Impaired cerebral vasoreactivity to CO₂ in Alzheimer's disease using BOLD fMRI

Abstract

Objective

To evaluate the cerebral vasoreactivity using blood oxygenation level dependent functional MRI during carbogen inhalation with 7% CO₂ in Alzheimer's disease and amnestic mild cognitive impairment.

Participants and methods

Thirty nine subjects were included to be investigated using blood oxygenation level dependent (BOLD) functional MRI at 1.5 T during a block-design carbogen inhalation paradigm, with a high concentration face-mask under physiological monitoring. Basal cerebral perfusion was measured using pulsed arterial spin labeling. Image analyses were conducted using Matlab® and SPM5 with physiological regressors and corrected for partial volume effect.

Results

Among selected participants, 12 subjects were excluded because of incomplete protocol, leaving for analysis 27 subjects without significant microangiopathy diagnosed for Alzheimer's disease (n = 9), amnestic mild cognitive impairment (n = 7), and matched controls (n = 11). No adverse reaction related to the CO₂ challenge was reported. Carbogen inhalation induced a whole-brain signal increase, predominant in the gray matter. In patients, signal changes corrected for gray matter partial volume were decreased (0.36 ± 0.13% BOLD/mmHg in Alzheimer's disease, 0.36 ± 0.12 in patients with mild cognitive impairment, 0.62 ± 0.20 in controls). Cerebral vasoreactivity impairments were diffuse but seemed predominant in posterior areas. The basal hypoperfusion in Alzheimer's disease was not significantly different from patients with mild cognitive impairment and controls. Among clinical and biological parameters, no effect of apoE4 genotype was detected. Cerebral vasoreactivity values were correlated with cognitive performances and hippocampal volumes. Among age and hippocampal atrophy, mean CVR was the best predictor of the mini-mental status examination.

Conclusion

This BOLD functional MRI study on CO₂ challenge shows impaired cerebral vasoreactivity in patients with Alzheimer's disease and amnestic mild cognitive impairment at the individual level. These preliminary findings using a new MRI approach may help to better characterize patients with cognitive disorders in clinical practice and further investigate vaso-protective therapeutics.

Introduction

Alzheimer's disease (AD) is the most common cause of dementia in the elderly and affects 25 million persons in the world. Recent epidemiological studies predict a doubling of the prevalence every 20 years (Ferri et al., 2005). Mild cognitive impairment (MCI) may correspond to a transition state before AD with an annual conversion rate of MCI to AD evaluated between 16% and 41% (Gauthier et al., 2006).

In pathology, AD is characterized by abnormal amyloid beta peptide deposition, neurofibrillary tangles, and neurodegeneration. Additionally, vascular disorders play a critical role in the occurrence of cognitive impairment and neurodegeneration (Bell and Zlokovic, 2009, Benarroch, 2007, Farkas et al., 2000, Iadecola, 2010, Zlokovic, 2005).

In AD, the underlying mechanisms of this vascular dysfunction rely on the association of acetylcholine depletion and amyloïd beta peptide deposition that induce functional vascular disorders (Claassen and Jansen, 2006, de la Torre and Stefano, 2000, Iadecola, 2010). The structural alterations of the neurovascular unit could be provoked by the functional alterations themselves (Farkas and Luiten, 2001).

Indeed, perfusion imaging studies showed a cerebral blood flow (CBF) decrease of 10–25% in temporal, parietal and frontal regions (Chao et al., 2010, Dai et al., 2009, Du et al., 2006, Johnson et al., 2005, Kogure et al., 2000). Nevertheless, these studies on basal perfusion have not investigated vasomotor properties. None of them affords individual diagnosis.

Besides regional hypoperfusion of the brain, vasomotor impairments have been detected in animal and humans (Iadecola, 2004, Mueggler et al., 2002, Princz-Kranz et al., 2010, Shin et al., 2007). In patients, functional alterations of the vessel contractility were assessed using a vascular property, the cerebral vasoreactivity (CVR) to circulating gases and particularly to CO₂. The capnic modulation is commonly induced by apnea or acetazolamide injection. Despite methodological and practical limitations, transcranial Doppler (TCD) and computed tomography during stable Xenon inhalation (CT-Xenon) have shown impaired CVR (Oishi et al., 1999, Silvestrini et al., 2006).

TCD may provide a breath holding index by measuring blood velocity in the proximal portion of the middle cerebral arteries. This parameter was related to the mini-mental status examination (MMSE) (Silvestrini et al., 2006). However, TCD during apnea is heavily dependent on the operator expertise and the patient cooperation. Moreover, TCD estimates a global parameter, and is not a brain imaging technique able to depict potential regional changes that may better characterize AD.

CT-Xenon after acetazolamide injection has shown a reduction of the consecutive hyperperfusion of 6–8% in frontal, parietal and temporal cortices (Oishi et al., 1999). However, CT-Xenon is not commonly performed in clinical settings and results were not demonstrative at the individual level.

In AD, recent advances in MRI confirmed impaired hemodynamics, especially using blood oxygenation level dependent (BOLD) contrast, a widely available and non-invasive technique to investigate brain function. Indeed, BOLD functional MRI (fMRI) allows mapping brain regions involved in cognitive tasks, by measuring signal changes induced by the neurovascular coupling. Besides potential neuronal dysfunction that may explain differences of activation across populations, a delayed hemodynamic response has been identified (Rombouts et al., 2005). Yet, signal modifications related to neuronal activity can hardly be distinguished from those related to hemodynamic changes (Brown et al., 2003, Buxton et al., 2004, Cohen et al., 2002). Thus, recent BOLD fMRI studies during hypercapnic challenge with CO₂ inhalation have been conducted to estimate CVR in steno-occlusive pathology (Haller et al., 2008, Mandell et al., 2008, van der Zande et al., 2005, Ziyeh et al., 2005) and neurooncology (Jiang et al., 2010). In this way, a recent work has shown a forebrain deficit of CVR in prefrontal, anterior cingulated and insular cortices in AD (Yezhuvath et al., in press). In AD, vascular dysregulation would advocate for further developments of vasoprotective treatments (de la Torre, 2010, Luzzi et al., 2010).

Therefore, we looked at whether basal perfusion and hemodynamic responses to controlled inhalation of carbogen (a gas mixture of 93% O₂ and 7% CO₂) could be altered in AD and patients at risk with amnestic MCI. These patients were compared to controls using basal perfusion imaging with arterial spin labeling and dynamic BOLD response to carbogen. We paid attention to select individuals without moderate or severe microangiopathy. We hypothesized regional CVR impairment in patients. We also estimated the potential interest of this method to help for diagnosis at the individual level.