Cerebrovascular reactivity in multiple sclerosis is restored with reduced inflammation during immunomodulation | Scientific Reports – Nature.com

Neurodegeneration in MS might be exacerbated by an inability of the cerebrovascular system, affected by inflammation, to maintain a constant nutrient supply16. This inability is expected to be causally linked to an altered CVR10. CVR in MS, therefore, may offer a relevant marker of disease status, progression and response to treatment. In this study, we demonstrated that a reduction of CVR exists in MS and that it can be, at least partially, reversed by spontaneous or treatment-induced reduction of inflammation.

We found a lower CVR in the GM and WM of MS patients, which were eligible to start treatment with IFN beta, compared to matched HC, a significant lower CVR within the Gd+lesions with respect to WM and a significant increase in CVR from the pre-treatment to the on-treatment period with IFN beta for both GM and WM which was also coupled with a reduction in the number of Gd+lesions (Figs.3, 4, 5, 6). IFN beta acts as an immunomodulatory agent, shifting cytokine networks in favour of an anti-inflammatory pattern17,18. CVR changes may reflect restoration of vascular integrity and its hemodynamic mechanisms induced by a reduction of inflammation. Several factors could mediate the association between reduction of inflammation and improvement of cerebrovascular function: oxidative damage resolution, with consequent stabilisation of the blood brain barrier (BBB)19; normalisation of endothelial function, with reduction of vasoconstrictive mediators, such as endothelin-120; modulation of vasodilatory mediators10. The hypothesis of the study is based on the latter factor: lowering levels of basal nitric oxide, a vasodilatory mediator whose high levels could desensitise endothelial and smooth muscle function (vascular habituation), may restore vasodilatory capacity. The extent of CVR increase in both GM and WM while on-treatment was negatively correlated with the pre-treatment CVR (Fig.4), supporting the notion of a restoration effect, in which greater recovery is associated with a greater initial impairment.

Diffuse increases in CVR were observed over the whole MS brain (Fig.5), but more significant clusters of positive change in CVR were observed in the posterior cingulate cortex, visual areas and cerebellum, in particular, the crus and lobules. This may reflect a selectively stronger impairment and thus greater recovery in regions more affected by MS, such as the visual cortex and cerebellum21,22,23, or it may reflect a larger sensitivity to BOLD signal changes in regions of greater cerebral blood volume (CBV)24. It is unlikely that this finding was related to a larger number of Gd+lesions in these regions, because lesions were mainly localised to the temporal and parietal lobes.

The spatial and longitudinal coupling between CVR and Gd+lesions strongly supports the hypothesis of an association of CVR with active lesions and inflammation (Fig.6). Patients whose MRI scans exhibited a larger reduction in the number of lesions showed a larger increase in CVR. Gd enhancement implies BBB breakdown, typically caused by an inflammatory response and damage due to tissue hypoxia25,26: this triggers further amplification of inflammatory processes and cerebrovascular impairment, resulting in leukocyte extravasation and release of cytokines27. Restoration of BBB integrity helps to reduce inflammation and to recover the neurovascular units function28. Of note, the positive findings were limited to regional and longitudinal effects. A limiting result of the study, probably related to methodological shortcomings, was the absence of significant differences in CVR between patients with and without Gd+lesions in any session. CVR imaging may therefore selectively complement Gd as a regional marker of brain inflammation or longitudinal marker of its changes.

We observed a pre-treatment association between low CVR and reduced GM volume in MS patients (Fig.7A,C). Assuming that greater brain atrophy is a marker of neurodegeneration, this result suggests an association of poor cerebrovascular status and inflammation with neurodegeneration. This association was lost while on-treatment. Since GM volume was not modified among the different sessions, this decoupling between CVR and volume in the GM is explained by the greater increase in CVR in patients with the larger initial vascular impairment, i.e., lower pre-treatment CVR (Fig.4). A strong selectivity of the effect in GM was observed. We hypothesise that, while degeneration of the WM is multifactorial and directly impacted by the dysregulated immune system, neurodegeneration of the GM is, as our study hypothesis suggested, associated with effects of inflammation on the vasculature which may lead to hypoxia-like tissue degeneration29,30. In fact, vascular effects may be less pronounced in tissue with diminished vascularization and slower metabolism, such as WM, compared to regions with higher vascularization and higher metabolic activity such as GM.

We did not detect a significant correlation of CVR and CVR changes with age, EDSS, disease duration and T2 lesion volume due to the eligibility criteria of the study that generated a statistically homogeneous cohort of MS patients: all patients had low EDSS score, disease duration and MRI lesion burden with an associated low variability. These criteria were designed to isolate the effects of acute inflammation on CVR, while minimising the influence of factors that could modify cerebral blood flow and hence, CVR31,32.

The study was designed to test for the effect of inflammation and its reduction on CVR. For this reason, we included a cohort of patients eligible to start treatment with IFN beta. This cohort was, by definition, inflamed and the changes observed with the introduction of IFN beta could be, at least partially, related to the natural history of MRI active lesions. The reasons for activity reduction, whether spontaneous or drug induced, are not separable in this study.

CVR was assessed through breath-holding, which is easy to implement but it can be affected by biases and diminished repeatability compared to the CVR measured using CO2 administration. To limit the shortcomings of this technique, we employed a standardised approach to perform breath holds, and we measured the PetCO2 as a surrogate of arterial CO2 pressure to account for intersubject variability in task performance. Finally, CVR was assessed in units of %BOLD signal modulation per mmHg of PetCO2 change. The change in BOLD signal during hypercapnia is determined by the relative change in CBF that may be altered both through stimulus-induced changes in CBF or by changes in baseline CBF values; these have been reported to be acutely increased in the WM, during active MS phase, but chronically decreased in GM6,33. The two factors affecting CVR cannot be decoupled without an independent measure of baseline CBF . The change in BOLD signal is also affected by baseline values of CBV and deoxy-haemoglobin concentration in blood34,35. Together with heterogeneous breath-holding performances and subtle temporal differences in the shape of the hemodynamic response following breath holds, the complex origin of the BOLD signal might explain the general robust spatial and longitudinal findings of the study and the weaker between-groups or between-subjects effects. Nonetheless, our CVR maps agreed in range and spatial distribution with previous studies reporting BOLD-CVR maps36. With these limitations in mind, our finding of increased BOLD CVR with diminished inflammation might be related to a decrease in baseline CBF, an increase in baseline CBV, or more complex changes in the shape of the hemodynamic response function, rather than an increased vascular responsiveness to the stimulus. These potential mechanisms are relevant as they too highlight the effect of inflammation on brain vasculature.

Original post:
Cerebrovascular reactivity in multiple sclerosis is restored with reduced inflammation during immunomodulation | Scientific Reports - Nature.com