Interplay of Darwinian Selection, Lamarckian Induction and Microvesicle Transfer on Drug Resistance in Cancer

Abstract

Development of drug resistance in cancer has major implications for patients' outcome. It is related to processes involved in the decrease of drug efficacy, which are strongly influenced by intratumor heterogeneity and changes in the microenvironment. Heterogeneity arises, to a large extent, from genetic mutations analogously to Darwinian evolution, when selection of tumor cells results from the adaptation to the microenvironment, but could also emerge as a consequence of epigenetic mutations driven by stochastic events. An important exogenous source of alterations is the action of chemotherapeutic agents, which not only affects the signalling pathways but also the interactions among cells. In this work we provide experimental evidence from in vitro assays and put forward a mathematical kinetic transport model to describe the dynamics displayed by a system of non-small-cell lung carcinoma cells (NCI-H460) which, depending on the effect of a chemotherapeutic agent (doxorubicin), exhibits a complex interplay between Darwinian selection, Lamarckian induction and the nonlocal transfer of extracellular microvesicles. The role played by all of these processes to multidrug resistance in cancer is elucidated and quantified.

Figure 1

Possible scenarios displaying the response of sensitive and resistant cancer cell subpopulations to the absence/presence of a specific drug and intercellular communication. (Left panel) Selection and/or induction drive the cell number and P-gp expression kinetics when no communication exists between these two subpopulations. (Right panel) In addition to selection and/or induction, a transfer mechanism may arise when both subpopulations are in contact via an extracellular medium through which they can exchange microvesicles (MVs). Color code: blue and red for constitutively sensitive and resistant cells, respectively.

Figure 2

Experimental results for H460 (left), H460/R (middle) and mix 1:1 (right) cells. Arrows account for the processes that are responsible for the changes observed in P-gp expression. Notice that the areas under each curve, which represent the number of cells, are equal in all cases (all cell samples were of the same size $\simeq 2\times {10}^4$ in the flow cytometry analyses).

Figure 3

Evolution of H460 and H460/R cell mix 1:1 (a) without and (b) with DOX. Cells were seeded at t = 0 h and measured at subsequent intervals of 24 h. Dashed and solid lines represent the experimental results and the numerical simulations of our model, respectively.

Figure 4

P-gp distribution changes after 72 h under (a) 0 nM and (b) 50 nM of DOX for initial sensitive, resistant and mixed cells (sensitive:resistant fractions 1:1, 3:1 and 7:1). Insets: Cell number versus time for initial populations of sensitive, resistant and mixed cells with 0 and 50 nM of DOX.

Figure 5

Detection of P-gp transfer during 48 h in different culture media with sensitive cells ($\simeq 2\times {10}^4$). Upper/lower rows represent the experimental results and the model predictions, with the corresponding calculated MV transfer function (inset). The culture media comprised only the H460 cells (blue solid curves), H460 cells grown in a conditioned medium (CM) exchanged from the H460/R cells to the H460 cells medium (red dashed curves), H460 cells grown in the presence of 50 nM of DOX (yellow dotted curves) and H460 cells grown both in the presence of 50 nM of DOX and CM (violet dashed-dotted curves). Asterisks in the upper row denote the p-values for pairwise comparisons: (**)p < 0.005 and (***)p < 0.0005.

Figure 6

Duration of P-gp changes. The upper four rows correspond to the numerical simulations over a time period of 240 h, while the last row shows the experimental results (at t = 240 h).

Figure 7

Treatment response of (a) H460, (b) RH460 and (c) Mix 1:1 cell groups to the three different protocols using a 100 nM drug concentration of DOX and a control group (no drug). The initial cell number was 2 × 10⁴ cells in all cases.

Figure 8

Portrait of the dynamics of P-gp expression and cell number under the action of Darwinian selection, Lamarckian induction and MV transfer on an initial population of sensitive cells with a continuous supply of MVs. The impact of these processes on P-gp expression and cell number can be quite dramatic in the presence of drugs (dotted, dashed-dotted and dashed curves). Darwinian selection yields the slowest dynamics whereas Lamarckian induction and MV transfer significantly accelerate the dynamics towards resistant populations.