Skip to main content

A drug strategy targeting heat shock protein 90 to combat metastatic, p53 mutant ovarian cancer.

Final Report Summary - GANNET53 (A drug strategy targeting heat shock protein 90 to combat metastatic, p53 mutant ovarian cancer.)

Executive Summary:
Epithelial ovarian cancer (EOC) is the most lethal gynaecological malignancy. The predominance of aggressive Type II tumours (comprised of high-grade serous, high-grade endometrioid, and undifferentiated carcinomas) that are characterised by nearly ubiquitous p53 mutations (> 96% of patients) and primary or acquired resistance to platinum-based chemotherapy is the reason for the high mortality rate. Recent data from the EUROCARE database showed a 5-year relative survival for European women diagnosed with EOC of only 38 % (range 31-41 % by European region). Across Europe 66,700 women are diagnosed with ovarian cancer and 41,900 die of the disease every year. This high mortality rate is due to the predominance of late-stage diagnoses, a high relapse rate after primary therapy and poor response of metastatic platinum-resistant tumours to current regimens. 70 % of EOC patients present with metastasised disease at the time of primary diagnosis (peritoneal carcinosis).
The current standard of primary therapy is cytoreductive surgery and adjuvant platinum-based chemotherapy. The addition of bevacizumab has been shown recently to improve progression-free survival in women with ovarian cancer. Initial response rates to primary therapy are high, but inevitably most patients will relapse within a short period of time and ultimately die of the disease.
Treatment strategy for recurrent ovarian cancer depends on the platinum-free interval, which is the time between the last platinum-based therapy and the detection of relapse. If the progression of the disease is more than 6 months after the last platinum-based therapy, the tumour is considered to be platinum-sensitive (Pt-S) and can be retreated with Carboplatin in different combinations such as with Paclitaxel, Gemcitabine or pegylated liposomal Doxorubicine (PLD). On the other hand, with current standard therapy options, e.g. Paclitaxel weekly, the median overall survival of metastatic platinum-resistant (Pt-R) ovarian cancer patients is only 14 months. Addressing the pressing need for more effective, innovative treatment strategies to improve the dismal survival of relapsed EOC patients, two clinical trials, the GANNET53 trial and the EUDARIO trial applied highly innovative treatment concepts whose scientific rationale directly grew out from solid basic research findings by members of the GANNET53 consortium. The drug strategy that targets the central driver of aggressiveness and metastatic ability of these EOC cancers – namely stabilised mutant p53 protein (mutp53) – for degradation via an innovative Hsp90 (heat shock protein 90) inhibition mechanism. The clinically most advanced, efficacious, and safest second-generation synthetic Hsp90 inhibitor currently available (used in >600 patients in unrelated clinical studies) named Ganetespib was used in both trials. In Pt-R ovarian cancer patients with mutp53 Type II EOC tumours Ganetespib was applied in a stratified approach in the GANNET53 trial. To achieve the most profound survival benefit, the GANNET53 trial added Ganetespib to standard Paclitaxel weekly therapy, since promising in vitro and in vivo data showed strong synergistic effects with this combination.
In Pt-S ovarian cancer patients the EUDARIO study is evaluating the therapeutic benefit of broad DNA repair pathway inhibition after induction of DNA damage. This is achieved by combining 1) standard induction platinum-based chemotherapy with an HSP90 inhibitor on one hand and 2) maintenance Poly(adenosine diphosphate-ribose) polymerase (PARP) inhibitor (PARPi) treatment combined with Ganetespib on the other hand. This approach of broad DNA repair inhibition is again applied in tumours with a mutant p53 background (corresponding to specific histological subtypes) to maximize potential therapeutic effects. EUDARIO (European Trial on enhanced DNA Repair Inhibition in Ovarian Cancer) is a multicentre, open-label, three-arm randomised Phase II trial assessing the safety and efficacy of Ganetespib in combination with Carboplatin followed by maintenance treatment with Niraparib versus Ganetespib plus Carboplatin followed by Ganetespib and Niraparib versus Carboplatin in combination with standard chemotherapy followed by Niraparib maintenance treatment in platinum-sensitive ovarian cancer patients.

Rationale for Combination of Paclitaxel with Ganetespib: The current standard of primary therapy for EOC is cytoreductive surgery and adjuvant chemotherapy with Carboplatin and Paclitaxel. Initial response rates are high, but inevitably the vast majority of patients will relapse in short time and ultimately die of the disease. The large subgroup of ovarian cancer patients with Pt-R EOC disease face particularly dismal survival rates with a median progression-free survival of 4 months and a median overall survival of only 14 months that did not improve in over 10 years. A major treatment obstacle are the 25-30% of patients who are resistant to first-line platinum-based chemotherapy. They experience progressive or persistent disease during initial platinum-based therapy (primary platinum-refractory), or relapse of disease after less than 6 months after completion of first-line platinum-based therapy (primary platinum-resistant). Eventually all patients will become resistant to platinum after reiterative therapy with platinum-based regimens (secondary platinum-resistant disease).
Treatment options are limited for platinum-resistant (Pt-R) patients. There is consensus that secondary cytoreductive surgery is only indicated in select cases, where palliation of symptoms has priority. No “standard” chemotherapy is currently available and systemic treatment is highly dependent on the physician’s choice. Several cytotoxic agents including non-pegylated or pegylated liposomal Doxorubicin (PLD), Topotecan, Gemcitabine and alkylating agents such as Treosulphan or Cyclophosphamide have shown a relatively modest anti-tumour activity as single agent. However, Paclitaxel given as single agent on a weekly basis at a dose of 80-90 mg/m2/week, proved to be one of the most effective regimens in that desperate situation, with response rates in the range of 20-60%. This efficacy is even seen in cases that exhibit resistance to paclitaxel administered via an ‘every-3-week’ schedule. It is noteworthy that the weekly schedule is by far less toxic. However, the progression-free interval may be short. The combination of Paclitaxel with Ganetespib has shown strong synergistic effects in vitro and in vivo. Ganetespip targets Hsp90 and destabilizes mutp53, leading to mup53 degradation. Acute depletion of mutp53.

Rationale for Combination of Carboplatin with Ganetespib: HSP90 inhibitors destabilise several HSP90 client proteins, such as those governing the Fanconi Anemia DNA repair pathway (e.g. FancA) and the G2/M checkpoint (e.g. Chk1 and Wee1). This raises the possibility for using HSP90 inhibitors in combination with DNA damaging chemotherapeutics to induce massive chromosome fragmentation followed by cell death.
Carboplatin is used as first line therapy in ovarian cancer patients as well as in the platinum-sensitive relapsed situation. It mainly acts by forming interstrand crosslinks (ICL) within the DNA double helix, which can only be removed by the Fanconi Anemia pathway. Since key proteins of the Fanconi Anemia pathway are HSP90 clients (e.g. FancA), HSP90 inhibitor Ganetespib virtually eliminates a functional Fanconi Anemia DNA repair complex, thereby preventing the repair of DNA interstrand crosslinks. In parallel, Ganetespib abrogates Chk1 and Wee1 expression thereby circumventing a G2/M arrest of DNA-damaged tumour cells. Consequently, cells with unrepaired DNA damage rush into mitosis, thereby inducing massive tumour cell death. mutp53 cancer cells were described to have increased Fanconi Anemia repair activity and human mutp53 tumours were shown to correlate with increased expression of Fanconi Anemia genes. Thus, in mutp53 tumours Ganetespib counteracts the aberrant upregulation of FA repair factors by inducing their degradation.

Rationale for Combination of Ganetespib with PARP-inhibitor Niraparib: Recent studies suggest the use of PARP inhibitors (PARPis) to treat EOC regardless of the BRCA status. Therefore, it will be more straight-forward to justify clinical studies using combinations of Ganetespib not only with Carboplatin but also with PARPis.
PARPi interfere with the repair mechanism of single strand DNA breaks, which allows DNA damage to progress and to result in double strand breaks. PARPi treatment of tumour cells with homologous recombination deficiency, for instance BRCA1/2 mutation, results in synthetic lethality. Furthermore, recently patients receiving Niraparib, a PARP1/2 inhibitor, as maintenance therapy, showed a significant longer progression free survival in the recently published NOVA study. This effect was independent of BRCA mutation status. PARPi and Ganetespib together induce more phosphorylated histone H2AX (indicative of DNA damage) than single drug treatment. This was observed by immunoblot analysis and by quantitative immunofluorescence. By Annexin V staining it was shown that PARPi and Ganetespib together induce apoptosis to a greater extent than each drug alone.
Mechanistically, it is anticipated that Ganetespib and PARP inhibitors both inhibit multiple pathways required for repair of DNA damage caused by Carboplatin (e.g. Fanconi anemia, non-homologous end joining and homologous recombination). Of note, Ganetespib strongly reduces the amount of BRCA1 in the cell. Thus, it creates a deficiency like BRCA1-mutant EOC cells, which have long been known for their exquisite sensitivity towards PARP inhibitors. Thus, Ganetespib will broaden the synergy of BRCA loss and PARPi to include ovarian carcinomas regardless of their BRCA status (broadening of the sensitive ovarian cancer spectrum). Furthermore, combination of Ganetespib and PARPi would prevent the evolvement of possible BRCA re-expression leading to acquired PARPi resistance.
All this argues for a combined treatment of HSP90 and PARP inhibitors.

Rationale for a p53 mutant background: The EUDARIO trial tests the approach of broad DNA repair inhibition in EOC with a mutant p53 background (certain histological subtypes, ie. high-grade serous, high-grade endometrioid, undifferentiated, carcinosarcoma). This offers the highest potential for achieving the most profound survival benefit, as these tumours lack a functional G1 checkpoint. While in untransformed, wildtype p53 expressing cells GC treatment leads to the induction of an p53 mediated G1 arrest, mutant p53 expressing cells lacking the option of a G1 arrest. Consequently, lack of wildtype p53 leads to high sensitivity towards GC induced loss of G2/M and accumulation of unrepaired DNA damage. Moreover, wildtype p53 can repress several components of the Fanconi anemia (FA) DNA repair pathway, the primary mechanism to eliminate Carboplatin-induced DNA crosslinks. Conversely, ovarian cancer cells lacking wildtype p53 are expected to have increased levels of FA pathway activity. This further supports the model of enhanced dependence of p53 mutant ovarian cancer cells on the FA pathway, especially in the context of Carboplatin treatment. Tumour cell addiction would thus increase their vulnerability towards Ganetespib, which strongly suppresses FA components and thus confers Carboplatin sensitivity.
Data from clinico-pathological and molecular studies performed to date led to a model in which EOC can be divided into two broad categories, designated type I and type II tumours. In this model, type I and type II refer to critical molecular tumourigenic pathways and not to specific histopathological patterns.
Type II tumours are highly aggressive. They evolve rapidly, have a high metastatic activity, and therefore have almost always already spread beyond the ovaries at the time of diagnosis. Thus, this tumour type is the most problematic from a clinical point of view. Moreover, type II tumours account for the overwhelming majority (>70%) of EOC. Histologically, type II tumours are mainly high-grade serous (HGS) carcinomas, and the remainder are high-grade endometrioid, undifferentiated or a subset of clear cell carcinomas. HGS carcinomas account for ~ 85% of all ovarian cancer deaths.
Importantly, type II tumours are characterised by the near ubiquitous presence of p53 missense mutations (mut p53) - their preeminent molecular hallmark. Mut p53 proteins highly stabilize in tumour cells and many of them actively promote oncogenicity (called Gain-of-Function mutants). This strongly suggests that mut p53 is a central oncogenic driver in the pathogenesis of these tumours.
In sharp contrast, type I tumours almost always lack p53 mutations, but often harbour somatic mutations of protein kinase genes including PIK3CA and ERRB2, and other signalling molecules including KRAS, BRAF, CTNNB1 and PTEN (13). Type I tumours are slow growing, often confined to the ovary at diagnosis, and develop in a stepwise fashion from well-recognised precursors, in most cases borderline tumours. Type I tumours include low-grade serous carcinomas, low-grade endometrioid carcinomas, mucinous carcinoma and a subset of clear cell carcinomas.

General objective: The general objective of thewo clinical trials was to combat metastatic Pt-S EOC (EUDARIO trial) and Pt-R EOC (GANNET53 trial) with novel drug strategies that target the central driver of aggressiveness and metastatic ability of these EOC cancers, namely stabilised mutant p53 protein, and one of the most important pathways in resistance to chemotherapies, namely the Fanconi Anemia DNA repair pathway, for elimination via an innovative Hsp90 inhibition mechanism in order to substantially improve survival.

Specific objectives:
• Completion of all legal, ethical, and administrational prerequisites for the execution of the planned GANNET53 and EUDARIO clinical trials.
• Definition of safety of Ganetespib in a new combination with the taxane Paclitaxel.
• Definition of safety of Ganetespib in combination with Carboplatin in platinum-sensitive ovarian cancer patients in the EUDARIO trial.
• Definition of safety of Ganetespib in a new combination with the PARPi Niraparib in the EUDARIO trial.
• Clinical Proof-of-Concept for the innovative mechanism of targeting mutp53 by Hsp90 inhibition in the GANNET53 trial.
• Determination of efficacy of our new therapeutic strategy in Type II platinum-sensitive ovarian cancer patients in comparison to standard therapy options in the EUDARIO clinical trial.
• Establishment of a unique biobank of archival (FFPE and fresh-frozen tissues) and prospectively collected ovarian cancer biosamples (tissue biopsies, ascites, blood) before and during experimental treatment.
• Development of innovative software for effective organisation of a large multi-centre biobank (virtual tumour-bank), real-time tracking and distribution of biosamples, and for handling of clinical data.
• Clinical Proof-of Concept for the innovative mechanism of enhanced DNA repair inhibition via Hsp90 inhibition following DNA damage by Carboplatin in the EUDARIO clinical trial.
• Clinical Proof-of Concept for 1) broadening sensitivity of ovarian cancers towards PARPi via generation of a BRCA like phenotype by Hsp90 inhibition and 2) preventing/circumventing development of PARPi resistance by combination with a Hsp90 inhibitor in the EUDARIO clinical trial.
• Evaluation of quality of life in ovarian cancer patients treated in the GANNET53 trial and in the EUDARIO trial.
• In vivo Genetic and Pharmacologic Proof-of-Principle for the mutp53-targeting concept in engineered knock-in mouse models.
• Coordination of collection, processing, storage and transfer of human biological samples by providing standard operating procedures.
• Stringent Causality proof for the mutp53-based mechanism of drug action of Ganetespib in human ovarian cancer models (cultured cells and xenografts.
• Implementation of Central Histopathological Review (CHR) to ensure Type II histology in all patients included into the Phase II clinical trials and for quality control of biosamples that will be used for p53 analysis in translational research tasks.
• Development of a functional molecular test to detect levels of mutp53-Hsp90 complexes in tumour tissues (proximity ligation assay), and the evaluation of its value to predict responsiveness to experimental therapy with Ganetespib in the GANNET53 trial.
• Evaluation of the value of circulating tumour cells (CTCs) for monitoring responsiveness to experimental therapy with Ganetespib in the GANNET53 trial.
• Determination of the exact mutational p53 status in patients enrolled in the Phase II GANNET53 trial.
• In vivo Genetic and Pharmacologic Proof-of-Principle for the mutp53-targeting concept in engineered knock-in mouse models.
• Stringent Causality proof for the mutp53-based mechanism of drug action of Ganetespib in human ovarian cancer models.




Project Context and Objectives:
GENERAL OBJECTIVE
The GANNET53 project with its two clinical trials, i.e. the GANNET53 and the EUDARIO trials, combats metastatic ovarian cancer with a drug strategy inhibiting the central chaperone Hsp90 in order to SUBSTANTIALLY IMPROVE SURVIVAL.

THE GANNET53 CONCEPT
The highly innovative approach of the GANNET53 project provides A MORE EFFECTIVE THERAPY, thereby improving survival:
a) The GANNET53 and EUDARIO clinical trials apply a stratified treatment approach in highly aggressive, p53 mutant Type II tumours to achieve the most profound survival benefit.
b) The GANNET53 clinical trial targets the central driver of tumour aggressiveness and metastatic ability in this disease, namely stabilised mutant p53 protein (mutp53) via the innovative mechanism of destabilising mutp53 via Hsp90 (heat shock protein 90) inhibition.
c) The EUDARIO clinical trial applies the concept of Hsp90 inhibition to crucially inhibit DNA repair by rapid decay of key components of the Fanconi anaemia pathway as well as of cell cycle checkpoint mediators following DNA damage by Carboplatin.
d) The GANNET53 and EUDARIO clinical trials apply the safest, most effective and most advanced Hsp90 inhibitor available, i.e. Ganetespib, to substantially improve survival.
e) The clinical trials apply a highly promising drug combination, i.e. Ganetespib with the taxane Paclitaxel in the GANNET53 clinical trial and Ganetespib with Carboplatin on one hand and with the PARPi Niraparib on the other hand in the EUDARIO clinical trial, respectively. These drug combinations have shown strong synergistic effects in vitro and in vivo.

SCIENTIFIC BACKGROUND OF THE GANNET53 CONCEPT
1. The GANNET53 and the EUDARIO clinical trials apply a stratified treatment approach in highly aggressive, p53 mutant Type II tumours to achieve the most profound survival benefit.
Data from clinicopathological and molecular studies led to a model in which EOC is divided into two broad categories, designated Type I and Type II tumours. Type I and Type II refer to critical molecular tumorigenic pathways and not to specific histopathologic patterns. Type II tumours are highly aggressive. They evolve rapidly, have a high metastatic activity, and therefore have almost always already spread beyond the ovaries at primary diagnosis. Thus, this tumour type is the most problematic from a clinical point of view. Moreover, Type II tumours account for the overwhelming majority (>70%) of epithelial ovarian cancer (EOC). Histologically, Type II tumours mainly are high-grade serous (HGS) carcinomas that account for ~ 85 % of all ovarian cancer deaths. Importantly, Type II tumours are characterized by the near ubiquitous presence of TP53 mutations - their preeminent molecular hallmark, which in contrast are very rare in Type I tumours. This strongly suggests that mutated p53 protein (mutp53) is a central oncogenic driver in the pathogenesis of these tumours. Based on the facts that Type II tumours are the most lethal and the most prevalent EOC type, and that mutp53 is the central oncogenic driver in these tumours, the novel “GANNET53” therapeutic approach was applied in a stratified molecularly defined patient population with Type II EOC. This offered the highest potential for achieving the most profound survival benefit.
2. The GANNET53 clinical trial targets the central driver of tumour aggressiveness and metastatic ability in this disease, namely stabilised mutant p53 protein (mutp53). This is achieved through an innovative mechanism of destabilising mutp53 via Hsp90 inhibition.
Stabilised mutp53 is a novel, rational and potent druggable target in cancer treatment. Missense mutp53 proteins (which make up > 85% of all p53 mutations) not only lose their tumour suppressor function, but often acquire new oncogenic functions (gain-of-function, GOF) to actively drive higher proliferation, metastatic ability and chemoresistance. Compelling evidence from mutp53 knockin mice carrying human hotspot mutations provide definitive genetic proof for GOF in vivo. Constitutive stabilisation is the hallmark of (full-length missense) mutp53 proteins in tumour cells and their aberrant accumulation is the prerequisite for exerting GOF. Most importantly, mutp53 cancers develop a strong dependency on high levels of mutp53 for survival (‘addiction’ to mutp53). Therefore, acute withdrawal of mutp53 triggers strong spontaneous cytotoxicity, blocking invasion and metastasis and restoring chemotherapy-induced cell death in human cancer xenografts in vivo. mutp53 proteins depend on permanent folding support by the multi-component HSP90 chaperone machinery (which in turn is constitutively activated in cancer but not in normal cells), and that it is this stable interaction between mutp53 and HSP90 that is largely responsible for mutp53 accumulation specifically in tumour cells. Pharmacological inhibition of the machine’s core ATPase Hsp90 (such as by the highly potent second generation Hsp90 inhibitor Ganetespib) destroys the complex between HSP90 and mutp53, thereby liberating mutp53 and inducing its degradation by MDM2 and CHIP E3 ubiquitin ligases. Thus, Hsp90 inhibition mediates effective destabilisation and degradation of mutp53 in human tumour cells, acutely withdrawing an oncoprotein these cells depend on for survival. Given on the advanced development of Hsp90 inhibitors, this new paradigm holds immediate strong translational potential for significantly improving outcome in mutp53-driven cancers such as Type II EOC.
3. The GANNET53 clinical trial applies a highly promising combination - Ganetespib with the taxane Paclitaxel - that has shown strong synergistic effects in vitro and in vivo.
In general, the combination of first-generation Hsp90 inhibitors and taxanes has shown synergy in preclinical evaluations with other Hsp90 inhibitors such as 17AAG. While taxanes disrupt the microtubules, an essential structural component of mitosis, Hsp90 inhibitors impact the regulatory checkpoint proteins controlling progression through the cell cycle. In addition, both drugs disrupt other critical facets of cell growth and proliferation, adding to their potential efficacy. Acute mutp53 knockdown mediated by 17AAG strongly chemosensitizes towards genotoxic drugs. Furthermore, Ganetespib was found to inhibit hypoxia-inducible factor-1alpha (HIF1-alpha), a regulator of resistance to taxanes. Moreover, Hsp90 inhibition can lead to AKT inactivation and sensitise tumour cells to induction of apoptosis by Paclitaxel.
4. The EUDARIO clinical trial crucially inhibits DNA repair by rapid decay of key components of the Fanconi anaemia pathway as well as of cell cycle checkpoint via the mechanism of Hsp90 inhibition
Hsp90 inhibitors destabilise several Hsp90 client proteins, such as those governing the Fanconi Anaemia DNA repair pathway. This raises the possibility for using Hsp90 inhibitors in combination with DNA damaging chemotherapeutics to induce massive chromosome fragmentation followed by cell death.
Carboplatin is used as first-line therapy in ovarian cancer patients as well as in the platinum-sensitive relapsed situation. It mainly acts by forming interstrand crosslinks (ICL) within the DNA double helix, which can only be removed by the Fanconi Anaemia pathway. Since key proteins of the Fanconi Anaemia pathway are Hsp90 clients, Ganetespib virtually eliminates a functional Fanconi Anaemia DNA repair complex, thereby preventing the repair of DNA interstrand crosslinks. Ganetespib sensitizes ovarian carcinoma cells specifically towards ICL-inducing drugs such as Carboplatin, Cisplatin, or Mitomycin C, by inhibiting DNA repair and blocking the induction of a G2/M cell cycle arrest. The combination of Carboplatin and Ganetespib strongly decreases the viability of a large panel of human ovarian cancer-derived cell lines. Effects occur synergistically when compared to single-drug treatment. Massive chromosome fragmentation is induced by combined Carboplatin and Ganetespib but not by the individual drugs. In ovarian cancer xenografts, this drug combination strongly synergises in the inhibition of tumour growth and induction of tumour cell death.
PARP inhibitors are a group of pharmacological inhibitors of the enzyme poly ADP ribose polymerase (PARP). They are developed for multiple indications, including the treatment of heritable cancers, like ovarian cancer. Ganetespib and PARP inhibitors both may inhibit multiple pathways required for repair of DNA damage caused by Carboplatin (e.g. Fanconi anaemia, non-homologous end joining and homologous recombination). Of note, Ganetespib strongly reduces the amount of BRCA1 and creates a deficiency like mutated BRCA1 that is known for sensitivity towards PARP inhibitors. Thus, Ganetespib might broaden the synergy of BRCA loss and PARPi to include ovarian carcinomas regardless of their BRCA status.

SPECIFIC OBJECTIVES
The GANNET53 project aimed to substantially improve survival in ovarian cancer patients with metastatic Type II tumours, specifically to increase median progression-free survival and median overall survival. Ganetespib was added to standard therapy and compared to standard therapy alone in two European, multicentre, randomised open label clinical trials, i.e. the Phase I/II GANNET53 and the Phase II EUDARIO clinical trials.
• Objective 1: Completion of all legal, ethical, and administrational prerequisites for the execution of the planned GANNET53 and EUDARIO clinical trials.
• Objective 2: Definition of safety of Ganetespib in a new combination with the taxane Paclitaxel in the Phase I GANNET53 trial. Established safety of the new drug combination is a prerequisite for conducting the Phase II clinical GANNET53 trial. Specifically, it was the aim to determine whether the recommended Phase II dose for Ganetespib combinations in solid tumours, i.e. 150mg/m² once weekly, proves safe in the combination with standard dose 80mg/m² Paclitaxel weekly in Type II Pt-R ovarian cancer patients. Thus, a Phase I escalation/de-escalation trial to establish safety of Ganetespib in the new combination with Paclitaxel was performed.
• Objective 3: Definition of safety of Ganetespib in combination with Carboplatin in platinum-sensitive ovarian cancer patients in the EUDARIO trial.
• Objective 4: Definition of safety of Ganetespib in a new combination with the PARPi Niraparib in the EUDARIO trial.
• Objective 5: Determination of efficacy of the new therapeutic strategy in Type II platinum-resistant ovarian cancer patients in comparison to the standard therapy option of single agent Paclitaxel weekly in the GANNET53 trial.
• Objective 6: Determination of efficacy of the new therapeutic strategy in Type II platinum-sensitive ovarian cancer patients in comparison to standard therapy options in the EUDARIO clinical trial.
• Objective 7: Clinical Proof-of-Concept that mutp53 is a critical target for cancer therapy. The GANNET53 trial is the first clinical trial to target mutp53 and thereby the expected proof-of-concept will establish mutp53 as critical druggable therapeutic target in mutp53-dominated solid tumours, with enormous potential for exploitation in oncology in general.
• Objective 8: Clinical Proof-of-Concept for the innovative mechanism of targeting mutp53 by Hsp90 inhibition. The GANNET53 trial is the first clinical trial to use an Hsp90 inhibitor for the mechanism of destabilising mutp53 protein leading to its degradation.
• Objective 9: Clinical Proof-of Concept for the innovative mechanism of enhanced DNA repair inhibition via Hsp90 inhibition following DNA damage by Carboplatin. The EUDARIO clinical trials is the first clinical trial to use an Hsp90 inhibitor for the mechanism of DNA repair inhibition following DNA damage.
• Objective 10: Clinical Proof-of Concept for 1) broadening sensitivity of ovarian cancers towards PARPi via generation of a BRCA like phenotype by Hsp90 inhibition and 2) preventing/circumventing development of PARPi resistance by combination with a Hsp90 inhibitor. The EUDARIO clinical trial for the first time combines a PARPi with and Hsp90 inhibitor.
• Objective 11: Evaluation of quality of life in ovarian cancer patients treated in the GANNET53 trial.
• Objective 12: Evaluation of quality of life in ovarian cancer patients treated in the EUDARIO trial.
• Objective 13: Establishment of a unique biobank of archival (FFPE and fresh-frozen tissues) and prospectively collected ovarian cancer biosamples (tissue biopsies, ascites, blood) before and during experimental treatment.
• Objective 14: Development of innovative software for effective organisation of a large multi-centre biobank (virtual tumour-bank), real-time tracking and distribution of biosamples, and for handling of clinical data.
• Objective 15: Coordination of collection, processing, storage, and transfer of human biological samples by providing standard operating procedures.
• Objective 16: Implementation of Central Histopathological Review to ensure Type II histology in all patients included into the Phase II clinical trials and for quality control of biosamples that will be used for p53 analysis in translational research projects.
• Objective 17: Development of a functional molecular test to detect levels of mutp53-Hsp90 complexes in tumour tissues (proximity ligation assay), and the evaluation of its value to predict responsiveness to experimental therapy with Ganetespib in the GANNET53 trial.
• Objective 18: Evaluation of the value of circulating tumour cells for monitoring responsiveness to experimental therapy with Ganetespib in the GANNET53 trial.
• Objective 19: Determination of the exact mutational p53 status in patients enrolled in the Phase II GANNET53 trial.
• Objective 20: In vivo Genetic and Pharmacologic Proof-of-Principle for the mutp53-targeting concept in engineered knock-in mouse models.
• Objective 21: Stringent Causality proof for the mutp53-based mechanism of drug action of Ganetespib in human ovarian cancer models.

Project Results:
>> GANNET53 trial: PHASE I (Work Package 3)
A total of 10 platinum-resistant ovarian cancer (PROC) patients were included in this dose escalation/de-escalation Phase I trial by the Medical University of Innsbruck, Austria (n=1), Katholieke Universiteit Leuven, Belgium (n=4), Universitätsmedizin Berlin Charité (n=2), Germany, Universitätsklinikum Hamburg Eppendorf (n=1), Germany and Centre Anticancereux Léon Bérard (n=2), Lyon, France. Criteria for dose-limiting toxicity (DLT) are provided in Table 1.

Patients characteristics
Patients characteristics are summarized in Table. 2.

Course of the Phase I GANNET53 trial, DLT and recommended dose for Phase II
In cohort 1 (Ganetespib dose level 100 mg/m²), one patient had to be replaced based on early disease progression after a single dosing of Ganetespib and paclitaxel weekly (cycle 1, day 1). This patient was not evaluable for DLT (DLT observation time-frame minimum of two complete cycles). This resulted in the inclusion of 4 patients in cohort 1. Cohorts 2 and 3 (both at Ganetespib dose level 150mg/m2) consisted of three patients, respectively (Figure 1). No DLT occurred in cohorts 1, 2 and 3.
The DSMC reviewed safety data of patients included into cohort 1 prior to the dose escalation step and concluded that there are no objections to continuing the study according to protocol. After all patients completed the DLT observation timeframe of 2 complete treatment cycles the DSMC concluded that the GANNET53 study can move forward to Phase II without any major concerns. The DSMC recommended to use a weekly Ganetespib dose of 150mg/m2 in combination with weekly paclitaxel 80mg/m2 in the randomized Phase II trial.

Safety
Incidences of grade 1/2 adverse events (AEs) which occurred in more than 1 patient and all ≥3 AEs are listed in Table 3.
The most common AE related to Ganetespib was a transient grade 1/2 diarrhoea (n= 6/10 patients). Furthermore, related grade 1/2 AEs occurring in more than 2 patients were QTc prolongation (n= 4), nausea (n=3), anemia (n=3), headache (n=3), fatigue (n=3) and dyspnoea (n=3). Related grade 3/4 AEs were diarrhoea (n=3), neutropenia (n=2), anemia, asthenia, syncope, and acute cardiac insufficiency (n=1, respectively). There was 1 death on study (after DLT period) caused by digestive tract haemorrhage from a duodenal ulcer. Three patients discontinued study treatment due to serious adverse reactions (SAEs; digestive haemorrhage n=1, cardiac failure n=1, abdominal pain and vomiting n=1), 6 patients due to progressive disease, and one patient due to physicians’ decision.

Serious adverse reactions
Five serious adverse events (SAE) related to Ganetespib were reported, i.e. serious adverse reactions (SARs), and are summarized in Table 4. One SARs occurred in a 71-year-old patient who died from a gastroduodenal haemorrhage and haemorrhagic shock originating from an ulcer in the duodenum. This patient was initially hospitalized for hypotension, hypovolemia, and grade 3 anemia. During hospitalization, the situation worsened, and haematochezia (with normal colonoscopy findings) and repeated vomiting of blood occurred. The patient received blood transfusions, medication with proton pump inhibitors and repeated emergency gastroscopies were performed. A duodenal bleeding was identified on gastroscopy which was impossible to stop. Ten days after hospitalization the patient died of a haemorrhagic shock. Autopsy confirmed gastrointestinal bleeding from a postpyloric ulcer with a central eroded vessel and an adhesive thrombus on the surface. Microscopic peritoneal carcinosis was present. This event was considered a SUSAR.
Another SAR occurred in a 61-year-old patient who presented with acute cardiac insufficiency stage IV, loss of systolic left ventricular function and atrial fibrillation. This event occurred on day 1 of cycle 3 at the end of the paclitaxel infusion given after the Ganetespib infusion. This patient suffered severe underlying conditions such as stage IV chronic renal failure (GFR of 30ml/min), preceding acute kidney failure one year ago, history of renal cell carcinoma (left nephrectomy) and hypertension. Also, the patient received previous angiotensin II receptor antagonist medication and beta-blockers suggesting pre-existing cardiovascular disease. A hydropic heart decompensation due to volume/chemotherapy was suspected by the cardiologists. The Sponsor evaluated this event as confounded by the study medication in addition to the multiple severe underlying conditions. Volume overload during treatment administration and a hypertensive crisis occurring after the paclitaxel infusion might possibly have contributed to the acute heart failure in this patient. In the follow-up this patient has recovered to a left ventricular ejection fraction (LVEF) of 55% (at screening LVEF of 60%). This SAE was assessed as SUSAR.
Three SARs involved grade 2 AEs resulting in hospitalizations and were therefore judged as serious. This consisted of two cases of one-day hospitalizations, one for grade 2 transient diarrhoea, in which the recommended prophylactic loperamide was not given, and one for grade 2 dyspnoea occurring 4 days after experimental treatment. Both patients were discharged the next day with complete recovery from symptoms. A third case concerned grade 2 abdominal pain and vomiting, for which the patient was hospitalized in an external hospital, not involved in the conduct of this Phase I study. A laparotomy was performed in which peritoneal carcinomatosis was seen and adhesiolysis and repair of a para-stomal hernia was performed. After 12 days of hospitalization the patient was completely recovered and discharged.

Adverse events of particular interest
Diarrhoea: The most frequent and well-known AE associated with the use of Ganetespib is diarrhoea, which is typically low grade and transient, lasting 24 - 48 hours after Ganetespib administration. Prophylactic medication with loperamide was strongly recommended in all patients. 9/10 patients included in this study experienced at least low-grade diarrhoea, which followed the classical transient course. In 3/10 patients grade 3 diarrhoeas occurred. One of these 3 patients had a pre-existing short bowel syndrome with constant grade 1 diarrhoea prior to study inclusion. After each Ganetespib application diarrhoea worsened transiently, one time to grade 3 diarrhoeas.

QT Prolongation: The results of a thorough QT study conducted in healthy volunteers (Study 9090-13) reported a maximum mean ΔΔQTcF of 21.5 ms at 24 hours post study drug administration. This finding places Ganetespib in a zone of clinical ambiguity. In the present trial echocardiography (ECG) assessments were performed during screening (average of triplicate ECG recording) on day 1 of each treatment cycle and 24-hours post-Ganetespib-dose on day 2 of cycle 1. Further 24-hours post-Ganetespib-dose ECGs were strongly recommended to be performed on day 2 of each subsequent cycle. Guidelines were provided in the study protocol for additional intensive ECG monitoring in case of QT prolongation. A thorough review of QT times in all Phase I patients was performed by the Sponsor. Solely grade 1 QT prolongations occurred in the Phase I GANNET53 trial in a total of 6/10 patients. In 4 of these patients the QT prolongation was possibly or probably related to Ganetespib. All 4 patients had already pre-existing grade 1 QT prolongation at the time of screening or before their first Ganetespib dose, which increased after Ganetespib application, yet remined within the grade 1 range. Two patients had pre-existing grade 1 QT prolongation which did not worsen after Ganetespib application. In the 4 patients with QT prolongation related to Ganetespib a total of 8 events occurred with a median ΔΔQTcF of 21,4 ms (range 8 – 32ms) at 24 h post study drug administration.

Treatment exposure and clinical activity
An overview on treatment exposure and clinical activity is provided in Table 5.
A total number of 42 treatment cycles (median: 2.5 per patient, range 1-11) were applied in the Phase I GANNET53 patients. Of 42 treatment cycles, 35 (83%) cycles were completed with study medication given on all 3 days (D1, D8, D15). The median treatment duration was 1.7 months (range: 1 day – 10.1 months). The patient who continued the experimental treatment the longest received 11 cycles of treatment.
The objective response rate (ORR) was 20% (2/10 patients). Two patients showed a partial remission (one assessed by RECIST, one by CA125 criteria due to non-measurable disease). The two responses lasted 8.5 and 6 months, respectively. Stable disease was seen in 4 patients, resulting in a disease control rate of 60% (6/10 patients). Both partial responses and all stable diseases occurred in the two cohorts with the escalated dose level of 150mg/m2 Ganetespib.
Median PFS in the 10 included patients was 2.9 months (1.6 months in cohort 1 dosed with 100mg/m2 Ganetespib, 5.1 months in cohorts 2+3 dosed with 150mg/m2 Ganetespib; Figure 2). Three patients had a PFS of > 6 months.


>> GANNET53 TRIAL: PHASE II (Work Package 4)

Patients
Of a total of 171 platinum-resistant ovarian cancer patients assessed for eligibility, 133 patients were randomized, on a 2:1 ratio, to either receive Ganetespib and Paclitaxel (G/P) in the experimental arm or Paclitaxel alone (P) in the control arm. The intention-to-treat (ITT) population consisted of 90 patients assigned to the G/P arm and 43 to the P arm. Four patients in the G/P arm never received study treatment (two patients withdrew consent and in two patients the investigator did not start study treatment because of safety concerns, i.e. diagnosis of grade 2 QTC prolongation and atrial fibrillation, respectively). One patient in the P arm had a critical protocol deviation (patient had primary platinum refractory disease and was thus not eligible for inclusion in the trial). Therefore, the per-protocol (PP) population consisted of 86 patients treated with G/P and 42 patients with P, respectively.
A consort diagram of patient disposition is presented in Figure 3.
Baseline characteristics of all randomly assigned patients in the ITT population were balanced between the two arms and are summarized in Table 6. Median age at enrolment was 61.4 and 62.1 years in the G/P and P arm, respectively. The median time between first diagnosis and enrolment was 2.5 years and 2.3 years, respectively. The big majority of patients had high-grade serous histology (97.8% in G/P and 95.3% in P arm). The median number of prior treatment lines was 2 in all included patients, with a range of 1-5 in the G/P arm and 1-4 in the P arm. For most patients included in this trial, study treatment was the first line of therapy in platinum resistant disease (62.2% in G/P and 72.1% in P arm), however 37.8% of patients in the G/P arm and 27.9% of patients in the P arm have already had 1-2 prior lines of treatment in platinum resistance.

Treatment exposure
Treatment exposure of the patients in the PP population is summarized in Table 7.
The median number of started, completed (administration of study drug on all days, i.e. days 1, 8, and 15) and optimal cycles (without dose reductions, without dose delays and with study drug administration on all days, i.e. days 1, 8, and 15) were significantly lower in the G/P arm compared to the P arm (p=0.021 p= 0.022 and p=0.003 respectively). In the G/P arm more patients had dose reductions (25.6% versus 14.3%), dose delays (17.4% versus 11.9%) and skipped doses (46.5% versus 35.7%) compared to patients in the P arm. However, this difference was statistically not significant. Also, there was no statistically significant difference in the median numbers of dose reductions, dose delays, or skipped doses between the treatment arms.

Efficacy
Efficacy analyses were performed in the ITT (n=133) and in the PP population (n=128).
In the ITT population the median duration of follow-up at data cut-off (04 December 2017) was 10.0 months (IQR 4.3-15.1) in the G/P arm and 11.9 months (IQR 6.6-18.1) in the P arm. By the time of data cut-off, PFS events were reported in 124/133 (93.2%) patients and OS events in 95/133 (71.4%) patients in the ITT population (PFS events per treatment arm: 82/90 in G/P, 42/43 in P; OS events per arm: 66/90 in G/P, 29/43 in P). Thirty patients (22.6%) were still in follow-up for OS at the time of data cut-off.
For the primary endpoint, PFS, no significant difference was demonstrated for patients treated in the G/P arm compared to the P arm. In the ITT population the median PFS was 3.5 months (95% CI 3.1-3.9) in the G/P arm and to 5.3 months in the P arm (95% CI 4.0-6.6) with a non-significant Hazard Ratio (HR) of 1.3 (95% CI, 0.90 to 1.90; p = 0.16). PFS rate at 6 months was 22% (95%CI, 14%-31%) in the G/P arm and 33% (95%CI, 20%-48%) in the P arm. Also, OS did not significantly differ between the two treatment arms. Median OS was 11 months (95% CI 9.2-12.7) in the G/P arm and to 14.9 (95% CI 7.6-22.2) in the P arm (HR 1.4; 95% CI, 0.90 – 2.17; p = 0.13). PFS, PFS at 6 months and OS data in the ITT and the PP population are summarized in Table 8. Kaplan-Meier curves on PFS and OS in both populations are shown in Figure 4.
PFS II could be computed for 114 patients in the ITT population (75 in G/P arm and 39 in P arm). The median PFS II was 8.5 months (95%CI 6.6-10.3) in the G/P arm and 11.3 months (95%CI 7.6-14.9) in the P arm with a non-significant Hazard Ratio (HR) of 1.3 (95% CI, 0.87 to 1.97; p = 0.20). In the PP population 111 patients were evaluable for PFS II (73 in the G/P arm and 38 in the P arm). The median PFS II was 8.4 (95%CI 6.5-10.2) and 10.7 (95%CI 7.1-14.3) in the G/P and P arm, respectively (HR 1.3 95%CI, 0.86-1.97; p = 0.21).
Objective response rates (ORR), disease control rates (DCR) and clinical benefit rates (CBR) for the ITT and the PP population are shown in Table 9. In the ITT population ORR was 25.6% (23/90) in the G/P arm and 39.5% (17/43) in the P arm (p= 0.10). 2/90 (2.2%) patients in the G/P arm and 3/43 (7%) patients in the P arm achieved a complete response, whereas 21/90 (23.3%) and 14/43 (32.6%) patients in the G/P and P arms achieved a partial response, respectively. Responses were confirmed by a second CT scan (after >4 weeks) in 48 patients (28 in the G/P arm and 20 in the P arm). The number of confirmed ORR patients were 14.4% (13/28) in the G/P arm and 27.9% (12/20) in the P arm (p = 0.05). The DCR comprising of complete responses (CR), partial responses (PR) and stable diseases (SD) was 58.9% (53/90) in the G/P arm and 67.4% (29/43) in the P arm (p=0.37). The CBR defined as CR, PR and SD lasting for >= 4months was 17.8% (16/90) in the G/P arm and 37.2% (16/43) in the P arm (p=0.02).

Safety
Safety was analysed in all patients who received at least one dose of study medication, this resulted in an analysis set of 129 patients (safety population). The safety population consisted of 86 patients in G/P arm (excluding 4 patients who did not receive the treatment due to withdrawal or safety concerns) and 43 patients in the P arm.
A summary of all treatment related Adverse Events (AEs) in the ITT population by treatment arm for grades 1-2 (occurring in at least 10% of the patients) and grades 3-5 (occurring in more than one patient) is given in Table 10. The three most common AEs with grades 1-2 in the G/P arm were diarrhoea (78.9%), anemia (45.6%), and nausea (41.1%), whereas in the P arm they were anaemia (51.6%), peripheral neuropathy (46.5%), and nausea (39.5%).
Serious Adverse Events (SAEs) and Serious Adverse Reactions (SARs) are presented in Table 11 and Table 12, respectively. SAEs were reported more commonly in the G/P arm (39.5%) compared to the P arm (23.3%).

Adverse Events of Particular Interest
Diarrhoea: The most frequent and well-known AE associated with the use of Ganetespib is diarrhoea, which is typically low-grade and transient, lasting 24–48 h after Ganetespib administration. Prophylactic medication with Loperamide was therefore recommended in all patients in the study. Diarrhoea was seen not only to be the most common AE of Ganetespib at grades 1-2 (78.9%), but it was also the second most common AE at grades 3-5 (11.1%) in this study. It was also seen as an SAE (3.5%) and as an SAR (2.3%) in the G/P arm whereas it was neither an SAE nor an SAR in the P arm.

Gastrointestinal perforation: A second event of particular interest was gastrointestinal perforation (GIP). It was observed in 2 of 65 patients in the G/P arm as of the cut-off date of 15-Jan-2016. GIP was identified as a new safety finding and was added to the reference safety information of the Investigators Brochure (edition 11, dated 13-Nov-2015).


>> EUDARIO TRIAL: PHASE II (WORK PACKAGE 4)

By the time of submission of this final report the EUDARIO trial is ongoing. Results on the primary endpoint PFS are expected in Q3 2021.
Recruitment in the EUDARIO trial was completed with the randomization of the 122nd patient on 07 May2020. During the recruitment period 132 patients were screened, 122 of which were randomised and started treatment. Figure 5 shows the final patient enrolment per centre in the EUDARIO trial.
The first clinical site initiated for the EUDARIO trial was KU Leuven (P2) on 10 October 2018. The first ovarian cancer patient was enrolled (signed informed consent) on 30 November 2018 in Belgium and first study treatment was applied in January 2019 at P2. For more than half a year P2 (participating as single high-volume centre in Belgium) was the only open and actively recruiting centre in the EUDARIO trial, until other countries obtained full approval and more sites could be activated and join the recruitment process. Between 17 July and 28 August all 3 participating Italian sites were initiated (P20, P21, P22) and the first Italian patient was randomised on 22 July 2019 at P21 (UCSC) in Rome. After full approval of the EUDARIO trial also in Austria, P1 (IMU) was initiated on 23 July 2019 and the first Austrian patient was randomised on 24 September 2019. Thus, at the end of the Fourth reporting period (30 September 2019) 5 clinical sites from 3 countries were active in the EUDARIO trial. The participating French sites were opened in the Fifth Reporting Period, namely in the fall of 2019: Caen (P16) on 02 October 2019, Lyon (P7) on 19 November 2019 and Paris (P6) on 25 November 2019. The first French patient was randomized on 06 December 2019 at P6 in Paris. Two German sites, Berlin (P3) and Essen (P14), were initiated on 06 and 12 November 2020, respectively. The first German patient was randomized on 17 December 2019 in P14 (Essen). Another three German centres were initially planned for participation in the EUDARIO trial, i.e. Hamburg (P4), Dresden (P15) and Bonn (P23), but were finally not initiated due to the overall high recruitment speed of the other centres and completion of recruitment (prior to initiation of these additional 3 German sites). The decision to not open the German sites P4, P15, P23 at the end of recruitment was jointly taken by all consortium partners during the Consortium Meeting on 23rd and 24th February 2020 in Vienna, Austria.


>> UNIQUE BIOBANK ESTABLISHED FROM BIOMATERIALS COLLECTED FROM PATIENTS TREATED IN THE PHASE II GANNET53 AND EUDARIO TRIALS (WORK PACKAGE 5)

GANNET53 Clinical Trial - BIOBANK
An outstanding biobank was established by the clinical partners from patients included in the randomised GANNET53 Phase II trial. Collected biomaterials include archival formalin-fixed, paraffin-embedded (FFPE) tumour tissues, biopsies of the actual relapse (fresh-frozen or FFPE), blood fractions (plasma, serum, cell pallets) collected taken at different time-points prior and during study treatment, circulating tumour cells (CTCs) in the blood, as well as ascites and pleural effusion samples.
Archival FFPE tumour tissues: For all 133 patients included in the Phase II GANNET53 archival FFPE tumour tissue is available and centrally stored. All clinical partners have provided FFPE samples to allow Central Histopathological Review prior to study inclusion (FFPE samples were mandatory according to study protocol). Furthermore, Tissue Microarrays with core biopsies from FFPE blocks were generated.
Blood fraction samples (plasma, serum, cell pellets): Blood samples (for blood fraction isolation) have been collected before treatment start and at different time-points during treatment in the Phase II GANNET53 study. Impressively, in 103/133 enrolled patients a complete set of sequential blood samples per patient has been successfully collected (at all pre-specified time-points according to the trial protocol and biomaterial collection manual). In 68% of patients, sequential blood samples from at least 4 different time points are available.
Circulating tumour cells in the blood: Sequential blood samples of 128/133 patients of the Phase II GANNET53 trial were collected. Samples were received from 11 different clinical centres in Austria, Belgium France and Germany. The average number of CTS shipments (2 blood tubes per shipment) per clinical centre was 47. A total of 521 blood samples for CTC analysis were received. Impressively, an average of 4 sequential CTC blood samples per patient (range: 1 - 9) were collected.
Biopsies of the actual relapse: Biopsies at time of study inclusion could be taken in 29/133 (22%) patients. In 20 cases, 2 to 4 biopsies per patients are available; in 9 cases 1 biopsy is available. In 24/29 patients, biopsies were stored as fresh-frozen samples, in 5 of 29 patients, biopsies were stored as FFPE samples.
Ascites and pleura effusions: Fourteen ascites samples of eight patients, as well as eight pleura effusion samples of two patients were collected. Ascites cells were prepared by centrifugation and frozen in liquid nitrogen, ascites supernatant was stored, as well as cytospins at -80°C. Furthermore, there was the attempt to cultivate all samples to create immortalized tumour cell lines for further experimental testing. Short time cultures of 7 samples from 6 patients could be established. The number of patients from whom ascites was successfully collected was low. Indeed, this was expected, as most of the included platinum-resistant patients do not present with a high volume of ascites at the time of relapse, requiring paracentesis.
TMAs, serum, plasma, and effusions samples are still available for future analysis in the GANNET53 Biobank.

EUDARIO Clinical Trial - Biobank
The EUDARIO biobank is, in general, an extension of the GANNET53 biobank already established. The existing infrastructure was adapted for the new clinical trial.
By the time of submission of this final report (March 2021) the EUDARIO Biobank consisted of 113 tumour tissues (FFPE) from primary diagnosis, 17 fresh-frozen biopsies, 3144 serum samples, 1513 plasma samples, 128 blood samples for CTC analyses, 1441 cell pellets, all collected from patients screened or randomised in the phase II EUDARIO clinical trial. Timepoints of samples being collected are illustrated in Table 13.


>> SOFTWARE FOR EFFICIENT ORGANISATION OF INTERNATIONAL, MULTICENTRE BIOBANKING (WORK PACKAGE 5)

One important result of the GANNET53 project is the developed software for the efficient organisation of international, multicentre biobanking. This software includes a real-time tracking system of biosample shipments. It was developed by the IT partner xailabs (P13) in close cooperation with Charité University (P3).
Based on this biobanking software the todays ENGOT biobank was built (ENGOT: European Network of Gynaecological Oncological Trial Groups).


>> COMPANION DIAGNOSTICS (WORK PACKAGE 6)

Initially, the focus was on choosing the most suitable methods and platforms to be applied for later testing of biosamples collected from patients included in the Phase II GANNET53 clinical trial, as well as on establishing and testing different protocols. To determine the mutational status of the TP53 gene, a next generation sequencing method was chosen and established. The protein expression of p53, Hsp90, and related proteins was determined using immunohistochemical staining (IHC) and immunofluorescence staining (IF). To analyse the interaction of Hsp90 with p53 a method termed proximity ligation assay (PLA) was chosen and established. PLA was also used for detecting p53 protein aggregates “p53 prions”.

TP53 mutational status
The GANNET53 project is based on the hypothesis that mutant p53 is stabilized through the chaperon HSP90 and can therefore neither fulfil its function, nor be degraded. This stabilization is a prerequisite for gain-of-function capabilities promoting tumour growth. Furthermore, tumours carrying a missense TP53 mutation develop a dependency on the high protein levels and withdrawal should result in reduced proliferation or cell death. The HSP90 inhibitor Ganetespib is used to release stabilized p53 and target it for degradation. The GANNET53 trial includes patients with Type II ovarian cancer. This type of cancer is characterized by an almost ubiquitous presence of TP53 mutations. As part of the companion diagnostics, the TP53 mutational status of each patient included in the study was determined in archival tissue specimens. Preferably, tissue from primary tumours were used; in some cases, tissue from a recurrence was available only. This analysis not only guaranteed that TP53 mutated patients were included in the study, it was also the basis for linking effects of the treatment to the presence of certain mutations in specific. Furthermore, the mutational information was taken into consideration when analysing further results, e.g. from IHC.
Next generation sequencing was performed on DNA that was isolated from archival (FFPE) tissue sections utilising a capture based SureSeq procedure (Oxford Gene Technology). This technology allowed determining the mutational status not only of TP53, but also of the tumour suppressor genes BRCA1, BRCA2, PTEN, ATM, ATR, and NF1.
Of 133 patients included in the trial, FFPE tissue sections were received from 131. In 2 patients, the available material was not sufficient for further analysis. DNA was isolated from the 131 samples using the Qiagen FFPE DNA kit (Qiagen). In cases where the DNA quantity was not sufficient, additional FFPE sections were used for DNA isolation with the Gene Read FFPE kit (Qiagen). DNA concentrations were determined using Qubit (Thermo Fisher Scientific) quantification, and the quality was assessed using a Fragment Analyser (Advanced Analytical). DNA samples with an available input amount of >500 ng and an average fragment length of >1.000 bp, were considered good quality. Samples only fulfilling one of those requirements were considered intermediate quality, and samples failing both were rated poor quality. The input amount of DNA for next generation sequencing and the number of samples per sequencing lane was adapted accordingly.
For seven samples, it was not possible to isolate DNA with sufficient quantity or quality, despite increasing the amount of FFPE material and using a kit specifically for FFPE DNA purification for subsequent next-generation sequencing. The average fragment length of those samples was 100bp and lower and the available DNA amount was less than 100ng.
In 118/124 (95.2%) of patients a TP53 mutation was detected, whereas 6/124 (4.8%) were found to carry TP53 wild type (wt) alleles only. Of the detected mutations, 5 were single nucleotide variants affecting splice acceptor or donor sites, 22 were insertions or deletions resulting in a frameshift. Furthermore, we detected 5 in frame deletions, 4 synonymous mutations, and in most cases (n=87) a single nucleotide variation resulting in a missense variant. In 4 tumours, more than one TP53 mutation was detected.
These results fit well into reports from current literature. The Cancer Genome Atlas (TCGA) Research Network reported TP53 mutations in 96% of high-grade serous carcinoma specimens.
The TP53 mutational status (mut vs. wt) had no influence on the outcome of the patients as determined by Log-rank testing, likely due to the very low number of wt-cases.

P53 and HSP90 expression and protein-protein interaction
Three Tissue micro arrays (TMA) were generated that included 2 cores of each FFPE tissue block. In the case of fine needle biopsies, only one core was possible sometimes. p53 IHC staining was performed. The staining intensity was classified into four categories negative, weak, intermediate, and strong) and correlated with the TP53 mutational status as well as survival data.
According to the literature, a mutation in the TP53 gene leads to increased stability and accumulation of mutant p53 protein. Especially missense mutations, which, in contrast to nonsense mutations result in full-length mutant p53 protein, are associated with abnormal accumulation of p53 protein. The statistical analysis revealed a moderate association between p53 protein expression determined by IHC and the mutational status (Cramer’s V = 0.602 p = 0.000). In 74% (55 out of 74 samples) of missense mutated samples stabilized p53 protein could be detected. In total 58% (65/112) of the tissues showed p53 expression detected using IHC. p53 expression was also determined by IF staining in FFPE TMA sections. p53 protein could not be detected in 45% (50/111) of patients, 7% (8/111) of patients showed weak staining, and intermediate to strong expression was detected in 20% (22/111) and 28% (31/111), respectively, of patients. Overall, there was good agreement between IHC and IF staining results.
However, p53 protein expression was neither associated with overall (OS) nor with progression-free survival (PFS).
Moreover, a semi-quantitative IF staining was performed for Hsp90. Hsp90 protein expression was detected in all, except one, patients. In 8% (9/111) of patients Hsp90 was weakly expressed, whereas in 24% (27/111) and 67% (74/111), respectively, intermediate to strong Hsp90 protein expression was detected.
As mutant p53 is one of Hsp90 client proteins, and the Hsp90 inhibitor Ganetespib destabilizes this interaction, the co-expression of both proteins was assessed. The concomitant expression of p53 and Hsp90 (p53+/Hsp90+) was neither associated with a benefit in PFS nor in OS.
The co-immunofluorescence (co-IF) staining revealed that not all p53-positive cells were Hsp90-positive as well, suggesting that not all tumour cells bear Hsp90-p53 complexes. Nonetheless, a co-immunofluorescence staining cannot detect specific interaction between two proteins, results are indicative only. Therefore, a proximity ligation assay for the protein-protein interaction of Hsp90 with p53 was developed. Only 17 out of 116 (14.7%) patients showed at least some PLA signals in single tumour cells. Furthermore, there was no statistical difference between chemonaïve and neoadjuvant treated samples regarding the presence of Hsp90-p53 complexes. Also, no association between primary or relapsed tumours and the detected number of Hsp90-p53 complexes was found. The rather low number of PLA positive patients suggests, that even if both proteins are overexpressed and/or stabilized in a tumour, they are not present in form of a complex. Therefore, the Hsp90 inhibitor Ganetespib, which is used to release stabilized p53 from such a complex and target it for degradation, probably will not work.
Summary and conclusions: The rationale of the GANNET53 clinical trials was to administer Ganetespib to block Hsp90 and release mutant p53 from the complex, which in turn leads to degradation of mutant p53 and cell cytotoxicity. The blockage of Hsp90 and the resulting elimination of stabilized p53 sensitizes p53 mutated cancer cells to chemotherapeutics. Hsp90 was highly expressed in almost all patients, whereas p53 protein was only stabilized in about half of all patients. Although, p53 is a known client protein of Hsp90, a stable interaction between these two could only be found in a small subgroup (14.7%) of patients. Overall, these findings indicate that the presence of Hsp90-p53 complexes in primary tumour tissue specimens cannot predict the responsiveness to Ganetespib treatment of relapsed platinum-resistant patients. Moreover, the results of three patients with consecutive ascites/pleural effusions gave no indication that Hsp90 blockage would resolve these complexes and that these patients have a superior outcome.

Molecular Ganetespib efficacy testing
The objective was to determine whether the presence, respectively absence of complexes between Hsp90 and p53 are predictive for the response to Ganetespib treatment. The plan was to collect ascites of Phase II patients and culture the epithelial cells in the ascites in the presence or absence of Ganetespib and to analyse Hsp90-p53 complexes in these cells. The hypothesis was that results from the PLA predict the responsiveness of patients to Ganetespib. It was anticipated that most cells contain Hsp90-p53 complexes. A strongly reduced signal or no signal after treatment was regarded an in vitro response. A robust procedure was developed using cell line models.
Unexpectedly, of 133 patients included in the GANNET53 trial, only few suffered from symptomatic ascites leading to clinical indication of ascites drainage. 14 ascites samples were collected from 8 patients, as well as 13 pleural effusion samples from 5 patients. In 4 patients more than one ascites/pleural effusion was available. From two patients both, ascites and pleural effusions were collected. Samples were processed and cell pellets, ascites supernatants and cytospins were prepared and stored. Given the low number of available (consecutive) ascites samples, it was not possible to determine the suitability of this test to predict response to Ganetespib. Short time cultures of cells from ascites samples were established and frozen for further expansion and analysis. However, these cells were characterised by extremely slow growth and a high death rate. Therefore, cytological preparations of the ascites and pleural effusions without prior cultivation / in vitro expansion were tested instead, as well as the primary tumour tissue using the established PLA.
Unfortunately, the Hsp90-p53 PLA could not be evaluated on ascites samples of three patients due to poor sample quality. In 4 out of 7 (57%) ascites samples Hsp90-p53 complexes could be detected. In contrast, Hsp90-p53 PLA signals were detected in all pleural effusions. Only three patients had consecutive ascites/pleural effusions. In two patients Hsp90-p53 complexes were present at all timepoints. The third patient was negative at the first timepoint but positive at the second. These few patients did not allow any correlation with therapy response or clinical data. Although anecdotal, these results indicate no predictive value of the presence of Hsp90-p53 complexes for responsiveness to Ganetespib.

Circulating tumour cell analysis
The objective of this translational research project was to determine whether the presence of CTCs in whole blood before and during treatment is a suitable marker for monitoring patients and determining their response to therapy. To tackle this question, blood samples were taken from the patients enrolled in the Phase II GANNET53 trial and the EUDARIO trial, the latter is still ongoing.
Of the 129 patients included in phase II of the GANNET53 study, blood samples were taken at start of the first cycle of treatment (C1D1) and on the following day (C1D2), at start of the second (C2) and third (C3) cycle, followed by blood draws at every second cycle thereafter (C5, C7, etc.) until progression of the disease.
Per patient, on average four (range 1-9) blood samples were available for the analysis of CTCs. Blood was drawn into Cell-free DNA BCT tubes (Streck, Inc.) and transferred to the central CTC laboratory for further analysis until noon of the following day. There, the samples were processed using a pre-enrichment step employing density gradient centrifugation, followed by a final microfluidic enrichment of the target cells using the Parsortix™ system (Angle plc., UK). In total, 522 blood samples were taken from the GANNET53 study patients. To detect the enriched CTCs at the molecular level and further characterize these cells, a panel of 28 genes including EpCAM and CK19 as universal markers for epithelial cells were selected based on the results obtained from previous studies and on in silico research (AGR2, CCNE2, CDH1, CDH2, CDH3, CDH5, CK19, EMP2, EPCAM, ERBB2, ERBB3, ERCC1, ESR1, FN1, FXYD3, GPX8, HJURP, LAMB1, MAL2, PGR, PLAT, PPIC, PRAME, S100A16, SCGB2A2, TFF1, TUSC3, and VIM).
From the blood samples CTCs were enriched as described above, total RNA was extracted from the enriched and lysed (possibly CTCs containing) cell fraction using the RNeasy Micro Kit (Qiagen) and converted into cDNA using the SuperScript VILO Mastermix (Invitrogen). To increase the sensitivity of the overall approach, a specific pre-amplification of all target genes was performed using the TaqMan PreAmp Master Mix (Life Technologies). qPCR was performed in duplicate reactions on the ViiA7 Real-Time PCR System with default cycling parameters. The raw data were analysed using the Viia7 Software v1.1 with automatic threshold setting and baseline correction. From each replicate the mean Ct-value was calculated. Replicates with just one Ct-value detected as well as mean Ct-values ≥35 were regarded as negative results. Due to the large number of samples and gene targets to be analysed, qPCR was done in three batches, with batch 1 including the very first patients to be off treatment (n=44), with batch 2 including patients treated with at least five cycles of chemotherapy (n=14), and with batch 3 including all remaining patients. Preliminary analyses of batch 1 and 2 aimed to identify potential candidate genes, which may indicate the progression of the disease. For this purpose, we compared the transcript levels of each gene with the results of the clinical assessment of the disease status performed at the time of the respective blood draw. From all 28 genes, we found seven genes (ERCC1, ERBB3, CDH1, ESR1, HJURP, CCNE2, and CDH3) to be potential candidates for the differentiation of progressive disease, stable disease, or partial remission. PCR results differed statistically between these three groups with p-values between 0.02 and 0.08 (ANOVA, Dunn’s multiple comparisons test).
An increase or decrease in gene expression levels during the treatment may reflect the direct or indirect effect of the administered treatment on CTCs. The data from all batches were normalized by subtracting the reference gene to remove batch effects and analysed using LIMMA (linear models for microarray data). Subsequently, data were compared to C1D1 as a control group using patient ID as a blocking factor. P values were adjusted for multiple testing using the Benjamini-Hochberg Method. Given the exploratory design of the study, an adjusted p-value of 0.25 was considered as significance limit. Four to 16 differentially expressed genes (DEGs) were identified for the time points C2, C5, and C7. No DEGs were found at C3. Significantly increased ERCC1 gene expression levels were observed at C5 and C7 as compared to C1D1 (adjusted p-values 0.0014 and 0.000002 respectively). It can be assumed that the presence of CTCs that express ERCC1 is indicative of therapy failure.
To assess the impact of the transcript levels on the patients’ outcome, results obtained from all batches were pooled. Blood samples from six patients were excluded from the final analyses because these patients withdrew their consent to participate in the study. Furthermore, samples with a high Ct-value of the reference gene indicating a poor RNA quality or insufficient amount of starting material were excluded, resulting in a final number of 114 samples taken at C1D1, 108 samples taken at C1D2, and 99, 78, 43, and 19 samples at C2, C3, C5, and C7, respectively. At each cycle of treatment, the samples were stratified into two groups: samples with a mean Ct-value of <35 were classified as positive for the respective gene transcript, and samples with a mean Ct-value ≥35 as negative. The mean Ct-value of 35 was chosen as threshold for all genes, except VIM and ERCC1. For these genes, the threshold value for stratifying the patients was the median Ct-value calculated from all samples, i.e. a Ct-value of 25. For TFF1 was excluded from further analysis because most blood samples were TFF1-negative. Overall survival (OS) and progression-free survival (PFS) were defined as the time from the date of the blood draw at each respective cycle to the date of last contact or documented death, and to the date of progression, respectively. The association of the two groups (positive vs. negative for gene X) and survival was assessed at every time-point of blood draw using Kaplan-Meier curves and log-rank (Mantel-Cox) tests. Patients with increased gene expression of LAMB1 and SCGB2A2 at baseline (C1D1) had a significantly higher risk to die from the disease than patients without or less LAMB1 (HR 1.63 95% CI 1.006-3.140; p=0.049) and SCGB2A2 (HR 2.03 95% CI 1.101-6.457; p=0.031) gene expression in the enriched CTCs. High ERCC1 gene expression before treatment (p=0.005) and furthermore at initiation of each further cycle of treatment until C5 (C2: p=0.005; C3: p<0.001; C5: p=0.028) was associated with a significantly higher risk for progression of the disease. At C7 the lack of statistical significance (p=0.092) is likely due to the small number of samples taken at that time point. In contrast, ESR1 gene expression was associated with better patient outcome. Similar to ERCC1, the difference in PFS was statistically significant from C1 (p=0,002) throughout to C5 (C2: p=0.023; C3: p=0.028; C5: p=0.003).
To further characterize the potential of ERCC1 and other gene transcripts as useful markers for monitoring the disease, it was investigated whether the presence of a specific gene transcript beyond the defined threshold value correlates with progressive disease (PD) proven by radiologic imaging at the same time point. The proportion of positive findings was assessed in all samples at initiation of treatment taken at C1D1 (these are patients with progressive disease per definition) or with radiologically confirmed PD during treatment (total n=160), and the proportion of negative samples in all samples with partial remission (PR), stable disease (SD), or complete response (CR; total n=115) as follows:
Sensitivity = (Σ positive samples) / (Σ all samples taken at C1D1 or PD)
Specificity = (Σ negative samples) / (Σ all samples taken at SD, PR, or CR)
Accuracy = (Σ positive samples+ Σ negative samples) / (Σ all samples)
ERCC1 was the gene transcript characterized by the largest numerical value (1.8) for sensitivity+specificity+accuracy; in addition, the presence of ERCC1 beyond the chosen cut-off was again significantly associated with PD (Fisher’s exact test, p=0.002). Furthermore, CDH1 (p=0.038) VIM (p=0.003) and ESR1 (p<0.001) were significantly related with PD, albeit at low sensitivity and specificity.
To test whether certain combinations of gene transcripts can detect PD at a higher sensitivity and specificity as single markers, seven markers characterized by their high specificity of >85%, namely LAMB1, SCGB2A2, AGR2, TUSC3, GPX8, CDH3, and PRAME (7-gene panel) were combined. In a second approach, we combined ERCC1, as the gene marker with the highest accuracy was combined with the seven-gene panel and the 7 genes individually. The addition of further gene transcripts to ERCC1 did not increase the accuracy in detecting PD substantially. However, the sensitivity of ERCC1 as a marker for PD was increased by adding the above mentioned seven gene markers from 72.5% to 83.8%, albeit on the cost of a reduced specificity (46.1 % vs. 29.6 %).
In contrast to ERCC1, that proved to be useful for detecting/predicting PD, ESR1 gene expression indicated a lower risk of PD and death. Thus, the combined effect of ERCC1 presence and ESR1 absence on PD was investigated. Indeed, that combination had the best accuracy to predict PD (p<0.001).
As the combination of ERCC1 presence and ESR1 absence had the highest accuracy (65.5 %) to predict PD, the possible impact of this combination on PFS was investigated. Blood samples were stratified into four groups: ERCC1+/ESR1+, ERCC1-/ESR1-, ERCC1+/ESR1-, and ERCC1-/ESR1+. The outcome of these four groups was evaluated in a landmark analysis, by designating the time point of the respective blood draw as landmark time and by analysing only those patients who have not progressed until the landmark time. PFS was significantly different between the four groups of patients, with the ERCC1-/ESR1+ group surviving longest without progression of the disease. The differences in PFS were most pronounced at C1 (log-rank p<0.001) but still statistically significant at C2 (p=0.004) C3 (p=0.003) and C5 (p=0.004).
The percentage of ERCC1-/ESR1+ samples increased with the cycles of treatment, suggesting that patients who survive longer and receive more cycles of treatment are more likely to be ERCC1-/ESR1+.
Summary and conclusions: These results strongly suggest that CTCs before and during treatment are a suitable marker for monitoring ovarian cancer patients and determining their response to therapy. Beyond enumeration, the molecular characterization of these cells generated valuable knowledge on prognostic and predictive markers, such as ERCC1. Recently, platinum-resistance was related to CTCs expressing ERCC1, a key gene of the nucleotide excision repair pathway and essential for the removal of platinum-induced DNA damage. Since early studies reported the association of ERCC1 with cisplatin resistance in ovarian tumours and cancer cell lines, clinical trials suggested that ovarian cancer patients with low ERCC1 levels benefit preferentially from cisplatin-based chemotherapy. In addition to ovarian cancer, the role of ERCC1 in the mechanism of platinum resistance has been evaluated in other types of cancer, including head and neck cancer, non-small cell lung cancer, and gastrointestinal cancer. It was reported that the appearance of ERCC1 mRNA in extracellular vesicles was significantly related with disease progression in metastatic breast cancer patients, and that ERCC1 gene expression was predominantly seen post treatment in CTCs.
In breast cancer, the capacity of DNA repair has been reported to be associated with oestrogen receptor expression. A recent meta-analysis including 35 publications and almost 6000 ovarian cancer patients prove that the expression of ER in the tumours, especially of ERα, was a positive predictor of OS. Furthermore, the expression of ER has been linked to epithelial-to-mesenchymal transition (EMT) in prostate cancer, and thus may foster the release of CTCs into the circulation.
In the Phase II GANNT53 clinical trial, the gene expression levels of ESR1 and VIM, a marker of EMT correlated significantly, whereas a negative correlation of ESR1 and ERCC1 was observed. ESR1 may downregulate DNA damage response in CTC-positive patients and thus contribute to a better survival. It cannot be excluded that some tumours that do not shed cells into the blood stream do express ESR1. However, since it is often very difficult to get access to tumour tissue, CTCs represent the diagnostic target of choice for molecular characterization.


>> MOUSE MODELS AND HUMAN XENOGRAFTS (WORK PACKAGE 7)

Acute removal of mutant p53 allele in SCID mice with transplanted mutant p53 tumours
The aim was to generate Genetic Proof-of-Principle data via acute ablation in mice to definitively answer whether stabilized mutp53 in established cancers is essential for tumour maintenance. Not seeing a mutp53-deletion effect would have weakened the notion that stabilized mutp53 R248Q is a strong cancer target but would still warrant to study other hotspot mutp53 alleles in KI models and human cancer cell systems (see below). R248Q is a frequently observed TP53 mutation and was chosen as a representative example.
The first conditionally inactivatable mutp53 KI mouse (harbouring a floxed humanized p53 R248Q knock-in (KI) allele (called ‘floxQ’) and a non-leaky Tamoxifen (4OHT)-responsive Rosa26CreERT2 knock-in gene) was generated and validated as the most stringent genetic system to definitively probe this clinically central question: Does acute removal of mutp53 lead to regression of established cancers, and/or render tumours less aggressive and more chemosensitive? Does it prevent metastasis? Thus, these are mutp53 KI mice until oral Tamoxifen/4OHT causes rapid and efficient deletion of the mutp53 allele. Mutp53 tumours of inactivatable floxQ/-; ERT2/+ mice were transplanted into SCID mice for acute allele removal. These floxQ/-; ERT2/+ KI mice spontaneously developed primarily aggressive T-lymphomas. (only a few B-lymphomas and sarcomas and no carcinomas were observed). The fate of lymphomas from floxQ/-; ERT2/+ KI mice after acute removal of mutp53 was determined by a transplantation approach, which offers the extra advantage of precise assay standardization.
Fresh tumours from 5 moribund floxQ/-; ERT2/+ mice were processed into single cells, split into aliquots and tail vein-injected (1 Mio. cells) into SCID recipients (16 recipients per each of the 5 original tumours) that did or did not receive 4OHT in drinking water (started 48 h prior, sugar water as mock). Orally administered 4OHT gave inconsistent results for removal of mutp53. Therefore, mice were treated with intraperitoneal injections of Tamoxifen in corn oil vs corn oil alone (vehicle/mock). Tamoxifen/mock-treated lymphomas were further split into Cyclophosphamide or mock treatment groups. SCID recipient mice were analysed for differential survival, cumulative lymph node tumour mass and by autopsy/histology. Mutp53 deletion in Tamoxifen-treated SCID tumours were verified by lack of p53 immunostaining (immunoblot, immunohistochemistry, and immunofluorescence).
Results and conclusions: To rigorously validate mutp53 as a drug target in a native organismal context, conditionally inactivatable p53 R248Q (‘floxQ’) mice were generated and crossed with Rosa26CreERT2 (‘ERT2’) transgenics as a definitive genetic system (and faithful human Li-Fraumeni syndrome model) to probe this central question. FloxQ;ERT2 mice mirrored the constitutive Q mice in all phenotypes, e.g. developed mainly T-lymphomas with some B-lymphomas and sarcomas and exhibited the same shortened survival compared to the p53null allele, revealing the gain-of-function (GOF) of the Q allele. Tamoxifen/4OHT treatment activated their CreERT2 recombinase reliably, causing rapid deletion of the floxQ allele in 100% of mice with ~90% efficiency. Concordantly, acute removal of the floxQ allele by 4OHT induced cell death in short-term primary T-lymphoma cultures, but not in various controls. Importantly, as shown by tumour transplantation assays into immunocompromised hosts (tail vein allografts into SCID mice), Tamoxifen-mediated floxQ deletion markedly curbed tumour growth in vivo and prolonged survival of recipient mice compared to Q/- and p53-/- control tumours, which did not respond. Moreover, removal of the mutp53 allele by Tamoxifen chemosensitised such tumours and extended survival of recipient mice.
In sum, the data indicates tumour dependence on sustained expression of high levels of mutp53 for maintenance within the context of T-lymphoma. These proof-of-principle data identify mutp53 protein as a cancer-specific drug target.

Is mutant p53 a biomarker for Ganetespib efficacy? Cultured cells and xenografts
It had been shown previously that the Hsp90/HDAC6-targeting drugs 17AAG or its hydrophilic derivative 17DMAG (first generation ansamycin-based Hsp90i) and SAHA (HDAC6i) preferentially kill mutp53 over wild-type p53 and p53 null human cancer cells, largely due to their ability to degrade mutp53 by reactivating (the mutant p53-degrading) E3 ligases Mdm2 and CHIP. This had been demonstrated in breast, colon, and prostate cancer cells in culture and in xenografts. Based on these data it was the aim to test whether mutant p53 ovarian cancer cell lines are also more sensitive to Ganetespib-induced cell killing than wild-type p53 or p53 null ovarian cancer cell lines.
To this end, 25 human ovarian cancer cell lines were characterized for growth in vitro and in some cases for growth as xenografts. The TP53 mutational status was determined by sequencing. p53 expression was analysed by immune-detection of p53 protein levels (Western Blot). Relative sensitivity to Ganetespib was assessed by using the ATP-content based cell viability assay (Promega) after treatment of the cultured cells with increasing concentrations of Ganetespib (ranging from 1-5000 nM) for 72 h. Based on these results the Ganetespib concentration where 50% of all cells die or stop growing was calculated (defined as IC50 value). The functional consequences of Ganetespib were analysed by immune-detection of mutant p53 levels and other Hsp90 clients upon treatment of the cultured cell lines with varying concentrations of Ganetespib.
It needed to be demonstrated that Ganetespib can block the tumour growth of ovarian cancer cells in vivo and that the effects of Ganetespib depend on mutant p53 expression. Thus, the in vitro results were validated in in vivo experiments by using the suitable cell lines for subcutaneous xenograft experiments. Of the 25 cell lines tested, 3 cell lines were suitable for this assay because of their fast in vitro growth and their tumourigenicity in vivo. The remaining cell lines were not suitable due to very slow in vitro growth and/or no outgrowth in xenografts. After establishment of subcutaneous tumours, the mice were treated with Ganetespib or vehicle as control.
Results and conclusions: Four cell lines (Caov-3, BG-1, SKOV-3 and Ovcar-5 cells) were found to constitute p53 non-expressors. Sequencing of the TP53 gene locus revealed that 3 cell lines (A2780, Cov434 and Colo704) harbour wild-type TP53. Cov504 and Ovsaho cells expressed truncated wild-type p53 with a deletion (P332del) or substitution of amino acids leading to the expression of a stop codon (R342stop). The remaining 16 cell lines were classified as missense mutant p53 expressors. The relative cytotoxicity response of these cell lines to increasing concentrations of Ganetespib was assessed using the cell viability assay that determines the relative ATP content in the cells. All cell lines respond to Ganetespib in a dose-dependent manner. The IC50 concentration for Ganetespib in all ovarian cancer cells ranged between 5 nM and 150 nM Ganetespib after 72 h incubation time, except for Ovsaho and Ovkate cells. Ganetespib sensitivity in ovarian cancer cells was independent of the p53 status. Therefore, mutant p53 expression does not seem to be a suitable biomarker for the sensitivity to Ganetespib in ovarian cancer.
The Ovsaho and Ovkate cell lines tolerated very high Ganetespib concentrations (IC50 > 5000 nM). Therefore, these cell lines were further characterised (see below).
To investigate the functional consequences of Ganetespib treatment in ovarian cancer cell lines the protein levels of known Hsp90 clients in Ganetespib-treated, mutant p53-expressing cell lines EFO21, Skov-6 and Ovcar-3 cells were analysed. mutp53 as well as the Hsp90 client proteins Akt, Chk1 and Wee1 were degraded upon Ganetespib treatment. It is concluded that degradation of these Hsp90 clients partially contribute to the cytotoxicity caused by Ganetespib treatment in ovarian cancer cells.
To determine how Ganetespib can kill ovarian cancer cells, the cell cycle distribution of Ganetespib-treated Ovcar-3, ES-2 and Ovcar-5 cells was analysed. Ganetespib caused an accumulation of cells in G2/M. Further analysis showed that a remarkable number of cells became positive for phospho-H3, a marker for mitotic arrest. These results suggest that Ganetespib treatment of ovarian cancer cells causes degradation of Hsp90-client proteins, including cell cycle regulators such as Chk1 and Wee1, leading to mitotic arrest and ultimately to mitotic catastrophe as the mechanism of tumour cell death.
Since in vitro results may not be representative for the in vivo situation in tumours, the effect of Ganetespib treatment on tumour growth in xenograft experiments of ovarian cancer cell lines was investigated. Cells were subcutaneously transplanted into SCID mice. After tumour growth to a size of 300 mm3, one group of mice was treated with vehicle (as control), whereas the other group of mice received Ganetespib treatment (every third day for ES-2 cells, every fifth day for BG-1 cells and once a week for Ovcar-5 cells). Ganetespib treatment effectively inhibited tumour growth of ES-2 cells in vivo. Established ES-2 tumours were prepared 24 h after acute Ganetespib treatment to analyse mutant p53 level by Western blot analysis. Mutant p53 was not degraded in Ganetespib-treated xenografts, although tumour growth was effectively inhibited. Immunohistochemistry detection of mutant p53 in treated ES-2 xenograft tumours validated findings, that mutant p53 is not degraded after Ganetespib treatment, at least not at Ganetespib concentrations that effectively inhibited ES-2 tumour growth in vivo. These results show that Ganetespib - mediated tumour growth inhibition is not dependent on degradation of mutant p53, but may be dependent on other HSP90 clients and their depletion upon Ganetespib treatment. In line, tumour growth of xenografts with p53 non-expressing ovarian cancer cells were effectively inhibited by Ganetespib treatment in a similar extent than mutant p53 expressing ES-2 tumours. Concluding, in vitro and in vivo that mutant p53 is not a biomarker for the Ganetespib-response in ovarian cancer cells. Instead, the majority of ovarian cancer cells are sensitive for Ganetespib treatment irrespectively of mutant p53 expression.

Characterization of Ganetespib resistance in Ovsaho and Ovkate cells
The Ovsaho and Ovkate cell lines tolerated very high Ganetespib concentrations and displayed cross-resistance to other HSP90 inhibitor classes such as 17-AAG and PU-H71. Thus, they were classified as Ganetespib-resistant ovarian cancer cell lines. It could be excluded that UGT1A overexpression leads to Hsp90 inhibitor resistance in these cell lines. Analysing the level of the HSP90 clients Wee1 and Chk1 in Ganetespib-treated revealed the degradation of Wee1 and Chk1 in Ovsaho whereas in Ovkate cells only Chk1 was getting degraded. Other markers for intracellular activity of Ganetespib, such as the accumulation of phosphorylated H3 (pH3) or a transient upregulation of the HSF1 transcription factor (pHSF1) validated a partial but weak activity of Ganetespib in both cell lines. In line, a prominent G2/M arrest with a loss of G1, which is normally displayed in other Ganetespib-treated ovarian cancer cell lines is not found in Ovkate and Ovsaho cells. Concluding, it is hypothesised that HSP90 inhibitors are active in Ovsaho and Ovkate cells leading to cell death for a portion of the cells by mitotic cell death. However, most of the cells are G1- or S-phase arrested which rendering these cell population insensitive to Hsp90 inhibitor treatment.

Causality: Determine if Ganetespib’s anti-tumour action depends on mutant p53
To test this causality in ovarian cancer, selected mutant p53 cell lines were used to generated mutant p53 knockdown cells by lentiviral transduction of these cells with constructs encoding shRNA against p53. These mutant p53 control and knockdown cells were then tested for their in vitro response to Ganetespib, using the cell viability assay and immune-detection of the apoptosis marker PARP. Following these in vitro experiments, mutant p53 control and knockdown cells in xenografts were tested treated with vehicle or Ganetespib. In a second experimental set-up the relative drug sensitivity of control and mutant p53 knockdown cells in response to Taxol and Cisplatin treatment, alone or in combination with Ganetespib was tested. Read out was the relative cell viability as assessed by the Cell Titer Glo assay (Promega) 72h after treatment as well as the level of PARP cleavage as a marker for apoptosis 48h after drug treatment.
Results and conclusions: To generate mutant p53 knockdown cells, two different lentiviral vector systems were used (GIPZ [Thermo Scientific Fisher] and plko [Sigma-Aldrich]). After viral transduction, luc control knockdown and mutant p53 knockdown cells were stably selected with Puromycin. Ovcar-3, Skov-6, EFO21 and ES-2 mutant p53 knockdown cells with an overall mutant p53 knockdown efficiency of 70-95% were successfully established.
Next, mutant p53 control and knockdown cells were treated for 72 h with 30 nM Ganetespib (IC50 value), followed by the assessment of cell viability using the Cell Titer Glo assay (Promega) (Figure 7-11A). In parallel, protein lysates from untreated and Ganetespib-treated mutant p53 control and knockdown cells were prepared and analysed for levels of cleaved PARP as a marker of apoptosis. Both assays showed that the presence or absence of mutant p53 in EFO 21, Ovcar-3, ES-2 and Skov-6 cells does not modify the response to Ganetespib. All cells respond to Ganetespib to the same extent, leading to the induction of apoptosis and decreased cell viability.
In conclusion, a causality between mutant p53 expression and sensitivity to Ganetespib treatment could not be demonstrated.
The GANNET53 project aimed at evaluating the benefit of a combination therapy with Ganetespib and Taxol (Paclitaxel). One cannot rule out that mutp53 might affect the responses to this combination therapy. It is known that mutp53 contributes to the chemo-resistance of tumour cells. Thus, mutant p53 expression might prevent an effective response to Cisplatin or Taxol, whereas addition of Ganetespib and therefore degradation of mutant p53 might re-sensitize the cells to Taxol or Cisplatin treatment. To address this question, cytotoxicity tests were performed as well as western blot analysis of mutant p53 control and knockdown cells that had been treated with Ganetespib + Taxol or Ganetespib + Cisplatin compared to the single drug treatments.
First, the IC50 drug concentrations of ovarian cancer cell lines were determined after 72 h treatment with Taxol (paclitaxel) or Cisplatin or the Cisplatin derivative Carboplatin. All cancer cell lines were sensitive for Carboplatin and Taxol (Paclitaxel) at various degrees. Moreover, a tendency towards a higher Taxol resistance in mutant p53 cell lines compared to p53 non-expressing or wild-type p53 expressing ones was visible.
Next, drug combinations were analysed for possible synergistic action. Cells were treated with Taxol, Cisplatin, Carboplatin or Ganetespib alone or in combination of Ganetespib and one of the other drugs. The relative cell viability was assessed 72 h after drug treatment using the Cell Titer Glo assay (Promega). The combination index (CI value) was calculated according to Chou and Talalay, to reveal possible antagonistic or synergistic effects of the drug combinations. The combinational treatment using Taxol and Ganetespib lead to additive or even antagonistic effects only, whereas Carboplatin and Ganetespib or Cisplatin and Ganetespib together acted highly synergistic in most of the cell lines. However, decreased cell viability does not necessarily mean increased induction of apoptosis, as senescence or cell cycle arrest might also lead to decreased cell viability as detected by the Cell titer glo assay. Thus, the level of apoptosis induction was assessed in single and combination treated cell lines by immunoblot detection of cleaved PARP1, a marker for apoptosis induction. The combined treatment of mutant p53 expressing ovarian cancer cell lines with Taxol and Ganetespib did not lead to an increased induction of apoptosis, as shown by comparing the amount of PARP cleavage in single and combination treated cell lines. Vice versa and in line with the cell viability assays, combination treatment of especially mutant p53 expressing cell lines with Cisplatin and Ganetespib led to a significantly higher induction of apoptosis.
In conclusion, in in vitro experiments using established ovarian cancer cell lines, Ganetespib treatment sensitized ovarian cancer cell lines for the cytotoxic effects of Cis- and Carboplatin, but not to the effects of Taxol. Moreover, especially mutant p53 expressing ovarian cancer cells seemed to be more sensitive to the combination treatment with Platin/Ganetespib. One possible explanation for this lack of synergism with Taxol and Ganetespib might be the fact that both drugs act in the same phase of the cell cycle (namely mitosis). Thus, the activity of one drug and the resulting mitotic cell death makes the activity of the other drug dispensable. Instead, the cytotoxic effects caused by platin drugs, such as DNA damage due to intrastrand crosslinks, is effectively potentiated by the addition of Ganetespib. In this setting, Ganetespib might inhibit the repair of these platin-induced lesions due to the depletion of Hsp90 clients involved in DNA repair and cell cycle regulation, and therefore abrogation of an intra S- and G2 arrest that normally gives a cell time to repair its damaged DNA.
One of the gain-of-functions of mutant p53 represents increased chemoresistance. In line, a tendency towards increased Taxol resistance in mutant p53 expressing ovarian cancer cell lines was observed as well as a higher chemo-sensitisation of mutant p53 expressing ovarian cancer cell lines for the combination treatment Ganetespib plus Carboplatin. It was investigated if the abrogation of mutant p53 renders cells more sensitive to Taxol or Carboplatin, thereby offering one explanation for the sensitization effects of the HSP90 inhibitor Ganetespib. To this end control and mutant p53 knockdown ovarian cancer cell lines were treated with either Cisplatin or Taxol and monitored cytotoxic effects by analysing the relative cell viability or level of apoptosis induction by immunoblot analysis of PARP cleavage. However, no significant changes in the sensitivity to Taxol or Cisplatin due to depletion of mutant p53 could be observed.
These results can be due to the use established ovarian cancer cell lines. Such cell lines are derived from advanced ovarian cancer patients who mostly had received several rounds of chemotherapy cycles beforehand. Th cells are extremely genomically unstable, leading to a high number of mutations and genomic rearrangements. It is proposed that mutant p53 is a uniform event in high grade serous tumours that happens very early in cell transformation. Mutation of p53 is an important early step in tumourigenesis, however later, at the stage when permanent cell lines can be established, other mutations have taken over as driver of tumour maintenance and metastasis, rendering mutant p53 dispensable for tumour cells to survive. In conclusion, ovarian cancer cell lines might not reflect the in vivo situation in patients.

Determine if Ganetespib blocks mutant p53-driven tumourigenesis in Knock-In mice
The aim of this task was to determine whether preventive and/or therapeutic inhibition of Hsp90 by Ganetespib can specifically intercept mutp53-driven tumour formation and progression in vivo. There is compelling evidence from genetic mouse models that the oncogenicity of missense mutant p53 alleles - and the survival of the ensuing tumours- profoundly depend on Hsf1-mediated (and thus largely Hsp90) chaperone support and proofs that a powerful co-oncogenicity between Hsf1 and mutp53 is at work in the organism. In contrast, Hsf1 cooperativity is not found in the case of the p53 null allele, indicating a fundamental difference in the genesis of tumour formation in the presence or absence of mutp53.
In full support of the GANNET53 treatment concept, highly encouraging results in mutp53 R172H knock-in mice (in short ‘H’ mice) treated with 17DMAG+SAHA were obtained earlier. Complementary Hsp90/HDAC6 axis blockade via this drug treatment dramatically suppressed T-lymphoma formation in H/H mice but had no effect in p53 null control mice. Constitutive mutp53 R248Q KI mice (‘Q/-‘) were injected i.v. with the potent second-generation Hsp90 inhibitor Ganetespib or vehicle once a week until endpoint. A p53 -/- control cohort of the same background/age received the same treatments. Mice were sacrificed when moribund and analysed for therapeutic efficacy (tumour incidence, latency, size, histology, grade, necrosis, apoptosis, invasiveness, Kaplan Meier survival analysis). Two-way comparisons were made: (1) within each p53 genotype to see if drug effects exist per se; and (2) between genotypes to see if treated mutp53 KI mice show stronger therapeutic benefit than the respective p53 null mice. This study was repeated as therapeutic protocol, starting as soon as mice appear sick by combining Ganetespib with cyclophosphamide versus cyclophosphamide alone. Ganetespib could improve efficacy.in combination with the conventional chemotherapeutic Cyclophosphamide, which is clinically used but is quite toxic to fast growing normal tissues and hence causes serious side effects. Again, using clinically advanced autochthonous T-lymphomas (of similar sizes by ultrasound imaging, about 300 mm3), tested tumour growth was tested in response to vehicle, Cyclophosphamide alone, Ganetespib alone, and the combination of both.
Results: Constitutive mutp53 R248Q knock-in mice (Q/-) at 16 weeks of age with well-developed T-lymphomas were treated once intravenously via tail vein injection with Ganetespib (100 mg/kg) or vehicle. T-lymphomas were harvested 24 h later and subjected to immunoblot analysis. The results showed that Ganetespib treatment degraded mutp53 protein in tumour tissues in vivo. Moreover, when mice were treated lifelong with one weekly i.v. dose of Ganetespib versus DMSO vehicle, starting at 8 weeks of age at the stage of early thymus-restricted disease, the animals still died of tumours. But importantly, only mutp53 Q/- mice, but not their p53 null control littermates, benefited from the Hsp90 inhibitor treatment with significantly extended survival. A similar mutantp53-specific drug effect was seen with mutant p53 knock-in mice carrying a different hotspot missense mutation – ‘structural’ mutp53 R172H. Likewise, when treatment was started only late - in clinically advanced autochthonous T-lymphomas – which models the clinical situation best, Ganetespib treatment suppressed tumour growth of clinically advanced autochthonous mutp53 tumours better than p53 -/- tumours. For the combination therapy with Cyclophosphamide the dose was chosen to be so low as to have little effect on its own (100 mg/kg). Combining this dose with a low dose of Ganetespib (50 mg/kg) produced the best results, yielding robust tumour regression for at least 60 days in 3 out of 5 mice. Most notably, one mouse survived for 121 days in remission with only a very small shrunken detectable tumour, and then died of unrelated other cause.

Potential Impact:
Intended results, specific uses, and impact of the GANNET53 project:
The major expected result of the GANNET53 project is a significant progression-free survival (PFS) benefit for ovarian cancer patients with aggressive histological subtypes harbouring p53 mutations treated with the new therapeutic concepts based on Hsp90 inhibition (in the GANNET53 and the EUDARIO trials), compared to standard treatment (WP4).
The Phase II GANNET53 trial was negative for its primary endpoint PFS. The Phase II EUDARIO trial is ongoing and applies two promising Ganetespib combinations, namely 1) with Carboplatin, and 2) with the PARP inhibitor Niraparib. First results in EUDARIO on the primary endpoint PFS are expected in Q3 2021.
Further results (and resulting specific uses and impacts) include
• First clinical proof-of concept for the innovative mechanism of DNA repair inhibition by Hsp90 inhibition after induction of DNA damage by Carboplatin (EUDARIO trial, WP4): Impact on future treatment of ovarian cancer patient and of other tumour entities for which Carboplatin is part of standard treatment strategy
• First clinical proof-of-concept for 1) broadening sensitivity of ovarian cancers towards PARPi via generation of a BRCA-like phenotype by Hsp90 inhibition and 2) preventing/circumventing development of PARPi resistance by combination with a Hsp90 inhibitor (EUDARIO trial, WP4): Substantial impact on use of PAPRi in the treatment of ovarian cancer patients and other tumour entities, for which PARPi treatment is part of standard treatment strategy/is in clinical testing, as resistance to PAPRi in one of the major obstacles in the treatment with this drug.
• Evaluated life quality of the new experimental therapies based on Hsp90 inhibition in comparison with standard treatment options (and WP4): Generation of knowledge on well-being of patients and therewith impact on implementation of Hsp90 therapy approaches in ovarian cancer and other tumour entities.
• Established safety of Ganetespib in new combination with the taxane Paclitaxel (GANNET53 trial, WP3), with Carboplatin (EUDARIO trial, WP4), and with PARPI Niraparib (EUDARIO trial, WP4), respectively: Impact on future clinical trials using this new combination in ovarian cancer and other tumour entities
• A unique biobank of archival (FFPE and fresh-frozen tissues) and prospectively collected (tissue biopsies, ascites, blood) ovarian cancer biosamples before and during treatment (from 133 platinum-resistant ovarian cancer patients treated in the Phase II GANNET53 trial and 122 relapsed, platinum-sensitive ovarian cancer patients treated in the EUDARIO trial, WP4): Strong scientific impact, several research projects based on the collected biomaterials are ongoing/will be initiated e.g. p53 basic research including comparison of exact p53 status at diagnosis and at relapse, analysis of different biomarkers, enables future genome-wide oncogenomics studies to identify additional resistance-mediating molecular changes etc.
• An innovative software for effective organisation of a large multi-centre biobank and real-time tracking and distribution of biosamples (WP5): Application in other tissue-banking networks such as e.g. TOC (Tumour Bank Ovarian Cancer)-Network (http://www.toc-network.de) or ENGOT biobanking (https://engot.esgo.org) sale of software, license
• Development of multiple methods, e.g. a functional molecular test to detect mutp53 Hsp90 complexes in tumour tissues, approaches to analyse prion-like behaviour of mutant p53 proteins, and a circulating tumour cell enrichment and detection/characterisation system: Scientific impact on the respective research field, license, use in basic-research and potential expansion to other tumour entities
• Value of circulating tumour cells (CTCs) for monitoring responsiveness to experimental therapy with Ganetespib (WP6): Potential application in new therapeutic strategies for relapsed ovarian cancer
• In vivo Genetic and Pharmacologic Proof-of-Principle for the mutp53-targeting concept in engineered knock-in mouse models (WP7): Potential application in tumour types with mutant p53 as disease driver

Lead users of the expected results are clearly ovarian cancer patients and clinicians who provide treatment to ovarian cancer patients. Other lead users of the projects’ results also include the scientific community and biotechnology companies for the novel concept 1) of targeting the cancer-specific mutp53 ‘addiction’ through state-of-the-art Hsp90 inhibition (in the GANNET53 trial) and 2) of crucially inhibiting DNA repair by rapid decay of key components of the Fanconi anemia pathway as well as of cell cycle checkpoint mediators, both through Hsp90 inhibition, following DNA damage (in the EUDRIO trial), and the pharmaceutical industry by providing a strong stimulus for the further development of more advanced Hsp90 inhibitors in Europe. Indirect users include policy makers, health systems and the society at large.

Expected impact:
The GANNET53 clinical trial is based on the highly innovative concept of targeting the oncogenic mutp53 protein by Hsp90 inhibition in platinum-resistant metastatic ovarian cancer to improve survival. The EUDARIO clinical trial is based on the highly innovative concept of enhanced DNA repair inhibition by Hsp90 after induction of DNA damage by Carboplatin in platinum-sensitive metastatic ovarian cancer patients. A survival benefit is aimed in our consortium by a close collaboration of leading European gynaecological oncology experts, the use of a clinically far advanced and safe Hsp90 inhibitor, the compelling existing network and administrative knowledge of national trial groups, the participation of world-renowned scientists in p53 basic and translational research, and of three innovative SMEs. This will guarantee fast bench-to-bedside translation of innovative basic research findings into ultimate survival benefits for patients with dismal prognosis. Through these collaborations and interactions, the basic/clinical European scientific excellence is fully integrated. The project is perfectly in line with and ideally suited for the objectives of FP7 Cooperation Work Programme Health-2013 in improving the health of European citizens, and tightly adheres to the aims of topic 2: Translating Research For Human Health.

Ovarian cancer is by far the most fatal among gynaecological malignancies causing 42,000 deaths annually in Europe. The two clinical trials within the GANNET53 project, i.e. the GANNET53 and the EUDARIO clinical trials apply an innovative therapeutic concept in a stratified patient population, i.e. ovarian cancer patients with Type II tumours to achieve its major goal of significantly improving PFS of ovarian cancer patients. Type II tumours ubiquitously harbour p53 mutations (> 95%) as THE defining molecular abnormality. They not only account for the overwhelming majority (>70%) of epithelial ovarian cancer (EOC), but also represent the most problematic tumour type from a clinical point of view: they are highly aggressive, evolve rapidly and are highly metastatic. More than 70% of EOCs present with advanced metastatic disease (peritoneal carcinosis) already at the time of primary diagnosis. All relapsed ovarian cancer patients have metastatic disease. Our novel therapeutic approaches target a major driver of tumour aggressiveness, namely mutant p53 (in the GANNET trial) and crucially inhibit DNA repair by rapid decay of key components of the Fanconi anemia pathway as well as of cell cycle checkpoint mediators following DNA damage (in the EUDARIO trial). Thus, both of our approaches critically fight metastatic ability via an innovative Hsp90 (heat shock protein 90) inhibition mechanism. Thus, the GANNET53 project fully satisfies topic 2.4.1-1: investigator-driven treatment trails to combat or prevent metastasis in patients with solid cancer.

The major expected result is a significant PFS benefit for ovarian cancer patients treated with the new therapeutic concepts compared to standard therapy. The Phase II GANNET53 trial applying Ganetespib in combination with Paclitaxel, was negative for its primary endpoint PFS. The Phase II EUDARIO trial is ongoing and applies two promising Ganetespib combinations, namely 1) with Carboplatin, and 2) with the PARP inhibitor Niraparib. First results in EUDARIO on the primary endpoint PFS are expected in Q3 2021. If successful, our concepts will then be advanced to Phase III clinical trials in Type II ovarian cancer patients and have the potential to become the new standard of care in this group of patients. Moreover, if successful, our concepts of targeting Hsp90 have the potential to move up-front to first-line therapy and be applied at primary diagnosis of ovarian cancer, being per se a metastatic disease in >70 % of cases. At primary diagnosis, our concepts might not only be able to combat metastasis but also prevent the occurrence of a metastatic relapse. Thereby our concepts might be able to increase the patients’ chance for robust long-term remission and to achieve a higher cure rate.
Most importantly, the GANNET53 trial is a proof-of-concept trial. The trial did not confirm mutant p53 as a critical drug target in ovarian cancer. However, mutant p53 might be a critical target in other tumour entities that are driven by a p53 mutation (such as HER2-positive and triple-negative breast cancers, colon cancer, head & neck cancers, NSCLC lung cancer, glioblastoma and others). The EUDARIO trial is also a proof-of-concept trial to establish that Hsp90 inhibition can substantially increase sensitivity towards Carboplatin treatment via critically enhanced DNA repair inhibition. This concept has enormous potential for exploitation for other tumours entities treated with standard platinum-based chemotherapy, particularly in a p53 mutant background. Considering that Carboplatin is a common treatment pillar in various different tumour types and given that over 50% of all cancer patients have tumours with p53 mutations, this proof-of-concept trial has enormous exploitation potential for cancer treatment at large. Furthermore, the EUDARIO trial will provide first clinical proof-of concept for a combination with a Hsp90 inhibitor to prevent/circumvent the development of resistance to PARPi. This is of highest clinical relevance as PARPi treatment has revolutionised ovarian cancer therapy and ovarian cancer patient’s prognosis in the past years and as resistance to PARPi is one of the major treatment obstacles of these drugs. Thus, this clinical proof-of-concept has substantial impact on use of PAPRi in the treatment of ovarian cancer patients in general and on other tumour entities, for which PARPi treatment is part of standard treatment strategy.
Further results of the GANNET53 and EUDARIO trials include the establishment of safety of Ganetespib in new combinations with the taxane Paclitaxel, with Carboplatin and, for the first time, with a PARPi, respectively. This opens the opportunity for application of these promising combinations in other cancer types.
The value of circulating tumour cells for monitoring responsiveness to the experimental therapy will be established in the GANNET53 project. This will allow monitoring of treatment success in relapsed ovarian cancer patients. The GANNET53 project has created/is creating a unique biobank of archival (FFPE and fresh-frozen tissues) and prospectively collected (tissue biopsies, ascites, blood) ovarian cancer biosamples (WP5) before and during treatment, and innovative software for the documentation, effective management and utilisation of the biobank including real-time tracking of sample analysis (WP5). This provides a strong scientific impact as several research projects based on the collected biomaterials are ongoing/will be initiated e.g. p53 basic research including comparison of exact p53 status at diagnosis and at relapse, analysis of different biomarkers, enables future genome-wide oncogenomics studies to identify additional resistance-mediating molecular changes etc.

Impact on EU policies:
The GANNET53 project aims at providing a more effective treatment for ovarian cancer patients and at improving PFS. It follows the Directive 2001/20/EC on Clinical Trials of the European Parliament and investigates the efficacy and safety of Ganetespib in Type II ovarian cancer patients. Type II ovarian cancer is the major causes of death from gynaecological cancers in Europe. Since there is no reliable method for early detection of ovarian cancer and most patients present with advanced metastatic disease, the treatment of ovarian cancer is a great challenge and burden for the health care system. Current clinical management fails to take the heterogeneity of ovarian cancer into account. Our therapeutic approach rests on the underlying molecular oncogenic pathway of these highly aggressive Type II tumours, namely mutp53 (in the GANNET53 trial) and on crucial inhibition of DNA damage repair following DNA damage (in the EUDARIO trial), to achieve the most profound survival benefit. The identification of new and more effective therapies has significant beneficial implications in health care, as well as having beneficial societal and financial implications at large. Therefore, GANNET53 fully addresses the EU policies for the optimisation of national health systems. Furthermore, our concept will have substantial impact on preventing insufficient therapies in ovarian cancer patients and provide a tailored molecular therapeutic concept in mutp53 tumours, and thus follows the EU policies for safeguarding public health.
The structure of our network also guarantees community-added value and contributes to EU policies from the economic viewpoint. Our Consortium includes three SMEs that are dedicated to 1) R&D for identifying, specifying, and developing tomorrows’ advanced medical-IT solutions, 2) manufacturing and marketing of medical in vitro diagnostics, and 3) excellence in clinical trial design and execution, respectively. Each member of the consortium benefits from the cohesive integration of clinical centres, research groups and commercial units by sharing materials, technologies, know-how and data to generate a superior competitiveness for all groups that otherwise would not be achievable by smaller one-on-one collaborations. This is expected to result in highly efficient new treatments, accompanied by the development of innovative data-handling systems and a marketable functional molecular test that can predict responsiveness to the new experimental therapy and thereby allows to further stratify patients who will benefit most from the new therapy. The exploitation of these results will lead to the generation of new jobs beyond the scope of this project. About 16% of the total EC contribution is allocated to three SMEs, which will drive the growth and development of these SMEs in their areas of expertise. This will fulfil the EU policy to help the SMEs realise their growth potential, to promote entrepreneurship and to create a healthier business environment for them, which will consequently contribute to the European economic development and to help increase the competitiveness of the European Union.
The GANNET53 and the EUDARIO clinical trial are highly innovative both from a conceptual point of view and by approach. GANNET53 is the first clinical trial aiming to target mutp53 stabilisation and to achieve this goal by drug inhibition of tumour-activated Hsp90 chaperone, using the clinically most-advanced state-of-the-art and safe drug. EUDARIO is the first clinical trial to enhance DNA repair inhibition after platinum-induced DNA damage via the innovative mechanism of Hsp90 inhibition. Furthermore, EUDARIO applies for the first time a combination of an Hsp90 inhibitor with a PARP inhibitor to broaden sensitivity of ovarian cancers towards PARPi and to prevent/circumvent PARPi resistance. The overall concept of this project is original and unique. Furthermore, the engineered mouse models for in vivo genetic and pharmacologic Proof-of-Principle, the human ovarian cancer model systems to gain causative knowledge on the mechanism of drug action of Ganetespib, and the data handling systems to be developed within the trial are highly innovative. Thereby, GANNET53 is fully in line with the EU policy to drive sustainable growth and competitiveness through the stimulation of innovation.

Societal impact:
Worldwide, Europeans have the highest incidence of ovarian cancer where it is the fifth most diagnosed female cancer. Over half of women diagnosed with ovarian cancer will not live beyond five years. The current standard of care at primary diagnosis is cytoreductive surgery and adjuvant platinum-based chemotherapy. Maintenance therapy with Bevacizumab and/or PARPi are also crucial treatment pillars. However, 25-30% of patients show primary resistance to first-line platinum-based chemotherapy. Even worse, eventually all relapsed patients will become resistant to platinum after reiterative therapy with platinum-based regimens (acquired ‘secondary’ platinum-resistant disease). The burden of ovarian cancer for the society at large is not only due to its morbidity and mortality but also to the treatment impact itself, which has significant side effects and a low response rate when administered in an unselected patient cohort, leading to enormous burden on healthcare budgets.
Our project offers two completely new therapeutic strategies with the potential benefit to markedly prolong progression-free survival and improve quality of life. The main difference to standard genotoxic treatment options is that our approach in the GANNET53 trial also aims to targets the essential molecular driver (mutant p53) of the disease and does so by targeting the tumour-specific factor (Hsp90) of mutant p53 stabilisation, thereby carrying a high therapeutic index to normal tissue. Also, the EUDARIO inhibits Hsp90, which is pivotally upregulated in ovarian cancer tumours. Furthemore, a higher binding affinity of tumour Hsp90 compared to Hsp90 extracted from normal tissues has been demonstrated for Hsp90 inhibitors, thus providing the advantage of a broader therapeutic window compared to chemotherapy. Our strategies in the GANNET53 and in the EUDARIO trials are applied in a stratified patient population with highly aggressive Type II tumours, which ubiquitously harbour p53 mutations. Our approach addresses the obvious urgent social need to expand the therapeutic armamentarium to fight ovarian cancer. Thus, the GANNET53 project has important social and economic impacts. It will give the patients in a desperate situation the opportunity to receive more effective therapy with little additional side effects. Clinicians will be able to offer effective treatment and patient stratification to selectively apply the therapy to those patients (mutant p53 Type II tumours) who are most likely to benefit from the respective treatment. Consequently, continuing treatment cycles with insufficient responses and with costs from debilitating side effects that require additional treatments including extra hospitalisations can be avoided. Thus, the cost of treatment and ongoing care for advanced ovarian cancer patients will achieve a more favourable cost-benefit ratio.

Economic impact:
The Commission Staff Working Document, Impact assessment report on the revision of the Clinical Trials Directive 2001/20/EC is cited: „Conducting clinical trials entails considerable investment and growth in the EU, including inward investment by sponsors from non-EU countries. In recent years, a range of publications have highlighted these tangible benefits of clinical trials. The main part of GANNET53 project are two randomised clinical multicentre trials, which is designed, conducted, and reported in accordance with the principles of Good Clinical Practice (GCP). Compliance with GCP provides assurance that the rights, safety, and well-being of our patients are protected and that the results will be credible. GCP means blessing and burden at the same time. Whereas the blessing is obvious, the costs for conducting clinical trials increased dramatically within the last ten years. Thus, the local economies in the 5 EU countries will benefit from our clinical trials as an additional source of cash flow and by providing enhanced employment opportunities. Administrative and legal requirements based on the Clinical Trials Directive are implemented in our research project and also generate cash flow. Therefore, the economic impact of the GANNET53 project is evident by contributing to state budgets via taxes in Austria, Germany, Belgium, France and Italy.
Moreover, our research will provide alternative cost savings that can relieve the public healthcare system, since Synta Pharmaceuticals Corp. (Lexington, MA, USA)/Aldeyra Therapeutics, Inc. (Lexington, MA, USA) and TESARO/GSK, who are the developers of Ganetespib and of the PARPi Niraparib, respectively, agreed to provide the drugs for this trial at no charge. Of note, the drug Ganetespib is physically produced in Halle/Westfalen, Germany, by a German company. Thus, the launch of GANNET53 clearly contributes to the inward foreign capital investment in five European countries.
Results from our research project will strongly stimulate the European pharmaceutical industry to perform further investigations and investments on the continuous development and/or improvement of more advanced, rationally designed Hsp90 inhibitors that lack the toxicity and efficacy issues that the current European-owned Hsp90 inhibitors struggle with. Furthermore, the GANNET53 and the EUDARIO are Proof-of-Concept trials for degrading stabilised mutant p53 as a rational target for anticancer treatment and for enhanced DNA damage repair inhibition following DNA damage after Carboplatin treatment, respectively. Both innovative treatment approaches are applied in a p53 mutant background to achieve the most profound benefit. Since more than 50% of all human cancers are p53 mutated, positive results will carry enormous potential for exploitation in the entire field of oncology in general.
Beyond these economic issues, the implementation of our research project designed and conducted as a Phase I and Phase II clinical trials will provide a variety of employment opportunities for researchers, clinical staff (as indicated by the 919 person months that are directly applied for funding from the 22 participants), support business, regulatory committees, and the related pharmaceutical industry. Three SMEs are participating in the project, each bringing unique expertise and innovation to the task assigned. They themselves, plus the exploitation of their deliverables, i.e. innovative software for multicentre clinical data retrieval and for biobanking with real-time sample tracking, as well as the response-predictive test for the new therapy, certainly contribute to the European economic development and increased global competitiveness of the European Union.
Overall, performing the clinical trials GANNET53 and EUDARIO will be a strong stimulus in several economic arenas with positive spill-over effects throughout the European Union.

New knowledge for the scientific community:
The major finding of interest for the scientific community from the GANNET53 trial is the proof-of-concept that mutant p53 is a rational target for cancer therapy. The Phase II GANNET53 trial did not confirm the clinical proof-of-concept in ovarian cancer patients, but preclinical data confirmed a Genetic and Pharmacologic Proof-of-Principle in engineered mouse models. The inactivatable and constitutive knock-in mouse models raised our concept to the highest level of evidence and provide the most stringent causal proof by genetic (allele removal) and pharmacologic (Ganetespib treatment of mutp53 knock-in versus p53 null mice) acute mutant p53 ablation in vivo. These models will be of highest interest for the scientific cancer research community and provides potential for exploitation in p53-related research and cancer research in general.
The key finding of interest for the scientific community from the EUDARIO trial is the proof-of concept for the innovative mechanism of DNA repair inhibition by rapid decay of key components of the Fanconi anemia pathway as well as of cell cycle checkpoint mediators via Hsp90 inhibition after induction of DNA damage by Carboplatin in mutant p53 cancers that have lost the wild-type p53-mediated G1 checkpoint function. EUDARIO will provide the clinical proof thereof. Furthermore, EUDARIO will provide clinical evidence for broadened sensitivity towards PARPi in ovarian cancer patients by Hsp90 inhibition via generation of a BRCA-like phenotype. These two models will be of highest interest for the scientific cancer community and provide substantial potential for exploitation e.g. in other cancers treated with platinum-based chemotherapy and for broader and sustained usage of PAPRi in ovarian cancer patients and other tumour entities treated with PARPi.
Another key issue of scientific interest is addressed by biobanking of materials from patients participating in the two Phase II clinical trials. The carefully curated and clinically annotated retrospective and prospective biosamples in GANNET53 project will be a very precious collection of ovarian cancer materials before and during treatment, representing a treasure of greatest scientific value for gaining molecular insights within our project as well as for exploitation in future research tasks.

Added value in carrying out the work at a European level:
Even though 66,700 cases are diagnosed annually in Europe, ovarian cancer is defined as a rare cancer and our trial targets the subgroup of patients with the most dismal prognosis. A critical number of patients need to be enrolled in a relatively short period of time to ensure statistical power and reliability of results. Thus, a scale at the European level is crucial. To reach sufficient recruitment a close multinational collaboration throughout five European countries is needed. We established a highly efficient consortium with previously proven capability and manpower to perform this multicentre clinical trial and to assess our innovative therapeutic concept in this deadly disease. Our consortium involves highly experienced national clinical trial groups in gynaecological oncology and a high-volume University Centre in Belgium. This composition of the consortium ensures the efficient recruitment of patients with Type II platinum-resistant (GANNET53 trial) and platinum-sensitive (EUDARIO trial) ovarian cancer patients within the short active enrolment time frames.
GANNET53 and EUDARIO are complex clinical trials, which cannot be realised within a single institution or at the national level. Since the members of the consortium hold different expertise in the field of oncology in general, the cooperation will establish a smart network of clinicians, national trial groups, scientists and innovative SMEs contributing to added value to the European Community. Participants share the same general objectives but hold different expertise, experimental systems and philosophy of approach, thus complementing each other extremely well. The composition of our consortium is not casual or accidental, but reflects an already established network of interactions (EUTROC, http://www.eutroc.org; ESGO, http://www.esgo.org; OVCAD (FP6), http://www.ovcad.eu; TOC, http://www.toc-network.de; EORTC, http://www.eortc.be; and many more), which will be strategically expanded in the GANNET53 trial. Our clinical trial reflects the intention to merge the expertise of leading European gynaecological oncologists and outstanding scientists to form a powerful synergy.
This proposal has therefore the added benefit of formally coalescing many of these interactions into an organised effective framework. It is obvious that such a complex network of interactions cannot be achieved at the individual group or even national level.

Building up further cooperation:
Through establishment of the external advisory board, the international visibility and connections of the consortium and each participant to clinical and research networks and to organisations abroad will expand the advances within the EU to a worldwide level. By “networking the networks”, we believe that clinical and basic research on ovarian cancer will reach a new dimension, which could finally break through the longstanding wall blocking therapeutic advance and bring a major leap forward to benefit the affected patients.
The local economies in the participating countries benefit from our clinical trials as an additional source of cash flow and by providing enhanced employment opportunities. Administrative and legal requirements based on the Clinical Trials Directive are implemented and generate cash flow.

Main dissemination activities and the exploitation of results:
The consortium and the coordinator especially aimed at the dissemination of the efforts of the EC and the consortium on improving survival in ovarian cancer patients and to inform on all aspects of the GANNET53 project including its goals, concept, perspectives, results, and the potential impacts on health care systems in a target-group adequate manner.
Various information of the GANNET53 and of the EUDARIO clinical trials was made available to clinical communities, ovarian cancer patients, scientific societies, the pharmaceutical industry, biotechnology companies, and the public.
The “GANNET53” website (“www.gannet53.eu”) was created, which is tailor-made for different user groups, namely for specialists and medical staff, for patients and the public, and for the partners involved in the GANNET53 project. To adequately address the different needs of the different interest groups, three separate (but interlinked) websites have been created providing targeted information on different aspects of the GANNET53 project (i.e. scientific webpage, patient webpage, project partner webpage).
Various press releases and articles were published in public media. Lectures were held to both, an interested non-scientific and scientific audience at several medical and scientific events, to make the project and the EC`s efforts to translating basic research results into patient benefits broadly known.
The scientific results were submitted in the form of manuscripts to scientific journals. Two manuscripts were published, others are under review or in preparation:
Part I of GANNET53: A European Multicenter Phase I/II Trial of the Hsp90 Inhibitor Ganetespib Combined With Weekly Paclitaxel in Women With High-Grade, Platinum-Resistant Epithelial Ovarian Cancer-A Study of the GANNET53 Consortium. Front Oncol 2019 Sep 10;9:832. doi: 10.3389/fonc.2019.00832. PMID: 31552170.
Strong antitumor synergy between DNA crosslinking and HSP90 inhibition causes massive premitotic DNA fragmentation in ovarian cancer cells. Cell Death Differ. 2017 Feb;24(2):300-316. doi: 10.1038/cdd.2016.124. PMID: 27834954.

The clinical trials “GANNET53” and “EUDARIO” trial were registered in four publicly accessible clinical trial registries.
List of Websites:
http://www.gannet53.eu/