CORDIS
EU research results

CORDIS

English EN

New approaches for the targeted therapy of Non-Small Cell Lung Cancer

Final Report Summary - LUNGTARGET (New approaches for the targeted therapy of Non-Small Cell Lung Cancer)

Executive Summary:
Non-Small Cell Lung Cancer (NSCLC) is a major health problem in Europe. Due to poor treatment opportunities, 350000 NSCLC patients die each year. Hence, it is clear that we need better molecular insights in the disease to provide novel therapeutic targets. 10% of NSCLC patients harbor mutations in the EGFR, and this has given an opportunity to treat these patients with the EGFR inhibitor erlotinib. Unfortunately, these patients soon develop resistance to the drug, and hence it is expected that combination treatment targeting the EGFR together with novel, yet to be identified targets, would be a more powerful therapeutic approach. Another 30% of patients harbor KRAS mutations, and unfortunately for these patients there are no targeted therapy options yet.
The LUNGTARGET consortium aims to first identify novel therapeutic targets for EGFR-mutant or KRAS mutant NSCLC, via a combination of several approaches. These novel targets are then evaluated for therapeutic relevance in the mouse NSCLC models (EGFR mutant of KRAS mutant) available in the consortium. Furthermore, the relevance of these targets will also be tested in patient tumor tissue material. Based on the combined outcome of these analyses, and also in view of already previoulsy obtained experimental data, we are also very actively developing novel therapeutics against various targets involved in NSCLC. Thus, the project has 3 levels of activities: a) discovery of novel therapeutic targets for NSCLC, b) validation of these targets in mice and patient material, and c) development of novel therapeutics.
The Lungtarget results within the above mentioned 3 levels of activities can be summarized as follows.
A) Discovery of novel therapeutic targets for NSCLC.
RNAi screening yielded the following novel targets: 1) The GATA2 transcriptional network, which can be targeted by combined NfkB and ROCK inhibition in KRAS driven NSCLC, 2) Loss of NF1 in EGFR resistant cells, which can be successfully treated with a combination of EGFR and MEK inhibitors. Secretome profiling showed higher secretion of tPA in resistance to EGFR inhibition, and higher AREG secretion in KRAS mutant cells, which is inducible by cisplatin. Kinome profiling revealed several protein kinases that are overexpressed and/or overactive in NSCLC, such as S6K2, SFK’s, MARK4, RSK4, DDR1
B) Validation of targets in mice and patient material.
We have developed a lentiviral system for faster validation of potential therapeutic targets and we validated several novel therapeutic approaches for KRAS mutant NSCLC via targeting of molecules such as DDR1 and MARK4. We also developed several effective therapeutic strategies for erlotinib resistant NSCLC, including a triple combination of antibodies that targets the entire EGFR family, restoration of tumor glutathione levels etc. We studied patient NSCLC tissue microarrays for expression of our various targets, and identified several clear correlations with clinical parameters such as survival, response to therapy etc.
C) Development of novel therapeutics.
We have developed several very promising novel therapeutics, such as antibodies blocking the EGFR ligand amphiregulin (AREG), c-Met targeting DARPins, novel EGFR, CDK, and RSK4 kinase inhibitors. We have also designed novel synergistic therapeutic combinations with our own novel drugs (or with existing drugs), e.g. the combination of AREG antibodies and cisplatin, the triple combination of EGFR family antibodies with new generation tyrosine kinase inhibitors.
Overall, the findings of the Lungtarget project will have a high therapeutic impact for both KRAS, as well as EGFR mutant NSCLC.

Project Context and Objectives:
Lung cancer is the most common cancer fatality in Europe causing about 335,000 deaths in 2006 (1). Non-small cell lung cancer (NSCLC) accounts for about 80-85% of cases, with 5 year survival < 15% (2). The survival rates are poor because most patients present with disease that is metastatic and too advanced for the possibility of curative surgical resection. Even in those who undergo surgery, cure is not guaranteed. Consequently there is an urgent need to improve the outcome for patients with NSCLC.
Currently, for advanced disease, chemotherapy with platinum (Pt)-based doublet regimens has been the standard of care for many years with or without additional radiotherapy (3). Recent data show that patients with tumours bearing activating EGFR mutations (10% of patients) respond better to the EGFR inhibitor gefitinib than to standard Pt-based chemotherapy. In contrast, for KRAS mutant NSCLC (30% of patients), there currently are no options for targeted therapy.
Current experience with targeted therapies clearly indicates that, even for the most promising therapies, resistance will in due course develop. It is therefore important to develop strategies to overcome resistance to these agents and this could be achieved through combinatorial approaches targeting multiple proteins at the same time.
The Lungtarget project has 3 levels of activities: a) discovery of novel therapeutic targets for NSCLC, b) validation of these targets in mice and patient material, and c) development of novel therapeutics. The Lungtarget objectives within these 3 levels of activities can be summarized as follows.
A) Discovery of novel therapeutic targets for NSCLC. This includes:
• Genome wide siRNA screening for the identification of targets for the treatment of EGFR mutant or KRAS mutant NSCLC. Here we want to find genes whose knockdown is capable of killing cells harbaouring mutant versions of EGFR or KRAS, thereby identifying proteins that may become targets for therapeutic intervantion.
• Profiling of the proteins secreted by EGFR and KRAS mutant NSCLC (i.e. secretome profiling). It is clear that cancer cells secrete various factors with autocrine and paracrine effects. We want to identify these factors and understand their effect on NSCLC behaviour. These factors will also be evaluated as potential therapeutic targets
• Profiling of the protein kinases that are activiated by EGFR and KRAS mutant NSCLC (i.e. kinome profiling). Protein kinases have proven to be excellent targets for the treatment a variety of cancer types (e.g. erlotinib for the treatment of EGFR mutant NSCLC). We want to identify additional kinases that are important for the regulation of NSCLC growth, apoptosis, cell motilty, invasion, metastasis, again with the ultimate aim of finding potential novel drug targets.
B) Validation of targets in mice and patient material. This includes:
• Evaluation of the therapeutic relevance of selected protein kinases in mouse models for the treatment of NSCLC. We want to find out whether genetic ablation of these kinases (or chemical inhibition in case drugs are already available) can inhibit NSCLC in mouse lung cancer models.
• Evaluation of the therapeutic relevance of the identified targets in mouse models for the treatment of NSCLC. Not all identified targets are kinases, they can also be other genes (including secreted factors). Our aim here is to develop a generally applicable (i.e. applicable to any gene to be tested) method for the fast evaluation of therapeutic potential of a target. This is done via the development of a novel lentiviral RNA interference vector that can be used for easy knockdown of genes in cells or mice, which is much faster than classical genetic knockout experiments.
• Development of mouse NSCLC cell lines and tissue arrays, to enable easier initial validation of targets. Again, in view of the objective above and to speed up target validation, we intend to create mouse NSCLC cell lines derived from our corresponding mouse tumor models. These can be used for much faster initial target validation experiments, after which further tests can be carried out in mice
• Evaluation of target expression and activity in patient tumor tissue and serum samples and correlation of target expression and activity with disease outcome for patients. It is obvious that conclusions about the vaildity of a potential therapeutic target cannot be based on cell and animal experimenta alone. For that reason, we use our NSCLC tumor microarray to detect the degree of expression (protein and/or mRNA) of the various targets that we identify. Since we also have the corresponding clinical data, we can also investigate whether there is a correlation between target expression and clinical parameters such as response to therapy, overall survival, tumor progression etc.
C) Development of novel therapeutics. The information gained in this project (as well as already existing information) on therapeutic targets is used foro the rational development of novel drugs and novel drug combinations. This includes
• Development of novel antibodies against various members of the broader family of EGFR ligands. NSCLC tumors often hypersecrete certain EGFR ligands, and we want to block these to improve therapeutic outcome.
• Development of DARPins targeting various NSCLC growth factor receptors (such as EGFR, cMET…). DARPins are small recombinant proteins that can serve as an alternative to antibodies, with some possibly very interesting advantages such as easy labeling, excellent pharmacokinetics, better tissue penetration, easy combination of several DARPins within one bigger DARPin molecule etc. We selected the EGFR and the cMET proteins to assess the therapeutic potential of DARPins in NSCLC.
• Design of synergism between the various developed molecules. We are very well aware that future modus of targeted therapy will be a combinatorial one (by analogy with the treatment of HIV infection, another example of a highly mutationally active system). Therefore, various modes to combine our novel therapeutic molecules with other molecules (including already existing drugs that can be repurposed for use in NSCLC) are tested for their effect on NSCLC.
• Development of protein kinase inhibitors against NSCLC

Taken together, the objective of the project is thus to identify novel targets, validate them in mouse models and patient materials, and develop novel therapeutics based on this knowledge. Interestingly, the various partners were already active at various stages of this discovery chain, which allowed us to start the project also at later stages (e.g. at the validation of already identified targets or at the development of drug molecules against already validated targets). This approach obviously strongly increased our chances of success.

Project Results:

4.1.3.1 WP1 RNAi based identification of targets for the treatment of EGFR- or KRAS-mutant NSCLC
We reasoned that genome wide screening for targets that selectively kill EGFR-mutant or KRAS mutant NSCLC cells would give us an excellent insight into novel potential therapeutic targets. Here we exploited the principle of synthetic lethality. In this phenomenon, one mutation may be very beneficial for a cell (e.g. oncogene mutation for a cancer cell) but combination of this mutation with inhibition of a particular pathway on which an oncogene mutated cell (in contrast to healthy cells) depends, is lethal for the mutated cell. Such targets are obviously extremely interesting from a therapeutic point of view, since their inhibition leaves healthy cells intact.

4.1.3.1.1 The GATA2 transcriptional network is a requisite for KRAS oncogene driven NSCLC
We performed a genome wide screen to identify candiate genes that are essential for the survival of cells bearing an oncogenic KRAS allele, and found that GATA2 is crucialy needed for the survival of mutant KRAS cells. GATA2 is a transcription factor, and as such, transcription factors as such are difficult therapeutic targets. Therefore, we investigated the network of signaling proteins that are dependent on GATA2, and we discovered that many genes controlled by GATA2 can be categorized as part of the proteasomal system, and others belong to the Rho-Rho kinase pathway. This opened therapeutic possibilities for treating KRAS mutant cancer via the combined administration of bortezumib (a proteasome inhibitor) and fasudil (a Rho kinase inhibitor). For further information on this, see WP3 where we did indeed validate this therapeutic approach in mouse KRAS mutant NSCLC model systems.

4.1.3.1.2 Loss of NF1 (which activates the Ras-Raf-MEK-ERK pathway) confers resistance to EGFR targeting therapy
We used siRNA to screen for cell killing in EGFR mutant cells, without killing EGFR wildtype cells. We found that silencing of 7 genes (CLSPN, CSPG6, FGF4, G6PC3, HGS, NFKBIIB and UNG) selectively reduce the survival of EGFR mutant cells. We also found that silencing the Ras GTPase activating protein NF1 conferred resistance to EGFR inhibitors. The consequence of loss of NF1, a Ras-GAP protein, is Ras hyperactivation and thus constitutive activation of the Ras-Raf-MEK-ERK cascade leading to uncontrolled proliferation. As a consequence, cancer cells who lose NF1 become resistant to EGFR inhibition. We could indeed confirm that loss of NF1 is a cause of erlotinib therapy resistance in a subgroup of NSCLC patients. An important conclusion is that hese patients might benefit from combination therapy with EGFR and MEK inhibitors, an approach that we validated in WP3.

4.1.3.2 WP2 Target identification via secretome and kinome profiling in EGFR-and K-Ras mutant NSCLC

We reasoned that a better knowledge of the proteins that are secreted by NSCLC, and of the kinase signaling pathways that are activated by NSCLC, may yield important novel insight for diagnosis and therapy of NSCLC. Indeed, knowledge of secreted factors that may affect NSCLC in an autocrine or paracrine fashion may provide clues for therapies designed to block the stimulatory effects of such secreted factors on NSCLC growth. Furthermore, better insights in the “wiring” of kinase signaling pathways may reveal signaling nodes that could become targets for therapeutic inhibition.

4.1.3.2.1 Secretome profiling of NSCLC

4.1.3.2.1.1 The EGFR ligand Amphiregulin (AREG) is overexpressed and secreted by NSCLC in response to cisplatin
Due to the established role of growth factors in cancer progression, we determined the secretion repertoire of some of the EGF-like ligands in lung cancer. Analysis of the conditioned media of 7 lung cancer cell lines revealed AREG as a common trait though no other ligand, with the exception of the secretion of HB-EGF by the cell line H522, was detected. To ensure that our results were valid, we performed a similar analysis on pleural effusions from lung cancer patients. Interestingly, the major growth factor detected in pleural effusions from patients was also AREG; 16 of 18 samples contained relatively high concentrations of it. Taken together, these results identify AREG, as the most prevalent EGF-like growth factor in lung cancer.
We have then investigated the functional consequences of this hypersecretion of AREG for NSCLC cells. This led to several important conclusions:
- AREG is less efficient at receptor downregulation and degradation than EGF or TGF-alpha. This means that signaling stimulated by AREG persists much longer, and that AREG is also much slower cleared from the extracellular space.
- AREG induces relatively weak ubiquitination of EGFR. Thus, unlike high-affinity EGFR ligands, AREG fails to induce receptor endocytosis and subsequent targeting for destruction via ubiquitination. As a consequence signaling desensitization is weakened, potentially explaining why advanced lung tumors frequently secrete this factor.
- AREG expression is increased after treatment of cancer cells with the chemotherapeutic drug cisplatin. Platinum-based antitumor agents have been the mainstay of KRAS mutant NSCLC chemotherapy, but various mechanisms of resistance limit therapeutic efficacy. Because the majority of body fluids we analyzed were derived from patients treated with chemotherapy, we examined the possibility that high level secretion of AREG by lung cancer cells is attributable to patient exposure to chemotherapeutic agents, such as cisplatin, and this was indeed the case. Therefore, we propose that the response to DNA adducts induced by platin analogs might specifically increase expression and secretion of AREG by lung cancer cells, potentially contributing to adaptive drug resistance.

4.1.3.2.1.2 Erlotinib resistant NSCLC hypersecretes tissue plasminogen activator (tPA).
We have also exploited an unbiased mass spectrometric method to identify protein factors that are secreted by NSCLC cells. Via stable isotope labeling (SILAC) followed by concentration of the tissue culture supernatants and subsequent mass spectrometry, we identified several proteins that are differentially secreted between erlotinib sensitive vs resistant cells (using isogenic clones PC9, which is erlotinib sensitive, and its derivative PC9-ER, which is erlotinib sensitive). One of the top hits in these experiments was tPA (sissue plasminogen activator). tPA is an interesting protein since this serine protease has been described to promote tumor angiogenesis, proliferation and invasion of various tumor cell lines. Thus, tPA could be both, a novel biomarker for the assessment of Erlotinib resistance in patients to be used to monitor drug-induced tumor evolution and potentially a novel target for Erlotinib-resistant NSCLC. To gain further insights into the proteins affected t-PA and its interactions with proteins secreted in PC9 and PC9ER cells, we have performed cross-linking mass spectrometry experiments. A number of candidate proteins have been identified for PC9 and PC9ER cells as potential t-PA interaction partners that are currently verified. These are: PLAT, ACTN1, FUCA2, TMEM132A, ACTN4, MAMDC2 and PAFAH183. We were also interested in the microenvironment conditions that modulate tPA secretion. Since hypoxia is a frequent event in tumors we examined as planned the effect of hypoxia on tPA secretion. Our data suggest that hypoxia may indeed promote tPA secretion in PC9 and PC9ER cells.

4.1.3.2.2 Kinome profiling of NSCLC

4.1.3.2.2.1 Antibody mediated profiling of kinase expression and activation status in NSCLC
Our results above have revealed that PKC isoforms and the mTOR pathway are deregulated in erlotinib-resistant NSCLC cell lines. We have also shown that targeting the class II PI3K isoform PI3KC2 by selective pharmacological inhibitors or downregulation by siRNA impairs the proliferation of NSCLC cell lines. In this new study, we have shown that the activation of the mTOR pathway is impaired upon PI3KC2 inhibition or silencing, which was not accompanied by decreased Akt activation in NSCLC cell lines. Targeting protein kinase C (PKC) also impaired the activation of the mTOR pathway, while co-targeting PI3KC2 and PKC resulted in enhanced effects on cell proliferation and downstream signaling in NSCLC cell lines. Conversely, ectopic over-expression of PI3KC2 enhanced mTOR pathway activation and promoted resistance to epidermal growth factor receptor (EGFR) inhibitors. A complex between PI3KC2 and PKC was identified in lung cancer cell lines and PI3KC2 was phosphorylated by PKC Together our data have uncovered the role of a novel PI3KC2/PKC complex in the activation of the mTOR pathway and drug resistance in NSCLC cells (Borgström et al., manuscript submitted).

4.1.3.2.2.2 Development of a method for the mass spectrometric profiling of protein kinase activation status
The analysis of kinase activation in cells, in particular if the purpose is to gain information about as many kinases as possible (i.e. kinome wide) can be quite cumbersome if classical methods using kinase activation recognizing antibodies are used. Therefore, we decided to design a kinae affinity matrix, consisting of broad spectrum kinase inhibitors covalently linked to beads, to capture as many kinases as possible from NSCLC cells. Upon proteolysis, kinase derived phosphopeptides were then analysed mass spectrometrically, to follow the activation status of the corresponding kinases (since the majority of kinases have well known phosphorylation sites that are diagnostic for their activation). We noticed that, although a high number of kinases could be detected mass spectrometrically, a considerable number of expected kinases were not detected. We concluded that this was due to undersampling of the peptides by the mass spectrometer. Indeed, the delivery of the huge amount of peptides present in the tryptic digest of the kinases may easily lead to preferential identification of the only most abundant or the most effectively ionised kinase peptides. Hence, we decided to adapt our kinase identification strategy. We reasoned that if we would focus our mass spectrometric detections solely on phosphpeptides that are “diagnostic” for kinase activation, our chances of detecting a broader range of kinases would increase significantly. Hence, we performed an extensive data mining of the literature to compile a list of kinases annotated with the corresponding peptides that are diagnostic for their activation. This were often multiple peptides per kinase. To determinde the peptides that are most closely associated with kinase activation, we then performed a datamining of mass spectrometry based databases of proteomics experiments, to determine which activated kinase peptides occur most frequently across a broad variety of experimental conditions. Based on all these data, we then compiled a phosphopeptide list, which is now ready for use for kinome wide mass spectrometric evaluation of kinase activation via MRM (multiple reaction monitoring).

4.1.3.2.2.3 Src family kinases are frequently overexpressed in NSCLC and their simultaneous targeting strongly inhibits the growth of NSCLC
We detected an overexpression and overactivity of several Src family tyrisine kinases in NSCLC. These mass spec, in vitro, and immunohistochemistry expression data prompted us to investigate the impact of downregulating SRC, LYN, FYN and YES on the growth of NSCLC cell lines. The growth inhibition obtained with the silencing of individual SFK isoforms was satisfactory for some NSCLC cell lines but poor for many other NSCLC llines, implying that the simultaneous targeting of multiple SFKs may be required to achieve an appreciable biological effect. Hence, siRNAs targeting SRC, LYN, FYN and YES were combined, leading to a substantial cell growth inhibition far exceeding that achieved by silencing individual SFK isoforms. These results have prompted us to investigate in WP2 whether dasatinib, a SRC family tyrosine kinase inhibitor, can inhibit the growth of NSCLC cells, and this was indeed the case.

4.1.3.2.2.4 S6K2 plays a crucial role in the chemoresistance induced by FGF-2, via the induction of anti-apoptotic protein expression
We found that S6K2 kinase activity is increased in therapy resistant NSCLC, and that this increase is induced by the growh factor FGF-2. We found that S6K2 causes an increased cap-independent translation of anti-apoptotic proteins and that is involved in the development of drug resistance in lung cancer. To better understand the mechanism of this increased translation, we set out to identify proteins that bind to S6K2 in NSCLC. We found that S6K2 binds and phosphorylates hnRNPA1 on novel Ser4/6 sites, increasing its association with mRNAs of the anti-apoptotic proteins BCL-XL and XIAP mRNAs, and that this promotes their nuclear export. In the cytoplasm, phosphoS4/6-hnRNPA1 dissociates from these mRNAs de-repressing their IRES-mediated translation. This correlates with the phosphorylation-dependent association of hnRNPA1 with 14-3-3 leading to hnRNPA1 sumoylation on K183 and its re-import into the nucleus. A non-phosphorylatible, S4/6A mutant prevented these processes, hindering the pro-survival activity of FGF-2/S6K2 signalling. Thus, phosphorylation on novel N-term sites of hnRNPA1 promotes translation of anti-apoptotic proteins and is indispensable for the pro-survival effects of FGF-2.

4.1.3.2.2.5 MARK4 is required for NSCLC migration and is more expressed in metastases
MARK4 is another kinase that is overexpressed in NSCLC, and when we inspected clinical samples, we found indications that it is more expressed in metastases. Therefore, we were interested whether this kinase would be involved in the regulation of cell migration. Transfection of NSCLC cell lines with MARK4-targeting siRNAs led to an increase in the cell area and clustering as can be observed following tubulin staining. The change in cell area was accompanied by a loss of cell polarity. We therefore proceeded to study possible associated changes in cell motility. Random migration (RM) and boyden chamber directional migration (DM) assays showed that MARK4 knockdown strongly inhibited cell migration. Therefore, MARK4 could be a potentially interesting target to limit further spread of NSCLC, an idea that we further tested in WP3.

4.1.3.2.2.6 RSK4 is needed for NSCLC cell migration and invasion
We also found overexpression of RSK4. We performed a phenotypic analysis of RSK4 knockdown and found that RSK4 knockdown strongly inhibits cell motility. Furthermore, knockdown of RSK4 also inhibited invasion of NSCLC cells in a collagen matrix. Taken together these experiments warranted further investigation of RSK4 as a potential target for the inhibition of tumor metastasis. For this reason, we have also started a RSK4 inhibitor program in WP5.

4.1.3.3 WP3 Evaluation of therapeutic relevance of identified targets in mouse models for treatment of NSCLC
Experiments in cell lines provide a straightforward access to a wealth of information about potential targets for the treatment of NSCLC, but the proof of their relevance requires thorough validation in mouse NSCLC models. For the purposes of this project, we use two excellent mouse models: one for NSCLC driven by mutation of the EGFR, and one driven by mutant KRAS. It is clear that genetic ablation of potential targets in such models provides strong indication that it may be worthwile to target these proteins for therapeutic purposes.
The knowledge on genes to be targeted in this WP came from several sources:
- genes identified in WP1 or identified in previous work of the partners
- kinases identified in WP2 or identified in previous work of the partners
- secreted proteins identified in WP2 or identified in previous work of the partners

4.1.3.3.1 Validation of therapeutic targets in mouse NSCLC models

4.1.3.3.1.1 Inhibition of c-Raf inhibits tumor growth in KRAS driven NSCLC mouse models
We identified the MAPK pathway as an essential mediator of K-RasG12V oncogenic signaling for the initiation of NSCLC in mouse. Therefore, we set up mouse experiments to validate the various components of the MAPK pathway as potential therapeutic targets.
We developed an adenovirus (Adeno-FLPase) based system that, upon inhalation by mice bearing NSCLC, allows for the selective and conditional Cre/Lox mediated knockout of genes in already established mouse tumors. In comparison with traditional experiments with mouse xenografts of human tumors, this much better mimicks the situation occurring in NSCLC patients for several reasons:
1) the tumors are located in situ (the lungs) and originated from the mice themselves
2) we look at the growth inhibitory effects of knocking out genes in already established tumors (as opposed to looking at growth inhibition of newly implanted xenografts of cell lines that were first infected in culture with lentiviral knockout constructs). It is clear that cancer therapies must be capable of dealing with already established tumors, since by definition a tumor is discovered in a patient when it is already established.
Using this Adeno FLPase system we found that c-Raf is a very promising therapeutic target and that not much therapeutic benefit is gained when B-Raf and/or A-Raf are also simultaneously knocked out. We are aware that genetic ablation of a putative druggable target may not faithfully recapitulate the effects of its chemical inhibition. Thus, in order to more faithfully mimic pharmacologic intervention we have generated catalytically inactive inducible alleles of C-Raf. Our results indicate that inhibition of the catalytic activity of C-Raf reduces tumour burden.

4.1.3.3.1.2 Blocking the mutant KRAS driven GATA2 transcriptional network abolishes NSCLC tumor growth in mice.
We have shown that conditional knockout of GATA2 is reqquired for KRAS driven tumorigenesis. Even more importantly, knockout of GATA2 in already establiushed KRAS driven mouse tumors induced a near complete abrogation of these tumors. In these regressing tumors, GATA2 was completely lost after 3 days and we found that the proteasome and Rho signaling patways were defective (in line with the involvement of the GATA2 transcriptional network in these systems). Combined inhibition of the GATA2 dependent proteasome (with the drug bortezomib) and Rho signaling (with the drug fasudil) pathways caused regression of established KRAS induced tumors, thereby demonstrating that this combination therapy holds promise for treatment of KRAS driven NSCLC.

4.1.3.3.1.3 The interaction between KRAS and PI3-kinase is a suitable target for KRAS driven NSCLC
We have shown that inducible expression of a PI3 kinase gene with a defective Ras binding domain causes regression of established KRAS driven NSCLC in mice. These results suggest that the interaction surface between Ras and PI3-kinase is a potential target for the development of drugs that block this protein-protein interaction. However, in this model system we observed only a partial regression (in contrast to the strong results observed with targeting GATA2). Although this could reflect redundancy between p110 isoforms, it might also be the case that coordinate inhibition of the MEK pathways downstream of RAS is also required. While this has a marked impact on the Pik3ca wild type tumour, effects on the Pik3ca RBD mutant tumours are only somewhat improved on vehicle controls. It is possible that the more modest effects on tumour regression of targeting RAS regulated signaling pathways seen here compared to targeting a non-oncogene addiction network such as GATA2 might represent a paradigm for therapeutic approaches to RAS mutant cancers in the future. Most likely, we speculate that targeting of both oncogene addiction and non-oncogene addiction networks simultaneously will be needed for the most effective treatment of KRAS mutant lung cancer in patients.

4.1.3.3.1.4 Neurofibromin 1 (NF1)-deficient, erlotinib resistant NSCLC can be resensitized to erlotinib via MEK inhibition
As mentioned above, downregulation of the RasGAP NF1 causes resistance to erlotinib in cellular NSCLC models, and reduced NF1 expression is also observed in a subpopulation of erlotinib resistant patient tumor samples. Long-term intermittent erlotinib treatment of mice bearing EGFR driven NSCLCleads to the outgrowth of resistant tumor. o assess whether the genes identified in our siRNA screen showed altered expression between untreated (erlotinib-sensitive) tumors and erlotinib-resistant tumors, we compared the expression levels of Nf1 in erlotinib-resistant tumors and corresponding adjacent normal lung using quantitative real-time PCR (qRT-PCR) analysis. These experiments revealed a decrease in Nf1 mRNA levels compared with normal lung in 10 of 18 erlotinib-resistant tumors. nterestingly, tumors bearing the EGFRT790M gatekeeper mutation, Kras mutations, or Met amplification did not show decreased Nf1 expression, suggesting that loss of neurofibromin could be selected for by EGFR TKIs in the absence of other mechanisms of resistance. NSCLC cells with reduced NF1 expression have a higher Ras-Raf-MEK-ERK signaling due to absence of the RasGAP activity of NF1. Therefore, treated NF1 deficient NSCLC tumors with erlotinib and a MEK inhibitor, and this combined treatment had a striking effect on most erlotinib-resistant, NF1 deficient mouse lung adenocarcinomas

4.1.3.3.1.5 Knockdown of the MARK4 kinase impairs metastasis in a mouse lung cancer model
We found that the expression of the MARK4 kinase is regulated by the miRNA miR-515-5p. To evaluate whether changes to the expression of miR-515-5p directly modulated the metastatic potential of breast and lung cancer cells, we performed animal experiments. we wished to examine whether directly changing levels of this miRNA could modulate tumor cells colonization in vivo. This analysis revealed that while cells transfected with a non-targeting miRNA mimetic (miR-NC) efficiently colonized the mice, miR-515-5p-transfected cells failed to establish tumors in the injected animals Hence, reduction of MARK4 levels is sufficient to modulate the metastatic potential of cancer cells.

4.1.3.3.1.6 DDR1 kinase is expressed in aggressive KRAS mutant subclones and is an excellent target
Solid tumors accumulate alterations during progression generating sub-clones with different genetic/epigenetic profile. This intra-tumor heterogeneity represents a major obstacle for the identification of targets whose inactivation could block the progression of the entire tumor. To circumvent this difficulty, it has been proposed that future treatments should target events that may occur early in tumor evolution. Following this rationale, we hypothesized that the identification of novel mediators of K-Ras signaling in early murine hyperplasias might unveil targets required at advanced tumor stages. Transcriptional profiling of K-RasG12V-driven hyperplasias revealed inter-tumor heterogeneity with a subset exhibiting an aggressive transcriptional profile analogous to that of advanced human adenocarcinomas. This signature identified the tyrosine kinase receptor Ddr1 as the top score. Genetic and pharmacological inhibition of Ddr1 in mice blocked progression of K-RasG12V-p53 proficient adenocarcinomas but not in the presence of mutant p53, But as we will see in section …, concomitant inhibition of Ddr1 and Notch signaling, a downstream mediator of Ddr1, thwarted progression of murine K-RasG12V;p53-null adenocarcinomas.

4.1.3.3.2 Experimental combination therapy in mouse in vivo and in vitro NSCLC models

4.1.3.3.2.1 A triple combination of antibodies against EGFR, HER2 and HER3 is highly effective against erlotinib resistant NSCLC
Our in vitro experiments for the design of synergy between EGFR family targeting antibodies in WP5 were very promising. In particular, since the triple mAb combination spares normal cells while it strongly inhibits tumor cells and possibly deprives them of stromal support. This prompted us to examine the triple combination in tumor-bearing animals. In line with our cell culture results, the EGFR-specific mAb partly inhibited tumor growth when singly applied, but neither HER2-specific nor HER3-specific mAb, nor their combination, showed inhibitory activity in mice. Although the paired combinations were partially inhibitory, only the triple mAb combination exerted strong and lasting inhibitory effects on tumor growth as well as on animal survival. Regardless of the exact mechanism of their action, the triple combination of mAbs we characterized in vitro and in animals offers a feasible pharmacological option for treating numerous lung cancer patients who inevitably develop resistance to the currently applied inhibitors of EGFR. Because mAbs targeting EGFR (like cetuximab) or HER2 (like trastuzumab) are already approved for clinical use and HER3-specific mAbs similar to NG33 (such as MM-121) are currently in clinical trials, our approach presents a realistic strategy for overcoming drug resistance.

4.1.3.3.2.2 Combined inhibition of Ddr1 and Notch signaling is an effective therapeutic strategy for human K-RAS driven lung adenocarcinoma
We treated mice with an oral Ddr1 tyrosine kinase inhibitor called 7rh and followed caspase 3 activation in the tumors, since Ddr1 exerts a pro-survival function. In colon cancer cells, Ddr1 exerts pro-survival functions through activation of Notch1. Moreover, Ras oncogenes increase expression and activity of Notch1 in a process essential to maintain the Ras-transformed phenotype. This consisted for us the rationale to investigate whether a combined inhibition of Ddr1 and Notch would have therapeutic benefit vor KRAS mutant NSCLC in mice. We treated tumor-bearing KRASmutant mice with the γ-secretase inhibitor LY-411575, a potent Notch inhibitor, alone or combined with 7rh. The effect of 7rh+LY-411575 was additive, leading to a higher level of apoptosis than with either agent alone. Interestingly, this effect was also observed in p53 deficient tumors, wich is very encouraging since these are more aggressive tumors. We then investigated how this combined therapy compared with standard of care chemotherapy. To further assess the therapeutic value of the combined treatment we performed a comparative study with cisplatin/paclitaxel chemotherapy, the standard of care for human patients. We found that cohorts of mice with KRAS driven tumors (either p53 or p53 deficient) displayed a higher apoptotic response to 7rh+LY-411575 inhibitors than to cisplatin/paclitaxel. To validate our findings in a clinically relevant setting we used lung orthotopic PDX derived from three independent human K-RAS mutant lung adenocarcinomas carrying different K-RAS codon 12 mutations accompanied by either p53 inactivation or PI3K activating mutations. We selected Dasatinib, a potent DDR1 inhibitor38, and Demcizumab, a humanized monoclonal antibody targeting the NOTCH ligand Delta-Like Ligand 4 (DLL4), which is currently in Phase II clinical trials for NSCLC. While both combined treatment and standard chemotherapy significantly prolonged survival, DDR1/NOTCH inhibition showed superior efficacy over chemotherapy as assessed by decreased uptake of 18F-fluoro-2-deoxy-glucose PET (FDG-PET) and reduced residual tumor burden irrespective of p53 status

4.1.3.3.2.3 Restoration of glutathione levels with ethacrynic acid causes re-sensitisation of erlotinib resistant tumors
Following our successful design of synergy between ethacrynic acid and erlotinib in in vitro models of erlotinib resistant NSCLC (see WP5), we wanted to test whether such an approach would also be valid in mouse models of erlotinib resistant NSCLC. Our in vitro findings could indeed be translated in an in vivo xenograft model, where increasing intratmoural GSH levels lead to resensitisation of resistant tumours to erlotinib. Thus, co-administration of EA may be a viable strategy for the management of erlotinib-resistant disease in humans, in particular since ethacrynic acid is an approved loop diuretic. These findings also stress the increasing potential of repurposing older drugs (that were previously used in other therapeutic fields) for use in the oncology therapeutic area.

4.1.3.3.3 Development of novel tools for the validation of drug targets in mouse NSCLC models

4.1.3.3.3.1 Development tissue micro-arrays, tumour cell lines and 3D cultures derived from K-Ras or EGFR driven tumours
We established tissue microarrays, cell lines and 3D cultures containing mutated EGFR and/or K-Ras oncogenes, so we can pre-screen in in vitro models for the therapeutic activity of therapeutic candidates such as Mabs, DARPins, kinase inhibitors. These reagents are also very valuable to validate newly identified targets.
We have succeeded in establishing permanent cell lines from the EGFRL858R mouse lung tumours. Regarding KRAS driven tumors, we followed various classical protocols and modifications based on the enzymatic and mechanical dissociation of mutant KRAS driven NSCLC tumours derived from animals. The generation of cell lines from mutant KRAS driven NSCLC with wildtype p53 was successful. 50% of human NSCLC cases retain a wild type p53 so NSCLC cell lines that are p53 proficient become an important tool. We have, for the first time, achieved this goal by sequential orthotopic implantation of small solid fragments from individual murine KRAS driven primary tumors into recipient nude mice. The sequential orthotopic transplantation increases tumour malignancy and has facilitated the generation of mutant KRAS NSCLC cell lines that retain a functional p53 response.

4.1.3.3.3.2 Development of a lentiviral RNAi vector for easy validation of novel therapeutic targets in NSCLC cells and mouse model systems.
Despite the pressing need for novel cancer treatments, our improved understanding of tumor biology is not being successfully translated into better therapies. Therefore, we developed a lentiviral vector that enables in vivo validation of cancer therapeutic targets when combined with existing cancer animal models that faithfully reproduce the natural history of human disease. Unlike the conventional genetic approaches with targeted alleles, the outlined experimental strategy could be used to assess the preclinical efficacy of a growing number of putative therapeutic hits in a rapid and cost-effective manner.
The lentiviral pTOSCA system was used to deliver tamoxifen inducible CRE recombinase (CRE-ERT2) and also doxycycline inducible shRNA into KP mice (CRE inducible expression of G12D Kras and deletion of p53). In addition, floxed alleles of target genes were used. Using these systems the following genes were validated as reducing Kras p53 mutant lung tumour burden when their expression was decreased in the tumours: DDR1, CRAF and CDK4.

4.1.3.4 WP4 Evaluation of the therapeutic relevance of the newly identified targets in NSCLC patient material

4.1.3.4.1 Lyn overexpression predicts poor survival in NSCLC patients
To assess the clinical relevance of our in vitro findings, we used tissue microarrays (TMAs) containing 469 lung cancer and 61 normal lung patient samples. These were analysed for expression of FYN and LYN. We could show that while SRC, LYN and FYN were undetectable in normal lung samples, SRC was over-expressed in 70%, FYN in 45% and LYN in 42% of NSCLC samples, respectively. Hence, our in vitro expression data for SFKs are mostly representative of the clinical setting and suggest that expression of SFKs may participate to lung cancer progression. We next determined whether over-expression of SFKs might impact on prognosis. As SRC expression in lung cancer has previously been examined we focused on the expression of FYN and LYN using two different tissue micro arrays. both LYN and FYN staining associated with poor prognosis but only LYN was statistically significant.

4.1.3.4.2 Increased S6K2 expression correlates with chemoresistance
The analysis of lung tumours from 204 patients revealed that an increase in S6K2 staining correlates with the concomitant upregulation of BCL-XL (R2 = 0.9968 P = 0.009521) and decrease in cytoplasmic hnRNPA1 staining (R2 = 0.9176 P = 0.009658). From these data it can be concluded that chemoresistance will increase with increasing S6K2 expression

4.1.3.4.3 Increased RSK4 expression correlates with expression of survival proteins and metastatic markers
We stained our NSCLC tissue arrays for RSK4 and demonstrated that 57% of lung tumour samples over-expressed RSK4 while this protein was undetectable in normal lung. Interestingly, NSCLC were more frequently positive for RSK4 than SCLC samples. In addition, adenocarcinoma stained more strongly than squamous cell carcinoma tumours. Increased expression correlates with expression of survival proteins and metastatic markers. We have not yet been able to show a correlation between RSK4 over-expression and poor survival outcome and are currently looking at whether over-expression correlates with poor response to therapy and/or early relapse after treatment.

4.1.3.4.4 Increased MARK4 expression is correlated with decreased patient survival
To assess the clinical relevance of our findings regarding MARK4, we first evaluated the potential of MARK4 as prognostic biomarkers in metastatic breast and lung cancer patients
We analyzed the overall survival (months) of a cohort of patients with metastatic lung cancer and correlated this with the levels of MARK4 mRNA present in their tumors. High levels of MARK4 mRNA correlated with a lower overall survival in lung cancer patients. MARK4 mRNA levels are increased in >50% of tumors as compared to normal lung tissues. In agreement with this, MARK4 protein was detectable in 37% of lung cancer samples while it remained undetected in normal lung tissue. In addition, comparison of metastatic and primary lesions form the same patients revealed that MARK4 expression was increased in secondary as compared to primary tumors in 41% of cases. Hence, these data suggest that MARK4 expression increases during tumorigenesis, impacting negatively on patients’ survival.
Taken together, MARK4 expression levels may be relevant prognosis markers in lung cancer.

4.1.3.4.5 Low expression of GSH synthesizing enzymes correlates with erlotinib resistance
We examined our collection of EGFR resistant vs sensitive tissues and measured mRNA expression levels of GSH synthesizing/degrading enzymes in the T790M positive vs negative biopsy samples. Strikingly, T790M tissue samples had lower mRNA expression levels for GSH synthesizing enzymes. Moreover, this was not seen in erlotinib resistant patients who did not carry the T790M mutation. The mechanism by which the EGFR and in particular the T790M mutation might regulate GSH levels and how low GSH might contribute to erlotinib resistance is the subject of our on-going research. However our data raises the possibility that patients who become resistant to erlotinib and who have low GSH levels might be resensitized to this drug through the addition of an inhibitor of enzymes involved in GSH degradation. Ethacrynic acid, an old diuretic, might be re-purposed for this, although newer agents may be more appropriate.

4.1.3.4.6 Amphiregulin expression is increased in pleural effusions of NSCLC
To confirm the high secretion of amphiregulin seen in NSCLC cancer cell cultures, we performed a similar analysis on pleural effusions from lung cancer patients. Interestingly, the major growth factor detected in pleural effusions from patients was also AREG; 16 of 18 samples contained relatively high concentrations of it. Taken together, these results identify AREG, as the most prevalent EGF-like growth factor in lung cancer.

4.1.3.5 WP5 Development of molecules to overcome resistance-conferring loops as experimental approach for NSCLC

4.1.3.5.1 Development of novel therapeutics

4.1.3.5.1.1 Development of antibodies against the EGFR ligand amphiregulin
When targeting growth factor receptor signaling with antibodies, there are 2 options: antibodies targeting the receptor or antibodies targeting the ligands. Drugs targeting EGFR or the homologous protein ERBB2/HER2 are already used for the treatment of breast, lung, gastric, pancreatic, colorectal, and head and neck cancers. We generate a mAb to human AREG, and EGFR ligand highly secreted by NSCLC tumors. First, we constructed a fusion protein comprising the EGF-like domain of AREG. Notably, the EGF-like domain of AREG and other EGF-like ligands consists of three highly conserved disulfide bonds, which are responsible for the correct folding and activity of all ERBB ligands. Therefore, we chose to fuse also the thioredoxin (TRX) protein to the EGF-like domain, since this ensures the correct formation of disulfide bonds., and immunized mice with tis protei. This led to the production of a high-affinity binding, AR30, that binds strongly with human AREG. We could also demonstrate AR30 reduced AREG induced EGFR phosphorylation, and thus this antibody truly functions as an effective blocker of the AREG induced cellular signaling.

4.1.3.5.1.2 Development of TRAP molecules capable of intercepting EGFR family ligands
Epidermal growth factor (EGF)-like growth factors control tumor progression as well as evasion from the toxic effects of chemotherapy. Accordingly, antibodies targeting the cognate receptors, such as EGFR/ErbB-1 and the co-receptor HER2/ErbB-2, are widely used to treat cancer patients, but agents that target the EGF-like growth factors are not available. To circumvent the existence of 11 distinct ErbB ligands, we constructed a soluble fusion protein (hereinafter: TRAP-Fc) comprising truncated extracellular domains of EGFR/ErbB-1 and ErbB-4. The recombinant TRAP-Fc retained high-affinity ligand binding to EGF-like growth factors and partially inhibited growth of a variety of cultured tumor cells. Consistently, TRAP-Fc displayed an inhibitory effect in xenograft models of human cancer, as well as synergy with chemotherapy. Additionally, TRAP-Fc inhibited invasive growth of mammary tumor cells and reduced their metastatic seeding in the lungs of animals. Taken together, the activities displayed by TRAP-Fc reinforce critical roles of EGF-like growth factors in tumor progression, and they warrant further tests of TRAP-Fc in preclinical models.

4.1.3.5.1.3 Development of anti-EGFR DARPins
Designed Ankyrin Repeat Proteins (DARPins) are recombinant proteins that can be selected in vitro for strong binding to any target of choice. They have several advantages such as ease of production at large scale in bacteria, small size, pharmacokinetic properties, possibility of (radio)labeling with other molecules, possibility to combine several small DARPins in a multivalent molecule etc. We selected against the EGFR, competing with the liganfd EGF, but without causing a signal. These DARPins were shown to inhibit proliferation by inducing G(1) cell cycle arrest, similar to cetuximab. To optimize the biological activity of the DARPins, we combined two DARPins binding to different sites on the EGFR. This combinatorial DARPin was able to reduce surface EGFR by inhibiting receptor recycling, leading to a dramatic decrease in cell viability. These results indicate that multispecific EGFR-specific DARPins are superior to cetuximab and may form the basis of new opportunities in tumor targeting and tumor therapy. Only these multispecific DARPins, but not cetuximab, were able to lead to the inhibition of receptor recycling. This multispecific DARPin can therefore form the starting point for the addition of novel functions by linking other effector domains. In contrast to antibodies, such DARPins can be prepared with very high yields in bacteria. Such DARPins may thus be able to form building blocks for future therapeutics.

4.1.3.5.1.4 Development of anti-HER2 DARPins
We also developed DARPins against HER2, and these have very remarkable properties that could be of great use for future therapies. In this category, we found that the best DARPins wee those consisting of regions recognizing two epitopes on the receptor, linked via a short linker. The biological properties of these DARPins were very different from those of HER2 targeting antibodies such as trastuzumab. Cell cycle analysis indicates that both Mabs and DARPins cause a G1 arrest, but only the DARPins lead to a sub-G1 DNA content and only the DARPins led to apoptosis. Bispecific DARPins inhibit HER2/HER3 complex formation more efficiently than TZB and PZB. Growth inhibition in the presence of Heregulin (HER3 ligand) is better for DARPins then for TZB and PZB. Only the DARPins inhibit ErbB2 phosphorylation, while both DARPins and PZB inhibit ErbB3 phosporylation. The DARPins very efficiently inhibit spheroid tumor growth, as well as heregulin induced acini budding. DARPins, but not trastuzumab, induce dephosphorylation of HER2 DARPins induce sustained inhibition of both ERK and PKB pathway, whereas Trastuzumab only inhibits PKB in a transient manner! DARPins control key players of cell cycle (cyclin D1, p27KIP1) and apoptosis (BIM), whereas Trastuzumab has much weaker effects on these players

4.1.3.5.1.5 Development of anti-cMet DARPins
Activation of cMet receptor signaling is an event that explains part of the cases of NSCLC therapy resistance. Hence, it is important to also design therapeutics that target this resistance pathway. We found DARPin binders against many different portions of the cMet extracellular domain. Interestingly, none of these binders had any biological activity. At this point, the unique advantage of the DARPins comes into play, however, as they can be easily and efficiently coupled to bi-valent and bi-specific proteins with great ease — much faster and much easier that with antibodies. When we then re-examined these bivalent DARPins, we found several ones with excellent cell killing activity. These DARPins can now be further developed into active reagents and their mechanism further studied in detail.

4.1.3.5.1.6 Development of EGFR and CDK kinase inhibitors for the potentiation of NSCLC therapy
Based on docking experiments Vichem designed and synthesized a novel EGFR inhibitor compound family. We have developed molecules with reasonable water solubility and some other preferable parameters such as the lack of cytotoxicity in non-cancerous cell model (NIH3T3) (Table 2.) and apoptosis induction in NSCLC model cell lines (HCC827 and H1993) (Figure 1.). Especially compound VCC868449 was with good parameters even comparing to reference material Erlotinib. These molecules, but definitely VCC868449 was the best and lead molecule, were efficient inhibitors of EGFR. In summary, this compound family can be considered as a good starting point to seek more effective EGFR inhibitors as drug candidates.
We designed, synthesized and analyzed a compound family (pyrido-pirimidine-7-one) with relatively wide kinase inhibiting spectrum, including CDK kinases. These compounds can also be good hit molecules for further inhibitor development against NSCLC.
Characterizing these molecules we have found that compound VCC251801 was the most promising with good ADME-Tox parameters in assays such as cytotoxicity, PAMPA permeability and MDCK cell permeability

4.1.3.5.1.7 Development of novel RSK4 kinase inhibitors
RSK4 silencing sensitised multiple NSCLC cell lines to cisplatin and taxol and prevented NSCLC invasion in vitro and in vivo, contrary to our findings for RSK1. This explains the failure of pan-RSK inhibitors to reproduce the beneficial effects of selected RSK4 silencing, and stresses the need for the development of RSK4 selective kinase inhibitors. Our small-molecule screen identified several molecules F of an existing and approved (for another application) drug family as potent allosteric inhibitors of RSK4 activation. Indeed, molecule F inhibits RSK4 activation at high nM concentrations and reproduces all beneficial effects of RSK4 silencing in NSCLC cells, without the adverse effects of RSK1 inhibition. There are limitations with the use of molecule F in patients however as long-term administration of this antibiotic in the clinic has shown significant hepatotoxicity in the past. Other molecule F’s either did not or only partially inhibited RSK4 activation with a matching reduction in biological effects.

4.1.3.5.2 Design of synergy between therapeutics

4.1.3.5.2.1 Design of synergy between AREG antibodies and cisplatin
Since cisplatin is an inducer of AREG, it can be anticipated that this constitutes a growth promoting autocrine loop (in particular since AREG knockdown impairs growth of NSCLC). Experiments to investigate the combined effect of cisplatin and AREG antibody treatment in NSCLC mice are currently ongoing. Similar experiments where we investigated this combination in mouse xenograft models for ovarian cancer indicate that AREG antibodies sensitize tumors to cisplatin, almost completely abolishing tumor growth.

4.1.3.5.2.2 Design of synergy between Src family kinase inhibitors and chloroquin
We observed that inhibition of Src family kinases (SFKs) induces autophagy in NSCLC. Autophagy is a cellular process that can either promote OR protect against cell death. Therefore, we first needed to investigate wether SFK induced autophagy contributes to (or is protective against) SFK induced apoptosis. We found that inhibition of autophagy with bafilomycin or with chloroquin led to an increased SFK inhibitor induced apoptosis. Again, this beneficial effect of chloroquin (a drug used for the prevention of malaria), is a nice example of repurposing old drugs for new applications in oncology. We also used a zebrafish model to test whether this combined treatment would be suitable for killing injected NSCLC tumors and this was indeed the case. Therefore, we conclude that the addition of chloroquine to SFK kinase inhibitor treatment regimens may be very beneficial.

4.1.3.5.2.3 Design of synergy between EGFR, HER2 and HER3 antibodies AND design of synergy between this triple combination and new generation TKI’s
Despite iniitial therapeutic success with EGFR tyrosine kinase inhibition (e.g. with erlotinib), resistance will ultimately develop in the majority of cases. For such cases treatment with EGFR Ab could potentially provide benefit as a backup, but clinical studies did not support this. We have found the reason for this, i.e prolonged treatment with EGFR antibodies leads to compensatory cellular mechanisms including expression of HER2 and HER3, activation of the ERK pathway, and expression of HGF which binds MET and induced MET-HER3 interaction. Based upon this , we reasoned that it could be beneficial to simulataneously block EGFR, HER2 and HER3 with a triple combination of antibodies. We found that a combination of three antibodies against EGFR, HER2, and HER3 targets all three receptors for degradation and prevents increased activation of ERK. The triple mAb combination also inhibits compensatory MET activation and disrupts physical interactions with HER3. Interestingly, the triple antibody mixture also reduces the ability of stromal cells to support NSCLC cells and better inhibits tumor cells than normal cells. And finally, we also found that the triple combination enhances the inhibitory effects of new-generation TKIs, specifically at a low concentration range of the kinase inhibitors. Taken together, these effects of the triple antibody combination may prove to be very beneficial for the potentiation of therapies targeting the EGFR family in NSCLC.

4.1.3.5.2.4 Design of synergy between erlotinib and ethacrynic acid
Using both NMR and MS based-metabolomics, we compared the profiles of erlotinib sensitive with erlotinib resistant cell lines provided by our partners. The resistant cell lines all contained the T790M additional mutant which is known to decrease erlotinib binding as the so called mechanism of resistance. Surprisingly, we found that in T790M erlotinib resistant cells, glutathione (GSH) levels were reduced. Treatment with ethacrynic acid (EA), a known GST inhibitor, increased GSH levels in erlotinib-resistant cells causing re-sensitisation of PC9ER and H1975 cells to erlotinib. Similarly, GPXs inhibition using mercaptosuccinate (MS) increased intracellular GSH levels and the response of H1975 cells to erlotinib. Conversely, GCLC inhibition using buthioninesulphoximine (BSO) in sensitive cells made these erlotinib-resistant (Fig1F-G), an effect associated with decreased GSH levels (FigH).
Potential Impact:
4.1.4.1 Contribution to the expected GENERAL impact of the HEALTH programme
The FP7-Health programme text states that it aims to have impact at “Improving the health of European citizens and increasing the competitiveness and boosting the innovative capacity of European health-related industries and businesses, while addressing global health issues including emerging epidemics. Emphasis will be put on translational research (translation of basic discoveries in clinical applications including scientific validation of experimental results), the development and validation of new therapies, methods for health promotion and prevention including promotion of child health, healthy ageing, diagnostic tools and medical technologies, as well as sustainable and efficient healthcare systems.”

The impact of the Lungtarget project in view of the above described desired GENERAL impact of the FP7-Health programme is described in the sections below.

4.1.4.1.1 To improve the health of European citizens
Lung cancer is a major European health problem. NSCLC is the most common form of lung cancer, accounting for approximately 80 per cent of all cases. In 2007, more than 341,000 cases of the disease were diagnosed in Europe. Lung cancer represents one-fifth of all cancer-related deaths in the European Union. NSCLC mortality rate has exceeded that of breast, prostate and colon cancers combined in Europe, accounting for almost 30 per cent of all cancer-related deaths. There is therefore an urgent need for effective NSCLC therapies. Currently, these therapies are either EGFR-targeting or Platinum based therapies. Despite some initial response, therapy resistance often develops very soon and 1 year survival is less than 5%. Our project has focused precisely on this problem, because we identified resistance factors and developed drugs that target these resistance factors. Therefore, the results of the Lungtarget project will ultimately have a high impact on the improvement of the health of European citizens.

4.1.4.1.2 To increase the competitiveness and innovative capacity of European health-related industries
Developing new drugs in general is a very expensive, high risk enterprise taking on average 10 years to develop a new drug according to legally-required safety standards. Every successful drug costs on average 500 million Euro in development. Because of the high risk, large pharmaceutical companies are more and more relying on licensing in promising drug candidates developed by SMEs or academic centres. Academic centres can produce new knowledge and discoveries and are more flexible in adapting new technologies.
The strategy of LUNGTARGET generated close collaboration between European academic centres and European pharma SMEs which are either embedded within the consortium or in close relation to the consortium, in order to bring together scientific and technological state-of-the-art expertise. Indeed, P8 (Vichem) is an SME that is embedded in our consortium precisely because of its expertise and capacitiy to generate novel protein kinase inhibitors against the kinase targets identified by the academic centres, and which in turn benefits from the general academic input from the rest of the consortium. Moreover, LUNGTARGET also strongly benefits from our are close contacts with another SME, Molecular Partners AG (http://www.molecularpartners.com) founded by P6 (Andreas Plückthun) for the development of therapeutic Darpins. This SME however has adopted another interaction pattern with our consortium. Indeed, while expressing an active interest in our ongoing research, Molecular Partners (for purely commercial reasons) is not investing in early stage discovery research but is more interested in a licensing of the endresults of the proposed project. This will be highly valuable for large scale production and marketing of DARPins developed by our consortium.
As a European SME partner with a focus on the development of protein kinase targeting drugs, Vichem (P8) has really brought the entrepreneurial spirit of a biotech company into the Lungtarget consortium, with special emphasis on drugability of targets, upscaling and automation, valorization, cost efficiency. Surpassing the primary objectives, this academic-industrial partnership also created a knowledge and technology platform on protein kinase targeting in Europe that will foster innovation and valorization by European industry. In conclusion, both sides of the partnership between our academic teams and the 2 above mentioned SMEs benefit strongly from this interaction.
The academic partners’ input have increased the innovation capacity of these SMEs (by delivering interesting therapeutically validated kinase targets to Vichem AND by delivering prototype DARPins against therapeutically validated growth factor or growth factor receptor targets to Molecular Partners AG). On the other hand, the academic teams have benefited strongly from the capacity of the SMEs to turn our findings into therapeutically usable drugs.

4.1.4.2 Contribution towards the expected impact of the original Call to which the Lungtarget project corresponds
In the text below, we will progressively zoom in (from general to more specific0 to the desired impacts of

→ Theme 1 Health, Area/Activity > 2. Translating Research for Human Health Call text:
“This activity aims at increasing knowledge of biological processes and mechanisms involved in normal health and in specific disease situations, to transpose this knowledge into clinical applications including disease control and treatment, and to ensure that clinical (including epidemiological) data guide further research.”
We have obtained results that contribute to the intended impact of this call:
- The LUNGTARGET poject discovered several mechanisms underlying therapy resistance of NSCLC.
- We have transposed these findings into novel clinical tools for the treatment of NSCLC
- Our current and further research is guided by clinical data that were analysed in the Lungtarget project. We started from the clinical observation that current NSCLC therapies are only weakly effective, and the therapeutics that we developed via our translational reseach effort, will be ultimately applied and continuously improved in accordance with clinical observations (therapy effectiveness, molecular pathology) in an effort to create a personalized medicine approach for NSCLC.

→ 2.4 Translational Research in other major diseases> 2.4.1 Cancer Call text:
“With an estimated 3.2 million new cases and 1.7 million deaths each year, cancer remains an important public health problem in Europe for cancer patients, their family as well as health care systems across Europe. With the ageing of the European population these numbers are predicted to steadily increase. Research in this policy area will focus on disease aetiology, identification and validation of drug targets and prevention, early diagnosis and treatment biomarkers as well as on assessment of preventive, diagnostic, prognostic, and therapeutic interventions. In the long term, this area will contribute to reducing cancer incidence and mortality and to improving quality of life and care with fewer side-effects to patients.”
We have obtained results that contribute to the intended impact of this call:
Our identification of molecular factors contributing to NSCLC therapy resistance, coupled with the development of novel therapeutics designed to potentiate current NSCLC therapy, will ultimately reduce cancer mortality. It is also anticipated that our precisely molecularly targeted therapeutics will allow us to create combination therapies with lower doses of the individual drugs (e.g. in the case of potentiation of Pt-based therapy), and that this will lead to fewer side effects to patients. We have also seen this with our triple antibody combination for the treatment of EGFR driven NSCLC: this combination allows for lower doses of either Pt or tyrosine kinase inhibitors, and this will undoubtedly contribute to improving quality of life and care with fewer side-effects to patients.

HEALTH.2010.2.4.1-6: Translational research on cancers with poor prognosis Call text:
The results of research in this area will contribute to ultimately reducing patient mortality for a number of difficult-to-treat cancers with dismal survival rates and integrate basic-clinical European scientific excellence.

The research proposed will integrate excellence in basic science and clinical practice to help improve the very low survival rate for NSCLC. Because of our particular focus, our program will reduce mortality for two NSCLC patient categories, constituting 10% (EGFR mutation+/erlotinib treatment) and 30% (K-Ras+/Platinum based treatment) of total NSCLC.
The research proposed will provide new insights regarding the causes of NSCLC treatment failure. Our RNAi/secretion/kinome studies will reveal a series of new druggable targets to reverse resistance to Erlotinib- or Pt-based therapies. By defining subgroups of patients most likely to benefit from particular combinations of new receptor (and non-receptor) targeting therapies and by generating DARPins, monoclonal antibodies, soluble receptors, and kinase inhibitors with novel activities we intend to contribute to providing more efficacious lung cancer therapies. Our reverse and forward translation will lead to the initiation of a series of novel clinical trials using various (sometimes already existing) agents. Moreover, our innovative transgenic mouse treatment models should provide a test bed to verify the value of further new agents for lung cancer.
Crucially, our consortium can rely on several SMEs for taking our results further into the exploitation and application stage. Thus, Andreas Plückthun (P6) founded the SME Molecular Partners (www.molecularpartners.com). The role of this SME is the mass production, clinical study and commercialization of our DARPins. Similarly, Yosef Yarden (P7) can rely on several SMEs with as role: turning our monoclonal antibodies (towards EGF-like growth factors that bind to ErbB family members) and soluble receptors (TRAPs, combining two receptors of the ErbB family) into therapeutic agents that are suitable to enter clinical studies. Finally, P8, Vichem (www.vichem.hu) is an SME with successful previous BIOMED and FP6 experience. The role of Vichem: is to take the identification of our protein kinase targets further into developing therapeutic kinase inhibitors to enter clinical studies.

4.1.4.3 Overview of the specific impacts of the Lungtarget project

4.1.4.3.1 Our findings regarding fundamental behavioral differences between therapy sensitive versus resistant NSCLC improves the understanding of therapy resistance in NSCLC
We have generated a more detailed signaling map that explains the therapy resistant phenotype for erlotinib resistant EGFR driven NSCLC en for Pt-resistant KRAS driven NSCLC. We have identified several secrted factors, protein kinases and transcription factors that are involved in these resistance processes.
This improved knowledge will not only be beneficial for the Lungtarget partners themselves, but also for other basic and applied researchers who are involved in research aimed to increase understanding of therapy resistance mechanisms.

4.1.4.3.2 Our validation of therapeutic targets in mouse models for NSCLC provides a safe starting point for the development of novel drugs to overcome therapy resistance
We have not only just identified factors that are involved in NSCLC therapy resistance: we have also validated them in mouse model systems for NSCLC. This is very important, since these validated targets can now serve as the starting point for the development of novel therapeutics via approaches that are similar to the ones we used. Because of the widespread involvement of the pathogenic proteins that we study here (mutated EGFR or mutated K-RAS), these results will also be important for other researchers working on tumors driven by these genes. Therefore, this validation of therapeutic targets will not only be beneficial for the Lungtarget partners themselves, but also for other applied researchers who are studying EGFR-driven or K-RAS driven tumors of other organs.

4.1.4.3.3 Our validation of therapeutic targets in clinical samples of NSCLC patients provides the foundations for novel tests to assess the mechanisms of resistance in a patient focused manner and to adapt therapy accordingly
We have validated several targets by our examination of tissue microarrays, serum and pleural effusions of NSCLC patients. These validated targets are proteins whose expression/secretion/activation status can be examined in clinical samples (tissue or serum). This will allow clinicians to make rational decisions for personalized medical treatment of NSCLC. If these targets are in due course also confirmed as therapeutic targets for other EGFR or K-Ras driven tumor types, which is very likely, then this will increase their impact even further. This will have strong impact for clinicians (pathologists and oncologists) involved in the diagnosis and treatment of NSCLC (but likely also other cancers).

4.1.4.3.4 The lead molecules developed by the Lungtarget consortium will greatly improve therapy of K-Ras or EGFR driven NSCLC
We identified several approaches for the therapy of erlotinib resistant EGFR driven NSCLC or for KRAS driven NSCLC (for which there are few options besides Pt).
For example, we discovered the following approaches (non-exhaustive list):
• Targeting the GATA2 network in KRAS driven NSCLC via combination therapy with the approved drugs bortezomib and fasudil
• Combined targeting of DDR1 (with dasatinib) and Notch (with a DDL4 antibody) for treatmentof KRAS driven NSCLC
• Treatment of EGFR driven NSCLC with our antibody against amphiregulin
• Treatment of erlotinib resistant NSCLC with our triple antibody combination against the broader EGFR family.
• Treatment of erlotinib resistant NSCLC with a combination of erlotinib and ethacrynic acid
• Treatment of NSCLC invasion and metastasis with RSK4 inhibitors.
The above described therapeutic approaches can be immediately applied for clinical studies.
Furthermore, we have several other therapeutic molecules (e.g. DARPins against EGFR, HER2, c-Met and kinase inihibitors) that can be further optimized (e.g. in terms of chemical synthesis, recombinant expression, pharmacodynamics) for ultimate use in clinical trials.
These therapeutic lead molecules will alo be highly relevant for the treatment of other K-Ras or EGFR driven tumors.

List of Websites:
www.lungtarget.eu

Coordinator: Prof Dr Johan Van Lint
e-mail: Johan.Vanlint@med.kuleuven.be

Project information

Grant agreement ID: 259770

Status

Closed project

  • Start date

    1 January 2011

  • End date

    30 June 2015

Funded under:

FP7-HEALTH

  • Overall budget:

    € 4 067 435,47

  • EU contribution

    € 2 997 332

Coordinated by:

KATHOLIEKE UNIVERSITEIT LEUVEN