Introduction

Disseminated cancer is treated with systemic chemotherapy, which is limited by toxic side effects and eventually by emergence of chemoresistance [1]. The so-called primary resistance is already present before therapy, and, in contrast, acquired resistance develops as tumor adaption during treatment cycles [2–4]. Alterations frequently associated with chemoresistance comprise changes in drug transport and metabolism, modification of drug targets, as well as an elevated threshold to initiate cell death. The genetic instability of cancer cells results in the generation of a range of heterogenous clones, which under therapeutic pressure leads to the expansion of the most chemoresistant subpopulations, thus driving tumor relapse. Aggravating the situation, resistance to one particular drug may result in diverse cellular changes that impair the efficacy of a whole range of chemotherapeutics, a phenomenon designated multidrug resistance (MDR). The majority of patients with disseminated disease undergoing systemic therapy will feature resistance and tumor progression. Additionally, primary resistance has been recognized in nearly half of all cancer patients, and early appearance of disease progression is almost invariably found in lung and ovarian cancer following platinum-based chemotherapy [5–8].

Mechanisms of drug resistance

Types of chemoresistance and drug disposition

Appearance of resistant tumor cells may follow different cellular pathways. Primary resistance is based on the initial presence of spontaneous resistant tumor cell clones that remain after deletion of the chemosensitive subpopulations and reconstitute the clinical observable lesions (Fig. 1). Whether these resistant clones have stem-like characteristics or random possession of different cellular characteristics or survive under special surroundings (niches) in a protected dormant stage is the focus of ongoing research [9]. Acquired resistance consists of stochastic events in the individual cancer cells predisposing for cell survival, generation of minimal residual disease, and clinical recurrence [1, 10, 11 ].

Fig. 1
figure 1

Schematic representation of the possible effects of chemotherapy on tumor cells. Sensitive cancer cells commit cell death and resistant cells survive through several possible mechanism: a as resistant, preexisting cancer stem cells, b as normal or cancer stem cells protected through interaction with extracellular matrix in niches, or c as cells randomly acquiring resistant phenotypes in the course of genetic instability and generation of heterogenous tumor subpopulations

Full size image

A primary mode of resistance comprises drug metabolism, covering cellular uptake, and extrusion as well as intracellular derivatization (Table 1). In general, hydrophobic compounds such as taxanes and camptothecines can pass the cell membrane freely, whereas other drugs depend on facilitated uptake by transporters or exert their activity at extracellular binding sites [10 ]. Resistance could be effected in the latter case by decreased expression or shedding of the cell surface receptor and mutations of the binding site. Alterations of nucleoside or folate transporters result in impaired uptake of nucleoside drugs, such as cytarabine and methotrexate, respectively [12–14 ].

Table 1 Most important mechanisms of chemoresistance at the cellular and tumor physiological level
Full size table

Drug transport

The so-called P-glycoprotein (P-gp) was the first member of the ATP-binding cassette (ABC) class of membrane transporters, actively extruding xenobiotics from cell interior [10 ]. As these transporters display a broad substrate specificity, they confer chemoresistance to a whole panel of unrelated drugs, a characteristic designated MDR. In contrast, atypical MDR stems from alterations in topoisomerase II, causing MDR to anthracyclines, podophyllotoxins, and other drugs [15]. Of the 48 known ABC transporters in humans, P-gp (ABCB1) transports a wide variety of hydrophobic anticancer drugs such as vinblastine, doxorubicin, vincristine, and taxol, among others, and the MDR-associated protein 1 (MRP1/ABCC1) and breast cancer resistance protein (BCRP/ABCG2) transport various other chemotherapeutics [10 ].

An important feature of drug resistance is that development of resistance to one drug can lead to resistance to other drugs [11, 16 ]. In particular, MRP1 transports negatively charged compounds and drugs that became derivatized by glutathione (GSH), glucuronic acid, or sulfate [17–19 ] , and BCRP participates in chemoresistance to topotecan, anthracyclines, and mitoxantrone [10]. Other transporters, such as MDR-associated protein 2 (MRP2) also called canalicular multispecific organic anion transporter 1 (cMOAT) or ABCC2, are expressed in normal liver tissue and affect chemoresistance indirectly by altering the elimination of chemotherapeutics [20]. Eventually, coadministration of chemotherapeutics and MDR inhibitors like verapamil, cyclosporine, and next-generation specific inhibitors was not successful in reversing clinical drug resistance due to the manifold other resistance mechanisms coexpressed [21, 22].

Recently, other families of membrane transporters, such as organic anion transport proteins (OATPs), were implicated in the uptake and extrusion of chemotherapeutics [23, 24]. OATPs are multispecific transport proteins, which means that they can transport a wide range of structurally unrelated compounds. In particular, OATPs 1B1 and 1B3 were demonstrated to be involved in paclitaxel transport in ovarian cancer cell lines [25]. OATPs has been implicated in transport of the camptothecin analogs, SN-38, flavopiridol, methotrexate, the taxanes paclitaxel and docetaxel, as well as imatinib (Gleevec®) [26].

Drug metabolism

A range of anticancer drugs exist as inactive prodrugs that have to be metabolically activated. For example, modified nucleosides such as cytarabine (AraC) require intracellular phosphorylation to yield the active triphosphate form, and the camptothecin irinotecan (CPT-11) is activated by carboxyesterases to its pharmacologic active metabolite SN-38. To decrease the efficacy of these drugs, cancer cells develop resistance through downregulation or mutation of the enzymes for metabolic activation, such as deoxycytidine kinase in the case of cytarabine and carboxyesterases for irinotecan [27–29 ]. Additionally, metabolic inactivation in the liver influences the pharmacokinetics of the active drugs [30].

Cells try to get rid of cytotoxic drugs via metabolic inactivation, employing protein binding, conjugation, and oxidation/reduction [1]. Platinum complexes are coupled to GSH and eliminated by ABC transporter-mediated efflux [31]. Members of the metallothionein family, small cysteine-rich proteins, bind and inactivate metals. The anticancer activity of resveratrol against breast cancer cells is reduced by sulfation, and the corresponding enzymes, namely sulfotransferases, are differently expressed in malignant breast tissues [32, 33]. Drugs, such as irinotecan, which is used for colon cancer treatment, are inactivated via phase I drug-metabolizing enzymes, CYP450 [34, 35]. Overexpression of CYP3A4 was found in a colon cancer cell line enriched for putative cancer stem cells [36].

Targeting of oncogenes

Instead of a relatively unspecific systemic chemotherapy, targeted therapy tries to specifically hit oncogenes that drive the growth of malignant tissues [37 ]. Most prominent examples of targeted therapy comprise imatinib directed to the BCR/ABL tyrosine kinase in leukemia, gefitinib and erlotinib against the epidermal growth factor receptor (EGFR) tyrosine kinase, trastuzumab targeting human epidermal growth factor (HER-2) in breast cancer, and crizotinib directed to the ALK kinase, among others in clinical use or under development [38–40 ]. Targeted therapies eliminate tumor cells and have fewer side effects, but under therapy, cells evade the pressure by mutating the target within a few months or years or show outgrowth of small resistant subpopulations present prior to treatment [41, 42]. Characterization of the mutations of the target proteins by sequence analysis and molecular modeling of the alterations is used to generate active second-generation targeted drugs [43, 44 ]. Additionally, cells have backup systems consisting of alternative parallel pathway supporting tumor survival and activation of alternative routes leaving the targeted approach ineffective [45]. In the case of successful tumor cell kill due to inhibition of both the original and alternative pathways, this constitutes a synthetic lethal relationship that may be exploited clinically [46]. One example is seen in breast and ovarian cancers carrying mutations in the BRCA1 and BRCA2 (breast cancer 1/2) genes, which are killed by targeting poly (ADP-ribose) polymerase (PARP) protein with specific inhibitors [47, 48].

Resistance to induction of apoptosis

Eventually, the fate of the cells is determined by its threshold to commit programmed cell death after irreparable damage to essential components. This type of chemoresistance results from downregulation/mutation of proapoptotic proteins, such as p53, Bak, Bax, SMAC/DIABOLO, and phosphatase and tensin homolog (PTEN), among others, or activation of prosurvival proteins, such as B cell lymphoma 2 (Bcl-2), inhibitors of apoptosis (IAPs), PI3K/Akt, and others [49]. Cyclosporine-binding proteins, the cyclophilin family, seems to be involved in the suppression of the apoptotic cascade in selected cancer cell types [50, 51]. Almost 50 % of tumors express mutated p53, which makes the cells less prone to initiation of apoptosis and induction of cell death after DNA damage [1]. In particular, prostate, lung, and ovarian cancer cells become highly resistant to any type of chemotherapy after failure of the first-line therapies, most likely to high resistance to induce cell death [8].

Tumor physiological mechanisms of resistance

Solid tumors exhibit irregular vascularization associated with intermittent perfusion and hypoxic conditions. As tumor cells switch to glycolysis, lactate accumulates in the interstitial space, and the resulting acidic condition in conjunction with high interstitial pressure impedes efficacy and diffusion of drugs significantly [8]. Cancer stem cells most likely survive in hypoxic and less perfused bone marrow niches in a dormant state [9]. Thus, the cellular mechanisms of chemoresistance are amplified at the tumor “organ” level.

Conclusion

Chemoresistance limits the survival of advanced cancer patients receiving chemotherapy. Pre-existing or induced resistant clones are present within the highly heterogenous tumor cell mass and trigger tumor relapses following treatment cycles solely eliminating the chemosensitive mass of cells [52, 53]. In contrast to the previous view of single or restricted cellular alterations to achieve resistance, gene expression analysis revealed the complex changes occurring in resistant variants, affecting a host of different pathways. For instance, changes in solute carrier family 31, member 1/copper transporter (SLC31A1), ABCC2 (MRP2), and copper-transporting ATPases 1/2 (ATP7A/B) modulate influx/efflux of cisplatin; mutS homolog 6 mismatch repair protein (MSH6), MutL homolog 1 DNA repair protein (MLH1), X-ray repair cross-complementing protein 1 (XRCC1), excision repair cross-complementation group 1/2 proteins (ERCC1/2), and DNA repair protein complementing XP-A cells (XPA) are responsible for increased DNA repair; myeloperoxidase (MPO), superoxide dismutase 1 (SOD1), GSH S-transferase µ1 (GSTM1), GSH S-transferase π1 (GSTP1), and metallothioneins impede the cytotoxic effects of cisplatin; and DNA polymerases POLH/B provide tolerance to DNA-directed drugs in conjunction with a host of other modified pathways [54]. Therefore, successful reversal of chemoresistance would require detailed knowledge of the individual causes of its molecular basis and specific targeting of critical components contributing to drug inactivation. Whether elimination of a small cancer stem cell population feeding the tumor will prove to be effective remains to be investigated [9, 55, 56].

Conflict of interest

The authors declare to have no conflict of interest.