Visualizing lactate in the ischemic hindleg of mice using localized proton 2D correlation spectroscopy (UHf-P-COSY) at ultrahigh field.
Devashish Das1, Andor Veltien1, and Arend Heerschap1

1Radiology, Radboud University Nijmegen, Nijmegen, Netherlands

Synopsis

buildup
Lactate is associated with mitochondrial and neuromuscular diseases. Therefore, its methylene (CH2) and methyl (CH3) proton assignment is necessary in the skeletal muscle. In particular for understanding mechanisms associated with the skeletal muscle ischemia in rodents. Unambiguous detection of lactate resonances in the ischemic and/or hypoxic rodent muscle remains challenging, however, necessary for the assessment of treatment in the dysfunctional muscle of transgenic animals. In vivo lactate resonances remain hidden under lipid pool. However, by tuning spinecho delays in the localized 2D proton correlation sequence we are able to separate lactate CH2 and CH3 protons from the lipid resonances in the F1 dimension of the 2D-spectrum.

Purpose

In vivo assignment of lactate remains ambiguous. However, unambiguous visualization is necessary for understanding mechanisms associated with skeletal muscle ischemia in rodents, and for the assessment of treatment in the dysfunctional muscle of transgenic animals. In this study we address unambiguous detection of lactate methylene and methyl groups in the ischemic hindleg of mice, subjected to prolong ischemia. Using localized 2D proton correlation spectroscopy [1-3] at ultra-highfield (UHf-P-COSY). Despite the strong contribution of Bo inhomogeneity, in vivo we can unambiguously assign (CH2) methylene and (CH3) methyl groups of lactate in the F1 dimension of localized 2D spectra.

Methods

Ultra-highfield localized 2D proton correlation spectroscopy (UHf-P-COSY) sequence (see Figure 1) was implemented on a Bruker Biospec (9cm horizontal bore) MR system 11.7T using a home built ischemia setup. This setup consists of a 31P three turn solenoid coil surrounded with a 1H Alderman-Grant coil. B1 fields of both coils are perpendicular to one another, which prevents coupling between both coils. At first we used (Bruker) NMRSim software for simulating optimal spin echo times for lactate spin system at 11.7T (500MHz) (see Figure 2). Next we obtained phantom spectra for 40mM fructose and 10mM lactate solutions at room temperature (see Figure 3 and 4). Finally, we acquired in vivo 2D-UHf-P-COSY spectrum of the mouse hindleg subjected to 35 minutes ischemia. For in vivo experiment optimal echo time was set to 88ms, which was efficient for coherence transfer between lactate protons and for lipid suppression (Figure 5). All 2D spectra were processed and analysed using Bruker Topspin 2.1 software.

Results and Discussion

Excellent water suppression was achieved in all proton spectra. Figure 3 shows 40mM fructose spectrum dissolved in water. In this figure clear cross peaks and diagonal peaks between J-coupled protons of fructose can be seen. Chemical shift differences between protons allow their separation in the F1 and F2 dimensions of the 2D spectrum. The 2D spectrum of fructose demonstrates superior resolving power of 2D UHf-P-COSY. Figure 4 is the 2D-UHf-P-COSY spectrum of the lactate, poor shimming reveals strong contribution of Bo inhomogeneity on the diagonal and cross peak volumes. Although we used suboptimal echo time for coherence buildup, despite, clear cross and diagonal peaks at 1.2 and 4.0 ppm between methyl (CH3) and methylene (CH2) protons can be seen. In Figure 5, the in vivo hindleg spectrum show distinct cross peaks for CH2 and CH3 lactate protons, which now can be unambiguously assigned. Both cross peaks have very similar intensity, which suggest significant levels of lactate (mM) accumulates in the ischemic muscle.

Conclusions

At ultra-highfield (11.7Tesla), there is an advantage to detect metabolites by exploiting increased chemical shift separation as well as high SNR despite the strong Bo inhomogeneity. Localized UHf-P-COSY can be useful in the assessment of metabolites in the resting and hypoxic hindleg of mice at high temporal and spatial resolution.

Acknowledgements

No acknowledgement found.

References

[1]Thomas et,al MRM, 26:58 (2001) [2] Blackband et.al JMR 79:184-189 (1988) [3] Braakman et.al ,MRM, 60:449-456(2008)

Figures

Figure 1 Scheme of the pulse sequence for UHf-P-COSY. Where delay TE is the sum of (2*TE/4+TE/2+t1+∆) compensation delays, and t1+∆= t1max=(TE/2-TE’/2). Delay t1 gets incremented for each F1 time domain experiment. For each experiment t1 is incremented by ∆t1=1/2*SW, such that t1max=(t1(0)+n*∆t1), n= number of indirect sampling points in the (F1) dimension. Coherence operators present at the start of t1 evolution and immediately after the 90 degree mixing pulse is indicated by blue arrow.

Figure 2 Simulated spectra of the lactate molecule at variable spin echo delays. Relaxation properties of CH2 and CH3 group differ at (11.7T) 500 MHz.

Figure 3 Fructose phantom (40mM solution in 2ml tube) spectrum, NS=2, Voxel (2x2x2)=8ul, TE=11.4ms, TR=1.5sec, F2=2500 Hz, F1=2500 Hz, TD2=4096, TD1=128

Figure 4 Lactate phantom (10mM solution in 2ml tube) UHf-P-COSY spectrum. In the absence of good shimming strong Bo inhomogeneity affects diagonal and crosspeak relaxation properties. Voxel=8µl (2x2x2), TE=24ms, NS=2, F1=5000 Hz, F2=5000 Hz, TD2=2048 and TD1=96. Shorter echo times seems suboptimal for coherence transfer between CH2 and CH3 protons.

Figure 5 In vivo spectrum of the mouse hindleg subjected to 35 min ischemia. A single voxel 175µl (7mmx5mmx5mm) in hindleg muscle can be seen in the image.TR=1.2s, TE=88ms, NS=16, F2=5000 Hz, F1=5000 Hz, TD2=2048, TD1=64.



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
4020