Accepted Manuscript

A Mathematical Model for Transmission Dynamics of HIV/AIDS with

effect of weak CD4+ T cells

Ajoy Dutta , Praveen Kumar Gupta

PII: S0577-9073(17)31343-6
DOI: 10.1016/j.cjph.2018.04.004
Reference: CJPH 498

To appear in: Chinese Journal of Physics

Received date: 21 October 2017

Revised date: 19 February 2018
Accepted date: 8 April 2018

Please cite this article as: Ajoy Dutta , Praveen Kumar Gupta , A Mathematical Model for Transmission
Dynamics of HIV/AIDS with effect of weak CD4+ T cells, Chinese Journal of Physics (2018), doi:
10.1016/j.cjph.2018.04.004

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service
to our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and
all legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Highlights
 A model is defined for HIV/AIDS with incorporated of weak CD4+ T-cells.

 The stability analysis is carried out for infection-free and infection equilibrium.

 Local stability is demonstrated by Routh-Hurwitz criterion.

 Global stability is verified by Lyapunov’s and the geometric approach method.

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

A Mathematical Model for Transmission Dynamics of HIV/AIDS with effect of weak

CD4+ T cells

Ajoy Dutta, Praveen Kumar Gupta

*Department of Mathematics, National Institute of Technology Silchar, Assam-788010.

T
Abstract

IP
A mathematical model is proposed for the dynamics of HIV/AIDS with incorporated of weak

CR
CD4+ T cells. The model considers three different categories of cells: uninfected CD4+ T-
cells, infected CD4+ T-cells, and virus. The anticipated model helps to illustrate the many of

US
mystifying features of HIV infection more clearly. This model demonstrates two steady
states: an infection-free equilibrium state, in which there is no virus, and an infection
AN
equilibrium state, in which virus and infected T-cells are present. We have also calculated the
local stability of the infection-free equilibrium and infection equilibrium for the model when
the valuable reproduction number is less than and greater than one. With the help of
M

Lyapunov’s second method and the geometric approach, we are defining the novel conditions
for the global stability of infection-free equilibrium state and infection equilibrium state. This
ED

study, which knocks off-balance the system, is articulated by a small variation of the
parameters conceded by the system from one stable state to an unstable state. The dynamics
PT

of this new steady state are calculated both numerically and via the stability analysis.

Keywords: HIV/AIDS, weak CD4+ T cells, Basic reproduction number, Stability Analysis.
CE

Mathematics Subject Classification (MSC) 2010 code: 34A34, 37B55, 45M10, 49M30.
AC

1. Introduction

In 1981, the first AIDS person was recognized and HIV has since spread worldwide.
According to World Health Organization (WHO), more than 2 million citizens pass away
globally up to 2015, and the epidemic continues to spread. AIDS is not yet fully curable and
dangerous infectious diseases that suppress the immune system of the body. For this reason
the human immune deficiency virus (HIV), which obliterates the body’s capability to fight
infections, because of its broad scope of infectiousness, sturdy mortality and strong
 ACCEPTED MANUSCRIPT

incurability. HIV has become a leading problem for human body. Literatures of the biological
sciences indicate that HIV has infected to CD4+ T helper cells. The primary stage of disease
is categorized by a strong viral response; which is speedily pursued by a powerful immune
riposte. In the second phase of HIV infection, infected cells exhibits no symptoms, although
has continual viral replications, which ultimately effected in the expansion of AIDS [1-3].

In the last two decades, many deterministic and stochastic mathematical models have been
described to point out the immunological behaviour to infectivity by HIV [4-7]. These

T
models have been used to explain diverse phenomena. In 2012, Joly and Pinto [8] presented a

IP
work based on a comprehensive, nevertheless deterministic, mathematical modeling basis,
the computational results to support important discoveries that are derived from clinical

CR
experience as well as introduce new theoretical possibilities for investigation. Mojaver and
Kheiri [9] have proposed a HIV infection models for both classical cell-free virus diffusion

US
and direct cell-to-cell transmission. They have discussed non-negative solutions for the
model; and derive the sufficient conditions for the asymptotic stability of equilibrium.
AN
The progressive damage of CD4+ T cell is the characteristic of HIV infection but the
mechanism is not clear due to the slow T cell decline. Some recent studies suggested that
M

pyroptosis is associated with the release of inflammatory cytokines, which can attract more
CD4+ T cells to be infected. In December 2015, Wang et al. [10] have offered two
ED

mathematical models to investigate whether cell death provoked by pyroptosis can clarify the
slow time scale of CD4+ T cell decline in untreated HIV patients. The results show that both
models produce a slow decline of CD4+ T cells in plasma, and many sympathetic agree with
PT

long-term CD4+ T cell data of untreated HIV patients.

CE

Here, we modify the previous model [11] and develop a mathematical model that involves
the effect of natural recovery of CD4+ T cells and due to extremely small transition period,
weak uninfected CD4+ T cells have been transferred directly into the virus class. The effort of
AC

this paper is as follows. The proposed model description and basic properties of the model are
described in section 2. Section 3 is devoted to study the existence of infection-free
equilibrium state and infection equilibrium (endemic) state, and the local and global stability
for both the states. Numerical simulations of the model are discussed in section 4. Section 5
contains the Conclusion remarks.
 ACCEPTED MANUSCRIPT

2. The model

2.1 System description

To construct the model, we have considered three different categories of cells: uninfected
CD4+ T cells as T (t ) , HIV infected CD4+ T cells as I (t ) , and the virus cells as V (t ) . Here,
we are modifying the model of Srivastava and Chandra [11] with the help of Essunger et al.
[12] and Rong et al. [13]. In the anticipated model, we have incorporated two new
assumptions: first one, natural recovery of CD4+ T cells; and the second one, due to

T
extremely small transition period of some weak uninfected CD4+ T cells, these CD4+ T cells

IP
transferred directly into the virus class. In inspection of this, the following model is proposed:

CR
dT
 r  1VT   2VT  1T   3 I
dt
dI
dt
dV
  2VT   3 I   2 I
US (2.1)
AN
 A 3 I  1VT   4V
dt

with T (0)  T0  0 , I (0)  I 0  0 and V (0)  V0  0 . (2.2)

M

Parameter Value Data Source

ED

r 10 mm 3 day1 [12]

1 0.000020 mm 3 day1 Assume

PT

2 0.000024 mm 3 day1 [12]

CE

3 0.2 day1 [12]

1 0.01day1 [12]
AC

2 0.5 day1 Assume

3 0.16 day1 [12]

4 3.4 day1 [12]

A 1940 Assume

Table1. Values of the parameters

 ACCEPTED MANUSCRIPT

T
IP
CR
Figure 1. Transfer diagram of the model (2.1)

US
Here, r is inflow rate of CD4+ T cells,  1 is the contact rate of CD4+ T cells and virus cells

become virus,  2 (  1 ) is rate of infected cells,  1 is the natural death rate of uninfected
AN
CD4+ T cells,  3 is rate at which infected cells return to uninfected cells,  2 ( 1 ) is the

death rate of infected CD4+ T cells,  3 is the Lytic death rate for infected cells,  4 is the death
M

rate of virus, and A is the average number of viral particles produce by an infected cell.
ED

2.2 Basic properties

PT

2.2.1 Positivity of solutions

Lemma 1. The exact solutions T (t ), I (t ), V (t ) of model (2.1) with initial values T (0)  0 ,
CE

I (0)  0 and V (0)  0 are positive for all t  0 .

AC

Proof. The first equation of the model (2.1) is as follows:

d  t  t 
T (t ) exp  (1   2 )V (  ) d  1 t   (r   3 I (t )) exp  (1   2 )V (  )d  1 t  (2.3)
dt  0  0 

Therefore,

t  t
 x 
T (t ) exp  ( 1   2 ) V (  )d  1 t   T (0)   (r   3 I ( x)) exp  ( 1   2 ) V (  )d   1 x  dx
0  
0  0 
 ACCEPTED MANUSCRIPT

so that,

 t  t x 
T (t )  exp   ( 1   2 ) V (  )d   1 t    (r   3 I ( x)) exp  ( 1   2 ) V (  )d   1 x  dx
 0  0  0 

 t 
 T (0) exp   ( 1   2 ) V (  )d  1 t   0 . (2.4)
 0 

Subsequently, we can also prove that I (t )  0 , V (t )  0 . Thus, the solutions T (t ), I (t ), V (t )

of the model (2.1) stay positive for all t  0 .

T
IP
2.2.2 Invariant regions

CR
Lemma 2. The feasible region  define by:
 r 
  (T (t ), I (t ),V (t ))  R3 : 0  T  I  , V   (2.5)
1

US

for some   0 , which is positively invariant with respect to model (2.1).
AN
Proof. After adding the first two equation of the model (2.1), we get

d
(T  I )  r  1 V T  1 T   2 I  r  1 (T  I ) ; ( 1   2 ) . (2.6)
M

dt

Here, r , 1 are non-negative, thus the sum of uninfected and infected T-cell populations are
ED

always bound. From Huo et al. [14], the solution of third equation of the model (2.1) also
restricted, i.e., V (t ) is bounded.
PT

So, we can define a bounded set

CE

 r 
  (T (t ), I (t ),V (t ))  R3 : 0  T  I  , V   ,
 1 
AC

for some   0 , which is positively invariant with respect to model (2.1).

3. Analysis of the model

3.1 The non-infected steady state and the basic reproduction number

It is straightforward to calculate and states the model (2.1) always exist a non-infected
equilibrium point
 ACCEPTED MANUSCRIPT

 r 
E 0  (T0 *, I 0 *, V0 *)   , 0, 0  . (3.1)
 1 

Let X  (I , V, T) , then system (2.1) can be written as

dX
 ( X )  ( X ) (3.2)
dt

where,  ( X ) is the rate of appearance of new infections, (X ) is the rate of transfer of

T
individuals,

IP
  2VT   ( 2   3 ) I 
   

CR
and  ( X )    1VT  ,  ( X )    4V  A 3 I . (3.3)
 0    VT   VT   T   I  r 
   1 2 1 3 

US
The Jacobian matrix of  ( X ) and (X ) at the infection free equilibrium E0 are,
respectively,
AN
 n22 0
m 0  
D ( E0 )   22 , D ( E0 )   r
 0 0    3 ( 1   2 ) 1  , (3.4)
 1 
M

 r 
0 2 
1   2  3 0 
ED

where, m 22   , n   . (3.5)

r  2 2
  A 3  4 
 0  1 
 1 
PT

From Driessche and Watmough [15], m n 1 is the next generation matrix of model (2.1). It
CE

follows that the spectral radius of matrix m n 1 is

r ( 1 3  A 2  3   1  2 )
 (m n 1 ) 
AC

.
( 3   2 ) 1  4

Therefore, the basic reproduction number of model (2.1) is

r (1 3   2 A 3  1 2 )
0  . (3.6)
(  2   3 ) 1  4
 ACCEPTED MANUSCRIPT

3.2 Local stability

3.2.1 Local stability of non-infected steady state

Theorem 1. The non-infected steady state E0 is locally asymptotically stable in  if 0  1

, but unstable if 0  1 .

Proof. For non-infected steady state E0 , the Jacobian matrix of model (2.1) is

T
 r 
  1 3  ( 1   2 )

IP
 1 
 r 
J  0   2  3 2  (3.7)

CR
 1 
 0 A 3  1   4 
r
 1
 
Therefore, the characteristic equation of matrix J is

(  1 )(2  a1  a2 )  0
US (3.8)
AN

From equation (3.8), one value of  is  1 . The other two values of  are illustrate by the
M

second order equation

(2  a1  a2 )  0
ED

 r  Ar 2 3
where, a1   2   3    4   1    2   3   (1   0 )
 1  1 ( 2   3 )
PT

r r r
a2   2  4   21  1 3   3  4  A 3 2   4 ( 3   2 ) (1  0 ) (3.9)
1 1 1
CE

From equation (3.9), we can state that if 0  1 then a1  0 and a 2  0 . Therefore, Routh-

Hurwitz criterion says that, when 0  1 , the non-infected steady state E0 is locally
AC

asymptotically stable. And, if 0  1 , E0 is unstable.

3.2.2 Local stability of endemic equilibrium state

Except for a non-infected steady state E0 , by simple computation, model (2.1) has the unique
endemic equilibrium state

E1  T 1 , I 1 , V 1
* * *
 for 0  1, (3.10)
 ACCEPTED MANUSCRIPT

( 3   2 )  4 r  1 
where, T1    ,
*

(1 3  1  2  A 2  3 ) 1   0 

 2 (r1 3  A r  2  3  r1  2   3 1  4  1  2  4 ) r2  0  1 

I1 
*
  ,
(1 3   2  2  1  2 )(1 3   2  2  A 2  3 ) (1 3  1  2   2  2 )   0 

(r1 3  r 2 A 3  r1  2   3 1  4  1  2  4 ) r (1 3  1  2  A 2  3 )  0  1 

V1 
*
  . (3.11)
(1 3  1  2   2  2 ) 4  4 (1 3  1  2   2  2 )  0 

T
Theorem 2. The positive endemic equilibrium E1 of the model (2.1) is locally asymptotically

IP
stable if 0  1 .

CR
Proof. For endemic equilibrium state E1 , the Jacobian matrix of model (2.1) is

  1   1V1*   2V1*

J 

 2V1*
 1V1*
3
  2  3
A 3
US
 ( 1   2 )T1 
 2T1*
 1T1   4 
*
*

.

(3.12)
AN


The eigenvalue equation of the Jacobian matrix then becomes

M

3  a12  a2   a3  0 , (3.13)
ED

where,
A 2  3  4
a1   1V1   2V1   3   2  1 
* *
,
( 1 3   1  2  A 2  3 )
PT

A 2 1 3  4
a2  0 (  2   3 ) 1  (1   2 )  4V1 
*
,
(1 3  1 2  A 2  3 )
CE

a3   2  2  4V1  1  2  4V1   31  4V1 ,

* * *

 A 2 1 3  4 
AC

a1a2  a3   2 3  4V1  ( 2   3 ) 0 ( 2   3 ) 1    (1V1*   2V1*  1

*

 ( 
1 3   
1 2  A  
2 3 
)

A 2  3  4  A  2 1  3  4 . (3.14)
    0 ( 2   3 )1  ( 1   2 ) 4V1*  
( 1 3   1  2  A 2  3 )   ( 1 3   1  2  A 2  3 ) 

Therefore, we can easily examine that if 0  1 , then a 1  0 , a 2  0 , a3  0 and

a1a2  a3  0 . Therefore, all conditions of the Routh-Hurwitz criterion are matched. Thus,

the infected steady state E1 is locally asymptotically stable.

 ACCEPTED MANUSCRIPT

3.3 Global stability

3.3.1 Global stability of non-infected steady state

Now, we have to define the global stability for the infection-free equilibrium E0 by
Lyapunov’s second method (see [14-16]).

Theorem 3. If 0  1 , then the non-infected steady state E0 is globally asymptotically stable

on  . If 0  1 , then E0 is unstable.

T
IP
Proof. Let, we introduce the Lyapunov candidate function with the help of [14],

CR
A 3
L I V . (3.15)
2  3
We have,

dL

A 3 dI dV

dt  2   3 dt dt
.
US (3.16)
AN

with the help of equations (2.1), the equation (3.16), is

M

  r (    2 A 3  1  2 ) 
L 1 3  1  4V
 (  2   3 ) 1  4 
ED

  4 (0  1)V . (3.17)

PT


It is simple to see from (3.17) that for 0  1 , L  0 . In addition, if  is the set of solutions

CE

of the model (2.1) where L  0 , in that case the Lyapunov-Lasalle theorem [17] reflects that
all paths in  move towards the leading positivity invariant subset of the set  , where  is
the set of population in which virus is not present i.e., V  0 . On the boundary of  where
AC

 r
V  0 , we have I  0 , T  r  1T . So, T  as t   . Therefore, all solution paths in
1
 move toward the non-infected study state E0 when 0  1 .

It can also be straight forwardly perceived from the Jacobian matrix of the model (2.1) at E0

that one root of eigenvalue equation (3.8) is positive if 0  1 . Thus the non-infected study

state E0 is unstable whenever 0  1 .

 ACCEPTED MANUSCRIPT

Remark 1. The numerical simulation of model (2.1) are shown in Figure 2 and Figure 3
which validates the Theorem 3, where the solution tend to steady state for 0  1 .

3.3.2 Global stability of endemic equilibrium state

Now, we have to discuss the global stability of E1 for model (2.1) using geometric approach
developed by Li and Muldowney [18].

Let the map x  f ( x)  R3 be a C 1 function for x in an open set   R3 . Consider the

T
differential equation

IP
x '  f ( x) (3.18)

CR
is uniquely determined by its initial value x(0)  x0 , and denote this solution by x(t , x0 ) . Li

US
and Muldowney [18] have described the following two assumptions:

(  1 ) There exist a compact absorbing set    .

AN
(  2 ) x is the unique equilibrium of Int  .
M

Lemma 3. Suppose that statement (  1 ) and (  2 ) hold and that equation (3.18) satisfies a

Bendixson criterion that is robust under C 1 local perturbations of f at all non-equilibrium

ED

non wandering points for (3.18). Then x is globally asymptotically stable in region 
provide it is stable.
PT

Theorem 4. If 0  1 , then the unique endemic equilibrium state E1 is globally

asymptotically stable in Int  .
CE

Proof. By Theorem 2, if 0  1 then E1 is the unique equilibrium in the interior of  .

AC

Hence the model (2.1) satisfies the postulate (  2 ). By Theorem 3, the uniform consistency of
the system (2.1) in the interior of  is equivalent to the existence of a compact set    ,
i.e., absorbing for model (2.1). So, it is also satisfies the postulate (  1 ).

We set S as the following diagonal matrix:

 I I
S (T , I , V )  diag 1, ,  . (3.19)
 V V
 ACCEPTED MANUSCRIPT

Followed by, S is under C 1 and non-singular in the interior of  . Let f denote the vector
field of model (2.1). Then

 I ' V ' I ' V '

S f S 1  diag  0,  ,   . (3.20)
 I V I V 

f
The second compound matrix J [ 2] of the Jacobian matrix J  can be defined as follows:
x

T
 1   1V   2V   2   3  2T  1T   2T 
  (3.21)

IP
J [ 2]
 A 3  1   1V   2V   1T   4 3 
   1V  2V   2   3   1T   4 

CR
Therefore, the matrix   S f S 1  S J [ 2] S 1 [18, Theorem 4.3] can be written in the

following block form:


   11
12 
US
 22 
(3.22)
AN
 21

with
M

 T V ( 1T   2T ) V   A I
T

11  1  1V   2V   2   3 , 12   2 ,  ,  21   3 ,  1V  ,
 I   V 
ED

I V 
    1   1V   2V   1T   4 3 
PT

 22 I V .
I V
  2V    2   3   1T   4 
 I V 
CE

The vector norm  in R3 is taken as

AC

(u , v , w)  sup u , v  w . (3.23)

Let  () be the Lozinskii measure with respect to this norm (3.23). Then we have the
following approximation (see [3], [18])

 ( B)  sup g1 , g 2  , (3.24)

where,
 ACCEPTED MANUSCRIPT

g1  1 ( B11 )  B12 , g 2  B21  1 ( B22 ) . (3.25)

Here, B12 and B 21 are the matrix norms induced from the l1 vector norm, and 1 is the

Lozinskii measure with respect to the norm. Therefore

( 1T   2T ) V  A 3 I 
1 (B11)  (1  1V   2V   2   3 ) , B12  , B21   .
I  V 

I V  I V 

T
1 ( B22 )  max    1  1V  1T   4 ,    2  1T   4 
I V I V 

IP
 I V 

CR
     2  1T   4  . (3.26)
I V 

Now, from the second and third equation of the model (2.1), we get

 2T V

I
 2  3 ,
US
A 3 I V 
   1T   4 .
AN
I I V V
Thus,

I  1T V I I I
M

g1    1  1V   2V   1 , g2   2   1 .
I I I I I
ED

At last we have to conclude the result as,

I
 (B )   1 . (3.27)
PT

Let T (t ), I (t ), V (t ) be any non-negative solution of model (2.1) and let t be the time such
CE

that T (t ), I (t ), V (t )    for all t  t . Then for t  t , we have

AC

1 I (t ) 1 (t  t )
t t
1 1
t0  ( B)ds    ( B)ds  ln
t0 t I (t )

t
. (3.28)

Consequently,

1
t

q2  lim sup   ( B)ds   1  0 . (3.29)
t  x t 2
0
 ACCEPTED MANUSCRIPT

Hence, we conclude that if 0  1 , then the infected steady state E1 of the model (2.1) is
globally stable in Int  .

Remark 2. The numerical simulation of model (2.1) are shown in Figure 4 and Figure 5
which validates the Theorem 4, where the solution tend to steady state for 0  1 .

4 Numerical Simulation

The parameter values characteristic of the in vivo circumstance is very complicated; all the

T
parameters in our model (2.1) have not been measured, we are taking several values from

IP
[11]. In this section, we are going to simulate model (2.1) by the numerical technique using

CR
Mathematica with the parameter values taken from Table 1. The numerical results are
demonstrated in Figures 2-5 for two different set of initial values I 1  (1000, 0, 0.001) and
I 2  (1000 ,10,10) .

US
First, we choose r  3 ,  1  0.00002 ,  2  0.000024 ,  3  0.2 ,  1  0.01 ,  2  0.5 ,
AN
 3  0.16 ,  4  3.4 , A  1940 . So, from equation (3.6), the numerical value of reproduction
number ( 0 ) is 0.94  1 . Therefore, Theorem 3 states that the disease dies out. The Figure 2
M

and Figure 3 have confirmed these outcomes. Another case, if we choose r  10 ,

 1  0.00002 ,  2  0.000024 ,  3  0.2 ,  1  0.01 ,  2  0.5 ,  3  0.16 ,  4  3.4 , A  1940 .
ED

Here, the value of reproduction number is 3.13  1. Hence, Theorem 4 shows that AIDS will
persists and in this situation the system goes to infected equilibrium state
PT

(318.90, 6.28, 575.24) . Figure 4 and Figure 5 confirm this conclusion. In this section, we
have also compared our results with the previous paper [11] for a special case and showed
CE

this comparison in Figure 2 (in dashed line) when r  10 ,  1  0 ,  2  0.000024 ,  3  0.2 ,

 1  0.01 ,  2  0 ,  3  0.16 ,  4  3.4 , A  1000 .

AC

For infection-free equilibrium state, when  1  0 , Figure. 2(a) shows that initially the
uninfected CD4+ T cells decreases rapidly with increase of time and after around 20 days it
*
will be lowest, after that it increases until it tends to positive equilibrium point T0 . Another
side, Figure 2(b) demonstrates that first 20 days infected CD4+ T cells increases, after that the
*
nature is reversed, so the infected CD4+ T cells population tending to the point I 0 . At last,
Figure 2(c) clearly depict that the virus population is increases too quickly compare to
 ACCEPTED MANUSCRIPT

infected CD4+ T cells and after around 20 days it will be maximum, later on virus cells
*
decreases drastically and reaches the point V0 .

1000

Uninfected CD4 T cells 800

600

400

T
200

IP
0

CR
0 100 200 300 400 500 600
Time in days
(a)

250

200
US
AN
Infected CD4 T cells

150

100
M

50
ED

0
0 100 200 300 400 500 600
Time in days
(b)
PT

12000

10000
CE

8000
Virus

6000
AC

4000

2000

0
0 100 200 300 400 500 600
Time in days
(c)

Figure 2: The infection free equilibrium E0 is globally asymptotically stable for initial values
I 1  (1000, 0, 0.001) and r  3 ,  2  0.000024 ,  3  0.2 ,  1  0.01 ,  3  0.16 ,  4  3.4 .
 ACCEPTED MANUSCRIPT

But if we consider  1  0 , there is not too much change (around 20 days) in the population
of uninfected CD4+ T cells compared to previous case, while infected CD4+ T cells (almost
double) and virus populations are very much increased (Figure 2).

1000

800
Uninfected CD4 T cells

600

T
400

IP
200

CR
0
0 100 200 300 400 500 600
Time in days
(a)

US
AN
300
Infected CD4 T cells

200
M

100
ED

0
0 100 200 300 400 500 600
Time in days
PT

(b)
CE

15000
AC

10000
Virus

5000

0
0 100 200 300 400 500 600
Time in days
(c)
Figure 3: The infection free equilibrium E0 is globally asymptotically stable for initial values
I 2  (1000, 10, 10) and r  3 ,  2  0.000024 ,  3  0.2 ,  1  0.01 ,  3  0.16 ,  4  3.4 .
 ACCEPTED MANUSCRIPT

For infection-free equilibrium state, again if  1  0 , Figure. 3(a) confirms that initially the
uninfected CD4+ T cells decreases speedily with increase of time and around 10 days it will
*
be lowest, afterward it increases until it tends to T0 . At the same time, Figure 3(b) shows
that first 10 days infected CD4+ T cells increases, later on behaviour is reversed, so the
*
infected CD4+ T cells population tending to the point I 0 . Eventually, Figure 3(c) clearly
illustrate that the virus population is increases quickly compare to infected CD4+ T cells and
around 10 days it will be highest, soon after virus cells decreases drastically and reaches the

T
point V0 . However if we assume  1  0 , Figure 3 shows that again there is not too much
*

IP
change (around 10 days) in the population of uninfected CD4+ T cells compared to earlier

CR
case, while infected CD4+ T cells and virus populations are enormously increased.

1000

800
US
Uninfected CD4 T cells

AN
600

400
M

200
ED

0
0 100 200 300 400 500 600
Time in days
(a)
PT

400
CE

300
Infected CD4 T cells
AC

200

100

0
0 100 200 300 400 500 600
Time in days
(b)
 ACCEPTED MANUSCRIPT

20000

15000

Virus
10000

5000

T
0 100 200 300 400 500 600
Time in days

IP
(c)
Figure 4: The endemic equilibrium E1 is globally asymptotically stable for initial values

CR
I 1  (1000, 0, 0.001) and r  10 ,  2  0.000024 ,  3  0.2 ,  1  0.01 ,  3  0.16 ,  4  3.4 .

US
For infection equilibrium state, when  1  0 , Figure. 4(a) shows that initially the uninfected
CD4+ T cells decreases rapidly with increase of time and around 20 days it will be least, after
AN
that it increases and around 90 days again it decreases. This behaviour is continuing with
*
increase of time until it tends to positive equilibrium point T1 . Simultaneously, Figure 4(b)
M

demonstrates that first 20 days infected CD4+ T cells increases after that it decreases and
around 70 days it again increases and after 90 days again it decreases. This behaviour is
ED

*
continuing until it tends to non-zero equilibrium point I1 . At last, Figure 4(c) clearly depict
that the virus population is increases too quickly compare to infected CD4+ T cells and
PT

reaches the point V1 . But if we consider  1  0 , there is not too much change (around 20
*

days) in the population of uninfected CD4+ T cells compared to previous case, while infected
CE

CD4+ T cells and virus populations are much higher than previous case (Figure 4).

For infection equilibrium state, when  1  0 , Figure. 5(a) shows that initially the uninfected
AC

CD4+ T cells decreases rapidly with increase of time and around 8 days it will be least, after
that it increases and around 90 days again it decreases. This behaviour is continuing with
*
increase of time until it tends to positive equilibrium point T1 . Simultaneously, Figure 5(b)
demonstrates that first 8 days infected CD4+ T cells increases after that it decreases and
around 70 days it again increases and after 90 days again it decreases. This behaviour is
*
continuing until it tends to non-zero equilibrium point I1 . At last, Figure 5(c) clearly depict
that the virus population is increases too quickly compare to infected CD4+ T cells and
 ACCEPTED MANUSCRIPT

reaches the point V1 . But if we consider  1  0 , there is not too much change (around 8
*

days) in the population of uninfected CD4+ T cells compared to previous case, while infected
CD4+ T cells and virus populations are much higher than previous case (Figure 5).

1000

800
Uninfected CD4 T cells

600

T
400

IP
200

CR
0
0 100 200 300 400 500 600
Time in days

400
US
(a)
AN
Infected CD4 T cells

300
M

200

100
ED

0
0 100 200 300 400 500 600
PT

Time in days
(b)
20000
CE

15000
AC
Virus

10000

5000

0
0 100 200 300 400 500 600
Time in days
(c)
E
Figure 5: The endemic equilibrium 1 is globally asymptotically stable for initial values
I 2  (1000, 10, 10) and r  10 ,  2  0.000024 ,  3  0.2 ,  1  0.01 ,  3  0.16 ,  4  3.4 .
 ACCEPTED MANUSCRIPT

Conclusion

In this study, the authors assume an extremely small transition period for a weak uninfected
CD4+ T cell. As a result, when weak CD4+ T-cells interacted with HIV virus then some of
these weak CD4+ T-cells has directly move in the virus class, which is one of the important
factor to rapidly spread the HIV. Another main result is that, we cannot ignore the natural
recovery of CD4+ T-cells because the good number of T-cells has recovered. Here, we also
discussed a detailed stability analysis of a novel model that describes the new dynamics of

T
infection HIV/AIDS with new conditions. The basic properties i.e., positivity of the solutions

IP
and invariant regions of the system has been derived, and locate the limitation of solutions
and the initial conditions. The basic reproduction number, R0 , is calculated by next generation

CR
matrix method and we validate that if it is less than one, the infection-free steady state is
globally stable. Otherwise, the infection-free steady state becomes unstable and there arises

US
an endemic steady state which is globally asymptotically stable. For two different set of
parameters, numerical simulations has graphically shown in Section 4 and confirm that
AN
analytical results of the model are acceptable.

Acknowledgement
M

The authors express their heartfelt thanks to the revered reviewers for their valuable
suggestions for the improvement of the article. The first author acknowledges the financial
ED

support from the MHRD, New Delhi, India under the JRF scheme and second author
acknowledges the NIT Silchar, Assam, India for providing the facility under the STIS
PT

scheme.

References
CE

1. G. Haas, A. Hosmalin, F. Hadida, J. Duntze, P. Debré, B. Autran, Imm. Lett., 57

(1997), p. 63.
AC

2. D. Burg, L. Rong, A.U. Neumann, H. Dahari, J. Theor. Bio., 259 (2009), p. 751.
3. J. Luo, W. Wang, H. Chen, R. Fu, J. Math. Anal. Appli., 434 (2016), p. 837.
4. A.S. Perelson, D.E. Kirschner, R-DE. Boer, Math. Biosci., 114 (1993), p. 81.
5. D.S. Callaway, A.S. Perelson, Bull. Math. Bio., 64 (2002), p. 29.
6. Y. Emvudu, D. Bongor, Int. Elec. J. Pure and App. Math., 4 (2012), p. 297.
7. Y. Emvudu, D. Bongor, R. Koïna, App. Math. Mod., 40 (2016), p. 9131.
8. M. Joly, J.M. Pinto, Math. Comp. Mod., 55 (2012), p. 342.
 ACCEPTED MANUSCRIPT

9. A. Mojaver, H. Kheiri, Appl. Math. Comp., 259 (2015), p. 258.

10. S. Wang, P. Hottz, M. Schechter, L. Rong, PLoS Comp. Bio., 11 (2015), p. e1004665.
11. P.K. Srivastava, P. Chandra, Nonlin. Anal.: Real World Appl., 11 (2010), p. 612.
12. P. Essunger, A.S. Perelson, J. Theor. Bio., 170 (1994), p. 367.
13. L. Rong, M.A. Gilchrist, Z. Feng, A.S. Perelson, J. Theor. Bio., 247 (2007), p. 804.
14. H.-F. Huo, R. Chen, X.-Y. Wang, App. Math. Mod., 40 (2016), p. 6550.
15. P.V. Driessche, J. Watmough, Math. Bioscie., 180 (2002), p. 29.
16. S. Wiggins, Introduction to Applied Nonlinear Dynamical System and Chaos, New

T
York City: Springer Verlag, 2003.

IP
17. H.J. Marquez, Nonlinear Control Systems Analysis and Design, Wiley, 2003.

CR
18. M.Y. Li, J.S. Muldowney, SIAM J. Math. Anal., 27 (1996), p. 1070.

US
AN
M
ED
PT
CE
AC