Introduction
Resistance of cancer cells
Resistance to chemotherapy is the most important reason for treatment failure in
cancer patients. Tumours may be intrinsically drug-resistant or develop
resistance to chemotherapy during treatment [1]. It
is well known that cancer cells are able to resist various cytotoxic agents
because they possess a set of anti-cell death mechanisms that counteract
chemotherapeutic responses. These protective mechanisms include the constitutive
activation of the phosphatidylinositide 3-kinase (PI3-K)/Akt and the nuclear
factor-kappa B (NF-κB) signalling pathways, which are interlinked [2], [3]. Treatment can
lead to the death of most tumour cells (drug-sensitive), but some cells
(drug-resistant) survive and grow. Cancer has the ability to become resistant to
many different types of drugs. Increased efflux of drug, enhanced repair and
increased tolerance to DNA damage, high anti-apoptotic potential, decreased
permeability and enzymatic deactivation allow cancer cells to survive
chemotherapy. Acquired resistance is a particular problem, as tumours do not
only become resistant to the drugs that are originally used to treat them but
may also become cross-resistant to other drugs with different mechanisms of
action.
A major obstacle to the effective treatment of cancer is the multidrug resistance
(MDR) phenomenon exhibited by many cancers [4], [5]. MDR can be an intrinsic
characteristic of malignant cells or acquired during drug therapy [5]. The most prominent mechanisms mediating MDR to
anti-neoplastic agents are (a) over-expression of members of three ATP-binding
cassette (ABC) transporter sub-families, ABCB, ABCC, and ABCG, (b) lung
resistance-related protein (LRP, identified as the major vault protein (MVP)),
and (c) loss of genes, such as p53, that control DNA integrity [5], [6], [7]. Thus, targeting or circumventing these proteinsʼ
activities would have a major impact on cancer chemotherapy and cancer patientsʼ
survival [8]. Although many efforts to overcome MDR
have been made, no outstanding breakthroughs have been achieved [8]. Consequently, there remains an urgent need to
identify new biological targets associated with cancer cell chemoresistance as
well as novel anti-cancer agents, with the goal of overcoming resistance to
chemotherapy. Previous unsuccessful approaches indicate the need to target
simultaneously multiple MDR-related targets and thus disable the cancer cellsʼ
ability to deploy escape strategies. Accordingly, a completely new way of
attacking resistant cancer cells might rely on targeting the sodium/potassium
pump (Na+/K+-ATPase; NaK) with its highly specific
ligands, i.e., cardiotonic steroids (CS).
The sodium/potassium pump (Na+/K+-ATPase; NaK)
NaK is an integral membrane protein composed of catalytic α and regulatory
β subunits; it is responsible for translocating sodium and potassium
ions across the cell membrane utilising ATP as the driving force [9]. Although the transport function of the
Na+/K+-ATPase has been investigated extensively in the
past, during the last decade multiple lines of evidence have suggested a number
of other functions for the sodium pump, revealing NaK as (i) a multifunctional
protein with key roles in the formation and maintenance of adhesion complexes,
induction of epithelial cell tight junctions and polarity, cell adhesion,
motility, and actin dynamics [10], [11], [12], [13], [14], [15], [16], [17], [18], (ii) a
signalling protein [19], [20], [21], [22], [23], [24], [25], [26], and (iii) a valuable novel target in anti-cancer therapy because
its aberrant expression and activity are implicated in the development and
progression of a growing number of cancers [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38].
In addition to the growing number of scientific publications, a number of
inventions (recently reviewed in [39]) have also
emphasised the potential usefulness of considering NaK expression for future
anti-cancer therapy by using it as a diagnostic and prognostic tool, as a
biomarker of a therapeutic response in cancer chemotherapy with CS, and as a
valuable new target. A recent, in-depth analysis of patent literature [39] revealed a large increase in the number of
inventions focusing on new NaK inhibitors and ligands designed or selected as
potential anti-cancer agents.
Cardiotonic steroids
The CS, which include cardenolides and bufadienolides ([Fig. 1]), are compounds that are able to bind to the extracellular
surface of the NaK [27] and are its natural
ligands. The best-known naturally occurring CS are digoxin, digitoxin, ouabain,
and oleandrin as cardenolides as well as bufalin, hellebrin, and marinobufagenin
as bufadienolides. The CS have long been used as positive inotropic agents in
the treatment of congestive heart failure [40].
Retrospective epidemiological studies conducted during the late 20th century
revealed some intriguing results: very few patients that underwent CS treatment
for heart problems died from cancer [41]. Over the
last 20 years, interest in developing the CS as anti-cancer agents has grown
progressively. CS were identified to be among the most potent inhibitors (out of
9000 screened chemicals) of the prostate cancer target genes investigated [42]. Furthermore, in a large investigation that
searched for new natural, cytotoxic anti-cancer compounds, Lindholm et al. [43] screened extracts from 100 different plants and
obtained seven plants with strong evidence of anti-tumour potential, among which
were three CS-enriched plants, Digitalis lanata, Digitalis purpurea, and
Helleborus cyclophyllus. By binding to the sodium pump, CS elicit
marked effects on cancer cell behaviour, and a number of studies have emphasised
their potential use in oncology [27], [37], [44], [45]. Some recent reviews [27], [37], [45], [46], [47], [48] summarise the anti-tumour
properties of this class of compounds as well as their multiple mechanisms of
action (briefly summarized in [Fig. 2]). We
recently reviewed the scientific literature to perform an in-depth
structure-activity relationship (SAR) analysis with respect to cardenolide-
versus bufadienolide-mediated anti-cancer effects [47]. In that review, we described the SAR of the CS based on a
molecular model of the NaK pump bound to ouabain [47]. After an analysis of the anti-cancer potency of the most
representative CS, we determined the key structural features that lead to
powerful cytotoxic agents and those that are deleterious for anti-tumour
activity.
Fig. 1 Classification and chemical structures of cardiotonic steroids.
Cardiotonic steroids are compounds presenting a steroid nucleus with a
lactone moiety at position 17. The aglycone moiety is composed of the
steroid nucleus and the R group (lactone ring) at position 17 that defines
the class of cardiotonic steroid: the cardenolides (with an unsaturated
butyrolactone ring) and the bufadienolides (with an α-pyrone ring).
The steroid nucleus has a unique set of fused ring systems that makes the
aglycone moiety structurally distinct from the other more common steroid
ring systems. The steroidal skeleton can be substituted at position 3 by the
third structural component, a sugar moiety (glycoside), leading to the
chemical classification of sub-families as glycosylated cardenolides or
glycosylated bufadienolides (depending on the lactone moiety). Up to 4 sugar
molecules may be present in cardiac glycosides; attached in many via the
3β-OH group.
Fig. 2 Summary of postulated mechanisms of CS-mediated anti-cancer
activity. Summary of the multiplicity of suggested molecular targets for the
action of most studied CS in human cancer cells. For more details see [19], [20], [26], [27], [37], [45], [46], [47], [48], [49], [93].
It is interesting that the CS tested in vitro induced potent
anti-proliferative effects in all of the human cancer cell lines examined;
consequently, there is no particularly resistant human cancer type. Indeed, the
cancer cell lines in the NCI 60 panel
(http://dtp.nci.nih.gov/dtpstandard/cancerscreeningdata/index.jsp) display
similar sensitivities to the CS tested (ouabain, digitoxin, and hellebrin), and
this effect was further confirmed with 19-hydroxy-2′′-oxovoruscharin (also known
as UNBS1450) [27], [37], [44], [49]. A growing number of reports document the ability of some CS to
circumvent cancer cell chemoresistance [50], [51], [52], [53], [54], making them
an interesting starting point for the development of new anti-chemoresistance
treatment strategies.
Fighting Resistant Cancer Cells through Na+/K+-ATPase
Targeting
Migrating cells are particularly resistant to cytotoxic agents: involvement
of the NaK β subunit in pro-cell attachment strategies
Resistance to chemotherapy is believed to cause treatment failure in more than
90 % of patients with metastatic cancer. Because metastatic cancers originate
from migrating cells, specific anti-migratory strategies should be added to
conventional radio- and/or chemotherapy.
The Na+/K+-ATPase associates with a number of signalling
molecules and with the actin cytoskeleton, forming a multiprotein complex
(recently reviewed in [17], [18]). The effect of CS, and particularly ouabain, on the adhesive
state of the cell was studied extensively, and signalling cascades involved in
the so-called P → A mechanism (pump → attachment) were deciphered by Contreras
et al. [12], [13], [14], [15], [16]. Contrerasʼ group demonstrated
that ouabain affects cell attachment through a complex signalling cascade and by
sending β-catenin to the nucleus, where it is known to act as a
transcriptional cofactor [12], [13], [14], [15], [16]. These
reports further emphasise that the interactions of CS with NaK could markedly
affect cell migration features. Furthermore, Rajasekaran et al. [10] presented evidence that NaK plays a crucial role
in E-cadherin-mediated development of epithelial polarity and suppression of
invasiveness and motility of carcinoma cells. Their results suggest that
E-cadherin-mediated cell-cell adhesion requires the function of the NaK β
subunit to induce epithelial polarisation and suppress the invasiveness and
motility of carcinoma cells. Tummala et al. [55]
revealed that reduced expression of the NaK β1 protein is associated with
oxaliplatin resistance in cancer cells and demonstrated a novel role for this
protein in sensitising the cells to oxaliplatin. Although the mechanism by which
NaK β1 increases sensitivity to oxaliplatin is not known, it is tempting
to speculate that the cell–cell adhesion function of NaK β1 might be
involved in this process. Importantly, it has been widely reported that NaK
β1 subunits are very frequently downregulated in human epithelial
cancer cells [10], [11], [28], [29], [30], [56]. The Rajasekaran group [10], [11], [28], [33], [56], [57] noted that when these cells downregulate
β1, they detach from each other as a result of a marked reduction in
cadherin expression, a process in which the Snail transcription factor plays a
major role [10], [29], [56]. Thus, downregulation of
β1 subunits seems essential for epithelial cancer cells to become
individually invasive and chemoresistant. The NaK β1 downregulation might
result from its rapid degradation in cancer cells. Yoshimura et al. [58] recently demonstrated that the α and
β subunits of NaK are assembled in the endoplasmic reticulum but are
disassembled in the plasma membrane and undergo different degradation processes,
leading to over-expression of the α subunits and faster degradation of
the β subunit. Thus, restoration of NaK β1 expression might
contribute to preventing cancer cell migration and the resulting invasion,
metastasis, and chemoresistance. Alternatively, compounds inducing NaK β1
expression might provide an interesting complement to the standard
anti-metastatic therapy. To the best of our knowledge, no such compound has been
reported.
A number of cancers display intrinsic resistance to pro-apoptotic stimuli:
targeting of NaK α subunit by CS
The malignant transformation of cells is associated with a constellation of
pro-survival mutations that increase the cellsʼ resistance to apoptosis. Because
most of the agents used in current anti-cancer therapies are pro-apoptotic
agents, agents that induce other types of cell death or act as apoptosis
sensitizers might offer better therapeutic results. Consistent with this idea,
as summarised in [Table 1], several reports [34], [59], [60], [61], [62], [63], [64], [65] have
documented the potential of CS, at least in vitro, to (i) act as
apoptosis sensitizers, (ii) act as anoïkis sensitizers, and (iii) be potent
inducers of autophagy-like cell death.
Table 1 Potential of CS to (i) act as apoptosis
sensitizers, (ii) act as anoïkis sensitizers, and (iii) be potent
inducers of autophagy-like cell death.
|
Function
|
CS
|
Mechanism
|
Reference
|
|
Apoptosis sensitizer
|
oleandrin
|
Apo2L/TRAIL-induced apoptosis via upregulation of death
receptors 4 and 5 in non-small cell lung cancer cells
|
[60]
|
|
Apoptosis sensitizer
|
oleandrin, ouabain,
digoxin
|
stimulate Ca2+ increases and apoptosis in
androgen-independent, metastatic human prostate
adenocarcinoma cells
|
[61]
|
|
Apoptosis sensitizer
|
oleandrin
|
oleandrin-mediated expression of Fas that potentiates
apoptosis
|
[62]
|
|
Apoptosis sensitizer
|
bufalin, bufotalin, gamabufotalin
|
TRAIL-sensitising agents, especially for the triple negative
breast cancer
|
[63]
|
|
Anoïkis sensitizers
|
ouabain, peruvoside, digoxin, digitoxin, strophanthidin
|
anoïkis sensitisation in anoïkis-resistant PPC-1 prostate
adenocarcinoma cells through the mitochondrial pathway of
caspase activation and by inducing hypoosmotic stress
|
[64]
|
|
Inducers of autophagy-like cell death
|
oleandrin
|
authophagic cell death of pancreatic cancer cells
|
[65]
|
|
Inducers of autophagy-like cell death
|
19-hydroxy-2′′-oxovoruscharin
|
disorganisation of the actin cytoskeleton and induction of
severe autophagic process
|
[34]
|
|
Inducers of autophagy-like cell death
|
19-hydroxy-2′′-oxovoruscharin
|
decrease of Hsp70 expression and induction of the lysosomal
membrane permeabilisation
|
[59]
|
Multidrug resistance as one of the major reasons for the failure of
anti-cancer therapy: targeting of the NaK α subunit by anti-MDR
CS
Available research data point to the divergent behaviour of CS with respect to
the induction and repression of MDR. Most known cardenolides have been reported
to antagonise the activity of several chemotherapeutic agents. Digoxin was shown
to up-regulate MDR1 mRNA, [66] and Huang et al.
[67] reported that ouabain and digitoxin
induced resistance to tubulin-dependent anti-cancer drugs such as paclitaxel,
colchicine, vincristine, and vinblastine in androgen-independent human prostate
cancer. It was suggested that these cardenolides inhibit the G2/M arrest induced
by tubulin-binding anti-cancer drugs via an indirect blockage of microtubule
function. Furthermore, a decline in the transport of these tubulin-dependent
anti-cancer drugs into the nucleus may explain the antagonistic action of these
cardenolides. Ouabain provokes reduced doxorubicin-mediated cytotoxicity in
human A549 non-small cell lung cancer (NSCLC), HT29 colon cancer, and U1
melanoma by decreasing doxorubicin-induced topoisomerase-mediated DNA strand
breakage [68]. This response indicates that altered
ionic gradients are a potential cause of resistance to drugs that use
topoisomerase II as a target [68]. Additionally,
Ahmed et al. [69] reported that cisplatin
accumulation in oral squamous carcinoma cells is regulated by NaK and thus, its
inhibition markedly reduced intra-cellular cisplatin accumulation. In contrast,
the reports on less thoroughly investigated CS indicate the potential usefulness
of these CS to combat chemoresistant cancers [52], [53], [54], [70]. Bufalin has been reported to
reverse multi-drug resistance in some human leukemia MDR cells. Indeed, Efferth
et al. [70] reported that bufalin caused a
significant increase in the accumulation of daunorubicin in CEM/VLB100 and
CEM/E1000 cells. Moreover, some cardenolides from Calotropis procera,
Pergularia tomentosa, and Nerium oleander can overcome MDR [52], [53], [54]. Interestingly, some of these compounds can
overcome MDR from multiple origins. Indeed, we previously reported that
19-hydroxy-2′′oxovoruscharin-mediated potent anti-cancer activity is not limited
by the intrinsic MDR conferred by the over-expression of key drug-transporter
proteins acquired as a result of exposure to a range of chemotherapeutic agents
or loss of wild-type p53 [52]. This was confirmed
in human cancer cell lines of different origin including HeLa-derived KB
carcinoma, MDA-MB-231 breast cancer, GLC4 small cell lung cancer, SW-1573 and
A549 NSCLC, S1 and HCT116 colon cancer, HL-60 leukaemia, and adenovirus
transformed HEK293 cells; these were selected given their resistance to various
chemotherapeutic agents (adriamycin, vincristine, cisplatin, oxaliplatin,
mitoxantrone, hydroxyurea) and/or their over-expression of different MDR-related
proteins (ABCB1, ABCC1 (MRP1), ABCC2, ABCC10, ABCG2 (BCRP), and MVP). In
general, the sensitivity of all tested cell lines to
19-hydroxy-2′′-oxovoruscharin was in the low nM range (IC50 range for
both sensitive and resistant cells: 7–32 nM). It must be emphasised that in
cardenolides from the Digitalis and Strophantus plant species
(such as digoxin and digitoxin), steroidal rings A/B and C/D are cis
fused, while rings B/C are trans fused. Such ring fusion gives the
aglycone nucleus of these cardiac glycosides a characteristic “U” shape. In
contrast, in cardenolides produced by plants from the milkweed family
Asclepiadaceae (such as calactin uscharin and 2′′-oxovoruscharin) A/B
rings are trans fused resulting in rather flat structures. Whereas the
cardiac glycosides from Digitalis and Strophantus species carry
sugar units linked through the 3β-OH of the steroid aglycone (single
link), some of those produced by plants from the milkweed family
Asclepiadaceae contain a single sugar in a unique “dioxanoid”
attachment (double link; [27], [34], [35], [49], [71], [72], [73]). The
consequences of these structural differences on the NaK binding of these
compounds have been reported previously [34], [35], [44] and indicate
the markedly more potent binding (particularly to NaK α1 subunits) of the
trans-trans-cis cardenolides.
Hypoxia-mediated drug resistance: targeting of the NaK α subunit by
CS
For decades, tumour hypoxia has been known to have a negative effect on therapy
outcomes (recently reviewed in [74]). Hypoxia
inhibits tumour cell proliferation and induces cell cycle arrest, ultimately
conferring chemoresistance because anti-cancer drugs preferentially target
rapidly proliferating cells. However, this knowledge has been largely neglected
during screening for anti-proliferative substances in vitro, resulting in
hypoxia-mediated failure of most newly identified substances in vivo. The
hypoxia-inducible factor (HIF) family of hypoxia-inducible transcription factors
represents the main mediator of the hypoxic response and is often upregulated in
human cancers. The oxygen-regulated HIF isoforms, HIF-1α and to a less
extent HIF-2α, have been associated with chemotherapy failure, and
interference with HIF function holds great promise for improving future
anti-cancer therapy (recently reviewed in [74]).
Accordingly, Zhang et al. [75] screened a library
of drugs that are in clinical trials or in use for inhibitors of HIF-1. Twenty
drugs inhibited HIF-1-dependent gene transcription by > 88 % at a
concentration of 0.4 µM. Eleven of these drugs were cardiac glycosides,
including digoxin, ouabain, and proscillaridin A, which inhibited HIF-1α
protein synthesis and the expression of HIF-1 target genes in cancer cells [75]. Digoxin administration increased the latency
and decreased the growth of tumour xenografts, whereas treatment of established
tumours resulted in growth arrest within one week. Enforced expression of
HIF-1α by transfection was not inhibited by digoxin, and xenografts
derived from transfected cells were resistant to the anti-tumour effects of
digoxin [75], demonstrating that HIF-1 is a
critical target of CS for cancer therapy.
Cytoprotective effects caused by constitutively activated NF-κB:
targeting of the NaK α subunit by CS
Constitutive or drug-induced activation of the NF-κB signalling cascade
represents one of the major pathways by which tumour cells avoid cytotoxicity
[76], [77], [78]. Many tumour cells display constitutively high
levels of nuclear NF-κB activity due to the hyper-activation of the
NF-κB signalling pathways or to inactivating mutations in the
regulatory Iκ-B subunits [76], [77], [78]. Several CS
have already been shown to interfere with the NF-κB pathway [51], [79], [80], [81]. We
previously reported that 19-hydroxy-2′′-oxovoruscharin (UNBS1450) is able to
sensitise chemoresistant, highly aggressive, and naturally therapy-resistant
A549 NSCLC cancer cells by deactivating the cytoprotective effects caused by
constitutively activated NF-κB [51]. This
UNBS1450-induced deactivation of the NF-κB pathways occurs at several
levels, including both the inhibitory I-κB portion of the NF-κB
signalling pathway and its stimulatory p65/Rel-A NF-κB portion. With
respect to the I-κB portion of the NF-κB signalling pathway, the
compound acts at the levels of i) the upregulation of inhibitory protein
expression (as observed for I-κBβ), ii) the downregulation of the
phosphorylation levels of I-κBα, and iii) the downregulation of
the expression of CDC34. With respect to the stimulatory p65/Rel-A NF-κB
portion, the compound induces i) the downregulation of the expression levels of
p65, ii) the downregulation of the DNA binding capacity of the p65 subunit, and
iii) the downregulation of the NF-κB transcriptional activity [51].
How might CS overcome cancer cellsʼ chemoresistance?
We were able to show that NaK α1 targeting by siRNA induced the death of
resistant cancer cells with the same morphologic features as those induced by
19-hydroxy-2′′-oxovoruscharin [35]. Thus, cancer
cells need abundantly expressed NaK for their survival, which seems not to be
the case for normal, non-tumour cells [35].
The observed hypersensitivity of some MDR cells to CS [52] suggests a rather specific MDR targeting. The multifactorial
nature of MDR indicates that it may be important to develop modulators that can
simultaneously inhibit the expression of the drug transporters and the key
signalling pathways, which are responsible for this phenomenon [8], [82]. The available,
yet scarce, data argue in favour of this double mechanism: (a) the inability of
tumour cells to acquire resistance to 19-hydroxy-2′′-oxovoruscharin, (b)
genome-wide microarray analyses performed after 19-hydroxy-2′′-oxovoruscharin
treatment of cancer cells revealed downregulation of different MDR-related mRNAs
(our unpublished data), and (c) by binding to the sodium pump, CS affect
multiple signalling pathways [27], [37], [45], [48], [50]. Furthermore,
post-translational modifications seem to play major roles in the MDR-related
regulation of protein expression. N-glycosylation was shown to contribute to the
stability of P-gp [83], and inhibiting
glycosylation reduced membrane-associated P-gp and altered the MDR phenotype
[84]. Consistent with this observation,
Beheshti Zavareh et al. [85] identified CS as the
most potent inhibitors of the N-glycosylation pathway. Zhang et al. [86] demonstrated that the stability and function of
P-glycoprotein can be regulated by the ubiquitin-proteasome pathway and
suggested that modulating the ubiquitination of P-glycoprotein might be a novel
approach to the reversal of drug resistance. Consistent with this suggestion, we
demonstrated that 19-hydroxy-2′′-oxovoruscharin induced an increase in the
accumulation of ubiquitinylated proteins in the MDR A549 tumour cells and that
some other ubiquitinylation-related enzymes are also affected by this CS [51].
Two major mechanisms might be responsible for CS-induced effects on
chemoresistant cancer cells. The first mechanism relates to the inhibition of
the glycolytic pathway and reduction of intra-cellular ATP levels [87], [88], [89] because these cancer cells have increased
metabolic requirements for ATP [87], [88], [89]. This
hypothesis is also supported by our data on the
19-hydroxy-2′′oxovoruscharin-induced drop in intra-cellular ATP concentrations
in cancer, but not in normal, cell lines [34], [35], [90]. It is interesting that aerobic glycolysis is linked to the
activity of Na+/K+-ATPase and that CS can inhibit aerobic
glycolysis (reviewed in [91]). The mechanism by
which a decrease in the activity of the Na+/K+-ATPase
produces glycolysis inhibition is not completely understood. However, it has
been reported that glycolysis is inhibited by ATP via an allosteric inhibition
of phosphofructokinase (PFK), a key enzyme in the control of glycolysis. Thus,
cells need to hydrolyse ATP in order to release PFK inhibition and activate
glycolysis. One of the major ATPases involved in the hydrolysis of ATP is indeed
Na+/K+-ATPase [91]. Thus,
Na+/K+-ATPase inhibition by CS could prevent the
hydrolysis of ATP, which in turn may inhibit PFK and glycolysis, leading
ultimately to cancer cell death. In addition, glucose transport into cells is
mediated by facilitative glucose transporters (GLUTs) and in some cell types
(such as small intestine and renal epithelial cells) by sodium glucose
transporters (SGLT), the activity of which depends on
Na+/K+-ATPase [91].
Therefore, Na+/K+-ATPase inhibition by CS may also reduce
glucose transport into these cells resulting in further inhibition of glycolysis
[91].
The second mechanism relates to CS-induced changes in cell ion concentrations,
with an increase of Ca2+
i following the Nai
increase due to NaK blockage contributing to the increase of MDR-1 mRNA [92]. In contrast, CS do not affect cell ion
concentrations when used at their IC50 or concentrations that
decrease MDR [52]. Additional data are, however,
needed to decipher the details of the mechanism(s) by which CS circumvent cancer
cell chemoresistance.
In summary, the multiplicity of potential targets might underlie the ability of
CS to overcome the multiple anti-cell death mechanisms established in
cancer.
Which signalling pathways are affected by NaK targeting in resistant cancer
cells?
Although CS-mediated signalling has been investigated in normal cells (indicating
the involvement of ERK, MAPK, PLC, PKC, and Ras-Raf), only a few studies of
NaK-mediated signalling in cancer cells in general and in chemoresistant cancer
cells in particular, have been reported.
By binding to the sodium pump, CS elicit several downstream signalling cascades
affecting a number of different targets (reviewed in [27], [37], [45], [48], [50], [93]). Among the multiple targets
are certain key markers. One pathway that might link NaK and MDR is the one
related to c-Myc because c-Myc is involved in regulating the expression of MDR
[94] and P-gp, the product of the MDR1 gene
[95]; c-Myc activates MDR-1 transcription by
binding the E-box motif (CACGTG) in the MDR1 gene promoter [96]. Our data indicate that (i) CS anti-tumour
efficiency is correlated with the ability to down-regulate c-Myc [97] and (ii) 19-hydroxy-2′′-oxovoruscharin impairs
the expression of five Myc-related genes [90],
suggesting a broad effect on the c-Myc pathway. As a reminder, the c-Myc
oncoprotein regulates transcription of genes associated with cell growth,
proliferation, and apoptosis [98]. The c-Myc
protein is required for activating ribosomal DNA transcription in response to
mitogenic signals, and it coordinates the activity of all three nuclear RNA
polymerases, thereby playing a key role in regulating ribosome biogenesis and
cell growth [99], [100]. Stimulation of ribosomal RNA synthesis by c-Myc is a key
pathway driving cell growth and tumourigenesis [99]. Furthermore, oncogenic signalling through the Myc pathways directly
controls glutamine uptake, which is of vital importance in cancer cells that
must satisfy the metabolic requirements associated with anabolism and rapid
growth rates [99]. Experimental evidence shows that
inhibiting c-Myc significantly halts tumour cell growth and proliferation [101].
The way cardiotonic steroids down-regulate c-Myc expression has not been
deciphered. Among the possible mechanisms are: (i) rapid compound-induced
increases in ROS (as we previously reported [90]),
which can inhibit gene expression partly by the oxidation of Sp1, which
decreases its DNA-binding activity and contributes to the suppression of a
number of genes, including c-Myc [102]; and (ii)
compound-induced STAT3 downregulation (as we previously reported [90]).
It is important to consider Na+/K+-ATPase as a signal
transducer able to mediate CS-induced effects in a compound, concentration, and
cell type-specific manner [27], [37], [45], [93]. Thus, while binding to the same receptor, CS
display different spectra of signatures indicating the differences in their
modes of action and subsequent effects on cell behaviour. Indeed, using Fourier
Transform Infrared (FTIR) analyses on the prostate cancer PC-3 cell line treated
with four different CS (two cardenolides and two bufadienolides), we
demonstrated the differences in the signatures of the metabolic changes induced
by these four compounds [103]. This could explain,
at least partly, the differences in CS behaviours toward the MDR of cancer
cells.
Finally, a question remains about the possible intracellular roles of NaK and CS.
Several studies showed NaK internalisation upon CS binding, and some of them
demonstrated NaK accumulation in the nuclei, suggesting a direct role of NaK in
gene expression [104], [105]. In contrast, the internalisation of CS together with NaK has
still not been demonstrated. If some CS could undergo internalisation, this
might explain, at least partly, why certain CS are substrates of P-gp and others
are not.
Potential NaK isoform-related specificities in overcoming cancer cellsʼ
resistance
Using a baculovirus expression system for studying
Na+/K+-ATPase-mediated ouabain effects, Pierre et al.
[106] showed that there were important
isoform-specific differences in NaK signalling. It is important to remember that
different CS display different NaK inhibitory properties and that most, if not
all, of them display higher binding affinity for the α2 and α3
isoforms compared to the α1 isoform [107].
Furthermore, conspicuous kinetic differences exist among sodium pump isozymes
from different species in their interaction with CS [107], [108], [109], [110]. According to Crambert et
al. [108], human α/β complexes
formed with α1 and α3 subunits have slow dissociation rate
constants corresponding to half-lives (t1/2) between 30 and 80 min,
whereas those formed with α2 have rapid dissociation kinetics with
t1/2 of about 4–5 min. Similarly, the association kinetics of
ouabain with human Na+/K+-ATPase isozymes followed the
order α2 >> α3 = α1, with the times required to reach
equilibrium binding being approximately 10 min (α2,β) and 60 min
(α1,β and α3,β). The association rate of ouabain
seems to depend on the steroid moiety, whereas the dissociation rate depends on
both the steroid and the sugar moieties. Several amino acids are involved in the
ouabain binding kinetics [108], [111]. Whether there is isoform-specific mediated
sensitivity towards the CS that display anti-cancer effects remains an open
question. Currently, most of the published data link the α1 NaK subunit
over-expression with cancer progression [27], [31], [32], [34], [35], [36], [37], [38]. Newman et al. [45]
suggested that rather than an increase or decrease in NaK α subunit
expression, the ratio of α3 to α1 should be used as the prognostic
indicator for candidate patients to be treated with CS. This proposal was based
on their data obtained with pancreatic cancer cell lines. The data suggest that
the higher the ratio of α3 to α1, the greater the sensitivity to
oleandrin. Unfortunately, this type of investigation cannot be conclusively
conducted with a large panel of human cancer cell lines because the NaK α
subunit expression is significantly influenced by culture conditions in
vitro
[112], [113], which
generally lead to the sole expression of α1.
CS-mediated NaK targeting: from bench to bedside – how far are we?
As already emphasized above, interest in developing the CS as anti-cancer agents
has grown progressively in the last two decades despite their potential
cardiotoxic effects and very narrow therapeutic index. Within the past 15 years,
there has been a marked increase in the number of reports of CS-induced
anti-cancer effects (recently reviewed in [26], [27], [37], [39], [45], [47], [48]). While in vitro anti-cancer properties of CS have been
widely studied, few publications have demonstrated their in vivo activity
in animal models or in clinical studies. Either these compounds demonstrated
appreciable in vivo anti-tumour activity but were quite toxic (e.g.,
ouabain) or they were found to be relatively devoid of anti-tumour activity at
the tolerated dose levels (e.g., digoxin). The studies published by Perne et al.
and Hallböök et al. [114], [115] raised the concern about the potential use of
CS in therapy since their results demonstrated that CS (digoxin and digitoxin)
induced cell death in human cells by inhibiting general protein synthesis,
pointing to the need of very detailed assessment of mechanism of action of
potential therapeutic CS. Despite, recently, Platz et al. [116] reported on a novel two-stage,
transdisciplinary study identifying digoxin as a possible drug for prostate
cancer treatment. They investigated whether any clinically-used drugs might have
utility for treating prostate cancer by coupling a high-throughput
laboratory-based screen and a large, prospective cohort study. Stage 1 was based
on an in vitro prostate cancer cell cytotoxicity screen of 3,187 compounds in
which digoxin emerged as the leading candidate given its potency in inhibiting
proliferation in vitro (mean IC50 = 163 nM) and common use.
Stage 2 was based on evaluating the association between the leading candidate
drug from stage 1 and prostate cancer risk in 47 884 men followed 1986–2006 and
uncovered that regular digoxin users had a ~ 25 % lower prostate cancer risk.
Thus this transdisciplinary approach for drug repositioning provides compelling
justification for further mechanistic and possibly clinical testing of this
class of compounds as drugs for cancer treatment. As a reminder, retrospective
epidemiological studies conducted by Stenkvist revealed some intriguing results:
very few patients that underwent CS treatment for heart problems died from
cancer [41]. In a 20-year follow-up [117], Stenkvist has reported that the death rate
from breast carcinoma (excluding other causes of death and confounding factors)
was 6 % (two out of 32) among patients who were treated with digitalis, compared
with 34 % (48 of 143) among patients who were not treated with digitalis
(p = 0.002). On the other hand, a very recent report from Biggar et al. [118] reported an increasing risk of breast cancer
in women taking digoxin for cardiac conditions: 2.05 % (2144 out of 104 648) of
women using digoxin developed breast cancer. Two oncology clinical trials
involving digoxin have recently been completed: i) a Phase I clinical trial
(ClinicalTrials.gov Identifier NCT00650910) combining digoxin with Lapatinib (an
oral receptor tyrosine kinase inhibitor that targets HER2 and the EGFR) in
treatment for metastatic ErbB2 breast cancer and ii) a Phase II clinical trial
(ClinicalTrials.gov Identifier NCT00281021) combining daily digoxin with
Erlotinib, an EGFR inhibitor, in treatment for NSCLC. Unfortunately, a
remarkable digoxin-mediated anti-tumour effect was not observed in any of these
trials, emphasising the need for more clinically efficient anti-tumour CS.
Interestingly, it is somewhat perplexing to observe the large number of patents
filed (see [39]) for novel anti-cancer CS and the
very limited number of these compounds being further assessed in pre-clinical
investigations and clinical trials. Indeed, a very limited number of new CS are
presently being evaluated in clinical trials: (i) Nerium oleander extract
(PBI-05240) is in Phase I clinical trials (ClinicalTrials.gov Identifier
NCT00554268) at the MD Anderson Cancer Center and an interim analysis presented
at 2009 ASCO Conference reported that 20 % of evaluable patients achieved stable
disease for more than 4 months [119]; (ii) one
modified cardenolide, UNBS1450, selected to minimize cardiotoxicity while
preserving potent anti-proliferative properties [49], is also currently in Phase I clinical trials in Europe (Belgium
and The Netherlands); and (iii) a traditional Chinese medicine Huachansu
(containing mainly bufadienolides) is currently being evaluated in a Phase II
clinical trial along with gemcitabine in pancreatic cancer patients
(ClinicalTrials.gov Identifier NCT00837239). Negative perceptions of CS toxicity
and reticence of medical community might be part of the explanation for the
observed discrepancy. Furthermore, elevated costs of pre-clinical investigations
might be one of the major reasons for the lack of translational research,
knowing that large number of the patent applications for novel anti-cancer CS
came from academic investigators. On the other hand, the lack of available
clinical data evidencing safety margin and therapeutic window of assessed new
anti-cancer CS prevent pharmaceutical industry to consider large investments in
order to investigate CS as potential new anti-cancer compounds.
