Introduction

AcidoCEST MRI is a promising non-invasive technique to directly measure extracellular pH in vivo by virtue of pH-dependent changes in the CEST signal from amide protons in injectable agents like iopamidol¹. This technique has a potential application in oncology to assess tumor pH considering that the Warburg effect can lead to tumor acidification². Therefore, the ability to measure tumor pH can allow for a better understanding of the microenvironment and unique conditions in which therapeutics can be more effective. Currently, there are only a handful of studies evaluating acidoCEST MRI in clinical settings and animal models of cancer. A major limitation is the limited demonstration of robust acidoCEST signal in vivo, the minimum agent concentration required, and paired CT data to confirm concentration levels. In this study, our goal was to experimentally determine the settings required for successful and robust in vivo acidoCEST MRI and to measure pH in different tumor xenograft models.

Methods

All animals were dosed and monitored according to the guidelines from the IACUC at Genentech, Inc.
Micro-CT
Micro-CT (MILabs U-CT^UHR) with an injection of isovue-370 (Bracco Diagnostics) was used to determine the local concentration of iopamidol in the tumor, using two different injection protocols: 1) intratumoral injection of 50µl; 2) intravenous bolus of a 300µl followed by infusion of 200µl over 30 minutes. Micro-CT included 1 whole body pre-injection scan, and 8 post-injection scans 4-minutes apart to capture the evolution of iopamidol concentration over a similar duration of the acidoCEST MRI acquisition.
MRI
MRI was done on a Bruker BioSpec 7T with a cryogenically-cooled receive-only surface array coil, and an 86-mm transmit volume coil. AcidoCEST MRI employed the parameters in Figure 5. Pixelwise pH was calculated by fitting the CEST z-spectra using Bloch McConnel equations modified to substitute pH for chemical exchange rates as a fitting parameter³.
In vitro acidoCEST MRI was performed on 16 iopamidol solutions of pH 6.1-7.5 (“pH phantom”) and compared with measurements from a calibrated pH meter. To estimate the minimum required concentration, a phantom comprised of 7 different concentrations of iopamidol (“concentration phantom”) was used.
In vivo acidoCEST MRI was conducted with intratumoral injection in 5 tumor xenograft models (KP4, KP4-CA9, KPL-4, MMTV-HER2.tg, KPR3070), and intravenous infusion in 2 models with highly vascularized tumors (SW480, 786-O). The same injection protocol as with micro-CT was used.

Results and Discussion

Micro-CT
Figure 1A shows micro-CT images post-injection of isovue-370 via intratumoral or intravenous delivery, and Figure 1B shows histograms of the iopamidol concentration within the tumors. Intratumoral injection yielded a concentration of 125.3±110.2mM, whereas intravenous delivery yielded 20.6±7.0mM and 22.1± 5.7mM in the two models. Figure 1C shows the temporal evolution of the iopamidol concentration. The average concentration after intratumoral delivery was initially ~200mM and decreased to ~100mM at the final time point. The large error bars indicate a highly heterogenous distribution. With intravenous delivery, the highest concentration was only ~25mM.
MRI
Figure 2A shows the acidoCEST MRI pH map in the pH phantom. MRI pH values were highly correlated with pH meter measurements (Figure 2B, r=0.98), but tended to be lower with a mean difference of –0.14. Therefore, the equation of the linear regression line was used to re-calibrate pH.
Figure 2C shows CEST z-spectra in the concentration phantom. The CEST resonances were only visible at concentrations >=10mM. This limit would be higher in vivo due to biological noise, susceptibility-induced spectral broadening, and heterogeneous uptake of the agent by the tumor.
Figure 3 shows in vivo acidoCEST MRI pH measurements and representative pH maps in different tumor xenografts after intratumoral delivery of isovue-370. The KPR3070 model exhibited the lowest pH, which could hypothetically be attributed to elevated glycolytic rate. Another notable finding was that the coefficient of variation of the pH in the KP4-CA9 model (overexpressed Carbonic Anhydrase IX) was the lowest among all tumors.
Figure 4 shows representative CEST z-spectra in a 100mM iopamidol solution (A), and in tumors with isovue-370 delivered intratumorally (B) and intravenously (C). Whereas the CEST resonances are clearly visible in the solution and intratumoral examples, they are inconspicuous in the intravenous example. This implies that even at our large intravenous infusion volume of 500µl, the average tumor concentration of ~20mM is most likely insufficient for in vivo acidoCEST MRI. In fact, only 50% of the animals imaged with intravenous delivery had >50% pixels in the tumor with sufficient signal for pH analysis.

Conclusion

We have demonstrated that acidoCEST MRI can be a robust technique to measure extracellular pH in vivo when delivered intratumorally. Moreover, we demonstrated that tumors can exhibit acidosis when considering the dynamic range of pH values in relation to different tumor xenografts and overexpression of CA9. However, a significant limitation of the technique is the low tumor concentration of iopamidol when delivered intravenously, which limits applications in orthotopic and metastases models where intratumoral injections are not practical. For clinical MRI studies, these high concentrations and doses might be very difficult to achieve. Under ideal conditions and highly vascularized tumors, pH measurements with acidoCEST MRI might be attainable. CT imaging could mitigate some risks by first understanding the concentrations and tumor heterogeneity prior to acidoCEST MRI.

Acknowledgements

No acknowledgement found.