Limits...
Temporally sequenced anticancer drugs overcome adaptive resistance by targeting a vulnerable chemotherapy-induced phenotypic transition.

Goldman A, Majumder B, Dhawan A, Ravi S, Goldman D, Kohandel M, Majumder PK, Sengupta S - Nat Commun (2015)

Bottom Line: Understanding the emerging models of adaptive resistance is key to overcoming cancer chemotherapy failure.This state is associated with a clustering of CD44 and CD24 in membrane lipid rafts, leading to the activation of Src Family Kinase (SFK)/hemopoietic cell kinase (Hck) and suppression of apoptosis.The use of pharmacological inhibitors of SFK/Hck in combination with taxanes in a temporally constrained manner, where the kinase inhibitor is administered post taxane treatment, but not when co-administered, markedly sensitizes the chemotolerant cells to the chemotherapy.

View Article: PubMed Central - PubMed

Affiliation: 1] Department of Medicine, Harvard Medical School, Boston, Massachusetts 02115, USA [2] Harvard-MIT Division of Health Sciences and Technology, Cambridge, Massachusetts 02139, USA [3] Division of Biomedical Engineering, Department of Medicine, Brigham and Women's Hospital, Boston, Massachusetts 02115, USA.

ABSTRACT
Understanding the emerging models of adaptive resistance is key to overcoming cancer chemotherapy failure. Using human breast cancer explants, in vitro cell lines, mouse in vivo studies and mathematical modelling, here we show that exposure to a taxane induces phenotypic cell state transition towards a favoured transient CD44(Hi)CD24(Hi) chemotherapy-tolerant state. This state is associated with a clustering of CD44 and CD24 in membrane lipid rafts, leading to the activation of Src Family Kinase (SFK)/hemopoietic cell kinase (Hck) and suppression of apoptosis. The use of pharmacological inhibitors of SFK/Hck in combination with taxanes in a temporally constrained manner, where the kinase inhibitor is administered post taxane treatment, but not when co-administered, markedly sensitizes the chemotolerant cells to the chemotherapy. This approach of harnessing chemotherapy-induced phenotypic cell state transition for improving antitumour outcome could emerge as a translational strategy for the management of cancer.

No MeSH data available.


Related in: MedlinePlus

Modelling the induction of CD44HiCD24Hi cells.(a) Schematic shows experimental design used to derive mathematical parameters of population dynamics. Treatment of MDA-MB-231 breast cancer cells with 25 nM docetaxel (DTX) for 24 h induces phenotype plasticity rather than providing a selection pressure. In parallel, starting cells with different permutations and combinations of CD24 and CD44 expression levels were used, and the expression of CD44CD24 was monitored over defined time points. (b) Population dynamics modelling derived from experimental data indicates temporal kinetics of breast cancer cells in distinct compartments over 5 days (CD44Lo described as non-CSC). Left panel shows dynamics of distinct phenotypes under basal conditions, right panel demonstrates population dynamics under chemotherapy pressure. (c) Schematic shows subpopulation transition dynamics and predictive contribution of each population under chemotherapy pressure or basal state, saturated to equilibrium. Arrow weights denote prevalence of conversion. Loops indicate propensity to replicate or transition. (d) Treatment-naive 231-parental cells were sorted into CD44HiCD24Hi, CD44HiCD24Lo, CD44LoCD24Hi and CD44LoCD24Lo subpopulations, which were subsequently exposed to high-dose docetaxel (100 nM) for 48 h and re-analyzed by FACS for CD44HiCD24Hi subset expressed as % of total population. ‘Basal’ denotes the change in % of CD44HiCD24Hi subset in parental cells treated with vehicle. Data are mean±s.e.m. (ANOVA analysis, N=7, #P<0.05, *P<0.05 **P<0.01 versus basal group). (e) Depletion of intrinsic CSC population with salinomycin (5 μM, 48 h) was confirmed by reduction of a CSC (CD44Hi/CD24Lo) and enrichment of a non-CSC phenotype (CD44LoCD24Hi) expressed as fold change from vehicle-treated cells (error bars indicate s.e.m., N=5, *P<0.05 **P<0.01). (f) Chemo-tolerant cells generated from parent (DTC) and salinomycin-selected (Sal-DTC) MDA-MB-231 cells were analyzed by FACS for CD44Hi/CD24Hi, and results are expressed as % of total population (Data shown are mean±s.e.m., n=8, ANOVA analysis ***P<0.001, NS, not significant). (g) Graph shows mean fluorescent intensity (MFI) from FACS analysis of CD44 and CD24 expression in MDA-MB-231-parent, -DTC, -Sal-DTC or in a population of -expanded (E)-DTC and -Sal-DTC, demonstrating a reversal to parental phenotype when the chemotolerant cells are expanded over time (Data shown are mean±s.e.m. n=5, *P<0.05 **P<0.01). (h) Graph shows cell viability of each indicated population to docetaxel or doxorubicin, quantified by MTS cytotoxicity assay as % of viability in vehicle-treated control. All data shown are mean±s.e.m. from independent replicates (ANOVA analysis, n=8, ***P<0.001 versus parent cells).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4339891&req=5

f2: Modelling the induction of CD44HiCD24Hi cells.(a) Schematic shows experimental design used to derive mathematical parameters of population dynamics. Treatment of MDA-MB-231 breast cancer cells with 25 nM docetaxel (DTX) for 24 h induces phenotype plasticity rather than providing a selection pressure. In parallel, starting cells with different permutations and combinations of CD24 and CD44 expression levels were used, and the expression of CD44CD24 was monitored over defined time points. (b) Population dynamics modelling derived from experimental data indicates temporal kinetics of breast cancer cells in distinct compartments over 5 days (CD44Lo described as non-CSC). Left panel shows dynamics of distinct phenotypes under basal conditions, right panel demonstrates population dynamics under chemotherapy pressure. (c) Schematic shows subpopulation transition dynamics and predictive contribution of each population under chemotherapy pressure or basal state, saturated to equilibrium. Arrow weights denote prevalence of conversion. Loops indicate propensity to replicate or transition. (d) Treatment-naive 231-parental cells were sorted into CD44HiCD24Hi, CD44HiCD24Lo, CD44LoCD24Hi and CD44LoCD24Lo subpopulations, which were subsequently exposed to high-dose docetaxel (100 nM) for 48 h and re-analyzed by FACS for CD44HiCD24Hi subset expressed as % of total population. ‘Basal’ denotes the change in % of CD44HiCD24Hi subset in parental cells treated with vehicle. Data are mean±s.e.m. (ANOVA analysis, N=7, #P<0.05, *P<0.05 **P<0.01 versus basal group). (e) Depletion of intrinsic CSC population with salinomycin (5 μM, 48 h) was confirmed by reduction of a CSC (CD44Hi/CD24Lo) and enrichment of a non-CSC phenotype (CD44LoCD24Hi) expressed as fold change from vehicle-treated cells (error bars indicate s.e.m., N=5, *P<0.05 **P<0.01). (f) Chemo-tolerant cells generated from parent (DTC) and salinomycin-selected (Sal-DTC) MDA-MB-231 cells were analyzed by FACS for CD44Hi/CD24Hi, and results are expressed as % of total population (Data shown are mean±s.e.m., n=8, ANOVA analysis ***P<0.001, NS, not significant). (g) Graph shows mean fluorescent intensity (MFI) from FACS analysis of CD44 and CD24 expression in MDA-MB-231-parent, -DTC, -Sal-DTC or in a population of -expanded (E)-DTC and -Sal-DTC, demonstrating a reversal to parental phenotype when the chemotolerant cells are expanded over time (Data shown are mean±s.e.m. n=5, *P<0.05 **P<0.01). (h) Graph shows cell viability of each indicated population to docetaxel or doxorubicin, quantified by MTS cytotoxicity assay as % of viability in vehicle-treated control. All data shown are mean±s.e.m. from independent replicates (ANOVA analysis, n=8, ***P<0.001 versus parent cells).

Mentions: To theoretically test the drug-induced phenotypic cell state transition versus clonal selection, we developed a phenotype switching model consisting of three cellular compartments, describing the population dynamics of CSCs (CD44HiCD24Lo), the induced cells (CD44HiCD24Hi) and non-stem cells (CD44LoCD24Hi and CD44LoCD24Lo). Experimental data for the population dynamics were obtained from FACS data describing the re-equilibration kinetics of both the parental cells as well as the DTC, generated using the experimental design shown in Fig. 2a. The obtained parameter sets for the cases of the parental population and the DTC populations are summarized in Supplementary Table 2, using the methods described in detail in the Supplementary Information. In addition, Fig. 2b depicts the curves describing the time-evolution of the system composition from an arbitrary steady state, and highlights the system dynamics as it reaches equilibrium. In both cases, the model was able to fit well to the experimental data, implying that the model is versatile enough to describe the system dynamics of both treatment naive and post-chemotherapy cases, although, given the phenomenological nature of the model, we note that the derived parameter sets are useful only in a comparative sense, and are not necessarily precisely representative of the underlying biology for individual cases. Interestingly, the parameter values for either system was found to be quantitatively distinct, giving rise to different system saturations in equilibrium, that is, after induction of chemotherapy, there is a deterministic shift in the parameters governing the growth and switching rates of the subpopulations of cells, such that different steady states are observed. The model predicted that within the DTC population, the rates of proliferation of CSCs and induced CD44HiCD24Hi cells are significantly increased, whereas the rate of proliferation of non-stem cells decreases to a negligible value. The rate of transition from stem to non-stem cells remains the same in both environments, but the rate of transformation directly from non-stem to stem cells does not occur to a great degree in the chemoresistant cells. In addition, in the DTC, we observe high rates of transition between the induced CD44HiCD24Hi cells and non-stem cell compartments in both directions, indicating high inter-conversion (with no net direction), whereas in the case of the parental cells, the transition rate between the CD44HiCD24Hi cells and non-stem cells is highly skewed in the direction of the former, predominantly switching into the latter and not in the reverse direction. Finally, in the parental population, the CD44HiCD24Hi cells and CSCs are able to transition between each other. In contrast, in the DTCs, CSCs do not appear to transition into CD44HiCD24Hi cells, which are however able to transition into CSCs (Fig. 2c).


Temporally sequenced anticancer drugs overcome adaptive resistance by targeting a vulnerable chemotherapy-induced phenotypic transition.

Goldman A, Majumder B, Dhawan A, Ravi S, Goldman D, Kohandel M, Majumder PK, Sengupta S - Nat Commun (2015)

Modelling the induction of CD44HiCD24Hi cells.(a) Schematic shows experimental design used to derive mathematical parameters of population dynamics. Treatment of MDA-MB-231 breast cancer cells with 25 nM docetaxel (DTX) for 24 h induces phenotype plasticity rather than providing a selection pressure. In parallel, starting cells with different permutations and combinations of CD24 and CD44 expression levels were used, and the expression of CD44CD24 was monitored over defined time points. (b) Population dynamics modelling derived from experimental data indicates temporal kinetics of breast cancer cells in distinct compartments over 5 days (CD44Lo described as non-CSC). Left panel shows dynamics of distinct phenotypes under basal conditions, right panel demonstrates population dynamics under chemotherapy pressure. (c) Schematic shows subpopulation transition dynamics and predictive contribution of each population under chemotherapy pressure or basal state, saturated to equilibrium. Arrow weights denote prevalence of conversion. Loops indicate propensity to replicate or transition. (d) Treatment-naive 231-parental cells were sorted into CD44HiCD24Hi, CD44HiCD24Lo, CD44LoCD24Hi and CD44LoCD24Lo subpopulations, which were subsequently exposed to high-dose docetaxel (100 nM) for 48 h and re-analyzed by FACS for CD44HiCD24Hi subset expressed as % of total population. ‘Basal’ denotes the change in % of CD44HiCD24Hi subset in parental cells treated with vehicle. Data are mean±s.e.m. (ANOVA analysis, N=7, #P<0.05, *P<0.05 **P<0.01 versus basal group). (e) Depletion of intrinsic CSC population with salinomycin (5 μM, 48 h) was confirmed by reduction of a CSC (CD44Hi/CD24Lo) and enrichment of a non-CSC phenotype (CD44LoCD24Hi) expressed as fold change from vehicle-treated cells (error bars indicate s.e.m., N=5, *P<0.05 **P<0.01). (f) Chemo-tolerant cells generated from parent (DTC) and salinomycin-selected (Sal-DTC) MDA-MB-231 cells were analyzed by FACS for CD44Hi/CD24Hi, and results are expressed as % of total population (Data shown are mean±s.e.m., n=8, ANOVA analysis ***P<0.001, NS, not significant). (g) Graph shows mean fluorescent intensity (MFI) from FACS analysis of CD44 and CD24 expression in MDA-MB-231-parent, -DTC, -Sal-DTC or in a population of -expanded (E)-DTC and -Sal-DTC, demonstrating a reversal to parental phenotype when the chemotolerant cells are expanded over time (Data shown are mean±s.e.m. n=5, *P<0.05 **P<0.01). (h) Graph shows cell viability of each indicated population to docetaxel or doxorubicin, quantified by MTS cytotoxicity assay as % of viability in vehicle-treated control. All data shown are mean±s.e.m. from independent replicates (ANOVA analysis, n=8, ***P<0.001 versus parent cells).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4339891&req=5

f2: Modelling the induction of CD44HiCD24Hi cells.(a) Schematic shows experimental design used to derive mathematical parameters of population dynamics. Treatment of MDA-MB-231 breast cancer cells with 25 nM docetaxel (DTX) for 24 h induces phenotype plasticity rather than providing a selection pressure. In parallel, starting cells with different permutations and combinations of CD24 and CD44 expression levels were used, and the expression of CD44CD24 was monitored over defined time points. (b) Population dynamics modelling derived from experimental data indicates temporal kinetics of breast cancer cells in distinct compartments over 5 days (CD44Lo described as non-CSC). Left panel shows dynamics of distinct phenotypes under basal conditions, right panel demonstrates population dynamics under chemotherapy pressure. (c) Schematic shows subpopulation transition dynamics and predictive contribution of each population under chemotherapy pressure or basal state, saturated to equilibrium. Arrow weights denote prevalence of conversion. Loops indicate propensity to replicate or transition. (d) Treatment-naive 231-parental cells were sorted into CD44HiCD24Hi, CD44HiCD24Lo, CD44LoCD24Hi and CD44LoCD24Lo subpopulations, which were subsequently exposed to high-dose docetaxel (100 nM) for 48 h and re-analyzed by FACS for CD44HiCD24Hi subset expressed as % of total population. ‘Basal’ denotes the change in % of CD44HiCD24Hi subset in parental cells treated with vehicle. Data are mean±s.e.m. (ANOVA analysis, N=7, #P<0.05, *P<0.05 **P<0.01 versus basal group). (e) Depletion of intrinsic CSC population with salinomycin (5 μM, 48 h) was confirmed by reduction of a CSC (CD44Hi/CD24Lo) and enrichment of a non-CSC phenotype (CD44LoCD24Hi) expressed as fold change from vehicle-treated cells (error bars indicate s.e.m., N=5, *P<0.05 **P<0.01). (f) Chemo-tolerant cells generated from parent (DTC) and salinomycin-selected (Sal-DTC) MDA-MB-231 cells were analyzed by FACS for CD44Hi/CD24Hi, and results are expressed as % of total population (Data shown are mean±s.e.m., n=8, ANOVA analysis ***P<0.001, NS, not significant). (g) Graph shows mean fluorescent intensity (MFI) from FACS analysis of CD44 and CD24 expression in MDA-MB-231-parent, -DTC, -Sal-DTC or in a population of -expanded (E)-DTC and -Sal-DTC, demonstrating a reversal to parental phenotype when the chemotolerant cells are expanded over time (Data shown are mean±s.e.m. n=5, *P<0.05 **P<0.01). (h) Graph shows cell viability of each indicated population to docetaxel or doxorubicin, quantified by MTS cytotoxicity assay as % of viability in vehicle-treated control. All data shown are mean±s.e.m. from independent replicates (ANOVA analysis, n=8, ***P<0.001 versus parent cells).
Mentions: To theoretically test the drug-induced phenotypic cell state transition versus clonal selection, we developed a phenotype switching model consisting of three cellular compartments, describing the population dynamics of CSCs (CD44HiCD24Lo), the induced cells (CD44HiCD24Hi) and non-stem cells (CD44LoCD24Hi and CD44LoCD24Lo). Experimental data for the population dynamics were obtained from FACS data describing the re-equilibration kinetics of both the parental cells as well as the DTC, generated using the experimental design shown in Fig. 2a. The obtained parameter sets for the cases of the parental population and the DTC populations are summarized in Supplementary Table 2, using the methods described in detail in the Supplementary Information. In addition, Fig. 2b depicts the curves describing the time-evolution of the system composition from an arbitrary steady state, and highlights the system dynamics as it reaches equilibrium. In both cases, the model was able to fit well to the experimental data, implying that the model is versatile enough to describe the system dynamics of both treatment naive and post-chemotherapy cases, although, given the phenomenological nature of the model, we note that the derived parameter sets are useful only in a comparative sense, and are not necessarily precisely representative of the underlying biology for individual cases. Interestingly, the parameter values for either system was found to be quantitatively distinct, giving rise to different system saturations in equilibrium, that is, after induction of chemotherapy, there is a deterministic shift in the parameters governing the growth and switching rates of the subpopulations of cells, such that different steady states are observed. The model predicted that within the DTC population, the rates of proliferation of CSCs and induced CD44HiCD24Hi cells are significantly increased, whereas the rate of proliferation of non-stem cells decreases to a negligible value. The rate of transition from stem to non-stem cells remains the same in both environments, but the rate of transformation directly from non-stem to stem cells does not occur to a great degree in the chemoresistant cells. In addition, in the DTC, we observe high rates of transition between the induced CD44HiCD24Hi cells and non-stem cell compartments in both directions, indicating high inter-conversion (with no net direction), whereas in the case of the parental cells, the transition rate between the CD44HiCD24Hi cells and non-stem cells is highly skewed in the direction of the former, predominantly switching into the latter and not in the reverse direction. Finally, in the parental population, the CD44HiCD24Hi cells and CSCs are able to transition between each other. In contrast, in the DTCs, CSCs do not appear to transition into CD44HiCD24Hi cells, which are however able to transition into CSCs (Fig. 2c).

Bottom Line: Understanding the emerging models of adaptive resistance is key to overcoming cancer chemotherapy failure.This state is associated with a clustering of CD44 and CD24 in membrane lipid rafts, leading to the activation of Src Family Kinase (SFK)/hemopoietic cell kinase (Hck) and suppression of apoptosis.The use of pharmacological inhibitors of SFK/Hck in combination with taxanes in a temporally constrained manner, where the kinase inhibitor is administered post taxane treatment, but not when co-administered, markedly sensitizes the chemotolerant cells to the chemotherapy.

View Article: PubMed Central - PubMed

Affiliation: 1] Department of Medicine, Harvard Medical School, Boston, Massachusetts 02115, USA [2] Harvard-MIT Division of Health Sciences and Technology, Cambridge, Massachusetts 02139, USA [3] Division of Biomedical Engineering, Department of Medicine, Brigham and Women's Hospital, Boston, Massachusetts 02115, USA.

ABSTRACT
Understanding the emerging models of adaptive resistance is key to overcoming cancer chemotherapy failure. Using human breast cancer explants, in vitro cell lines, mouse in vivo studies and mathematical modelling, here we show that exposure to a taxane induces phenotypic cell state transition towards a favoured transient CD44(Hi)CD24(Hi) chemotherapy-tolerant state. This state is associated with a clustering of CD44 and CD24 in membrane lipid rafts, leading to the activation of Src Family Kinase (SFK)/hemopoietic cell kinase (Hck) and suppression of apoptosis. The use of pharmacological inhibitors of SFK/Hck in combination with taxanes in a temporally constrained manner, where the kinase inhibitor is administered post taxane treatment, but not when co-administered, markedly sensitizes the chemotolerant cells to the chemotherapy. This approach of harnessing chemotherapy-induced phenotypic cell state transition for improving antitumour outcome could emerge as a translational strategy for the management of cancer.

No MeSH data available.


Related in: MedlinePlus