Dynamics of a stochastic cell-to-cell HIV-1 model with distributed delay

https://doi.org/10.1016/j.physa.2017.11.035Get rights and content

Highlights

  • A stochastic cell-to-cell HIV-1 model with distributed delay is proposed and investigated.

  • We establish sufficient conditions for extinction of the disease.

  • We establish sufficient conditions for the existence of an ergodic stationary distribution.

  • The stationary distribution implies that the disease can be persistent in the mean.

Abstract

In this paper, we consider a stochastic cell-to-cell HIV-1 model with distributed delay. Firstly, we show that there is a global positive solution of this model before exploring its long-time behavior. Then sufficient conditions for extinction of the disease are established. Moreover, we obtain sufficient conditions for the existence of an ergodic stationary distribution of the model by constructing a suitable stochastic Lyapunov function. The stationary distribution implies that the disease is persistent in the mean. Finally, we provide some numerical examples to illustrate theoretical results.

Introduction

Human immunodeficiency virus-1 (HIV-1) continues to be a major global public health issue and priority. Many scientists have made great effort against HIV-1, and they are still going on. Mathematics and biological researchers also contribute to this by revealing its transmission and dynamics. Recently, many mathematical models have been formulated to describe the immunological response to infection with HIV-1. Most of these models focus on cell-free viral spread in a compartment such as the bloodstream, see for example, Callaway and Perelson [1], Spouge, Shrager and Dimitrov [2], Kirschner, Lenhart and Serbin [3], Kirschner and Webb [[4], [5], [6] ], McLean and Kirkwood [7], McLean and Nowak [8], Müller et al. [9], Nowak and Bangham [10], Nowak and May [[11], [12]], Perelson, Kirschner and De Boer [13], Perelson [14], Perelson and Nelson [15], Wodarz et al. [16], etc. Some most advances in areas of modeling cell-to-cell transmission and of modeling physical processes with distributed time delay [[17], [18], [19], [20] ]. And these models have been used to explain different phenomena. This is because HIV-1 mathematical models can provide insights into the dynamics of viral load in vivo and can play an important role in the development of a better understanding of HIV/AIDS and drug therapies. Especially, by assuming that infection is spread directly from infected cells to healthy cells and neglecting the effects of free virus, Culshaw et al. [17] considered a two-dimensional model of cell-to-cell spread of HIV-1 in tissue cultures dC(t)dt=rCC(t)(1C(t)+I(t)CM)κIC(t)I(t),dI(t)dt=κItC(u)I(u)F(tu)duμII(t),where C(t) and I(t) denote the concentration of healthy cells and infected cells at time t, respectively and all of the parameters are positive constants. rC is the effective reproductive rate of healthy cells (the term is the total reproductive rate for healthy cells r minus the death rate for healthy cells μC), and so rCC(t) denotes the number of effective reproductive cells per unit time, CM is the effective carrying capacity of system (1.1), κI denotes the infection of healthy cells by the infected cells in a well-fixed system, κIκI is the fraction of cells surviving the incubation period, μI is the death rate of the infected cells. It is assumed that the cells, which are productively infectious at time t, were infected u time units ago, where u is distributed according to a probability distribution F(u), called the delay kernel.

Taking the weak kernel function F(u)=αeαu (α>0) and letting X(t)=tαeα(tu)C(u)I(u)du,system (1.1) is equivalent to the following system dC(t)dt=rCC(t)(1C(t)+I(t)CM)κIC(t)I(t),dI(t)dt=κIX(t)μII(t),dX(t)dt=αC(t)I(t)αX(t).There are three equilibria in system (1.2): the trivial equilibrium E0=(0,0,0), the healthy equilibrium E1=(CM,0,0), and the infected equilibrium E¯=(C¯,I¯,X¯) provided κI>μICM, where C¯=μIκI, I¯=rC(κICMμI)κI(κICM+rC), X¯=μIκII¯. When CM<μIκI, the healthy cells predominate and infected cells die exponentially. In this case E0,E¯ are unstable, E1 is asymptotically stable. When μIκI<CM<rCCM, E0 remains unstable and E1 is also unstable. In this situation, healthy cells and infected cells co-exist. Furthermore, if a1(α)>0, a3(α)>0 and a1(α)a2(α)a3(α)>0, then the positive steady state E¯ is asymptotically stable, where a1(α)=rCCMC¯+μI+α, a2(α)=α(rCCMC¯)+μIrCCMC¯, a3(α)=α(κI+rCCM)μII¯ [17].

On the other hand, in the real world, epidemic models are inevitably subject to the environmental noise, which is an important component in an ecosystem (see e.g. [[21], [22]]). Hence the deterministic models have some limitations in predicting the future dynamics of the system accurately. When modeling biological phenomena such as HIV dynamics, different cells and infective virus particles reacting in the same environment can often give different results. Lately, by incorporating the effects of a fluctuating environment, many authors have studied epidemic models with parameter perturbations (see e.g. [[23], [24], [25], [26], [27], [28], [29], [30], [31] ]). For example, Ji and Jiang [23] considered a stochastic HIV-1 infection model with cell-mediated immune response. They established a sufficient condition for the stochastic asymptotic stability in the large of the infection-free equilibrium and gave the conditions for the solution fluctuating around the two infection equilibria (one without CTLs being activated and the other with). Sánchez-Taltavull et al. [24] presented a stochastic model of the dynamics of the HIV-1 infection and studied the effect of the rate of latently infected cell activation on the average extinction time of the infection. Liu [30] analyzed a model of cell-to-cell HIV-1 infection to CD4+ T cells perturbed by stochastic perturbations. He studied the asymptotic behavior of the solution and he also investigated the existence of ergodic stationary distribution.

There are different approaches to introduce random perturbations into the model, both from a mathematical and biological perspective. In this paper, we assume that the environmental noise is proportional to the variables C(t) and I(t). For convenience in mathematics, we also assume that the environmental noise is proportional to X(t) (see Remark 4.1). Then the stochastic version corresponding to system (1.2) takes the following form dC(t)=[rCC(t)(1C(t)+I(t)CM)κIC(t)I(t)]dt+σ1C(t)dB1(t),dI(t)=[κIX(t)μII(t)]dt+σ2I(t)dB2(t),dX(t)=[αC(t)I(t)αX(t)]dt+σ3X(t)dB3(t),where Bi(t) (i=1,2,3) are mutually independent standard Brownian motions defined on a complete probability space (Ω,F,{Ft}t0,P) with a filtration {Ft}t0 satisfying the usual conditions (i.e., it is increasing and right continuous while F0 contains all P-null sets), σi2>0(i=1,2,3) denote the intensities of white noise, the other parameters in system (1.3) have the same meaning as in system (1.2).

The initial value of system (1.3) is as follows C(s)=ϕ(s)0,I(s)=ψ(s)0,s(,0],where ϕ and ψ are continuous functions on (,0].

Throughout this paper, we let R+3={x=(x1,x2,x3)R3:xi>0,i=1,2,3}. This paper is organized as follows. In Section 2, we show that there exists a unique global positive solution of system (1.3) with the initial value (1.4). In Section 3, we establish sufficient conditions for extinction of the disease. In Section 4, by constructing a suitable stochastic Lyapunov function and a rectangular set, we verify the existence of an ergodic stationary distribution of system (1.3). Finally, we provide some numerical examples to illustrate theoretical results and give a brief discussion.

Section snippets

Existence and uniqueness of the positive solution

To study the dynamical behavior of an epidemic model, the first concern is whether the solution is global and positive. In this section, motivated by the methods mentioned in [21] and [[32], [33], [34], [35] ], we show that there is a unique global positive solution of system (1.3).

Theorem 2.1

For any initial value (1.4), there is a unique solution (C(t),I(t),X(t)) of system (1.3)on R and the solution will remain in R+3 with probability one, that is to say, (C(t),I(t),X(t))R+3 for all

Extinction of the disease

In this section, we shall consider the extinction of the disease. For convenience and simplicity in the following analysis, define the following notations. R̂0=2κICM(rCσ122)rC[(μI+σ222)σ322]and ft=1t0tf(s)ds,when f is an integrable function on [0,).

Theorem 3.1

Let (C(t),I(t),X(t)) be the solution of system (1.3)with any initial value (1.4). If rC>σ122 and R̂0<1, then limtCt=CM(rCσ122)rCa.s. and lim supt1tln[αI(t)+κIX(t)]12[(μI+σ222)σ322](R̂01)<0a.s.

Stationary distribution of system (1.3)

When considering epidemic dynamical systems, we are also interested in when the disease will persist and prevail in a population. In the deterministic models, it can be solved by proving that the endemic equilibrium of the corresponding model is a global attractor or is globally asymptotically stable. But for system (1.3), there exists no endemic equilibrium. In this section, based on the theory of Has’minskii [39], we prove that there is a stationary distribution [40], which reveals that the

Discussion and conclusion

In this section, we give some examples to illustrate the obtained theoretical results and give a brief discussion. For the numerical simulation, we use Milstein’s Higher Order Method mentioned in [43] to obtain the following discretization transformation of system (1.3) Cj+1=Cj+[rCCj(1Cj+IjCM)κICjIj]Δt+σ1CjΔtε1,j+σ122Cj(ε1,j21)Δt,Ij+1=Ij+[κIXjμIIj]Δt+σ2IjΔtε2,j+σ222Ij(ε2,j21)Δt,Xj+1=Xj+[αCjIjαXj]Δt+σ3XjΔtε3,j+σ322Xj(ε3,j21)Δt,where the time increment Δt>0, σi2>0 denote the intensities

Acknowledgments

This work was supported by NSFC of China Grant Nos. 11371085, 11601043, the Fundamental Research Funds for the Central Universities (No. 15CX08011A), China Postdoctoral Science Foundation (No. 2016M590243) and sponsored by Jiangsu Province “333 High-Level Personnel Training Project” (Grant No. BRA2017468), Qing Lan Project of Jiangsu Province of 2016 and 2017.

References (43)

  • MengX. et al.

    Dynamics of a novel nonlinear stochastic SIS epidemic model with double epidemic hypothesis

    J. Math. Anal. Appl.

    (2016)
  • RazookyB.S. et al.

    Mapping the architecture of the HIV-1 Tat circuit: A decision-making circuit that lacks bistability and exploits stochastic noise

    Methods

    (2011)
  • Sánchez-TaltavullD. et al.

    Stochastic modelling of viral blips in HIV-1-infected patients: Effects of inhomogeneous density fluctuations

    J. Theor. Biol.

    (2015)
  • LiuQ.

    Asymptotic behaviors of a cell-to-cell HIV-1 infection model perturbed by white noise

    Physica A

    (2017)
  • ImmonenT. et al.

    A hybrid stochastic-deterministic computational model accurately describes spatial dynamics and virus diffusion in HIV-1 growth competition assay

    J. Theor. Biol.

    (2012)
  • ZhangS. et al.

    Dynamics analysis and numerical simulations of a stochastic non-autonomous predator–prey system with impulsive effects

    Nonlinear Anal. Hybrid Syst.

    (2017)
  • LiuL. et al.

    Optimal control strategy for an impulsive stochastic competition system with time delays and jumps

    Physica A

    (2017)
  • ZhaoY. et al.

    Stochastic periodic solution of a non-autonomous toxic-producing phytoplankton allelopathy model with environmental fluctuation

    Commun. Nonlinear Sci. Numer. Simul.

    (2017)
  • JiC. et al.

    Analysis of a predator–prey model with modified Leslie-Gower and Holling-type II schemes with stochastic perturbation

    J. Math. Anal. Appl.

    (2009)
  • ZhaoY. et al.

    The stationary distribution and ergodicity of a stochastic phytoplankton allelopathy model under regime switching

    Commun. Nonlinear Sci. Numer. Simul.

    (2016)
  • LiuQ. et al.

    Dynamical behavior of a stochastic HBV infection model with logistic hepatocyte growth

    Acta Math. Sci.

    (2017)
  • Cited by (0)

    View full text