0996

Quantification of trans-endothelial water exchange and vessel geometry using contrast-enhanced MRI and alterations in a transgenic rat model of Alzheimer’s disease

Ben Dickie¹, Hervé Boutin^1,2, Jose Ulloa^3,4, Laura M Parkes¹, and Geoff JM Parker³

¹Division of Neuroscience and Experimental Psychology, University of Manchester, Manchester, United Kingdom, ²The Wolfson Molecular Imaging Centre, University of Manchester, Manchester, United Kingdom, ³Division of Informatics, Imaging and Data Sciences, University of Manchester, Manchester, United Kingdom, ⁴Bioxydyn, Manchester

Synopsis

MRI measurements of cerebrovascular function and structure are important for understanding neurodegenerative disease mechanisms. In this study, a novel contrast-enhanced multi-flip angle multi-echo (MFAME) MRI technique capable of simultaneously quantifying vessel permeability surface area product to water (PS_w), blood water population fraction (p_b), and contrast agent r₂* is presented, and applied to a transgenic rat model of Alzheimer’s disease (TgF344-AD). Transgenic rats exhibit higher p_b and lower PS_w in the hippocampus compared to wild-types, suggesting MFAME MRI may be sensitive to regional pathologic microvascular alterations in this model.

Purpose

MRI measurements of cerebrovascular function and structure are important to aid understanding of neurodegenerative disease mechanisms and for use as imaging biomarkers sensitive to the action of novel targeted therapeutics. In particular, sensitive measurements of blood-brain-barrier (BBB) damage appears important^1,2. This work presents a rapid contrast-enhanced multi-flip angle multi-echo MRI protocol (MFAME-MRI) for combined measurement of mean blood water residence time (τ_b [s]), water permeability surface area product (PS_w [mL min^-1 mL^-1]), blood water population fraction (p_b), and in-vivo contrast agent r₂* [mM s]^-1 (a parameter theoretically dependent on vessel geometry)^3,4. Signal sensitivity and error analyses were performed. Optimised sequences were then applied to an aged rat model of Alzheimer’s disease⁵ (TgF344-AD) and differences between transgenic and wild-type animals evaluated.

Methods

Sensitivity and error analysis: Sensitivity of tissue MRI signals to τ_b , p_b , and PS_w (PS_w = p_bλ/τ_b, λ is the blood-tissue partition coefficient = 0.8) were evaluated using a generalized two-site exchange spoiled gradient echo signal model (2SX-SPGR)⁶:

$$S_{t}(TE = 0) = S_{0}\big[a_{s}\frac{sin(\theta)(1-e^{-\frac{TR}{T_{1,S}}})}{1-cos(\theta)e^{\frac{-TR}{T_{1,S}}}} + (1-a_{s})\frac{sin(\theta)(1-e^{-\frac{TR}{T_{1,L}}})}{1-cos(\theta)e^{\frac{-TR}{T_{1,L}}}}\big]$$

where S₀ is a scaling constant, θ is the flip angle, TR the repetition time, T_1,S and T_1,L, are the apparent T₁ relaxation times of short and long T₁ compartments, and a_s is the apparent water population fraction of the short T₁ compartment. Dependence of S_t(TE = 0) on p_b and τ_b arises through T_1,S, T_1,L, and a_s via known analytic expressions⁷. Monte Carlo simulations were performed to assess the effect of image noise, transmit B₁ error, and non-zero K^trans on p_b, τ_b, and PS_w.

MR Imaging: Figure 1 shows the MRI protocol. A rat model of early-onset Alzheimer’s disease⁵ (TgF344-AD; transgenic [n = 8], wild-type controls [n = 8]) was imaged aged 18 months on a 7T Bruker BioSpec MRI scanner (Bruker Corporation, Billerica, USA. At 18 months of age, transgenic TgF344-AD rats display increased amyloid-beta, tau, and neuronal loss, relative to wild-type controls⁵. The manufacturer supplied rat head coil was used for signal reception. Rats were anesthetised with 2-2.5% isoflurane + 100% O₂. Contrast agent (Gd-DOTA, 0.5 mmol/kg) was injected through a tail vein cannula at 1mL/min.

Image analysis: Figure 2 provides details of the image analysis pipeline. Briefly, flip angles were mapped to allow correction for B₁ error⁸. Regional MRI signals were extracted from the hippocampus, cerebellum, cortex, thalamus, and striatum using the Schwarz rat brain atlas⁹. Blood signal was taken from the superior sagittal sinus. An exponential model was fitted to each multi-echo readout, giving estimates of T₂*, and also providing decay correction to TE = 0. Native T₁ of blood and tissue were estimated from pre-contrast decay-corrected signals, assuming the fast exchange limit (FXL). Blood contrast agent concentration was estimated by fitting a bi-exponential clearance model to post-contrast decay-corrected blood signal. Parameters p_b and τ_b were estimated by fitting the 2SX-SPGR model to post contrast decay-corrected tissue signals. PS_w was computed as above. Contrast agent r₂* was estimated by fitting a linear model to 1/T₂*(t) = r₂*C_b(t). To quantify leakage of contrast agent across the BBB, K^trans was estimated by fitting the Patlak model² to image sets a, b and d (Figure 1). The null hypothesis of no difference in parameter estimates between transgenic and wild-type groups was tested.

Results

Sensitivity of tissue signal to p_b, τ_b, and PS_w, was maximal at the Ernst angle and at long TRs (Figure 3). Parameter estimate coefficient of variation for in-vivo data (both at region of interest and voxel level) was estimated to be less than 15% (Figure 4, top row). Biases in p_b, τ_b, and PS_w increased with flip angle error and K^trans (Figure 4, bottom row). Figure 5 shows parameter estimates in selected brain regions. Transgenic rats showed significantly elevated p_b (P = 0.027) and reduced PS_w (P = 0.019) in the hippocampus, relative to wild-type controls (Figure 5). No group differences in r₂* were observed. K^trans was < 2 ×10^-4 min^-1 in all rats. Estimates of p_b were in the range of previously published values for rat brain¹⁰. PS_w was overestimated by a factor of 1.5-2¹⁰.

Conclusion

Estimates of p_b, τ_b, and PS_w were shown to be precise and accurate for sources of error investigated. Decreased PS_w in the hippocampus of transgenic rats may reflect a reduction in the number of perivascular aquaporin channels¹¹. Overestimation of PS_w relative to published values could be due to animal age (i.e. age-related alterations), however it is possible that other sources of error not investigated in this study are responsible. Further work will aim to correlate PS_w with aquaporin-4 expression in the hippocampus.

Acknowledgements

Karen Davies for guidance and expertise during pre-clinical scanning. The Engineering and Physical Sciences Research Council for funding (grant EP/M005909/1).

References

1. Wardlaw et al. Is breakdown of the blood-brain barrier responsible for lacunar stroke, leukoaraiosis, and dementia? Stroke. 2003 Mar;34(3):806-12

2. Heye et al. Tracer kinetic modelling for DCE-MRI quantification of subtle blood–brain barrier permeability. Neuroimage. 2016. 15;125:446-55

3. Sourbron et al. Bolus-Tracking MRI with a Simultaneous T1- and T2*-Measurement. Magnetic Resonance in Medicine. 2009. 62:672–681

4. Yablonskiy et al. Theory of NMR Signal Behavior in Magnetically Inhomogeneous Tissues: The Static Dephasing Regime. Magnetic Resonance in Medicine. 1994 Dec;32(6):749-63.

5. Cohen et al. A transgenic Alzheimer rat with plaques, tau pathology, behavioral impairment, oligomeric aβ, and frank neuronal loss. J Neurosci. 2013 Apr 10;33(15):6245-56

6. Buckley et al. Cellular-interstitial water exchange and its effect on the determination of contrast agent concentration in vivo: dynamic contrast-enhanced MRI of human internal obturator muscle. Magnetic Resonance in Medicine. 2008 Nov;60(5):1011-9.

7. Landis et al. Equilibrium TranscytolemmalWater-Exchange Kinetics in Skeletal Muscle In Vivo, Magnetic Resonance in Medicine, 1999, 42:467–478

8. Balezeau et al. Mapping of low flip angles in magnetic resonance, Phys Med Biol, 2011, 21;56(20):6635-47.

9. Schwarz et al. A stereotaxic MRI template set for the rat brain with tissue class distribution maps and co-registered anatomical atlas: application to pharmacological MRI. Neuroimage. 2006 Aug 15;32(2):538-50.

10. Schwarzbauer et al. Quantitative Magnetic Resonance Imaging of Capillary Water Permeability and Regional Blood Volume With an Intravascular MR Contrast Agent. Magnetic Resonance in Medicine.1997;37(5):769-77.

11. Wilcock et al. Vascular amyloid alters astrocytic water and potassium channels in mouse models and humans with Alzheimer’s disease, Neuroscience. 2009 Mar 31;159(3):1055-69

Figures

The MRI protocol. T₁w images were acquired to provide an anatomic reference for regional signal extraction. Flip angle maps were acquired using the method of Balezeau et al. Pre-contrast MFAME images were collected to estimate native T₁ in blood and tissue (a). Post-contrast MFAME images were acquired for two purposes: to estimate p_b, τ_b, PS_w and r₂* in tissue (c) and to track the concentration of contrast agent in the superior sagittal sinus during injection of contrast agent and after acquisition of tissue signals, allowing the concentration during acquisition of tissue signals to be inferred (b and d).

Analysis pipeline with representative MFAME data and fits.

Tissue signal sensitivity plots for variable flip angles and repetition time (TR). Intermediate flip angles (20°-60°) show greatest sensitivity to p_b, τ_b and PS_w. Maximal sensitivity occurs at the Ernst angle. At low flip angles (0°-20°), signals are insensitive to all parameters, and large flip angles (> 70°) signals are insensitive to τ_b and PS_w. Therefore, using flip angles across the full range should provide parameter estimates that are both determinable and identifiable. Sensitivity to all parameters increases with TR, reflecting increased T₁-weighting.

Precision and bias in τ_b, p_b, and PS_w. Ground truth values were p_b = 0.05, τ_b = 0.1 s, PS_w = 24 mL min^-1 mL^-1 (a) and p_b = 0.05, τb = 0.5 s, PS_w = 4.8 mL min^-1 mL^-1 (b). Vertical dashed lines indicate signal to noise ratios for in-vivo data at the voxel and region of interest (ROI) levels. Flip angle errors led to overestimation of p_b and τ_b, and underestimation of PS_w (c). Non-zero K^trans resulted in a negative bias in PS_w (d). In (d), the vertical line indicates typical K^trans for intact BBB.

Regional parameter estimates in transgenic and wild-type animals. Transgenic animals display increased p_b and decreased PS_w in the hippocampus, relative to wild-type controls (P < 0.05). Hippocampal r₂* is elevated transgenic animals, but is not statistically different from wild-types. No other statistically significant differences were observed.

Proc. Intl. Soc. Mag. Reson. Med. 25 (2017)

0996