Citation
Human Calf Muscle Metabolites Evaluated by 31Phosphorous-Magnetic Resonance Spectroscopy as Biomarkers in Duchenne Muscular Dystrophy: Assessing the Reproducibility of Non-localized and Multivoxel Sequences

Material Information

Title:
Human Calf Muscle Metabolites Evaluated by 31Phosphorous-Magnetic Resonance Spectroscopy as Biomarkers in Duchenne Muscular Dystrophy: Assessing the Reproducibility of Non-localized and Multivoxel Sequences
Creator:
Rasmussen, Hannah Camille
Publication Date:
Language:
English

Notes

Abstract:
Duchenne Muscular Dystrophy (DMD) is an X-linked recessive disorder affecting males. It causes progressive muscle weakness, engendering early loss of ambulation and culminating in death. Many ongoing clinical trials show promise in finding a cure. Thus, there is a need for biomarkers which can detect drug efficacy and disease progression. 31Phosphorus-magnetic resonance spectroscopy (31P-MRS) has potential to be used as a sensitive, non-invasive biomarker of metabolic status in disease pathology. However, data concerning replicability of 31P values is lacking. Further, studies which observe muscle-specific metabolites are scarce. In this study, a 31P-MRS protocol was implemented to investigate metabolites in calves of control subjects using non-localized and multivoxel sequences. The objective was to determine whether 31P spectra showed day-to-day reproducibility and which of the indices were the most sensitive across subjects. In this experimental study, 31P-MRS spectra were collected from calf muscles of 5 unaffected control subjects (4 adults mean age 24.3±6.5 years,1 boy age 5 years) using non-localized and multivoxel sequences. A 3T MR scanner with 31P circular surface coil (14cm diameter) was used to gather data at baseline and after 1 week. Spectra were graphed using jMRUI software as 7 peaks representing metabolite values. Metabolites measured included phosphocreatine (PCr), inorganic phosphate (Pi), ATP[(αATP+γATP+βATP)/3], and Ptot(Pi+PCr+αATP+γATP+βATP). The chemical shift between PCr and Pi was used to calculate pH. Variability was quantified by coefficient of variation (CV), and data compared using two-tailed t-tests. 31P data demonstrated acceptable day-to-day reproducibility in several indices of muscle metabolism in both non-localized and multivoxel scans. ATP/Ptot, PCr/Ptot, and pH showed the best day-to-day consistency in adult subjects and the child (CV<10%). Reproducibility in metabolite ratios was also noted across subjects. Values influenced by ßATP tended to associate with increased variability. For many values, reproducibility of multivoxel scans was similar (p>0.05) to reproducibility of non-localized.These findings characterized 31P-MRS indices in the calf by indicating that metabolite values, most notably those of ATP/Ptot, PCr/Ptot, and pH have potential to be sensitive and reproducible, both in multivoxel and non-localized sequences. ( en )
General Note:
Awarded Bachelor of Health Science, summa cum laude, on May 8, 2018. Major: Health Science. Emphasis/Concentration: General Health Sciences
General Note:
College or School: College of Public Health & Health Professions
General Note:
Advisor: Sean Forbes. Advisor Department or School: Physical Therapy

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Hannah Camille Rasmussen. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

UFDC Membership

Aggregations:
UF Undergraduate Honors Theses

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

Running head: HUMAN CALF MUSCLE METABOLITES EVALUATED BY 31 P MRS 1 Human Calf Muscle Metabolites Evaluated by 31 Phosphorous Magnetic Resonance Spectroscopy as Biomarkers in Duchenne Muscular Dystrophy: Assessing the Reproducibility of Non localized and Multivoxel Sequences Hannah C. Rasmussen University of Florida

PAGE 2

Human Calf Muscle Metabolites Evaluated by 31 P MRS 2 Abstract Background Duchenn e Muscular Dystrophy (DMD) is a n X linked recessive disorder affecting males. It cau ses progressive muscle weakness engendering early l oss of ambulation and culminating in death. Many ongoing clinical trials show promise in finding a cure. Thus, there is a need for biomarkers which can detect drug efficacy and disease progression. 31 Phosphorus magnetic resonance spectroscopy ( 31 P MRS) has potential to be used as a sensitive non invasive biomarker of metabolic status in disease pathology. However, data concerning replicability of 31 P values is lacking. Further, studies which observe muscle specific metabolites are scarce. In this study, a 31 P MRS protocol was implemented to investigate metabolites in calves of control subjects using non localized and multivoxel se quences The objective was to determ ine whether 31 P spectra showed day to day reproducibility and which of the indices were the most sensitive across subjects. Methods In this experimental study, 31 P MRS spectra were collected from calf muscles of 5 unaffected control s ubjects (4 adults mean age 24.36.5 years,1 boy age 5 years) using non localized and multivoxel sequences A 3T MR scanner with 31 P circular surface coil (14 cm diameter) wa s used to gather data at baseline and after 1 week. Spectra were graphed using jMRUI software as 7 peaks representing metabolite values Metabolites measured included phosphocreatine (PCr), inorga nic phosphate (P i and P tot (P i P i was used to calculate pH. Variability was quantified by coefficient of variation (CV ) and data compared using two tailed t tests Results 31 P data demonstrated acceptable day to day reproducibility in several indices of muscle metabolism in both non localized and multivoxel scans ATP/P tot PCr/P tot and pH showed the best day to day consistency in adult subjects and the child (CV<10%) Reproducibility in metabolite ratios was also noted across subjects. Values influenced by ATP

PAGE 3

Human Calf Muscle Metabolites Evaluated by 31 P MRS 3 tended t o associate with increased variability. For many values, r eproducibility of multivoxel scan s was similar (p> 0.05) to reproducibility of non localized. Conclusion These findings characterized 31 P MRS indices in the calf by indicating that metabolite values, most notably those of ATP/P tot PCr/P tot and pH have potential to be sensitive and reproducible, both in multivoxel and non localized sequences. Public Health and/or Health Professions Relevance 31 P MRS has the potential to qualify as a sensitive and n on invasive biomarker of DMD pathology. Funding NIH NIAMS : R01AR070101 Student Role Over the course of this study, Hannah Rasmussen took part in subject recruitment, data collection, image processing, and quantitative analysis.

PAGE 4

Human Calf Muscle Metabolites Evaluated by 31 P MRS 4 Human Calf Muscl e Metabolites Evaluated by 31 Phosphorous Magnetic Resonance Spectroscopy as Biomarkers in Duchenne Muscular Dystrophy: Assessing the Reproducibility of Non localized and Multivoxel Sequences Literature Review Duchenne M uscular Dystrophy (DMD) is a severe and progressive muscle wasting disease which impacts 1 in every 3,500 6,000 boys ( Mendell et al., 2012 ) DMD is the consequence of an X linked recessive mutation in the gene that codes for dystrophin such that the protein is either defective o r absent (Vohra et al., 2015). As dystrophin acts to link the myofibril cytoskeleton to the extracellular matrix, dystrophic muscle is vulnerable to damage from mechanical stress When satellite cells cannot maintain repair, rapidly damaged muscle is repla ced by fatty and fibrotic tissue (Latroche et al., 2015). Myocytic degeneration, regeneration and replacement are concurrent with inflammation, abnormalities in ionic homeostasis and energy metabolism, and decreased regenerative capabilities ( Reyngoudt, Turk, & Carlier, 2018) Pro gressive fatty infiltration and muscle weakness engender an early loss of ambulation, around the age of 8 12 years, and a reduced life expectancy due to cardiopulmonary complications, around the age of 20 30 years ( Deconinck & Da n 2007; Humbertclaude et al. 2012 ). A lthough there is no cure for DMD, upcoming therapeutic interventions such as exon skipping, utrophin upregulation, and stop codon suppression compounds show promise in compensating for the defective dystrophin coding gene (Willcocks et al., 2016). Corticosteroids, which aim to mitigate inflammation and thereby improve quality of life, are considered to be the current standard of care (Reinig, Mirzaei, & Berlau, 2017) T reatment is most beneficial when begun prior to severe muscle wasting, around the age of 5, to delay the onset of cardiac, orthopedic, and respiratory dysfunction (Forbes et al., 2014). Therefore, there is a need for sensitive, reliable biomarkers which detect disease progression even in young boys, and which are responsive to treatment in clinical trials

PAGE 5

Human Calf Muscle Metabolites Evaluated by 31 P MRS 5 Functional tests are the current canon of disease progression, preferred by many over invasive measures such as muscle biopsies (Willcocks et al., 2016) However, the FDA approved 6 Minute Walk Test has been criticized for having narrow inclusion criteria, for only showing disease progression in boys over the age of 7 and for having low inter operator reproducibility (Forbes et al., 2014). Recent studies have shown magnetic resonance imaging (MRI) and spectroscopy (MRS) to be sensitive and non invasive measures of inflammation and fat infiltration in muscles of young boys with DMD (Willcocks et al., 2016) These indices have displayed reproducibility when measured across multiple institutions, and over the span of several days ( Triplett et al., 2013 ) Fu rthermore, these MRI and MRS quantifications have been used t o detect the positive impact of corticosteroids on dystrophic muscle, validating their usefulness as indices for other treatments (Arpan et al ., 2014 ; Wary et al., 2015 ). However, many concur that there is still a need for a re plicable biomarker of DMD in young boys which is both able to characterize pathology in the muscle itself and able to discriminate disease status in spite o f corticosteroid use (Hoo ij man s et al., 2017) 31 Phosphorous magnetic resonance spectroscopy ( 31 P MRS) has been used to identify numerous aberrations in phosphate metabolite ratios in DMD research over the years These ratios of phosphate energetics and mem brane metabolites reflect disease anomalies at a cellular level (Wary et al., 2015). 31 P MRS principally observes spectra from myocytes, as bioenergetic values are almost undetectable in fat (Hooijman et al., 2017). A recent study by Wary et al. (2015) showed that metabolic indices were all exaggerated in more affected boys with DMD and were already altered relative to normal in younger boys with DMD Additionally, increased phosphodiester (PDE) ratios have been show n to reflect pathologic changes in dystrophic muscle as early as 5 years of age, independ ent of fatty infiltration (Hooijmans et al., 2017; Reyngoudt et al., 2018). Furthermore, the sensitivity of 31 P indices to therapeutic intervention has been assessed u sing Golden Retriever Muscular Dystrophy (GRMD) canines,

PAGE 6

Human Calf Muscle Metabolites Evaluated by 31 P MRS 6 the animal model most comparable to human DMD These studies have reported on improved phosphate metabolism values in the forelimbs of GRMD canines following exon skipping therapy ( Le Guiner et al., 2014; Le Guiner et al., 2017; Thibaud et al., 2012). Banerjee et al. (2010) showed that 31 P MRS measures were sensitive enough to detect short term changes in the calves of boys with DMD following oral creatine supplementation including in young boys under the age of 7. T hese results infer the use of 31 P MRS as a valuable indicator with the potential to detect changes in metabolic status following therapeutic intervention even from a young age The structure of clinical trial s renders gathering 31 P MRS data across several subjects and over time necessary for making accurate comparisons. However, confounding variables across individuals and over days may reduce the reliability of MR measures. These confounds include inter opera tor variability, and differences in subject positioning and sequence configurations. One procedure which acts to reduce the impact of these variables on data is the utilization of a phantom. A phantom is a canon which can be tested regularly to detect devi ations in scanner performance, and to correct data to make valid comparisons. The inorganic phosphate ( P i ) phantom was designed as part of this project to resemble P i stored in the skeletal muscle of the human leg. Human controls are also utilized in MR data acquisition for many purposes. The first is in detecting variability in 31 P data both day to day and across subjects. Metabolic indices which vary minimally across scans are more reproducible measures. Human controls are also useful for developing standardized methods for positioning subjects and for selecting muscles of interest (Forbes et al., 2013). Finally, 31 P MRS spectra collected from the skeletal muscle of h uman controls can be compared to spectra from boys with DMD, and can be held as a standard against which to be measured in future trials. In this study, a 31 P MRS protocol was implem ented to investigate metabolite values in a 31 P Lego Phantom and in the calves of unaffected human control subjects The aims o f this

PAGE 7

Human Calf Muscle Metabolites Evaluated by 31 P MRS 7 study were as follows: 1 ) to determine whether the 31 P Lego Phantom showed acceptable rep rod ucibility over several weeks; 2 ) to assess whether non localized 31 P MRS spectra collected simulta neously from all calf muscles of human control subjects differed across days scanned ; 3 ) to observe whether multivoxel 31 P MRS spectra collected from certain calf muscles of control s were reproducible and further, to determine whether any s ignificant differences existed among selected calf muscles ; 4 ) to evaluate which of the metabolite ratios collected from both the spatially localized and un localized scans we re the m o st sensitive across subjects and had the highest day to day reproducibility ; 5 ) to consider whether significant differences existed between the reproducibility of data gathered from the multivoxel and non localized 31 P MRS sequences Methods This experimental study was approved by the institutional review board at the University of Flor ida, and was conducted in compliance with the Health Insurance Por tability and Accountability Act. Informed written consent was obtained from the subject or guardian before study participation Phantom Studies The P i Lego P hantom was created with a t w o compartment coaxial geometry. The inner compartment contained potassium monophosphate, whil e the outer compartment contained potassium diphosphate, both diluted to simulate P i values in the skeletal muscle of the human calf. One phantom was created and s canned seven times over the course of 7 months to detect day to day variability in data Non localized and m ultivoxel (4 4 voxel and 8 8 voxel ) sequences were collected and analyzed. P i m onobasic and P i d ibasic were each represented by a Lorentzian curve with only a sing le peak. If more than one peak was present, the data was

PAGE 8

Human Calf Muscle Metabolites Evaluated by 31 P MRS 8 excluded from analysis. Data was acquired and quantified using standardized protocols which paralleled those used for human controls (see below). P articipants Four unaffected adult subjects (mean age, 24.3 6.5 year s [s tandard deviation]; range, 21 3 5 years) and one young boy (5 years of age) were recruited to participate in this study P articipants were asked to avoid physical activity beyond their normal level for three days prior to each scan All controls completed mea surements both at baseline and 1 week follow up 31 P MRS Data Acquisition All spectra were collected by a Philips 3T whole body MRI scanner using a circular phosphorou s surface coil (diameter=14cm). Standardized operating procedures were implemented across scans including subject positi oning and routine scan protocol. The imaging protocol contained a 2D Chemical Shift Imaging ( 2D CSI) data set to assess energy metabolism Total time in the scanner, including time for set up, did n ot exceed 40 minutes per subject. During the scan, patients lay supine, feet first, their calf centered in the bore of the magnet and elevated over rice bags The widest portion of the over the phosphorous surface coil and secured by foam padding to minimize movement Proper positioning was assessed using a T1 weighted survey scan w ith 2.5mm slice thickness (1.34 min TR=7.2ms, TE=3.5ms ) and a T2 weighted spin echo sequence with 7mm slice thick ness (slices=4, NSA=1, TR=3,000 ms, TE=20 ms n=16, 20 320 ms, each separated by 2 0ms ) Non localized data. Un localized MR sequences enable analysis of metabolite ratios gathered simultaneously from all skele tal muscle s of the calf. Non localized data was collected from the right calf of each subject at a repetition time (TR) of 5,00 0 m s and an echo time (TE) of

PAGE 9

Human Calf Muscle Metabolites Evaluated by 31 P MRS 9 0.10 ms. The number of signal averages (NSA) collected was 4, totaling a 20 second acquisition time for each patient. Non localized data was graphed as 7 Lorentzian resonances shown in Figure 1. Multivoxel data Spatially localized sequences allow for the collection of muscle specific metabolite ratios. Multivoxel data was collected from the right calf a nd voxels from the soleus (Sol) and medial gastrocnemius (MG ) were analyzed as free induction decays as shown in Figure 2. For these scans, the TR was 3000ms, the TE was 0.31ms, and the NSA collected was 8. For all except for the 55 sequence, spatial z ero filling was set to 4 for an increased matrix size. The first scan consisted of a 4 by 4 voxel matrix ( zero filled to 8 8 voxel matrix, each voxel was 35mm 35mm with 60mm slice thickness) with a total acquisition time of 6.4 minutes. The four adult control subjects each completed a 4 4 multivoxel scan both at baseline and 1 week follow up. Of note, the young control underwent a repeated 5 5 voxel scan (each voxel was 20mm 20mm with 60mm s lice thickness) with a total acquisition time of 7.4 minutes The second scan consisted of an 8 8 voxe l matrix ( zero filled to 16 16 voxel matrix, each voxel was 17.5mm 17.5mm with 30mm slice thickness) for a total acquisition time of 25.6 minutes. L ocalized 8 8 scans were collected on one acquisition day each for volunteers 1, 3 and 4. T he tibialis anterior (TA) was analyzed in addition to the Sol and MG during these scans. Spectra collected by the Philips 3T were converted to digital imaging and communications in medicine (DICOM) files and analyzed using ImageJ (NIH) software. The spin echo image taken at 20ms TE was overlaid with spectra using jMRUI, and muscles of interest were carefully selected by standardizing the number and position of voxels chosen from the

PAGE 10

Human Calf Muscle Metabolites Evaluated by 31 P MRS 10 slices. Voxels with at least 75% of their surface area overlaying the muscle of interest were considered acceptable while voxels overlapping multiple muscles or vo xels containing primarily fat or bone were not used for analysis (Figure 3 ) 31 P MRS Data Analysis Data analysis procedures were standardized across scans including the selection of key voxels, the use of prior knowledge, the setting of PCr as the zero reference, zero filling, apodizing, and phasing Metabolite resonances were evaluated using jMRUI ( http://www.jmrui.eu/welcome to the new mrui website ) and the total signal was f i tted with AMARES, a non linear least squares quantitation algorithm. Starting values and prior knowledge were created and used to tighten constraints for more accurate quantitation. Line widths were kept under 40 Hz Metabolic values were calculated by fit ting in the time domain for seven Lorentzian 31 P spectral peaks (Figure 4). Phosphorous peaks corresponded to phosphomonoester (PME), inorganic phosphate (P i ATP, respectively. Shimming was only considered satisfactory if the PCr full line width at half maximum (FWHM) was less than 100Hz. Resonances which were not visible due to low signal to noise ratio were excluded from analysis. For some of the multivoxel scans, the signal to ATP was not acceptable for accurate quanti tation Collected values were then analyzed as 15 ratios of interest (ATP/P tot ATP, PCr/ ATP, PCr/ATP, PCr/P tot PDE/ ATP, PDE/ATP, PDE/P tot P i ATP, P i ATP, P i /ATP, P i /P tot P i /PCr, and PME/P tot ATP)/3], and P tot was defined as ( P i ATP ) by the chemical shift between PCr and P i

PAGE 11

Human Calf Muscle Metabolites Evaluated by 31 P MRS 11 Statistical Analysis Variability in m etabolite ratios across scans was computed and coefficient of variation (CV). CV was calculated individually fo r each subject by dividing the standard deviation of the repeated measures by the mean of the repeated measures, and multiplying the result by 100%. The mean of the CV of all five subjects was then calculated and used to infer variability across participants and days scanned. A CV<10% was necessary for measures to be con sidered reproducible The CV was calculated in all phantom scans the non localized human scan s and the multivoxel human scan s for each ratio of interest and for pH Correlation between all other key data sets was calculated using two tailed Student T test s Comparisons were made between the reproducibility of un localized and multivoxel d ata (44 voxel ) by conducting a two tailed t test each array being composed of the CV of the five subjects. These t tests were conducted f or each of the metabolic indices and the mean of the non localized data was compared to the mean of the Sol and the MG separately T tests were also used to compare metabolic indices from the scan at baseline and the scan at follow up for each of the sequences to analyze whether any systematic differenc es occurred across days scanned. Finally, a two tailed t test was performed for each of the ratios and pH to determin e whether the 31 P values collected from the MG and Sol were similarly reproducible. P values>0.05 were considered necessary in determining that no significant differences existed between data sets of interest. Results The P i Lego Phantom was scanned seven times over the course of 7 months. Of the 36 resonances collected, 34 met criteria for analysis. 31 P MRS spectra were collected from the calf muscl e s of five unaffected controls both at baseline and after 1 week for a total of 10 scans A ll 90 spectra collected from un localized scans at baseline and follow up met quality control

PAGE 12

Human Calf Muscle Metabolites Evaluated by 31 P MRS 12 criteria. For the multivoxel data, 199 out of 203 resonances collected from baseline and follow up met quality control criteria. CV<10% was considered a noteworthy value in assessing for replicability and sensitivity of metabolic indices A p value>0.05 was necessary to establish that no significant differences existed between two data sets. Phantom Studies In the non localized phantom scans P i monobasic/ P i dibasic had a CV of 13.9% for the six scans which met criteria for analysis Interestingly, the amplitude ratio s of scans 1, 2 and 5 were identical at 1.72 each As seen in figure 5, s can 3 was an outlier in that it had a high er amplitude of 3.40 owing to an increased P i dibasic value The finding that the CV of the non localized scans was not consistently low suggests that this measure may be sensitive to potential variations in set up or acquisition from day to day, such as small changes in coi l position shimming, or signal to noise ratio. In the way of the seven 44 multivoxel phantom scans a CV of 9.14% was observed for the outer compartment ( =0.245), and a CV of 1 8.20 % was observed for the inner compartment ( =0.0992) CV is more affected by outliers when analyzing quantitative variables less than 1, as was the case when analyzing the amplitude ratio of the phantom inner compartment Therefore, the standard deviation i s also important to consider in determining how spect ra varied across scans In this case, the low inner compartment CV, as well as the low outer compartment were sufficient to suggest that there may have been minimal variation across 44 scans (Figure 6). The 8 8 multivoxel phantom scan s had a CV of 29.3% for the outer compartment amplitude ratio ( =1.33), and a CV of 11.9% for the inner compartment amplitude ratio ( =0.036) The variation in the outer compartment was likely influenced by an increased P i monobasic amplitude on scan 5 as the 88 sequence w as only performed during three scans Non Localized Human Control Studies

PAGE 13

Human Calf Muscle Metabolites Evaluated by 31 P MRS 13 For all 15 metabolite ratios col lected as well as for pH, p>0.05 for the t test which was conducted to compare day 1 scans to day 2 scans for volunteers 1 5. Thi s suggests that t here were likely no systematic differences between spectra collected at baseline and 1 week follow up. Thus, non localized sequences may have displayed high day to day replicability in adult controls, as well as in the young child. Multivoxel Human Control Studies Data also showed that it was probable that no systematic differences existed between the repeated 31 P MR S scans in the Sol, as was evidenced by p>0.05 for all metabolic indices evaluated in the 44 scans In the MG, all bio energetic values except for those of PDE were associated with p>0. 05 when comparing the scan at baseline to the follow up scan. Ra tios of PDE ( ATP ATP PDE/ATP and PDE/P tot ) rendered p values<0.05. Therefore, some day to day variability in PDE may have exist ed among the 44 MG scans. 15 metabolic indices and pH were also compared between the Sol and the MG via two tailed t test All 16 of these data sets were characterized by a p>0.05 suggesting that spectra gathered from the two muscles of interest ha ve the potential to be similarly reproducible Day to day Reproducibility in Human Controls To determine day to day replicability and variation across subjects the CV was calculated individually for each repeated measure Then the CV of all 5 volunteers were averaged for each ratio and compared. Regarding non localized data, ATP/P tot (CV=5.93 3 56 % ) ATP (CV=9.64 5 85 % ) PCr/ P tot (CV=6.07 3 42 %), and pH (CV=0.28 0 238 % ) had CV which were s ufficiently low to associate these values with minimal day to day variability and high sensitivity across subjects ATP tended to be associated with i ncreased CV (23.4 13 %
PAGE 14

Human Calf Muscle Metabolites Evaluated by 31 P MRS 14 miniscule amplitude ratios (<1) such that they were associated with a relatively large CV ( 10.4 6 7 %0.05. A p value>0.05 implies that no systematic differences exist ed between the reproducibility of the u n localized scan and the scan localized to the Sol. Six of the 16 t tests conducted to compare the reproducibility of non

PAGE 15

Human Calf Muscle Metabolites Evaluated by 31 P MRS 15 localized values to those of the MG were associated with p values<0.05 ATP, PDE/ATP, PDE/P tot P i /ATP, an d P i/ P tot ). Therefore, u n localized PDE and P i values were different than localized PDE and P i values in the MG These findings may indicate that multivoxel scans are sensitive enough to detect subtle changes in PDE and P i values in specific skeletal muscle s supporting the value in conducting localized mea s ures Discussion This experimental study evaluated the sensitivity and day to day reproducibility of metabolites quantified using non localized and multivoxel 31 P MRS sequences. A standardized MR protocol using 31 P hosphorus 2D CSI was implemented in both a coaxial P i Lego Phantom and the calf muscles of 5 unaffected human control subjects (4 adults, age range 21 35; 1 boy age 5 years) The P i phantom was scanned 7 ti mes over 7 months, and c ontrol subjects were scanned at baseline and 1 week follow up. T he time domain of the 31 P resonances collected from the scan w as fitted to quanti fy myocytic metabolites These amplitudes were then presented as ratios and compared to observe for day to day variability and accuracy. Ou r primary findings were that: 1 ) the P i phantom accurately detected variability between scans in both non localized and multivoxel 31 P MRS sequences; 2 ) there were no s ystematic differences between indices quantified by repeated non localized scans (p>0.05); 3 ) most values collected by multivoxel 44 scans displayed minimal variability (p>0.05) across days except for metabolic ratios in volving PDE in the MG. Further, metabolic indices in the MG and So l quantified by 44 multivoxel sequences were similarly reproducible (p>0.05); 4 ) ATP/P tot, PCr/P tot and pH were found to be the most highly reproducible indices across days scanned, and the most sensitive across subjects (CV<10%). Moreover, bioenergetic ratios influenced by ATP tended to have lower sensitivity and repeatability (CV>10%). These findings were true for the 4 adult contro ls and the young child alike; 5 ) spatially resolved human scan s have day to day reproduc ibility that is comparable to the reproducibility of non localized scans. This was true

PAGE 16

Human Calf Muscle Metabolites Evaluated by 31 P MRS 16 for all bioenergetic ratios except for those involving PDE and some of those involving P i in the MG. Upcoming t herapeutic interventions with the potential to alter the course of DMD would benefit from sensitive, non invasive bi omarkers of disease progression Recent studies have observed 31 P MRS values of metabolic function in DMD patients which differ significantly from those in controls (Wary et al., 2015). DMD is quantified by 31 P MRS as a graph with archetypal metabolite anomalies, including: a double P i peak (cytosolic and po oled inorganic p hosphate); increased total P i PME and PDE; and decreased PCr (Wary, Naulet, Thibaud, M onnet, Blot, & Carlier, 2012). Many have speculated that dystrophin deficiency in DMD is the origin of these abnormal indices, as d ystrophin is implicated in the regulat ion of calcium ion concentrations. B ecause Ca 2+ coordinates enzyme function, dystrophin deficient muscle may be susceptible to alterations in energy metabolism (Sharma et al., 2003). In establishing how 31 P indices in boys with DMD differ from those in unaffected controls studies have specifically analyzed how 31 P MRS ratios differ with age, inflammation and fat infiltration in both animal and human models of the disease (Latroche et al., 2015 ; Sharma et al., 2003 ; Wary et al., 2012 ). Although these aberrations in DMD bioenergetics have been well established, they are poorly understood For instance, r esearchers have delineated that phosphodiester (PDE) values are increased in the skeletal muscle of subjects with DMD compared to control s ( Reyngoudt et al., 2018). However, data regarding the exact origin of these increased products of membrane degradation as well as their evolution with disease progression has yet to be established (Banerjee et al., 2010; Hooijmans et al., 2017). The stud y at hand was useful in better characterizing metabolite values in the calf by detecting which were the most sensitive and replicable, and by determining how bioenergetic ratios differ among muscles, using both a P i phantom and unaffected human control sub jects. Replicability and Accuracy of 31 P Indices in Human Controls

PAGE 17

Human Calf Muscle Metabolites Evaluated by 31 P MRS 17 While a growing body of evidence has shown how bioenergetic values differ in boys with DMD compared to unaffected boys, n ot much has been done in the way of evaluating the day to day variability of metabolic indices as measured by 31 P MRS sequences. It is beneficial to better characterize these values by establish ing which of the phosphorous ratios is the most sensitive and repeatable. By so doing, we can expand up on previous research and further evaluate the usefulness of 31 P MRS a s a n efficacious biomarker of DMD pathology Replicability was evaluated in the calves of five human controls by observing how seven metabolic resonances vary at baseline and after one week The presentation of the seven peaks ATP peak down ATP often nearly o verlapping PCr. PCr was characterized by a large, centrally located peak. Quantification of ATP/P tot and PCr/P tot proved to be highly reproducibl e and sensitive with a mean within subject CV<10% in both the non localized and 44 Sol scan s This was true for all four adult subjects, as well as for the young child. These ratios were not as reproducible in the 44 MG scan, (CV>10%) likely because the MG had a lower signal to noise ratio than the Sol since it had a greater offset from the center of the coil ( Figure 2). Notably, PCr/Ptot showed high reproducibility for all three selected muscles in the 88 multivoxel scans as well, although these sequences were ATP was shown to be a less sensitive biomarker compared to the other metabolites as revealed by its high CV across all sequences and in all subjects. We hypothesized that this was because its resonance frequency was further from the zero reference (center frequency), and because it was often associated with a lower signal to noise ratio than were other peaks. ATP and PCr are both high energy phosphate stores essential for muscle contraction. PCr stored in the muscle donates its phosphate group to ADP to generate ATP and the rate of ATP production and utiliza A decreased PCr peak may

PAGE 18

Human Calf Muscle Metabolites Evaluated by 31 P MRS 18 be indicative of a pathologic status such as in DMD (Banerjee et al., 2010). Reduced PCr may be associated with membrane instability, which leads to enzyme dysregulation and an inability of the myocyte to uptake glucose for glycolysis (Sharma et al., 2003). ATP resonances do not vary significantly in DMD pathology (Raymond et al., 1982). Because PCr spectra were shown to be both sensitive and replicable, even in the yo ung subject this high energy metabolite may be used to characterize DMD pathology in future trials. The PDE and PME peaks were both slight in all subjects and across both non localized and multivoxel scans. Because their amplitude ratios were so small, they were associated with a relatively large CV Further, the PDE peak did not meet criteria for analysis in the 44 scan of the MG for volunteer 3, and PME did not meet criteria in the 44 scan of the Sol for volunteer 5 These excluded resonances may have also impacted the magnitude of the CV. The PME peak is composed primarily of gluc ose 6 phosphate, an intermediate molecule in the breakdown of glucose during glycolysis. Some suggest that an increased PME peak is characteristic of dystrophic muscle, which is easily fatigue d and lacks normal levels of lactic acid production due to an inability to use glucose efficiently (Wehling Henricks, Oltmann, R inaldi, Myung, & Tidall, 2009). PDE represents catabolized components of the lipid bilayer. It has been suggested that high PDE amplitudes may correlate to a higher rate of cell wall turnover, reflecting a diseased state (Banerjee e t al., 2010; Wary et al., 2012). A recent study by Hooijman et al. (2017) using a 7T MR system revealed that PDE was highly reproducible with a mean wit hin subject CV of 4.3% in datasets with high signal to noise ratio, and 5.7% in datasets with low signal to noise ratio, in four adult controls. This study also found that in DMD, PDE was elevated independent of fat fraction, even in boys as young as 5 yea rs old. Thus, a lthough PDE was not found to be highly reliable at 3T, this resonance has the potential to be sensitive to disease pathology even from a young age

PAGE 19

Human Calf Muscle Metabolites Evaluated by 31 P MRS 19 the distance between PCr and P i In th is study, pH had the lowest CV of all metabolic indices and across all sequences both in the adult subjects and the young child Alkaline pH is one of the most consistent indicators of DMD pathology across age, muscle of interest and severity as well as one of our most sensit ive and reproducible biomarkers (Hooijman et al., 2017). It has been suggested that a more basic P i resonance would be reflective of oxidative phosphorylation dysregulation or leaky membra nes which would lead to energy wasting ( Reyngo udt et al., 2018; Wary et al., 2012) By establishing which of the myocytic metabolites quantified by 31 P MRS are the most sensitive and repeatable, we can expand upon and verify previous research. For instance, Banerjee et al. (2010) found that PCr/P i ATP showed the most significant difference between boys with DMD and unaffected human controls. Findings from the research at hand suggest that follow up studies should focus more on the PCr/P i metabolic ratio, as it is the more sensitive ratio. Multivoxel Studies of 31 P Indices in Human Controls There is also a shor tage of research which analyzes metabolite ratios via spatially localized sequences. The collection of muscle specific 31 P indices is necessary because muscles in patients with DMD become affected at different stages and progress at different rates within the disease process (Hooijmans et al., 2017). Typically, type II fast twitch muscle fibers such as the MG are more affected by DMD (Wary et al., 2012) Recent DMD research has benefitted greatly from observing proton transverse relaxation time (T 2 ) among individual muscles. Researchers found that in DMD, the MG and Sol had an increased T 2 while the TA (medially) was relatively spared e ven in boys from a young age Further, m uscle specific T 2 values have been shown to be more sensitive than are mean T 2 measures. W ater T 2 has also been implicated as a biomarker capable of detecting subtle

PAGE 20

Human Calf Muscle Metabolites Evaluated by 31 P MRS 20 changes among specific muscles in diseas e pathology, even from a young age (Arpan et al., 2013). Overall, these MR studies support the importance of studying disease progression in individual muscles. A recent study by Hooijmans et. al (2017), assessed muscle specific 31 P values in the calves of boys with DMD compared to age matched controls. The study found that in DMD, P i /PCr was increased in all muscles at baseline, and increased only in the Sol at 24 months follow up. It was also observed that pH was more alkaline in the Sol, MG and TA of boys with DMD at all time points. Many m etabolite ratios were found to be impacted differently among muscles at varying time points in DMD pathology These findings indicate that 31 P values may detect differences in muscle tissue stat us throughout the disease process. The study at hand has also helped to establish the use of multivoxel sequences at 3T as efficacious biomarkers of DMD progression by showing that, overall, they have reproducibility which is similar to that of non localized scans. One disadvantage to the use of multivoxel sequences is that they result in added time in the MR scanner. While the non localized scan takes only 20 seconds, the 44 scan ta kes 7 minutes, and the 88 scan takes 26 minutes. Thus, these spatially resolved scans may be more susceptible t o effects from movement and loss of participant attention. Nonetheless, there is potential for multivoxel measures of bioenergetic indices to be specific, and to have the ability to detect subtle changes in varying muscles of boys with DMD even from a young age. This study observed the replicability of phosphorous metabolites in the calves of human control subjects using both localized and non localized 31 P MRS sequences. Results revealed that ATP/P tot PCr/P tot and pH showed the greatest day to day reproducibility of the metabolite ratios examined Further, spectra collected from the non localized scan and the multivoxel scan were shown to be similarly reproducible. However, because some spectra failed to meet quality control criteria, they were excluded from analysis. This may have biased some of the CV and p

PAGE 21

Human Calf Muscle Metabolites Evaluated by 31 P MRS 21 values of the 16 metabolic indices Additionally, this study was limited by its small sample size which consisted primarily of adult subjects. Fut ure studies should verify whether this adult population is fully translatable to a population of young boys. The use of larger sample sizes and young controls may help to better elucidate whether multivoxel and non localized 31 P MRS sequences may serve as sensitive, replicable b iomarkers with the potential to detect drug efficacy and disease progression in young boys with DMD

PAGE 22

Human Calf Muscle Metabolites Evaluated by 31 P MRS 22 References Arpan, I., Forbes, S. C., Lott, D. J., Senesac, C. R., Daniels, M. J., Triplett, W. T., & ... Vandenborne, K. (2013). T2 mapping provides multiple approaches for the characterization of muscle involvement in neuromuscular diseases: a cross sectional study of lower leg muscles in 5 15 year old boys with Duchenne muscular dystrophy. NMR In Biomedicine 26(3), 320. doi:10.1002/nbm.2851 Arpan, I., Willcocks, R. J., Forbes, S. C., Finkel, R. S., Lott, D. J., Rooney, W. D., . Vandenborne, K. (2014). Examinati on of effects of corticosteroids on skeletal muscles of boys with DMD using MRI and MRS. Neurology, 83(11), 974 980. doi:10.1212/WNL.0000000000000775 Banerjee, B., Sharma, U., Balasubramanian, K., Kalaivani, M., Kalra, V., & Jagannathan, N. R. (2010). Effe ct of creatine monohydrate in improving cellular energetics and muscle strength in ambulatory Duchenne muscular dystrophy patients: a randomized, placebo controlled 31P MRS study. Magnetic Resonance Imaging (0730725X), 28(5), 698 707. doi:10.1016/j.mri.201 0.03.008 Deconinck, N., & Dan, B. (2007). Pathophysiology of Duchenne Muscular Dystrophy: Current Hypotheses. Pediatric Neurology, (1), 1. Forbes, S., Willcocks, R., Triplett, W., Rooney, W., Lott, D., Wang, D., . Vandenborne, K. (2014). Magnetic resonance imaging and spectroscopy assessment of lower extremity skeletal muscles in boys with duchenne muscular dystrophy: A multicenter cross sectional study. Plos One 9(9), e106435. doi:10.1371/journal.pone.0106435 Hooijmans, M. T., Doorenweerd, N., Baligand, C., Verschuuren, J. M., Ronen, I., Niks, E. H., & ... Kan, H. E. (2017). Spatially localized phosphorous metabolism of skeletal muscle in

PAGE 23

Human Calf Muscle Metabolites Evaluated by 31 P MRS 23 Duchenne muscular dystrophy patients: 24 month follow up. Plos ONE 12(8) 1 15. doi:10.1371/journal.pone.0182086 Humbertclaude, V., Hamroun, D., Bezzou, K., Brard, C., Boespflug Tanguy, O., Bommelaer, C., & ... Tuffery Giraud, S. (2012). Original article: Motor and respiratory heterogeneity in Duchenne patients: Implication f or clinical trials. European Journal Of Paediatric Neurology 16149 160. doi:10.1016/j.ejpn.2011.07.001 Latroche, C., Matot, B., Martins Bach, A., Briand, D., Chazaud, B., Wary, C., . Jouvion, G. (2015). Structural and functional alterations of skeleta l muscle microvasculature in dystrophin deficient mdx mice. American Journal of Pathology 185(9), 2482 2494. doi:10.1016/j.ajpath.2015.05.009 Le Guiner, C., Montus, M., Servais, L., Cherel, Y., Francois, V., Thibaud, J., & ... Voit, T. (2014). Original Ar ticle: Forelimb Treatment in a Large Cohort of Dystrophic Dogs Supports Delivery of a Recombinant AAV for Exon Skipping in Duchenne Patients. Molecular Therapy, 221923 1935. doi:10.1038/mt.2014.151 Le Guiner, C., Servais, L., Montus, M., Larcher, T., Frays se, B., Moullec, S., & ... Dickson, G. (2017). Long term microdystrophin gene therapy is effective in a canine model of Duchenne muscular dystrophy. Nature Communications 816105. doi:10.1038/ncomms16105 Mendell, J. R., Shilling, C., Leslie, N. D., Flaniga n, K. M., al Dahhak, R., Gastier Foster, J., & ... Weiss, R. B. (2012). Evidence based path to newborn screening for Duchenne muscular dystrophy. Annals Of Neurology 71(3), 304 313. doi:10.1002/ana.23528 Raymond J., N., Peter J., B., Lawrence, C., David G., G., Peter, S., Doris, T., & George K., R. (1982). Nuclear Magnetic Resonance Studies Of Forearm Muscle In Duchenne Dystrophy. British Medical Journal (Clinical Research Edition), (6322), 1072.

PAGE 24

Human Calf Muscle Metabolites Evaluated by 31 P MRS 24 Reinig, A M., Mirzaei, S., & Berlau, D. J. (2017). Advances in the Treatment of Duchenne Muscular Dystrophy: New and Emerging Pharmacotherapies. Pharmacotherapy: The Journal Of Human Pharmacology And Drug Therapy (4), 492. doi:10.1002/phar.1909 Reyngoudt, H., Tur k, S., & Carlier, P. G. (2018). 1H NMRS of carnosine combined with 31P NMRS to better characterize skeletal muscle pH dysregulation in Duchenne muscular dystrophy. NMR In Biomedicine (1), doi:10.1002/nbm.3839 Sharma, U., Atri, S., Sharma, M., Sarkar, C., & Jagannathan, N. (2003). Regular article: Skeletal muscle metabolism in Duchenne muscular dystrophy (DMD): an in vitro proton NMR spectroscopy study. Magnetic Resonance Imaging, 21145 153. doi:10.1016/S0730 725X(02)00646 X Thibaud, J., Wary, C., Moullec, S., Azzabou, N., Le Guiner, C., Garcia, L., . Carlier, P. (2012). Quantitative evaluation of locoregional high venous pressure rAAV8 U7 ESE6 ESE8 exon skipping therapy in the GRMD dog using NMR H 1 imaging and P 31 spectroscopy. Neuromuscular Disorders 22(9 10), 859 859. doi:10.1016/j.nmd.2012.06.188 Triplett, W., Forbes, S. C., DeVos, S., Lott, D. J., Willcocks, R. J., Senesac, C., & ... Sweeney, H. L. ( 2013 ). Skeletal Muscles of Ambulant Children with Duchenne Muscular Dystrophy: Validation of Multic enter Study of Evaluation with MR Imaging and MR Spectroscopy. Radiology 269(1), 198 207. Wary, C., Naulet, T., Thibaud, J., Monnet, A., Blot, S., & Carlier, P. G. (2012). Splitting of pi and other 31P NMR anomalies of skeletal muscle metabolites in cani ne muscular dystrophy: 31P NMR anomalies in skeletal muscle of dystrophic dogs. NMR in Biomedicine 25(10), 1160 1169. doi:10.1002/nbm.2785

PAGE 25

Human Calf Muscle Metabolites Evaluated by 31 P MRS 25 Wary, C., Azzabou, N., Giraudeau, C., Le Lour, J., Montus, M., Voit, T., . Carlier, P. (2015). Quantitative NMR I and NMRS identify augmented disease progression after loss of ambulation in forearms of boys with duchenne muscular dystrophy. NMR in Biomedicine 28(9), 1150 1162. doi:10.1002/nbm.3352 Wehling Henricks, M., Oltmann, M., Rinaldi, C., Myung, K., & Tidball J. (2009). Loss of positive allosteric interactions between neuronal nitric oxide synthase and phosphofructokinase contributes to defects in glycolysis and increased fatigability in muscular dystrophy. Human Molecular Genetics 18(18), 3439 3451. Willcocks, R. J., Rooney, W. D., Triplett, W. T., Forbes, S. C., Lott, D. J., Senesac, C. R., & ... Vandenborne, K. (2016). Multicenter prospective longitudinal study of magnetic resonance biomarkers in a large Duchenne muscular dystrophy cohort. An nals of Neurology 79(4), 535 547. doi:10.1002/ana.24599 Vohra, R. S., Lott, D., Mathur, S., Senesac, C., Deol, J., Germain, S., & ... Vandenborne, K. (2015). Magnetic Resonance Assessment of Hypertrophic and Pseudo Hypertrophic Changes in Lower Leg Muscles of Boys with Duchenne Muscular Dystrophy and Their Rela tionship to Functional Measurements. Plos ONE 10(6), 1 17. doi:10.1371/journal.pone.0128915

PAGE 26

Human Calf Muscle Metabolites Evaluated by 31 P MRS 26 Appendix Figure 1. Presentation of the seven peaks of interest in the non localized day 1 scan of volunteers 1 5. Fig ure 2. Multivoxel 44 spectra from the Sol (left) and MG (right) of volunteer 1 from both days scanned coil coil Day 1 Day 2 Day 1 Day 2

PAGE 27

Human Calf Muscle Metabolites Evaluated by 31 P MRS 27 Figure 3 A comparison of 44 (left) and 88 (right) voxels selected from the Sol of volunt eer 3 on the second day scanned. Voxels which had 75% of their surface area overlapping the muscle of interest were utilized for analysis. Figure 4 Non localized sp ectra collected from volunteer 5 on the first day scanned

PAGE 28

Human Calf Muscle Metabolites Evaluated by 31 P MRS 28 Figure 5. Amplitude ratios for non localized phantom scans labeled with error bars representing standard deviation (scan 1: 5/18/2017, scan 2: 5/24/2017, scan 3: 7/24/2017, scan 4: 8/4/2017, scan 5: 9/29/2017, scan 6: 10/20/2017). Scan 7 was excluded from this data set due to poor shim. Figure 6 Amplitude ratios for spatially localized phantom scans labeled with error bars representing standard deviation (scan 1: 5/18/2017, scan 2: 5/24/2017, scan 3: 7/24/2017, scan 4: 8/4/2017, scan 5: 9/29/2017, scan 6: 10/20/2017, scan 7: 11/16/2017). 0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5 2.75 0 1 2 3 4 5 6 Apliitude Pi Dibasic/ Pi Monobasic Scan Amplitude Ratios for Non Localized Phantom Scans 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 1 2 3 4 5 6 7 8 Amplitude Pi Dibasic/Pi Monobasic Scan Amplitude Ratios for 4 by 4 Multivoxel Phantom Scans Outer Compartment Inner Compartment

PAGE 29

Human Calf Muscle Metabolites Evaluated by 31 P MRS 29 Table 1. Note. ATP was defined as ATP)/3], and P tot was defined as (P i ATP). Averages of Day 1 and Day 2 N on Localized CV(%) for A ll Subjects Volunteer 1 Volunteer 2 Volunteer 3 Volunteer 4 Volunteer 5 Average St Dev ATP/P tot 12 % 2.1 % 6.1 % 5.2 % 4.5 % 5.9 % 0.036 PCr/ ATP 42 % 37 % 35 % 6.9 % 17 % 28 % 0.15 PCr/ ATP 14 % 5.9 % 7.2 % 18 % 3.7 % 9.6 % 0.059 PCr/ATP 22 % 9 % 9.9 % 13 % 6.4 % 12 % 0.059 PCr/Ptot 10 % 7 % 3.8 % 7.6 % 1.6 % 6.1 % 0.034 PDE/ ATP 25 % 11 % 42 % 12 % 27 % 23 % 0.13 PDE/ ATP 3.7 % 21 % 14 % 12 % 14 % 13 % 0.063 PDE/ATP 4.7 % 18 % 17 % 7.2 % 17 % 13 % 0.063 PDE/Ptot 7 % 20 % 11 % 2 % 12 % 10 % 0.067 Pi tot / ATP 28 % 2.7 % 40 % 31 % 29 % 26 % 0.14 Pi tot / ATP 0.62 % 29 % 12 % 6.5 % 16 % 13 % 0.11 Pi tot /ATP 7.7 % 26 % 15 % 12 % 19 % 16 % 0.07 Pi tot /Ptot 4 % 28 % 8.9 % 17 % 14 % 14 % 0.09 Pi b /PCr 31 % 41 % 31 % 13 % 46 % 32 % 0.13 Pi a /PCr 13 % 34 % 0.96 % 29 % 8.6 % 17 % 0.14 Pi tot /PCr 14 % 35 % 5.1 % 24 % 12 % 18 % 0.12 Pi b /Pi a 18 % 7.7 % 30 % 41 % 38 % 27 % 0.14 Pi b /P tot 20 % 35 % 35 % 20 % 47 % 31 % 0.12 Pi a /P tot 2.3 % 27 % 4.8 % 22 % 10 % 13 % 0.11 PME/P tot 7.5 % 66 % 17 % 3.1 % 19 % 23 % 0.25 pH 0.29 % 0 % 0.40 % 0.6 % 0.099 % 0.28 % 0.0024

PAGE 30

Human Calf Muscle Metabolites Evaluated by 31 P MRS 30 Table 2. Note. PDE data from the Sol of volunteer 5 on the first day scanned was excluded from analysis due to poor shim. Averages of Day 1 and Day 2 Multivoxel Soleus CV(%) for A ll Subjects Vol 1 (44) Vol 2 (44) Vol 3 (44) Vol 4 (44) Vol 5 (55) Average St Dev ATP/P tot 5.1 % 1.8 % 15 % 3.4 % 11 % 7.4 % 0.057 PCr/ ATP 5.9 % 3.4 % 44 % 25 % 29 % 21 % 0.17 PCr/ ATP 13 % 2.6 % 27 % 6.1 % 12 % 12 % 0.095 PCr/ATP 14 % 0.33 % 32 % 6.7 % 20 % 15 % 0.12 PCr/P tot 8.7 % 1.5 % 17 % 3.2 % 9. 2 % 8 % 0.063 PDE/ ATP 33 % 78 % 43 % 20 % --43 % 0.25 PDE/ ATP 26 % 76 % 27 % 0.28 % --32 % 0.32 PDE/ATP 26 % 72 % 31 % 0.84 % --32 % 0.29 PDE/P tot 31 % 68 % 17 % 2.7 % --30 % 0.28 P i / ATP 33 % 58 % 17 % 21 % 49 % 36 % 0.17 P i / ATP 26 % 57 % 0.12 % 2.2 % 34 % 24 % 0.24 P i /ATP 25 % 53 % 5.1 % 2.7 % 41 % 25 % 0.22 Pi/P tot 30 % 49 % 10 % 0.73 % 31 % 24 % 0.19 Pi/PCr 38 % 52 % 27 % 4 % 22 % 29 % 0.18 PME/P tot 37 % 2.6 % 22 % 35 % 86 % 36 % 0.31 pH 0.30 % 0.10 % 0.30 % 0.10 % 0.30 % 0.22% 0.0011

PAGE 31

Human Calf Muscle Metabolites Evaluated by 31 P MRS 31 Table 3. Averages of Day 1 and Day 2 Multivoxel Medial Gastrocnemius CV(%) for all Subjects Vol 1 (44) Vol 2 (44) Vol 3 (44) Vol 4 (44) Vol 5 (55) Average St Dev ATP/P tot 14 % 2.1 % 0.28 % 23 % 21 % 37 % 0.11 PCr/ ATP 18 % 16 % 19 % 64 % 67 % 13 % 0.26 PCr/ ATP 23 % 13 % 0.32 % 30 % 0.04 % 23 % 0.13 PCr/ATP 24 % 9.6 % 5.8 % 38 % 37 % 11 % 0.15 PCr/P tot 10 % 7.6 % 5.5 % 16 % 17 % 64 % 0.05 PDE/ ATP 68 % 58 % 60 % 60 % 75 % 63 % 0.072 PDE/ ATP 72 % 61 % 44 % 25 % 110 % 63 % 0.34 PDE/ATP 72 % 63 % 49 % 34 % 97 % 59 % 0.24 PDE/P tot 62 % 65 % 49 % 11 % 110 % 48 % 0.35 P i / ATP 33 % 31 % 29 % 66 % 78 % 33 % 0.23 P i / ATP 38 % 35 % 47 % 32 % 15 % 42 % 0.12 P i /ATP 39 % 38 % 41 % 41 % 50 % 31 % 0.05 Pi/P tot 26 % 39 % 41 % 18 % 31 % 25 % 0.095 Pi/PCr 16 % 46 % 46 % 2.5 % 15 % 57 % 0.20 PME/P tot 59 % 2.2 % --43 % 120 % 57 % 0.5 pH 0.20 % 1.7 % 0.60 % 0% 0 % *0.5 % 0.0071 Note. PME data from the MG of volunteer 3 on the first day scanned was excluded from analysis due to poor shim.

PAGE 32

Human Calf Muscle Metabolites Evaluated by 31 P MRS 32 Table 4. CV and St andard Dev iation of Volunteers 1, 3 & 4 Multivoxel 88 Scans Sol 88 MG 88 TA 88 CV (%) St Dev CV(%) St Dev CV(%) St Dev ATP/Ptot 6.6 % 0.0075 11 % 0.014 18 % 0.018 26 % 4.4 66 % 13 96 % 33 11 % 0.39 29 % 0.96 20 % 0.59 PCr/ATP 8.4 % 0.42 18 % 0.81 23 % 1.3 PCr/Ptot 2.1 % 0.012 7.4 % 0.042 9.2 % 0.054 42 % 0.54 28 % 0.08 79 % 0.43 47 % 0.12 73 % 0. 08 89 % 0.14 PDE/ATP 44 % 0.17 97 % 0.098 130 % 0.34 PDE/Ptot 36 % 0.016 90 % 0.011 130 % 0.032 2 2 % 0.55 71 % 1.3 105 % 6.6 26 % 0.13 29 % 0.095 9.9 % 0.049 Pi/ATP 23 % 0.17 12 % 0.05 6.3 % 0.057 Pi/Ptot 17 % 0.014 1.2 % 0.00064 8.1 % 0.0079 Pi/PCr 17 % 0.024 13 % 0.013 12 % 0.021 PME/Ptot 70 % 0.025 120 % 0.0082 49 % 0.012 pH 0.16% 0.012 0.41% 0.029 3.8 % 0.27