Magnetic resonance imaging of sugar moieties and PH in metabolic dysfunction

Abstract

Provided herein is a noninvasive method for measuring pH, glucose and glycogen levels using a single acquisition method, as well as systems for performing the method. This method for glycogen / glucose (and pH) detection has high specificity enabled by superior chemical shift separation in high-strength magnetic fields. The methods have applications in detecting changing tumor microenvironments in cancerous cells for measuring important metrics for noninvasive and quantitative metabolic profiling.

Description

[0001] Attorney Docket No.: 047162-7396WO1 (02764)

[0002] TITLE OF THE INVENTION

[0003] Magnetic Resonance Imaging of Sugar Moieties and pH in Metabolic Dysfunction

[0004] CROSS-REFERENCE TO RELATED APPLICATION

[0005] This application claims priority to U.S. Provisional Patent Application No. 63 / 735,438 entitled "MAGNETIC RESONANCE IMAGING OF SUGAR MOIETIES AND PH IN METABOLIC DYSFUNCTION," filed December 18, 2024, the disclosure of which is incorporated herein by reference in its entirety.

[0006] STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH

[0007] This invention was made with government support under MH067528 and EB023366 awarded by National Institutes of Health. The government has certain rights in the invention.

[0008] BACKGROUND

[0009] Unlike healthy cells, cancer cells produce only two (2) adenosine triphosphate (ATP) molecules per glucose, mainly through aerobic glycolysis, instead of the about 36 ATP molecules generated from oxidative metabolism in healthy cells. This behavior leads to noticeably more glucose uptake by cancer cells to support their survival. The abnormal dependence on aerobic glycolysis is exploited to detect tumors in vivo by positron emission tomography (PET), magnetic resonance imaging (MRI), or magnetic resonance spectroscopy (MRS).

[0010] A variety of MRS methods can measure glucose in vivo, including proton (1H ) MRS.. A disadvantage ofXH-MRS for glucose detection is signal-to-noise ratio (SNR), especially in conditions where the signal is too small to be reliably distinguished from noise. Further progress into studying glucose metabolism came with the use of deuterium (2H) labeled glucose and nitrogen-15 (15N) enriched ammonia allowing measurements of intermediary metabolites of glucose breakdown. A major advance in measuring glucose metabolism uses carbon-13 (13C) labeled glucose infusion while simultaneously collecting13C spectra. However, the13C-MRS method requires sufficient glucose enrichment in blood and brain for detection with adequate SNR.13C-MRS has superior spectral resolution compared to 'H- MRS, but it has lower SNR compared to 'H-MRS

[0011] - 1 -

[0012] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764)

[0013] With the explosive growth of genomic and proteomic technologies, it has become clear that acquisition of molecular information could help better understand the pathological changes of disease.

[0014] BRIEF SUMMARY OF THE INVENTION

[0015] In one aspect, a method for magnetic resonance imaging is provided. In certain embodiments, the method includes exposing a sample to a saturating radiofrequency (RF) pulse with a frequency that is off-resonance with bulk water and a corresponding Bo field. In certain embodiments, the method includes exposing the sample to at least one sampling RF pulse having an amplitude (Bi) of about 0.5 to about 3 pT. In certain embodiments, the method includes exposing the sample to one or more RF pulses to obtain chemical exchange saturation transfer (CEST) data. In certain embodiments, the method includes generating one or more images from the CEST data at a plurality of frequency offsets.

[0016] Also provided are systems and storage media for performing the methods of magnetic resonance imaging described herein. Advantageously, these systems and methods simultaneously enhance sensitivity of exchanging hydroxyl protons from both glycogen and glucose at high magnetic field strength. Additionally, the RF power levels used herein minimize non-specific magnetization transfer effects arising from protons in semisolid tissues, while at the same time revealing mobile proton components.

[0017] BRIEF DESCRIPTION OF THE FIGURES

[0018] The drawings illustrate generally, by way of example, but not by way of limitation, various embodiments of the present application.

[0019] FIGs. 1 A-1D show the effect of concentration and Bi amplitude on Z-spectra for glycogen and glucose. (FIG. 1 A) Z-spectra of glycogen bovine liver at Bi = 3 pT shows one broad peak at +1.2 ppm, whose intensity decreases with decreasing concentration of glycogen. (FIG. IB) Z-spectra of gly cogen bovine liver at Bi = 0.5 pT shows that the detection sensitivity of the other -OH groups from glycogen are heightened. At this lower Bi amplitude, the CEST peak is visible at +0.7 ppm and is well separated from the peak at +1 .2 ppm. The peak at -1.2 ppm corresponds to aliphatic protons [NOE peaks from glycogen protons (H3), (H5), and (H2) + (H4-1)] and shows a linear dependence on glycogen concentration (FIG. IB). Similar Z-spectra obtained from glucose samples at Bi amplitudes of (FIG. 1C) 3 pT and (FIG. ID) 0.5 pT indicate that glucose can also be detected by CEST

[0020] - 2 -

[0021] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) with good sensitivity at lower Bi (0.5 pT). Similarly, to glycogen (FIG. 1A and FIG. IB), at higher Bi amplitude a broad peak is present at +1.2 ppm for glucose. At Bi = 0.5 pT two distinct CEST peaks are visible at +1.2 and +2.1 ppm. All experiments were obtained at 11.7T from samples at pH=7.0 and ~30 °C. See FIGs. 8A-8B forXH spectra of glycogen and glucose at 11.7T.

[0022] FIG. 2 shows Z-spectra of mixture samples containing glycogen and glucose at various Bi amplitudes. Results at different Bi amplitudes are shown different shades of blue lines. In the mixture of equimolar concentration of glycogen (5 mM) and glucose (5 mM), the effect of higher Bi amplitude is similar to that observed in individual samples of glycogen and glucose (FIG. 1). All peaks for glycogen (arrows at +0.7 ppm and -1.2 ppm) and glucose (arrow at +2. 1 ppm) are also visible in the mixture at Bi = 0.5pT (dark line). The center arrow at +1.2 ppm shows overlapping peaks from glucose and glycogen. Lower intensity of CEST peaks are visible at Bi = 1 pT (dark line). At Bi = 3 pT, the peak at +1.2 ppm from glycogen and glucose becomes broad and overlap with the peak at +0.7 ppm from glycogen, while the glucose peak at +2. 1 ppm and glycogen peak at -1.2 ppm are not visible. The glycogen peak upfield from water (-1.2 ppm) is due to nuclear Overhauser effect (NOE) attributed to the aliphatic protons. See FIG. 8 for 'H spectra of glycogen and glucose mixture at 11.7T.

[0023] FIGs. 3A-3B show Z-spectra of glycogen and glucose mixtures at different concentrations. (FIG. 3A) To demonstrate reproducibility and sensitivity of the method, mixture samples containing different concentration of glycogen and glucose at pH 7. 12 were prepared and CEST imaging was conducted (see Material and Methods). (FIG. 3B) Z-spectra of individual samples with different concentrations of glycogen and glucose. The CEST intensity at +0.7, +1.2 and -1.2 ppm from glycogen and +1.2 and +2. 1 ppm from glucose, in the mixture have the same increasing trend with increasing concentrations without losing specificity. At higher concentration of glycogen / glucose, the common peak at +1.2 ppm starts to overlap with the peak at +0.7 ppm because of the broadening effect, which is more dominant at higher concentrations. A concentration of 1 mM for glycogen / glucose is not sufficient to generate visible CEST peaks under these experimental conditions. The experiments were conducted with Bi of 0.5 pT at 11.7T. See FIGs. 10A-10E for water saturation shift referencing (WASSR) method used to provide Bo field correction in each voxel.

[0024] FIGs. 4A-4H show in vitro validation of GGCID for different mixtures of glycogen and glucose. Selectively generated CEST maps from mixture samples (see FIG. 3A for

[0025] - 3 -

[0026] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) details of each tube) for (FIG. 4A) glycogen with saturation at +0.7 ppm, (FIG. 4B) glucose with saturation at +2. 1 ppm, and (FIG. 4C) common CEST peak for both molecules with saturation at +1.2 ppm. (FIG. 4D) Glycogen CEST at +0.7 ppm (black dots and line) and glucose CEST at +2. 1 ppm (red dots and line) show' a linear dependence with concentration of each metabolite. The CEST sensitivity7(slope) for glycogen (black line, R2=0.96) is 30% smaller than for glucose (red line, R2=0.99). The CEST contrast range for glycogen (2.5-10 mM) was wider than for glucose (2.5-5 mM). pH maps were calculated from ratiometric CEST using a combination of tw'O CEST peaks: (FIG. 4E) +0.7 and +2.1 ppm, (FIG. 4F) +0.7 and +1.2 ppm, and (FIG. 4G) and +2.1 and +1.2 ppm. While the pH data for E and F are quite homogenous in each tube, the pH data for G is more heterogeneous. (FIG. 4H) Comparison of pH measured with the pH electrode (black dashed line) with calculated pH by GGCID for +0.7ppm / +2. 1 ppm (black bars) and for +0.7 ppm / +1.2 ppm (red bars). For samples where glycogen / glucose concentration is 1 mM, the calculated pH by GGCID is much less accurate, being outside the pH range 6-8 used for calibration of GGCID. See FIGs. 12A-12E for the ratiometric CEST maps with glycogen NOE peak at -1.2 ppm.

[0027] FIGs. 5A-5B show representative maps of glycogen and glucose concentrations and pH maps from GGCID in human derived U87 and U251 GBMs. The left column show s the tumor location (black outline inside the brain) on T2-weighted MR images for (FIG. 5A) U87 and (FIG. 5B) U251 tumors. The other three columns show maps for glycogen, glucose and pH. In all tumors glycogen and glucose concentrations are higher and pH is lower inside tumor, compared to normal (contralateral) tissue. The higher glycogen CEST signal can only be attributed to glycogen increase in tumor cells, but the higher glucose CEST signal can be due to higher blood glucose within tumor blood vessels also. See FIGs. 13A-13B for a more detailed analysis of glycogen, glucose, and pH inside and outside the tumors.

[0028] FIGs. 6A-6E show histochemical confirmation of brain glycogen with periodic acid Schiff (PAS) and dimedone periodic acid Schiff (D-PAS) staining in U87 tumors. Comparison of CEST-derived glycogen using Bi = 0.5 pT with (FIG. 6A) saturation at +0.7 ppm and (FIG. 6B) saturation at -1.2 ppm (or NOE*) shows (FIG. 6C) similar levels of glycogen inside the tumor. However, beyond the tumor boundary, the NOE* signal is widespread, suggesting less specificity7of this CEST peak for glycogen only. (FIG. 6D) PAS staining showing heterogeneous glycogen distribution through the whole brain. (FIG. 6E) The same slice section stained for pretreatment with dimedone to digest glycogen yielded background staining from non-glycogen sources (D-PAS). Selected tumor area (black square) examined by *20 magnification shows accumulation of glycogen inside tumor (arrows in

[0029] - 4 -

[0030] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764)

[0031] FIG. 6D) as deep magenta color. The same part of tumor on the section treated with dimedone (FIG. 6E) show significantly reduced magenta color, which supports the presence of glycogen inside the tumor. See FIGs. 14A-14D for detailed analysis of staining data.

[0032] FIGs. 7A-7E show a comparison between CEST-derived ratiometric pH based on glycogen / glucose (GGCID) and amide / amine (AACID) in U87 tumor. Comparison of pH maps measured using ratiometric CEST with (FIG. 7A) GGCID with Bi = 0.5 LIT and (FIG. 7B) AACID with Bi = 1.5 pT. Z-spectra from tumor defined ROI (black contour) with main peaks identified from (FIG. 7C) GGCID and (FIG. 7D) AACID experiments. The average pH measured with GGCID or AACID (FIG. 7E) shows similar values inside (black bars) and outside (white bars) the tumor.

[0033] FIGs. 8A-8B show chemical structures and 'H spectra of glycogen and glucose. (FIG. 8A) Representative glycogen structure and1H spectra of five glucose monomer subunits in a (1-4) and a (1-6) linkage (arrows) with proton numbers following assignments by Zang and co-workers. The bovine liver gly cogen sample was prepared in 70% H2O and 30 % D2O, 15 mM concentration with 2% NaN3 and adjusted to pH=7 by adding NaOH / HCl. The Hl proton with a 1-4 linkage gives rise to the downfield resonance at +0.7 ppm from water (set to 0 ppm). The other composite peaks correspond to H3 (-0.7 ppm), H5 (-0.85 ppm) and H2+H4-1 (-1.1 ppm). The H4 protons gives rise also to an upfield resonance at -1.4 ppm. (FIG. 8B) Structure and 'H spectra of D-glucopyranose in the form of pyranose ring with a anomeric conformation. Glucose sample was prepared similarly to the glycogen sample in 70% H2O and 30 % D2O, 15mM concentration with 2% NaN? and adjusted pH=7 by adding NaOH / HCl. The H6 has a chemical shift upfield from water resonance at -0.35 ppm. The composite H2+H3+H4+H5 peaks have chemical shifts spreading over a wider range from 0.9 to 1.8 ppm downfield from water. The Hl proton resonance for a anomeric conformation gives rise to the +2. Ippm resonance. All spectra were obtained at 1 1.7T, pH = 7.0 and ~30 °C.

[0034] FIG. 9 is aXH spectrum of a mixture of 5 mM glycogen and 5 mM glucose at 11.7T and 30°C. The peaks are labeled accordingly, and pH was 7.0.

[0035] FIGs. 10A-10E show in vitro water saturation shift referencing (WAS SR) method. The CEST effect is notoriously susceptible to the inhomogeneity of the static magnetic field (Bo). (FIG. 10A) Bo map, in Hz, where dark blue (less than 20 Hz on the scale) indicates good shim of the sample. For regions with Bo inhomogeneity, the spillover effect is no longer symmetric. Direct water saturation imaging allows measurement of the absolute water frequency in each voxel, allowing accurate repositioning of Z-spectra on a voxel-by-voxel

[0036] - 5 -

[0037] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) basis independent of spatial Bo field variations. Bo inhomogeneity in CEST was corrected on a voxel-by-voxel basis through shifting ofZ-spectra, where (FIG. 10B) and (FIG. 10C) show examples of voxels (49,14) and (32,26) in (x,y) dimensions from two samples with poor (top) and good (bottom) shims. This method requires acquisition of saturation images at 20 to 40 different frequencies. The shape of the direct water saturation Z-spectrum is not affected by Bo inhomogeneities making it symmetric with respect to its center frequency (point of minimum intensity). To determine the Bo offset in each voxel, the center of the observed signal (red line) is measured by fitting wdth a Lorentzian function. Next, the entire Z- spectrum is shifted (blue line in FIG. 10B and FIG. 10C) for alignment of water signal in all voxels. Images before and after Bo correction with WASSR are shown in (FIG. 10D) and (FIG. 10E), respectively. The images before Bo correction (FIG. 10D) show a small spatial intensity variation over the top samples compared to all samples after Bo correction (FIG. 10E).

[0038] FIG. 11 shows glycogen and glucose chemical exchange saturation transfer dependence on pH. Glycogen / glucose concentration independent detection (GGCID) CEST were measured directly from Z spectra from mixture samples of 5mM glycogen and 5 mM glucose with different pH using equation lb (see material and methods). Ratiometric GGCID values decreased linearly when increasing pH from 6.6 to 7.6.

[0039] FIGs. 12A-12E show in vitro ratiometric CEST and NOE* CEST maps of different mixtures of glycogen and glucose. (FIG. 12A-12C) Calculated ratiometric maps using saturations at +0.7 / +2. 1 ppm, +0.7 / +1 .2 ppm and +2. 1 / +1 .2 ppm. See FIG. 3A for details of each tube. (FIG. 12D) Glycogen NOE* maps (wdth saturation at -1.2 ppm) for a mixture of glucose and glycogen of different concentrations (see FIG. 3A for details of each tube). Glycogen NOE* CEST intensity shows a linear dependence on concentration (R2= 0.98). This peak is not visible at 1 mM concentration. (FIG. 12E) Glycogen and glucose CEST dependence on pH w ith GGCID were measured directly from Z spectra from mixture samples of 5 mM glycogen and 5 mM glucose with different pH using equation lb (see Material and Methods). Ratiometric GGCID values decreased linearly when increasing pH within the physiological range See FIG. 4 for the pH maps derived from the ratiometric CEST of glucose / glycogen peaks.

[0040] FIGs. 13A-13B show" distribution analysis of glycogen, glucose and pH in rat brains with U87 and U251 tumors. Histograms for glycogen, glucose, and pH were created, where x-axis represents the concentration or pH value, and y-axis represents the voxel count for (FIG. 13A) U87 and (FIG. 13B) U251 tumors. Higher values of glycogen concentration were

[0041] - 6 -

[0042] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) measured inside both U87 and U251 tumors (8. 12±1.97 mM and 7.16±2.01mM, respectively) compared to normal (contralateral) tissue (2.98±2.13 and 3.83±2.41mM, respectively). Same trend was observed for glucose, where higher concentrations was measured inside tumor (6.96±2.53 mM for U87 and 6.43±2.32 mM for U251) comparing to normal tissue (3.51±2.14 mM for U87 and 4.01±2. 15 mM for U251). The average pH was also lower inside both U251 and U87 tumors (6.9±0.2 for U87 and 6.8±0.3 for U251) compared to normal tissue (7.2±0.2 for both tumor types). See FIGs. 5A-5B for spatial distributions of glycogen, glucose, and pH inside and outside the tumors.

[0043] FIGs. 14A-14D show PAS and D-P AS-stained sections analyzed with brightfield microscopy. Brain sections stained with PAS (FIG. 14A) and D-P AS (FIG. 14B) pre-treated with dimedone. Images were transformed by decomposing the RGB image into "Blue” channel (middle row) and ‘'Red” channel (bottom row) based on the absorbance values of stained mixtures into absorbance values of single stains (for detailed description see Material and Methods). The PAS stain in RGB shows cell bodies and poly saccharides, whereas the D- PAS stain in RGB shows cell bodies with glycogen digested. For “Blue” channel, the PAS stain shows only the cell bodies, whereas the D-P AS stain shows cell bodies with glycogen digested. For “Red” channel, the PAS stain shows polysaccharides, whereas the D-PAS stain shows the glycogen digested. (FIG. 14C, FIG. 14D) Glycogen is determined by thresholding the RGB image at magenta level (50% Blue, 50% Red) because the green channel is empty. Analyzing the “Red” channel images at 20x magnification allows comparison of PAS and D- PAS stain in (FIG. 14C) tumors and (FIG. 14D) normal tissue. Tumor tissue shows -50% level drop in intensity7for D-PAS compared to PAS, whereas normal tissue shows -15% level drop in intensity for D-PAS compared to PAS. This suggests that glycogen in tumor tissue is significantly higher than in normal tissue. See FIG. 6 for in vivo CEST-derived glycogen maps.

[0044] FIGs. 15A-15D show a comparison of glucose derived from CEST and 'H-MRS in RG2 tumor. (FIG. 15A) The left image shows the tumor location (black outline inside the brain) on Tz-weighted MR images for RG2 tumor. The other two images show maps for glycogen and glucose. (FIG. 15B) Localized short echo time spectra of normal (black spectrum) and tumor (red spectrum) tissues m the same brain with 3.03 ppm creatine peak normalized. (FIG. 15C) The difference between the normal and tumor tissues (blue spectrum) shows the prominence of lactate, macromolecules (MM), myo-inositol (MI) as well as glucose (* at 3.44 ppm), whereas the N-acetyl aspartate (NAA) peak is low’er. (FIG. 15D) Glucose difference spectra from Gruetter R et al. (1996) J Cereb Blood Flow Metab. 16:427-

[0045] - 7 -

[0046] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764)

[0047] 438 show that glucose can be identified by a peak at 5.2 ppm and several peaks over 3.4-3.8 ppm, where the peak at 3.44 is most reliably measured.

[0048] DETAILED DESCRIPTION OF THE INVENTION

[0049] Reference will now be made in detail to certain embodiments of the disclosed subject. While the disclosed subject matter will be described in conjunction with the enumerated claims, it will be understood that the exemplified subject matter is not intended to limit the claims to the disclosed subject matter.

[0050] Throughout this document, values expressed in a range format should be interpreted in a flexible manner to include not only the numerical values explicitly recited as the limits of the range, but also to include all the individual numerical values or sub-ranges encompassed within that range as if each numerical value and sub-range is explicitly recited. For example, a range of "about 0.1% to about 5%" or "about 0.1% to 5%" should be interpreted to include notjust about 0.1% to about 5%, but also the individual values (e.g., 1%, 2%, 3%, and 4%) and the sub-ranges (e.g., 0.1% to 0.5%, 1.1% to 2.2%, 3.3% to 4.4%) within the indicated range. The statement "about X to Y" has the same meaning as "about X to about Y," unless indicated otherwise. Likewise, the statement "about X, Y, or about Z" has the same meaning as "about X, about Y, or about Z," unless indicated otherwise.

[0051] In this document, the terms "a," "an," or "the" are used to include one or more than one unless the context clearly dictates otherwise. The term "or" is used to refer to a nonexclusive "or" unless otherwise indicated. The statement "at least one of A and B" or "at least one of A or B" has the same meaning as "A, B, or A and B." In addition, it is to be understood that the phraseology or terminology employed herein, and not otherwise defined, is for the purpose of description only and not of limitation. Any use of section headings is intended to aid reading of the document and is not to be interpreted as limiting; information that is relevant to a section heading may occur within or outside of that particular section.

[0052] In the methods described herein, the acts can be carried out in any order, except when a temporal or operational sequence is explicitly recited. Furthermore, specified acts can be carried out concurrently unless explicit claim language recites that they be carried out separately. For example, a claimed act of doing X and a claimed act of doing Y can be conducted simultaneously within a single operation, and the resulting process will fall within the literal scope of the claimed process.

[0053] - 8 -

[0054] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764)

[0055] Definitions

[0056] The term "about" as used herein can allow for a degree of variability in a value or range, for example, within 10%, within 5%, or within 1% of a stated value or of a stated limit of a range, and includes the exact stated value or range.

[0057] The term "substantially " as used herein refers to a majority of, or mostly, as in at least about 50%. 60%, 70%, 80%, 90%, 95%, 96%, 97%, 98%, 99%, 99.5%, 99.9%, 99.99%, or at least about 99.999% or more, or 100%.

[0058] Methods of Magnetic Resonance Imaging using CEST Signals

[0059] In one aspect, the present disclosure addresses the need for better imaging techniques by providing a method of magnetic resonance imaging (MRI) that utilizes a CEST (Chemical Exchange Saturation Transfer) scheme to simultaneously enhance sensitivity of exchanging hydroxyl protons from both glycogen and glucose at high magnetic field strength, while further simultaneously quantifying pH.

[0060] In various embodiments, a method for magnetic resonance imaging is provided. In some embodiments, the method includes (i) exposing a sample to a saturating radiofrequency (RF) pulse with a frequency that is off-resonance with bulk water and a corresponding Bo field. In some embodiments, the method includes (ii) exposing the sample to at least one sampling RF pulse having an amplitude (Bi) of about 0.5 to about 3 pT. In some embodiments, the method includes (iii) exposing the sample to one or more RF pulses to obtain chemical exchange saturation transfer (CEST) data. In some embodiments, the method includes (iv) generating one or more images from the CEST data at a plurality of frequency offsets.

[0061] The number of frequency offsets is not particularly limited, and can be at least, greater than, or equal to about 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50,

[0062] 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75,

[0063] 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88. 89. 90, 91, 92, 93, 94, 95, 96, 97, 98, 99,

[0064] 100, or more separate frequency offsets. The frequency offsets can be made between the values of + 15 ppm and -15 ppm (relative to the water resonance).

[0065] In various embodiments, step iii) includes simultaneously obtaining CEST data for glucose and glycogen protons in the sample. In various embodiments, the glucose and glycogen protons are -OH protons (the hydrogen atom in the hydroxyl group) in each such

[0066] - 9 -

[0067] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) molecule present in the sample, where previous field correction methods can be used for correcting small field inhomogeneity’ across voxels in the CEST data.

[0068] In various embodiments, the sampling RF pulse has a Bi amplitude of 0.5, 1, 1.5, or 3 pT. In various embodiments, the sample is exposed to sequential RF sampling pulses with a Bi amplitude at each of 0.5, 1, 1.5, or 3 pT. The sequential RF sampling pulses can, in various embodiments, include a sequence of two, three, or four pulses, where each of the pulses has a different Bi amplitude. The sequential RF sampling pulses can be generated in any order, for example, three pulses with amplitudes of 0.5, 1, and 1.5 pT, or three pulses with amplitudes of 1.5, 0.5, 1 pT, and so on.

[0069] In various embodiments, the method includes correcting the inhomogeneity the Bo field in each voxel of the image using water saturation shift referencing. In various embodiments, the image is generated at glycogen or glucose proton magnetic resonance frequencies of at least one of +2.1, +1.2, +0.7, and -1.2 ppm relative to the bulk water resonance.

[0070] In various embodiments, the images are acquired in a magnetic field strength of at least about 4.7T, 7.0T, 9.4T, 10.5T, 11.7T, 15.2T, or about 17.2T. In various embodiments, the images are acquired in a magnetic field strength of at least about 11.7T.

[0071] In various embodiments, the method includes computing an absolute pH in a region of the sample using the ratiometric equation

[0072] [Mz(0.7ppm)]x[Mz(8 ppm)-Mz (2.1 ppm)]

[0073] GGCID-

[0074] [Mz(2.1 ppm)]x[Mz(8 ppm)-Mz(0.7 ppm)]

[0075] The computation of the absolute pH, when performed, is performed simultaneously with the acquisition of the CEST data for glucose and glycogen protons in the sample. In various embodiments, a pH map is generated using a plurality of absolute pH values.

[0076] In various embodiments, the absolute pH is determined at a Bi amplitude of 0.5 and 1.5 pT. In various embodiments, the absolute pH is determined at a Bi amplitude selected from the group consisting of 0.5, 1, 1.5, and 3 pT. In various embodiments, the absolute pH is determined at any two Bi amplitudes selected from the group consisting of 0.5, 1, 1.5, and 3 pT.

[0077] In various embodiments, the sample includes in vitro or in vivo tissue. In vitro tissue includes tissue samples taken from any organism, such as a mammal. In various embodiments, the mammal is a human. In vivo tissue includes an entire organism, such as any living mammal. In various embodiments, the in vivo sample is a human, or any portion of a human that is placed in the magnetic resonance scanner, such as a limb, head, chest,

[0078] - 10 -

[0079] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) abdomen, pelvis, and the like. In various embodiments, the sample includes cancerous tissue.

[0080] The cancerous tissue can be in vitro or in vivo.

[0081] System for Magnetic Resonance Imaging using CEST Signals

[0082] The methods described herein are implemented on an MRI scanner system with an imaging controller with a computer processor coupled to the imaging scanner and a non- transitory computer-readable memory storing instructions executable by the computer processor. Control over the scanner operation and imaging are provided by the controller computer and software. In various embodiments, a system for performing chemical exchange saturation transfer (CEST) magnetic resonance imaging (MRI) includes providing computer coupled to an MRI scanner adapted to image a sample. In certain embodiments, the computer comprises an imaging controller. In certain embodiments, the computer comprises a computer processor. In certain embodiments, the computer comprises a non-transitory computer-readable memory7storage that stores instructions executable by the computer processor. In certain embodiments, executing the instructions comprises exposing a sample to a saturating radiofrequency (RF) pulse with a frequency that is off-resonance with bulk water and a corresponding Bo field. In certain embodiments, executing the instructions comprises exposing the sample to at least one sampling RF pulse having an amplitude (Bi) of about 0.5 to about 3 pT. In certain embodiments, executing the instructions comprises exposing the sample to one or more RF pulses to obtain chemical exchange saturation transfer (CEST) data. In certain embodiments, executing the instructions comprises generating one or more images from the CEST data at a plurality7of frequency offsets.

[0083] In various embodiments, the computer processor is configured to cause the MRI scanner to simultaneously obtain CEST data for glucose and glycogen protons in the sample. In various embodiments, the computer processor is configured to cause the MRI scanner to apply a sampling RF pulse with a Bi amplitude of 0.5, 1, 1.5, or 3 pT. In various embodiments, the computer processor is configured to cause the MRI scanner to correct inhomogeneity the Bo field in each voxel of the image using water saturation shift referencing. In various embodiments, computer processor is configured to cause the computer and MRI scanner to generate the image at magnetic resonance frequencies of at least one of +2.1, +1.2, +0.7, and -1.2 ppm relative to the bulk water resonance. In various embodiments, the computer processor is configured to cause the MRI scanner to acquire the images at a magnetic field strength of at least 3.0T, 4.7T, 7.0T, 9.4T, 10.5T, 11.7T. 15.2T, 17.2T . In

[0084] - 11 -

[0085] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) various embodiments, the computer processor is configured to cause the MRI scanner to acquire the images at a magnetic field strength of at least 11 T.

[0086] In various embodiments, the computer processor is configured to cause the computer to compute an absolute pH in a region of the sample using the ratiometric equation

[0087] [Mz(0.7ppm)]x[Mz(8 ppm)-Mz (2.1 ppm)]

[0088] GGCID- [Mz(2.1 ppm)]x[Mz(8 ppm)-Mz(0.7 ppm)]

[0089] In various embodiments, the computer processor is configured to cause the computer to generate a pH map using a plurality of absolute pH values. In various embodiments, the computer processor is configured to cause the computer to determine the absolute pH at a Bi amplitude of 0.5 and 1.5 pT. In various embodiments, the computer processor is configured to cause the computer to determine the absolute pH at a Bi amplitude selected from the group consisting of 0.5, 1, 1.5, and 3 pT. In various embodiments, the computer processor is configured to cause the computer to determine the absolute pH at any two Bi amplitudes selected from the group consisting of 0.5, 1, 1.5, and 3 pT. In various embodiments, the sample includes in vitro or in vivo tissue as described herein. In various embodiments, the sample includes cancerous tissue.

[0090] In various embodiments, a storage medium for storing instructions executable by a computer processor communicatively connected to an MRI scanner to perform the method described herein is provided. In certain embodiments, when executed, the instructions cause the computer processor communicatively connected to an MRI scanner to expose a sample to a saturating radiofrequency (RF) pulse with a frequency that is off-resonance with bulk water and a corresponding Bo field. In certain embodiments, when executed, the instructions cause the computer processor communicatively connected to an MRI scanner to expose the sample to at least one sampling RF pulse having an amplitude (Bi) of about 0.5 to about 3 pT. In certain embodiments, when executed, the instructions cause the computer processor communicatively connected to an MRI scanner to expose the sample to one or more RF pulses to obtain chemical exchange saturation transfer (CEST) data. In certain embodiments, when executed, the instructions cause the computer processor communicatively connected to an MRI scanner to generate one or more images from the CEST data at a plurality of frequency offsets.

[0091] The CEST scheme described herein provides the capability to simultaneously enhance sensitivity7of exchanging hydroxyl protons from both glycogen and glucose at high magnetic field strength. This approach uses a long RF saturation pulse with low RF power to satisfy the slow-intermediate exchange condition required for CEST imaging of some diamagnetic

[0092] - 12 -

[0093] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) endogenous pools. This saturation scheme increases the sensitivity for those exchangeable pools resonating close to water, e.g.. hydroxyls in glucose and glycogen. In some embodiments, the experiments were performed in a high magnetic field of 11.7 T, enabling superior chemical separation of exchangeable pools. The low RF power minimizes nonspecific magnetization transfer effects arising from protons in semisolid tissues, while at the same time revealing the fine structure of the mobile proton components.

[0094] Furthermore, simultaneously measured CEST signals for glycogen and glucose at two distinct frequencies provides a way to apply a ratiometric approach and to quantify pH. This method glycogen / glucose concentration-independent detection can be called GGCID. To test the present hypothesis, CEST signals of glucose and glycogen in vitro were characterized by varying their concentration in solution samples, and then used the in vivo CEST signals to evaluate glucose and glycogen levels along with pH in rat brain. This GGCID method was applied in GBM models implanted in rat brains and significantly higher levels of sugar moieties and lower pH were found inside tumor compared to normal brain tissue. These findings can be further exploited as improved biomarkers of impaired sugar homeostasis in characterization of the different diseases such as neurological / psychiatric disease, stroke, etc. Because GGCID uses a low RF power for saturation, it can be safely translated to clinical settings.

[0095] Results

[0096] The GGCID method can be used to simultaneously enhance sensitivity of exchangeable hydroxyl (-OH) protons from both glycogen and glucose at high magnetic field. FIG. 1 shows sample Z-spectra for samples of bovine liver glycogen and a-D-(+)- glucose at concentrations ranging from 2.5 to 20 mM acquired with two different RF saturation amplitudes. In all spectra the water resonance is set to 0 ppm. Z-spectra of bovine liver glycogen at Bi of 3 pT shows one broad peak at +1.2 ppm, and the intensify of the peak decreases as concentration decreases (FIG. 1 A). At the lower Bi of 0.5 pT detection sensitivity of the other -OH groups from glycogen are heightened (FIG. IB). At lower power, the CEST effect from the Hl proton with a (1-4) linkage (FIG. 8A) is visible at +0.7 ppm and is well separated from the peak at +1.2 ppm and its intensify increases with higher concentration of glycogen in the sample. At 0.5 pT the peak at +1.2 ppm is well separated from the +0.7 ppm peak, while at higher Bi this peak becomes broader and overlaps with the +0.7ppm peak. The peak at -1.2 ppm visible at 0.5 pT originate from aliphatic H3, H5. and

[0097] - 13 -

[0098] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764)

[0099] H2+H4-1 glycogen protons (FIG. 8A) either through direct H1-H3 and H1-H5 dipolar interactions or cross-relaxation nuclear Overhauser effect (NOE*).

[0100] FIG. 1 also shows experiments with glucose samples for concentrations ranging from 2.5 to 20 mM and Bi amplitudes of 0.5 and 3 pT. At a Bi of 3 pT (FIG. 1C) glucose demonstrates a broad peak at +1.2 ppm whose intensity decreases with decreased glucose concentration. This broad CEST peak originates from combined contributions of H2+H3+H4+H5 glucose (-OH) protons (FIG. 8B). At Bi of 0.5 pT, the glucose CEST peaks separate into two distinctive signals (FIG. ID). Compared to glycogen which has three hydroxyl proton groups, glucose has five hydroxyl proton groups leading to the slightly larger frequency range. The proton resonance at +2. 1 ppm emanates from the a anomeric conformation of glucose (FIG. 8B). The glucose CEST peaks at +1.2 and +2.1 ppm become separated at lower Bi and which increased with rising concentrations. We were not able to detect glucose H6 peak due to its proximity to water.

[0101] There were no significant differences between CEST spectra of individual glycogen or glucose samples (FIGs. 1 and 8) when compared to the mixture samples (FIGs. 2 and 9). FIG. 2 shows that all resonances of glycogen (arrows at +0.7 ppm and -1.2 ppm) and glucose (arrow at +2. 1 ppm) are also visible in the mixture at lower Bi (dark blue line). The center arrow at positive ppm offset in FIG. 2 indicates the peak at +1.2 ppm is common for both glucose and glycogen. Similar to individual samples (FIGs. IB and ID), the resonance at +2. 1 ppm (from glucose) and +0.7 ppm (from glycogen) are not visible at higher Bi amplitude (FIG. 2, indicated by 3 pT arrow label). The proton resonance from glucose at +2.1 ppm has reduced intensity in the mixture (FIG. 9), but the detection sensitivity is good in the Z-spectra (FIG. 2).

[0102] We prepared mixture samples containing different concentrations of glycogen and glucose at pH 7.12 for validation purposes (FIG. 3 A) and conducted CEST imaging under the experimental conditions, which included Bo inhomogeneity corrections (FIGs. 10A-10E). Consistent with results shown in FIGs. 1 and 2, the CEST peaks at +0.7, +1.2 and -1.2 ppm from glycogen and +1.2 and +2. 1 ppm from glucose, have the same trend of increasing intensities with rising concentrations, without changes in their CEST properties in the mixture. For both mixtures and individual samples, at higher concentrations of glycogen and glucose, the peak at +1.2 ppm, characteristic of both molecules, overlaps with the peak at +0.7 ppm because of CEST signal broadening dominant at higher concentrations (FIG. 3B). Under these experimental conditions, a concentration of 1 mM for either glycogen or glucose is not sufficient to generate an observable CEST signal.

[0103] - 14 -

[0104] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764)

[0105] CEST maps from the mixture samples (see FIG. 3A for details of each tube) by saturation at +0.7 ppm, +2.1 ppm, and +1.2 ppm are shown in FIG. 4A-C. The ratiometric CEST maps for these saturations are shown in FIG. 12A-C. CEST contrast for glycogen and glucose were best by saturation at +0.7 ppm and +2.1 ppm, respectively, which showed linear dependence with increasing concentration (FIG. 4D). However, CEST contrast generated bysaturation at +1.2 ppm was less conclusive for either molecule (FIG. 4C). The concentration range of glycogen in our experiment (2.5-10 mM) is slightly larger than that for glucose (2.5- 5 mM). The CEST sensitivity for glycogen (black line, R2=0.96) is ~30% smaller than for glucose (red line, R2=0.99). In the same samples NOE* CEST maps (saturation at -1.2 ppm) showed linear dependence with glycogen concentration (FIG. 12D).

[0106] The pH maps based on the ratiometric CEST intensities (GGCID) with saturations at +0.7 / +2.1 ppm, +0.7 / +1.2 ppm and +2.1 / +1.2 ppm are shown in FIGs. 4E-4G (for calibration curve see FIG. 12E). The pH maps with saturations +0.7 / +2.1 ppm and +0.7 / +1.2 ppm (FIGs. 4E and 4F) are relatively homogenous, whereas the pH map with saturations +2.1 / +1.2 ppm (FIG. 4G) is more heterogeneous. The pH values measured with the pH electrode (black dashed line, pH=7. 12) are similar to those calculated by GGCID with saturations +0.7 ppm / +2.1 ppm (black bars) or +0.7 / +1.2 ppm (red bars) (FIG. 4H). For glycogen and / or glucose concentrations of 1 mM, the pH calculation is not very accurate due to the weak CEST signal.

[0107] Following promising sensitivity from in vitro data, we tested the feasibility of our new CEST methods in vivo using different GBM cell lines (FIG. 5). The first column from the left shows anatomical images for both an U87 (FIG. 5A) and U251 tumor (FIG. 5B). The second and third columns show maps of glycogen and glucose CEST intensities, with an average glycogen concentration of 8. 1+2.0 mM, glucose concentration of 7.0+2.5 mM, and low tissue pH of 6.9+0.2 recorded in U87 tumors compared to contralateral tissue (FIG. 5A). Similar trends are observed for U251 tumors (FIG. 5B) with average glycogen concentration of 7.2+2.0 mM, glucose concentration of 6.4+2.3 mM, and low tissue pH around 6.8+0.3 in the tumor. For a more in-depth analysis and comparison of tumor and normal tissue for the two GBM cell lines we generated histograms of glycogen and glucose concentrations, as well as of pH (FIG. 13).

[0108] Glycogen concentrations calculated from CEST maps for glycogen peaks at +0.7 ppm (FIG. 6A) and -1.2 ppm (FIG. 6B) show- similar, high concentrations inside tumor (FIG. 6C). However, the NOE* map shows widespread signal beyond the tumor boundary suggesting that this CEST peak could be representing a complex mixture of molecular

[0109] - 15 -

[0110] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) moieties attributed to the NOE* peak in vivo. On PAS stained microscopy images obtained with using a lOx objective (FIG. 6D), a heterogeneous distribution of PAS positive structures can be seen throughout the whole brain indicating a high proportion of carbohydrate macromolecules (glycogen, glycoprotein, proteoglycans, mucins or other molecules with polysaccharide moieties). On the sections treated with amylase (FIG. 6E) diminished PAS signal, due to glycogen digestion, yielded background staining from nonglycogen sources. Representative areas of the tumor imaged using 20x magnification (black boxes on FIGs. 6D and E) show deep magenta color from glycogen inside tumor (black arrow s) which disappeared after enzymatic digestion, confirming the presence of glycogen in the corresponding section. Comparison of PAS and D-PAS stains in tumor tissue shows -50% level drop in intensity for D-PAS compared to PAS, whereas for normal tissue this decrease was only -15% (FIG. 14).

[0111] To validate the tissue pH measurements calculated by the GGCID method, we obtained pH maps using the already established AACID method. Tissue pH maps were calculated using both GGCID (FIG. 7A) and AACID (FIG. 7B) in the same brain with U87 tumor. The tissue pH maps with GGCID are slightly more diffusive, compared to AACID, presumably due to the compartmentation of glycogen being purely intracellular whereas glucose is contained across compartments. Averaged Z-spectra from the tumor region in vivo show detectable peaks for GGCID (FIG. 7C) and AACID (FIG. 7D). Calculated pH values using these two CEST methods show good agreement for both tumor and normal tissue (FIG. 7E). Inside tumors, the average pH values measured by GGCID and AACID were 6.83±0.02 and 6.75±0.03, respectively (FIG. 7E). To validate glucose level inside and outside the tumor, we conducted in vivo1H-MRS single voxel experiments to obtain localized short TE spectra from normal and tumor tissues of the same rat brain bearing RG2 tumor (FIG. 15). Similar to U251 and U87 tumors (FIG. 5), in RG2 tumor we observe higher sugar moieties by CEST (FIG. 15 A). Comparison of short TE spectra of normal and tumor tissues (by normalization of creatine) show that the tumor has higher lactate and glucose, but lower N-acetyl aspartate (FIGs. 15B, 15C). Specifically, the 3.4 pm peak is most reliable for glucose in the difference spectra . and which has been used in previous functional activation studies in human brain although other peaks located at 3.4-3 8 and 5.2 ppm can also be attributed to glucose (FIG. 15D).

[0112] - 16 -

[0113] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764)

[0114] Selected Discussion

[0115] GBM (glioblastoma multiforme) is a very aggressive primary brain tumor with poor prognosis, high resistance to chemotherapy and short survival. Thus, identification of new molecular biomarkers to detect therapeutic targets, improve diagnostic accuracy and monitor treatment efficacy is urgently needed.

[0116] CEST is mainly used for studying exchangeable protons, most of which are in the slow exchange regime where the exchange rate constant k is much smaller than their chemical shift. With a growing interest in studying molecules where k is close to intermediate / fast exchange, CEST imaging has been employed to optimize acquisition protocols to target myo-Inositol, glutamate, glucose, creatine, glycosaminoglycan or glycogen. To detect these molecules with CEST, it is crucial to optimize the acquisition protocol based on their exchange properties, but relaxation times, concentration, pH and temperature should also be considered. Furthermore, given that the CEST frequency for the hydroxyl group is close to water resonance, contamination by direct water saturation (DWS) and magnetization transfer (MT) effects with high Bi power are to be expected. Thus, to minimize the DWS and MT effects, a lower saturation power is advantageous, but this comes at the cost of contrast. Surprisingly, the present disclosure shows the feasibility of using a lower Bi power and longer pulse to avoid DWS / MT and to achieve higher CEST sensitivity for in vivo glycogen / glucose detection in GBM tumors in rat brain.

[0117] The current saturation scheme allowed detection of glycogen Z-spectra with resonances of interest at +0.7, +1.2 and -1.2 ppm (FIGs. IB and ID; FIG. 8A and 8B). In vitro Z-spectra of glucose, under the same conditions, showed resonances at +1.2 and +2.1 ppm, similarly to results reported by Zaiss and coworkers who showed that at higher pH and higher Bi, the +1.2 ppm glucose resonance gets stronger and broader while the +2.1 ppm resonance decreases. They also showed that the H6 resonance of glucose at +0.6 ppm becomes visible at pH<6 but cannot be seen at pH>6.5. Without being bound by theory, this explains why this resonance was not observed in our glucose Z spectra at pH~7, and why we assumed the glucose H6 resonance has no contribution to glycogen peak at +0.7 ppm and low Bi. All glycogen / glucose properties were well preserved in the mixture (FIG. 2) and all peaks were visible from their individual assignments (FIG. 1). A trend of increasing intensity for the +1.2 ppm peak with increasing Bi amplitude was observed, where exchange rates are faster and glucose and glycogen peaks converge. The +1.2 ppm peak is common for both molecules with the same broadening behavior at higher Bi.

[0118] - 17 -

[0119] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764)

[0120] CEST parameters to separate glucose and glycogen signals and use them for tissue pH quantification with GGCID were sought. Thus, the joint peak at +1.2 ppm could not be used for pH quantification. The detection sensitivity of glucose and glycogen peaks from a mixture was used to imitate the in vivo tumor microenvironment for translation at higher magnetic fields and was tested using in vitro samples with various glucose and glycogen concentrations. Mixtures of glucose and glycogen with different concentration ratios (FIG. 3) demonstrated enough sensitivity and specificity, especially at higher concentrations (FIG. 4). We hypothesized higher glucose and glycogen concentrations in tumors. Although glucose is multi-compartmental, tumors are expected to have higher levels due to upregulated glucose transporters and blood volume. Histology shows that tumors are enriched in glycogen.

[0121] In this disclosure it was found that in brain tumors, the basal levels of glycogen and glucose are higher compared to normal tissue (FIG. 5), where presence of elevated tumoral glycogen was confirmed by PAS staining (FIG. 6) and higher glucose was measured by1H- MRS (FIG. 14). Without being bound by theory', the high glycogen levels may be explained by the fact that glycogen synthesis occurs in response to the lack of oxygen and nutrients, which are both a hallmark of solid tumors. In the healthy brain, glial (but also neuronal) glycogen is protective against hypoxia-induced cell death. Distinctly for tumor cells, the level of the two enzy mes responsible for glycogen breakdow n and synthesis are increased, suggesting that both are important for tumor progression. The most important mediators of these responses are HIFs, as shown by Favaro and coworkers in U87 cell lines. This Favaro study demonstrated that glycogen levels experience significant temporal change in U87 cells exposed to hypoxic stress. They found that depletion of PYGL and consequent glycogen accumulation reduces the input into the pentose phosphate pathway - a major biosynthetic pathway - thereby significantly weakening tumor growth. In the present disclosure, the glycogen level in U87 tumors was estimated to be higher than that of glucose, and similar values were observ ed for U251 (FIG. 5) and RG2 (FIG. 14) tumors.

[0122] Under hypoxic conditions cancer cells increase glucose uptake and subsequently store more glycogen. Pelletier et al. reported that cancer cell lines had several-fold increase in glycogen content when cultured in 1% oxygen, which was HIF-la dependent. Once glucose w as removed from the media, cells that had accumulated glycogen stores under hypoxia had greater viability7(70-80%) after 24 h than cells cultured in normoxia (20-60%). It appears that glucose consumption via glycogen is favorable to directly enter downstream pathways, which are necessary for optimal tumor growth and could explain the higher glycogen concentration observed inside tumors. Several earlier as well as recent studies estimated

[0123] - 18 -

[0124] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) glycogen concentrations to be up to 200 pg / mg in malignant human intestine epithelial cells and up to 400 pg / mg in kidney tumors, both of which are much higher than found in neighboring normal tissues.

[0125] The main feature of solid tumors is not only local hypoxia, but rather a highly heterogenous microenvironment resulting from the combination of poor vascular perfusion and local hypoxia. Recently, glycogen metabolism has been recognized as an important metabolic pathway for reprogramming cancer cells. It is well established in cancer cells in vitro and tumor models in vivo that glycogen metabolism is upregulated in tumors in response to hypoxia, and hypoxia induces an early accumulation of glycogen, followed by a gradual decline. Favaro and colleagues reported that GYSI showed a rapid induction (followed by a later PYGL increase). PYGL depletion and consequent glycogen accumulation led to increased reactive oxygen species (ROS) levels that contributed to a p53-dependent induction of senescence and markedly impaired tumorigenesis in vivo. Therefore, an essential role of glycogen breakdown has been demonstrated in cancer cells. Taken together with these findings, an enhanced induction of senescence could be promoted in cancer cells through glycogen accumulation, or through suppression of glycogen depletion. A precise evaluation of glycogen content in cancer cells in vivo could provide a way for early detection of tumors, and perhaps track their response to therapies.

[0126] As described herein, metabolic adaptations such as catabolism of sugars represents a canonical response of tumor cells to survive, but this results in an acidic pH in tumors. Pyruvate made by glycolysis can either enter the tricarboxylic acid (TCA) cycle or be reduced to lactate, which occurs under hypoxic conditions. In attempt to balance this acute low' intracellular pH, several families of proton transporters such as the Na+ / H+exchanger (NHE). vacuolar H+ATPases and sodium bicarbonate cotransporter (Na / HCOs ) excrete protons and lactate into the extracellular / extravascular compartment resulting in the reversal of extracellular pH gradient in tumors. We have discussed previously how these channels constitute a complex interplay between H+and sodium ions (Na+) . Therefore, the mechanisms behind this reversed extracellular pH gradient also affect Na+balances across the plasma cell membrane and blood vessel wall. We demonstrated these phenomena using MRS techniques employing a paramagnetic probe whose spectroscopic properties have been rigorously analyzed.

[0127] Using two ratiometric CEST methodologies, AACID and GGCID, acidic tissue pH (6.7-6.9) was found inside tumors compared to normal tissue (7-7.2). which was in good agreement with previous results by APT CEST. Low extracellular pH in gliomas w as also

[0128] - 19 -

[0129] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) observed in several studies using chemical shift imaging with biosensor imaging of redundant deviation in shifts (BIRDS), hyperpolarized13C bicarbonate and PET. Due to the nonspecific compartmentation of glucose (i.e., in blood, extracellular, and intracellular) vs. glycogen (i.e., intracellular), the GGCID derived pH is more representative of “bulk” pH rather than any specific compartment. Although AACID signal originate from mobile peptides and proteins, found predominantly intracellularly, it is more accurate to consider the GGCID obtained pH as an average tissue pH that is weighted by blood, intracellular, and extracellular compartments. However, applying a ratiometric approach in pH measurement eliminates the concentration dependence of in vivo CEST imaging.

[0130] Moreover, GGCID has potential applications not only for cancers, but also for other diseases.

[0131] Examples

[0132] Various embodiments of the present application can be better understood by reference to the following Examples which are offered by way of illustration. The scope of the present application is not limited to the Examples given herein.

[0133] Materials and Methods

[0134] Sample Preparation

[0135] In vitro experiments were performed using bovine liver glycogen (Type IX G0885, CAS no. 9005-79-2, Sigma-Aldrich. St. Louis, MO. USA) and D-(+)-Glucose (G7021. CAS no.50-99-7, Sigma- Aldrich, St. Louis, MO, USA). The solutions of bovine liver glycogen, with an average molecular mass of 50 kD and glucosyl unit concentrations of I, 2.5, 5, 10, 15 and 20 mM were prepared by dissolving the glycogen in water (70% H2O and 30% D2O), adjusting pH to 7.0±0.2 with NaOH / HCl and adding 2% Nabfi for preservation purpose. The glycogen concentration was expressed in glycosyl units, each glucosyl unit contributing with 168 g / mol. Glucose samples were prepared in the same manner as bovine liver glycogen.

[0136] In vitro and in vivo1H-MRS

[0137] The 'H-MRS and Z-spectra for glycogen and glucose were collected on a Bruker Avance III HD 500 MHz vertical-bore spectrometer with TopSpin v2. 1 software (Bruker, Billerica, MA, USA). The samples were stabilized at 30 °C using the probe’s internal heating and cooling devices. High resolution ’H spectra were measured using a 90° rectangular pulse and a spectral width of 5000 Hz. 1024 FIDs were collected, each having 43252 points and an acquisition time of 2.1626 s. For CEST measurements, the off-resonance saturation was

[0138] - 20 -

[0139] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) achieved by a rectangular, continuous wave RF pulse of 4s duration and different RF amplitudes (Bi): 0.5. 1, 1.5 and 3 pT. For signal acquisition a 90° rectangular pulse was used. A 14 s delay was used to allow complete recovery of water signal before the next saturation event. The use of 8 dummy scans established a steady state before data acquisition. The CEST effect was quantified by measurement of water resonance signal (Mz) and normalization with equilibrium signal (Mo). Z-spectra were sampled using 201 frequency offsets in equal, 0. 1 ppm steps between ±10 ppm relative to the water peak.

[0140] In vivo1H-MRS single voxel spectra were obtained on a 11 7T horizontal-bore Bruker Avance system interfaced with Bruker ParaVision v6.0.1 software (Billerica, MA) using a 14-mm diameter surface coil and positioned on top of the animal's head. The ’H spectra for both normal and tumor regions were obtained using localization by adiabatic selective refocusing (LASER) sequence with a repetition time (7R) of 3 s, echo time (7 ) of 8 ms, a spectral window of 6 kHz, 2048 acquisition points and 256 averages, where voxel sizes ranged from 2x2x2 mm to 3x3x3 mm. We used variable power and optimized relaxation delays (VAPOR) for water suppression. A 5 Hz line broadening was applied. The creatine peak at 3.03 ppm in normal and tumor spectra were aligned and scaled to match. Previous functional activation studies in human brain used difference of short TE spectra obtained before and during stimulation, where the 3.44 pm peak is reliable for measurement of glucose level in the difference spectra.

[0141] In vitro and in vivo CEST imaging

[0142] For validation of in vitro CEST imaging, separate samples were prepared containing mixtures of glycogen and glucose with different concentrations at pH 7.12: (i) 1 mM glycogen + 1 mM glucose. (ii)l mM glycogen +2.5 mM glucose, (iii) 2.5 mM glycogen + 2.5 mM glucose, (iv) 2.5 mM glycogen + 5 mM glucose, (v) 5 mM glycogen + 2.5 mM glucose, (vi) 10 mM glycogen + 2.5 mM glucose.

[0143] CEST imaging was conducted on a 11.7T horizontal-bore Bruker Avance system interfaced with Bruker ParaVision v6.0.1 software (Billerica, MA) using a ‘H birdcage volume coil (4 cm diameter) for the in vitro experiments, and a hybrid ‘H volume transmit / surface receive coil (volume: 8 cm diameter, surface: 3.5 cm diameter) for the in vivo experiments. Sample positioning and power optimizations for 'H signals were performed using Bruker’s gradient-echo (GE) and fast spin-echo (FSE) sequences. Shimming was done using an ellipsoid region to bring the water linewidth to less than 30 Hz using static magnetic field (Bo) mapping with second-order shim. 'H anatomical MRI was first performed using a

[0144] - 21 -

[0145] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) spin-echo sequence in axial orientation (FOV: 42x42 mm3, 128x 128 in-plane resolution). T2 mapping was done with 10 echo times TE (10 - 100 ms) with TR = 4000 ms. Voxelwise T2 maps for 'H were calculated by fitting the MRI voxel intensities versus TE values to a single exponential function eTK r2_

[0146] Similar to a recent ratiometric MRI technique that measured absolute pH using the ratio of CEST-mediated contrast from amine and amide protons (AACID: amine / amide concentration-independent), the feasibility of absolute pH imaging with ratiometric CEST was tested using glucose and glycogen signals (GGCID: glycogen / glucose concentrationindependent detection). Single-slice GGCID CEST MRI data were acquired using a 2D spinecho echo-planar imaging (EPI) pulse sequence, with a continuous wave RF saturation pulse (0.5 pT for 4 s), with the following parameters: matrix size = 64 x 64; FOV = 32 x 32 mm2; slice thickness 2 mm, TR=8000 ms; TE=12. 16 ms and 2 averages. The Z-spectrum was collected with 73 frequency offsets in unequal steps betw een ±8 ppm relative to the water peak, the smaller offsets steps being concentrated around previously defined resonances of interest from glucose and glycogen. For Bo correction, water saturation shift-referencing (WASSR) images with 61 frequency offsets were acquired, with an RF saturation Bi amplitude of 0.2 pT and 2 s duration, which w as chosen to minimize magnetization transfer (MT) effects while allowing sufficient direct saturation. The same parameters w ere used for in vitro and in vivo experiments, and all 'H- RI and CEST data were processed using inhouse written code in MATLAB (MathWorks, Natick, MA, USA).

[0147] The same slice geometry and orientation was used for GGCID and ACCID, where the MRI data w ere acquired using 2D spin echo EPI sequence with a slice thickness of 2 mm, FOV of 32x32 mm2, a matrix size of 42x42 interpolated to 64x64 during post-processing to match GGCID. TR / TE=7000 / 10 ms and 2 averages. For AACID a 2s RF saturation pulse with a Bi amplitude of 1.5 pT and 70 frequency offsets between ± 15 ppm was used. Before each CEST experiment a Bi field map w as obtained using a two flip angles imaging sequence (TR=10000 ms, TE=6 ms, flip angles 60° and 120°, 1 average, 1 slice, FOV= 32x32 mm2and matrix size 42x42. total acquisition time = 10 min). Bi was calculated for each voxel and used to determine the RF power levels required to generate the Bi fields of 0.5 and 1.5 pT for GGCID and ACCID, respectively. Since Bi variation was small throughout the sample (<6%), no Bi correction w as needed. Using the Bo-corrected Z-spectra, the ACCID and GGCID values w ere calculated on a voxel-by-voxel basis:

[0148] - 22 -

[0149] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764)

[0150] [Mz(0.7ppm)]x[Mz(8 ppm)-Mz (2.1 ppm)]

[0151] GGCID = (lb)

[0152] [Mz(2.1 ppm)]x[Mz(8 ppm)-Mz(0.7 ppm)] where 3.5 and 2.75 ppm are the frequencies of amide and amine, respectively (for AACID), and 0.7 and 2.1 ppm correspond to glycogen and glucose frequencies, respectively (for GGCID). The signals at 6 ppm and 8 ppm were used as references for AACID and GGCID, respectively.

[0153] Tumor implantation

[0154] The in vivo protocol was approved by the Institutional Animal Care and Use Committee (IACUC) at Yale University. Rats (athymic / nude and Fischer 344) were purchased through Yale University' vendors. RG2, U251, and U87 cell lines were purchased from American Type Culture Collections (ATCC, Manassas, VA, USA). The cells were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% heat- inactivated fetal bovine serum (FBS) and 1% penicillin- streptomycin. U87 cells were cultured in high glucose (4.5 g / L) DMEM and U251 cells were grown in low glucose (1.5 g / L) DMEM. All cells were maintained in an incubator in a 5% CO2 atmosphere at 37°C. Cells for tumor inoculation were harvested upon reaching at least 80% confluence and were prepared in FBS-free DMEM. Nude rats were injected intracranially with 2 - 5 x 106tumor cells either from the U2 1 (n = 2) or the U87 (n = 6) cell line (5-pL aliquot). Similarly, Fischer 344 rats were prepared for RG2 tumor (n=5). During the procedure, animals were anesthetized via isoflurane inhalation (2 - 3%). Injections were performed using a 10-pL Hamilton syringe with a 26G needle into the right striatum, 3 mm to the right of bregma and 3 mm below the dura. Animals were given bupivacaine (2 mg / kg at incision site) and carprofen (5 mg / kg, subcutaneously) during the tumor inoculation to relieve pain. Carprofen was given daily for two days post-inoculation. When the tumor had reached a minimum mean diameter of 3 mm each animal was imaged using the described imaging protocol. Before imaging the animal was given Puralube Vet Ointment (Dechra, Overland Park, KS, USA) over the eyes. The breathing rate was measured by placement of a pressure sensitive sensor under the torso, and the temperature was monitored using a rectal fiber-optic probe.

[0155] Brain Tissue Preparation

[0156] After the MRI experiments, rats were exposed to focused microwave irradiation using a microwave fixation system (Muromachi Kikai, Tokyo, Japan). Anesthetized rats were positioned inside the microwave system in a plastic animal holder with a hollow wall filed

[0157] - 23 -

[0158] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) with iced water. High energy (4.5 kW) micro waves were focused on the head of the rat for 1.2 s. Immediately after, the brain was removed from the skull and fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (PBS) overnight at 4°C. After fixation, the brains were transferred to 70% ethanol at 4°C. After 48 hours the brains were sliced at 3 mm in the axial orientation for paraffin embedding and cut to thin sections of 6 pm.

[0159] Histology

[0160] Before tissue staining, the sections were deparaffmized by 3-5 minutes of wash in xylene three times, followed by 100% ethanol tw ice, and then once with 95%, 70%, and 50% ethanol, and then three times with deionized water (DI). Glycogen staining kit (Polyscience, Warrington, PA, USA; Product number - 24200-1) contained 0.5% Periodic acid aq (250ml), Schiff s reagent (250 mL), 0.55% Potassium Metabisulfite (250 ml), Harris hematoxylin Acidified (250 mL), phosphate Citrate buffer pH 5.0 (250 mL) and diastase (refers to any a-, P-, or y-amylase) powder (0.5 mg)). Periodic acid Schiff (PAS) staining was performed as described previously. Briefly, 6 pm-sectioned slices were oxidized by a treatment with 0.5% periodic acid for 10 minutes. The oxidative process results in formation of aldehyde groups. Free hydroxyl groups should be present for oxidation to take place and completed upon reaching the aldehyde stage. Slice sections used as controls were incubated in saturated dimedone (a cy clic diketone) solution at 60°C for 30 minutes to depolymerize glycogen into smaller sugar units (maltose and glucose), which were washed out of the section. Dimedone is an aldehyde blocking reagent for histochemical depiction of glycogen. The dimedone-PAS (D-PAS) stain detects where glycogen is located, which is otherwise difficult to observe because of proximity of diastase-fast PAS positive materials. Schiff s reagent w as then used for 5 minutes at room temperature to localize the aldehyde groups. A colorless, unstable dialdehyde compound formed, which then transformed to the colored final product by restoration of the quinoid chromophore groups. The last step w as to counterstain slides with hematoxylin, for no more than ~3 minutes. After staining, the dehydration was done in reverse order (DI water - increasing ethanol - xylene) as to the washes mentioned above. The PAS and D-PAS sections were then visualized using a microscope (Keyence BZ-X, Itasca, IL, U.S.A). Microscopy images were analyzed in ImageJ (NIH and the Laboratory for Optical and Computational Instrumentation, LOCI, University of Wisconsin) applying color deconvolution with color defined vectors. ImageJ allowed deconvolution of color information acquired with red-green-blue (RGB) cameras and calculation of each stain contribution based on stain-specific RGB absorption. The “Blue channel” image represented hematoxylin

[0161] - 24 -

[0162] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764)

[0163] (stained cell nuclei appear bluish), whereas the “Red channel'’ reflected the PAS stain (positive PAS reflecting polysaccharides). The PAS stain in RGB shows cell bodies and polysaccharides, whereas the D-PAS stain in RGB shows cell bodies with glycogen digested. The PAS stain in “Blue channel” shows only the cell bodies, whereas the D-PAS stain shows cell bodies with glycogen digested. The PAS stain in “Red channel” shows polysaccharides, whereas the D-PAS stain shows the glycogen digested. Glycogen is determined by thresholding the RGB image at magenta level (50% Blue, 50% Red) as the “Green channel” is empty. Analyzing the “Red channel” images allows comparison of PAS and D-PAS stain in tumors and normal tissue. Tumor tissue shows -50% drop in intensity for D-PAS compared to PAS, whereas normal tissue shows -15% drop in intensity for D-PAS compared to PAS. In other words, the greater the loss of the magenta color in the D-PAS section, higher the amount of glycogen present.

[0164] The terms and expressions employed herein are used as terms of description and not of limitation, and there is no intention in the use of such terms and expressions of excluding any equivalents of the features shown and described or portions thereof, but it is recognized that various modifications are possible within the scope of the embodiments of the present application. Thus, it should be understood that although the present application describes specific embodiments and optional features, modification and variation of the compositions, methods, and concepts herein disclosed may be resorted to by those of ordinary skill in the art. and that such modifications and variations are considered to be within the scope of embodiments of the present application.

[0165] Enumerated Embodiments

[0166] The following enumerated embodiments are provided, the numbering of which is not to be construed as designating levels of importance:

[0167] Embodiment 1 provides a method for magnetic resonance imaging. In certain embodiments, the method comprises: i) exposing a sample to a saturating radiofrequency (RF) pulse with a frequency that is off-resonance with bulk water and a corresponding Bo field. In certain embodiments, the method comprises ii) exposing the sample to at least one sampling RF pulse having an amplitude (Bi) of about 0.5 to about 3 pT. In certain embodiments, the method comprises iii) exposing the sample to one or more RF pulses to obtain chemical exchange saturation transfer (CEST) data. In certain embodiments, the method comprises iv) generating one or more images from the CEST data at a lurality of frequency offsets.

[0168] - 25 -

[0169] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764)

[0170] Embodiment 2 provides the method of embodiment 1 , wherein step iii) comprises simultaneously obtaining CEST data for glucose and glycogen protons in the sample.

[0171] Embodiment 3 provides the method of any one of embodiments 1-2, wherein the sampling RF pulse has a Bi amplitude of about 0.5, 1, 1.5, or 3 pT.

[0172] Embodiment 4 provides the method of any one of embodiments 1-3, wherein the image is generated at magnetic resonance frequencies of at least one of +2.1, +1.2, +0.7, and -1.2 ppm relative to the bulk water resonance.

[0173] Embodiment 5 provides the method of any one of embodiments 1-4, wherein the images are acquired in a magnetic field strength of at least 11.7 T.

[0174] Embodiment 6 provides the method of any one of embodiments 1-5, further comprising computing an absolute pH in a region of the sample using the ratiometric equation

[0175] Embodiment 7 provides the method of any one of embodiments 1-6, wherein a pH map is generated using a plurality of absolute pH values.

[0176] Embodiment 8 provides the method of any one of embodiments 1-7, wherein the absolute pH is determined at a Bi amplitude of 0.5 and 1.5 pT.

[0177] Embodiment 9 provides the method of any one of embodiments 1-8, wherein the sample comprises in vitro or in vivo tissue.

[0178] Embodiment 10 provides the method of any one of embodiments 1-9. wherein the sample comprises cancerous tissue.

[0179] Embodiment 11 provides a system for performing chemical exchange saturation transfer (CEST) magnetic resonance imaging (MRI). In certain embodiments, the system comprises providing computer coupled to an MRI scanner adapted to image a sample. In certain embodiments, the computer comprises an imaging controller. In certain embodiments, the computer comprises a computer processor. In certain embodiments, the computer comprises a non-transitory computer-readable memory storage that stores instructions executable by the computer processor. In certain embodiments, executing the instructions performs the step of i) exposing a sample to a saturating radiofrequency (RF) pulse with a frequency that is off-resonance with bulk water and a corresponding Bo field. In certain embodiments, executing the instructions performs the step of ii) exposing the sample to at least one sampling RF pulse having an amplitude (Bi) of about 0.5 to about 3 pT. In certain embodiments, executing the instructions performs the step of iii) exposing the sample

[0180] - 26 -

[0181] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) to one or more RF pulses to obtain chemical exchange saturation transfer (CEST) data. In certain embodiments, executing the instructions performs the step of iv) generating one or more images from the CEST data at a plurality of frequency offsets.

[0182] Embodiment 12 provides the system of embodiment 11, wherein the computer processor is configured to cause the MRI scanner to simultaneously obtain CEST data for glucose and glycogen protons in the sample.

[0183] Embodiment 13 provides the system of any one of embodiments 11-12, wherein the computer processor is configured to cause the MRI scanner to apply a sampling RF pulse with a Bi amplitude of 0.5, 1, 1.5, or 3 pT.

[0184] Embodiment 14 provides the system of any one of embodiments 11-13, wherein the computer processor is configured to cause the computer and MRI scanner to generate the image at magnetic resonance frequencies of at least one of +2.1, +1.2, +0.7, and -1.2 ppm relative to the bulk water resonance.

[0185] Embodiment 15 provides the system of any one of embodiments 11-14, wherein the computer processor is configured to cause the MRI scanner to acquire the images at a magnetic field strength of at least 11 T.

[0186] Embodiment 16 provides the system of any one of embodiments 11-15, wherein the computer processor is configured to cause the computer to compute an absolute pH in a region of the sample using the equation

[0187] Embodiment 17 provides the system of any one of embodiments 1 1-16, wherein the computer processor is configured to cause the computer to generate a pH map using a plurality of absolute pH values.

[0188] Embodiment 18 provides the system of any one of embodiments 11-17. wherein the computer processor is configured to cause computer to determine the absolute pH at a Bi amplitude of 0.5 and 1.5 pT.

[0189] Embodiment 19 provides the system of any one of embodiments 11-18, wherein the sample comprises in vitro or in vivo tissue.

[0190] Embodiment 20 provides the system of any one of embodiments 11-19, wherein the sample comprises cancerous tissue.

[0191] Embodiment 21 provides storage medium for storing instructions executable by a computer processor to perform a method. In certain embodiments, the method comprises i) exposing a sample to a saturating radiofrequency (RF) pulse with a frequency that is off-

[0192] - 27 -

[0193] 56933733.1 Attorney Docket No.: 047162-7396WO1 (02764) resonance with bulk water and a corresponding Bo field. In certain embodiments, the method comprises ii) exposing the sample to at least one sampling RF pulse having an amplitude (Bi) of about 0.5 to about 3 pT. In certain embodiments, the method comprises iii) exposing the sample to one or more RF pulses to obtain chemical exchange saturation transfer (CEST) data. In certain embodiments, the method comprises iv) generating one or more images from the CEST data at a plurality of frequency offsets.

[0194] - 28 -

[0195] 56933733.1

Claims

Attorney Docket No.: 047162-7396WO1 (02764)CLAIMSWhat is claimed is:

1. A method for magnetic resonance imaging, the method comprising: i) exposing a sample to a saturating radiofrequency (RF) pulse with a frequency that is off-resonance with bulk water and a corresponding Bo field; ii) exposing the sample to at least one sampling RF pulse having an amplitude (Bi) of about 0.5 to about 3 pT; iii) exposing the sample to one or more RF pulses to obtain chemical exchange saturation transfer (CEST) data; and iv) generating one or more images from the CEST data at a plurality of frequency offsets.

2. The method of claim 1, wherein step iii) comprises simultaneously obtaining CEST data for glucose and glycogen protons in the sample.

3. The method of claim 1, wherein the sampling RF pulse has a Bi amplitude of about 0.5, 1, 1.5, or 3 pT.

4. The method of claim 3, wherein the image is generated at magnetic resonance frequencies of at least one of +2.1 , +1 .2, +0.7, and -1.2 ppm relative to the bulk water resonance.

5. The method of claim 1, wherein the images are acquired in a magnetic field strength of at least 11.7 T.

6. The method of claim 2, further comprising computing an absolute pH in a region of the sample using a ratiometric equation:[Mz(0.7ppm)]x[Mz(8 ppm)-Mz (2.1 ppm)]GGCII[Mz(2.1 ppm)]x[Mz(8 ppm)-Mz(0.7 ppm)]7. The method of claim 6, wherein a pH map is generated using a plurality of absolute pH values.- 29 -56933733.1Attomey Docket No.: 047162-7396WO1 (02764)8. The method of claim 6, wherein the absolute pH is determined at a Bi amplitude of 0.5 and 1.5 pT.

9. The method of claim 1, wherein the sample comprises in vitro or in vivo tissue.

10. The method of claim 9, wherein the sample comprises cancerous tissue.

11. A system for performing chemical exchange saturation transfer (CEST) magnetic resonance imaging (MRI), the system comprising: providing computer coupled to an MRI scanner adapted to image a sample, wherein the computer comprises: an imaging controller; a computer processor; and a non-transitory computer-readable memory storage that stores instructions executable by the computer processor, wherein executing the instructions performs the steps of: i) exposing a sample to a saturating radiofrequency (RF) pulse with a frequency that is off-resonance with bulk water and a corresponding Bo field; ii) exposing the sample to at least one sampling RF pulse having an amplitude (Bi) of about 0.5 to about 3 pT; iii) exposing the sample to one or more RF pulses to obtain chemical exchange saturation transfer (CEST) data; and iv) generating one or more images from the CEST data at a plurality' of frequency offsets.

12. The system of claim 11, wherein the computer processor is configured to cause the MRI scanner to simultaneously obtain CEST data for glucose and glycogen protons in the sample.

13. The system of claim 11, wherein the computer processor is configured to cause the MRI scanner to apply a sampling RF pulse with a Bi amplitude of 0.5, 1, 1.5, or 3 pT.- 30 -56933733.1Attorney Docket No.: 047162-7396WO1 (02764)14. The system of claim 13, wherein the computer processor is configured to cause the computer and MRI scanner to generate the image at magnetic resonance frequencies of at least one of +2.1, +1.2, +0.7, and -1.2 ppm relative to the bulk water resonance.

15. The system of claim 11, wherein the computer processor is configured to cause the MRI scanner to acquire the images at a magnetic field strength of at least 11 T.

16. The system of claim 11, wherein the computer processor is configured to cause the computer to compute an absolute pH in a region of the sample using the equation17. The system of claim 16, wherein the computer processor is configured to cause the computer to generate a pH map using a plurality of absolute pH values.

18. The system of claim 16, wherein the computer processor is configured to cause computer to determine the absolute pH at a Bi amplitude of 0.5 and 1.5 pT.

19. The system of claim 11, wherein the sample comprises in vitro or in vivo tissue.

20. The system of claim 19, wherein the sample comprises cancerous tissue.

21. A storage medium for storing instructions executable by a computer processor to perform a method comprising: i) exposing a sample to a saturating radiofrequency (RF) pulse with a frequency that is off-resonance with bulk water and a corresponding Bo field; ii) exposing the sample to at least one sampling RF pulse having an amplitude (Bi) of about 0.5 to about 3 pT; iii) exposing the sample to one or more RF pulses to obtain chemical exchange saturation transfer (CEST) data; and iv) generating one or more images from the CEST data at a plurality of frequency offsets.- 31 -56933733.1

Images