper cells use as an alternative to Il2 to growth in
vivo
2. That Il2 depletion treatments can be used to reestablish tolerance in an autoimmune individual. (interface b
Mabs)
3. That IL2 treatment could only reinforce a preexistent immune state, thus is could be used to adjuvate vaccines,
but applying it preferentially after the vaccine and not before.
4. That Il2 can not be used as a therapy to autoimmune diseases, despite being the main growth factors for Tregs.
(The same goes for immunocomplexes)
5. That the lack of success of Il2 based therapies on cancer is a consequence of the heterogeneity of the
Acknowledgements
CIM
Lic. Tania Carmenate
M.Sc. Karina García
Lic. Ileana Gonzalez
Lic. Jose Antonio
Dr. Enrique Montero
Dr. Agustín Lage
IGC (Portugal)
Dr. Jorge Carneiro
Financial support
IGC and CIM