Figure 4 - uploaded by Graziella Bernocchi
Content may be subject to copyright.
PMCA1 immunocytochemistry in the DG granule cell layer. Histograms show the OD values and the significance of differences is reported (* p <0.05). At PD11 , the intensity of labelling decreases after cisPt and increases after PtAcacDMS. At PD17 , decreased immunolabelling intensity is found after cisPt, while increased intensity is present after PtAcacDMS. ml: Molecular layer; gl: Granule layer. 

PMCA1 immunocytochemistry in the DG granule cell layer. Histograms show the OD values and the significance of differences is reported (* p <0.05). At PD11 , the intensity of labelling decreases after cisPt and increases after PtAcacDMS. At PD17 , decreased immunolabelling intensity is found after cisPt, while increased intensity is present after PtAcacDMS. ml: Molecular layer; gl: Granule layer. 

Source publication
Article
Full-text available
Platinum compounds cause significant clinical neurotoxicity. Several studies highlight neurological complications especially in paediatric oncology patients with Central Nervous System (CNS) and non-CNS malignancies. To understand the toxicity mechanisms of platinum drugs at cellular and molecular levels in the immature brain, which appears more vu...

Contexts in source publication

Context 1
... the hippocampus labelling, PMCA1 was detected in granule and pyramidal cell layers, and reflected the low density typical of plasma membranes in these layers, while intense punctate staining was detected in the neuropil [98,100,112]. Data obtained here indicated the presence of PMCA1 immunolabelling in the DG granule cells and molecular layer ( Figure 4), both at PD11 and PD17; no changes in intensity of immunoreactions were detectable in the DG granule cells. Tables 1 and 2 summarize the trend of immunolabelling for CB and PMCA1 in the Purkinje and DG granule cell layers. In the Purkinje cell layer, the treatment with cisPt caused at PD11 a strongly decreased CB immunoreactivity in the soma and dendrites, which appeared atrophic in several Purkinje neurons (Figure 1; Table 1). PMCA1 on the plasma membranes as well inside the cytoplasm markedly decreased (Figure 3). The decreased immunoreactivity of PMCA1 in the ML was localized in distinct puncta, likely due to synaptic contacts especially on Purkinje dendrite ...
Context 2
... of calcium homeostasis in the Purkinje neurons was described in mutants [71,72] and under experimental conditions in animal models as well as in human patients having neurological disorders. Notably, loss of CB (and PV) produces an alteration of the spine morphology which may be considered as compensatory mechanism. CB staining decreased in spinocerebellar ataxia and downregulation might cause alterations in the calcium homeostasis finally leading to Purkinje cell death. Moreover, the pivotal role of PMCA2 in the function of the cerebellum is shown by the phenotype of the PMCA2-null mouse and the deafwaddler 2 J, which is a mouse with a spontaneous mutation in the PMCA2 gene, and a consequent null pump activity. Both mice manifested cerebellar pathology consisting in motor deficit and ataxia [113]. In the cerebellum of PMCA2-null mouse, there was a reduction in the levels of mGlur1, which plays several essential roles in processes like motor coordination and associative learning. On the other hand, the dendrite branching of Purkinje cells as revealed by MAP2 immunocytochemistry was similar in wild type and knockout mouse; in addition, alterations in total and non-phosphorylated neurofilament-heavy (NF-H) immunoreactivity were shown [113]. These perturbations suggested altered formation of synaptic contacts, which might contribute to cerebellar dysfunction in absence of PMCA2. On PD17, the strong loss of CB in the dendrite of Purkinje cells (Figure 1) was accompanied by a lower expression of PMCA1 in the soma (Figure 3). Thus, there was a down regulation of both calcium homeostasis systems. After cisPt treatment, the early injury of both the systems regulating calcium homeostasis and their persistent imbalance in the most critical phase of Purkinje cell differentiation might alter deeply the growing of dendrite tree and synaptogenesis process. These alterations have been previously shown by the labelling for ionotropic and metabotropic glutamate receptors, and for the GABA enzyme GAD65 [84,85,114]. All these alterations are responsible for degeneration we observed several years ago in 20% of the Purkinje cell population [115]. In DG granule cell layer, the treatment with cisPt caused an early (one day after injection) decrease in CB immunoreactivity (Figure 2; Table 2). The loss of CB immunoreactivity was accompanied by decreased PMCA1 immunoreactivity (Figure 4). The loss and imbalance of calcium efflux and/or calcium buffering might be a possible or concomitant cause of cisPt neurotoxicity during DG postnatal maturation. Alteration of calcium homeostasis occurred in the superficial differentiating layers of DG and was accompanied by degeneration of mitotic granule cells that, on the contrary, are located in the deep granule cell layer [116]. Disturbance of the calcium ATPase pump in the DG persevered at PD17 (Figure 4), while CB immunoreactivity followed by recovery of calcium buffering (Figure 2). The effects of PtAcacDMS on normal nerve cells differed from those of cisPt (Tables 1 and ...
Context 3
... of calcium homeostasis in the Purkinje neurons was described in mutants [71,72] and under experimental conditions in animal models as well as in human patients having neurological disorders. Notably, loss of CB (and PV) produces an alteration of the spine morphology which may be considered as compensatory mechanism. CB staining decreased in spinocerebellar ataxia and downregulation might cause alterations in the calcium homeostasis finally leading to Purkinje cell death. Moreover, the pivotal role of PMCA2 in the function of the cerebellum is shown by the phenotype of the PMCA2-null mouse and the deafwaddler 2 J, which is a mouse with a spontaneous mutation in the PMCA2 gene, and a consequent null pump activity. Both mice manifested cerebellar pathology consisting in motor deficit and ataxia [113]. In the cerebellum of PMCA2-null mouse, there was a reduction in the levels of mGlur1, which plays several essential roles in processes like motor coordination and associative learning. On the other hand, the dendrite branching of Purkinje cells as revealed by MAP2 immunocytochemistry was similar in wild type and knockout mouse; in addition, alterations in total and non-phosphorylated neurofilament-heavy (NF-H) immunoreactivity were shown [113]. These perturbations suggested altered formation of synaptic contacts, which might contribute to cerebellar dysfunction in absence of PMCA2. On PD17, the strong loss of CB in the dendrite of Purkinje cells (Figure 1) was accompanied by a lower expression of PMCA1 in the soma (Figure 3). Thus, there was a down regulation of both calcium homeostasis systems. After cisPt treatment, the early injury of both the systems regulating calcium homeostasis and their persistent imbalance in the most critical phase of Purkinje cell differentiation might alter deeply the growing of dendrite tree and synaptogenesis process. These alterations have been previously shown by the labelling for ionotropic and metabotropic glutamate receptors, and for the GABA enzyme GAD65 [84,85,114]. All these alterations are responsible for degeneration we observed several years ago in 20% of the Purkinje cell population [115]. In DG granule cell layer, the treatment with cisPt caused an early (one day after injection) decrease in CB immunoreactivity (Figure 2; Table 2). The loss of CB immunoreactivity was accompanied by decreased PMCA1 immunoreactivity (Figure 4). The loss and imbalance of calcium efflux and/or calcium buffering might be a possible or concomitant cause of cisPt neurotoxicity during DG postnatal maturation. Alteration of calcium homeostasis occurred in the superficial differentiating layers of DG and was accompanied by degeneration of mitotic granule cells that, on the contrary, are located in the deep granule cell layer [116]. Disturbance of the calcium ATPase pump in the DG persevered at PD17 (Figure 4), while CB immunoreactivity followed by recovery of calcium buffering (Figure 2). The effects of PtAcacDMS on normal nerve cells differed from those of cisPt (Tables 1 and ...
Context 4
... an early stage, at PD11, the intense immunolabelling for CB ( Figure 1) showed an almost normal dendrite branching of Purkinje cells. In particular, the reaction intensity was the same as in controls. The expression of PMCA1 (Figure 3) in the soma and on the growing dendrite changed versus control rats; the immunoreaction intensity decreased. Later, at PD17, the PMCA1 (Figure 3) immunoreactivity was significantly lowered as compared with control rats, while no significant change in the CB labelling was found (Figure 1). Therefore, PtAcacDMS treatment did not alter early the efficiency of both calcium homeostasis systems. The presence of the CB calcium buffering protein alone might guarantee the correct differentiation of Purkinje neurons [41,117] and the formation on synaptic contacts on them [118]. At PD11, as in the Purkinje cells CB expression was unchanged in the DG granule cells (Figure 2), which experienced an increase in the PMCA1 immunoreactivity (Figure 4). This indicates efficient calcium efflux from cells. At PD17, the decreased CB concentration and subsequent decreased calcium buffering was balanced by an increase in calcium efflux. Unequivocally, the balance involves a mechanism of calcium homeostasis in the DG. Multiple acute effects of a single injection of cisPt have been demonstrated during the postnatally developing cerebellum. CisPt induced morphological and molecular changes, including the early damage of proliferating and postmitotic differentiating neurons. Apoptotic cell death in the EGL with concomitant alteration of migratory process, atrophy of Purkinje cell dendrite branches and formation of altered synaptic contacts were detected [114]. Our studies on the effects of cisPt on neurotransmitter molecules connected with the brain maturation, showed that neurotransmitters such as GABA, glutamate, and monoamines affected the cisPt-induced changes of developing cerebellum architecture [84,85,115,119], and therefore in the formation of synaptic contacts. The whole cerebellar cortex presented numerous hemorrhagic foci, mainly seven days after the treatment that was carried out at postnatal day 10. Findings demonstrate that cisPt alters the blood vessels endothelium and could pass the BBB [115]. A recent report [116] has shown that cisPt alters the differentiation and maturation of synaptic contacts of some types of cells and interneurons of the rat hippocampus formation. These evidences help to explain its neurotoxicity on the developing brain. The investigations on the toxic effects of PtAcacDMS on normal tissues pointed to a reduced neurotoxicity of the new platinum complex. PtAcacDMS induces less serious changes than cisPt on fundamental events in cerebellum and hippocampus development, such as no significant apoptotic events and less injured neuroarchitecture. In particular, the balance between Bcl2/Bax proteins [117] favours the PtAcacDMS treated rats, ensuring normal features in the highly actively proliferating cells in the cerebellum without cell death and cell migration [41]. The Bcl2 family proteins, which play a key role in regulating apoptotic cell death of many cell types [120] showed that after both platinum compounds treatments, Bax (pro-apoptotic protein) expression in Purkinje cells was more intense in respect to the controls. Nevertheless, the labelling for Bcl2 (anti-apoptotic protein) in the same neurons was lowered after cisPt treatment but maintained the steady state after the PtAcacDMS ...

Citations

... Additionally, these metallic protrusions can break off during fabrication or use with cell cultures. Although platinum is widely used in biological and medical electrode applications due to its inertness and high electrical conductivity (24), studies have shown that platinum nanoparticles can cause cytotoxic and inflammatory effects in neurons both ex vivo and in vivo (25,26). Therefore, it is desirable to have uniformly layered platinum deposits without discernible protrusions that could break off. ...
Preprint
Full-text available
Microelectrode arrays are invaluable tools for investigating the electrophysiological behaviour of neuronal networks with high spatiotemporal precision. In recent years, it has become increasingly common to functionalize such electrodes with highly porous platinum to increase their effective surface area, and hence their signal-to-noise ratio. Although such functionalization significantly improves the electrochemical performance of the electrodes, the impact of various electrode morphologies on biocompatibility and electrophysiological performance in cell cultures remains poorly understood. In this study, we introduce reproducible protocols for depositing highly porous platinum with varying morphologies on microelectrodes designed for neural cell cultures. We also evaluate the impact of morphology and electrode size on the signal-to-noise ratio in recordings from rat cortical neurons cultured on these electrodes. Our results indicate that electrodes with a uniform layer of highly nanoporous platinum offer the best trade-off between biocompatibility, electrochemical, and electrophysiological performance. While more microporous electrodes exhibited lower impedance, nanoporous electrodes detected higher extracellular signal amplitudes from neurons, suggesting reduced distance between perisomatic neuronal areas and the electrodes. Additionally, these nanoporous electrodes showed fewer thickness variations at their edges compared to the more porous electrodes. Such edges can be mechanically broken off during cell culturing and contribute to long-term cytotoxic effects, which is highly undesirable. We hope this work will contribute to better standardization in creating and utilizing nanoporous platinum microelectrodes for neuroscience applications. Improving the accessibility and reproducibility of this technology is crucial for enhancing the quality of electrophysiological data and advancing our understanding of neuronal network function and dysfunction.
... For instance, CDDP induced a sustained elevation in [Ca 2+ ] i , which reflects into enhanced [Ca 2+ ] ER , mitochondrial depolarization, and apoptosis, in multiple types of tumor cell lineages, including neuroblastoma , ovarian, and cervical cancer [27] (Shen et al. 2016). Nevertheless, the mechanism(s) whereby CDDP elevates the [Ca 2+ ] i remains to be elucidated, although this drug may severely rewire the intracellular Ca 2+ toolkit in both cancer Schrodl et al. 2009) and non-cancer (Piccolini et al. 2013;Bernocchi et al. 2015) cells. Interestingly, PtAcacDMS was also recently shown to increase the [Ca 2+ ] i by inhibiting Plasma Membrane Ca 2+ -ATPase (PMCA) activity in breast cancer MCF-7 cells, while it did not affect Orai1mediated Ca 2+ entry (Muscella et al. 2011). ...
Article
Full-text available
Cisplatin (CDDP) is one of the most effective chemotherapeutic agents, used for the treatment of diverse tumors, including neuroblastoma and glioblastoma. CDDP induces cell death through different apoptotic pathways. Despite its clinical benefits, CDDP causes several side effects and drug resistance.[Pt(O,O′-acac)(γ-acac)(DMS)], namely PtAcacDMS, a new platinum(II) complex containing two acetylacetonate (acac) and a dimethylsulphide (DMS) in the coordination sphere of metal, has been recently synthesized and showed 100 times higher cytotoxicity than CDDP. Additionally, PtAcacDMS was associated to a decreased neurotoxicity in developing rat central nervous system, also displaying great antitumor and antiangiogenic activity both in vivo and in vitro. Thus, based on the knowledge that several chemotherapeutics induce cancer cell death through an aberrant increase in [Ca2+]i, in the present in vitro study we compared CDDP and PtAcacDMS effects on apoptosis and intracellular Ca2+ dynamics in human glioblastoma T98G cells, applying a battery of complementary techniques, i.e., flow cytometry, immunocytochemistry, electron microscopy, Western blotting, qRT-PCR, and epifluorescent Ca2+ imaging. The results confirmed that (i) platinum compounds may induce cell death through an aberrant increase in [Ca2+]i and (ii) PtAcacDMS exerted stronger cytotoxic effect than CDDP, associated to a larger increase in resting [Ca2+]i. These findings corroborate the use of PtAcacDMS as a promising approach to improve Pt-based chemotherapy against gliomas, either by inducing a chemosensitization or reducing chemoresistance in cell lineages resilient to CDDP treatment.
... As mitochondria are important for neuronal Ca 2+ dynamics we suggest that astrocytic mitochondrial transfer is key to normalizing the neuronal mitochondrial network. It remains to be determined whether the normalization of Ca 2+ dynamics results from a direct role of the donated mitochondria in calcium buffering or a secondary effect on other key regulators of calcium homeostatis such as the endoplasmic reticulum (ER), Golgi apparatus, and peroxisomes, or the functioning of ion channels and pumps [55][56][57][58][59]. However, our findings do show that transfer of mitochondria normalizes neuronal Ca 2+ dynamics in neurons damaged by cisplatin. ...
Article
Full-text available
Neurodegenerative disorders, including chemotherapy-induced cognitive impairment, are associated with neuronal mitochondrial dysfunction. Cisplatin, a commonly used chemotherapeutic, induces neuronal mitochondrial dysfunction in vivo and in vitro. Astrocytes are key players in supporting neuronal development, synaptogenesis, axonal growth, metabolism and, potentially mitochondrial health. We tested the hypothesis that astrocytes transfer healthy mitochondria to neurons after cisplatin treatment to restore neuronal health. We used an in vitro system in which astrocytes containing mito-mCherry-labeled mitochondria were co-cultured with primary cortical neurons damaged by cisplatin. Culture of primary cortical neurons with cisplatin reduced neuronal survival and depolarized neuronal mitochondrial membrane potential. Cisplatin induced abnormalities in neuronal calcium dynamics that were characterized by increased resting calcium levels, reduced calcium responses to stimulation with KCl, and slower calcium clearance. The same dose of cisplatin that caused neuronal damage did not affect astrocyte survival or astrocytic mitochondrial respiration. Co-culture of cisplatin-treated neurons with astrocytes increased neuronal survival, restored neuronal mitochondrial membrane potential, and normalized neuronal calcium dynamics especially in neurons that had received mitochondria from astrocytes which underlines the importance of mitochondrial transfer. These beneficial effects of astrocytes were associated with transfer of mitochondria from astrocytes to cisplatin-treated neurons. We show that siRNA-mediated knockdown of the Rho-GTPase Miro-1 in astrocytes reduced mitochondrial transfer from astrocytes to neurons and prevented the normalization of neuronal calcium dynamics. In conclusion, we showed that transfer of mitochondria from astrocytes to neurons rescues neurons from the damage induced by cisplatin treatment. Astrocytes are far more resistant to cisplatin than cortical neurons. We propose that transfer of functional mitochondria from astrocytes to neurons is an important repair mechanism to protect the vulnerable cortical neurons against the toxic effects of cisplatin.
... Neurotoxicity associated with the treatment of platinum compounds has been thoroughly studied for the wide use of these substances as chemotherapeutic agents [182][183][184][185]. Here, we focus briefly on experimental works carried out by our research group at the Department of Biology and Biotechnology (formerly Animal Biology) of the University of Pavia from 1986 to the present, which have revealed alterations in the immunopositivity for CBPs in developing and adult rats exposed to two platinum compounds, cisplatin (CisPt) and [Pt(O,O -acac)(γ-acac)(DMS)] (PtAcacDMS) ( Table 2). These studies have highlighted a dynamic decrease in two buffer CBPs, CB and PV, following treatment with both compounds in the hippocampus and cerebellum, and also detected a less severe damage by PtAcacDMS compared to CisPt, demonstrating that CBPs can help discriminate between compounds by providing a reliable estimation of neuronal damage [19,39,[186][187][188]. Interestingly, contrary to what was observed in hibernating animals, there is no correlation between changes in CBPs and cytoskeletal detection, as CB immunolabelling decreases while NF-H phosphorylation increases following CisPt treatment [18,188], and is likely responsible for the marked morphological signs of degeneration observed in Purkinje cells of the developing rat cerebellum [189]. ...
Article
Full-text available
Calcium-binding proteins (CBPs) can influence and react to Ca2+ transients and modulate the activity of proteins involved in both maintaining homeostatic conditions and protecting cells in harsh environmental conditions. Hibernation is a strategy that evolved in vertebrate and invertebrate species to survive in cold environments; it relies on molecular, cellular, and behavioral adaptations guided by the neuroendocrine system that together ensure unmatched tolerance to hypothermia, hypometabolism, and hypoxia. Therefore, hibernation is a useful model to study molecular neuroprotective adaptations to extreme conditions, and can reveal useful applications to human pathological conditions. In this review, we describe the known changes in Ca2+-signaling and the detection and activity of CBPs in the nervous system of vertebrate and invertebrate models during hibernation, focusing on cytosolic Ca2+ buffers and calmodulin. Then, we discuss these findings in the context of the neuroprotective and neural plasticity mechanisms in the central nervous system: in particular, those associated with cytoskeletal proteins. Finally, we compare the expression of CBPs in the hibernating nervous system with two different conditions of neurodegeneration, i.e., platinum-induced neurotoxicity and Alzheimer’s disease, to highlight the similarities and differences and demonstrate the potential of hibernation to shed light into part of the molecular mechanisms behind neurodegenerative diseases.
... Nimodipine was able to protect the dorsal root ganglion neurons from the toxic effects of CDDP [28]. However, CGCs were different from the dorsal root ganglion neurons because the cerebellum is one of the regions of the central nervous system that develops postnatally [29]. CGC progenitors still migrate into the internal granular layer of the cerebellum even during early postnatal period [30]. ...
Article
Full-text available
Objective: The aim of the present study was to investigate the role of ionic homeostasis in cisplatin (cisdiamminedichloroplatinum (II), CDDP)-induced neurotoxicity. CDDP is a severely neurotoxic antineoplastic agent that causes neuronal excitotoxicity. According to some studies, calcium influx increases, whereas potassium efflux decreases neuronal death. Nimodipine and glibenclamide were used to analyze the role of ionic flows in CDDP-induced neurotoxicity in rat primary cerebellar granule cell (CGC) culture. Materials and methods: CGC culture was prepared from the cerebella of Sprague Dawley 5-day-old pups. The submaximal concentration of CDDP was determined and then given with 1, 10, or 50 µM of drugs into culture. Neurotoxicity was investigated using the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, a tetrazole) assay. One-way analysis of variance, Kruskal-Wallis H test, and Tukey test were applied for statistical analysis. Results: CDDP induced neurotoxicity in a concentration-dependent manner. Neither nimodipine nor glibenclamide was able to protect CGCs against CDDP neurotoxicity. Conclusion: By blocking L-type voltage-gated calcium channels, nimodipine did not prevent CDDP neurotoxicity in CGCs. Ca2+ influx via these channels seemed to be insufficient to cause a change in CDDP-induced neurotoxicity. Similarly, glibenclamide failed to prevent CDDP neurotoxicity. Further studies are needed to elucidate the mechanisms of these preliminary results.
... Bernocchi and coworkers studied the effects of cisplatin on the immature brain, which appears to be more vulnerable to injury than the adult brain. Changes in the intracellular calcium homeostasis within the central nervous system architecture after cisplatin exposure demonstrates that the equilibrium and synergy between calcium proteins to limit neuroarchitecture damages [4] is essential. 5. ...
Article
Full-text available
This Special Issue on “Toxicities of Therapeutic Agents Used in Medicine” reports on some peculiar cases of toxicities related to widely and commonly employed drugs.[...]
Article
Introduction The main toxicity of cisplatin is nephrotoxicity, but more and more studies have highlighted and unveiled the mechanisms of cisplatin toxicity, and the neurotoxicity has been standing out. Objective We aimed to bring together the main studies that highlight the neurotoxicity of cisplatin in the treatment of cancer patients. Methods We performed a literature review using the keywords “Neurotoxicity”, “Cisplatin”, “Oncology Therapy”, and “Chemotherapy” in the SciELO, PubMed, Sciencedirect, MEDLINE, Scifinder, and CAplus databases. Results We selected 60 articles published between 1983 and 2021 that report the mechanisms of cisplatin toxicity or which provide clinical data on the neurotoxicity profile of cisplatin as monotherapy and as a combination therapy, highlighting that one of the main neurotoxicity of cisplatin is in the development of peripheral neuropathy. Conclusion Cisplatin is neurotoxic and can induce the development of peripheral neuropathy and the combination with neurotoxic drugs such as paclitaxel and vincristine only contributes to the increase in neurological toxicity. Thus, we emphasize the importance of evaluating the neurotoxicity of cisplatin, especially in patients who use protocols that contain other antineoplastic agents that are also neurotoxic.
Article
Cisplatin (cisPt), among the best known components of multi-drug front-line therapies used for the treatments of solid tumors, such as the childhood neuroblastoma, acts through DNA linking. Nevertheless, the cisPt effectiveness is compromised by the onset of severe side effects, including neurotoxicity that results in neurodegeneration, cell death, and drug-resistance. In the field of experimental oncology, aimed at overcoming cytotoxicity and chemoresistance, great efforts are devoted to the synthesis of new platinum-based drugs, such as [Pt(O,O′-acac)(γ-acac)(DMS)] (PtAcacDMS), which shows a specific reactivity with sulfur residues of enzymes involved in apoptosis. Autophagy, an evolutionary conserved degradation pathway for recycling of cytoplasmic components, represents one of the mechanisms adopted by cancer cells which contribute to drug-resistance. In the present study, standard acute (48 h-exposure) and long-term effects (7 day-recovery after treatment or 7 day-recovery followed by reseeding and 96 h-growth), of cisPt and PtAcacDMS (40 and 10 μM, respectively) were investigated in vitro employing rat B50 neuroblastoma as a cancer model. Using fluorescence and electron microscopy, as well as biochemical techniques, our data highlight a key role of the autophagic process in B50 cells. Specifically, long-term effects caused by cisPt lead to inhibition of the apoptotic process and paralleled by the activation of autophagy, thus evidencing that autophagy has a protective role after cisPt exposure, allowing cells to survive. Whereas, long-term effects produced by PtAcacDMS lead toward both apoptosis and autophagy activation. In conclusion, autophagy may represents an alternative cell death pathway, circumventing drug-resistance strategies employed by cancer cells to survive chemoterapy.