15th European Molecular Imaging Meeting
supported by:

To search for a specific ID please enter the hash sign followed by the ID number (e.g. #123).
*.ics

Imaging the Tumour Microenvironment

Session chair: Tim Witney (London, UK); Inigo Martinez (Tromso, Norway)
 
Shortcut: PS 01
Date: Wednesday, 26 August, 2020, 10:00 a.m. - 11:30 a.m.
Session type: Parallel Session

Contents

Abstract/Video opens by clicking at the talk title.

10:00 a.m. PS 01-01

Introductory Lecture

Henry C. Manning1

1 Vanderbilt University, Nashville, United States of America

 
10:18 a.m. PS 01-02

MMP-independent cancer cell invasion guided by tissue microchannels

Bettina Weigelin2, 1, Lianne Beunk2, Esther Wagena2, Marco Scianna3, Marit de Beer2, Jack Fransen2, Mark Ellisman4, Katarina Wolf2, Luigi Preziosi3, Peter Friedl2, 5

1 University of Tübingen, Werner Siemens Imaging Center, Tübingen, Germany
2 Radboud University, Cell Biology, Nijmegen, Netherlands
3 Politecnico di Torino, Mathematical Sciences, Torino, Italy
4 University of California, Center for Research in Biological Structure, San Diego, United States of America
5 University of Texas, MD Anderson Cancer Center, Houston, United States of America

Introduction

Cancer cell invasion is an adaptive process based on cell-intrinsic properties to migrate individually or collectively, and further thought to depend on tissue remodeling provided by matrix metalloproteinases (MMPs). Whereas molecular and physical mechanisms of cancer invasion in vitro are well-established, the mechanisms governing interstitial tissue invasion in vivo remain incompletely understood.

Methods

Using intravital multiphoton and higher harmonic generation microscopy, we mapped tissue spaces and geometry in live melanoma and fibrosarcoma tumors in the mouse dermis and in relation to tumor cell invasion routes and efficiency. Structural tissue components and interfaces were detected by in vivo immunohistochemistry and correlated serial block-face scanning 3D electron microscopy (SBEM). Tissue spaces and the interstitial fluid compartment were visualized by injection of fluorescent contrast agents and solidifying resin to generate 3D casts of tissue microchannels. Mathematical modeling and molecular intervention were used to test whether tumor invasion was protease-dependent or primarily determined by preformed interstitial space and tissue biomechanics.

Results/Discussion

Tumor cells developed adaptive invasion patterns into pre-existing tissue tracks with linear confined or complex topography. 3D ultrastructural mapping by SBEM did not show migration-associated anatomic tissue remodeling. Invasion-promoting niches (i) were lined by basement membrane-covered myofibers, vessels or adipocytes, (ii) contained hyaluronic acid and interstitial fluid, and (iii) enabled outward drainage of tumor-secreted chemokine SDF1α-mCherry. This interstitial fluid network promoted barrier-free cell movement with passive components, such as outward pushing of cells by invasive growth or externally applied pressure. MMP-independent invasion was confirmed by in vivo interference, using Crisp/Cas9-mediated MMP14 k.o. or broad-spectrum MMP inhibitor Batimastat. Mathematical simulations predicted non-destructive widening of tissue spaces by moving cells, which was confirmed by time-lapse intravital microscopy and 3D SBEM showing cells unfolding narrow tissue clefts.

Conclusions

In aggregate, the data suggests non-destructive individual and collective tumor cell migration along pre-existing tissue conduits which drain fluids, soluble factors, and support invasion by a contact-guidance and ‘chimneying’ mechanism.

Keywords: intravital multiphoton microscopy, tumor invasion, tumor microenvironment
10:30 a.m. PS 01-03

EGFR-mediated photoimmunotherapy to potentiate anti-GBM immune response

Justyna Mączyńska1, Florian Raes1, Chiara Da Pieve1, Stephen Turnock1, Julia Hoebart1, Jessica K.R. Boult1, Marcin Niedbala2, Simon P. Robinson1, Kevin J. Harrington1, Wojciech Kaspera2, Gabriela Kramer-Marek1

1 The Institute of Cancer Research, 1Division of Radiotherapy and Imaging, London, United Kingdom
2 Medical University of Silesia, 2Department of Neurosurgery, Sosnowiec, Poland

Introduction

Despite rigorous therapeutic regimens, glioblastoma (GBM) patients eventually relapse. This is most likely due to a relatively immune-depleted (“cold”) GBM microenvironment characterised by high levels of immunosuppressive cytokines which inhibit T cells activity [1]. In a clinical context a desirable outcome would be to restore an immunological environment to re-sensitise GBM cells to immuno-oncology agents. We postulate that photoimmunotherapy (PIT) targeting epidermal growth factor receptor (EGFR) can promote T cell activation, whilst overcoming the immunologically cold status of GBM.

Methods

The phthalocyanine dye, IR700, was conjugated to affibody molecule ZEGFR:03115. Cell viability, reactive oxygen species (ROS) production, and major damage associated molecular patterns (calreticulin (CRT), heat shock protein 70 (Hsp70), high mobility group box 1 (HMGB1), and ATP) were studied in human and murine glioma cell lines post-ZEGFR:03115–IR700 PIT using flow cytometry (FC), Western blot and ELISA. Dendritic cells (DCs) maturation (CD86 and HLA-DR upregulation) was verified by FC. Xenograft and syngeneic murine tumour models (subcutaneous and orthotopic) were used to determine therapeutic and tumour-specific immune response following PIT. MRI, PET (18F-AlF-NOTA-ZEGFR:03115) and optical images were acquired to characterise the tumours before and post-treatment.

Results/Discussion

ZEGFR:03115–IR700 PIT showed a significant decrease in cell viability in EGFR positive cells. Generation of ROS post-PIT resulted in translocation of CRT to the cell membrane and the release of HMGB1, Hsp70 and ATP. FC analysis showed significant increase in the expression of CD86 and HLA-DR molecules on DCs co-cultured with ZEGFR:03115–IR700 PIT treated cells compared to the controls. PET, MRI and enhanced fluorescence signals confirmed the presence of well-defined tumour masses, and the ability of ZEGFR:03115 to target and visualise EGFR-expressing cells in vivo. PIT led to a significant delay in subcutaneous tumour growth. Therapeutic efficacy of the conjugate was observed in brain tumours as early as 24 h post-irradiation. Ki-67 and H&E staining of treated-tumour sections showed a reduced cell proliferation index, extensive tumour necrosis and microhaemorrhage. In addition, tumour-infiltrating lymphocytes were elevated in the PIT-treated mice as early as 4 h post-treatment.

Conclusions

ZEGFR:03115-IR700-based PIT activate and mature DCs in vitro and promotes T cell attraction in vivo. Therefore, it is an attractive therapeutic strategy for GBM as following complete or cytoreductive resection in patients could i) lead to the elimination of residual or surgically inaccessible EGFR positive cancer cells and ii) trigger tumour-infiltrating cytotoxic T-cell immune response.

Acknowledgment

The authors gratefully thank AffibodyAB (Stockholm, Sweden) for supplying the affibody molecule (ZEGFR:03115). We owe special thanks to Ian Titley and Fredrik Wallberg for flow cytometry technical support. This work was funded by the CRUK Convergence Science Centre - Development Fund and  National Science Centre, Poland (Project No. 2015/19/N/NZ7/01336).

References
[1] Zhao J et al., ‘Immune and genomic correlates of response to anti-PD-1 immunotherapy in glioblastoma’, Nat Med, 25, 462-469, 2019
Photoimmunotherapy (PIT) targeting epidermal growth factor receptor (EGFR).

ZEGFR:03115–IR700-mediated PIT promotes T cell activation, whilst overcoming the immunologically cold status of glioblastoma cells.

 

Keywords: photoimmunotherapy, EGFR, glioblastoma, affibody
10:42 a.m. PS 01-04

Immunomodulation of GAMs by inhibition of the CSF-1 receptor in a syngeneic glioma mouse model: a multimodal molecular imaging study

Claudia Foray1, 2, 3, Cristina Barca1, 2, 3, Christian Döring1, Michael Schäfers1, 4, 5, Brian West6, Bastian Zinnhardt1, 2, 4, Andreas H. Jacobs1, 2, 7

1 European Institute for Molecular Imaging - EIMI, Münster, Germany
2 PET Imaging in Drug and Development - PET3D, Münster, Germany
3 Westfälische Wilhelms-Universität Münster, Münster, Germany
4 University Hospital Münster, Nuclear Medicine, Münster, Germany
5 Cells in Motion (CiM) Cluster of Excellence, Münster, Germany
6 Plexxikon Inc., Berkeley, United States of America
7 Johanniter Hospital, Evangelische Kliniken, Geriatrics, Bonn, Germany

Introduction

Gliomas are characterized by a highly heterogeneous tumour microenvironment (TME) causing therapy resistance and escape. Within the TME, glioma-associated microglia/macrophages (GAMMs) are important in creating an immunosuppressive TME.  The recruitment and activity of GAMMs can be efficiently targeted by inhibition of the colony stimulating factor 1 receptor (CSF-1R) signaling cascade. We employed non-invasive PET/MRI using a combination of [18F]FET (aa-metabolism) and [18F]DPA-714 (TSPO) to monitor CSF-1R inhibitor-induced changes in the glioma TME and subsequent modulation of GAMM activity.

Methods

N=18 female C57BL/6 mice underwent MRI and PET/CT scans after intra-striatal implantation of 2x10^5 syngeneic GL261 glioma cells. 7 days p.i. mice were treated with the CSF-1R inhibitor (PLX5622; Plexxikon, Berkeley CA, 1200 ppm chow) or control diet for 1 week. Mice were subjected to PET imaging at d0 (before) and at d7 after treatment with [18F]DPA-714 (TSPO; 15 MBq; 60-80 min p.i.) and [18F]FET (amino acid transport; 10 MBq, 20-30 min p.i.). Tumour-to-background uptake ratios (T/B) were calculated using an atlas-based volumetric approach. ROIs defining the area of interest were placed using the brain atlas as reference and a threshold based on 3.5xSD of the contralateral striatum area was applied. After the last scan, brains were harvested for histology and other ex-vivo investigations.

Results/Discussion

[18F]FET-derived tumour volume was significantly increased in both, controls (0.027±0.014; p≤0.01) and PLX5622 treated mice (0.025±0.015; p≤0.01) (Fig 1. A, B). Analysis of the T1w MR images based on gadolinium enhancement also showed a significant volume increase, both in mice fed with control diet (0.031±0.02; p≤0.01) and with PLX5622 (0.025±0.01; p≤0.01) (Fig. 1A, C). [18F]DPA-714 T/B ratios were significantly reduced in PLX5622 treated mice at day 7 (0.466±0.11; p≤0.001) (Fig 1. A, D). Histological analysis confirmed the almost complete absence of Iba1 positive cells after PLX5622 treatment (Fig 2. A). Immunofluorescence of CSF-1R/Iba1 and TSPO/Iba1 confirmed ex vivo the decrease of CSF-1R and TSPO after treatment with PLX5622 (Fig 2. A). Increased infiltration of immune cells into the TME after CSF-1R inhibition could be confirmed by CD163 (macrophages) and Ly6G (neutrophils) immunofluorescence (Fig 2. B).

Conclusions

[18F]FET- and [18F]DPA-714-PET with MRI allow the non-invasive assessment of glioma growth and response to CSF-1R inhibition. PLX5622 therapy significantly reduced TSPO expression without an effect on tumour growth. Interestingly, CSF-1R inhibition leads to an increase in immune cell infiltration in the tumour microenvironment. Further analyses are ongoing to characterize the changes in the TME and the importance of microglia cells in gliomas.

AcknowledgmentThis work was supported by the Horizon2020 Programme under grant agreement n° 675417 (PET3D).
Multi-tracer images to investigate CSF-1R inhibition.
Fig.1 A) T1w Gd MRI and PET/CT images of [18F]FET and [18F]DPA-714 in control and treated mice at d0 and after 7 days of treatment. The dotted line highlight the tumour mass. B) Quantitative analysis of the [18F]FET-derived tumour volume and (C) of the MR-derived tumour volume (cm3/ml) after 1 week of PLX5622 treatment. D) Quantitative analysis of [18F]FET and [18F]DPA-714 T/B uptake ratio pre- and post-CSF-1R inhibition. Differences intra- and inter-groups were tested for significance using one-way ANOVA, paired t-test, Wilcoxon test and Mann-Whitney test with Bonferroni correction.
Modulation of the tumour microenvironment

Fig.2. A) Immunofluorescence images showing co-localization of CSF-1R (green) and Iba1 (red) or TSPO (green) and Iba1 (red), in tumour-bearing mice not-treated (GL261) and treated with PLX5622 (GL261+PLX5622). B) Immunofluorescence staining for CD163 (M2 macrophages marker) and Ly6g (neutrophils) in tumour-bearing mice not-treated (GL261) and treated with PLX5622 (GL261+PLX5622). The white box represent the detailed images at a higher magnification (40x).

Keywords: glioma, DPA-714, CSF-1R
10:54 a.m. PS 01-05

Characterising the tumour vascular microenvironment of breast cancer patient-derived xenografts using optoacoustic imaging

Emma L. Brown1, 2, Isabel Quiros- Gonzalez1, 2, 3, Ziqiang Huang2, Dominique Laurent-Courtier2, Joanna Brunker1, 2, 4, Lina Hacker1, 2, Alejandra Bruna2, Elena Provenzano5, Carlos Caldas2, 6, 7, Sarah E. Bohndiek1, 2

1 University of Cambridge, Department of Physics, Cambridge, United Kingdom
2 University of Cambridge, Cancer Research UK Cambridge Institute, Cambridge, United Kingdom
3 University of Oviedo, Cell Biology Department-FINBA-IUOPA, Oviedo, Spain
4 University College London, Wellcome/EPSRC Centre for Interventional and Surgical Sciences, London, United Kingdom
5 Addenbrookes Hospital, Department of Histopathology, Cambridge, United Kingdom
6 University of Cambridge, Department of Oncology, Cambridge, United Kingdom
7 Cambridge University Hospitals NHS Foundation Trust, Cambridge Breast Unit, NIHR Cambridge Biomedical Research Centre and Cambridge Experimental Cancer Medicine Centre, Cambridge, United Kingdom

Introduction

Vascular phenotypes in breast cancer may differ drastically among different subtypes and influence tumour biology and therapy response1. Optoacoustic imaging (OAI) is an emerging modality that allows non-invasive visualisation of tumour vascular features through the endogenous contrast of deoxy- and oxy-haemoglobin (Hb and HbO2)2, filling a gap in the clinic. We have selected two different subtypes of breast cancer patient-derived xenografts (PDXs) and hypothesised that OAI could provide a non-invasive tool to longitudinally monitor the vascular microenvironment in those subtypes.

Methods

One luminal B PDX and one basal PDX were implanted in vivo (nLumB= 13, nBasal= 19). Raster-Scanning Optoacoustic Mesoscopy (RSOM) (Fig.1a) and Multispectral Optoacoustic Tomography (MSOT) (Fig.1b) were conducted weekly to yield data on tumour blood volume (BVRSOM), total haemoglobin concentration (THbMSOT=Hb+HbO2) and blood oxygen saturation (SO2MSOT= HbO2/THb) respectively. Additionally, mice were subjected to a change of breathing gas between air and 100% O2 during their MSOT scan; kinetic SO2 profiles were smoothed using a moving average and ΔSO2MSOT calculated as the baseline SO2MSOT under O2 minus baseline SO2MSOT under air to reveal hypoxic status3. Once excised, tumours underwent immunohistochemistry (IHC) to stain CD31 (endothelial cells) and CAIX (hypoxia).

Results/Discussion

BVRSOM was modelled linearly as tumour volume increased, and was found to increase at a significantly higher rate in the basal PDX compared to the luminal B PDX (p=0.03, Fig.2a). Supporting this, THbMSOT , measured from MSOT images, was significantly higher in the basal PDX at the final time-point ( 2.32±0.24a.u. vs. 0.82± 0.10a.u., p<0.001, Fig.2b). This was corroborated with a higher microvessel density measured from CD31 stained sections in a subset of tumours analysed so far (1.64*10-4±2.76*10-5 vessels/μm2 vs. 6.59*10-5±2.29*10-5 vessels/μm2, p=0.05, Fig.2c). Despite a bloodier phenotype in the basal PDX, the luminal B PDX had higher SO2MSOT under O2 (0.55±0.026 vs. 0.46±0.016, p=0.006, Fig.2d), and responded better to the gas challenge with higher ΔSO2 (0.10±0.024 vs. 0.046±0.013, Fig.2e). The luminal B PDX was less hypoxic upon ex vivo IHC (CAIX staining, 37.76±2.53% vs. 56.50±1.80%, p<0.001, Fig.2f) supporting the in vivo SO2 measurements.

Conclusions

Optoacoustic imaging is sensitive to different vascular phenotypes between basal and luminal B PDXs, underscored by the vascular phenotypes measured using IHC. Our preliminary results suggest that breast PDXs with a lower blood volume have better vascular function and oxygenation. Future work will aim to further understand the subtype-dependency of vascular phenotypes.

Acknowledgment

This work was supported by Cancer Research UK (C14303/A17197).

References
[1] Junttila, MR, de Sauvage, FJ, 2013, 'Influence of tumour micro-environment heterogeneity on therapeutic response', Nature, 501, 346
[2] Wang, LV, Hu, S 2012, 'Photoacoustic Tomography: In Vivo Imaging from Organelles to Organs', Science, 335, 1458–1462
[3] Tomaszewski, MR, Gonzalez, IQ, O’Connor, JP, Abeyakoon, O, Parker, GJ, Williams, KJ, Gilbert, FJ and Bohndiek, SE 2017, 'Oxygen Enhanced Optoacoustic Tomography (OE-OT) Reveals Vascular Dynamics in Murine Models of Prostate Cancer', Theranostics, 7, 2900–2913
Figure 1. Optoacoustic imaging of breast cancer PDXs in vivo

(A) Representative RSOM image shown as a maximum intensity projection in the xy plane. Ultrasound frequencies are spilt between low (11-33MHz) and high (33-99MHz) frequencies to highlight large and small vessels in red and green respectively. (B) Representative MSOT image (tumour outline in red) showing a 2D tomographic slice.

Figure 2. Comparison of vascular phenotypes measured between basal and luminal B breast PDXs

(A) BVRSOM modelled linearly against tumour volume. (B) THbMSOT, (C) MVD measured using CD31 IHC, (D) SO2MSOT , (E) ΔSO2 measured during a gas challenge in the MSOT, (F) CAIX IHC. For (A), nLumB= 8, nBasal= 13. For (B), (D) and (E) nLumB= 7, nBasal= 10. For (C) nLumB= 3, nBasal= 5. For (F) nLumB= 11, nBasal= 18

Keywords: Optoacoustic Imaging, Breast Cancer, Patient-derived xenografts, Vasculature
11:06 a.m. PS 01-06

Imaging CD80 by PET to assess the immune status of the tumour and its microenvironment

Claudia A. Castro Jaramillo1, Marco F. Taddio1, Peter Runge2, Ioannis Kritikos2, Sandra Kreis1, Federica Petruzzelli1, Nicholas van der Meulen3, Alain Blanc4, Martin Béhé4, Cornelia Halin2, Roger Schibli1, 3, Stefanie D. Krämer1

1 Center for Radiopharmaceutical Sciences ETH, PSI and USZ, Department of Chemistry and Applied Biosciences of ETH Zurich, Zürich, Switzerland
2 Pharmaceutical Immunology, Department of Chemistry and Applied Biosciences, Zürich, Switzerland
3 Laboratory of Radiochemistry Paul Scherrer Institute, Paul Scherrer Institute, Zürich, Switzerland
4 Center for Radiopharmaceutical Sciences ETH, PSI and USZ, Paul Scherrer Institute, Zürich, Switzerland

Introduction

The tumor microenvironment plays a pivotal role in cancer development and progression. The polarization of macrophages in the microenvironment (TAM) into tumor-suppressive (M1) or tumor-promoting (M2) types is of particular interest. CD80 is a co-stimulatory molecule present on the surface of M1 TAMs. As well CD80 can be expressed by cancer cells directly, inducible by anti-cancer treatments. High CD80 cell surface density is required for T-cell activation. Radiotracers able to imagine CD80+ cells can be used as powerful indicators of tumor prognosis and response to immunotherapy

Methods

We inoculated CT26 cells subcutaneously in Balb/c mice to grow syngenic tumors in immunocompetent mice. We characterized the tumors by flow cytometry and RT-qPCR for relevant markers, and their correlation with CD80. Heterogeneity of the tumours was evaluated based on degree of vascularization and differential presence of M1 and M2 TAM biomarkers by confocal microscopy. The affinity of abatacept to murine CD80 was determined by surface plasmon resonance and the effect of abatacept on murine T-cell proliferation was evaluated in an in vitro co-culture system. PET imaging was performed with [64Cu]NODAGA-abatacept, specific binding was determined with an excess of unlabelled abatacept. The effect of the anti-cancer drug decitabine on CD80 expression of CT26 cells was evaluated in vitro.

Results/Discussion

The distribution of M1 (CD80+) and M2 (CD206+) TAMs between and within the individual tumors was highly heterogenic, highlighting the power of molecular imaging over traditional biopsy, which provides spatially limited information. The affinity of abatacept to murine CD80 was in the one-digit nanomolar range, ideal for PET imaging. Abatacept at a concentration corresponding to the maximal expected plasma concentration during PET imaging did not affect T cell proliferation in vitro, indicating that abatacept PET will not negatively affect the immune system. In PET experiments, [64Cu]NODAGA-abatacept accumulated specifically in CD80 positive tissues, including the CT26 syngrafts (Figure 1). Treatment with decitabine increased the expression of CD80 in CT26 cells in vitro, indicating that the CT26 syngraft model may serve as a proof-of-concept to study treatment effects on CD80 levels by PET.

Conclusions

We have successfully taken steps towards the application of CD80 PET radiotracers for non-invasive assessment of the immune state of tumors based on the expression of CD80. We were able to specifically image CD80 in a CD80-positive syngraft model. The heterogenic distribution of CD80 positive cells in the tumour suggests that CD80 PET may provide more information on the immune state of a cancer than biopsies.

AcknowledgmentThe project is funded by the Molecular Imaging Network Zurich (KFSP – MINZ).
References
[1] Sounni N.E. and Noel A. Targeting the Tumor Microenvironment for Cancer Therapy. Clin Chem. 2013;59:85-93.
[2] Taddio M.F., Mu L., Castro Jaramillo C.A., Bollmann T., Schmid D.M., Muskalla L.P., Gruene T., Chiotellis A., Ametamey S.M., Schibli R. and Krämer S.D. Synthesis and structure-affinity relationship of small molecules for imaging human CD80 by positron emission tomography. J Med Chem. 2019;62:8090-8100.
[3] Meletta R, Muller Herde A, Dennler P, Fischer E, Schibli R, Kramer SD. Preclinical imaging of the co-stimulatory molecules CD80 and CD86 with indium-111-labeled belatacept in atherosclerosis. EJNMMI Res 2016; 6(1):1.
Figure 1
PET coronal images of mice with a CT26 syngraft (indicated by the arrows) 48 hours after 64Cu-NODAGA-abatacept injection. B) As A but 64Cu-NODAGA-abatacept was co-injected with 0.5 mg/kg unlabeled abatacept to show specificity (blocking). One representative out of 3 mice each. SUV, standardized uptake value.
Figure 2
Ex vivo biodistribution of 64Cu-NODAGA-abatacept 48 hours after injection. For blocking conditions 0.5 mg/kg unlabeled abatacept was coinjected to show specificity. SUV, standardized uptake value.
Keywords: PET, CD80, Precision medicine, tumor microenvironment