Is proton magnetic resonance spectroscopy a valid method to quantify muscle carnosine in humans?

5 1 Applied Physiology and Nutrition Research Group, School of Physical Education and Sport; 6 Rheumatology Division; Faculdade de Medicina FMUSP, Universidade de Sao Paulo, Sao Paulo, SP, BR. 7 2 Rheumatology Division; Faculdade de Medicina FMUSP, Universidade de Sao Paulo, Sao Paulo, SP, BR. 8 3 LIM44 Institute of Radiology and Oncology, School of Medicine, University of Sao Paulo, Sao Paulo 9 01246903, Brazil. 10 4 Sport, Health and Performance Enhancement Research Centre; Musculoskeletal Physiology Research 11 Group; School of Science and Technology, Nottingham Trent University, Nottingham, United Kingdom. 12 5 Department of Biochemistry and Technological Chemistry, Universidade Estadual Paulista (UNESP), SP, 13 BR. 14


7
The 50 mmol·L -1 carnosine phantom was used as an external reference. To that end, an acquisition was made 0 8 using the same parameters as those used in vivo, except for the TR, which was 12000 ms. In each in vivo 0 9 measurement, the right leg of each participant was positioned and was firmly immobilized in the knee coil, 1 0 such that the gastrocnemius muscle was in the centre of the coil. The left leg was supported outside the coil 1 1 to improve comfort and thus minimize leg movement. Voxel location was standardized on the larger calf 1 2 region in the centre of the medial portion of the gastrocnemius muscle of the right leg. The same well-1 3 trained and experienced biomedical technician was responsible for placing the voxel in all conditions. After  Absolute quantification of the carnosine resonance was determined using the following equation (32): CM is the concentration of the metabolite in vivo, SV and ST are the signals of the water-corrected 2 2 metabolite in vivo and in vitro; SM is the integral of the carnosine peak in vivo and SR is the integral of the The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/568923 doi: bioRxiv preprint 1 0 1 0 water peak in vivo and S H2OT is the integral of the water peak in vitro; F H2OMT1 is the correction factor for T1 2 6 relaxation of water in vivo; And F H2OMT2 and F H2ORT2 are the correction factors for T2 relaxation of water in 2 7 vivo and in vitro. Pt is the temperature correction factor applied as the signal decreases by 6% between the 2 8 room temperature (i.e., phantom temperature) and body temperature (11). For the in vitro signal, it is not 2 9 necessary to correct the T1 relaxation, since the acquisition was performed with a sufficiently long TR 3 0 (TR=12000 ms) to neglect this factor. Signals were also corrected by water content; since the water content 3 1 in phantoms is ~100%, a correction factor=1 was used. For the in vivo analyses, a correction factor=0.66 3 2 was used, assuming that ~2/3 of the muscle is water (36).

3
The relaxation correction factors were calculated using the following equations: and T2 values for water and carnosine were taken from the literature, and were assumed to be  The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/568923 doi: bioRxiv preprint To examine the potential interference of other sources of imidazole ring to the carnosine signal 9 9 obtained in vivo, we performed a series of in vitro 1H-MRS acquisitions in phantoms containing pure 0 0 imidazole (12.5 mmol·L -1 ), pure histidine (equivalent to imidazole 12.5 mmol·L -1 ), pure carnosine 0 1 (equivalent to imidazole 12.5 mmol·L -1 ) and pure protein (bovine serum albumin, BSA, equivalent to 0 2 imidazole 12.5 mmol·L -1 ).

3
Signals were quantifiable in the phantoms containing imidazole, histidine and carnosine, but not in   In order to confirm that HPLC could be used as a reference method, and to account for all major 2 3 sources of error associated with this method, we determined intra-assay reliability (i.e., same extract, from The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/568923 doi: bioRxiv preprint 1 4 1 4 the same muscle sample, analysed on two separated runs; n=15 m. vastus lateralis samples), inter-extract 2 5 reliability (i.e., two different extracts, from the same muscle sample, analysed on two separated runs; n=11 2 6 m. vastus lateralis samples) and inter-biopsy reliability (two different extracts from two different biopsies of 2 7 the same muscle; n=7 m. gastrocnemius samples).

8
Inter-assay reliability showed very similar values between measurements (mean difference: 2 9 0.6±4.0%). No statistically significant differences between measurements were shown (t=0.144; p=0.887),  To thoroughly assess the validity of 1H-MRS to quantify carnosine in human skeletal muscle, a 4 2 series of in-vivo studies were conducted in young healthy men aiming to examine test-retest reliability,  The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/568923 doi: bioRxiv preprint 1 5 1 5 before the second test. When the participants were not removed from the scanner, good reliability was 5 1 obtained between measurements, (ICC=0.924, 95%CI=0.0.451-0.992; CV=6.6%; Figure 5, upper panel).

2
However, reliability indexes were poorer when individuals were removed and subsequently repositioned on The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/568923 doi: bioRxiv preprint 1 7 1 7 causes a phenomenon known as spin-spin coupling, where the magnetic field generated by each proton can be better observed for the C4-H peak, since its position in the imidazole ring and its proximity with the 0 5 nearest protons makes its signal more susceptible to spin-spin coupling effects (7). These features represent 0 6 a challenge when trying to use 1H-MRS to quantify carnosine in human muscle. As herein demonstrated, 0 7 carnosine peaks already present a small amplitude in the 1H-MRS (21); when these signals are divided or 0 8 further weakened by the spin-spin coupling, a decrease in signal, resulting in poorer signal-to-noise ratio 0 9 (19). In individuals with low muscle carnosine content, increased error is to be expected, since peak 1 0 amplitude is naturally lower and, therefore, very close to the basal noise. This is supported by the increased 1 1 disagreement between 1H-MRS and the reference method shown in the Bland-Altman plot when carnosine 1 2 concentrations are near to the lowest range. One could suggest to measure, as an alternative, the carnosine 1 3 subpeaks. However, identifying all subpeaks in the spectrum may not be possible, since they might be 1 4 totally covered by noise, especially in volunteers who present low muscle carnosine levels.

5
To investigate the potential impact of the imidazole ring present in other molecules (e.g., free 1 6 imidazole, free histidine, carnosine analogues and histidine residues in proteins) on the carnosine signal 1 7 detected by 1H-MRS, a series of acquisitions were performed in phantoms containing pure carnosine, 1 8 imidazole, histidine, and BSA. Although the best signal quality was obtained with carnosine, quantifiable 1 9 signals were also obtained with imidazole and free histidine. This indicates that small imidazole-containing 2 0 molecules might constitute a potential source of error, although they are likely of low relevance for the 2 1 skeletal muscle since they are expressed in very low concentrations in comparison with carnosine (33).

2
However, this may be problematic when carnosine is to be measured in tissues expressing higher 2 3 concentrations of small imidazole-containing molecules. Conversely, no signal was obtained with BSA, 2 4 probably due to its large size (~66 kDa). Increasing molecular size leads to slower tumbling and 2 5 correspondingly shorter spin-spin relaxation times (T2), resulting in to a more complex spectrum with very 2 6 broad peaks of low amplitude that do not surpass noise level. Accordingly, 1H-MRS experiments become 2 7 unreliable at room temperature for proteins larger than 30 kDa and largely fail for proteins above 35 kDa in 2 8 . CC-BY-NC 4.0 International license is made available under a The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/568923 doi: bioRxiv preprint 1 8 1 8 the absence of elevated temperature (40). These results indicate that imidazole-rich proteins such as 2 9 haemoglobin and other large proteins do not represent a source of error. However, it is still possible that 3 0 other smaller histidine-rich proteins, such as myoglobin (17 kDa, 4.7 mg.g -1 wet muscle, 11% histidine) 3 1 might contribute to the in vivo signal. Unfortunately, purified myoglobin is not easily accessible and we 3 2 could not prepare a phantom containing myoglobin for further verification. Hence, whether myoglobin 3 3 constitutes a source of error requires future clarification.

4
To assess reliability, convergent and discriminant validity of in vivo 1H-MRS, a series of analyses 3 5 were conducted. No significant differences were shown between test and retest values, indicating that 1H-3 6 MRS is free of systematic errors and that the variation is explained by random error. Importantly, a 3 7 remarkable increase in test-retest variation (6.6% vs. 16.9%) was shown when the retest was performed with 3 8 the participant being removed from and then relocated to the scanner, which is a more "real-world" 3 9 representation of studies assessing muscle carnosine before and after an intervention. Such an increase in shimming is somewhat expected for a metabolite with a broad signal and of low amplitude, such as 4 7 carnosine, since similar levels of variation have been reported for other metabolites of much sharper and 4 8 high amplitude signals (1). In addition, the ~17% variation reported in this study is not too dissimilar to the 4 9 ~23% previously shown (32).

0
One explanation for the larger variation of 1H-MRS is the non-homogeneity of carnosine distribution 5 1 in the skeletal muscle. Carnosine content is greater in type II than type I muscle fibres (12); therefore, the 5 2 inevitable change in muscle site when performing two 1H-MRS may cause sampling sites to have different 5 3 fiber type composition, adding a source of measurement error. However, slight changes in sampling sites The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/568923 doi: bioRxiv preprint 1 9 1 9 likely occurred with muscle biopsies, but the variation for HPLC was lower, nonetheless. This means that 5 5 other factors may play a role in the increased variability in 1H-MRS, such as the proximity with tissues that 5 6 may cause signal interference (e.g., adipose tissue). The fat signal often appears bright in many important 5 7 clinical imaging sequences and can obscure other signals (6, 27). Thus, adipose tissue near or inside the data 5 8 acquisition site can contribute to the increased variability. Additionally, 1H-MRS appears to be more 5 9 sensitive to changes in sampling sites because shimming and spin-spin coupling are dependent on the angle 6 0 between spins and the magnetic field, which may alter with slight changes in sampling sites. Such 6 1 orientation-dependence is particularly true for the carnosine signal in human skeletal muscle (23). Finally, 6 2 participants' motion during the exam could also disrupt data acquisition, thereby contributing to 1H-MRS 6 3 variability (26), although the participants were instructed not to perform any movement during the entire 6 4 duration of the exams.

5
The discriminant validity study showed that 1H-MRS, despite having large variation owing to 6 6 random errors, is sensitive to detect group-mean increases in muscle carnosine in response to β -alanine