Key words ovarian cancer - metastasis - molecular pathway - β-adrenergic signaling - β-Blocker
- hallmarks of cancer
Schlüsselwörter Ovarialkarzinom - Metastase - molekulare Signalwege - β-adrenergische Signalgebung
- β-Blocker - Kennzeichen von Krebs
Background
Ovarian carcinoma (OC) occupies seventh place on the list of female cancers in Germany
[1 ]. Approximately 7500 women are diagnosed with the disease annually in Germany alone
[1 ] and it is the fifth most common cause of cancer-related death among women [1 ]. This is partly due to the fact that OC is often diagnosed at an advanced stage.
A large majority of cases (84 %) are diagnosed at FIGO stage IIIC, i.e. the carcinoma
has already spread beyond the pelvis and extrapelvic tumour size is larger than 2 cm
[2 ], [3 ].
Complete operative tumour resection has been identified as a decisive prognostic factor
[4 ]. Systemic treatment with monoclonal antibodies such as bevacizumab (Avastin®) has
been shown to prolong progression-free survival [5 ], [6 ], however actual survival advantage is not more than a few months [5 ], [6 ]. With a relative 5-year survival rate for all ovarian carcinomas of only around
42 %, treatment results on the whole are unsatisfactory [1 ], [7 ] and there is a dire need for new treatment concepts.
A recent retrospective study showing a survival advantage for ovarian carcinoma patients
treated with nonselective beta blockers met with a very positive response both in
the lay and specialist press [8 ], [9 ], [10 ] ([Fig. 1 ]).
Fig. 1 Kaplan-Meier curve (overall survival) of patients with ovarian carcinoma treated
with/without beta blockers. Median survival for patients without beta blockers was
34.2 months, with selective beta blockers 38.2 months (p = 0.005) and with nonselective
beta blockers 90 months (p < 0.001) ([8 ], with kind permission).
Contrary to the very optimistic, near sensational impression given by the lay press,
numerous studies on the subject already exist [8 ], [11 ], [12 ], [13 ], [14 ], [15 ] whose findings have been critically appraised in the specialist literature [16 ]. All existing clinical studies are retrospective in nature, have heterogenous patient
groups and to some extent present inconsistent results. The following article gives
an overview of the latest preclinical data on the pathophysiology of catecholamines
in ovarian carcinoma and summarizes the available clinical data on the use of beta
blockers in this context.
Studies using animal model stress regimes and those focusing on psychological aspects,
such as patient distress and the possible role of psychotherapeutic agents or psychiatric
support, were excluded from this review due to capacity constraints. For a detailed
discussion of these issues the reader is referred to an article by Hefner et al. [17 ]. Here we focus on pathophysiology at the biochemical level.
Basic Pathophysiologic Principles
Basic Pathophysiologic Principles
Research on cell cultures and animal models from the past 15 years has consistently
illustrated the detrimental effects of catecholamines on ovarian carcinoma and the
possibility of blocking these effects. This is applicable to both the direct effects
of catecholamines on tumour cells (anoikis, cell migration and invasion) [18 ], [19 ], [20 ], [21 ], [22 ] and indirect effects on the tumour microenvironment (inflammation, angiogenesis)
[23 ], [24 ], [25 ], [26 ] ([Table 1 ]).
Table 1 Preclinical studies: effects of agonists and antagonists on adrenoreceptors in ovarian
carcinoma.
Author, year published
C/X
Agonist
Effect
Effect blockade
No blockade
C = cell culture, X = xenograft
Sood 2010 [18 ]
C
A, N
Reduced anoikis (by a factor of 0.5)
Propranolol
Rangarajan 2003 [19 ]
C
I
Increased cell adhesion (factor: 1.5)
Enserink 2004 [20 ]
C
I
Increased cell migration (factor: 3)
Sood 2006 [21 ]
C
A, N
Increased propensity to invade (198 %)
Landen 2007 [22 ]
A, N
Raised STAT3-concentration
Propranolol
Prazosin
Nilsson, 2007 [23 ]
C
A, N, I
Increased IL-6 secretion (by a factor of 200)
Propranolol
Shazad 2010 [24 ]
C
A, N
Increased IL-8 secretion (factor: 3)
Propranolol
Lutgendorf 2003 [25 ]
C
A, N, I
Increased VEGF production
Propranolol
Thaker 2006 [26 ]
C
N, I
Increased VEGF mRNA transcription (factor: 8.4)
Propranolol
X
I, T
Increased tumour mass (factor: 2.5)
Propranolol
Armaiz-Pena 2015 [27 ]
C
A, N, I
Raised IL-8, IL-8, VEGF, MCP1 levels
Propranolol
Atenolol
Nagaraja 2015 [28 ]
C
N, I, T
Raised PGE concentration
Propranolol
C
N
Increased p65 and p50 in cell nuclei
X
N
Increased number of tumours
Choi 2015 [31 ]
C
N
Increased hTERT expression
X
N
Increased propensity to metastasise
Kang 2015 [36 ]
C
N, I before paclitaxel or cisplatin
Reduced apoptosis rate (43 %)
Propranolol
Atenolol
C
N, I, T
Increased DUSP1 promoter activity
Propranolol
Metoprolol
Newly Discovered Pathomechanisms
Newly Discovered Pathomechanisms
The most recent preclinical studies on catecholamines in cancer provide deeper insight
into some pathophysiologic interrelationships. In the context of ovarian carcinoma,
studies of inflammation, cell senescence and chemoresistance have been particularly
important. Propranolol, a nonselective beta blocker acting on both β1 and β2 adrenergic
receptors has been used most in research ([Table 1 ]). In one study of inflammatory reactions the application of catecholamines to an
ovarian carcinoma cell line lead not only to a rise in IL-6 and IL-8 but also to increased
levels of monocyte chemotactic protein 1 (MCP1) [27 ]. MCP1 contributes to increased monocyte recruitment into tumour tissue, and raised
MCP1 blood concentrations were associated with higher stage of disease, shorter progression-free
survival and worse overall survival [27 ]. The catecholamine induced rise in IL-6, IL-8 and MCP1 concentrations observed in
the cell line was inhibited by beta blockers [27 ] ([Table 1 ]).
Another study discovered a previously unknown interconnection between the signal pathways
of catecholamine and prostaglandin metabolism [28 ]. In the experiment by Nagaraja et al. noradrenaline application lead to increased
PGE2 production via the ADRB2–NF-κB–PTGS2–PGE2 signal cascade in cell lines with β
adrenergic receptors, and to increased activity of the PTGS2 and PTGES genes necessary
for this to occur [28 ]. In an orthotopic mouse model experimental deactivation of the PTGS2 gene lead to
reduced tumour load and metastasis [28 ]. A genome analysis of patients found that strong expression of β2 adrenergic receptors,
PTGS2 and PTGES was associated with reduced progression-free survival and overall
survival [28 ] ([Table 1 ]).
Another previously unknown pathomechanism was discovered during studies of telomerase.
In up to 95 % of ovarian carcinoma cells the catalytic subunit of telomerase (hTERT)
is upregulated in order to stabilise tumour cell telomeres [29 ], [30 ]. A complex signal pathway to hTERT via β2 adrenergic receptors/PKA/Src/HIF-1α/c-Myc
was demonstrated on addition of noradrenaline to ovarian carcinoma cells. Simultaneously,
hTERT induced the expression of Slug, a central gene in epithelial-mesenchymal transition
[EMT]) [31 ] ([Table 1 ]). EMT itself is regarded as essential for the development of newly discovered ovarian
carcinoma cancer stem cells of [32 ], [33 ]. In a study by Choi et al. using a mouse model noradrenaline administration lead
to increased hTERT expression and pulmonary metastasis of ovarian carcinoma cells
[31 ] ([Table 1 ]).
Studies on chemoresistance in association with catecholamines have proven to be particularly
significant with respect to clinical disease course. It is already known from preclinical
studies on other tumour entities that catecholamines increase chemoresistance of tumour
cells and that beta blockers can potentiate chemotherapeutic effects [34 ], [35 ]. The most recent work on ovarian carcinoma now shows similar results [36 ]. Various cell lines were treated with catecholamines and thereafter exposed to paclitaxel
or cisplatin. The apoptosis rate usually observed under these chemotherapeutic agents
was reduced [36 ]. The effect was only demonstrable on application of substances with β2 receptor
agonist properties, and only in cell lines possessing β2 receptors [36 ]. This disadvantageous effect on chemotherapeutic action was mediated by the dual
specificity phosphatase 1 (DUSP1), whose expression was increased by the stress hormones
[36 ]. In addition a further signal pathway was described that mediates JNK-dependent
phosphorylation of c-Jun via cAMP-PLC-PKC-CREB, protecting ovarian carcinoma cells
from apoptosis [36 ]. There was no loss of chemotherapeutic effect after application of a β2 receptor
blocker [36 ] ([Table 1 ]).
Beta Blockers and the Clinical Course of Ovarian Carcinoma
Beta Blockers and the Clinical Course of Ovarian Carcinoma
Negative or inconsistent results
First abstracts on the clinical application of beta blockers in ovarian carcinoma
were published in 2012 by Eskander et al. [11 ], [12 ]. In a retrospective, single centre study of overall survival no survival advantage
was shown for the use of beta blockers in a study population of 680 newly diagnosed
patients from all disease stages [11 ]. Prolonged beta blocker use for more than 2.5 years was associated with a 47 % reduced
likelihood of dying from ovarian carcinoma. Overall and progression-free survival
were determined retrospectively using 489 data sets from the same patient collective
[12 ]. Here the analysis showed significantly reduced survival with beta blocker use,
especially among younger patients (< 61 years); there was a nonsignificant negative
trend for progression-free survival with beta blocker use [12 ] ([Table 2 ]).
Table 2 Clinical studies on beta blockers and disease course in ovarian carcinoma (OC).
Author, year, design
Tumour entity, n
Type of beta blocker
Duration of use (UD)
Results with/without beta blocker
AGO = study group of the working group for gynaecological oncology of the German Society
of Obstetrics and Gynaecology† 95 % CI: 0.47–0.81; p < 0.0005, ‡ 95 % CI: 0.48–0.83; p = 0.001
Eskander 2012 [11 ]
Initial diagnosis
Undefined
UD > 30 d prior to diagnosis
Overall survival
Eskander 2012 [12 ]
Initial diagnosis
Undefined
UD > 30 d prior to diagnosis
Overall survival
Johannesdottir 2013 [13 ]
Initial diagnosis OC
Undefined
UD = most recently < 90 d prior to diagnosis
Compared to no beta blocker
Heitz 2013 [14 ]
Recurrence
Sel. β1 blocker n = 32
Undefined
Overall survival
Diaz 2012 [15 ]
Initial diagnosis
Sel. β1 blocker n = 17
Undefined
OC specific survival
Watkins 2015 [8 ]
First diagnosis
Sel. beta blocker n = 194
Overall survival†
‡
Johannesdottir et al. performed a far more extensive, retrospective analysis of 6626
data sets from the Danish Cancer Registry of newly diagnosed ovarian carcinoma patients
at all stages of disease. Disease course was compared between patients who had never
taken beta blockers, those who had used beta blockers less than 90 days prior to data
acquisition, and those who had used them more than 90 days previously [13 ]. There was no difference in mortality risk between the groups [13 ] ([Table 2 ]).
In the context of recurrent ovarian carcinoma, Heitz et al. found no advantage for
the use of β1 receptor blockers in a retrospective analysis of the prospective, multicentre
Ovar-2.4 and Ovar-2.5 studies that were initiated by the working group for gynaecological
oncology (AGO) of the German Society of Obstetrics and Gynaecology [14 ], [37 ], [38 ] ([Table 2 ]).
Positive results
Diaz et al. reported a statistically significant benefit for both disease-specific
and progression-free survival at disease stages III and IV with the use of beta blockers
[15 ]. In their retrospective, single centre study the authors calculated that beta blockers
lead to a 54 % reduced chance of dying [15 ] ([Table 2 ]).
Recently the much discussed retrospective, multicentre study by Watkins et al. including
1425 ovarian carcinoma patients at all stages of disease also showed a survival advantage
for the use of beta blockers [8 ] and a distinction between selective and nonselective beta blockers was documented
for the first time [8 ]. Although use of selective beta blockers produced a survival advantage overall,
median survival was significantly worse (38.2 months) than with nonselective beta
blockers, and in some cases the use of selective beta blockers was even associated
with reduced survival [8 ]. In contrast, median survival using nonselective beta blockers was 90 months compared
to 34.2 months in patients not receiving any beta blocker [8 ]. The hazard ratio (HR) for death following a diagnosis of ovarian carcinoma was
0.26 with beta blockers overall, 0.32 for selective beta blockers and 0.08 for nonselective
beta blockers [8 ] ([Table 2 ]).
Discussion
Ovarian carcinoma remains one of the most commonly occurring, and one of the most
commonly fatal malignancies in women [1 ]. Treatment options developed over the past 50 years have not improved disease prognosis
significantly [1 ], [7 ] and innovative treatment alternatives are urgently needed.
In the realm of preclinical research impressive studies of first-rate quality have
been published for most of the hallmarks of cancer [39 ].
These include studies on chemoresistance, invasivity, migration and adhesion tendency,
inflammation reactions and angiogenesis [18 ], [19 ], [20 ], [21 ], [23 ], [24 ], [25 ], [26 ], [27 ], [36 ], [40 ]. Recent discoveries such as interconnections between the metabolism of catecholamines
and pain mediators [28 ], or EMT and cancer stem cell development [31 ] provide new targets for potentially innovative treatments ([Table 1 ]).
Despite these successes at the pathophysiological level many questions remain open,
such as the significance of the autonomic innervation of tumour tissue [41 ], the roll of β3 receptors [42 ] and apoptosis pathways via protein p53 [41 ]. The sporadically observed positive effects of catecholamines and negative effects
of beta blockers remain completely unexplained and require urgent further study [40 ].
Important points of criticism of the preclinical work to date include the use of pharmacological
doses of catecholamines and xenografts, both of which complicate the assessment of
clinical significance. However this applies to preclinical research on ovarian carcinoma
in general, which requires innovative studies of pathomechanisms using modified cell
lines and animal models [43 ]. These studies could be usefully expanded on through studies of catecholamines and
beta blockers. As an example, on the basis of experience with MCP1, the combination
of checkpoint inhibitors and beta blockers could constitute an innovative design to
enable the study of immune therapy synergism [27 ], [44 ], [45 ].
Despite the limitations and justified criticism of this preclinical data it has nevertheless
convinced many researchers that catecholamines do promote relevant aspects of tumour
progression, and that especially nonselective beta blockers could reduce these effects
[46 ].
And indeed the latest clinical work on the influence of beta blockers not only on
ovarian carcinoma but also on breast cancer and malignant melanoma, does prima facie
support this conclusion [8 ], [9 ], [10 ], [46 ], [47 ], [48 ], [49 ], [50 ], [51 ], [52 ], [53 ]. In a recent multicentre study including 1425 patients with ovarian carcinoma at
all different stages, on retrospective analysis beta blockers were shown to provide
a significant survival advantage [8 ], and for the first time, an advantage of nonselective beta blockers over selective
beta blockers was demonstrated [8 ]. This result fulfills the hypothesis of preclinical studies where the main beta
blocker effect was shown to occur via β2 receptors [36 ], [41 ], [42 ]. At the same time it provides a possible explanation for the nonsignificant findings
of studies that either did not stratify by β receptor type [11 ], [12 ], [13 ] or in which patients mainly took β1 receptor blockers [14 ]; the findings of Diaz et al. are in disagreement though, showing a survival advantage
for beta blocker use even though the majority of their patients took β1 receptor blockers
[15 ] ([Table 2 ]). The most controversial issues, however, surround prognosis. Diaz et al. found
survival advantages for patients in the more advanced disease stages III and IV in
particular, and in the study by Watkins the hazard ratio for death for patients at
all disease stages following diagnosis of ovarian carcinoma was 0.26 for those taking
beta blockers, 0.32 for selective beta blockers, and 0.08 for nonselective beta blockers
[8 ]. In stark contrast, the HR for the use of platinum-based chemotherapy in advanced
disease was calculated at 0.88 [54 ]. If true, this would make beta blockers a sort of “wonder drug”, their effects far
surpassing those of standard treatments [16 ]. This is seriously doubted by commentators [16 ] however, who suspect the results may have been skewed by a statistical bias (so-called
“immortal person-time bias”) [16 ]. This occurs when the definition of an exposure or a covariable is dependent on
an event (e.g. starting beta blocker treatment) occurring after the start of the follow-up
period; in the time between the beginning of follow-up and e.g. starting a beta blocker
the patient is statistically “immortal” and their data will distort the groupʼs survival
time [16 ].
Watkins and his co-authors dispelled this criticism in their case stating that only
an estimated 5 % of study participants had started beta blockers after the beginning
of follow-up [55 ]. In addition they referred to preclinical studies on ovarian carcinoma and other
tumour entities where beta blockers helped to sensitise malignant cells to chemotherapeutic
agents, potentiating chemotherapy effects [34 ], [35 ], [36 ], [56 ], [57 ]. Initial groundbreaking prospective clinical work on pancreas carcinoma has shown
nearly doubled survival rates with the addition of a combination of beta blockers
and COX-2 inhibitors to standard chemotherapy [58 ].
Despite the euphoria, however, it should not be forgotten that all clinical studies
on ovarian carcinoma to date have been retrospective in nature, and at best should
be considered as contributing towards the generation of hypotheses. In view of the
poor prognosis associated with ovarian carcinoma it is very possible that a publication
bias/“file drawer problem” exists, where nonsignificant or negative results are not
published, and that positive findings even from retrospective studies receive undue
acclaim both in the speciality and lay press [59 ], [60 ]. In addition, further distortion of results due to the previously mentioned “immortal
person-time bias” must be assumed, since, according to a recent review, all positive
effects of beta blockers in cancer are subject to this bias [61 ]. Further limitations of the reviewed studies of ovarian carcinoma are their retrospective
nature, limited patient numbers and the fact that the various disease stages were
not considered separately. No study has yet considered the beneficial effects of catecholamines
in the context of peritoneal carcinomatosis and severe tumour recurrence, or the possibility
of perioperative beta blockade [62 ], [63 ], [64 ]. Before beta blockers can be widely implemented in clinical practice for ovarian
carcinoma a “second wave” of clinical studies is required [65 ] that are at least prospective in design with a focus on relevant biomarkers [66 ]. As is also the case with other tumour entities it will be necessary to study the
receptor profile and density in ovarian carcinoma in order to select suitable beta
blockers [41 ]. The expression profiles of catecholamine dependent genes before application of
beta blockers have also not yet been determined [41 ]. Most importantly, however, when selecting a beta blocker increased attention must
be paid to the individual patientʼs comorbidities and relevant drug indication restrictions
and side effect profiles. Although beta blockers in general are known to be safe and
economic from decades of use in other areas of medicine, selective beta blockers,
which have been preferred in other medical fields in view of their favourable side
effect profile, appear to be less effective in ovarian carcinoma and may even be detrimental
[8 ], [14 ]. Also, without in-depth knowledge of possible drug interactions beta blockers used
as co-medication with standard chemotherapies increase the risk of side effects. Pharmacokinetic
characteristics should also be investigated in vivo since beta blocker degradation
via the cytochrome system is well known and could contribute to increased excretion
with consequent reduced efficacy on an individual basis [8 ]. Lastly, the consideration of specific time points in the disease course may prove
innovative: preclinical data suggest so-called “windows of opportunity” (e.g. during
chemotherapy or when metastasis or recurrence occur) during which beta blockers may
be particularly effective [41 ], [42 ], [47 ]. To our knowledge, both a feasibility study and a prospective study on the clinical
application of beta blockers in ovarian carcinoma are currently underway [67 ], [68 ]; we eagerly await their results as they may provide first data justifying the use
of beta blockers in ovarian carcinoma.
Conclusion
Preclinical data clearly indicate that catecholamines influence ovarian carcinoma
unfavourably. In vitro these catecholamine effects can be inhibited with the aid of
beta blockers. Recent studies also report benefits from beta blockers in clinical
practice, however these optimistic reports are based on retrospective data analyses.
Existing studies assist the generation of new hypotheses, e.g. on pathophysiologic
interrelationships, and form a basis for future prospective clinical studies with
a focus on relevant biomarkers. The evidence published to date, however, does not
justify the widespread clinical application of beta blockers in ovarian carcinoma.
