Abstract
Resistance to chemotherapy is a major cause of cancer treatment failure. The processes of resistance induction and selection of resistant cells (due to the over-expression of the membrane transporter P-glycoprotein, P-gp) are well documented in the literature, and a number of mathematical models have been developed. However, another process of transfer of resistant characteristics is less well known and has received little attention in the mathematical literature. In this paper, we discuss the potential of simple mathematical models to describe the process of resistance transfer, specifically P-gp transfer, in mixtures of resistant and sensitive tumor cell populations. Two different biological hypotheses for P-gp transfer are explored: (1) exchange through physical cell–cell connections and (2) through microvessicles released to the culture medium. Two models are developed which fit very well the observed population growth dynamics. The potential and limitations of these simple “global” models to describe the aforementioned biological processes involved are discussed on the basis of the results obtained.
Similar content being viewed by others
References
Aktipis CA, Kwas VS, Johnson KA, Neuberg SL, Maley CC (2011) Overlooking evolution: a systematic analysis of cancer relapse and therapeutic resistance research. PLoS One 6:e26100
Ambudkar SV, Sauna ZE, Gottesman MM, Szakacs GG (2005) A novel way to spread drug resistance in tumor cells: functional intercellular transfer of P-glycoprotein (ABCB1). Trends Pharmacol Sci 26:385–387
Boelens MC, Wu TJ, Nabet BY, Xu B, Qiu Y, Yoon T, Azzam DJ, Wiemann BZ, Ishwaran H, Ter Brugge PJ, Jonkers J, Slingerland J, Minn AJ (2014) Exosome transfer from stromal to breast cancer cells regulates therapy resistance pathways. Cell 159:499–513
Chisholm RH, Lorenzi T, Lorz A, Larsen AK, de Almedia LN, Escargueil A, Clairambault J (2015) Emergence of drug tolerance in cancer cell populations: an evolutionary outcome of selection, nongenetic instability, and stress-induced adaptation. Cancer Res 75:930–939
Gerlinger M, Swanton C (2010) How Darwinian models inform therapeutic failure initiated by clonal heterogeneity in cancer medicine. Br J Cancer 103:1139–1143
Gillet GP, Gottesman MM (2010) Mechanisms of multidrug resistance in cancer. Methods Mol Biol 596:47–76
Gillies RJ, Verduzco D, Gatenby RA (2012) Evolutionary dynamics of carcinogenesis and why targeted therapy does not work. Nat Rev Cancer 12:487–493
Gong J, Jaiswal R, Mathys JM, Combes V, Grau GE, Bebawy M (2012) Microparticles and their emerging role in cancer multidrug resistance. Cancer Treat Rev 38:226–234
Greene J, Lavi O, Gottesman MM, Levy D (2014) The impact of cell density and mutations in a model of multidrug resistance in solid tumors. Bull Math Biol 76:627–653
Holohan C, Van Schaeybroeck S, Longley DB, Johnston PG (2013) Cancer drug resistance: an evolving paradigm. Nat Rev Cancer 13:714–725
Korolev KS, Xavier JB, Gore J (2014) Turning ecology and evolution against cancer. Nat Rev Cancer 14:371–380
Lavi O, Greene JM, Levy D, Gottesman MM (2013) The role of cell density and intratumoral heterogeneity in multidrug resistance. Cancer Res 73:7168–7175
Levchenko A, Mehta BM, Niu X, Kang G, Villafania L, Way D, Polycarpe D, Sadelain M, Larson SM (2005) Intercellular transfer of P-glycoprotein mediates acquired multidrug resistance in tumor cells. Proc Nat Acad Sci 102:1933–1938
Limame R, Wouters A, Pauwels B, Fransen E, Peeters M, Lardon F, De Wever O et al (2012) Comparative analysis of dynamic cell viability, migration and invasion assessments by novel real-time technology and classic endpoint assays. PloS one 7(10):e46536
Lorz A, Lorenzi T, Hochberg M, Clairambault J, Perthame B (2013) Populational adaptive evolution, chemotherapeutic resistance and multiple anti-cancer therapies. Math Model Num Anal 47:377–399
Murray JD (2007) Mathematical biology: I. An introduction. Springer, New York
Obenauf AC, Zou Y, Ji AL, Vanharanta S, Shu W, Shi H, Kong X, Bosenberg MC, Wiesner T, Rosen N, Lo RS, Massagué J (2015) Therapy-induced tumour secretomes promote resistance and tumour progression. Nature 520:368–372
Oswald M et al (2015) Brain tumour cells interconnect to a functional and resistant network. Nature. doi:10.1038/nature16071
Pasquier J, Magal P, Boulange-Lecomte C, Webb G, Le Foll F (2011) Consequences of cell-to-cell P-glycoprotein transfer on acquired multidrug resistance in breast cancer: a cell population dynamics model. Biol Direct 6:5
Perko L (1991) Differential equations and dynamical systems. Springer, New York
Pešić M, Marković J, Janković D, Marković I, Rakić LJ, Kanazir S, Ruždijić S (2006) Induced resistance in the human non small cell lung carcinoma (NCI-H460) cell line in vitro by anticancer drugs. J Chemother 18(1):66–73
Pisco AO, Brock A, Zhou J, Mojtahedi M, Jackson D, Huang S (2013) Non-Darwinian dynamics in therapy-induced cancer drug resistance. Nat Commun 4:2467
Sousa D, Lima RT, Vasconcelos MH (2015) Intercellular transfer of cancer drug resistance traits by extracellular vesicles. Trends Mol Med 21:595–608
Turner NC, Reis-Filho JS (2012) Genetic heterogeneity and cancer drug resistance. Lancet Oncol 13:e178–185
Zhareddine H, Borden LB (2013) Mechanisms and insights into drug resistance in cancer. Front Pharmacol 4:28
Zomer A, Maynard C, Verweij FJ, Kamermans A, Schäfer R, Beerling E, Schiffelers RM, de Wit E, Berenguer J, Ellenbroek SI, Wurdinger T, Pegtel DM, van Rheenen J (2015) In vivo imaging reveals extracellular vesicle-mediated phenocopying of metastatic behavior. Cell 161:1046–1057
Acknowledgments
We would like to acknowledge Gabriel F. Calvo (Universidad de Castilla-La Mancha) for discussions. The authors would also like to thank the anonymous referees for valuable suggestions. This research has been partially supported by Grants MTM2015-71200-R (Ministerio de Economía y Competitividad/FEDER) and PEII-2014-031P (Junta de Comunidades de Castilla-La Mancha and FEDER).
Author information
Authors and Affiliations
Corresponding author
Appendix
Appendix
1.1 Basic Properties of Model Eq. (1)
First we state the following:
Theorem 1
For any nonnegative initial data \((S_{0},R_{0},S_{R0})\) and \(\tau _{s}> 0\), \(\tau _{r}> 0\), the solutions to Eq. (1) exist for \(t>0\), are nonnegative, bounded and unique.
Proof
Let \(N=S+R+S_{R}\); then, defining \(N=S+R+S_{R}\) we can write Eq. (1) as
with \(\gamma = 1/\min \{ \tau _{s}, \tau _{r}\}\). Therefore N(t) is bounded and thus S, R and \(S_{R}\) are bounded as well. Since Eq. (1) has bounded coefficients, and the right-hand side of Eq. (5) is a continuous function in \((S,R,S_{R})\), it is straightforward to obtain the existence of solutions of Eq. (1) (as detailed in Perko (1991)). As the partial derivatives of the right-hand side of the system are continuous and bounded, uniqueness follows from the Picard–Lindelof theorem.
Representing the solutions of Eq. (1) in integral form we get:
Similarly, it is straightforward to prove the same inequalities for S(t) and \(S_{R}(t)\). Thus, if the initial conditions are nonnegative, the components of the solution are nonnegative for all finite time t. \(\square \)
In order to understand better the dynamics of Eq. (1), we will first calculate the fixed points and determine their stability. These are the isolated point \(P_{1}=(0,0,0)\) and the family of points of the intersection between the plane and the surface
which can be written in parametric form as
To analyze the stability of these points, we calculate the Jacobian matrix of Eq. (1):
Equilibrium point \(P_{1}=(0,0,0)\). The Jacobian matrix is
and the eigenvalues \(\mu _{1}= 1/\tau _s, \mu _{2}=1/\tau _{r}\) and \(\mu _{3}=1/\tau _r-1/\tau _{*}\). As in our model, because of the biological meaning of the parameters, \(1/\tau _r < 1/\tau _*\), \(\mu _1, \mu _2>0\) and \(\mu _3<0\) and point \(P_{1}\) is a saddle point and therefore, an unstable point.
Equilibrium points \(P_{2}(\lambda )= \left( \lambda , \frac{K-\lambda }{1+\tau _*\lambda /\tau _c}, \frac{\lambda (K-\lambda )}{\lambda +\tau _c/\tau _*}\right) \).
The Jacobian matrix for all these points is:
The eigenvalues of the matrix are
Since \(\mu _{1}=0\), these points are non-hyperbolic points. Moreover, as \(\mu _{2}\) and \(\mu _{3}\) are negative, for the parameters calculated in the article, the fixed points \(P_{2}(\lambda )\) possess two local stable manifolds (corresponding to the eigenvalues \(\mu _{2}\) and \(\mu _{3}\)) each of dimension 1 and a local center manifold (corresponding to the eigenvalue \(\mu _{1}\)) also of dimension 1.
At this point, since these stationary points are non-hyperbolic, the linear approximation is not valid. As \(\mu _{2}\) and \(\mu _{3}\) are negative, all the orbits starting near these equilibrium points approach the center manifold (see Fig. 11). Since the center manifold is formed by the curve of equilibrium points (8), all the orbits starting near these points tend to them. Thus, these points are stable.
The phase portrait for solutions of system (1) is shown in Fig. 11.
1.2 Basic Properties of Model Eq. (2)
For model Eqs (2), we have:
Theorem 2
For any nonnegative initial data \((S_{0},R_{0},S_{R0}, Q_{0})\) and \(\tau _{s}> 0\), \(\tau _{r}>0\), the solutions to Eq. (2) exist for \(t>0\), are nonnegative and unique.
Proof
The proof mimics the steps of the proof of theorem 1 with the exception of the boundedness of the solution Q, which is not bounded. The remainder of the steps work in the same way so the repetitive details are omitted. \(\square \)
The equilibrium points of Eq. (2) are grouped in two different sets:
Equilibrium points \(E_1=\left( S,R,S_R, Q\right) \) are given by
For these points, their corresponding eigenvalues are:
where \(f(\alpha )\) is a continuous function depending on \(\alpha \). As there is at least one positive eigenvalue, \(\mu _{2}\), these points are all unstable equilibrium points. It has a biological explanation: Without cells, the concentration of MVs will not be affected, and the system will never tend to equilibrium.
Equilibrium points \(E_2(\alpha )=\left( S,R,S_R, Q\right) \), these points satisfy the condition \(S+R+S_{R} = K\) and they can be written as points of the following parametrized curve
where \(0 \le \alpha < \infty \). For these points, their corresponding eigenvalues are:
and
where \(g(\alpha )\) and \(h(\alpha )\) are continuous functions depending on \(\alpha \). A graph of these eigenvalues is shown in Fig. 11, for the values of the parameters previously calculated.
It is straightforward to check that \(\beta _{2}\) and \(\beta _{4}\) do not cross the \(\alpha \)-axis. Thus, we obtain two negative \(\beta _2\), \(\beta _4\) and one null \(\beta _1\) eigenvalues.
Figure 11 shows that the eigenvalue \(\beta _3\) is positive or negative depending on the value of \(\alpha \). If \(\alpha \le 0.67\), all eigenvalues are non-positive, and, similarly, as in Appendix, we conclude that this subset of equilibrium points is stable. On the other hand, if \(\alpha > 0.67\), then the eigenvalue \(\beta _3\) is positive, and the equilibrium points \(E_2(\alpha )\) are unstable.
Therefore, the point \(E_2(0.67)\) is a bifurcation point.
Rights and permissions
About this article
Cite this article
Rosa Durán, M., Podolski-Renić, A., Álvarez-Arenas, A. et al. Transfer of Drug Resistance Characteristics Between Cancer Cell Subpopulations: A Study Using Simple Mathematical Models. Bull Math Biol 78, 1218–1237 (2016). https://doi.org/10.1007/s11538-016-0182-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11538-016-0182-0