Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Mar 20.
Published in final edited form as: Transl Res. 2026 Feb 18;289:40–46. doi: 10.1016/j.trsl.2026.02.008

Mapping glycogen accumulation and treatment effect in Pompe disease with saturation transfer MRI

Qing Zeng 1,2, Yuguo Li 1,2, Derek Timm 3, Tyler Johnson 3, Nickita Mehta 3, Lukas Martin 3, Brian Fox 3, Peter C M van Zijl 1,2, Glenn A Walter 4, Jens T Rosenberg 5, Craig W Vander Kooi 6,7, Manuela Corti 8, Matthew S Gentry 6,7, Ramon C Sun 6,7, Barry J Byrne 8, Nirbhay N Yadav 1,2,*
PMCID: PMC13001076  NIHMSID: NIHMS2151322  PMID: 41720458

Abstract

Pompe disease is a glycogen storage disease caused by the impaired breakdown of glycogen in lysosomes, leading to abnormal glycogen accumulation in tissue. Here we use glycogen nuclear Overhauser effect (glycoNOE) MRI to detect glycogen levels in skeletal muscle in a mouse model of Pompe disease. Moreover, we evaluated if glycoNOE MRI could detect changes in glycogen load after enzyme replacement therapy. The results show that glycoNOE MRI can distinguish between Pompe mice and wildtype controls. Furthermore, the technique detected treatment-dependent changes in muscle glycoNOE signals, which were validated with ex vivo biochemical assays. To demonstrate potential human translation, glycoNOE MRI was applied to two Pompe patients and revealed elevated glycogen levels in patients compared to healthy controls.

Introduction

Glycogen is an important energy reserve in mammals and plays a central role in glucose homeostasis.1 Major deposits of glycogen are in skeletal muscle and liver, although glycogen is present in many other tissue types such as the kidney, heart, and brain.2 The role of glycogen in fueling muscle contraction and providing glucose for export into the bloodstream has been well characterized.1–3 More recently, it has emerged that the role of glycogen in the body is far more complex and it is involved in learning and memory4, neurological disorders 5, and tumor proliferation6–8.

The synthesis and breakdown of glycogen involves the coordinated action of multiple enzymes.9 Synthesis can occur from glucose, after it is transported into the cell from the bloodstream, or via gluconeogenic precursors such as lactate (the Cori cycle) or alanine (the glucose–alanine cycle).10 Glycogen breakdown takes place both in the cytoplasm and inside the lysosomes. In the cytosol, glycogen is catabolized into glucose via glycogen phosphorylase and glycogen debranching enzyme. In lysosomes, glycogen is degraded by the enzyme acid alpha-glucosidase (also known as acid maltase) that releases glucose. Perturbations in these pathways can lead to diseases with severe and devastating symptoms.9, 11

Glycogen storage diseases (GSDs) are a broad group of inherited diseases where the body’s ability to synthesize or break down glycogen is impaired by the loss-of-function mutation of relevant genes, leading to aberrant glycogen structure and accumulation.9 Depending on the gene and pathway affected, aberrant glycogen metabolism may be observed in different tissue types such as skeletal muscle, liver, heart, and brain. Among them, Pompe disease, also known as glycogen storage disease type II, is a rare inherited metabolic disease caused by mutations in the GAA gene leading to a functional deficiency of lysosomal enzyme acid alpha-glucosidase. Patients with Pompe disease usually exhibit excess glycogen accumulation in multiple tissues, particularly in skeletal and cardiac muscle as well as in the central nervous system, leading to progressive motor, cardiac and respiratory dysfunction.9, 12–14 Pompe disease is broadly classified into two main phenotypes: infantile-onset Pompe disease (IOPD) and late-onset Pompe disease (LOPD). IOPD results from complete or almost complete deficiency of acid alpha-glucosidase and can lead to death within the first few years of life if left untreated, while LOPD is characterized by onset of symptoms later in life and has less severe clinical outcomes.15, 16 Currently, there is no cure for Pompe disease. Enzyme replacement therapy (ERT) with recombinant human GAA (rhGAA) is the only approved disease-modifying treatment modality for Pompe disease and slows disease progression.14, 16 Current clinical methods for tracking the efficacy of ERT in clearing skeletal muscle glycogen include muscle biopsy12, 17 for biochemical glycogen quantitation and histopathological assessment of cellular and lysosomal glycogen. However, biopsy carries risk to patients and has sampling bias if the glycogen distribution is nonuniform. A noninvasive method with high detection sensitivity and spatial resolution is needed for the comprehensive assessment of Pompe disease.

Magnetic Resonance Imaging (MRI) is highly suited for monitoring treatment response due to its noninvasiveness, allowing multiple repetitive examinations, relatively high spatial resolution (millimeter scale), and wide clinical availability. Previous MRI studies monitoring the effects of Pompe disease in skeletal muscle have relied on T1-weighted structural imaging, which can clearly demarcate the loss of muscle fibers and its replacement by fat, a key hallmark of Pompe disease.16, 18 An important limitation is that structural imaging, such as T1-weighted MRI, shows the effects of the disease and not the main pathologic feature of glycogen accumulation within the cells. 13C magnetic resonance spectroscopy (MRS) methods can quantify glycogen in vivo19, 20 but is hampered by low sensitivity and the need for specialized RF coils that are not commonly used in clinical settings.

Saturation transfer MRI has proven to be a useful tool to detect metabolite levels in vivo with enhanced sensitivity by taking advantage of the coupling between metabolite and water protons.21 Saturation transfer MRI detection of glycogen using standard MRI equipment was proposed almost two decades ago using chemical exchange saturation transfer (CEST). CEST detection of glycogen (glycoCEST) relies on the magnetic labeling (“saturation”) of hydroxyl protons in glycogen and then observing the direct transfer of these labeled protons to bulk water.22 High transfer rates of glycogen hydroxyl protons result in broad glycoCEST signals that are difficult to quantify and, consequently, its application in vivo has been limited. More recently, glycogen nuclear Overhauser effect (glycoNOE) MRI was proposed to detect glycogen through the magnetic coupling between glycogen aliphatic protons and water protons via relayed NOE (rNOE).23, 24 As the rNOE transfer originates from irradiating the narrow aliphatic signal24 and the overall transfer rate is determined by its slowest step (NOE part), efficient labeling and signal detection can be achieved at lower B1 values with less background interference from direct water saturation and its intensity is easier to quantify at lower magnetic fields. rNOE’s are one of several signal sources in saturation transfer experiments where the MRI water signal is measured as a function of the saturation frequency, creating a so-called Z-spectrum.25 This spectrum, analogous to an MRS spectrum, contains signals from different molecular origins. It was recently demonstrated that glycoNOE MRI can assess hepatic glycogen and muscle glycogen in vivo in mouse models of glycogen storage disease type III 26 and Pompe disease27 respectively. It was also used to monitor human muscle glycogen change post-exercise.28 Here we used GAA knockout mice29 as a mouse model of Pompe disease and employed glycoNOE MRI to detect glycogen levels in skeletal muscle. The efficacy of different treatments on the mouse model was also investigated. Finally, we demonstrate human translation by imaging two Pompe disease patients and compare the images to controls.

Materials and Methods

Animal preparation

All animal studies were performed with the approval of and in accordance with Johns Hopkins University Animal Care and Use Committee guidelines. GAA knockout (Gaatml −/−) and wildtype littermate mice on 129/Sv background with mixed sex were used in this study.

Ten-week-old mice were assigned to four separate groups and treated with ERT or vehicle every two weeks for eight weeks (4 doses) prior to undergoing MRI scans. The wild-type (WT) control group (n = 8) and the GAA knockout (KO) control group (group a, n = 7) were administered a vehicle (0.9% saline) via both intravenous (IV) injection and oral gavage (OG). The remaining two GAA KO groups received either 20 mg/kg alglucosidase alfa IV with OG vehicle (group b, n=7) or 20 mg/kg cipaglucosidase alfa IV combined with 10 mg/kg miglustat OG (group c, n=8). Additionally, all four groups were pre-treated with 10 mg/kg diphenhydramine via intraperitoneal injection 15 minutes before each IV dose, starting prior to the second dose, in order to avoid any potential anaphylactic reaction.

Before the MRI scan, the mice were fasted for between 8 and 16 hours. Anesthesia for the mice was induced by 3% isoflurane and maintained by 1–1.5% isoflurane during the scan. The mouse RF coil was positioned around rear legs to scan quadriceps, triceps and gastrocnemius.

Human study

Human studies were approved by the Johns Hopkins Medicine Institutional Review Board. Two Pompe disease patients (25-year-old female with clinical characteristics: 6-minute walk test – 300 m, forced vital capacity – 3.5 L 91% predicted, and on ERT for 15 years; and a 39-year-old male with clinical characteristics of 6-minute walk test – 450 m, forced vital capacity – 3.7 L 95% predicted, and on ERT for 10 years) and a healthy control (35-year-old male) were scanned after informed consent was obtained.

MRI experiment

Preclinical MRI experiments were conducted on an 11.7 T MRI scanner (Bruker Biospec). Multi-slice T2-weighted images were acquired across the rear legs of each mouse and used as a reference for selecting slices containing quadriceps, triceps, and gastrocnemius respectively from rear legs. Each slice was acquired separately using a single-slice steady-state ultrashort-spin-echo saturation-transfer (UTE-ST) pulse sequence with radial acquisition.30 During each spoke, a Gaussian-shaped saturation pulse (20 ms, 0.7 μT) was applied, followed by the UTE image readout. The repetition time (TR) was 30 ms. The effective echo time was about 0.3 ms. The number of spokes was 302, resulting in a scan time of 12 s for each offset. A total of 115 saturation offsets were acquired (200, 200, 200, −8, −7, −6, −5, −4.9 −4.8, −4.7, …, 0, 200, 200, 200, 200, 8, 7, 6, 5, 4.9, 4.8, 4.7, …, 0 ppm) and the 7 saturation offsets at 200 ppm were used for S0. In-plane field of view (FOV) was 28×28 mm2 with a matrix size of 96×96 (in plane resolution 0.29×0.29 mm2), and slice thickness was 2 mm. The acquisition time for each glycoNOE scan was about 17 min and 22 s. High-resolution T2-weighted images (matrix size 256×256) at the same slice were also acquire for reference.

Human subjects were scanned on a 3 T Philips Elition RX system (Philips Healthcare). A multi-slice (10 slices) steady-state glycoNOE sequence was used consisting of 50 ms sinc-Gaussian pulses with a flip angle of 310° for saturation. The slice thickness was 10 mm. In-plane FOV was 150×150 mm2 (resolution 1.5×1.5 mm2) for calves, and 162.62×161.62 mm2 (resolution 2×2 mm2) or 171.44×171.44 mm2 (resolution 2.5×2.5 mm2) for thighs. Forty saturation frequencies from 10 to −10 ppm with uneven intervals were used (10, 5, 4, 3, 2, 1.5, 1, 0.9, 0.8, …, −0.8, −0.9, −1, −1.1, −1.2, −1.4, −1.6, −1.8, −2, −2.2, −2.6, −3, −3.5, −4, −5, −10 ppm), and 3 extra frequencies at 100 ppm were used for S0.

Ex vivo quantitation

All mice were euthanized after their MRI scan. Then the quadriceps, triceps and gastrocnemius were excised from left rear leg immediately and snap frozen. The frozen tissues were stored at −80 °C until analysis.

Glycogen contents were measured by a biochemical assay. Tissue samples were homogenized to a concentration of 100 mg/mL in water. In duplicate, homogenized tissues and suitability controls (glycogen standards) were acidified with 4N trifluoroacetic acid (TFA) and hydrolyzed at 100°C for 4 hours. At the completion of the acid-heat hydrolysis, the dried samples were resuspended in degassed deionized water, vortexed and transferred to high performance liquid chromatography (HPLC) vials. Ten microliters of each sample were injected onto a Dionex CarboPac PA 10 2 × 250 mm analytical column equipped with a 2 × 50 mm guard column and analyzed by high pH anion exchange chromatography with pulsed amperometric detection (HPAEC-PAD) along with two sets of calibration standards and quality control samples prepared in water at eight and three concentration levels, respectively, using a 5.5 millimolar glucose standard stock solution. Raw data was extracted using Chromeleon software (Thermo Fisher). Glycogen concentrations were calculated in excel using the extracted raw data and normalized to protein data (μg/mg of protein) determined by BCA assay.

Data analysis

Z-spectra in each voxel were interpolated to 512 points and corrected for B0 field inhomogeneity by fitting the water direct saturation and assigning the minimum to 0 ppm. Principal component analysis (PCA)-based denoising method was applied prior to spectral fitting.31 The Z-spectra were then fitted using a Voigt and polynomial hybrid lineshape model based on R1ρ relaxation theory.32 A two-step fitting was employed to the Z-spectrum between −0.3 and −2 ppm. First, the background was fitted by a Lorentzian and a linear function using data points excluding the glycoNOE peak (−0.7 ~ −1.3 ppm). Then the background parameters were fixed and the glycoNOE peak was fitted by a Voigt lineshape function. The glycoNOE signal was estimated from the fitted ΔZ (the fitted background minus the fitted Z spectrum). Statistical significance was evaluated using an unpaired t test. Regions of interest (ROIs) were drawn manually for quadriceps, triceps and gastrocnemius respectively using the high-resolution T2-weighted image in reference to the mouse anatomy. The high-resolution image and glycoNOE image were co-registered and then the ROI masks were transformed to the glycoNOE image space for quantitative analysis. Z-spectral signals were averaged within the ROI before glycoNOE signal estimation.

Results

Figure 1A shows averaged Z-spectra of triceps for the wildtype and GAA knockout control groups. A clear signal difference is observed at approximately −1 ppm, which corresponds to the glycoNOE peak. A much smaller signal difference is observed at around +1 ppm corresponding to the glycogen CEST (glycoCEST) signal22, which is small for this low-B1 pulse sequence. Figure 1B shows representative upfield Z-spectral regions from the triceps of a wildtype mouse and a GAA knockout mouse, together with data fits. In the glycoNOE spectral range, the difference between the background fit and the data points is much larger in the GAA knockout mouse than in the WT. The residual Z-spectra (ΔZ spectra) after removing the fitted background and their fittings are shown in the bottom row and are used for glycogen estimation.

Figure 1.

Figure 1.

Open in a new tab

(A) Averaged Z-spectra of triceps for the wildtype control group (blue, n = 8) and GAA knockout control group (orange, n = 7). Error bars show the standard derivation within the group. Note, Z-spectra follow the standard nuclear magnetic resonance (NMR) convention for displaying frequency with decreasing values from left to right. The largest component (bulk water protons) are used as a reference and assigned to 0 ppm. The glycoNOE signal originates from aliphatic protons in glycogen (centered at −1 ppm, right-hand side of water) and the glycoCEST signal originates from hydroxyl protons in glycogen (approximately +1 ppm, left-hand side of water). (B) Representative upfield sections of the Z-spectra for a wildtype mouse (left) and a GAA knockout mouse (right) and corresponding fittings. The bottom shows the residual Z spectra after removing the fitting background and the fitted ΔZ spectra.

Figure 2 shows representative glycoNOE maps of muscles from each slice obtained by fitting Z-spectral data voxel by voxel. Varying glycoNOE contrast levels can be seen for different mouse groups. The GAA knockout control (group a) treated with placebo has the highest glycoNOE contrast, while the WT control treated with placebo has the lowest glycoNOE contrast. The other two groups were treated with different drugs and show treatment-dependent contrasts.

Figure 2.

Figure 2.

Open in a new tab

Representative glycoNOE maps of muscles from three slices, overlaid on the T2-weighted images. The regions of interest (ROIs) for quadriceps (top), triceps (middle) and gastrocnemius (bottom) are show in red outlines. Treatments for GAA knockout mice: group a (placebo), group b (alglucosidase alfa), group c (cipaglucosidase alfa plus miglustat).

GlycoNOE signals from region of interest (ROI) in different muscle types are shown in Figure 3. ROIs were drawn (red outlines in Figure 2) within the quadriceps, triceps and gastrocnemius in slices 1, 2, and 3 respectively to determine GlycoNOE signals, as shown in Figure 3. For all muscles, glycoNOE signals in GAA KO control mice (group a) are significantly higher than the WT control. For quadriceps, both treated groups (group b-c) show significantly lower glycoNOE signals compared to GAA KO control. For triceps and gastrocnemius, group c has a significant reduction of glycoNOE signals compared to GAA KO control. Generally, group c shows lower glycoNOE signals compared to group b, indicating better treatment efficacy. For validation, an ex vivo biochemical assay was conducted for glycogen quantitation. These ex vivo results confirm the presence of higher glycogen contents in GAA KO mice compared to WT control and reduced glycogen levels due to the treatments. There is a good linear correlation between glycoNOE signal and ex vivo measurement of glycogen content, as shown in Figure 3B.

Figure 3.

Figure 3.

Open in a new tab

GlycoNOE signals (A) and the correlation between glycoNOE signals and glycogen contents obtained from ex vivo biochemical assay (B) for the three different muscle types. Statistical analysis is versus GAA knockout control (group a). **, p < 0.05; ***, p < 0.005

To demonstrate the potential for translation to human patients, leg muscles from two LOPD patients were scanned and the glycoNOE contrast compared to a healthy control. Structurally, the most prominent difference between the LOPD patients and control is the much larger fraction of fat in patients (Figure 4). Voxels containing primarily fat were masked out and the remaining tissue mapped for glycoNOE signal. The glycoNOE maps show much higher contrast in LOPD patients (Figure 4 B and D) compared to the healthy subject in both calves and thighs.

Figure 4.

Figure 4.

Open in a new tab

GlycoNOE maps for the control (left column) and Pompe patients (right column) in the calves (A-B) and thighs (C-D). Clinical characteristics for the patient in B are: 6-minute walk test (6MWT) – 300 m, forced vital capacity (FVC) – 3.5 L 91% predicted, and on ERT for 15 years. The patient in D has a 6MWT – 450 m, FVC – 3.7 L 95% predicted, and on ERT for 10 years.

Discussion

In Pompe disease, GAA enzyme deficiency leads to the lysosomal accumulation of glycogen. ERT with recombinant human GAA has been approved since 2006 and has been shown to slow the progression of the disease33–35, but it is not a cure. Patients continue to develop muscle weakness and recent studies have shown symptoms in the central nervous system.15 Non-invasive imaging is critical to monitoring disease progression and treatment efficacy. Fiber degeneration and fat replacement in muscle is a classic feature of the disease. The replacement of muscle by fat is the basis for current MRI methods for characterizing this disease, but this occurs in the latter stages of the disease when there is permanent muscle weakness. Further, MRI imaging of fat infiltration is an indirect measure of disease load and does not directly report on the main pathologic feature of the disease, which is the glycogen accumulation. GlycoNOE MRI reports on glycogen content directly and thus may be more sensitive to disease progression at early timepoints. Here, we showed that glycoNOE MRI can distinguish between Pompe disease and controls in mice and in humans. Our preclinical results show that glycoNOE MRI can distinguish between GAA KO (Pompe) mice and WT controls, treated mice showed reduced glycoNOE contrast. We found that glycoNOE MRI method is able to distinguish between the efficacy of different types of treatments.

Ex vivo biochemical assays confirmed increased glycogen content in GAA knockout mice compared to wildtype mice and reduced glycogen levels after treatment. The glycoNOE signals are in good agreement with the ex vivo results but not perfect. Several reasons may be contributing to this discrepancy including the imperfect matching of the imaging volume and excised tissue. Muscle excised for ex vivo analysis was almost whole muscle for each muscle type, while the glycoNOE scan covered only one slice. Another source for variability is fiber type. Skeletal muscle comprises of slow-twitch oxidative fibers, fast-twitch oxidative/glycolytic fibers, and fast-twitch glycolytic fibers with different glycogen utilization.36 The imaging slice used in our animal study may have bias in fiber types. This issue can be alleviated in future studies by using multi-slice or 3D image acquisitions that include the whole muscle of interest, but it would also increase acquisition time, and development of faster approaches is needed to accomplish this. Moreover, glycogen metabolism during the time between the MRI scan and the tissue extraction can affect the glycogen content, and it has also been reported that glycogen rapidly degraded postmortem37, 38. These combined factors can affect the ex vivo quantitation of glycogen content. The glycoNOE signals in gastrocnemius show larger variation compared to quadriceps and triceps. We believe this is likely due to the much smaller size of the muscle and ROI in the acquired slices used in our analysis. We also noted that there was much larger B0 field variation across the slice containing the gastrocnemius.

Another important result to highlight is the nonzero intercept in the correlation analysis. This indicates that the MRI signal likely contains contributions from other components such as mobile proteins, lipids, and other polysaccharides. It has been reported that there are alterations in protein and lipid metabolism in Pompe disease that may progress independently of glycogen burden.39 These factors combined with glycogen degradation postmortem could be contributing to the baseline shift. We have tried to minimize these confounding factors by using a fitting algorithm that separates the glycogen signal from other signal components, but some overlap may persist.

GlycoNOE maps were heterogenous across the muscle tissue, which was more clearly seen in the human LOPD patients. It has been reported that glycogen content is nonuniform in skeletal muscle fiber and its distribution depends on several factors such as muscle fiber function.40 Another reason for this heterogeneity could be fat replacement and fiber degeneration in muscle. It was also reported that IOPD patients had higher glycogen levels compared to LOPD patients.17 Still, studies on additional LOPD and IOPD patients are required for more robust statistical analysis and investigation of the effects of fat and fibrous tissue on glycoNOE signal characterization.

As muscle biopsy carries risk in an invasive way and suffer from sampling bias, it is not a practical way to monitor disease load and treatment effect. In contrast, glycoNOE is noninvasive and can be used for long-term trials. Its capability of assessing treatment efficacy can also facilitate new drug development for Pompe disease and other GSDs.

Conclusion

We demonstrate that glycoNOE MRI can be used to detect glycogen levels in a mouse model of Pompe disease compared to wildtype mice and, for the first time, show that glycoNOE is sensitive to treatment in the mouse model. The potential for human translation was demonstrated in two LOPD patients, who had higher glycoNOE contrast compared to healthy controls. These results indicate that glycoNOE has potential to monitor disease load due to glycogen accumulation and the efficacy of treatments in Pompe disease. Due to the ability to detect elevated glycogen levels, we expect it to also be applicable to a wide variety of glycogen storage diseases.

Acknowledgment

This research was supported by the National Institutes of Health grants (R01 NS127280 and P41 EB031771) and Amicus Therapeutics Inc. The MRI equipment used in this study was funded by NIH grant S10OD032188.

Footnotes

Disclosure of conflicts of interest

Derek Timm, Tyler Johnson, Nickita Mehta, Lukas Martin, and Brian Fox are employees in Amicus Therapeutics Inc. Nirbhay N. Yadav and Peter C. M. van Zijl are patent holders for the glycoNOE technology.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Reference

  • (1).Adeva-Andany MM; González-Lucán M; Donapetry-García C; Fernández-Fernández C; Ameneiros-Rodríguez E Glycogen metabolism in humans. BBA Clinical. 2016;5:85–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (2).Roach Peter J; Depaoli-Roach Anna A; Hurley Thomas D; Tagliabracci Vincent S. Glycogen and its metabolism: some new developments and old themes. Biochem J. 2012;441:763–787. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (3).Hargreaves M; Spriet LL. Skeletal muscle energy metabolism during exercise. Nature Metabolism. 2020;2:817–828. [DOI] [PubMed] [Google Scholar]
  • (4).Alberini CM; Cruz E; Descalzi G; Bessières B; Gao V Astrocyte glycogen and lactate: New insights into learning and memory mechanisms. Glia. 2018;66:1244–1262. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (5).Duran J; Guinovart JJ. Brain glycogen in health and disease. Mol Aspects Med. 2015;46:70–77. [DOI] [PubMed] [Google Scholar]
  • (6).Curtis M; Kenny HA; Ashcroft B; Mukherjee A; Johnson A; Zhang Y; Helou Y; Batlle R; Liu X; Gutierrez N; et al. Fibroblasts Mobilize Tumor Cell Glycogen to Promote Proliferation and Metastasis. Cell Metab. 2019;29:141–155.e149. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (7).Young LEA; Conroy LR; Clarke HA; Hawkinson TR; Bolton KE; Sanders WC; Chang JE; Webb MB; Alilain WJ; Vander Kooi CW; et al. In situ mass spectrometry imaging reveals heterogeneous glycogen stores in human normal and cancerous tissues. EMBO Mol Med. 2022;14:e16029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (8).Clarke HA; Hawkinson TR; Shedlock CJ; Medina T; Ribas RA; Wu L; Liu Z; Ma X; Xia Y; Huang Y; et al. Glycogen drives tumour initiation and progression in lung adenocarcinoma. Nature Metabolism. 2025;7:952–965. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (9).Hannah WB; Derks TGJ; Drumm ML; Grünert SC; Kishnani PS; Vissing J. Glycogen storage diseases. Nature Reviews Disease Primers. 2023;9:46. [DOI] [PubMed] [Google Scholar]
  • (10).Radziuk J; Pye S Hepatic glucose uptake, gluconeogenesis and the regulation of glycogen synthesis. Diabetes Metab Res Rev. 2001;17:250–272. [DOI] [PubMed] [Google Scholar]
  • (11).Özen H Glycogen storage diseases: new perspectives. World J Gastroenterol. 2007;13:2541. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (12).Kohler L; Puertollano R; Raben N Pompe Disease: From Basic Science to Therapy. Neurotherapeutics. 2018;15:928–942. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (13).Kishnani PS; Steiner RD; Bali D; Berger K; Byrne BJ; Case LE; Crowley JF; Downs S; Howell RR; Kravitz RM; et al. Pompe disease diagnosis and management guideline. Genet Med. 2006;8:267–288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (14).Cabello JF; and Marsden D Pompe disease: clinical perspectives. Orphan Drugs: Research and Reviews. 2017;7:1–10. [Google Scholar]
  • (15).Korlimarla A; Lim J-A; Kishnani PS; Sun B An emerging phenotype of central nervous system involvement in Pompe disease: from bench to bedside and beyond. Annals of Translational Medicine. 2019;7:289. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (16).Díaz-Manera J; Walter G; Straub V Skeletal muscle magnetic resonance imaging in Pompe disease. Muscle Nerve. 2021;63:640–650. [DOI] [PubMed] [Google Scholar]
  • (17).Schoser B; Raben N; Varfaj F; Walzer M; Toscano A Acid α-glucosidase (GAA) activity and glycogen content in muscle biopsy specimens of patients with Pompe disease: A systematic review. Mol Genet Metab Rep. 2024;39:101085. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (18).Burakiewicz J; Sinclair CDJ; Fischer D; Walter GA; Kan HE; Hollingsworth KG. Quantifying fat replacement of muscle by quantitative MRI in muscular dystrophy. J Neurol. 2017;264:2053–2067. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (19).Gruetter R; Magnusson I; Rothman DL; Avison MJ; Shulman RG; Shulman GI. Validation of 13C NMR measurements of liver glycogen in vivo. Magn Reson Med. 1994;31:583–588. [DOI] [PubMed] [Google Scholar]
  • (20).Harris T; Mishra K; Vaknin H; Friedman-Ezra A; Weil M; Rosenmann H 13C-MR as a Diagnostic and Non-invasive Tool for Quantification of Brain Glycogen in Adult Polyglucosan Body Disease (APBD): A Pilot Case Study in Mouse. Journal of Case Reports in Medicine. 2022;10:13–19. [Google Scholar]
  • (21).van Zijl PCM; Yadav NN. Chemical exchange saturation transfer (CEST): What is in a name and what isn’t? Magn Reson Med. 2011;65:927–948. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (22).van Zijl PCM; Jones CK; Ren J; Malloy CR; Sherry AD. MRI detection of glycogen in vivo by using chemical exchange saturation transfer imaging (glycoCEST). Proc Natl Acad Sci U S A. 2007;104:4359–4364. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (23).Zhou Y; van Zijl PCM; Xu X; Xu J; Li Y; Chen L; Yadav NN. Magnetic resonance imaging of glycogen using its magnetic coupling with water. Proc Natl Acad Sci U S A. 2020;117:3144–3149. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (24).Zhou Y; van Zijl PCM; Xu J; Yadav NN. Mechanism and quantitative assessment of saturation transfer for water-based detection of the aliphatic protons in carbohydrate polymers. Magn Reson Med. 2021;85:1643–1654. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (25).Zhou Y; Bie C; van Zijl PCM; Yadav NN. The relayed nuclear Overhauser effect in magnetization transfer and chemical exchange saturation transfer MRI. NMR Biomed. 2023;36:e4778. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (26).Zeng Q; Machado M; Bie C; van Zijl PCM; Malvar S; Li Y; D’souza V; Poon KA; Grimm A; Yadav NN. In vivo characterization of glycogen storage disease type III in a mouse model using glycoNOE MRI. Magn Reson Med. 2024;91:1115–1121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (27).Bie C; Bo S; Yadav NN; van Zijl PCM; Wang T; Chen L; Xu J; Zou C; Zheng H; Zhou Y Simultaneous monitoring of glycogen, creatine, and phosphocreatine in type II glycogen storage disease using saturation transfer MRI. Magn Reson Med. 2025;n/a:1782–1792. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (28).Bie C; Ma Y; van Zijl PCM; Yadav NN; Xu X; Zheng H; Liang D; Zou C; Areta JL; Chen L; et al. In vivo imaging of glycogen in human muscle. Nature Communications. 2024;15:10826. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (29).Raben N; Nagaraju K; Lee E; Kessler P; Byrne B; Lee L; LaMarca M; King C; Ward J; Sauer B; et al. Targeted Disruption of the Acid α-Glucosidase Gene in Mice Causes an Illness with Critical Features of Both Infantile and Adult Human Glycogen Storage Disease Type II*. J Biol Chem. 1998;273:19086–19092. [DOI] [PubMed] [Google Scholar]
  • (30).Chen L; Wei Z; Chan KWY; Cai S; Liu G; Lu H; Wong PC; van Zijl PCM; Li T; Xu J Protein aggregation linked to Alzheimer’s disease revealed by saturation transfer MRI. Neuroimage. 2019;188:380–390. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (31).Breitling J; Deshmane A; Goerke S; Korzowski A; Herz K; Ladd ME; Scheffler K; Bachert P; Zaiss M Adaptive denoising for chemical exchange saturation transfer MR imaging. NMR Biomed. 2019;32:e4133. [DOI] [PubMed] [Google Scholar]
  • (32).Chen L; Zeng H; Xu X; Yadav NN; Cai S; Puts NA; Barker PB; Li T; Weiss RG; van Zijl PCM; et al. Investigation of the contribution of total creatine to the CEST Z-spectrum of brain using a knockout mouse model. NMR Biomed. 2017;30:e3834. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (33).Ploeg ATvd; Clemens PR; Corzo D; Escolar DM; Florence J; Groeneveld GJ; Herson S; Kishnani PS; Laforet P; Lake SL; et al. A Randomized Study of Alglucosidase Alfa in Late-Onset Pompe’s Disease. N Engl J Med. 2010;362:1396–1406. [DOI] [PubMed] [Google Scholar]
  • (34).van der Ploeg AT; Barohn R; Carlson L; Charrow J; Clemens PR; Hopkin RJ; Kishnani PS; Laforêt P; Morgan C; Nations S; et al. Open-label extension study following the Late-Onset Treatment Study (LOTS) of alglucosidase alfa. Mol Genet Metab. 2012;107:456–461. [DOI] [PubMed] [Google Scholar]
  • (35).Kishnani PS; Corzo D; Nicolino M; Byrne B; Mandel H; Hwu WL; Leslie N; Levine J; Spencer C; McDonald M; et al. Recombinant human acid α-glucosidase. Neurology. 2007;68:99–109. [DOI] [PubMed] [Google Scholar]
  • (36).Hokken R; Laugesen S; Aagaard P; Suetta C; Frandsen U; Ørtenblad N; Nielsen J Subcellular localization- and fibre type-dependent utilization of muscle glycogen during heavy resistance exercise in elite power and Olympic weightlifters. Acta Physiologica. 2021;231:e13561. [DOI] [PubMed] [Google Scholar]
  • (37).Xu H; Stapleton D; Murphy RM. Rat skeletal muscle glycogen degradation pathways reveal differential association of glycogen-related proteins with glycogen granules. J Physiol Biochem. 2015;71:267–280. [DOI] [PubMed] [Google Scholar]
  • (38).Rauckhorst AJ; Borcherding N; Pape DJ; Kraus AS; Scerbo DA; Taylor EB. Mouse tissue harvest-induced hypoxia rapidly alters the in vivo metabolome, between-genotype metabolite level differences, and (13)C-tracing enrichments. Mol Metab. 2022;66:101596. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • (39).Sidorina A; Catesini G; Levi Mortera S; Marzano V; Putignani L; Boenzi S; Taurisano R; Garibaldi M; Deodato F; Dionisi-Vici C Combined proteomic and lipidomic studies in Pompe disease allow a better disease mechanism understanding. J Inherit Metab Dis. 2021;44:705–717. [DOI] [PubMed] [Google Scholar]
  • (40).Ørtenblad N; Nielsen J Muscle glycogen and cell function – Location, location, location. Scand J Med Sci Sports. 2015;25:34–40. [DOI] [PubMed] [Google Scholar]

RESOURCES