Direct Quantitative 13C-Filtered 1H Magnetic Resonance Imaging of Pegylated Biomacromolecules In Vivo

Rohan Alvares¹, Justin Lau^2,3, Peter Macdonald¹, Charles Cunningham^2,3, and R. Scott Prosser¹

¹Department of Chemistry, University of Toronto, Toronto, ON, Canada, ²Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada, ³Physical Sciences, Sunnybrook Research Institute, Toronto, ON, Canada

Synopsis

We demonstrate a new platform technology in which macromolecular constituents, such as proteins and drug delivery systems, are observed directly and quantitatively in vivo using ¹H MRI of ¹³C-labeled polyethylene glycol (¹³C-PEG) tags. The 28 kDa ¹³C-PEG tags are non-immunogenic, and each bears approximately 2500 spectroscopically equivalent ¹H nuclei appearing at a single resonance position. By filtering the ¹H PEG signal through the directly coupled ¹³C nuclei, background water and fat signals are largely eliminated. We demonstrate the approach by monitoring in real-time the distribution of ¹³C-PEG and ¹³C-pegylated albumin injected into the hind leg of a mouse.

Purpose

Direct in vivo detection and quantitation of specific molecules, metabolites or macromolecules by MRI methods at concentrations below the millimolar range is difficult.¹ One option is to tag the species of interest with a relaxation reagent which acts to modify the ¹H signal of water in its vicinity.^2–5 While sensitivity is enhanced in the presence of such contrast agents, detection is indirect and quantitation of the biomolecule of interest is complicated by the range of factors and conditions which influence the degree of contrast produced in the local water signal.⁶ In this study, we demonstrate the use of uniformly-labelled ¹³C-poly(ethylene glycol) (¹³C-PEG) tags to directly and quantitatively detect biomolecule biodistribution by ¹H MRI at concentrations in the low micromolar range.

Methods

¹³C PEG synthesis. ¹³C-labelled PEG (27.6 kDa) was synthesized by anionic ring opening polymerization of ¹³C-labelled ethylene oxide⁷ (Sigma Aldrich) using inert techniques.⁸ The labelled protein, ¹³C-PEG-BSA, was synthesized by converting the PEG alcohol end groups to aldehydes which were then conjugated to BSA amino terminal groups and lysine residues using reductive amination.⁹

Imaging. Experiments involving animals were performed in accordance with institutional animal care protocols. Imaging experiments were carried out using a Bruker BioSpec 70/30 USR 7 Tesla small animal system with a dual-tuned ¹H/¹³C volume mouse coil (RAPID MR International). Sensitivity limits were determined in 30% poly(acrylamide gel) phantoms with the ¹³C PEG concentrations varied between 0 – 500 µM. ¹³C-Filtered images were acquired using a spoiled gradient echo (FLASH) sequence¹⁰ modified with a slice-selective ¹³C filter (HFLASH sequence). Axial images were acquired with the following parameters: 8 × 8 cm FOV, 4 mm slice thickness, 64 × 64 acquisition matrix, 100 ms TR, 6.9 ms TE, 1 ms sinc ¹H 44° pulse, 1.37 ms Gaussian ¹³C 90° pulses, 1 dummy scans, and 10 – 200 transients, depending on ¹³C-PEG concentration. In vivo experiments were conducted by injecting ¹³C-PEG (4.1 mM; 50 µL; saline) and ¹³C-PEG-BSA (2.4 mM; 50 µL; saline) solutions into the hind legs of separate adult male BALB/cJ mice. ¹³C-filtered images were acquired similar to the phantom experiments above except with a 32 × 32 acquisition matrix, 50 averaged transients, and 10 dummy scans.

Results

Figure 1 illustrates the general approach developed to demonstrate the feasibility of direct ¹H MRI of ¹³C-pegylated macromolecules. The ¹³C-filtering step is based on the heteronuclear multiple-quantum coherence (HMQC) technique, as illustrated schematically in Figure 2, while the scheme for implementing the HMQC filter within a ¹H MRI sequence is illustrated in Figure 3. To demonstrate the potential of in vivo ¹³C-filtered direct ¹H MRI of ¹³C-PEG, images were obtained from mice after intravenous injections of either free ¹³C-PEG (Figure 4 a-c) or ¹³C-PEG-BSA (Figure 4d). Figure 4a-c compares anatomical (water), filtered (¹³C-PEG), and co-registered (superimposed filtered and anatomical images) obtained following administration of 27.6 kDa ¹³C-PEG into the right leg. Unlabelled PEG of similar size and concentration was injected simultaneously into the left leg. Detection limits were established using known concentrations of 27.6 kDa ¹³C-PEG in poly(acrylamide) gel phantoms as shown in Figure 5. High concentrations of ¹³C-PEG (500 µM, 3 nmol in a 6.25 mm³ imaging volume) yielded images with visual quality similar to standard water gradient echo images. Concentrations as low as 25 µM (160 pmol in a 6.25 mm³ imaging volume) were detectable with a contrast/noise ratio (CNR) of 6.0 in ~ 21 minutes.

Discussion

We have demonstrated here a new ¹H MRI approach to detect and quantify biomacromolecular constituents in vivo based on using ¹³C-PEG tags. Pegylation is now the norm in antibody-drug conjugates, therapeutic proteins, antibodies, aptamers, and many drug delivery systems, because it is non-immunogenic, tends to prolong retention of the macromolecule of interest in the desired tissue, confers stability, and helps avoid renal filtration, opsonisation and consequent phagocytic clearance.^11–13 PEG itself also represents an ideal candidate for detection via ¹H MRI. With four spectroscopically equivalent protons per ethylene oxide monomer unit, a single PEG tag has the potential to yield a signal amplification of 4N where N is the degree of polymerization: for a 28 kDa PEG tag, N≈640. Furthermore, virtually all the signal from the PEG tags appears under a single, narrow ¹H resonance line, and its resonance appears at a ¹H chemical shift distinct from that of water or fat.

Conclusion

The current study demonstrates that in vivo pegylated protein concentrations may be quantitatively studied by ¹H MRI and that biodistribution can be monitored in real time over the course of many hours.

Acknowledgements

R. S. Prosser and P. M. Macdonald acknowledge NSERC Canada for Discovery Awards.

References

1. Cassidy, P. J. & Radda, G. K. J. R. Soc. Interface 2, 133–144 (2005).

2. Louie, A. Y. et al.Nat. Biotechnol. 18, 321–325 (2000).

3. Weissleder, R. et al. Nat. Med. 6, 351–354 (2000).

4. Genove, G., DeMarco, U., Xu, H., Goins, W. F. & Ahrens, E. T. Nat Med. 11, 450-454 (2005).

5. Zhao, M., Beauregard, D. A., Loizou, L., Davletov, B. & Brindle, K. M. Nat. Med. 7, 1241–1244 (2001).

6. Caravan, P. Chem. Soc. Rev. 35, 512–523 (2006).

7. Price, C. C. & Carmelite, D. D. J. Am. Chem. Soc. 88, 4039–4044 (1966).

8. Ndoni, S., Papadakis, C. M., Bates, F. S. & Almdal, K. Rev. Sci. Instrum. 66, 1090 (1995).

9. Borch, R. F., Bernstein, M. D. & Durst, H. D. J. Am. Chem. Soc. 93, 2897–2904 (1971).

10. Haase, A., Frahm, J., Matthaei, D., Hanicke, W. & Merboldt, K.-D. J. Magn. Reson. 67, 258–266 (1986).

11. Chan, A. C. & Carter, P. J. Nat. Rev. Immunol. 10, 301–316 (2010).

12. Ginn, C., Khalili, H., Lever, R. & Brocchini, S. Future Med. Chem. 6, 1829–1846 (2014).

13. Veronese, P. F. M. & Mero, A. BioDrugs 22, 315–329 (2008).

Figures

¹³C filtered ¹H imaging of ¹³C-pegylated molecules using MRI. An MRI active probe, ¹³C-labeled poly(ethylene glycol) (¹³C-PEG), is synthesized and then conjugated to a protein, such as BSA, or other macromolecule, thereby rendering it MRI visible. Following intramuscular injection into the hind leg, separate ¹³C-filtered and standard ¹H images yield molecular and anatomical information, respectively, where superposition of the two gives the distribution of the molecule within the animal.

Comparison of the standard ¹H and the ¹³C-filtered pulse sequences used to eliminate background proton signals. In the ¹H experiment, both water (blue) and ¹³C-PEG (orange) protons are excited and detected. With the HMQC filter, proton magnetization is passed to carbon and back again. During this process, the phase of the non-¹³C coupled protons can be inverted, or not, relative to those that are coupled. This allows the non-coupled (background) signal to be cancelled out in a two scan process.

¹³C-filtered gradient echo ¹H MRI pulse sequence. The standard slice selection elements are shown in blue. In the ¹³C filtered version, a ¹³C filter is added, shown in red, consisting of two frequency selective 90° ¹³C pulses. The 180° ¹H inversion pulse in the standard HMQC sequence may be eliminated. The spatial encoding gradient elements that follow the rf pulses are identical to a standard gradient echo sequence.

(a) Standard ¹H gradient echo axial image of the mouse leg. ¹³C-PEG (4.1 mM) or ¹³C-PEG-BSA (2.4 mM) was injected into the left leg of the mouse, while unlabeled PEG was injected into the right leg. (b) ¹³C filtered ¹H gradient echo axial image of the mouse leg after the same ¹³C-PEG injection. Only the ¹³C-PEG is visible in this filtered image. (c) Superimposed (co-registered) anatomical and ¹³C-PEG filtered images. (d) Superimposed (co-registered) anatomical and ¹³C-PEG-BSA filtered images.

Quantitative ¹H MRI of ¹³C-PEG using a ¹³C filtered ¹H gradient echo sequence. Phantoms were constructed consisting of 30% poly(acrylamide) gels formed in Eppendorf tubes into which various concentrations of ¹³C-PEG were dispersed. ¹H images transverse to the cylindrical long axis, having a voxel size of 1.25 x 1.25 x 4 mm³, were acquired via the ¹³C filtered ¹H gradient echo sequence. Representative images acquired in this fashion are shown on the right.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)

1022