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

MRI & Multimodal Technologies

Session chair: Jessica Bastiaansen (Lausanne, Switzerland); Uwe Himmelreich (Leuven, Belgium)
 
Shortcut: PW13
Date: Wednesday, 25 August, 2021, 7:15 p.m. - 9:00 p.m.
Session type: Poster

Contents

Click at talk title to open the abstract

1016

Time-lapse MRI single cell tracking of iron labeled monocytes with velocities of 1 µm/s

Enrica Wilken1, Felix Freppon1, Max Masthoff1, Cornelius Faber1

1 University Hospital Muenster, Translational Research Imaging Center, Clinic of Radiology, Muenster, Germany

Introduction

With increasing knowledge on cell-based diagnostics and therapies, the understanding of cell motion becomes equally important.1
In contrast to established means of cell tracking, like intravital microscopy 2, time-lapse MRI provides a non-invasive tool for the three-dimensional tracking of single cells with unlimited tissue penetration and whole-brain coverage. 3,4 However, the time-lapse contrast generated by iron-labeled cells is motion-dependent. Therefore, we address the detectable velocity range of single cells by simulations based on artificial k-space and tracking immune cells in vivo.

Methods

In vivo time-lapse MRI of 8 mice brain was performed on a 9.4 T Biospec (Bruker Biospin). 24h before scanning, monocytes were labeled by injection of Ferucarbotran (Resovist Bayer AG) via the tail vein. 20 repetitive T2* weighted images were acquired with a gradient echo sequence (in-plane resolution: 67 x 55 µm2; scan time: 8 mins 12 s). Cells were counted manually and tracked with ImageJ.3 Blurring of moving cells was simulated using Matlab by stepwise increasing a signal voids position in an artificial phantom and filling synthetic k-space with fractions of the Fourier transforms of the individual images for cartesian and radial sampling. The final image was acquired by the inverse Fourier transform of the assembled k-space (Fig. 2a) and multiplied with real MRI to create an overlay.

Results/Discussion

In vivo time-lapse MRI enabled the detection and tracking of intravascular moving monocytes represented by hypointense spots in repetitive T2*-weighted images (Fig. 1a). In total 61 cells were detected moving across several voxels in consecutive timeframes and their velocity was determined to be 0.19 ± 0.01 µm/s (Fig. 1b), matching the speed of patrolling monocytes.5 Furthermore, through assembled artificial k-space, time-lapse contrast of moving cells with growing velocities was simulated showing the increasing blurring (Fig. 2b,c). The superpositions with real MRI data account for noise, anatomic structures and imaging artifacts, and show that the simulated cells reproduce experimentally observed signal voids well. Further, they demonstrate that cells with a speed of up to 1 µm/s create enough contrast to be distinguishable from the background with both, cartesian and radial, sampling schemes. Consequently, rolling cells cannot be seen as they are related to a much higher speed.5

Conclusions

We conclude that time-lapse MRI allows the visualization of iron-labeled monocytes in mice brain, and, therefore, can be used to non-invasively analyze immune responses. The generated contrast is sufficient for cells moving with up to 1 µm/s to be detectable, enabling the detection of patrolling monocytes. To reveal rolling immune cells as well, we aim at accelerating image acquisition by compressed sensing and radial sampling.

Acknowledgement

This study was supported by the German Research Foundation (DFG; SFB 1009 TP-Z02 to CF), the Joachim Herz Foundation (Add-on Fellowship for Interdisciplinary Life Sciences to MM), the Interdisciplinary Centre for Clinical Research (IZKF, core unit PIX) and the Medical Faculty of the University of Muenster (MEDK dissertation fellowship to FF).

Disclosure

I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

References
[1] Perrin, J., Capitao, M., Mougin-Degraef, M., Guérard, F., Faivre-Chauvet, A., Rbah-Vidal, L., Gaschet, J., Guilloux, Y., Kraeber-Bodéré, F., Chérel, M., & Barbet, J. (2020), ‘Cell Tracking in Cancer Immunotherapy.’, Frontiers in medicine, 7, 34.
[2] Karreman, M. A., Hyenne, V., Schwab, Y., & Goetz, J. G. (2016), ‘Intravital Correlative Microscopy: Imaging Life at the Nanoscale.’, Trends in cell biology, 26(11), 848–863.
[3] Masthoff, M., Gran, S., Zhang, X., Wachsmuth, L., Bietenbeck, M., Helfen, A., Heindel, W., Sorokin, L., Roth, J., Eisenblätter, M., Wildgruber, M., & Faber, C. (2018), ‘Temporal window for detection of inflammatory disease using dynamic cell tracking with time-lapse MRI.’, Scientific reports, 8(1), 9563.
[4] Mori, Y., Chen, T., Fujisawa, T., Kobashi, S., Ohno, K., Yoshida, S., Tago, Y., Komai, Y., Hata, Y., & Yoshioka, Y. (2014), ‘From cartoon to real time MRI: in vivo monitoring of phagocyte migration in mouse brain.’, Scientific reports, 4, 6997.
[5] Auffray, C., Fogg, D., Garfa, M., Elain, G., Join-Lambert, O., Kayal, S., Sarnacki, S., Cumano, A., Lauvau, G., & Geissmann, F. (2007), ‘Monitoring of blood vessels and tissues by a population of monocytes with patrolling behavior.’, Science (New York, N.Y.), 317(5838), 666–670.
Figure 1: In vivo detected cells in motion.
(a) Image details, extracted from a representative example of one slice of a mouse brain, showing a migrating cell as a moving hypointense spot (red arrowhead) in four consecutive timeframes.
(b) The mean velocity of monocytes moving across several voxels was determined to be 0.19 ± 0.01 µm/s. The horizontal bar represents the mean value, the closed circles the individual values.
Figure 2: Simulations of time-lapse contrast.
(a) Basic concept of how artificial k-space is created in order to simulate image contrast of moving cells of different speed.
(b,c) Image details of a superposition of a T2*-weighted image acquired with the time-lapse MRI protocol and the contrast simulation for various cell motion velocities, showing a real (red arrowhead) and a simulated (blue arrowhead) cell in the brain cortex. The simulations were performed using a (b) cartesian and (c) radial sampling scheme.
Keywords: MRI, cell tracking, preclinical, time-lapse MRI, monocytes
1017

In vivo preclinical measurements of water transport across the cell membrane by filter-exchange imaging

Athanasia Kaika1, Geoffrey J. Topping1, Irina Heid2, Luca Nagel1, Mathias Schillmaier1, Rickmer Braren2, Franz Schilling1

1 Technical University of Munich, School of Medicine, Klinikum rechts der Isar, Department of Nuclear Medicine, Munich, Germany
2 Technical University of Munich, School of Medicine, Klinikum rechts der Isar, Institute of Radiology, Munich, Germany

Introduction

Plasma membrane permeability plays a key role in cell vitality1. Filter-Exchange Imaging (FEXI) is performed using a double-diffusion magnetic resonance pulse sequence, separated by a mixing time, which encodes information about transmembrane water exchange, quantified by the apparent exchange rate (AXR)2-5. The aim of this work was to establish a preclinical in vivo imaging protocol to map cell membrane permeability by using FEXI. This protocol was tested in mice within brain and within abdominal and subcutaneous tumors.

Methods

Subjects

  1. Mouse brain
  2. Mouse bearing abdominal PDAC tumor6
  3. Mouse bearing subcutaneous EL4 lymphoma

MR measurement

MRI scanner: 7T small animal (Bruker/Agilent/GE)

FEXI acquisition

1: bf = 1496s/mm2; bi: 54,904,1304s/mm2; tm= 13.5-303.4ms; Acq. time:15 min

2, 3: bf = 1288s/mm2; bi: 54,604s/mm2; tm= 13.5-303.4ms; Acq. time:17 min

Where tm: mixing time between diffusion filter (bf) and diffusion encoding (bi) module application.

Analysis

Over repetitions, signal median value of each tm and filter was calculated. Outliers were excluded if  >1 median from the ADC curve fit to all data.

AXR maps were generated according to [2-4].

The AXR curves were calculated from ROI-mean ADCs on FEXI ADC maps for each mixing time.

ROIs were drawn on T2W images.

Results/Discussion

Figure 1 depicts T2W and FEXI images and parametric maps of mouse brain. In the low part of FEXI (1b, red arrow) and DWI images (not shown), an artifact was observed, possibly due to poor fat suppression. FEXI ADC maps (1c) show ADC reduction after the diffusion filter application and gradual increase with the mixing time. This increase can also be seen in the AXR fit (1e). The ADC of ventricles was lower after the filter application, in contrast to previous experiments in water (not shown), in which the filter didn’t affect ADC. The non-zero AXR values in the cerebrum are consistent with the presence of structures with fast and slow diffusivity.

Figure 2 shows T2W images and AXR maps of an endogenous mouse PDAC tumor (2ai, ii) and a mouse subcutaneous EL4 lymphoma tumor (2bi, ii).

  1. PDAC tumor: inhomogeneous AXR values were observed.
  2. EL4 lymphoma: was easily distinguished from its background because of its high AXR values. Lower AXR values were observed in the edges than in the center.

Conclusions

The FEXI sequence, acquisition protocol and an analysis tool were developed and tested in vivo on a mouse brain, a mouse PDAC tumor and a mouse EL4 lymphoma. The EL4 lymphoma showed higher AXRs than surrounding tissues. The future work includes follow-up studies of pathologies that are subject to changes in membrane permeability, as well as histological correlation.

Acknowledgement

We thank Martin Grashei, Sandra Sühnel, Anna-Maria Schmidmüller, Simon Baller and Irina Skuratovska for their technical assistance. The present work was supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation – 391523415, SFB 824).

Disclosure

I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

References
[1] Galluzzi L, Vitale I, et al., 2018, Molecular mechanisms of cell death: recommendations of the Nomenclature Committee on Cell Death, Cell Death Differ., 25(3):486–541
[2] Nilsson M, Lätt J, et al., 2013, Noninvasive mapping of water diffusional exchange in the human brain using filter-exchange imaging, Magn Reson Med., 69(6): 1573–1581
[3] Lasič S, Nilsson M et al., 2011, Apparent exchange rate mapping with diffusion MRI, Magn Reson Med., 66(2):356-65
[4] Aslund I, Nowacka A, et al.,  2011, Filter-exchange PGSE NMR determination of cell membrane permeability, Magn Reson Med., 66(2):356-65
[5] Schilling F, Ros S, et al., 2017, MRI measurements of reporter-mediated increases in transmembrane water exchange enable detection of a gene reporter, Nat Biotechnol., 35(1):75-80
[6] Heid I, Steiger K, et al., 2017, Co-clinical Assessment of Tumor Cellularity in Pancreatic Cancer, Clin Cancer Res., 23(6):1461-1470
Figure 1: FEXI application in mouse brain.
(a) T2-weighted image with ROI depicted in red. (b) FEXI image with diffusion filter off. The red arrow indicates artifact due to poor fat suppression. (c) FEXI ADC maps with the diffusion filter off/on (i/ii respectively) and different mixing times. (d) AXR map with the same ROI as in (a). (e) AXR relaxation curve with AXR fit calculated from the mean ADCs in the red ROI, which was copied to the FEXI ADC maps of (c). The blue point represents the mean unfiltered ADC (ci) and the red points represent the filtered ADC (cii) at different mixing times.
Figure 2: FEXI applications in a mouse PDAC tumor (a) and in a mouse EL4 lymphoma.
(a, b i) T2-weighted images with the tumor ROIs depicted in red. (a, b ii) AXR maps with the same ROIs. (a, b iii) AXR relaxation curves with AXR fits calculated from the mean ADCs in the red ROIs, which were copied to the FEXI ADC maps. The blue point represents the mean unfiltered ADC and the red points represent the filtered ADC at different mixing times.
Keywords: exchange, AXR, FEXI, permeability, ADC
1018

MRI-imaging and numerical modeling for stent implantation in the aorta

Dandan Ma1, 3, Mueed Azhar1, Ansgar Adler1, Michael Steinmetz1, Yong Wang2, 3, Martin Uecker1, 3

1 University Medical Center Göttingen, Goettingen, Germany
2 MPI for Dynamics and Self-Organization, Goettingen, Germany
3 DZHK (German Center for Cardiovascular Research) Partner Site Göttingen, Goettingen, Germany

Introduction

Implantation of a stent is the recommended therapy for the coarctation of the aorta (CoA), which is generally performed based on clinical experience without a theoretical evaluation to predict the outcome.[1] Magnetic resonance imaging (MRI) has been recently proposed as a promising tool for flow analysis of the thoracic aorta.[2,3] However, MRI can only visualize blood flow while computational fluid dynamics (CFD) can predict the result after surgery.[4] In this work, we compared MRI measurements of turbulent blood flow in 3D-printed models with numerical simulations.

Methods

3D models of the aorta of a 13 year-old patient with CoA, before and after surgery, were reconstructed from MRI images obtained at University Medical Goettingen. Those models were then 3D-printed using a Connex3 (Stratasys) with a biocompatible material MED610. Flows inside the models were obtained experimentally using two-dimensional (2D) and four-dimensional (4D) flow sequences on a 3T Siemens MRI scanner. Using the same geometries, the lattice Boltzmann method (large eddy simulation) was then used to simulate the complex blood flow in the aorta before and after surgery and the numerical results were then compared with the experimental data.

Results/Discussion

The flows of the whole aorta were resolved within the two geometries. To perform quantitative comparisons between the MRI measurements and numerical results for the mean velocity magnitude and velocity contour, three locations, in the ascending aorta, the arch, and the descending aorta respectively, were considered. Good agreement between the MRI and CFD results is observed (Figs. 1,2) with relative errors less than 15% on the three planes. A visualization of streamlines shows the complexity of the flow within the patient’s aorta, especially in the region of the stenosis. Jet flow and recirculation, which may cause high blood pressure and high wall shear stress, are observed. As the aorta is deformed and flatted after surgery, the flow resistance therein is reduced and the pressure drop (CFD) is also reduced. Those results indicate that the stent implantation restored the aortic flow effectively.

Conclusions

By comparing the results from MRI and CFD, we conclude that the MRI experimental data agree well with CFD results. The results show that both pressure drop and maximum WSS are reduced after stent implantation. MRI and CFD is a powerful combination for quantifying and predicting aortic blood flows in CoA before and after surgery.

Disclosure

I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

References
[1] Doshi AR, Chikkabyrappa S. (2018). Coarctation of aorta in children. Cureus.10: e3690.
[2] Frydrychowicz A, Francois CJ, Turski PA. (2011). Four-dimensional phase contrast magnetic resonance angiography: potential clinical applications. Eur J Radiol, 80:24-35.
[3] Hope TA, Herfkens RJ. (2008). Imaging of the thoracic aorta with time-resolved three-dimensional phase-contrast MRI: a review. Semin Thorac Cardiovasc Surg, 20: 358-364.
[4] Mathias N, Martin G, Leonid G, Marcus K, Titus K, Anja H. (2015). Interactive virtual stent planning for the treatment of coarctation of the aorta. International Journal of Computer Assisted Radiology and Surgery, 11(1): 133-144.
Figure 1 Comparison of streamline within pre-surgery aorta (MRI, CFD)
Figure 2 Comparison of streamline within post-surgery aorta (MRI, CFD)
Keywords: MRI, numerical modeling, blood flow, coarctation of the aorta
1019

Non-invasive Detection of Necrotic and Apoptotic Cell Death Using Filter-Exchange Spectroscopy

Jakob Wohlhüter1, Athanasia Kaika1, Mathias Schillmaier1, Rickmer Braren2, Franz Schilling1

1 Technical University of Munich, School of Medicine, Klinikum rechts der Isar, Department of Nuclear Medicine, Munich, Germany
2 Technical University of Munich, School of Medicine, Institute of Radiology, Munich, Germany

Introduction

Living cells actively control their plasma membrane water exchange using aqueous pore proteins such as aquaporins or urea transporters1. During cell death, the membrane degenerates and becomes permeable for water2. Based on magnetic resonance ‘filter-exchange spectroscopy’ (FEXSY) the water exchange between the cell and the surroundings can be measured as an apparent exchange rate (AXR) non-invasively3. In this work, we show that AXR measurements sensitively detect early stages of apoptotic and necrotic cell death in vitro.

Methods

Phantoms

Experiments in Fig. 1 used CHO-K1 cell suspensions at 16-18°C. Cells were treated for 24 hours with okadaic acid (50nM,100nM) to induce apoptosis4. Experiments in Fig. 2 used yeast cells treated with different concentrations of isopropanol (20%, 25%, 30%) for 30 minutes prior to measurements being performed at 16°C.

MR measurement
MRI hardware: 7T small animal scanner (Bruker/Agilent)
FEXSY acquisition parameters: bfi= 1300s/mm2; bi= 58-977s/mm2; tm= 24-305ms
For every sample five to six serial measurements were performed.

Image analysis
For the image analysis the theory of Lasič et al.5 was used to model the signal intensity, the filtered ADC and the filter efficiency.
Dead cells were determined using an automated cell counter (Countess II FL) (Fig. 1 d,e) and Trypan blue staining.

Results/Discussion

Experiments with yeast cells in solutions with increasing concentrations of isopropanol showed at an increase of cell membrane permeability and a higher AXR (Fig. 2) for treated cells up to isopropanol concentrations of 25%. Interestingly, for 30% isopropanol, AXR values are higher compared to control cells but lower compared to cells treated with 25% isopropanol. This might be induced by cell degeneration and rupture of the cell membrane at higher isopropanol concentrations causing the AXR to decrease due to homogenous mixing of cell interior and water. In another experiment we induced apoptosis in CHO-K1 cells and observed an increase of the AXR in parallel with an increasing number of dead cells (Fig. 1 a,c). The mean AXR increases from 3.11 s-1 to 4.45 s-1 as the percentage of dead cells grows from 1% to 70% (Fig. 1 b,c).

Conclusions

We have shown that AXR measurements are sensitive to detect changes in cell membrane permeability in cells undergoing apoptotic as well as necrotic cell death. In addition, our experiments indicate that AXR correlates with the fraction of dead cells. Since FEXSY can be implemented on common clinical MRI scanners and does not require exogenous contrast agents it could provide a novel way of non-invasive detection of cell death.

Disclosure

I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

References
[1] Agre, Peter 2004, 'Aquaporin water channels' (Nobel Lecture), Angewandte Chemie, 43, 4278–4290
[2] Ziegler, U.; Groscurth, P. 2004, 'Morphological features of cell death', News in physiological sciences, 19, 124–128
[3] Aslund, Ingrid; Nowacka, Agnieszka; Nilsson, Markus; Topgaard, Daniel 2009, 'Filter-exchange PGSE NMR determination of cell membrane permeability', Journal of magnetic resonance , 200(2), 291-295
[4] Huynh-Delerme, C.; Fessard, V.; Kiefer-Biasizzo, H.; Puiseux-Dao, S. 2003, 'Characteristics of okadaic acid--induced cytotoxic effects in CHO K1 cells', Environmental Toxicology 18 (6), 383–394
[5] Lasič, Samo; Nilsson, Markus; Lätt, Jimmy; Ståhlberg, Freddy; Topgaard, Daniel 2011, ‘ Apparent exchange rate mapping with diffusion MRI ’, Magnetic Resonance in Medicine , 66 (2) , 356–365
Figure 1: Induced apoptosis leads to a higher AXR in CHO-K1 cells.
(a) AXR values of two untreated (1,2) and two with Okadaic acid treated (3,4) samples of CHO-K1 cells. (b) AXR fits of the first measurements of the samples 2,3 and 4. The error bars shows the standard deviation of the different measurements of the same sample.[FS1]  (c) Percentage of dead cells in the samples before the measurements, determined with the automatic cell counter. (d,e) Images from the cell counter of sample 2(d) and 4(e).

[FS1]Move this sentence to the figure caption

Figure 2: AXR of yeast samples with by isopropanol induced necrosis.
Keywords: AXR, FEXSY, permeability
1020

Ex vivo bioreactor setup for multi-sample use in hyperpolarized 13C based nuclear magnetic resonance studies

Nichlas V. Christensen1, Christoffer Laustsen1, Mathilde H. Lerche2, Juan  D. Sanchez3, Jan Henrik Ardenkjær-Larsen3, Lotte B. Bertelsen1

1 Aarhus University, Department of Clinical Medicine, Aarhus N, Denmark
2 Technical University of Denmark, Department of Health Technology, Kgs. Lyngby, Denmark
3 Technical University of Denmark, Department of Health Technology, Kgs. Lyngby, Denmark

Introduction

Dissolution dynamic nuclear polarization (DNP) can improve the sensitivity of nuclear magnetic resonance spectroscopy (NMR) more than 10,000-fold[1]. This has proven very useful in metabolic studies of in vitro systems however these are limited by low throughput[2]. We here investigate a proof-of-concept bioreactor and microcoil setup that allows the DNP-based NMR study of multiple ex vivo samples with the feasibility to perform multiple measurement simultaneously, creating an easy-to-use and robust way to simultaneously run multi-sample experiments.

Methods

The bioreactor setup consists of a central computer for control of the various systems, an Elveflow setup for media perfusion and bioprobe injection, a 9.4 T Agilent imaging system, and a custom built multichannel microcoil system (Figure 1). The microcoil system consists of a 3D-printed housing module equipped with a triple-well plate attached to three receiver coils and the associated electronics. Using this system, hyperpolarized [1-13C]pyruvate (SpinAligner[3]) was injected into the perfusion medium resulting in  a final concentration of 10-15 mM upon arrival in the wells after about 40 seconds. A 1D carbon-13 timeseries experiment with a flip angle of 10 degrees were performed on a control sample of media-filled wells and a sample with slices of rat brain tissue in each well.

Results/Discussion

Summing the time-series of the various data in each of the respective wells showed that all coils were sensitive enough to detect lactate, a downstream product of pyruvate produced in the rat brain tissue (Figure 2). This is both evident from visual inspections of the spectra as well as when comparing the normalized lactate area integrals of control [MHL1] and brain tissue samples. Although small variances exist between the coils in terms of their proximity and angle of position to their respective wells, which contributes to a varying sensitivity among them, this experiment demonstrated a proof-of-concept for the setup with perfusion of brain tissue in multiple channels simultaneously. With further optimization and upscaling, this setup will undoubtedly be of high potential in high throughput studies of various bioprobes and tissue types.

Conclusions

The potential of our customized bioreactor and microcoil setup for use in high throughput DNP-based NMR has demonstrated the feasibility to detect downstream products of pyruvate in ex vivo tissue with all three coils, giving the proof-of-concept. Continued optimization will strengthen the potential uses of the setup.

Acknowledgement

Mette Dalgaard and Duy Anh Dang is acknowledged for their expert laboratory assistance.

Disclosure

I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

References
[1] Ardenkjaer-Larsen JH, Fridlund B, Gram A, et al. ”Increase in signal-to-noise ratio of > 10,000 times in liquid-state NMR”, Proc Natl Acad Sci U S A, 2003 Sep, vol. 100(18), pp. 10158-63.
[2] Cavallari E, Carrera C, Di Matteo G, Bondar O, Aime S, Reineri F, “In-vitro NMR Studies of Prostate Tumor Cell Metabolism by Means of Hyperpolarized [1-13C]Pyruvate Obtained Using the PHIP-SAH Method”, Front Oncol., 2020 Apr, vol. 10:497.
[3] Ardenkjær-Larsen JH, Bowen S, Petersen JR, Rybalko O, Vinding MS, Ullisch M, Nielsen NC., “Cryogen-Free dissolution Dynamic Nuclear Polarization polarizer operating at 3.35 T, 6.70 T and 10.1 T”, Magnetic Resonance in Medicine, 2019 Mar, vol. 81(3), pp. 2184-2194.
Figure 1
Illustration of bioreactor setup.
Figure 2
Resulting sum of hyperpolarized time-series for individual wells and samples. Control sample in deep-blue and rat brain tissue sample in brown-red.
Keywords: Hyperpolarized 13C NMR, Bioreactor, Ex vivo metabolism, Microcoil
1023

The effects of nitroxide structure upon Overhauser dynamic nuclear polarization efficacy at ultralow-field

Kai Buckenmaier1, Paul Fehling1, Sergey Dobrynin2, Denis Morozov2, Yulia Borozdina1, Jörn Engelmann1, Yuliya Polienko2, Yulia Khoroshunova2, Klaus Scheffler1, Goran Angelovski1, Igor Kirilyuk2

1 Max Planck Institute for Biological Cybernetics, Tübingen, Germany
2 Novosibirsk Institute of Organic Chemistry, Novosibirsk, Russian Federation

Introduction

Overhauser dynamic nuclear polarization (ODNP) is a hyperpolarization method, which transfers electron spin order to protons. In contrast to other DNP mechanisms (solid effect, cross effect, etc.), the Overhauser effect allows for the hyperpolarization of liquids and can be used for in vivo Overhauser MRI (OMRI) at ultralow-fields (< 10 mT). For OMRI usually trityl radicals or nitroxides such as carboxy-PROXYL and TEMPO in mM concentrations are used. In this study a great variety of nitroxide free radical molecules are being investigated in respect to their physical ODNP properties.

Methods

All measurements were performed with a home-made superconducting quantum interference device based field cycling ULF-MRI setup operating at B0 = 92 ± 0.8 mT and hyperpolarization field Bp = 1—10 mT. For radical characterization, the hyperpolarization field was set to 2 – 4 mT, matching an electron Larmor frequency of we = (120 ± 1) MHz. The spin probes were dissolved in PBS (Phosphate Buffered Saline) and pH adjusted to 7.3. For TEMPO a 2 mM concentration leads to the highest enhancement at a moderate HF power level P. For the sample characterization, the leakage factor f, the product of the coupling constant and the maximum saturation factor xsmax, the maximum theoretical possible enhancement Emax and the power P1/2 needed to reach 0.5∙Emax have been determined.

Results/Discussion

More than 25 different nitroxide free radicals were synthesized and characterized. The focus was set on the interpretation of P1/2 and Emax. These are the most relevant parameters for application. It was found that P1/2 is strongly dependent on the line width of the ODNP spectrum. Additionally, 15N labelled nitroxides will have improved P1/2 over 14N counterparts since the losses from mixing of the energy levels are reduced. The neighboring substituents of the nitroxide group seem to have an important influence on the linewidth, where methyl substituents (blue circles in fig.) lead to the lowest P1/2. In our experiments, pyrrolidines showed better P1/2 than piperidines. The molecular weight has a significant influence on Emax, but not on P1/2 (see fig. A and B). This indicates a correlation between the tumbling rate and Emax. The results also suggest that polyradicals with significant spin-spin coupling exhibit too much spectral line broadening to produce a relevant ODNP enhancement.

Conclusions

The ideal nitroxide radical would be a lightweight, deuterated 15N-pyrrolidine monoradical, with neighboring methyl substituents and a narrow linewidth ODNP spectrum. Such a radical is likely to have a high maximum enhancement Emax at a low P1/2. The presented results provide a list of ODNP properties of different spin probes, which helps to investigate the functionalization of free radicals by incorporation into macro- or carrier molecules.

Acknowledgement

The synthesis of 2,2,5,5-tetraethylpyrrolidine-1-oxyls was supported by the Russian Foundation for Basic Research (grant 18-53-76003 within the framework of the ERA.Net RUS+ project ST2017-382: NanoHyperRadicals).

Disclosure

I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

Figure

Emax A and P1/2 B of non-deuterated and 14N Nitroxide monoradicals over molecular weight.

Keywords: ODNP, Nitrioxide free radicals, ULF MRI
1024

A variable resolution sequence for hyperpolarized 13C MRI in kidney disease

Camilla W. Rasmussen1, Nikolaj Bøgh1, Esben S. S. Hansen1, Rolf F. Schulte2, Thomas H. Thorsen1, Christoffer Laustsen1

1 Aarhus University, The MR Research Centre, Department of Clinical Medicine, Aarhus N, Denmark
2 GE Healthcare, Munich, Germany

Introduction

Hyperpolarized (HP) MRI shows potential for metabolic evaluation of kidney disease.1 However, the signal-to-noise-ratio (SNR) is often a limiting factor when imaging downstream metabolites of [1-13C]pyruvate. Pyruvate and its metabolites have a 10-100-fold difference in SNR.2 Achieving high resolution pyruvate data compromises the imaging of the metabolites. This hinders differentiation of the intra-renal metabolic differences between the medulla and cortex. We aimed to evaluate a novel strategy using variable time and spatial resolution in the porcine kidney to overcome these challenges.2

Methods

Variable resolution hyperpolarized [1-13C]MRI was performed in one healthy pig and in one pig 7 days after kidney ischemia-reperfusion-injury (n=2). Data were measured on a 3T MRI (GE MR750, GE Healthcare) equipped with a 13C clamshell transmit and a 16-channel receive coil (RAPID Biomedical). Spectral-spatial imaging was performed with a spiral readout. Pyruvate/metabolites were excited with 10°/70° (TR = 500 ms) in an interleaved fashion, yielding 1s and 3s time-resolution, respectively (Fig. 1). Different spirals (FOV = 34 cm2, TE = 10 ms) were used for pyruvate (matrix size = 90x90, 32 ms readout, 2 arms) and the metabolites (matrix size = 40x40, 32 ms readout, 1 arm). Images were reconstructed in Matlab and zero-filled to 180x180 before quantification.

Results/Discussion

For the healthy pig, the SNR was 15.2 ± 3.5, 26.6 ± 6.4, 10.8 ± 3.3, and 49.5 ± 12.5 for pyruvate, lactate, bicarbonate, and alanine, respectively. These were comparable in the injured pig (16.3 ± 3.1, 61 ± 20.5, 9.9 ± 4.1, and 56.5 ± 21). Thus, the approach provided sufficient SNR for all resonances. The contrast-to-noise ratios between medulla and cortex across all kidneys were 1.9 ± 0.4 for pyruvate, 8.5 ± 5.5 for lactate, 2.7 ± 0.4 for bicarbonate, and 12.1 ± 10.4 for alanine, highlighting the ability to differentiate between cortex and medulla. The high-resolution pyruvate data allowed visualization of the larger renal vessels (Fig. 2a); further, the pyruvate mean transit time was longer in the medulla than the cortex (5.36 ± 2.4 vs 6.29 ± 2.6). The cortical lactate-to-pyruvate ratio was 0.34 ± 0.15 in the healthy kidneys vs 0.85 in the injured kidney. The alanine-to-pyruvate ratio was 0.37 ± 0.07 in healthy vs 0.25 in the injured.

Conclusions

A variable resolution strategy refines HP MRI resulting in higher resolution of pyruvate and quantification of lower SNR metabolites at coarser resolution. This strategy improves differentiation of the renal medulla from cortex, and a superior estimate of perfusion is obtained by an improved time-resolution of pyruvate. This results in a better foundation of assessing kidney metabolism in preclinical and clinical studies with HP MRI.

Disclosure

B: Rolf F. Schulte is working at GE Healhtcare. Otherwise, nothing to declare.

References
[1] Schroeder, M. & Laustsen, C., 2017, Imaging oxygen metabolism with hyperpolarized magnetic resonance: a novel approach for the examination of cardiac and renal function, Biosci Rep, Feb 28
[2] Gordon, J. W. et al., 2020, A variable resolution approach for improved acquisition of hyperpolarized (13) C metabolic MRI. Magn Reson Med, Dec, 2943-2952
Figure 1

Figure 1: Illustrating the hyperpolarized MRI setting with imaging of pyruvate in interleaves between its metabolites (A). Spirals used for imaging of pyruvate and its metabolites (B).

Figure 2

Figure 2: The right kidney represents ischemia-reperfusion-injury and the left kidney serves as the healthy control. Pyruvate was imaged at high spatial (matrix = 90x90, field-of-view = 34 cm2) and temporal resolution (1s, A), while the lower SNR metabolites were sampled at lower spatial (matrix = 40x40, field-of-view = 34 cm2) and temporal resolution to ensure accuracy (3s, B). The higher resolution of pyruvate is used in quantification of metabolic ratios or perfusion estimates such as the mean transit time (MTT) (C).
 

Keywords: MRI, hyperpolarization, carbon-13, pyruvate, kidney disease
1025

Short-term effects of transcranial direct current stimulation in rat auditory cortex: a 9.4 T MR-spectroscopy study

Fahmida Akter1, Patricia Wenk1, Kai Heimrath2, Tino Zähle2, Eike Budinger1

1 Lebiniz-Institut für Neurobiologie, COMBINATORIAL NEUROIMAGING CORE FACILITY, MAGDEBURG, Germany
2 Otto von Guericke University Magdeburg, Department of Neurology, MAGDEBURG, Germany

Introduction

Transcranial direct current stimulation (tDCS) is an increasingly popular approach to modulate neuronal activity non-invasively. However, potential changes in GABA and glutamate (Glu) levels associated with tDCS remain still unclear. Magnetic resonance spectroscopy (MRS) is a powerful technique to quantify small molecules in vivo. So far, there are only a few MRS studies reporting effects of tDCS on neurotransmitter level in sensory cortex1. In this study, we investigated changes in neurotransmitter concentrations in auditory cortex (AC) before, during and shortly after tDCS in rats using MRS.

Methods

During examination inside the scanner (Bruker 94/20 UHR,  1H volume resonator 075/040), rats were anaesthetized with 2 vol% isoflurane. Gold plate electrodes were placed on the skin over both AC regions to apply anodal and cathodal tDCS (0.3mA) to right and left AC in separate sessions (Fig. A). A spectroscopy voxel covering either the left or right AC region (measured in separate sessions as well) was placed based on a T2-weighted anatomical reference scan before acquiring the spectra using a STEAM sequence. In each session, MRS spectra were obtained before (baseline), during tDCS and two times after tDCS (15 min, 30 min; Fig. A). MRS data were analyzed using LCModel.

Results/Discussion

Data from left and right AC were pooled together for anodal and cathodal stimulation respectively. Neurotransmitter quantification from ten animals was used for further statistical analysis. We performed two-way ANOVA of Glu and GABA content and the Glu/GABA ratio within subject factor (anodal and cathodal stimulation) and time factor. Multiple comparison test (Turkey) between the time points showed that anodal tDCS significantly increases the Glu/GABA ratio (Fig. D). Thereafter, this ratio significantly drops, returning to the baseline level within 30 min. There is a trend of a slightly decreasing Glu/GABA ratio with and after cathodal tDCS (Fig. D). Individual GABA levels initially start to decrease with anodal tDCS but largely return to baseline after stimulation, whereas there is a trend of increasing GABA levels with and after cathodal tDCS (Fig. C). We did not find any significant changes in individual Glu levels neither with anodal nor with cathodal tDCS in rat AC (Fig. B).

Conclusions

Our study demonstrates profound tDCS-induced changes in neurotransmitter concentration in rat AC. Thereby, the modulatory effect of anodal tDCS is more evident in the Glu/GABA ratio than in single transmitter level changes. Our results represent an important step towards understanding the short-term effects of non-invasive electrical stimulation for the therapy of neurotransmitter-related neurological diseases.

Acknowledgement

The study was supported by a CBBS NeuroNetwork NeeMo (cbbs.eu). We would like to thank our master student Joseph Stokes and technical assistant Janet Stallmann.

Disclosure

I have no financial interest or relationship to disclose regarding the subject matter of this presentation.

References
[1] Heimrath K, Brechmann A, Blobel-Lüer R, Stadler J, Budinger E, Zaehle T (2020) Transcranial direct current stimulation (tDCS) over the auditory cortex modulates GABA and glutamate: a 7 T MR-spectroscopy study. Sci Rep 10:1–8 
Experimental paradigm, study design and results
Keywords: tDCS, MRS, spectroscopy
1026

Photo-CIDNP-based hyperpolarization of kynurenic acid – a neuroactive substrate

Markus Plaumann1, Christian Bruns1, Frederike Euchner1, Johannes Bernarding1

1 Otto-von-Guericke-University Magdeburg, Institute of Biometry and Medical Informatics, Magdeburg, Germany

Introduction

Due to its neuroactive activity kynurenic acid has an influence on important neurophysiological and neuropathological processes.[1] Elevated levels of kynurenic acid have been linked to tick-borne encephalitis, schizophrenia, and HIV-related diseases. It can be detected in urine as a metabolic product of L-tryptophan. In the present study, we are investigating the nuclear spin hyperpolarization of this compound inside a strong magnetic field of 7 T. Therefore, we used photo-Chemical Induced Dynamic Nuclear Polarization (photo-CIDNP)[2,3], which is established in our laboratory.[4]

Methods

For the first investigations, a stock solution consisting of 2.8 mg kynurenic acid, 0.5 mg riboflavin 5’-monophosphate sodium salt hydrate and 3 ml D2O was used. 600 µL of this saturated solution were used for the first experiments. MR-spectroscopic measurements were performed in a 5 mm NMR tube on a 7 T NMR system (Bruker WB-300 Ultrashield). In all examinations an optical fiber is connected to a Cree XP E high power LED (455 nm) and was centrally positioned in the aqueous solution.[4] Four irradiation times (0.5 s, 2.0 s, 6.0 s and 12 s) were chosen and 32 scans per measurement were used for the detection of 1H (P1=29.0 µs, PL1=17 W) NMR spectra.

Results/Discussion

Previous studies have shown that photo-CIDNP can hyperpolarize molecules, e.g.,  tyrosine or tryptophan in aqueous solutions. The kynurenic acid (Figure 1) is an excellent example to investigate the polarization distribution in ring systems. This is due to the following advantages: a) a hydrogen nucleus adjacent to the hydroxyl group that does not couple with other protons, b) a hetero nucleus (N) in the ring system and c) another aromatic ring system. The spectroscopic data (Figure 2) clearly show that under the given conditions, only one signal in the 1H-NMR spectrum is influenced by the light irradiation and can be detected increasingly with negative amplitude. A structure elucidation shows that this signal can be assigned to the isolated proton. These measurements show that proton signal amplifications (possibly with phase reversal) can also be detected in a strong magnetic field without the addition of biologically incompatible compounds such as metal complexes.

Conclusions

Biomolecules can be hyperpolarized inexpensively (LED construction) and reproducibly in high magnetic fields. The gain factor depends on the radiated light output and the lighting duration. For the last aspect, with the given system, a maximum can be determined with an irradiation time of 6 seconds. In addition to the measurements shown here, studies based on this have already been carried out.

Disclosure

I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

References
[1] Ocampo, JR, Huitrón, RL, González-Esquivel, D, Ugalde-Muñiz, P, Jiménez-Anguiano, A, Pineda, B, Pedraza-Chaverri, J, Ríos, C, Pérez de la Cruz, V, ‘Kynurenines with Neuroactive and Redox Properties:Relevance to Aging and Brain Diseases’, Oxidative Medicine and Cellular Longevity, 2014, ID 646909.
[2] Bargon, J, Fischer, H, Johnsen, U, ‘Kernresonanz-Emissionslinien während rascher Radikalreaktionen. I. Aufnahmeverfahren und Beispiele’, Zeitschrift Naturforschung Teil A, 1967, 22, 1551-1555.
[3] Goez, M, ‘Chapter 3 Photo-CIDNP Spectroscopy’, Annual reports on NMR Spectroscopy, 2009, 66, 77-147.
[4] Bernarding, J, Euchner, F, Bruns, C, Ringleb, R, Müller, D, Trantzschel, T, Bargon, J, Bommerich, U, Plaumann, M, ‘Low-cost LED-based Photo-CIDNP Enables Biocompatible Hyperpolarization of 19F for NMR and MRI at 7 T and 4.7 T’, ChemPhysChem., 2018, 19, 2453–2456.
Figure 1
Molecule structure of kynurenic acid.
Figure 2
1H NMR spectra measured during light exposure. Each shown spectrum is an average of 32 scans. In these measurements, the rise in the baseline can be explained by a large water signal at 4.8 ppm.
Keywords: hyperpolarization, photo-CIDNP, kynurenic acid, riboflavin 5’-monophosphate sodium salt hydrate
1027

ParaHydrogen polarized ethyl-[1-13C]pyruvate in water, a key substrate for fostering the PHIP-SAH approach to metabolic imaging

Carla Carrera2, Eleonora Cavallari1, Giuseppe Digilio3, Oksana Bondar1, Silvio Aime1, Francesca Reineri1

1 Universita' degli studi di Torino, BMSS, Torino, Italy
2 National Research Council, IBB, Torino, Italy
3 Università del Piemonte Orientale, Science and Technologic Innovation, Alessandria, Italy

Introduction

Ethyl pyruvate, hyperpolarized (HP) by means of d-DNP, was used as a metabolic probe in the brain,[1,2] thanks to its ability to cross the blood-brain barrier with greater efficiency than pyruvate. ParaHydrogen hyperpolarization is simpler, faster and cheaper than d-DNP. In this work, HP ethyl‑[1‑13C]pyruvate was obtained from vinyl‑[1‑13C]pyruvate by means of the PHIP-SAH procedure.[3] The vinyl ester of pyruvate was synthesized thanks to an efficient synthetic procedure which overcomes the intrinsic instability of this molecule.

Methods

In order to allow the trans-vinylation reaction of pyruvate, the ketal derivative of pyruvate was synthesized and reacted with vinyl acetate (pyridine complex of palladium diacetate as catalyst). The vinyl-[1-13C]pyruvate thus obtained, after deprotection with Trifluoroacetic acid, was hydrogenated with parahydrogen (2bar, pH2 85%) using a rhodium(I) catalyst, in CDCl3 and in water solution. Net 13C magnetization was obtained by means of magnetic field cycling (MFC)[4]. The metabolic transformation of HP ethyl-[1-13C]pyruvate into [1-13C]-pyruvate was assessed using an aqueous solution of esterase.

Results/Discussion

The 13C polarization on the carboxylate signal of pyruvate ester, obtained from parahydrogenation of vinyl‑[1‑13C]pyruvate in chloroform, was 3.8 ± 0.3%, while 6.2 ± 0.3 % was obtained from the propargylic ester, in the same condition. In order to obtain HP ethyl-[1-13C]pyruvate in an aqueous solution, hydrogenation has been carried out using the water-soluble analog of the rhodium complex. In this case, 13C polarization on ethyl pyruvate was 12.3 ± 0.3 %, while on the hydrated form (76% of the total) is 2.4 ± 0.1 %. The average 13C polarization is 4.8 % (average of hydrated and non-hydrated form). HP ethyl-[1-13C]pyruvate in aqueous solution was added to a buffered solution of esterase: the build-up of the signals of [1‑13C]pyruvate was clearly observed in the series of 13C-NMR spectra acquired immediately after the addition of the HP-ester to the enzyme containing solution.

Conclusions

Vinyl-[1-13C]pyruvate was synthesized and 13C HP ethyl-[1-13C]pyruvate was obtained, by means of PHIP-SAH method, in both organic and aqueous solutions. Its polarization level was measured and compared with that of allyl-[1-13C]pyruvate. The in vitro metabolic transformation of the HP ethyl-[1-13C]pyruvate, catalyzed by an esterase, was observed. This substrate is a good candidate as PHIP HP probe for in vivo metabolic investigations.

Acknowledgement

This project has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie (Grant Agreement No. 766402) and the FETOPEN program (Grant agreement 858149, proposal acronym Alternatives to Gd).

Disclosure

I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

References
[1] Hurd R.E., Yen Y.F., Mayer D., Chen A., Wilson D., Kohler S., et al., 'Metabolic Imaging in the Anesthetized Rat Brain Using Hyperpolarized [1-13C] Pyruvate and [1-13C] Ethyl Pyruvate', Magn. Reson. Med. 2010, 63, 1137–1143.
[2] Miller J. J., Grist J. T., Serres S., Larkin J. R., Lau A. Z. et al. '13C Pyruvate Transport Across the Blood-Brain Barrier in Preclinical Hyperpolarised MRI', Sci. Rep., 2018, 8, 1–15
[3] Reineri F., Boi T., Aime S., 'ParaHydrogen Induced Polarization of 13C carboxylate resonance in acetate and pyruvate' Nat. Commun. 2015, 6, 1–6
[4] Cavallari E., Carrera C., Sorge M., Bonne G., Muchir A., Aime S., Reineri F., 'The 13C hyperpolarized pyruvate generated by ParaHydrogen detects the response of the heart to altered metabolism in real time', Sci. Rep., 2018, 8, 2–10
Figure 1
(A) Series of 13C-NMR spectra acquired upon mixing the pyruvate esterase containing solution (400 U) with the aqueous solution of hyperpolarized ethyl‑[1-13C]pyruvate. Spectra were acquired using small flip angle pulse (10°) and 2 s delays between successive scans. (B) expanded 13C‑NMR spectrum at maximum intensity of the [1-13C]pyruvate signal; (C) Plot of the time dependent changes of the signals of ethyl-[1-13C]pyruvate, its hydrated form, and of the [1‑13C]pyruvate (obtained from the integrals of the peaks in the 13C-NMR spectra reported in (A))
Scheme 1
Synthetic pathway for the preparation of vinyl-[1-13C]pyruvate and HP ethyl-[1-13C]pyruvate. The 13C label is indicated in red, the p-H2 in blue.
Keywords: hyperpolarization, para-hydrogen, pyruvate, metabolism
1028

Hyperpolarization of bisfluorinated pyridine derivatives

Isabell Prediger1, Christian Bruns1, Frederike Euchner1, Johannes Bernarding1, Markus Plaumann1

1 Otto-von-Guericke-University Magdeburg, Institute of Biometry and Medical Informatics, Magdeburg, Germany

Introduction

Since Adams et al. published the first measurements of parahydrogen-induced nuclear spin hyperpolarization of pyridine in 2009[1], there have been numerous advances.[2],[3] The SABRE (signal amplification by reversible exchange) method, allows a repeatable hyperpolarization. However, influences of individual substituents were rarely examined. For this reason, we have investigated the effect of different substituents such as cyano or amino groups in the 2-position of bisfluorinated pyridines. The 1H and 19F NMR data provide information about the signal enhancements to be achieved.

Methods

Three pyridine derivatives (3,5-difluoropyridine (3), 2-cyano-3,5-difluoropyridine (4), 2-amino-3,5-difluoropyridine (5)) were selected for the measurements in 2 ml methanol-d4 and measured individually. After addition of the Ir-IMes-catalyst, the sample was degassed. The hyperpolarization was realized in a 10 mm NMR tube with about 50% enriched parahydrogen and 6 bar pressure. Directly after hydrogenation at about 6 mT, the NMR spectra were detected on a Bruker wide bore 300 MHz spectrometer. The obtained signal enhancements (SEs) were calculated from signal-to-noise ratios of the thermal and the hyperpolarized spectra. The results are compared with previous investigations for e.g. pyridine (1) and 3-fluoropyridine (2) compared.

Results/Discussion

The 1H NMR spectra of 3,5-difluoropyridine (3) and 2-amino-3,5-difluoropyridine (5) are shown as examples in this abstract in Figure 1. Both show enhanced proton signals with a negative sign. From older studies in our group it is known that the phase position depends on the selection of the Ir complex. Thus, opposite signs are observable when choosing the Crabtree catalyst. Signal enhancements of >150 were calculated for the 1H signals.

Figure 2 shows the 19F NMR spectra of 2-cyano-3,5-difluoropyridine (4) and 2-amino-3,5-difluoropyridine (5). Only in the case of 2-amino-3,5-difluoropyridine (5) the 19F signals are enhanced. If the amino group is exchanged for a CN group, no signal enhancement can be observed. Also in the 1H NMR spectrum of 2-cyano-3,5-difluoropyridine (4) no signals of hyperpolarized nuclei can be detected (not shown here). It can be stated that not only steric reasons have an influence on the polarizability of individual nuclei (1H, 19F and also 13C).

Conclusions

Knowing how functional groups such as F, CN, NH2, OH and CH3 influence nuclear magnetic hyperpolarization is of great interest. With this knowledge, tailor-made molecules can be synthesized and the polarization increased or redirected to specific groups. For example, as expected, mesomeric/inductive pushing and pulling effects influence nuclear spin hyperpolarization.

Disclosure

I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

References
[1] Adams, RW, Aguilar, JA, Atkinson, KD, Cowley, MJ, Elliott, PIP, Duckett, SB, Green, GGR, Khazal, IG, López-Serrano, J, Williamson, DC, ‘Reversible Interactions with para-Hydrogen Enhance NMR Sensitivity by Polarization Transfer’, Science, 2009, 323,1708-1711.
[2] Skinner, JG, Menichetti, L, Flori, A, Dost, A, Schmidt, AB, Plaumann, M, Gallagher, FA, Hövener, J-B, ‘Metabolic and Molecular Imaging with Hyperpolarised Tracers’, Mol. Imaging Biol., 2018, 20, 902–918.
[3] Hövener, J-B, Pravdivtsev, AN, Kidd, B, Bowers, CR, Glöggler, S, Kovtunov,  KV, Plaumann, M, Katz-Brull,  R, Buckenmaier, K, Jerschow, A, Reineri, F, Theis, T, Shchepin  RV, Wagner, S, Bhattacharya, P, Zacharias,  NM, Chekmenev, EY ‘Parahydrogen-Based Hyperpolarization for Biomedicine’, Angew. Chem. Int. Ed., 2018, 57, 11140 – 11162.
Figure 1
1H NMR spectra of hyperpolarized molecule (blue) and the corresponding spectrum in thermal equilibrium (red). Left: 3,5-difluoropyridine and right: 2-amino-3,5-difluoropyridine.
Figure 2
19F NMR spectra of hyperpolarized molecule (blue) and the corresponding spectrum in thermal equilibrium (red). Left: 2-cyano-3,5-difluoropyridine (4) and right: 2-amino-3,5-difluoropyridine.
Keywords: hyperpolarization, SABRE, parahydrogen, 19F, bisfluorinated
1029

Non-invasive blood half-life determination of fluorinated nanoparticles by 19F MRS

Francesca Garello1, Francesca Arena2, Lorena Consolino1, 3, Enzo Terreno1

1 University of Turin, Molecular and Preclinical Imaging Centers, Department of Molecular Biotechnology and Health Sciences, Torino, Italy
2 Italian National Research Council (CNR), Institute of Biostructures and Bioimaging (IBB), Torino, Italy
3 RWTH Aachen University, Department of Nanomedicine and Theranostics, Institute for Experimental Molecular Imaging, Faculty of Medicine, Aachen, Germany

Introduction

The blood half-life time of some fluorinated systems has been already evaluated in the past.1 However, invasive techniques consisting of blood collection followed by ex-vivo 19F Magnetic Resonance Imaging (MRI)/Magnetic Resonance Spectroscopy (MRS) were applied. The low sensitivity of the method, in fact, requires the collection of large blood volumes, reducing the sampling and the time window in which the blood clearance can be studied. In the present work, a non-invasive method to estimate the blood half-life of fluorinated compounds is presented.

Methods

C57BL/6J healthy mice were anesthetized and a catheter was inserted into the tail vein. The mice were positioned into a 7 T MRI scanner. Allocation of animals in the scanner was performed accurately to ensure that positioning and imaging could be consistent through all imaging sessions. A perfluorocarbon-based nanoemulsion (PFCE-NE) was injected via the tail vein (1 mmol/kg b.w). Immediately after the injection, 19F MR spectra were acquired every 43 s for 1h. Additional spectra were acquired 3,24, 48, and 72 h after the injection. A standard reference tube containing TFA was used to normalize the signal acquired during each imaging session. Pre-injection of liposomes 10 min before PFCE-NE was also performed to evaluate blood half-life variations in the case of Kupffer cells pre-saturation.

Results/Discussion

Since no extravasation of fluorinated particles has been reported in the brains of healthy mice, by placing in the volume coil only the head of the animal, all the signal detected by 19F MRI/MRS merely referred to the particles circulating in blood vessels. In this way, a large amount of data can be acquired, performing a precise sampling. The acquired spectra were integrated, and the resulting values were expressed in terms of % injected dose and plotted against time. Data obtained were fitted with a bi-exponential fitting curve. Resulting blood half-life times were respectively t1/2 fast = 46 min and t1/2 slow = 11.4 h, R2 = 0.99 (Figure 1). When liposomes were pre-injected to saturate Kupffer cells, available PFCE-NE was significantly higher only in the first hours post-injection (82.3 vs. 60.4% of the injected dose, 3 h post PFCE-NE injection, p < 0.05, Figure 2), but not in the following hours and days. No differences were found in blood half-life times.

Conclusions

The 19F-MRS method herein proposed for determining the blood half-time of fluorine-based systems in vivo has the advantage of allowing an assiduous sampling, for long time windows and without blood collection. However, it can be applied only to compounds with enough fluorine concentration and no extravasating in the central nervous system of healthy mice.

Acknowledgement

The authors acknowledge the Italian Ministry of Research for FOE contribution to the Euro-BioImaging MultiModal Molecular Imaging Italian Node (www.mmmi.unito.it).

Disclosure

I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation.

References
[1] Santaella, C, Frézard, F, Vierling, P, Riess, J.G. Extended in Vivo Blood Circulation Time of Fluorinated Liposomes. FEBS Lett. 1993, 336, 481–484
Figure 1
Blood half-life time calculation of PFCE-NE in healthy mice carried out by 19F MRS (n = 3). Data points fitted a bi-phasic exponential decay (R2 = 0.99).
Figure 2
Blood clearance of PFCE-NE with (green squares) or without (black circles) pre-injection of liposomes in order to pre-saturate Kupffer cells, calculated by 19F MRS in healthy mice (n = 6, * p < 0.05).
Keywords: Blood halflife, Fluorinated compounds, 19F MRS
1030

Clinically feasible B1 field correction of multi-organ sodium images at 3T

Michael Vaeggemose1, 3, Rolf F. Schulte2, Christoffer Laustsen3

1 GE Healthcare, Brøndby, Denmark
2 GE Healthcare, Advanced Science Lab, Munich, Germany
3 Aarhus University, MR Research Centre / Department of Clinical Medicine, Aarhus N, Denmark

Introduction

Recent years has seen a renewed interest in sodium MRI [1]. One important determinant of the sodium signal level when using surface coils is the transmit and receive B1 fields. The Bloch-Siegert off-resonance pulse approach [2] is shown to be faster, more robust, and yield higher SNR as to the dual angle approach [3]. The aim of this study is to evaluate Bloch-Siegert off-resonance B1 transmit correction of sodium images in thigh muscle, heart, kidney, and brain with the use of MRI in healthy human subjects using a 3D FLORET readout trajectory.

Methods

MRI examinations were performed on a 3T MRI scanner, allowing proton (1H) MRI as well as sodium (23Na) imaging. Scans were performed at the thigh muscle, heart, kidney, and brain on two healthy subjects. Sodium phantoms were placed in the field of view (32 and 80 mmolL-1) to determine tissue sodium concentration. Proton scans was applied as anatomy images followed by evaluation of the main magnetic field inhomogeneities. The radio frequency pulses were calibrated (amplitude transmit gain and frequency) to optimize the sodium imaging acquisition. B1 transmit field map was acquired using a Bloch-Siegert off-resonance approach [2] based on the Fermat Looped, Orthogonally Encoded Trajectories (FLORET) technique [4]. Sodium imaging was acquired as described by Nagel et al [5].

Results/Discussion

Table 1 indicates sodium levels before and after B1 transmit corrections as comparable with an increase in the corrected ROIs. The images become more homogeneous with improved anatomy (Figure 1). 

[Please review Table 1 in the attached figures]

Conclusions

The preliminary results indicated great potential of the Bloch-Siegert off-resonance based B1 transmit correction. This is seen in thigh muscle, heart, kidney, and brain. The initial benefits were most pronounced in kidney and brain. The complete quantitative 23Na imaging protocol, including B1 correction is acquired at 3T within 15min. This support the use of the fast B1 transmit mapping (2:07min) in the clinical setting.

Acknowledgement

The authors would like to thank radiographer Tau Vendelboe for contributing to scanning of the participating subjects.

Disclosure

M.V. and R.F.S. are employees of GE Healthcare. The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

References
[1] D. Burstein and C. S. Springer, “Sodium MRI revisited,” Magn. Reson. Med., vol. 82, no. 2, pp. 521–524, 2019, doi: 10.1002/mrm.27738.
[2] L. I. Sacolick, F. Wiesinger, I. Hancu, and M. W. Vogel, “B1 mapping by Bloch-Siegert shift,” Magn. Reson. Med., vol. 63, no. 5, pp. 1315–1322, May 2010, doi: 10.1002/mrm.22357.
[3] S. P. Allen et al., “Phase-sensitive sodium B1 mapping,” Magn. Reson. Med., vol. 65, no. 4, pp. 1126–1131, 2011, doi: 10.1002/mrm.22700.
[4] J. G. Pipe, N. R. Zwart, E. A. Aboussouan, R. K. Robison, A. Devaraj, and K. O. Johnson, “A new design and rationale for 3D orthogonally oversampled k-space trajectories,” Magn. Reson. Med., vol. 66, no. 5, pp. 1303–1311, 2011, doi: 10.1002/mrm.22918.
[5] A. M. Nagel, F. B. Laun, M. A. Weber, C. Matthies, W. Semmler, and L. R. Schad, “Sodium MRI using a density-adapted 3D radial acquisition technique,” Magn. Reson. Med., vol. 62, no. 6, pp. 1565–1573, 2009, doi: 10.1002/mrm.22157.
Table 1

Sodium levels at two region-of-interest (ROI) and contrast-to-noise ratio before and after B1 transmit correction at different anatomical locations (thigh muscle, kidney, heart, and brain). Sodium levels are in mmolL-1.  Sodium1 = ROI_1. Sodium2 = ROI_2.

Figure 1

B1 correction of sodium images. Original (first column), corrected (middle column), and proton image (last column). Anatomies: brain (A), heart (B), thigh muscle (C) and kidney (D). Circles are first region of interest (ROI) (yellow), second ROI (white) and noise ROI (red).

Keywords: Sodium imaging, MR Spectroscopy, B1 map, Clinical feasible protocol
1032

Imaging pH, metabolism and hypoxia using hyperpolarized 13C-MRI and [18F]FMISO-PET to assess metabolic connectivity and heterogeneity of the tumor microenvironment in glioblastoma

Martin Grashei1, Carolin Kitzberger2, Jason G. Skinner1, Sandra Sühnel1, Geoffrey J. Topping1, Elisabeth Bliemsrieder1, Christian Hundshammer1, Katja Steiger3, Rainer Glass4, Wolfgang A. Weber1, Christine Spitzweg2, Franz Schilling1

1 Technical University of Munich, Department of Nuclear Medicine, School of Medicine, Klinikum rechts der Isar, Munich, Germany
2 Ludwig-Maximilians-University Munich, Medizinische Klinik und Poliklinik IV-Campus Großhadern, University Hospital of Munich, Munich, Germany
3 Technical University of Munich, Department of Pathology, Klinikum rechts der Isar, Munich, Germany
4 Ludwig-Maximilians-University Munich, Neurosurgical Research University Clinics, Munich, Germany

Introduction

Glioblastoma (GBM) is a highly malignant and heterogeneous tumor of the central nervous system. Despite a large variety of existing treatment approaches1, prognosis and outcome are still devastating. However, therapy efficacy often depends on the tumor microenvironment such as extracellular pH and oxygenation and is often driven by tumor metabolism. Here, we studied GBM by multimodal imaging of pH2,3, pyruvate-lactate conversion and hypoxia using hyperpolarized (HP) 13C-MRI and 18F-PET to investigate links between hypoxia, pH and lactate production in these tumors.

Methods

Model: 10 CD-1-nu/nu mice injected subcutaneously with 1∙106 patient-derived GBM cells undergoing the protocol in Fig. 1a.

HP: 27 mg [1,5‑13C2,3,6,6,6-D4]zymonic acid and 25 mg 13C-urea were polarized by DNP3 and dissolved in 80 mM TRIS and D2O. 25 mg [1‑13C]pyruvate was polarized and dissolved in 80 mM TRIS and H2O.

HP MR(S)I: pH-imaging at 7T used FIDCSI with FA 15°, resolution 2x2x5 mm³, BW 3201 Hz, 256 points. Metabolic imaging used bSSFP4 with FAP/L = 4/90°, resolution (1.75 mm)³, 1.05 s per frame.

PET-Imaging: Mice were imaged with voxel size (0.8 mm)³ 3 hours after injecting 10‑15 MBq [18F]Fluoromisonidazole ([18F]FMISO).

Data Processing: pH maps were calculated in MATLAB2. Metabolism was quantified by AUC-ratios5.

Histology: Tumors (FFPE) were stained for carbonic anhydrase 9 (CAIX).

Results/Discussion

Tumor pH maps (Fig. 1b) reveal pH heterogeneity with acidified hotspots as low as pH = (7.17±0.11, n = 10) and light overall acidification pH = (7.34±0.02, n = 10). Imaging of pyruvate (Fig. 1c) and lactate (Fig. 1d) shows strong, heterogeneous lactate production quantified by AUC = (1.14±0.17, n = 8). No correlation between acidification and pyruvate-lactate metabolism is observed (Fig. 2b). [18F]FMISO-PET indicates increased tumor uptake with SUVmean = (0.49±0.05, n = 7) relative to muscle tissue (SUVmuscle = 0.11±0.01, n = 7)  (Fig. 2a), correlating strongly with lactate production (Fig. 2c). Histological staining shows strong expression of CAIX (Fig. 2d). These results suggest that this tumor model uses anaerobic glycolysis to overcome hypoxia. Buffer capacity appears sufficient such that produced lactate only mildly acidifies the extracellular tumor microenvironment, whereas under normoxic conditions, overexpression of CAIX might lead to stronger extracellular acidification6.

Conclusions

We demonstrated a multimodal imaging characterization of a patient-derived GBM model in mice regarding pH, metabolic pyruvate-to-lactate conversion, and hypoxia. Observed hypoxia occurred together with increased lactate-production and mild acidification. Therefore, the tumor microenvironment can be assessed with the described imaging modalities, suggesting their use as predictive imaging biomarkers in the context of GBM therapy.

Acknowledgement

We acknowledge help from Sybille Reder, Markus Mittelhäuser and Hannes Rolbieski for help with PET-Acquisitions, Michael Herz for PET-Tracer Synthesis and Marion Mielke, Olga Seelbach und Tanja Groll from pathology department (CeP) for help with histology. Further, we acknowledge support from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation – 391523415, SFB 824).

Disclosure

I or one of my co-authors have no financial interest or relationship to disclose regarding the subject matter of this presentation

References
[1] Anjum K, Shagufta B I, Abbas S Q, et al. Current status and future therapeutic perspectives of glioblastoma multiforme (GBM) therapy: A review. Biomed Pharmacother 2017; 92:681-689
[2] Duewel S, Hundshammer C, Gersch M, et al. Imaging of pH in vivo using hyperpolarized 13C‑labelled zymonic acid. Nature Commun 8, 15126 (2017)
[3] Hundshammer C, Duewel S, Koecher S, et al. Deuteration of Hyperpolarized 13C‐Labeled Zymonic Acid Enables Sensitivity‐Enhanced Dynamic MRI of pH. Chemphyschem. 2017; 18(18): 2422-2425
[4] Skinner J G, Topping G J, Heid I, et al. Fast 3D hyperpolarized 13C metabolic MRI at 7T using spectrally selective bSSFP. Digital Poster at ISMRM2020 International Conference 2020
[5] Hill D K, Orton M R, Mariotti E, et al. Model Free Approach to Kinetic Analysis of Real-Time Hyperpolarized 13C Magnetic Resonance Spectroscopy Data. PloS ONE 2013; 8(9): e71996
[6] Lee S-H, Griffiths J R, How and Why are Cancers Acidic? Carbonic Anhydrase IX and the Homeostatic Control of Tumour Extracellular pH. Cancers 2020, 12, 1616
Figure 1: Study Protocol and hyperpolarized 13C-MR(S)I

a: Imaging study protocol showing the temporal sequence and spacing of the applied modalities and injected tracers.

b: Mean pH map weighted by the signal intensity of each compartment in the corresponding voxel overlaid with anatomical image shows mildly acidified tumor regions. Tumor (white ROI) and a [1‑13C]lactate‑phantom (white arrow) were covered with carbomer gel for shim improvement.

c: Axial [1-13C]pyruvate intensity image overlaid with an anatomical image.

d: Axial [1-13C]lactate intensity image overlaid with an anatomical image shows high lactate production within the tumor (white ROI).
Figure 2: [18F]FMISO-PET, Correlation Plots and Histology

a: Axial image of [18F]FMISO-uptake overlaid with CT. Tumors (white ROI) show increased and heterogeneous uptake compared to healthy tissue (spine muscle).

b: Correlation plot of tumor mean pH with mean pyruvate-lactate AUC ratio. Linear regression shows no correlation (r = 0.02).

c: Correlation plot of tumor metabolic conversion rate AUC ratio and SUVmean for [18F]FMISO. Linear regression shows strong correlation (r = 0.84).

d: Immunohistochemical staining for CAIX. Tissue was sliced according to 13C-MR(S)I. Brown areas indicate positive staining.

Keywords: Hyperpolarized 13C-MRSI, Zymonic Acid, pH, Glioblastoma, [18F]FMISO
1033

Preclinical nanoparticle X-ray fluorescence computed tomography

Kian Shaker1, Giovanni M. Saladino1, Bertha Brodin1, Carmen Vogt1, Muhammet S. Toprak1, Marie Arsenian-Henriksson2, Hans M. Hertz1

1 KTH Royal Institute of Technology, Department of Applied Physics, Stockholm, Sweden
2 Karolinska Institutet, Department of Microbiology, Tumor and Cell Biology, Solna, Sweden

Introduction

In recent years, laboratory X-ray fluorescence (XRF) computed tomography (XFCT) using nanoparticles (NPs) as contrast agents has reached technical maturity, paving the way for imaging of molecular markers with higher spatial resolution than present techniques (e.g., PET, SPECT). Today the technique is mostly focused on preclinical imaging (i.e., imaging of small animals) and has been experimentally demonstrated in vivo on mice [1-3]. Here we present on the latest developments and results acquired from our laboratory XFCT arrangement capable of low-dose in vivo preclinical imaging.

Methods

Our laboratory XFCT arrangement consists of a liquid metal-jet X-ray source (D2, Excillum AB, Sweden) combined with multilayer optics providing an X-ray pencil-beam centered around 24 keV (cf. Fig. 1). Mice to be imaged are placed in the pencil-beam focus (100 µm FWHM) and scanned across the stationary pencil-beam using motorized stages. The energy of our pencil-beam allows suitable excitation of NPs injected in mice based on Mo, Ru or Rh (K-absorption edges around 20, 22 and 23 keV respectively). XRF generated from the NPs are detected with a 3-element silicon-drift detector (SDD), while transmitted X-rays are simultaneously detected using a separate SDD. By rotating and translating the mice, projection images are acquired at different angles to allow for XFCT and CT reconstruction.

Results/Discussion

Our imaging arrangement can be used to study long-term NP biodistribution dynamics in mice. Projection images are acquired at 200 µm scanning step sizes, which translate to the size of the pixels in the image. Here shown as an example (cf. Fig. 2), we injected in-house developed Mo NPs without any active-targeting ligand coating resulting in initial unspecific accumulation in major organs such as lungs, liver, and spleen. At longer timepoints, the XRF signal from these organs is significantly reduced which indicates long-term clearance of the NPs. Radiation dose for each of these 2D projection images was estimated using Monte Carlo simulations to ~1 mGy. While the unspecific accumulation of NPs in major organs is generally undesirable, we demonstrate that our laboratory XFCT arrangement can be used for in vivo whole-body localization of NP accumulations at few-100-µm-range spatial resolution.

Conclusions

We demonstrate that low-dose XFCT can be acquired longitudinally on mice with a laboratory arrangement. We also demonstrate multiplexed XFCT by showing that our arrangement can separate XRF signals from Mo, Ru, and Rh NPs simultaneously [4] which could serve as a path towards simultaneous imaging of different targets.

Acknowledgement

This study has been funded by the Wallenberg Foundation. We thank Kenth Andersson for assistance with animal handling.

Disclosure

I or one of my co-authors have the following financial interest or relationship(s) to disclose regarding the subject matter of this presentation:
Hans M. Hertz is a shareholder in Excillum AB.

References
[1] Zhang, S et al, 'Quantitative imaging of Gd nanoparticles in mice using benchtop cone-beam X-ray fluorescence computed tomography system', Int. J. Mol. Sci. 20, 2315 (2019)
[2] Jung, S et al, 'Dynamic in vivo X-ray fluorescence imaging of gold in living mice exposed to gold nanoparticles', IEEE Trans. Med. Im. 39, 526-533 (2019)
[3] Shaker, K et al, 'Longitudinal In-Vivo X-Ray Fluorescence Computed Tomography with Molybdenum Nanoparticles', IEEE Trans. Med. Im. 39, 3910-3919 (2020)
[4] Saladino, GM et al, 'Optical and X-Ray Fluorescent Nanoprobes for Dual Mode Bioimaging', ACS nano 15, 5077-5085 (2021)
Figure 1.

Experimental arrangement for laboratory in vivo preclinical XFCT.

Figure 2

Longitudinal imaging performed in vivo post-injection of Mo NPs. Recorded XRF photons per pixel (colour) overlayed on transmission image (grayscale). Acquisition time was 15 minutes for each combined image with 200 µm pixel size. Imaging timepoint post NP-injection denoted in the title. Percentage values represent the integrated whole-body XRF signal compared to the 1-hour timepoint and can be regarded as a measure of NP signal clearance over time.

Keywords: XRF, XFCT, microCT, Nanoparticles, X-ray Fluorescence