Skip to main content

Tackling the Achilles Heel of Immunotherapy: Validating imaging biomarkers and targeting the immunological niche of tumour hypoxia

Periodic Reporting for period 2 - HYPOXIMMUNO (Tackling the Achilles Heel of Immunotherapy: Validating imaging biomarkers and targeting the immunological niche of tumour hypoxia)

Reporting period: 2018-06-01 to 2019-11-30

The absence of effective anticancer treatment, in combination with the steeply rising prevalence of cancer, underscores the urgent need for more specific and effective cancer therapies coupled to biomarkers to select the right patient to the right treatment. Overall, this is a typical project of convergence sciences sitting at the intersection of clinical sciences (WP4), biology (WP1-2) and technology (WP3).

Immune checkpoint inhibitors (ICIs) are monoclonal antibodies that function by blocking the inhibitory interactions between immunosuppressive molecules, thereby relieving T-cells from negative regulation. This therapeutic approach is capable of eliciting robust anti-tumour immune responses in advance-staged cancer patients. However, clinical studies have shown that durable responses are limited to a minority of patients and malignancies, indicating underlying resistance mechanisms to ICIs. Accumulating evidence suggests that oxygen deficiency (tumour hypoxia) within tumours is an important phenomenon that can suppress the anti-tumour immune response via multiple mechanisms. It was therefore hypothesised that targeting hypoxia may improve the efficacy of ICIs and immunotherapy in general (IO) in murine syngeneic tumour models through lowering of tumour hypoxic fraction.

Hypoxia is a common feature of the majority of solid tumors and arises due to a disturbed balance between growth, proliferation and oxygen supply. Besides contributing to malignant progression, invasion and metastasis, tumor hypoxia has also been associated with resistance to several therapies including immunotherapy (IO). Using hypoxia-activated prodrugs (HAPs), hypoxic regions of the tumor can be selectively targeted (watch the animation: We have been using a next generation HAP (HAPng) that is reduced by endogenous one-electron reductases to an oxygen-sensing intermediate. Only under severe hypoxic conditions, further reduction of this intermediate leads to the generation of its active metabolite to exert its cytotoxic effect as DNA crosslinker. We aim to identify the key factors contributing to the cytotoxicity of HAPng.

In this project, we want to overcome treatment resistance and improve curative response by using appropriate biomarkers and addressing the issue of tumour hypoxia leading to immune evasion. Since up to half of solid tumors exhibit hypoxic tissue areas that are heterogeneously distributed within the tumor stroma, we hypothesized a beneficial therapeutic outcome of a trimodal therapy, more specifically the combination of immunotherapy (“pushing the accelerator” with L19-IL2 and “releasing the break” with checkpoint inhibitor (watch the animation: )) with HAPng. We have obtained experimental evidence that HAPng can increase the immunological visibility of tumors in which cancer cells secrete immunostimulatory molecules upon their death, a process called immunogenic cell death (ICD). In addition, using the concept of HAPs, we will bring immunomodulators targeting TLR7/8 in these hypoxic areas in order to stimulate the activation of pro-inflammatory cytokines.

Furthermore, to establish an ideal individualized therapy, the identification of biomarkers of therapy efficacy is similarly important. Thus, it would be beneficial for patients as well as being cost-effective if the outcome of therapy could be predicted based on predictive biomarker(s) assays to guide the most optimal treatment selection for each individual patient. We envision biomarker(s), preferably tri-dimensional, imaging-based and non-invasive such as radiomics approaches, can be a powerful tool to maximize a therapeutic benefit for patients.

Another way to tackle hypoxic cells is to use radiotherapies with a high linear energy transfer (hLET) such as carbon ions. Compared to conventional photon radiotherapy, carbon ions will have a higher deposition of the actual radiation dose at the tumor site without b
We have demonstrated that the synergistic effect of radiation is greater with immunocytokine then with checkpoint inhibitor and that a immunotherapy combining a “release the break approach” with checkpoint inhibitor and a “push the accelerator approach” with immunocytokine is more efficient then one of the two approaches.

The HAPng has been investigated thoroughly as monotherapy in several tumor models showing a therapeutic efficacy independent of the baseline hypoxic fraction found before treatment initiation, indicating the involvement of three potential resistance mechanisms currently under investigation: 1) lack of hypoxia, 2) DNA repair (intrinsic sensitivity to alkylating agents) and 3) reductase activity. Overexpression of sPOR, a key Nitroreductase, in H1299 resulted in an increased sensitivity, justifying investigating in more details which reductases are the most important ones. Additionally, using knockout cell lines we were able to show in vitro that HAPng activity depends on Homologous Recombination (HR) and Fanconi Anemia (FA) status, but not on Non-homologous End Joining (NHEJ). HAPng also is capable of sensitizing tumors to irradiation and immunotherapy. We have demonstrated that HAPng combined with immunocytokines, but not with checkpoint inhibitors, results in a synergistic anti-tumor effect. Based on these data, we will also investigate a different approach, namely bringing the immunomodulators in the radiotherapy and immunotherapy resistant hypoxic tumor areas, by using new “HAP-immunomodulators”. The synthesis of these compounds is currently ongoing.

It is well known that the efficacy of conventional X-ray based radiotherapy depends on the oxygen status of the tumor. Increasing Linear Energy Transfer (LET) of irradiation causes more complex DNA damage independent of hypoxia. We therefore tested in collaboration with Heidelberg the combination of Carbon ion irradiation with different LET (high vs low, for which low mimics the LET of X-rays) and immunocytokines. Although high LET irradiation had a therapeutic, benefit comparable to X-rays, the combination with immunocytokines did not result in increased benefit. Currently, we are investigating the molecular mechanisms in order to explain these results. Clonogenic survival analyses using different cell lines have been performed in parallel to enable calculation of the relative biological effectiveness (RBE) of the different LET irradiations, allowing in the future irradiation with RBE adjusted dose.

Thanks to numerous advances in the field of radiomics image analysis, such as the quality control on FDG-PET prior to image acquisition, on CT prior to the extraction of radiomics features and the so called radiomics quality score (RQS) (Lambin et al., Nat Rev Clin Oncol, 2017), we are making great progress in our ability to implement radiomics imaging biomarkers to improve patient selection. In parallel in radiomics there are multiple steps still requiring further rigorous attention, such as volume-dependency of several groups of radiomics features, lack of standardization of image acquisition despite proposed guidelines, lack of standardization in feature calculation and statistical analysis methods. With regard to the statistical analysis our group has contributed with a valuable published feature selection (selection of most informative radiomics features) methods for longitudinal (cone beam) CT radiomics and management of breathing artefacts.

We have managed to generate validated hypoxia signatures on both CT and FDG-PET, with a radiomics approach. We found that the use of another promising method to target tumor hypoxia, nitroglycerin (a repurposed vasodilating drug) patching, did not result in a decrease in tumoral FDG uptake in advanced non-small cell lung cancer patients. We gathered and curated images data to validate currently published immunotherapy response CT signatures. Worthwhile to mention are key technological achievements such
A. Pre-clinical work (WP1-2)
We expect to achieve following main results from now until the end of the reporting period:
- Further characterization of reductases and their contribution to the cytotoxicity of HAPng.
- Further characterization of the DNA damage response to HAPng in a wide range of cell lines deficient in DNA repair pathways. This will be tested in 2D and 3D in vitro cell models, as well as in in vivo models.
- Identification of potential biomarkers of response to HAPng and immunocytokine.
- Testing in vitro and in vivo the combination of HAPng with inhibitors of DNA repair.
- Testing in vivo the combinational treatment of HAPng with radiotherapy and standard-of-care treatments.
- Provide evidence that removing the immunoresistant hypoxic areas will sensitize tumors to immunotherapy.
- Reveal the molecular mechanism supporting the immunotherapy outcome.
- Test the therapeutic efficacy of C-12 high LET based on in vitro determined RBE adjusted dose levels.
- Test the therapeutic efficacy of high LET irradiation (RBE-adjusted dose) in combination with immunotherapy.
- Reveal the immunological underlying mechanisms supporting high LET irradiation efficacy.
- Test in biological systems the HAP-immunomodulators.
- Test intratumoral clostridium as delivery system of immunotherapeutics (watch the animation: a project for which we received a ERC PoC

B Radiomics as biomarker (WP2)
- A fully automated lung tumor segmentation algorithm that can delineate a wide variety of lung tumors with a higher reproducibility than expert radiologists can.
- A radiomics signature that can detect lung cancer histologic subtype (adenocarcinoma and squamous cell carcinoma).
- A radiomics signature that can accurately predict tumor oxygenation status on CT and FDG-PET on a wide array of solid tumors.
- A radiomics signature that can accurately evaluate immunotherapy response in NSCLC lung cancer patients and a radiomics signature that can predict patients likely to develop pneumonitis after immunotherapy.

C. Clinical trial (WP4)

Regarding ImmunoSABR (H202-733008; we expect:
- The combination radiotherapy and immunotherapy alone (Van Limbergen et al, British J. Radiology, 2017; Van Limbergen et al, ImmunoSABR phase 1, in progress) vs triple therapy.
- Useful information on the triple therapy (radiotherapy + Immunocytokine + checkpoint inhibitor) that can be compared or used as control group for the next trial adding an HAP.
- All CT-scans will be used to validate the Hypoxia radiomics signature
- Predicting early death in NSCLC patients. (Jochems et al, Int J. Radiation Oncol, 2017; Jochems et al, Acta Onco, 2017)
- Predicting tumor hypoxia in NSCLC patients based on FDG and HX4 PET scan. (Even et al, Acta Onco, 2017)
- Retrospectively prediction tumor hypoxia based on CT and FDG-PET in NSCLC lung cancer patient (Sanduleanu et al, submitted Green Journal, 2019).