Periodic Reporting for period 1 - CHyMERA (Monitoring cancer heterogeneity based on the dynamic assessment of the Warburg effect under metabolic perturbation)
Reporting period: 2019-05-02 to 2021-05-01
CHyMERA aimed to provide a noninvasive molecular magnetic resonance imaging methodology able to contrast and quantify tumor heterogeneity with regards to proliferation status and metastatic potential. This entails dramatic implications for clinical management, particularly in the context of brain tumors (glioblastoma) and breast cancer, respectively. Thus, the initial aims of the project were to develop the CHyMERA methodology, based on dynamic monitoring of glucose and lactate kinetics in tumors, and use it to assess tumor proliferation in mouse models of glioblastoma. Additionally, we proposed a pilot study to investigate CHyMERA’s ability for assessing metastatic potential in breast cancer models.
The initial approach for CHyMERA was assessing the dynamics of nonradioactive glucose utilization in the tumor in response to metabolic perturbation (acute hyperglycemia and hypoxia), by monitoring the metabolic kinetics of glucose and lactate non-invasively using magnetic resonance imaging and spectroscopy.
For this we initially tested several pulse sequences, which performed generally well in vitro but not in vivo in the mouse brain for the purposed of CHyMERA (SECSI, PRESS-CSI, LASER-CSI, EPSI, and CEST). These problems were circumvented with deuterium magnetic resonance spectroscopy (DMRS), a novel magnetic resonance modality based on injection of non-toxic deuterated glucose. Therefore, we contacted NeosBiotec to design a custom-made radiofrequency coilset for the mouse brain (Fig 1).
Then, we successfully harnessed DMRS with volume localization and spectral denoising to perform dynamic glucose-enhanced (DGE) experiments in mouse brain tumors at 9.4 Tesla with unprecedented voxel resolution (15-35 mm3, Fig 2 A-C). This enabled the quantification of simultaneous glucose flux through glycolysis and mitochondrial oxidation in a well-established, immunocompetent mouse model of glioblastoma (GL261) in vivo (Fig 2 D-H). Importantly, we detected a strong correlation between glucose oxidation rate and cell proliferation in a relatively small, non-gender biased tumor population, presenting heterogeneous vascular-stromal phenotypes (R=0.829 p=0.021) which is consistent with additional results on a another immunocompetent GBM model with homogeneous phenotypes, CT-2A (cell line obtained from Dr. Thomas Seyfried) – Fig 2 I. The methodology is also being applied to a third GBM mouse model, in this case the conditional model Pten;P15/16b;EGFRvIII (obtained from Dr. Olaf van Tellingen), therefore addressing the main objective of CHyMERA – developing a non-invasive methodology to assess tumor proliferation in mouse models of glioblastoma.
We are also exploring ways to extend DGE-DMRS to map glycolytic and oxidative rates of glucose metabolism within the GBM microenvironment, using deuterium molecular imaging (DGE-DMI, Fig 3). This is particularly challenging for the latter, which were talking with new methods for spectral denoising. This will enable us to address another important aim of CHyMERA – to provide hotspot images of tumor proliferation.
Finally, we have successfully tested DGE-DMRS on an orthotopic, immunocompetent mouse model of metastatic breast cancer (4T1), also proposed in CHyMERA. Our preliminary results confirm the ability to measure the glycolytic turnover of glucose in these tumors (Fig 4), and thus the feasibility to pursue the final aim of the project: imaging the metastatic potential of breast cancer. However, since our current deuterium RF coilset is custom-made for the mouse brain, and therefore suboptimal for breast tumors, we will need a new coil design to continue with these studies. The same applies to the tissue perfusion studies initially proposed at 16.4 Tesla, for which we would also need an additional preamplifier for deuterium.
The results obtained during CHyMERA are under revision for publication or under preparation for additional publications, in an estimated total of 3 papers in high impact journals. This work has been consistently highly rated at international scientific meetings, where it has been regularly submitted and selected for oral presentations; namely, at ISMRM 2020, ISMRM 2021, and ESMI 2021. Once accepted for publication, the results will be disseminated to the public through the various media channels available at the Champalimaud Center for the Unknown, such as Ar magazine, and promoted for external diffusion to local and national news channels through the Communications Team, according to standard institutional procedures.
For this we initially tested several pulse sequences, which performed generally well in vitro but not in vivo in the mouse brain for the purposed of CHyMERA (SECSI, PRESS-CSI, LASER-CSI, EPSI, and CEST). These problems were circumvented with deuterium magnetic resonance spectroscopy (DMRS), a novel magnetic resonance modality based on injection of non-toxic deuterated glucose. Therefore, we contacted NeosBiotec to design a custom-made radiofrequency coilset for the mouse brain (Fig 1).
Then, we successfully harnessed DMRS with volume localization and spectral denoising to perform dynamic glucose-enhanced (DGE) experiments in mouse brain tumors at 9.4 Tesla with unprecedented voxel resolution (15-35 mm3, Fig 2 A-C). This enabled the quantification of simultaneous glucose flux through glycolysis and mitochondrial oxidation in a well-established, immunocompetent mouse model of glioblastoma (GL261) in vivo (Fig 2 D-H). Importantly, we detected a strong correlation between glucose oxidation rate and cell proliferation in a relatively small, non-gender biased tumor population, presenting heterogeneous vascular-stromal phenotypes (R=0.829 p=0.021) which is consistent with additional results on a another immunocompetent GBM model with homogeneous phenotypes, CT-2A (cell line obtained from Dr. Thomas Seyfried) – Fig 2 I. The methodology is also being applied to a third GBM mouse model, in this case the conditional model Pten;P15/16b;EGFRvIII (obtained from Dr. Olaf van Tellingen), therefore addressing the main objective of CHyMERA – developing a non-invasive methodology to assess tumor proliferation in mouse models of glioblastoma.
We are also exploring ways to extend DGE-DMRS to map glycolytic and oxidative rates of glucose metabolism within the GBM microenvironment, using deuterium molecular imaging (DGE-DMI, Fig 3). This is particularly challenging for the latter, which were talking with new methods for spectral denoising. This will enable us to address another important aim of CHyMERA – to provide hotspot images of tumor proliferation.
Finally, we have successfully tested DGE-DMRS on an orthotopic, immunocompetent mouse model of metastatic breast cancer (4T1), also proposed in CHyMERA. Our preliminary results confirm the ability to measure the glycolytic turnover of glucose in these tumors (Fig 4), and thus the feasibility to pursue the final aim of the project: imaging the metastatic potential of breast cancer. However, since our current deuterium RF coilset is custom-made for the mouse brain, and therefore suboptimal for breast tumors, we will need a new coil design to continue with these studies. The same applies to the tissue perfusion studies initially proposed at 16.4 Tesla, for which we would also need an additional preamplifier for deuterium.
The results obtained during CHyMERA are under revision for publication or under preparation for additional publications, in an estimated total of 3 papers in high impact journals. This work has been consistently highly rated at international scientific meetings, where it has been regularly submitted and selected for oral presentations; namely, at ISMRM 2020, ISMRM 2021, and ESMI 2021. Once accepted for publication, the results will be disseminated to the public through the various media channels available at the Champalimaud Center for the Unknown, such as Ar magazine, and promoted for external diffusion to local and national news channels through the Communications Team, according to standard institutional procedures.
DGE-DMRS is a novel magnetic resonance modality recently reported by others to measure glucose oxidation rates in normal rat brain, and glycolytic rates in mouse lymphomas. We have tuned this methodology beyond its current state-of-the-art, harnessing it with spatial localization (outer-volume suppression) and an unbiased noise-reduction technique (Machenko-Pastur PCA denoising), which was tested in vitro and in vivo. This enabled us to quantify, for the first time in mouse GBM tumors in vivo, the metabolic heterogeneity of glucose flux through glycolysis and mitochondrial oxidation in vivo and its association with the respective histopathologic phenotype and cell proliferation index. These results are consistent across two immunocompetent GBM models (and currently being tested with a genetically modified conditional model), suggesting glucose oxidation rate as a potential non-invasive metabolic marker of cell proliferation in GBM.
Therefore, DGE-DMRS demonstrates strong potential for in vivo metabolic phenotyping of GBM, which represents a current clinical need for patient stratification and early monitoring of therapeutic efficacy of novel drugs targeting mitochondrial metabolism, in GBM and other cancers. Importantly, extension of this approach to its imaging modality – DGE-DMI – is currently underway and should further help defining tumor borders, to improve the efficacy of surgical resection, when feasible, and guiding biopsy collection and/or radiation therapy to more aggressive regions, to improve diagnostic accuracy and treatment efficacy, respectively. Altogether, the future translation of DGE-DMRS/DMI to GBM patients is expected to improve their clinical outcome.
Moreover, our preliminary results with 4T1 tumor-bearing mice demonstrate the successful application of DGE-DMRS to breast cancer, and the feasibility of exploring this methodology for assessing metastatic potential in this tumor type. This would also have a strong social impact, as it would enable clinicians to tailor treatment at early stages, improving patient outcome and reducing their exposure to unnecessary toxicity.
Therefore, DGE-DMRS demonstrates strong potential for in vivo metabolic phenotyping of GBM, which represents a current clinical need for patient stratification and early monitoring of therapeutic efficacy of novel drugs targeting mitochondrial metabolism, in GBM and other cancers. Importantly, extension of this approach to its imaging modality – DGE-DMI – is currently underway and should further help defining tumor borders, to improve the efficacy of surgical resection, when feasible, and guiding biopsy collection and/or radiation therapy to more aggressive regions, to improve diagnostic accuracy and treatment efficacy, respectively. Altogether, the future translation of DGE-DMRS/DMI to GBM patients is expected to improve their clinical outcome.
Moreover, our preliminary results with 4T1 tumor-bearing mice demonstrate the successful application of DGE-DMRS to breast cancer, and the feasibility of exploring this methodology for assessing metastatic potential in this tumor type. This would also have a strong social impact, as it would enable clinicians to tailor treatment at early stages, improving patient outcome and reducing their exposure to unnecessary toxicity.