Abstract

There are tantalizing pieces of evidence suggesting that the function of myocardial tissue post-infarction may be restored with the use of tissue engineering therapies. An injectable hydrogel (UPy-gel) based on poly(ethylene glycol) modified with fourfold hydrogen bonding ureido-pyrimidinone (UPy) moieties (UPy-PEG-UPy hydrogelator, Fig.1) was proposed as a new delivery system for the therapeutic molecules and cells[1]. UPy-gel is a pH-switchable hydrogel, i.e. at a pH≥8.5 it is in a liquid state and can be injected through a catheter. At a pH close to 7, e.g. in contact with tissue, it forms a gel. While pH-switchable properties have been confirmed in a test tube, it remained to be seen if gel forming properties and retention of the biomaterial takes place upon injection in myocardial tissue. In this study we developed a non-invasive MRI protocol not only capable of imaging hydrogel biomaterials but also their pH-switchable sol-gel behavior.

Methods

All MRI experiments were done on a 9.4T Varian MR-platform and data processing was done using in-house developed scripts.

Sample preparation

Liquid biomaterials (pH≈9) were prepared by dissolving UPy-gel in PBS to yield the following concentrations: 5% (w/w, UPy5liquid), 7.5% (w/w, UPy7.5liquid), and 10% (w/w, UPy10liquid). A drop of 10µL 1M HCl was added to a total volume of 1ml liquid UPy-gel to induce a sol-gel transition resulting in the corresponding gelated biomaterials: 5% (w/w, UPy5gel), 7.5% (w/w, UPy7.5gel), and 10% (w/w, UPy10gel).

In vitro MRI experiments

For Z-spectra acquisition in vitro, the MT saturation module was composed of 100 gaussian pulses (20ms each) followed by a final spoiler. A single-shot spiral readout was used with a center-out k-space trajectory with the following parameters: FOV 40x40mm², matrix size 128x128, 1mm slice thickness, and a post-readout T1-recovery of 20s. Z-spectra were acquired by applying the saturation module (with varying power levels) at 137 frequency offsets.

Ex vivo MR experiments

Adult female rats were terminated with an overdose of isoflurane and hearts were exposed by thoracotomy. UPy10liquid (50µL or 100µL) was injected into the myocardial tissue (ventricular wall) of the still beating heart. The hearts were flushed with a perfluorocarbon liquid and placed in a sample holder for scanning.Ex vivo MRI protocols for Z-spectra acquisition were similar to those for in vitro experiments with a slight change in the readout parameters: FOV 19.2x19.2mm² (matrix size 128x128 with NSA=5 and matrix size 192x192 with NSA=10 for an injection volume of 100µL or 50µL, respectively), 0.4mm slice thickness, and a post-readout T1-recovery of 20s. Z-spectra were acquired by applying the saturation module (24dB power) at 137 frequency offsets.

Data analysis

B0-correction of Z-spectra was done pixel-wise by searching for the minimum of spline interpolated Z-spectra and shifting the minimum accordingly. The magnetization transfer (MT) effect was quantified as an amplitude of the Lorentzian function fitted to Z-spectra. Pixel-wise MT fitting was used to generate MT maps.To distinguish two different states of the UPy10 biomaterial, i.e. liquid versus gelated, an artificial neural network (NN) was trained on the Z-spectra from in vitro (Z-spectra for UPy10liquid and UPy10gel) and ex vivo (Z-spectra for the myocardial tissue) experiments. The different states or the outputs in NN, namely UPy10liquid, UPy10gel and myocardial tissue are color-coded in Fig. 3 (a and d)

Results and Discussion

The UPy-gel generated a detectable MT effect in vitro (Fig.2). The MT effect was higher in the gelated state when compared with the corresponding liquid biomaterial of the same concentration. Also, the MT effect scaled with the power of the saturation module. A power level as low as 15dB was sufficient to distinguish two different states of the UPy10 biomaterials: liquid (UPy10liquid) versus gelated (UPy10gel). For comparison, we also performed imaging with more traditional MRI contrast, i.e. quantitative T1, T2-weighted, and diffusion-weighted MRI but none of those was able to differentiate between the liquid and gelated states of the biomaterials.To test the pH-switchable behavior of the UPy-gel, two experiments were performed where UPy10liquid (100µL and 50µL, respectively) was injected in the myocardial tissue of a beating rat heart (Fig.3, a and d). The retention of UPy-gel at the injection site was confirmed with MT-MRI (Fig.3, b and e). A liquid core and a gelated rim were identified at the injection site of 100µL of UPy10liquid (Fig.3c), whereas at the injection site of 50µL of UPy10liquid, only the gelated state was identified. The liquid core at the injection site with the larger UPy10liquid volume (100µL) is likely a result of hindered diffusion. The calibrated NN was able to correctly identify the control (UPy10liquid) as the liquid biomaterial (Fig.3 d and f).

Conclusions

We developed and validated MT-MRI protocols for imaging of hydrogel biomaterials and their pH-switchable behavior.

Acknowledgements

No acknowledgement found.

References

[1] Bastings, M. M. C. et al. A fast pH-switchable and self-healing supramolecular hydrogel carrier for guided, local catheter injection in the infarcted myocardium. Adv Healthc Mater 3, 70–78 (2014).