Aarhus CTH Meeting

Cenk Ayata

Neurovascular coupling during cortical spreading depolarizations

Spreading depression (SD) is an intense wave of pan-depolarization of virtually all cells within the brain tissue, which is the electrophysiological substrate of migraine aura, and an important pathophysiological component in brain injury states such as stroke. Besides the neuronal and glial ionic and metabolic changes, SD has a profound impact on cerebral vasculature. Although the prototypical cerebral blood flow (CBF) response to SD has traditionally been considered a monophasic hyperemia, accumulating data show a range of vascular responses that are highly reproducible but heterogeneous depending on the species, systemic physiology, drug effects, and the presence of pathological circumstances such as tissue ischemia or subarachnoid hemorrhage. Evidence strongly suggests that SD–induced CBF response is composed of multiple vasomotor components vary in magnitude and time course depending on local tissue and systemic physiological state. In my talk, I will review the range of CBF responses and discuss the vasomotor components in context of health and disease.

Janine Berkholz
Charité Berlin

The endothelial surface layer

In recent years the presence of a substantial layer on the luminal endothelial surface (endothelial surface layer, ESL) exhibiting a thickness in the range of 0.4 to over 1 µm has become widely accepted. The endothelial surface layer includes a deeper layer, directly bound to the plasma membrane called endothelial glycocalyx (EG) and a superficial, adsorbed layer, which is presumed to be bound to the endothelial cell membrane indirectly via the endothelial glycocalyx. The typical thickness of the glycocalyx, as seen by electron microscopy, amounts to only about 50-100 µm. Therefore it is assumed that additional associated components, e.g. adsorbed plasma proteins or hyaluronan, are essential in constituting the main part of the endothelial surface layer.

This very dynamic layer, which normally covers the surface of the endothelial cells, severely restricts the access of plasma and cellular components of the blood to the surface of the endothelium. Over the past decade, insight has been gained into the role of the thick endothelial surface layer in vascular physiology and pathology, including haemodynamic conditions, mechanical stresses acting on red cells in microvessels, oxygen transport, vascular control, blood coagulation, inflammation and atherosclerosis. Despite the increasing experimental evidence for the presence of the endothelial surface layer, its exact biochemical organization and the distribution of the involved macromolecules as well as their functional roles are not yet sufficiently understood. 

Richard B. Buxton
University of California San Diego, USA

Modeling neurovascular coupling and the BOLD signal

Functional MRI is based on measuring the blood oxygenation level dependent (BOLD) signal, which is driven by the mismatch of changes in cerebral blood flow (CBF) and oxygen metabolism (CMRO2). Quantitative fMRI methods, combining measurements of the BOLD signal and CBF with an arterial spin labeling method provide a window on the dynamics of blood flow and oxygen metabolism in the human brain. Recent experimental and modeling studies from our group highlight the difficulty of interpreting the BOLD signal on its own, finding that the coupling ratio of the CBF and CMRO2 responses to a stimulus varies within the same brain region depending on stimulus strength, attention, adaptation and the effects of caffeine. Three speculative hypotheses related to these findings will be discussed:

Excitation/Inhibition: Does variability of the CBF/CMRO2 coupling ratio reflect the ratio of evoked inhibitory to excitatory neural activity?

Preserving a high tissue pO2: Is the basic function of the rapid and large CBF change to prevent a drop in tissue pO2?

Capillary flow heterogeneity: Could CBF control mechanisms acting to preserve tissue pO2 in the face of growing capillary heterogeneity lead to the Jesperson-Ostergaard effect—an initial CBF increase followed by a decrease as the pathology evolves?

Turgay Dalkara 
MD, PhD, Professor, Hacettepe University, Ankara

Cortical spreading depression and migraine

Initial phase in the development of migraine headache is still poorly understood. In this presentation, we will review a previously unknown signaling pathway between stressed neurons and trigeminal afferents during cortical spreading depression (CSD), the putative cause of migraine aura (Karatas et al, 2013). Recent findings from the mouse brain suggest that neuronal cellular stress may induce headache by activating the trigeminovascular system via a complex parenchymal inflammatory cascade initiated by transient opening of pannexin1 large-pore channels, formation of the inflammasome complex and activation of caspase-1 in neurons. Pro-inflammatory mediators (e.g. HMGB1, IL-1β) are released from neurons and cause translocation of NFκB to nucleus in astrocytes, which leads to induction of inflammatory enzymes such as COX2 and iNOS. The activity of these enzymes cause continuous release of cytokines, prostanoids and NO from the astrocyte end-feet forming the glia limitans to promote sustained activation of trigeminal nerve fibers around pial vessels in the subarachnoid space and triggers headache 20-60 minutes after CSD. Inhibiting pannexin-1 channels and HMGB1 with genetic or pharmacological means as well as NFκB activation and COX inhibitors prevented trigeminal activation. These findings suggest that pannexin-1 channels may serve to detect neuronal stress and alarm the organism with headache by activating a parenchymal inflammatory response. The above summarized inflammatory cascade also explains the delayed appearance of headache 20–60 min after aura (CSD) in migraineurs.

The CSD-induced trigeminal activation also initiates a parasympathetic reflex, release of vasoactive peptides from the trigeminal nerve endings and dural mast cell degranulation, which lead to dilation of the middle meningeal artery and plasma protein extravasation (Bolay et al., 2002). This sterile dural inflammation (also known as neurogenic inflammation) is though to amplify and sustain the algesic signals coming from the parenchyma and, hence, lead to sensitization of the first, second and finally third order neurons in pain pathways causing hours to days lasting throbbing headache and allodynia (perception of non-painful sensory signals such as simple touch as painful). The neurogenic inflammation theory has been instrumental for development of two classes of highly effective anti-migraine drugs, triptans and CGRP antagonists. CSD hypothesis has provided insight to the mechanism of migraine prophylaxis; more than one month (but not acute) administration drugs used in prophylaxis such as topiramate, valproate, propranolol, amitriptyline and methysergide have been found to suppress CSD frequency and increase CSD induction threshold in the rat in parallel with the time required for effective prophylaxis in their clinical use (Ayata et al., 2006).

  • Karatas et al., Science. 2013; 339:1092-5
  • Bolay et al., Nature Medicine 2002; 8:136–42.
  • Ayata et al., 2006; Ann Neurol, 59:652–61

Simon Fristed Eskildsen
CFIN, Aarhus University

Capillary dysfunction in Alzheimer's Disease

Disturbances in the neurovascular control mechanisms that secure the blood supply of the brain have long been suspected of playing a part in the development of Alzheimer’s disease (AD), giving rise to the hypothesis of neurovascular dysfunction (Girouard and Iadecola, 2006). In support of this hypothesis, evidence suggests that vascular changes and hypoperfusion are intimately involved in the etiopathogenesis of AD (Iadecola, 2010; Kalaria, 2010; Pantoni, 2010; Zlokovic, 2011). In AD, however, blood vessels and blood flow is only mildly affected when cognitive symptoms develop, making it difficult to explain the development of AD in terms of limited blood flow or oxygen supply.

A less studied feature of AD is the morphological changes that happens in the walls the capillaries. Such changes are likely to disturb, but not to block, the flow of blood through tissue. Changes in capillary flow patterns invariably increase the proportion of erythrocytes that pass through the capillary bed too fast to permit any significant extraction of oxygen by the tissue (Jespersen and Ostergaard, 2012). Direct microscopy studies indeed show that capillary flows shift to a more homogeneous pattern during episodes of increased metabolic needs. This mechanism serves to maintain efficient oxygen extraction during episodes of high flow by reducing this effective 'shunting'. If, however, changes in capillary morphology prevent this homogenization, the consequences are striking: the resulting effective shunting of blood was shown to invalidate the conventional paradigm, that normal cerebral blood flow (CBF) values signals ample oxygen supplies, and that increased CBF leads to increased oxygen availability at the level of neurons (Jespersen and Ostergaard, 2012). The classical, single capillary model of tissue oxygenation has recently been extended to include the effects of capillary transit time heterogeneity (CTH) (Jespersen and Ostergaard, 2012). Increasing CTH is referred to as capillary dysfunction and envisioned to be the result of the gradual changes in capillary wall morphology and function that accompany AD risk factors (Ostergaard et al., 2013).

Perfusion properties can reliably be estimated from dynamic susceptibility contrast magnetic resonance imaging (DSC-MRI) (Mouridsen et al., 2006), and recently it has been shown that the maximum oxygen extraction fraction from the blood can be estimated (Mouridsen et al., 2014) using the extended capillary model of tissue oxygenation (Jespersen and Ostergaard, 2012). In this presentation, I will report our latest findings with DSC-MRI to measure cerebral perfusion properties in AD patients and healthy controls for the purpose of evaluating the mechanisms predicted by the capillary dysfunction hypothesis of AD (Ostergaard et al., 2013).

Eszter Farkas
Department of Medical Physics and Informatics, University of Szeged, Hungary

Blood flow and oxygenation during global and focal ischemia

Spontaneous, recurrent spreading depolarizations (SD) occur in patients of subarachnoid hemorrhage, traumatic brain injury and ischemic stroke during the days following the primary injury. These SDs are associated with the progression of secondary, delayed brain injuries subsequent to the primary insult, and worsen clinical outcome, probably via atypical cerebral blood flow responses.

We have developed a multi-modal imaging technology for the synchronous visualization of SD-related local changes in (i) membrane potential (voltage-sensitive dye method), (ii) cerebral blood flow (laser speckle contrast analysis), (iii) cerebral blood volume (intrinsic optical signal, IOS at 540-550 nm illumination), and (iv) hemoglobin (Hb) saturation (IOS at 620-640 nm illumination) in relevant rat models of cerebral ischemia. We aimed to directly relate hemodynamic responses to membrane potential changes typical of ischemic SDs, with high spatial and temporal resolution.

Incomplete global forebrain ischemia was induced by the permanent, bilateral occlusion of the common carotid arteries, while multifocal forebrain ischemia was created by the unilateral intracarotid infusion of polyethylene microspheres (d=45 to 53 mm).

In the incomplete global ischemia model, the dominant type of neurovascular coupling was inverse (i.e. hypoperfusion rather than hyperemia with SD). Moreover, we propose that Hb had been completely desaturated by the time SD occurred in these experiments, as the SD-related changes in red IOS appeared almost identical with those of green IOS.

In the multifocal forebrain ischemia model, transient functional hyperemia developed with most SDs. The SD-related red IOS variation appeared to be biphasic: an early increase in signal intensity (Hb saturation) coincided with the initial phase of CBF elevation, which was followed by a sharp decrease in red IOS below baseline (Hb desaturation), the start of which preceded the peak of hyperemia. The hyperemic CBF responses to propagating SDs were coupled with three types of Hb saturation kinetics: (1) dominant initial phase (i.e., accentuated transient increase in signal intensity), (2) approximately equal magnitude of the two phases, and (3) dominant later phase (i.e., major transient reduction in signal intensity).Based on these findings, a fraction of ischemia-induced SDs is suggested to have metabolic consequences that may be beneficial to the ischemic tissue (increase of red IOS with SD, implying Hb saturation), and and SDs coupled with dominantly decreasing IOS signatures (sign of Hb desaturation) are suspected to be malignant.

Jozien Goense
University of Glasgow

Neurovascular coupling and the BOLD signal across cortical layers

Although fMRI is one of the most widely used techniques in cognitive neuroscience, exactly how the blood oxygenation level dependent (BOLD) signal reflects neural processing remains unclear. This limits our interpretation of fMRI signals and the potential of clinical applications of fMRI. There are many neural processes that influence the energy balance and the hemodynamic response, such as synaptic activity (neurotransmitter release and associated processes), action potentials, excitatory vs. inhibitory activity, feedforward vs. feedback processing, and neuromodulation. A way to differentiate these neural processes is by exploiting the functional segregation in the cortical layers. The different cortical layers have different functions, energy use and connectivity. We use high-resolution fMRI in combination with electrophysiology to study laminar processing, with the aim of gaining insight into the different layer-dependent neural processes and into the mechanisms of neurovascular coupling. The vascular responses are regulated at sufficiently small spatial scales to allow the use of fMRI to distinguish responses in the cortical layers. I will discuss our work on high-resolution fMRI and electrophysiology in the primary visual cortex (V1) and temporal lobe of awake and anesthetized macaques, and our most recent work on the mechanisms of the BOLD response and the effect of neuromodulators on these processes. We exploit the well-studied layered structure of V1 to investigate whether laminar differences in the BOLD, cerebral blood flow (CBF) and cerebral blood volume (CBV) responses can be detected for excitatory and inhibitory stimuli. We found that the mechanisms for the positive and negative BOLD responses differ, but also that neurovascular coupling differed in the cortical layers. Furthermore, neuromodulators such as dopamine can alter neurovascular coupling. Our results suggest that neurovascular coupling differs depending on factors such as the stimulus and neuromodulation, and that the combination of high-resolution fMRI with electrophysiology can be used to resolve neurovascular coupling in functional microcircuits. 

Clare Howarth
Co-authors: B.A. Sutherland2, H.B. Choi1, C. Martin2,3, B. L. Lind4, L. Khennouf4, J.M.P. Pakan1, G.C.R. Ellis-Davies6, M.J. Lauritzen4,5, N.R. Sibson2, A.M. Buchan2 & B.A. MacVicar1 

1University of British Columbia, Vancouver, 2University of Oxford, Oxford, 3University of Sheffield, Sheffield, 4University of Copenhagen, Copenhagen, 5Glostrup Hospital, Glostrup, 6Mount. Sinai School of Medicine, New York

Hypercapnia-evoked cerebral blood flow responses: a role for astrocytes and glutathione

Astrocyte calcium signalling has been shown to increase the diameter of local brain arterioles. However, it remains unclear under what circumstances such calcium signals contribute to cerebral blood flow regulation. In this work, we demonstrate that hypercapnia evokes an increase in both astrocyte calcium transients and cerebral blood flow in vivo.

Following an increase in calcium within the astrocyte, arachidonic acid can be generated which is then metabolised to vasoactive substances including PgE2, a vasodilator. Previous work has suggested that PgE21 and cyclooxygenase-12,3 (an enzyme involved in the metabolism of arachidonic acid to PgE2) are involved in the cerebral blood flow response to hypercapnia.

In vitro, we show that increased astrocyte calcium transients result in the downstream release of PgE2 via cyclooxygenase-1. This production of PgE2 is sensitive to glutathione levels within the tissue. When glutathione levels are pharmacologically depleted, PgE2 release and astrocyte calcium-evoked vasodilations in brain slices are reduced.

Using in vivo multiphoton microscopy we reveal that hypercapnia evokes an increase in calcium transients within astrocytes, but not within neurons. In vivo, we demonstrate that hypercapnia-evoked increases in cerebral blood flow are attenuated when glutathione levels are depleted. We further confirm that this increase in cerebral blood flow is dependent on cyclooxygenase-1 activity.

These data suggest that astrocyte calcium transients and subsequent glutathione-dependent PgE2 release are involved in hypercapnia-evoked cerebral blood flow increases.


  1. Wagerle, L. C. & Degiulio, P. A. Indomethacin-sensitive CO2 reactivity of cerebral arterioles is restored by vasodilator prostaglandin. Am J Physiol 266, H1332-1338 (1994).
  2. Wagerle, L. C. & Mishra, O. P. Mechanism of CO2 response in cerebral arteries of the newborn pig: role of phospholipase, cyclooxygenase, and lipoxygenase pathways. Circ Res 62, 1019-1026 (1988).
  3. Niwa, K., Haensel, C., Ross, M. E. & Iadecola, C. Cyclooxygenase-1 participates in selected vasodilator responses of the cerebral circulation. Circ Res 88, 600-608 (2001).

Patrick Jenny
ETH Zürich

Multi-scale modeling of oxygen transport in the cerebral microvasculature

Understanding the brain’s energy metabolism is crucial for a wide range of research topics. It is well known that the vasculature of the brain is able to adapt to a locally increased energy demand resulting from neuronal activation. However, the detailed signaling pathways and the level of changes in the cerebral microvasculature are still fairly unknown, especially in the capillary bed, where most of the gas exchange occurs. We developed a multi-scale approach to study the oxygen supply at capillary level and focus on the impact of individual RBCs.

On the larger scale, we simulate blood flow in capillary networks with discrete tracking of individual RBCs. Our numerical model accounts for three well-established hemodynamic effects in the microcirculation: the Fahraeus, the Fahraeus-Lindqvist and the phase-separation effect, which occurs at bifurcations. We study the impact and interplay of these phenomena on the overall network dynamics. The distribution of RBCs throughout the network is strongly affected by the behavior of RBCs at divergent bifurcations. Our results suggest that the distribution of RBCs at divergent bifurcations influences the bulk flow velocities in the daughter branches, and vice versa. Given certain conditions the bulk flow velocities in both outflow branches will be equalized due to the impact of RBCs (well-balanced bifurcation). In artificial capillary networks we could show that at well-balanced divergent bifurcations capillary dilation leads to a local accumulation of RBCs, whereas the flow rate remains nearly constant. Currently, we are analyzing the impact of capillary dilation in realistic capillary networks. In contrast to previous models, our numerical approach enables us to comment on transient effects. Furthermore, we are able to measure transit times of single RBCs and to compute trajectories of individual RBCs.

On the smaller scale, we use the provided RBC trajectories for the simulation of oxygen transport to tissue. In each moving RBC, the non-equilibrium kinetics between hemoglobin and oxygen are simulated. The resulting hemoglobin saturation is used to obtain a detailed map of the oxygen partial pressure (PO2) in both the vasculature and the surrounding tissue. In particular, the oxygen flux out of each RBC is known. While most oxygen transport models rely on the averaged hematocrit in each capillary, our model accounts for the well-known heterogeneity of blood flow in capillaries. We are now able to quantify the impact of RBC flow variations on tissue oxygenation. For example, our model produces erythrocyte-associated transients (EATs) that were successfully compared to recent measurements (Parpaleix et al. Nature Medicine, 19(2):241:246 (2013)). The comparison of longitudinal PO2 gradients in capillaries with these experimental data also led us to the hypothesis that capillary networks are denser on the venular side than on the arteriolar side. In the future, we will investigate the impact of capillary dilations, arteriolar flow increase and metabolic rate changes on oxygen transport in realistic capillary networks.

This multi-scale model for RBC dynamics and oxygen transport will be an essential tool to quantitatively assess different scenarios for functional hyperemia in health and disease.

Tess Kornfield
University of Minnesota

Regulation of Blood Flow in the Retinal Trilaminar Vascular Network

Blood flow in the retinal vasculature is closely tied to neuronal activity. Though large arterioles are known to modulate blood flow, the role of capillaries in actively regulating local blood flow is currently under debate. Pericytes, contractile cells that wrap around capillaries, may actively control capillary diameter, but results from brain studies are conflicted and in vivo studies in the retina have not been conducted. Moreover, blood flow regulation within the three distinct vascular layers of the retina has not been characterized. Our goal was to determine which vessels in the retina actively regulate blood flow and to investigate the role of capillaries in controlling blood flow. To accomplish this, we characterized flickering light-evoked diameter and flux changes throughout the retinal vascular network in the in vivo rat retina.

The largest light-evoked dilations occurred in the primary and secondary arterioles, although capillaries and primary venules also dilated. The magnitude of capillary dilations was not correlated with pericyte proximity. All blood vessels displayed light-evoked flux increases. These flux increases were most heterogeneous in capillaries. Strikingly, capillaries in the intermediate capillary layer, which lies at the outer border of the inner plexiform layer, displayed large, slowly developing dilations that were not seen in the other two capillary layers. Flux increases in the intermediate capillary layer were larger than increases in either the superficial or deep layers.

Due to their large dilations, we conclude that primary and secondary arterioles are primarily responsible for driving neuronally induced blood flow increases in the rat retinal vascular network. The differential dilation and flux responses within the three layers of the retinal vascular network suggests that blood flow in capillaries is actively regulated. Additional work will be required to determine what mechanisms generate this active regulation.

Jonghwan Lee
Harvard Medical School

Assessing the microcirculation by optical ­ coherence tomography

As capillaries exhibit heterogeneous and fluctuating dynamics even during baseline, techniques measuring red blood cell (RBC) flow properties over many capillaries at the same time will be very useful. Here, we report that dynamic OCT imaging can capture individual RBC passage; utilize this finding to develop a technique of quantifying RBC flux, speed and density over many capillaries at the same time; and introduce another technique for rapid volumetric imaging of capillary network flux dynamics during functional activation.

We repeated 1024 B-scans with a 4-ms time gap at a fixed cross-sectional plane of the cortex embracing many capillaries, and found a number of peaks representing RBC passage in the OCT intensity signal time courses at the voxels of the capillary centers. This conclusion that the peaks represent individual RBC passage is supported by (i) such peaks were not observed in the time courses at the neighboring voxels 10-μm apart from the capillary center; and (ii) these peaks moved through capillary paths so that stripe patterns appeared when scanned along the paths. This finding enabled us to quantify RBC speed (mm/s) and flux (RBC/s) simultaneously over many capillaries located at different depths. The flux was estimated from the number of peaks per unit time, and the speed was estimated from the mean peak width. The speed measurement was validated by comparing with another estimation from the slope in the stripe pattern (similar to the traditional two-photon line scanning technique). By repeating cross-sectional dynamic imaging and its analysis over the volume of interest, the presented technique enabled us to obtain 3D maps of capillary RBC flux, speed, and linear density. The RBC linear density was obtained by the simple relation (flux) = (speed) × (density). Compared to Doppler OCT, our technique identifies individual RBC passage and thus enables direct measurements of RBC speed, flux, and density in capillaries even when they lay in the transverse direction.

Based on the above findings, we developed another technique for more rapid volumetric imaging of capillary network flow. Rapid volumetric imaging is required for functional studies where capillary flow dynamics should be imaged with 1-s temporal resolution during functional activation. For this purpose, we defined a novel metric, statistical intensity variation (SIV), and validated that its mean averaged along a capillary path is proportional to RBC flux, through theoretical simulation and experimental comparison. Acquiring SIV volume data only requires repeating two B-scans, rather than continuously repeating B-scans at a fixed plane as in the first technique, so that much higher temporal resolution can be achieved. We developed algorithms to automatically trace and vectorize capillaries from the volume data and thereby average the SIV and estimate the RBC flux for each capillary. In consequence, SIV imaging enabled us to measure RBC flux over hundreds of capillaries with 1-s temporal resolution during functional activation.

The techniques will be used to study how capillary network flow varies in response to brain activation, and also will be applicable to human ophthalmology studies.

Fiona Moreton
University of Glasgow

CADASIL: A neurovascular perspective

Cerebral autosomal dominant arteriopathy with subcortical infarcts and leucoencephalopathy (CADASIL) is the most common inherited cause of stroke, and is characterised by a widespread vasculopathy. Mutations in the NOTCH3 receptor classically lead to loss or gain of cysteine residues in the extracellular domain.

This talk will review the current disease hypothesis that abnormal NOTH3 extracellular domain is improperly cleared and acts as a nidus for aggregation of proteins, leading to the development of granular osmiophilic material. Both vascular smooth muscle cells and pericytes express NOTCH3 and are therefore affected by CADASIL pathology, but it is proposed abnormal pericytes may be responsible for blood-brain barrier breakdown and changes in brain haemodynamics. The evidence of vascular functional abnormalities in vitro and in vivo will be examined and recent unpublished myography studies on gluteal biopsy samples from CADASIL patients will be discussed. These have shown evidence of impaired endothelium-dependent and independent vasodilatation.  Work from an ongoing longitudinal study examining peripheral and cerebral perfusion and reactivity in CADASIL will be discussed, which suggests impaired reactivity may be associated with volume of lacunes. Finally some recent collaborative work examining capillary transit time heterogeneity in CADASIL patients will be reviewed and some proposed future directions for CADASIL research discussed. 

Roland Pittmann
Dept. of Physiology and Biophysics, Virginia Commonwealth Univ, Richmond, Virginia

Regulation and Oxygen Transport in the Microvasculature

Most cells, because of their reliance on oxidative phosphorylation to produce adenosine triphosphate (ATP), need a continuous supply of O2.  Thus, when the activity of a group of cells changes, the rate at which they consume O2 will change and this altered O2 consumption will be met by a change in O2 delivery and/or O2 extraction.

The Krogh-Erlang model (1918) for O2 diffusion from blood to tissue remains a key conceptual element in models of O2 transport.  This model was designed for skeletal muscle and assumed that blood was a continuum and O2 diffusion and consumption were uniform within the tissue; it has since been applied to other tissues.  In 1970 Duling and Berne reported the first systematic examination of O2 levels in microvascular networks by using microelectrodes to measure the partial pressure of O2 (PO2).  They found that PO2 fell progressively from large arteries to the capillaries, suggesting a substantial pre-capillary loss of O2.  Subsequent development of spectrophotometric techniques to measure hemoglobin O2 saturation (SO2) in arterioles, venules and capillaries demonstrated that indeed SO2 also declined in step with PO2 as predicted by the O2 dissociation curve.  The introduction of the phosphorescence quenching technique to measure PO2 opened the possibility to extend PO2 measurements to new areas.  Examples are measurement of “erythrocyte associated transients” in PO2 (EATs), local O2 consumption and its PO2 dependence.

The role of O2 in local regulation of blood flow has been studied since 1876.  When the activity of a group of cells increases (e.g., muscle contraction, activation of neurons), there is an increase in blood flow in the nearby microvessels.  How do those microvessels know to increase flow and how is this information/signal transmitted?  The classic metabolic vasodilator mechanism postulates that O2 consumption increases in the activated cells, making them hypoxic.  Vasodilator products of the increased metabolism are released from the active cells and they diffuse through the interstitial space to the nearby arterioles, relaxing the smooth muscle of those vessels with an accompanying increase in blood flow and O2 delivery.  The concept of a mobile O2 sensor was proposed in 1995 by Ellsworth et al and independently by Stamler and co-workers.  The idea is that, as hemoglobin becomes deoxygenated, there is a conformational change from the oxy to the deoxy form and this triggers the release of a vasodilator from the erythrocyte, either ATP or nitric oxide (NO).  A recent proposal to explain O2-linked metabolic regulation of blood flow involves the two signaling radicals NO and superoxide (O2-) (Golub & Pittman, 2013).  In this case NO is the vasodilator, produced continuously through eNOS in endothelial cells.  Superoxide is produced by NAD(P)H oxidase in the plasma membrane of parenchymal cells.  The O2- diffuses through the interstitium and reacts avidly with NO, thereby modulating the local NO concentration and the associated vasodilator influence.  When the activity of a group of cells increases, the increased O2 consumption causes a fall in local PO2 and NADH, the other substrate for O2-, moves from the cytosol to the mitochondria.  This reduction in both substrates for O2- production leads to a rapid rise in NO, resulting in vasodilation and increased blood flow.

Franck Plouraboué
IMFT Toulouse

Modelling micro-vascular blood flow from  tissue imaging: methods and results

Advances in physiological imaging provides astonishing new measurements to nourish and challenge modeling's predictions. Yet most valuable, local measurements are difficult to embrace in a more global picture, so that image analysis and modeling  is needed. Many methodological issues are raised by image/model coupling approaches. Some of them have been already partially addressed in specific contexts, whilst others are still open.

Since Image analysis issues depend on acquisition protocols and modalities, the solutions to be found are case specific. Nevertheless there is some generic foot-step to be follow for vessel segmentation and vectorization that will be reviewed. We will underline methodological bottlenecks associated with the quality of the network reconstruction and will also discuss artifact corrections. Furthermore, the basic steps for micro-vascular blood flow modelling will also be described, so as to stress their limits and challenges.

Finally, the presentation will expose some modelling results as well as introducing operational concepts of vascular territory, couplings, robustness and overlap  to be useful in a very broad normal and pathological contexts.

Axel Pries
B. Reglin1 and A.R. Pries1,2
1Dept. of Physiology, Charité Berlin, Germany, 2Deutsches Herzzentrum Berlin, Germany

Oxygen sensing and the metabolic control of microvascular networks

Purpose: Metabolic regulation of blood flow is central to guarantee adequate substrate supply of tissues and microvascular network stability. Vascular reactions to local oxygenation are assumed to match blood supply to tissue demand in steady state and in response to exercise and/or tissue hypoxia via negative feedback regulation: Low oxygen tension is assumed to induce the release of vasoactive substances which trigger increase of vessel number and diameter and thus of blood flow and oxygen supply (angioadaptation). Here, we try to integrate experimental and modelling findings to derive general schemes of metabolic regulation of angioadaptation, with a focus on control of vessel diameter.

Methods: The different components and signalling chains of feedback control of vessel diameter were analysed with respect to their suitability to result in functionally adequate and stable complex vascular networks under steady state and non-steady state conditions.

Results and Conclusions: A functional scheme of metabolic regulation is derived considering a) main requirements for vascular beds: sufficiently low diffusion distances between vessels and tissue cells, sufficiently homogenous perfusion under steady state conditions, adequate perfusion increase during increased demand, b) possible biological mechanisms: angiogenesis, structural diameter adaptation, change of vessel tone, and c) implementation of these mechanisms via feedback loops on different time scales.

The following hypotheses are proposed: 1) In addition to oxygen dependent metabolic signalling, metabolic vascular regulation independent of local oxygenation or metabolic situation is established by the ‘dilution effect’. 2) Control of resting vessel tone, and thus perfusion reserve, which has been shown to differ between tissues, can be explained by vascular responses to transient phases of insufficient blood supply: ‘vascular memory’. 3) While for steady state adjustment of luminal vessel diameter a location of structures producing metabolic signals in or near to the vessel wall is optimal, signalling by red blood cells may be well suited to amplify perfusion during transient increases of tissue oxygen demand.

Sava Sakadzic
Optics Division, Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Harvard Medical School, Charlestown, Massachusetts, USA

Large arteriolar component of oxygen delivery implies safe margin of oxygen supply to cerebral tissue

The global architecture of the blood supply to the cortex consists of a planar mesh of pial arteries and veins that dive into the cortex supplying the complex microvascular network and draining the blood back to the surface. However, in spite of extensive efforts in brain and in other organs, the detailed intravascular oxygen distribution along the microvascular paths that connect pial arteries and veins remains largely unknown. Therefore, we have limited knowledge about the mechanisms that secure sufficient oxygen delivery in microvascular domains during brain activation, and provide some metabolic reserve capacity in diseases that affect either microvascular networks or the regulation of cerebral blood flow (CBF). Such information is therefore critical for our understanding of not only normal brain physiology, but also the relation between progression of microvascular dysfunction and neurodegeneration in various brain diseases, and for attempts to develop a quantitative interpretation of existing and emerging brain imaging modalities.

To start addressing these questions, we used Two-Photon Microscopy to measure PO2 in a large subset of arterioles, venules, and capillaries at different levels of CBF, and to obtain microvascular morphology. We exploited a Doppler Optical Coherence Tomography to acquire CBF in penetrating arterioles and surfacing venules. The measurements were combined with a detailed analysis of the microvascular morphology and with computation of oxygen delivery from an anatomical vascular model under different levels of oxygen metabolism. We have found that arterioles are responsible for 50% of the extracted O2 at baseline activity. Most of the remaining O2 exchange is taking place at the level of the first few capillary branches after precapillary arterioles, while majority of the capillaries (those of higher branching orders) on average release little O2 at rest. Our measurements and modeling results support this finding showing that high branching order capillaries may act as a dynamic O2 reserve that is recruited on demand to ensure adequate tissue oxygenation during increased neuronal activity or decreased blood flow. Our results challenge the common perception that O2 is almost exclusively released from the capillaries and provide a novel understanding of the distribution and dynamics of O2 extraction along the capillary paths in the cortex.

Timothy W. Secomb
University of Arizona

Network hemodynamics and oxygen diffusion in the brain

Oxygen transport to the brain may be regarded as the most critical function of the circulatory system. Because oxygen can diffuse only a short distance (of order 50 microns) into oxygen-consuming tissue, a dense network of microvessels carrying oxygenated blood is necessary to ensure that all tissue points are adequately supplied. Using a Green’s function method, we simulated oxygen delivery by a three-dimensional network of microvessels in rat cerebral cortex, and predicted the distribution of partial pressure of oxygen (PO2) in tissue and its dependence on blood flow and oxygen consumption rates. This simulation includes the effects of the radial gradients of PO2 within and outside each microvessel that drive diffusive transport into the tissue, and of the axial decline of PO2 along each microvessel as oxygen is extracted. In a typical control state with consumption 10 cm3O2/100cm3/min and perfusion 160 cm3/100cm3/min, the predicted minimum tissue PO2 was 7 mmHg. In comparison, a Krogh-type model with the same density of vessels, but with uniform spacing, predicted a minimum tissue PO2 of 23 mmHg. With a 40% reduction in perfusion, tissue hypoxia (PO2 < 1 mmHg) was predicted. These results suggest that the normal microcirculation operates with a relatively small ‘safety’ margin of excess supply relative to basal requirements. Although one might intuitively expect that hypoxia provides a feedback signal for the short-term regulation of blood flow to ensure tissue oxygenation, a substantial amount of evidence argues against this mechanism. Nonetheless, it appears that the structure of the brain microvasculature is finely tuned for oxygen delivery. As a resolution of this apparent paradox, we suggest that the structural control of the brain vasculature, through the processes of angiogenesis and vascular remodeling, is sensitive to the occurrence of tissue hypoxia, thus providing the necessary feedback control on a slow timescale. Supported by NIH grant HL070657.

Aaron Simon

Challenges in measuring blood flow and oxygen metabolism dynamics in humans 

Blood oxygenation level dependent (BOLD) functional magnetic resonance imaging (fMRI) has fundamentally changed the practice of human neuroscience research over the last two decades, providing a simple, robust, and non-invasive method of localizing brain regions associated with cognitive tasks and sensory stimuli. Recently, BOLD fMRI has also begun to reveal the dynamic nature of the brain’s functional anatomy as naturalistic experiments demonstrate the intricate patterns of BOLD signal fluctuations that occur under conditions as varied as listening to music, watching film, and resting quietly. However, interpretation of these dynamic BOLD fluctuations is made difficult by the complex nature of the BOLD signal, which is driven chiefly by fluctuations in blood flow and oxygen metabolism associated with neural activity. In this talk I will discuss our group’s efforts to apply quantitative fMRI techniques to less constrained experimental conditions with the goal of more directly measuring the blood flow and oxygen metabolism changes associated with natural human behaviors. In particular I will discuss the promise of and challenges facing BOLD Constrained Perfusion, a technique that combines BOLD and perfusion-weighted MRI to produce high signal-to-noise estimates of blood flow and oxygen metabolism changes under naturalistic experimental conditions.

Tomas Strömberg
Linköping University

Assessing the microcirculation by diffuse reflectance spectroscopy and conventional laser Doppler flowmetry 

We have developed a new method combining white light diffuse reflectance spectroscopy (DRS; 450-850 nm) and laser Doppler flowmetry (LDF; 780 nm) in a fiber optic probe that is in contact with the tissue. The LDF and DRS data are analyzed and compared to Monte Carlo simulated photon transport data from an adaptive multilayered tissue model including geometrical and optical properties, as well as red blood cell (RBC) flow-speed information. Output data from the model are: The tissue fraction of RBC, the Hb oxygenation/ saturation and speed-resolved perfusion, in absolute units. The method will be presented as well as results from skin measurements during standardized skin provocation protocols such as local heating hyperemia, arterial occlusion and release. The method offers several advantages such as simultaneous quantification of RBC tissue fraction and oxygenation and perfusion from the same, predictable, sampling volume. The method can be extended to evaluate the effect of different RBC speed distribution assumptions and to other tissues than skin. 

Anna Tietze
CFIN, AU / Department of Neuroradiology, AUH

Perfusion MRI derived indices of microvascular shunting and flow control correlate with tumor grade and outcome in patients with cerebral glioma.  

Shunting of oxygenated blood through abnormal neo-vasculature is thought to limit oxygen extraction in tumors, causing them to grow and metastasize faster, and to be more resistant to radio- and chemotherapy. We developed a physiological model that estimates oxygen extraction efficacy in terms of capillary transit time heterogeneity (CTH), and a method that permits its measurement during diagnostic magnetic resonance imaging (MRI) of brain tumor patients. We investigated CTH in 72 glioma patients and show that CTH images increase diagnostic accuracy in the differentiation of high-grade from low-grade gliomas and allow better prediction of patient outcome than cerebral blood volume, the existing diagnostic index of tumor neovascularization. We suggest that CTH may provide additional information when evaluating tumor angiogenesis, tumor oxygenation, or responses to anti-angiogenic therapy.  

Hans Vink
Maastricht University

Physiological and pathophysiological properties of the glycocalyx.
Clinical assessment of glycocalyx: A tool to monitor vascular risk in patients?

The innermost lining of blood vessels is formed by endothelial cells, and numerous studies have demonstrated that many interactions between blood components and the vascular wall are mediated by the endothelial glycocalyx, a complex mesh of proteoglycans, glycosaminoglycans and plasma proteins that is anchored to the plasma membrane of vascular endothelium. The glycocalyx contributes to the permeability barrier and limits interaction of platelets and leukocytes with the endothelial surface and thereby forms a protective compartment that lowers vascular vulnerability for atherogenic challenges.

Recent studies have established that the glycocalyx is perturbed by hypercholesterolemia, cigarette smoke, endotoxin, hyperglycemia and in diabetes. These findings led to the concept that atherogenic damage to the glycocalyx marks an increased pathogenic risk in individuals exposed to cardiovascular risk factors, while prevention of glycocalyx loss or restoration of glycocalyx protective properties might be a possible therapy against acute or chronic vascular complications. Despite the major role of the glycocalyx in vascular wall homeostasis, convenient methods allowing for glycocalyx measurements in a clinical setting are lacking.

We recently developed an approach for standardized assessment of changes in glycocalyx dimensions in clinical microscopic video recordings of the sublingual microvasculature, based on the principle that an intact glycocalyx provides only limited access to erythrocytes, while loss of glycocalyx protective properties is reflected by greater penetration of circulating cells towards the endothelial surface.

Recent clinical findings demonstrate that automated measurement of dynamic changes in erythrocyte column width allows for the identification of early vascular vulnerability in cardiovascular risk patients with e.g. diabetes and renal failure. Moreover, preliminary data indicate that these measurements may also allow for the identification of cardiovascular risk in apparently healthy individuals exposed to low risk profiles, but with increased vascular vulnerability due to genetic predisposition.