High Resolution pH e Imaging of Rat Glioma Using pH-Dependent Relaxivity Maria L. Garcia-Martin, 1 Gary V. Martinez, 1 Natarajan Raghunand, 1 A. Dean Sherry, 3 Shanrong Zhang, 3 and Robert J. Gillies 1 * Previous studies using MR spectroscopy have shown that the extracellular pH (pH e ) of tumors is acidic compared to normal tissues. This has a number of important sequelae that favor the emergence of more aggressive and therapy-resistant tumors. New MRI methods based on pH-sensitive T 1 relaxivity are an attractive alternative to previous spectroscopic methods, as they allow improvements in spatial and temporal resolution. Recently, pH-dependent GdDOTA-4AmP 5- and a pH-indepen- dent analog, GdDOTP 5- , were used to image renal pH in mice. The current study has used a similar approach to image pH e in rat gliomas. Significant differences were observed compared to the renal study. First, the relaxivity of GdDOTP 5- was found to be affected by the higher extracellular protein content of tu- mors. Second, the pixel-by-pixel analysis of the GdDOTP 5- and GdDOTA-4AmP 5- pharmacokinetics showed significant disper- sion, likely due to the temporal fluctuations in tumor perfusion. However, there was a robust correlation between the maximal enhancements produced by the two boluses. Therefore, to ac- count for the local time-courses differences, pH e maps were calculated at the time of maximal enhancement in each pixel. Finally, the comparison of the pH e and the time to maximal intensity maps revealed an inverse relationship between pH e and tumor perfusion. Magn Reson Med 55:309 –315, 2006. © 2006 Wiley-Liss, Inc. Key words: acid-base; MRI; relaxivity; gadolinium; tumor, pH Over the past decades, non-invasive imaging techniques have been developed to characterize the metabolic and phys- iologic environments of living tissues. An important appli- cation to emerge in this area has been the measurement of tumor pH using magnetic resonance. This application began early on with the acquisition of 31 P MR spectra, which pro- vided information not only on the in vivo nucleoside triphos- phate (NTP) levels, but also inorganic phosphate, P i , whose chemical shift is pH dependent (1,2). A neutral-to-alkaline tumor pH calculated from the chemical shift of P i has been reported in numerous 31 P MRS studies of human and animal tumors (i.e., 7.0 –7.4). This was discrepant with previous microelectrode reports of an acid pH in tumors, and was not resolved until it was shown that P i reported primarily the intracellular pH (3). The extracellular pH (pH e ) can be inter- rogated using an exogenous 31 P NMR reporter, 3-amino pro- pylphosphonate, 3-APP (4). Studies with 3-APP have con- firmed that the average pH e of tumors is acidic, with values reaching as low as 6.0 (5). However, 31 P investigations are limited in both spatial and temporal resolution. Even at high field, voxel size is limited to ca. 4  4  4 mm 3 , with a temporal resolution on the order of 5 min. An improvement in spatial resolu- tion has been achieved using 1 H NMR and 2-imidazol-1- yl-3-ethoxycarbonyl-propionate, IEPA, which has a hydro- gen attached to C-2 of the imidazole ring whose chemical shift is pH-sensitive (6). This compound is non-toxic and is membrane impermeant, and thus restricted to the extra- cellular compartment. Using magnetic resonance spectro- scopic imaging (MRSI), IEPA has been used to measure pH e in breast cancer tumors (7,8) and gliomas (9) with spatial resolution up to 1  1  1 mm 3 . These studies showed, for the first time, that the pH e in tumors was heterogeneous, with differences in pH e of as much as 0.5 pH unit across 8 mm in distance. The relationship between the low pH e and perfusion has been inferred (see refs. (5,8), yet cannot be directly tested by this method due to differences in the spatial resolution between dynamic contrast MR images and MRSI. Although the use of IEPA has led to an increase in the understanding of tumor pH e regulation, it still suffers from relatively poor spatial and temporal resolution. In recent years, methods have been developed that allow pH to be rapidly determined at spatial resolutions compa- rable to standard MRI. These methods fall into two broad categories: magnetization transfer and relaxation en- hanced contrast (10). Magnetization transfer can be achieved with either endogenous (11) or exogenous (12– 14) hydrogen donors, and has great potential, as the con- trast can be turned on or off through on- and off-resonance irradiation. Relaxation enhanced pH measurements in- volve the use of gadolinium-based contrast reagents (Gd- CR) whose relaxivity is pH dependent. A number of such compounds have been developed (15,16) where the relax- ation enhancement is primarily achieved through in- creased hydrogen exchange rates at the lanthanide center. This approach was first used by Brindle’s group (17), who realized that the ability to quantitatively determine pH e with relaxivity required accurate knowledge of the Gd-CR concentration in each voxel. More recently, this approach has been used to measure the pH e in kidneys using Gd- DOTA-4AmP 5- (18,19). In that study, the concentration of GdDOTA-4AmP 5- at a given time post-injection was as- 1 Department of Biochemistry and Molecular Biophysics, Arizona Cancer Cen- ter, The University of Arizona, Tucson, AZ, USA. 2 Department of Chemistry, The University of Texas at Dallas, Richardson, TX, USA. 3 The Rogers Magnetic Resonance Center, Department of Radiology, Univer- sity of Texas Southwestern Medical Center, Dallas, TX, USA. Grant Sponsor: NCI; Grant Numbers: R01 Ca 077575 and R24 088578 to RJG. Grant Sponsor: Robert A. Welch Foundation; Grant Number: AT-584. Grant Sponsor: National Institutes of Health; Grant Number: CA-84697. Grant Sponsor: Division of Research Resources and the National Institutes of Health; Grant Number: RR-02584 to ADS. *Correspondence to: Robert J. Gillies, Department of Biochemistry and Mo- lecular Biophysics, Arizona Cancer Center; 1515 N. Campbell Ave., Tucson, AZ, USA 85724-5024. E-mail: rgillies@email.arizona.edu Received 11 July 2005; revised 3 October 2005; accepted 14 October 2005. DOI 10.1002/mrm.20773 Published online 9 January 2006 in Wiley InterScience (www.interscience. wiley.com). Magnetic Resonance in Medicine 55:309 –315 (2006) © 2006 Wiley-Liss, Inc. 309