Heterogeneity of intracellular and extracellular tumor pH: A novel approach enabling simultaneous multiparametric tissue pH analysis

Abstract

Event Abstract Back to Event Heterogeneity of intracellular and extracellular tumor pH: A novel approach enabling simultaneous multiparametric tissue pH analysis Norbert W. Lutz1* and Patrick J. Cozzone1 1 Université Aix-Marseille, Faculté de Médecine, CRMBM, France Background and aim. Alterations in tissue pH play an important role in disease, notably in cancer. However, none of the methods currently available for measuring tissue pH in vivo provides quantitative parameters describing pH distributions in heterogeneous tissues. Statistical methods have long been used to determine parameters that characterize frequency distributions, e.g. based on histograms. This prompted us to analyze the shapes of pH-sensitive 31P NMR signals according to the same principle. Methods. First, we noninvasively acquired in vivo 31P NMR spectra of tumor xenografts in the thighs of nude mice (ras-transformed CCL39 hamster fibroblasts). Endogenous inorganic phosphate (Pi) served as an intracellular, and exogenous (i.p.-injected) 3-aminopropylphosphonate (3-APP) as an extracellular pH reporter molecule [1]. Second, for validation purposes 31P NMR spectra were acquired for phantoms with up to three compartments containing 3-APP solutions at different pH values. Third, for further validation numerical simulations were performed using an EXCEL spreadsheet. These calculations were based on Gaussian-shaped 31P NMR signals or pH distributions. Results. Heterogeneous tissue pH affected the lineshape of pH-sensitive 31P NMR signals. This was verified by analogous measurements using multiple-pH phantoms. In many tumors, pH distributions were asymmetric and/or composed of several distinct pH ranges. In these cases, the conventional method of determining one overall maximum of the pH curve yielded an unrepresentative tissue pH value, as a result of discarding information on tissue pH distribution contained in pH curves. Therefore, we exploited the entire pH distribution curve. The digital points under a pH curve represent a histogram-like graph. This "pH histogram" was used to calculate representative, weighted (wt.) pH values (wt. mean pH, wt. median pH) based on the corresponding classical equations. The pH histogram was also used to calculate quantitative measures of the asymmetry, flatness and evenness of tissue pH distributions (skewness, kurtosis and entropy, respectively). For pH distributions with distinct pH regions, the relative area of each region was calculated to estimate the proportion of the underlying tumor volumes. Conclusions / Discussion. Tissue pH distributions can be characterized by multiple statistical parameters based on pH-sensitive in vivo 31P NMR resonances. The underlying procedures can be applied to 3-APP signals for extracellular pH (pHe), and to Pi signals for intracellular pH (pHi) [2]. Acknowledgements Support from Centre National de la Recherche Scientifique (CNRS UMR 6612 and 7339) is greatfully acknowledged.