Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/307945097
CITATIONS READS
0 201
7 authors, including:
Some of the authors of this publication are also working on these related projects:
Time-Delay Model For Tumour-Immune System With Optimal Control View project
Call for Papers, Special Issue of title "Applications of Delay Differential Equations in Biological
Systems " in Complexity View project
All content following this page was uploaded by Fathalla A Rihan on 09 September 2016.
se
rn
arc
h
Research
Research Article
Article Open
OpenAccess
Access
Abstract
Most of biological systems have long-range temporal memory. Modeling of such systems by fractional-order
(or arbitrary-order) models provides the systems with long-time memory and gains them extra degrees of freedom.
Herein, we suggest a simple fractional-order model to describe the dynamics of tumor-immune interactions. Two
effector cells are considered, in the model, with a Holling function response of type-III. The model is extended to
include treatment terms which represent an external source of the effectors cells by ACI and an external input of IL-
2. Asymptotic stabilities of tumour-free steady state and persistent- tumour steady state are studied. The threshold
parameter R0 (average number of newly infected cells produced by a single councerous cell) is deduced. The
numerical simulations show that the fractional-order derivative enriches the dynamics of the system and increases
the complexity of the observed behaviours, which confirms that the fractional-order may play the role of memory in
the system.
Keywords: Cancer; Fractional-order; Numerical simulation; Assume that model of cancer-immune system includes two
Stability; Steady states; Tumor-immune system immune effectors: E1(t) and E2(t) (such as cytotoxic T-cells, and
natural killer cells), interacting with the cancer cells, T(t) with tumours
Introduction functional response of Holling Type-III [25]. The model takes the form
Tumors are a family of high-mortality diseases, exhibiting a a1 T 2 (t)E1 (t)
D= E1 (t) -d1E1 (t) + ,
derangement of cellular proliferation which often lead to uncontrolled T 2 (t) + k1
cell growth [1,2]. Research efforts are being devoted to understand Da2 T(t)
= aT(t) r1T(t)E1 (t) r2T(t)E 2 (t), 0 < a i =
1, i 1, 2,3. (1)
the interaction between the tumour cells and the immune system T 2 (t)E 2 (t)
[3-6]. Mathematical models, using ordinary differential equations Da2 E 2 (t) =
d 2 E 2 (t) + 2 ,
T (t) + k 2
with integer-order, have been proven valuable in understanding the
The parameter d1 represents natural decay rate of the effect cells
dynamics of tumour-immune system and how host immune cells and
E1(t). k1 and k2 are half saturation parameters, a is the growth rate of
cancerous cells evolve and interact; See e.g. [7-12]. However, modeling
tumour cells. r1 and r2 are the reduction rate of tumour cells due to
of biological systems by fractional-order differential equations has
more advantages than classical integer-order mathematical modeling, presence of the effector cells E1(t) and E2(t) respectively. d2 represents
in which such affects the memory are neglected. Accordingly, the natural decay rate of the effect cells E2(t). All of these parameters are
subject of fractional calculus (that is, calculus of integral and derivatives supposed to be positive constants. The interaction terms in the first and
of arbitrary order) has gained popularity and importance, mainly due third equations of model (1) satisfy the cross reactivity property of the
to its demonstrated applications in system biology [13,14] and other immune system. It has been assumed that (d1k1/(1d1))<<(d2k2/(1
fields of sciences [15-18]. The Fractional-Order Differential Equations d2)), to avoid the non-biological interior solution where both immune
(FODEs) models are more consistent with the biological phenomena effectors coexist. For more details about model (1), we refer to [16].
than those of integer-orders. This is due to the fact that fractional- Here, we have modified model (1) to include external sources of
order derivatives the description of the memory and hereditary effector cells and immune stimulation effects by treatment Interleukin-2
properties inherent in the processes [19]. It has been deduced in [13]
(IL-2). Assume that three populations of the activated immune-system
that the membranes of the organism have fractional-order electrical
cells, E(t); the tumor cells, T(t); and the concentration of IL-2 in the
conductance, which are classified under the non-integer order models.
single tumor-site compartment, IL(t) (Figure 1). To ease the analysis,
Fractional Models of Tumor-Immune System consider a classic bilinear model that includes Holling Type-I function
Page 2 of 6
Death + ve Death
External input Effector Cells, E(t) synthesis IL-2, I L (t) External input
+ ve + ve
Stimulate
Tumour Cells, T(t)
Death
Figure 1: Key-players in tumour-immune interactions described in model (2).
and external effector cells s1 and input of IL-2, s2. The interactions Equilibria and local stability of model (1)
of the three populations are then governed by the fractional-order
differential model: The equilibrium points of system (1) are:
Da1 E(t) =
s1 + p1E(t)T(t) p 2 E(t) + p3E(t) I L (t) =0 (0,0,0);
= 1 a / r1 , d1k1 /(1 - d1 ), 0);
a2
D= T(t) p 4 T(t)(1 p5T(t)) p 6 E(t)T(t), 0 <a= 1, 2,3. (2)
(5)
i 1, i
2 =(0, d 2 k 2 /(1 - d 2 ),a / r2 ).
a3
D I L (t) =
s 2 + p 7 E(t)T(t) p8 I L (t),
Here 0 is the naive first equilibrium, 1 is the memory equilibrium
with initial conditions: E(0) = E 0 , T(0) = T0 , I L (0) = I L0 . The first and the 2 is endemic according to the value of the tumor size. Stability
equation describes the rate of change in the effector cells population. analysis shows that the naive state is unstable.
The parameter p1 represents the antigenicity rate of the tumor (immune
response to the appearance of the tumor), p3 is the cooperation rate However, the memory state is locally asymptotically stable if
of effector cells with Interleukin-2 parameter, and s1 represents the d1
external source of the effector cells, with rate of death p2. The second R0 = < 1, and 0<d1 <1
d2
equation shows the rate of change of the tumour cells which follows a
While the endemic state is locally asymptotically stable if R 0 > 1
logistic growth (a type of limiting growth) in the absence of immune
and 0 < d 2 < 1. There is bifurcation at R0=1. The stability of the
response. The parameter p4 incorporates growth rate of tumor cells.
memory state depends on the value of one parameter namely the
The maximal carrying capacity of the biological environment for tumor
1 immune effector death rate.
cell is p5 . Whereas, p6 is the rate of tumor cells. The third equation
gives the rate of change for the concentration of IL-2. Its source is the Equilibria and local stability of model (4)
effector cells, which are stimulated by interaction with the tumour. The
The steady states of the reduced model (4) are again the
parameter p7 is the competition rate between the effector cells and the
tumor cells. The external input of IL-2 into the system is s2 and the intersection of the null-clines Da1x = 0 Da 2 y = 0. If y=0, the tumor-
loss-rate parameter of IL-2 cells is p8.
free equilibrium is at 0 = (x, y) =( ,0) . This steady state always exist,
In the absence of immunotherapy with IL-2, we have since > 0 . From the analysis, it is easy to prove that the tumor-
Da1 E(t) =
s1 + p1E(t)T(t) p 2 E(t),
free equilibrium 0 = ( ,0) of the model (4) is asymptotically stable
a2
D= T(t) p 4 T(t)(1 p5T(t)) p 6 E(t)T(t), 0 < a=
i 1, i 1, 2. (3)
if threshold parameter (the minimum tumor-clearance parameter)
To minimize sensitivity (or robustness) of the model, we non-
dimensionalize the bilinear system (3) by taking the following re- a
R 0 ( , < 1) , and unstable if R 0 > 1 .
=
scaling:
However, if y 0, the steady states are obtained by solving
E(t) T(t) p1T0 p2
x(t) = , y(t)
= =, = , , aby 2 -a(+ b)y + a - = 0.
E0 T0 t sE0 ts
s1 p4 p6E 0 * In this case, we have two endemic equilibria, 1 and 2
= =,a = , b p5T=
0, 1 =, t t s t.
t sE0 ts ts
-a(b-)- a(b + ) +
=1 (x1=
, y1 ), where x1 = , y1
After the above substitution into (3), we have 2 2ab
-a(b-) + a(b + )
Da1x(t) = + x(t)y(t) - x(t), = 2 (x 2=
, y 2 ), where x 2 = , y2
(4) 2 2ab
Da 2 y(t) = ay(t)(1 - by(t)) - x(t)y(t)
with = a 2 (b ) 2 + 4ab > 0 . The Jacobian matrix of system (4)
(Here t is replaced by t*.) The analytical stability region of fractional- at the endemic equilibrium 1 is
order system (4) is given in Figure 2.
We then study the stability of the steady states of models (1) and (4). (
J(1 ) = y1
y 1
x1
a 2ab y1 x1 ) (6)
Page 3 of 6
3
E1(t) 2.5
T(t) E1(t)
2.5 E2(t) T(t)
E2(t)
2
2
E1(t), T(t), E2(t)
1.5
1.5
1 1
1
0.5
0.5
0
0 20 40 60 80 100
0
Time 0 20 40 60 80 100
Time
Figure 3: Shows the numerical simulations of model (1) when =1.0 and a=r1=r2=1; d1=0.3, d2=0.7, k1=0.3, k2=0.7 (left banners) where the system converges to steady
state 1; and when d1=0.7, d2=0.3 (right banner), where the system converges to steady state . We note sustained oscillations.
2
Page 4 of 6
2 1.8
E1(t) E1(t)
1.8 T(t) 1.6 T(t)
E2(t) E2(t)
1.6
1.4
1.4
1.2
E1(t), T(t), E2(t)
0.2 0.2
0 0
0 20 40 60 80 100 0 20 40 60 80 100
Time Time
Figure 4: Shows the numerical simulations of model (1) when =0.95 and a=r1=r2=1; d1=0.3, d2=0.7, k1=0.3, k2=0.7 (left banners) where the system converges to
steady state 1 ; and when d1=0.7, d2=0.3 (right banner), where the system converges to steady state 2. The fractional derivative damps the oscillation behaviour.
4 2.5
E1(t) E1(t)
3.5 T(t) T(t)
E2(t) E2(t)
2
3
2.5
E1(t), T(t), E (t)
1.5
2
1
1.5
1
0.5
0.5
0 0
0 10 20 30 40 50 0 10 20 30 40 50
Time Time
Figure 5: Shows the numerical simulations of model (1) when =0.75 and a=r1=r2=1; d1=0.3, d2=0.7, k1=0.3, k2=0.7 (left banners) where the system converges to
steady state 1 ; and when d1=0.7, d2=0.3 (right banner), where the system converges to steady state 2 . The fractional derivative damps the oscillation behaviour.
the stability regions of the solutions. Although the fractional-order description of memory and hereditary properties of inter and intra
model (4) is simple, it displays up to three steady states. Figure 6 cells. It is possible that the tumour may result in either equilibrium
shows different types of steady states: Tumour-free steady state 0 , with (dormancy) or escape from the immune system. The fractional-
dormancy-, medium and high persistent-tumour steady states (right) order differential equations are, at least, stable as their integer-order
counterparts. The presence of a fractional differential order in the
1,2 , for the model (4). differential equation can lead to a notable increase in the complexity
As a result, in the endemic steady states, Figure 7 shows that the of the observed behaviour, as the solution continuously depends on
fractional-order derivative kills the oscillation behaviour. Figure 8 all the previous states of the solutions. The analysis can be extended to
displays the numerical simulations for the model (4) with different include more components of immune response and control variables.
values of the fractional order and parameter values given in the caption. It can also be extended to describe the dynamics of hepatitis B and C
The tumour-free 0 is locally asymptotically stable as R 0 = a / < 1 virus infections.
.From the graphs, it can be seen that FODEs have rich dynamics and
are better descriptors of biological systems than traditional integer- Appendix A: Preliminaries
order models. The equilibria for infection-free and endemic fractional- The subject of fractional calculus deals with the investigations
order cases are the same as integer-order counterparts. of derivatives and integrals of any arbitrary real or complex order,
which unify and extend the notions of integer-order derivative and
Conclusion n-fold integral. Here, we provide some definitions of fractional-order
In this paper, the authors conclude that the fractional-order integration and fractional-order differentiation [26]. There are several
derivatives in the models provide an excellent instrument for the approaches to the generalization of the notion of differentiation to
Page 5 of 6
1000 1000
+3
900 900
800 800 +
2
700 700
600 600
T(t)
T(t)
500 500
400 400
+1
300 300
200 200
100 100
0
0
0 0
0 100 200 300 400 500 600 0 0.5 1 1.5 2 2.5
E(t) E(t) 4
x 10
Figure 6: Shows (left) tumour-free steady state and low-, medium and high persistent-tumour steady states (right) 1,2 ,for the model (4). The stable steady states
0
are [], unstable steady states are [], stable manifold is [] and initial conditions are [+], with external treatment doses >0 and particular parameters.
4
=1 12
=0.9 =1
3.5 =0.7 =0.9
=0.7
10
3
8
2.5
Effector Cells, E(t)
2
6
1.5
4
2
0.5
0 0
0 50 100 150 0 50 100 150
Time Time
Figure 7: Shows numerical simulations of model (4) when s=0.1181, =0.1184, =0.1747, r=0.636, b=0.002. _1=2=1, 0.9, 0.7. Note that 2 is locally asymptotically
stable as
a
R 0 ( , < 1) . The fractional-order derivative kills the oscillation behaviour.
=
0.75 0.35
=1
=0.9
0.7
0.3 =0.8
0.65
0.6 0.25
Effector Cells, E(t)
0.55
Tumor Cells, T(t)
0.2
0.5
0.15
0.45
0.4 0.1
0.35
0.05
=1
0.3
=0.9
=0.8
0.25 0
0 20 40 60 80 100 120 140 160 180 200 0 20 40 60 80 100 120 140 160 180 200
Time Time
Figure 8: Shows the numerical simulations of model (4) when =0.1211, =0.1184, =0.3746, a=1.635, b=0.002. 1=2=1, 0.9, 0.7. 0 is locally asymptotically stable
a
as R 0 ( , < 1) .
=
Page 6 of 6
D*a f (t) C(a, b] for 0 < a 1, then we have 21. Bellomo N, Bellouquid A, Nieto J, Soler J (2010) Multiscale biological tissue
models and flux limited chemotaxis from binary mixtures of multicellular
1 growing systems. Math. Models Methods Appl Sci 20: 1179-1207.
f(t) = f(a) + D*a f ()(t a)a , with a < < t t (a, b). (10)
T(a) 22. Gkdoan A, Yildirim A, Merdan M (2011) Solving a fractional order model of
HIV infection of CD4+ T cells. math. Comput Modeling 54: 2132-2138.
If (v) holds, and Da* f(t) 0 t [a, b], then f(t) is nondecreasing for
each t [a, b]. If Da* f(t) 0 t [a, b], then f(t) is non-increasing for each 23. Kirschner D, Panetta J (1998) Modeling immunotherapy of the tumor immune
t [a,b]. interaction. J Math Biol 37: 235-252.