Biologically targeted therapies (1,363 words)
Since less than 20% of all PDAC patients exhibit surgically resectable disease at time of presentation, systemic chemotherapy is currently the most frequently applied treatment option [
21]. Albeit the development of novel poly-chemotherapy protocols, the overall prognosis, and survival rate of PDAC patients still remain poor. Hence, there is a strong demand for novel, biologically motivated treatment strategies with higher specificity for PDAC-relevant, tumor-driving targets. The genomic landscape of PDAC is dominated by a handful of signature genes which are affected by aberrations and mutations at high frequencies:
KRAS,
CDKN2A,
TP53, and
SMAD4 [
49,
51]
. All of these genes are still basically considered to be undruggable, although agents targeting mutant
TP53 have been developed, and attempts to pharmacologically manipulate
RAS function are constantly increasing [
137,
138]. So far, substances targeting downstream effectors of these major PDAC drivers or other regulators which are also frequently altered, including
BRAF,
ERK, PI3K/AKT, and
mTOR, are in the focus of investigation.
The
mitogen-activated protein kinase (
MAPK) signaling cascade offers promising perspectives in this regard, because PDAC cells are known to depend on
MAPK signaling, both in terms of progression and metastasis formation [
139,
140]. The most apical possibility to interfere with
MAPK signaling is targeting the
epidermal growth factor receptor (
EGFR). However, a phase III trial evaluating the efficacy of anti-
EGFR treatment with cetuximab in addition to gemcitabine-based chemotherapy showed no significant improvement in clinical outcome [
141]. Recent data attributed this to a compensatory activation of
Integrin β1 signaling [
142]. Downstream of
EGFR, KRAS constitutes a near-perfect target for PDAC treatment as revealed by preclinical RNA interference experiments [
143]. However, clinical RNA interference is challenging, and no reliable
KRAS inhibitors have been described so far [
144]. Nevertheless, pharmacological disruption of the interaction between
KRAS and phosphodiesterase
PDEδ was shown to efficiently suppress PDAC progression
in vitro and
in vivo [
145]. The only targeting approach for
MAPK signaling that has entered the clinical routine thus far is the combination of gemcitabine and the
EGFR-specific tyrosine kinase inhibitor erlotinib [
146]. Although
EGFR is considered to be its only target, erlotinib was reported to be similarly effective in tumors with wildtype or hyperactive mutants of
KRAS, respectively [
147]. This implies that either inhibition of tyrosine kinases other than
EGFR or feedback regulatory mechanisms between hyperactivated
KRAS and
EGFR may be involved, respectively [
148‐
151]. Sunitinib, a tyrosine kinase inhibitor that does not target
EGFR, failed to show similar performance when combined with gemcitabine [
152], and preclinical data support the notion that indeed inhibition of gemcitabine-induced
MAPK signaling by erlotinib accounts for the observed clinical benefits [
153]. Several other inhibitors of
MAPK signaling, including inhibitors of
EGFR, MEK,
ERK, and corresponding protein phosphatases, have shown convincing performance in preclinical studies [
154‐
156], but their potential for clinical implementation remains to be examined, as for instance in ACCEPT, a randomized phase II trial combining gemcitabine with the
EGFR inhibitor afatinib (NCT01728818).
Single-drug treatments – most likely – will not be sufficient to improve the therapeutic outcome of PDAC [
157]. Instead, dual or even multiple targeting strategies appear to be required in order to achieve significant advances. One example is the concomitant inhibition of
MAPK and
PI3K/AKT signaling. Preclinical data revealed that inhibition of
MAPK signaling results in potent compensatory activation of
PI3K/AKT signaling and
vice versa, each being of importance for PDAC progression [
158,
159]. Indeed, concomitant inhibition of
MAPK and
PI3K/AKT signaling did interfere with tumor progression to significantly greater extent than the single-drug treatments in preclinical PDAC models [
158,
160]. However, other studies reported only modest effects of combined
MAPK and
PI3K/AKT inhibition [
161‐
163], and clinical trialing of this combination failed [
164]. One potential explanation could be that inhibitors of different target specificities were employed. A more detailed characterization of the target spectrum of these inhibitors would clarify this and could also help to find new targets for mechanism-based therapies. In this regard, upstream and/or transcriptional regulators of
PI3K expression, such as
transducin beta-like 1 (TBL1), may also be of interest as studies in genetic mouse models have identified them as crucial checkpoints in PDAC development and progression [
165]. Nevertheless, if this mechanism can be exploited therapeutically remains unclear [
166].
The
mammalian target of rapamycin (
mTOR) pathway is best known for its functions in cell survival, proliferation, motility, and evasion of apoptosis [
167]. In several preclinical studies,
mTOR inhibitors revealed promising results [
168‐
171], but it was also reported that inhibition of
mTOR stimulates feedback activation mechanisms involving
MEK/ERK or
AKT signaling, respectively, further emphasizing the need for combinatorial treatment regimens [
172‐
176]. Not surprisingly, multi-pathway inhibition regimens are commonly associated with higher levels of toxicity [
177]. This toxicity often interferes with clinical implementation. Nevertheless, clinical trials evaluating
mTOR inhibition as monotherapy in PDAC altogether failed [
178‐
180], and combined modality approaches of
mTOR inhibition in conjunction with capecitabine revealed only limited improvements as compared to capecitabine alone [
181]. These findings raise the question whether
mTOR inhibitors, despite their successful clinical implementation for the treatment of neuroendocrine pancreatic tumors, may at all represent a therapeutic alternative for the treatment of PDAC [
182], or whether such approaches have been inadequately tested in the clinic.
PDAC is commonly considered a hypovascularized tumor [
183], but relevant expression of
vascular endothelial growth factor A (
VEGF-A) has been observed [
184]. Therefore, the
VEGF-A-specific antibody bevacizumab was tested in combination with gemcitabine in a randomized phase III trial with locally advanced PDAC but failed to show improved outcome [
185]. A possible explanation could be the expression of other
VEGF isoforms. However, complementary phase III trials which evaluated the
VEGF receptor tyrosine kinase inhibitor axitinib in combination with gemcitabine, or the combination of bevacizumab, gemcitabine, and erlotinib, respectively, also failed [
186,
187]. In summary, these results render therapeutic targeting of angiogenesis a questionable approach for the treatment of PDAC [
188].
A subset of PDAC tumors (approximately 15% of all cases) is characterized by mutations in genes that are related to the DNA damage response [
54]. Amongst those, PDAC tumors carrying mutations in
BRCA1/2 genes are of highest interest as they are supposed to be defective in homologous recombination DNA damage repair [
189]. Accordingly, patients with
BRCA1/2-mutated tumors were reported to benefit significantly more from platinum-based chemotherapy than patients with
BRCA1/2 wildtype tumors [
190,
191]. For
BRCA1/2-deficient tumors, the inhibition of
Poly-(ADP-ribose)-polymerase (
PARP) may be promising, since this enzyme shares an axis of synthetic lethality with
BRCA1/2 [
192]. Initial trials examining the therapeutic potential of
PARP inhibitors in patients with
BRCA1/2-deficient PDAC reported promising results [
193‐
196]. Currently, the randomized phase III POLO trial is evaluating
PARP inhibition in patients who received first-line platinum-based chemotherapy, and results are awaited in 2019 (NCT02184195). Beyond
BRCA1/2, mutations in other genes of the DNA damage response, including
ATM, may select for PARP inhibitor sensitivity [
197].
In addition to the described genetic alterations, PDAC tumors display relevant changes in epigenetic modifications, including DNA methylation, histone post-translational modification, nucleosome remodeling, and regulation by non-coding RNAs [
56]. In contrast to genetic alterations, epigenetic modifications are in principle reversible, and it is plausible to assume that pharmacological interference with epigenetic mechanisms underlying PDAC pathology and progression could open new therapeutic perspectives [
198]. Preclinical results of epigenetic therapies have so far been promising, PDAC cell plasticity could be reduced, and resistance against standard chemotherapy was attenuated. However, in mono-agent settings, epigenetic therapeutics did not provide any measurable benefits, demanding for combined modality settings, e.g. in conjunction with chemotherapy or in form of multi-agent combinations, such as combined inhibition of
bromodomain and extra-terminal motif (BET) proteins and
histone deacetylases (HDACs) [
199]. Currently, various phase I/II trials are ongoing which will determine the clinical perspectives of such approaches. Despite all efforts, individualized, mechanism-based treatment strategies for PDAC are still far from being clinical standard [
200].
Therapeutic targeting of hypoxia and metastasis formation appears to be very attractive in the PDAC context, since hypoxia is a principal determinant of therapy resistance and metastasis formation, and metastases are the major cause of death [
20,
74]. Regardless of all preclinical efforts [
201], however, no therapeutic strategy could so far be established. Sort of alternatively, efforts to (re-)activate the immune system in order to detect and combat macro- and micro-metastases have been undertaken and will be discussed in the following.
Immunotherapy
Immunotherapy implementing immune checkpoint inhibitors has revolutionized cancer treatment in the last years [
202]. Therapeutic antibodies targeting
cytotoxic T-lymphocyte-associated protein 4 (
CTLA-4) or the axis of
programmed cell death protein 1 (PD-1) and its corresponding ligand
PD-L1 have shown compelling results in several different cancer types, including metastasized melanoma and lung cancer [
36,
203]. Hence, immune checkpoint inhibition was also tested in PDAC [
35,
39], but compared to melanoma and lung cancer, considerably smaller numbers of patients (approximately 2%) exhibited clinical benefits [
40,
204]. Consistently, the responding tumors showed high levels of microsatellite instability, providing a mechanistic explanation as well as a potential future stratification marker, since microsatellite instability is known to increase the number of tumor-associated neo-antigens [
205].
A major determinant of the immunotherapeutic success are tumor-specific T cells and their (re-)activation. Although their numbers have been described to be rather low in PDAC patients [
90], recent data suggest that the tumor-reactive T-cell repertoire is similar to the one found in melanoma where T cell-based therapies meanwhile have relevant therapeutic impact [
91]. Further studies showed that neo-antigen quality rather than quantity, and strong intra-tumoral CD8
+ T cell infiltration are associated with prolonged survival, indicating that the stimulation of anti-tumor T cell responses can indeed be a promising strategy for the treatment of PDAC [
60,
206,
207]. Along these lines, different vaccination strategies employing various kinds of antigens have already been tested [
208‐
210]. The Algenpantucel-L vaccine consisting of irradiated, allogeneic pancreatic tumor cells stably expressing
alpha-1,3-galactosyltransferase 2 (
A3GALT2), a glycosylating enzyme that mainly targets lipids and extracellular proteins, turned out to be the most promising candidate for a PDAC-targeting vaccine [
209]. However, this vaccine failed to improve treatment efficacy when being tested in a randomized phase III trial combined with the standard of care [
211]. Other antigens that were examined include peptides derived from human
telomerase 1 (
TERT1) and GVAX, a vaccine comprised of autologous or allogeneic tumor cells expressing the dendritic cell-stimulating cytokine
GM-CSF [
212,
213]. Unfortunately, none of these vaccines achieved convincing clinical results. In principle, common PDAC driver mutations, such as
KRASG12D, can harbor tumor-specific, T cell epitopes [
214]. An ongoing phase II trial first predicts such neo-antigens using exome-sequencing of tumor biopsies, followed by production of personalized dendritic cell vaccines loaded with the respective epitopes (NCT03300843) [
215]. Whether this strategy turns out to be successful needs to be awaited. Overall, several vaccination approaches could successfully elicit measurable anti-tumor T cell responses, yet so far none of these strategies resulted in clear clinical benefits [
216].
Antigen-independent immunostimulatory therapies aim at the activation of antigen-presenting cells. Diverse receptor-ligand-axes have been explored in this regard. As such, treatment with agonistic anti-CD40 antibodies is well known to activate antigen-presenting cells and to polarize macrophages towards the pro-inflammatory M1-like state [
217,
218]. However, clinical evaluation of this strategy in PDAC patients disclosed only short-term responses, and no long-term anti-tumor immunity was observed [
219]. Nevertheless, CD40 stimulation in combination with chemotherapy and immune checkpoint blockade is currently under clinical investigation in a phase I/II trial (NCT03214250). Complementary approaches to achieve activation of antigen-presenting cells involve ligand-dependent stimulation of
pattern recognition receptors (PRRs) [
220]. Indeed, agonists of
toll-like receptors (TLRs),
RIG-I-like helicases (RLHs), and the
stimulator of interferon genes (
STING) revealed encouraging results in preclinical PDAC models [
221‐
223], but their clinical potential remains to be elucidated.
Bypassing the
in situ steps of T cell priming by antigen-presenting cells, adoptive transfer of T cells carrying chimeric antigen receptors (CARs) has proven powerful clinical performance in B-cell malignancies [
224]. CAR T cells recognize specific cancer cell surface antigens through a single-chain variable fragment (scFv) whose ligation stimulates T cell activation via the intracellular domains of the CAR construct, resulting in efficient T cell-mediated killing of the target cell [
225]. PDAC exhibits several tumor-specific antigens, such as
carcinoembryonic antigen (CEA),
mesothelin (MSLN), and
mucin 1 (MUC1), which are promising determinants for CAR T cell therapy [
226,
227]. However, for solid cancer entities, intra-tumoral recruitment and trafficking of CAR T cells as well as the commonly observed immunosuppressive tumor microenvironment appear to be major challenges. Intelligent combinations, thus, are needed in order to overcome these obstacles.
A cardinal feature of the immunosuppressive PDAC microenvironment is its massive stromal content and the excessive deposition of extracellular matrix, including hyaluronan [
72]. Early phase clinical trials combining recombinant human
hyaluronidase 20 (
rHuPH20) with gemcitabine and nab-paclitaxel revealed promising results, particularly in those patients whose tumors were characterized by high levels of hyaluronan [
228]. Reporting of the HALO-109-301 phase III trial (NCT02715804) is awaited in order to fully assess the clinical performance of this approach [
229]. Inhibition of
FAK1, a tyrosine kinase involved in the process of CAF generation, constitutes another approach to interfere with stromal function in PDAC, and pharmacological
FAK1 inhibition eventually rendered preclinical PDAC model systems more susceptible to T cell immunotherapy and immune checkpoint inhibition [
73]. Other studies showed that genetic ablation or inhibition of
FAK1 also increases PDAC responsiveness to gemcitabine and nab-paclitaxel [
230,
231]. In rather strong contrast, genetic deletion of stromal myofibroblasts in PDAC mouse models led to disease exacerbation and diminished animal survival due to enhanced regulatory T cell-mediated immunosuppression, clearly calling for caution when targeting components of PDAC stroma [
78].
On a cellular level, massive infiltration by myeloid cells, such as MDSCs, and resulting exclusion of CD8
+ T cells are major hallmarks of the immunosuppressive PDAC microenvironment [
86,
232]. Several myeloid cell-targeting approaches have been investigated in recent years in order to overcome these mechanisms of immunosuppression [
82,
233,
234].
Chemokine receptor 2 (CCR2), for instance, is known to contribute to the infiltration of pancreatic tumors by monocytes and macrophages, and this is associated with reduced patient survival and poor outcome [
235]. Strikingly, the combination of
CCR2 blockade and gemcitabine/nab-paclitaxel chemotherapy showed promising results in phase I trials [
85,
236]. However, the follow-up phase Ib/II trial (NCT02732938) was discontinued due to strategic considerations, and instead phase I/II trials with combined modality approaches of CCR2 blockade in conjunction with pre-operative SBRT and immune checkpoint inhibition were recently initiated (NCT03778879, NCT03767582). Another target that regulates the function of macrophages and MDSCs in PDAC is
M-CSF. Preclinical data suggest that
M-CSF blockade can indeed reprogram macrophages and thus, synergize with immune checkpoint inhibition, but the clinical potential of this strategy remains to be examined [
237].
In summary, (re-)activating anti-PDAC immunity in order to improve the overall clinical outcome appears clearly more challenging than extrapolated experiences from other cancer entities have suggested. Probably the most promising strategies would incorporate combinations of different immunotherapeutic approaches and/or combinations with other (classical) treatment modalities, such as chemotherapy and/or radiotherapy [
238].