Homogeneity of the vasculature of the subthalamic nucleus: A
post-mortem study in Parkinson’s disease patients and
non-demented elderly
Lieke E. W. Vermaat
Master Thesis Clinical Neuropsychology, 20 ECTS
Faculty of Behavioural and Social Sciences, Leiden University July, 2017
Student number: 1889028
Daily supervisor: Dr. Anneke Alkemade, Integrative Model-Based Cognitive Neuroscience (IMCN) Research Unit, University of Amsterdam
CNP-supervisor: Kaya Peerdeman, Department of Health, Medical and Neuropsychology, Leiden University
2
Abstract
Objective: The aim of this study was to investigate the homogeneity of the microvasculature of the subthalamic nucleus (STN) and to compare the STN’s vasculature between Parkinson’s disease (PD) patients and non-demented elderly. Studying the homogeneity of the STN’s vasculature can inform about whether the blood-oxygen level dependent (BOLD) signal used in functional magnetic resonance imaging (fMRI) could originate homogeneously from all putative STN parts.
Methods: Immunohistochemistry with structural vasculature staining on post-mortem material of 6 Parkinson disease patients and 9 non-demented elderly was used to outline the
microvasculature. Intensity measurements were performed for each STN in a central section and separately for the ventromedial and dorsolateral part, with two correlated outcome measures: the mean intensity and percentage stained.
Results: No difference was found between the ventromedial and dorsolateral part of the STN for both outcome measures. Also, the intensity measurements of the central section of the STN did not differ between PD patients and non-demented elderly.
Conclusions: This study did not find a difference in intensity of a structural vasculature staining within the STN. This could imply that all putative STN parts can provide equally strong BOLD signals. Future research should expand upon our findings with larger samples to further validate the use of BOLD fMRI to study neural activity in subcortical structures such as the STN and the processes behind the pathology and side effects of treatment in
3
Introduction
Functional magnetic resonance imaging (fMRI) studies can be used to non-invasively investigate altered brain activity and structure in neurological disorders such as Parkinson’s disease (PD) (Boertien et al., 2011; Kahan et al., 2012; Prodoehl, Spraker, Corcos, Comella, & Vaillancourt, 2010). PD is a neurodegenerative disorder, characterized by motor symptoms such as resting tremor, bradykinesia, and rigidity (Cacabelos, 2017; Hirsch, Jette, Frolkis, Steeves, & Pringsheim, 2016). It is the second most prevalent neurodegenerative disorder after Alzheimer’s disease, with European prevalence rates ranging from 66 to 12,500 per 100,000 and incidence rates from 5 to 346 per 100,000 (Von Campenhausen et al., 2005). The incidence of PD increases with age and is especially a burden on health care for countries with an increasing elderly population (Hirsch et al., 2016).
In PD, a selective loss of dopaminergic neurons in the substantia nigra (SN), which normally inhibit the subthalamic nucleus (STN), results in dysinhibition of the STN’s activity (Figure 1, p. 4). The STN is a small nucleus in the subcortex with sensorimotor, cognitive, and limbic functions, and part of the basal ganglia (Alkemade & Forstmann, 2014;
Brunenberg et al., 2012; Temel, Blokland, Steinbusch, & Visser-Vandewalle, 2005). Thus, in PD the STN is overactivated, and then strongly excites the output nuclei of the basal ganglia. These output nuclei in turn inhibit the thalamus’ excitatory stimulation of the cortex more, leading to less excitation of the primary motor, premotor, and somatosensory cortex by the thalamus (Temel et al., 2005). This lower excitation is associated with PD’s sensorimotor problems (Bronstein et al., 2011; Fernández-Seara et al., 2015; Temel et al., 2005).
4
Figure 1. Adapted from Temel et al., 2005. In PD, the STN is overactive, leading to less excitation of the
primary motor, premotor, and somatosensory cortex (c).
fMRI studies on PD suggest that deep brain stimulation (DBS) of the STN can normalize the altered activation patterns found in PD (Benabid, 2003; Hamani, Saint-Cyr, Fraser, Kaplitt, & Lozano, 2004; Hesselmann et al., 2004; Jech et al., 2001; Kahan et al., 2012; Min et al., 2014; Phillips et al., 2006; Stefurak et al., 2003). However, the exact mechanisms behind the therapeutic effects of DBS are unclear (Kahan et al., 2012). A
possible means by which DBS exerts its effects could be that it causes a depolarization block, by which it limits the occurrence of action potentials and thus neural activation (Beurrier, Bioulac, Audin, & Hammond, 2001). Also, DBS could lower neural activity by stimulating GABAergic afferents of the STN or other stimulated nuclei, which results in a release of GABA. This inhibitory neurotransmitter then hyperpolarizes postsynaptic terminals, thereby lowering the chances of action potentials occurring and consequently deactivating the
stimulated nucleus (Moser, Gieselberg, Ro, Keller, & Qadri, 2003). Remarkably, it was found that high frequency DBS inhibits cell bodies of neurons by activating presynaptic terminals close to the electrode, but that it enhances output activity of a targeted nucleus by stimulating action potentials in the axons further away from the cell body (Grill & McIntrye, 2001; Grill, Snyder, & Miocinovic, 2004; Hashimoto, Elder, Okun, Patrick, & Vitek, 2003; McIntyre et al., 2004; Windels et al., 2003). This mechanism is different from the first two proposed in
5 which DBS has effects similar to a lesion (Temel et al., 2005). Moreover, many side effects are reported for DBS treatment of PD, with cognitive dysfunctions such as decreased verbal memory and executive functions mostly reported, followed by depression, occasionally accompanied by suicidal ideation (Brunenberg et al., 2012; Temel et al., 2005). These side effects might be caused by unintentional stimulation of putative non-sensomotoric parts of the STN (Alkemade, 2013; Plantinga et al., 2016; Temel et al., 2005). In sum, DBS is an effective technique for improving motor symptoms of PD, but its working mechanisms remain to be understood and side effects can be severe (Temel et al., 2005).
Studying the effects of DBS of the STN in PD or activity of subcortical structures such as the STN can be done with fMRI studies (Forstmann, de Hollander, van Maanen,
Alkemade, & Keuken, 2017; Turner, 2016). fMRI measures the blood oxygen-level dependent (BOLD) signal, which increases with neural activity (Logothetis, 2008; Turner, 2016). Neural activity is related to an increase in metabolic demand, which results in more oxygen extraction from the blood (Buxton & Frank, 1997; Gagnon et al., 2015; Hillman, 2014). Dilation of arterioles and capillaries is thought to be caused by the molecules released during neural activity, usually overcompensating for the higher oxygen demands (Hall et al., 2014; Hillman, 2014; Turner, 2016). This overoxygenation of the active region results in a decrease in deoxyhemoglobin (Hillman, 2014). Relating to deoxyhemoglobin’s paramagnetic properties which distort the magnetic field measured in fMRI, the BOLD signal increases with lower deoxyhemoglobin levels, thus with higher neural activity (Hillman, 2014). It was suggested that pericytes might be a crucial factor mediating the BOLD signal (Hall et al., 2014; Hillman, 2014). Pericytes are contractile cells located in capillaries which regulate cerebral blood flow by dilating capillaries through relaxation during neural activity (Hall et al., 2014). Dilation of capillaries by pericytes was estimated to be responsible for as much as 84% of the increase in blood flow following enduring neural activity by one study (Hall et al., 2014). Neurovascular coupling might also be mediated by astrocytes, supporting cells located in between neurons and blood vessels (Hillman, 2014). An increase in glutamate, associated with neural activity, might be sensed by receptors on astrocytes which then initiate
vasodilating processes (Hillman, 2014). However, recent findings have found the involvement of astrocytes in neurovascular coupling less likely than previously thought (Nizar et al., 2013). Furthermore, interneurons have been proposed as an alternative mechanism involved in neurovascular coupling by releasing vasoactive substances (Cauli, 2004; Piché, Uchida, Hara, Aikawa, & Hotta, 2010). Last, vasodilation might be mediated by a part of the vasculature itself: the vascular endothelium, which lines the inner wall of all types of blood
6 vessels (Bagher & Segal, 2011; Figueroa & Duling, 2009; Wölfle et al., 2011).
Hyperpolarisation can propagate along the endothelium, thereby self-initiating vasodilation. It was suggested that probably a combination of the described methods is responsible for the BOLD response (Attwell et al., 2010). In addition, the BOLD signal is criticized for not being spatially specific and an indirect measure of brain activity, the hemodynamic response lacking a few seconds behind actual neuronal activity (Hillman, 2014; Logothetis, 2008; Turner, 2016). In short, the BOLD signal is probably mediated by several microvascular mechanisms involving the capillaries, but the contribution of each and the exact mechanism still needs to be ascertained.
Currently, it is unclear whether the just described BOLD signal used in fMRI studies in PD (e.g., Brunenberg et al., 2012) can provide insights into the internal structure of the STN. Studying the STN’s internal structure is relevant for successful DBS stimulation of the STN in PD: whether there are clear boundaries and subparts within the STN can influence the spreading of the DBS stimulation and therefore its planned and unintentional effects
(Alkemade, 2013; de Hollander et al., 2014). The putative dorsolateral part of the STN, associated with sensorimotor functions, has been reported to be an effective target for DBS, lowering the sensorimotor problems often already seen early in PD (Bronstein et al., 2011; Fernández-Seara et al., 2015; Min et al., 2014; Plantinga et al., 2016). There is no consensus in the literature on the amount of anatomical subparts of the STN and on whether the potential boundaries between subparts are distinct or gradual (Alkemade & Forstmann, 2014; de
Hollander et al., 2014; Keuken et al., 2012). The tripartite hypothesis seems most popular, dividing the STN into a dorsolateral sensorimotor part, a limbic part in the medial tip, and a central cognitive part (Parent & Hazrati, 1997; Temel et al., 2005), but this idea has been substantially opposed by others due to unconvincing evidence in favour of this hypothesis (Alkemade & Forstmann, 2014; de Hollander et al., 2014; Keuken et al., 2012).
To determine whether the BOLD signal can be informative about the internal structure of the STN, it needs to be ascertained whether the BOLD signal could homogeneously
originate from the STN, in other words, whether all putative parts of the STN have a similar amount of local blood flow and can, in theory, provide equally strong BOLD signals. This is especially relevant with the increasing availability of ultra-high field (UHF) 7T MRI, enabling visualization of subcortical structures such as the STN which is difficult at lower field
strengths (Forstmann et al., 2017; Turner, 2016). Little research has been conducted on the vasculature of the STN, but one previous study found increased blood flow within the STN in the direction of the SN (Wårdell, Hemm-Ode, Rejmstad, & Zsigmond, 2016). In addition,
7 higher density of markers in the ventromedial tip of the STN was reported (Alkemade,
Balesar, Zhao, Swaab, & Forstmann, n.d.; Foix & Nicolesco, 1925; Fussenich, 1967; Kodama, 1928), which might translate into higher cell density in the ventromedial tip and heterogeneous vessel density across the STN.
To our knowledge, a study in which the main aim is to investigate the homogeneity of the vasculature of the STN has not been conducted before. A heterogeneous vasculature would indicate that all putative STN parts might not be able to provide equally strong BOLD signals. Then, fMRI BOLD studies might not be a valid method for studying neural activity in the STN, relevant for studies on DBS in PD or any other studies targeting potential
subdivisions of the STN. In addition, comparing the STN’s vasculature between PD patients and non-demented elderly could associate structural vascular pathology within the STN with PD. Again, preceding research is limited, and to our knowledge, this study is the first using structural markers visualizing the STN’s vasculature. Nevertheless, one study found the number of capillaries to be smaller in PD patients than in non-demented elderly in the SN (Guan et al., 2013), a subcortical structure related to and proximal to the STN (Mai, Majtanik, M., & Paxinos, 2015).
This study aimed to investigate the vasculature of the STN up to the small-scale level of capillaries, from which the BOLD signal might originate, by using immunohistochemistry on human post-mortem material. First, the homogeneity of the STN’s vasculature was investigated, to determine whether the BOLD signal could originate homogeneously from all putative STN parts. Based on the few previous studies in which the internal structure of the STN was investigated, we expected the distribution of the vasculature marker within the STN to be heterogeneous, with an increased intensity in the direction of the ventromedial tip (Alkemade et al., n.d.; Fussenich, 1967; Kodama, 1928; Wårdell et al., 2016). Second, we aimed to compare the STN’s vasculature between PD patients and non-demented controls, possibly providing insight into the anatomy of the pathology of PD as suggested in a previous study (Plantinga et al., 2016). Based on the single study that found a smaller number of capillaries in PD patients than in non-demented elderly (Guan et al., 2013), the
immunoreactivity to the vasculature marker in the STN was expected to be lower in PD patients than in non-demented elderly.
8
Methods Design & Participants
For this study, 15 post-mortem brains were studied, of which 6 from PD patients (3 female, 3 male) and 9 from non-demented controls (4 female, 5 male). The post-mortem tissue blocks were obtained from the Netherlands Brain Bank, Amsterdam. PD patients were included if PD was diagnosed clinically in vivo according to standard clinical guidelines. The non-demented controls registered voluntarily at the Netherlands Brain Bank to donate their brains to research after death. Neuropathological examination post-mortem confirmed their non-demented status. Medical history was collected by reviewing patients’ clinical records. All material was obtained in accordance with ethical and legal guidelines of the Netherlands Brain Bank, ensuring informed consent and medical confidentiality.
Procedure
The brains were dissected at autopsy and fixed in 10% phosphate-buffered formalin at room temperature. First, multimodal structural MRI scans using a 7T Magnetom MRI system were performed with the tissue blocks (see Figure 3 in Forstmann et al., 2017 and Weiss et al., 2014 for the post-mortem data acquisition pipeline). Then, the tissue blocks were dehydrated and embedded in paraffin for conservation of the post-mortem material (see Figure 2 for an example of a tissue block). The tissue blocks were serially cut, rostral to caudal, on a microtome at 6 μm prior to block face imaging and 2D staining. This resulted in on average about 2000 sections per STN. Nissl staining was carried out on every 50th section, for
anatomical orientation. Block face imaging was performed on the Nissl stained sections after glass cover slipping the sections.
For the immunohistochemistry used in this study, 25 to 30 sections per person were obtained of the STN, spaced 300 μm apart. 2D immunocytochemical staining was performed with CD34 on consecutive sections. CD34 is a glycoprotein, found on hematopoietic
progenitor cells and the endothelium of small blood vessels, and its function remains unclear (Satterthwaite, Burn, Le Beau, & Tenen, 1992; Simmons, Satterthwaite, Tenen, & Seed, 1992). Since CD34 is present in the endothelium, it can be visualized using antibodies post-mortem, demonstrating the vasculature. The primary antibody was mouse monoclonal anti-CD34 (cat. no.sc-52312, lot. no. J2512, Santa Cruz, 1:50), the secondary antibody
biotinylated anti-mouse IgG (1:400). Staining was amplified using the ABC vector elite kit. After completion of staining, sections were cover slipped, and digital images were obtained from the 2D stained sections. Then, these images of the stained tissue were brought together
9 in a common space with the block face images, allowing us to locate the STN in our images. Linear transformations were performed to register the images of the stained tissue to the corresponding block face images.
Figure 2. Example of a tissue block. The drawing indicates that the left side of the block is the rostral side, the
bottom ventral.
Measures
Fiji, a distribution of ImageJ2 developed for analysing biological images, an open source Java image processing program, was used to perform densitometry on the stained image data (Abràmoff, Magalhães, & Ram, 2004; Schindelin et al., 2012). After thresholding the images, intensity measurements were done for each section. Intensity measurements of the ventromedial and dorsolateral part of the STN per participant for approximately each middle section along the rostro-caudal axis were quantified in ImageJ2. The intensity values for each complete STN in the same middle section per subject enabled a comparison of the STN’s vasculature between the non-demented elderly and PD patients.
Statistical analyses
The data were analyzed with Rstudio, version 1.0.143. A Student’s t-test compared age between the non-demented elderly and PD patients. Normality was tested with Shapiro-Wilk tests (ɑ = 0.05) for both hypotheses. Some resulting p-values were larger than .1, which makes the approximation of the p-values unreliable: our sample is probably too small for the Shapiro-Wilk test to pick up deviations from normality (Royston, 1995). Therefore, and due to the low number of observations in our post-mortem sample, we decided to use non-parametric tests, since those are more conservative than non-parametric tests.
10 marker, was measured with ImageJ with two variables: the mean intensity, which is the sum of the gray values of all the pixels within the STN divided by the total number of pixels in the STN and therefore a relative measure, and the percentage stained, which indicates the
percentage of pixels in the STN that have been highlighted (Ferreira & Rasband, 2012). These two measures are thus highly related, and both were used to verify the outcomes.
For the first hypothesis, the distribution of the vasculature marker within the STN was tested by the Wilcoxon-Signed Rank test, first for all participants, and then for PD patients and non-demented elderly separately. Including all participants, the Pearson’s correlation of the mean intensity and percentage stained ventromedially and dorsolaterally were both calculated. The intensity of the vasculature marker was compared between the ventromedial and dorsolateral part of the STN. The Wilcoxon-Signed Rank test takes into account that the data used to test this hypothesis are matched: observations of the intensity of the vasculature marker in the ventromedial and dorsolateral part of the STN are not independent, since they were measured within the same STN per participant. The location within the STN,
ventromedial or dorsolateral, was the independent variable, the intensity of the vasculature marker the dependent variable. This hypothesis was tested twice with both outcome measures. Thus, the null hypothesis was that there is no relation between the location within the STN and the intensity of vessels found. We expected this null hypothesis to be rejected.
For the second hypothesis, group differences between post-mortem material from PD and non-demented elderly were tested with Mann-Whitney U tests. A Pearson’s correlation between the mean intensity and percentage stained was calculated including all participants. The independent variable was group (PD or non-demented elderly), the dependent variable the vessel intensity in the STN, again tested twice with the mean intensity and percentage stained as outcome variables. Thus, the null hypothesis was that there is no relation between group (PD or non-demented elderly) and vessel intensity, and we expected this hypothesis to be rejected.
Results Participant characteristics
The average age at death was 86.40 years (SD = 10.31) for the total sample, 83.67 years (SD = 5.65) for the PD patients and 88.22 years (SD = 12.52) for the non-demented elderly. Age did not differ between PD patients and non-demented elderly (p = .37, W = 19) Further demographics of the participants can be found in Table 1.
11 Table 1
Clinicopathological data per participant
NBB Age
(y)
Sex Brain type PMD
(h:m)
Fix (days)
Medical history
12062 88 M Control 05:40 nd ischemic bowel rupture (COD), aortic
stenosis, femoropopliteal bypass, hypercholesterolemia, cardiorenal syndrome, ischaemic cardiomyopathy, atrical fibrillation, ischeamic bowel rupture, braak 4, tauopathy
12082 101 F Control 05:10 nd cachexia (COD), cataract, TIA, mitral
valve insufficiency, osteoporosis, coxarthrosis, cachexia, kyfose, decubitus dehydration
13054 91 F PD 04:40 56 pneumonia by advanced LBD (COD),
6yPD, LBD, Braak 6 aSYN, pneumonia, obstipation, tremor and saccadic eye movements, urinary tract infection, exelon 9,5mg 1dd1, madopar 125mg 2dd1, dopa aggrevated the symptoms, hip replacement
13058 77 F PD 07:20 53 sepsis by urinary tract infection (COD),
PD, 15y, sepsis, hypertension, pulmonary embolism, 4dd levo/carbi/entacapon 100/25/100
13060 84 M PD 04:50 48 pneumonia (COD), severe decubital
ulcer, LBD. PD/LBD, 7y pneumonia, hypertension, COPD, type 2 diabetes, cardiac failure, aneurysm, MI, cholecystectomy,
hypercholesterolemia, dementia hypothyroidism, rivastigmine 3dd3m, diabetic gastroparesis, collum fracture
13072 87 F PD 08:10 57 uremia by dehydration (COD), 35y PD,
M Raynaud, left onset PD, biperidene 1mg 1dd1, levodopa/benserazide 125 mg 4dd1, braak 3, note that PD and Raynaud medications are often contra-indicated, making titration difficult
13077 86 M PD 04:10 57 respiratory insufficiency (COD), PD,
18y, respiratory insufficiency, levo/carbidopa 125mg 7dd1,
rivastigmine 13,3mg/h, cholesteatoma, presbyacusis, thalamic DBS
13061 77 M PD 03:10 56 aspiration pneumonia (COD), 31y PD,
‘95 left pallidotomy, ‘97 right pallidotomy, Levodopa/carbidopa 125mg 5dd1, clozapine 100 mg 3dd1, apomorphine 25mg 1dd1, rivastigmine
12 9,5mg dd, tremor stiffness
bradykinesia, left onset,
hypersexuality, oral behavior, 2 sybs also affected, braak 6,, glaucoma, visual hallucinations
13095 101 F Control 06:15 57 pneumonia (COD), cateract, uterus
extirpation, cholecystectomy, DM2, coxarthrosis, spondylosis,
conjunctivitis, angina pectoric, COPD, cataract, cardiac falure, bullous pemphigoid, aspiration pneumonia, braak 2
14037 101 F Control 07:27 57 renal insufficiency (COD), cataract,
atrial fibrillation, hip replacement, scoliosis, dicopathy, gasteroenteritis, urinary tract infection, renal insufficiency
14051 92 M Control 07:45 57 cardiac failure (COD), hip
replacements, DMII, COPD, cataract liver cirrhosis, ascites, decubitus, polyneuropathy, heart failure
14069 73 M Control 04:25 56 pneumonia (COD), COPD,
hypercholestoerolemia, atrial fibrillation, anrtofemoral bypass, PTCA, cataract, spondylodese, oesophagitis, pneumonia, prostate carcinoma, hyperthyroidism, decubitus
15033 93 M Control 07:40 59 cardiac failure (COD), aortic stenosis,
decompensatio cordis, macular degeneration, basal cell carcinoma
15035 73 M Control 08:00 56 pneumonia (COD), cardiac failure,
myelodysplastic syndrome, fungal infection
15055 72 F Control 06:50 55 respiratory insufficiency (COD), polio,
ovarium carcinoma, ileus, osteoporosis, polymyalgia
Note. COD = cause of death, F = female, Fix = fixation duration, LBD = Lewy Body Dementia, M = male, NBB
= Netherlands Brain Bank number, nd = no date, PD = Parkinson’s disease, PMD = post-mortem delay, TIA = transcient ischemic attack, y = years.
Distribution of the vasculature marker within the STN
Including all participants, the mean intensity and percentage stained ventromedially correlated highly (r = .97, p < .001), as did the mean intensity and percentage stained dorsolaterally (r = .97 , p < .001), affirming the use of both as outcome variables to double-check our results. The average mean intensity of the staining was ventromedially 4.70 (SD = 1.94), and dorsolaterally 5.10 (SD = 2.11) (see Figure 3). The percentage stained was
13 ventromedially on average 32.19% (SD = 11.71), and dorsolaterally on average 33.52% (SD = 11.90) (see Figure 4). The Wilcoxon-Signed Rank test showed no difference between the ventromedial and dorsolateral part of the STN for both the outcome variables mean intensity (p = .15, V = 34, 95% CI 1.17, 0.15]) and percentage stained (p = .23, V = 38, 95% CI [-5.67, 1.57).
When including only PD patients in the analysis, the mean intensity was
ventromedially on average 3.74 (SD = 2.18) and dorsolaterally on average 3.93 (SD = 2.22). The percentage stained was ventromedially on average 26.68% (SD = 13.18) and
dorsolaterally on average 26.79% (SD = 11.18). No difference was found both for the mean intensity (p = 1.00, V = 10, 95% CI [-2.41, 3.14]) and the percentage stained (p = 1.00, V = 10, 95% CI [-12.48, 19.45]). Including only the non-demented elderly resulted in an average mean intensity ventromedially of 5.35 (SD = 1.56) and dorsolaterally of 5.87 (SD = 1.74). The percentage stained was ventromedially on average 35.86% (SD = 9.66) and dorsolaterally on average 38.01% (SD = 10.62). Again, no difference was found for the mean intensity (p = .07, V = 7, 95% CI [-1.17, .06]) and the percentage stained (p = .13, V = 9, 95% CI [-5.73, 1.09]).
Figure 3. Mean CD34 staining intensity ventromedially and dorsolaterally within the STN of Parkinson’s disease
14
Figure 4. Percentage staining by the CD34 staining of the ventromedial and dorsolateral part of the STN in
Parkinson’s disease patients and demented elderly. PD = elderly with Parkinson’s disease, CTRL = non-demented elderly.
Vasculature marker in PD patients compared to non-demented elderly
The mean intensity and percentage stained of the middle section of the STN across all participants correlated highly (r = .97, p < .001), thus using both for double-checking the outcomes appears valid. The mean intensity of the staining was 5.48 (SD = 1.57) in the non-demented elderly and 3.79 (SD = 2.01) in the PD patients (see Figure 5). The percentage stained was 36.39% (SD = 9.78) in the non-demented elderly and 26.58% (SD = 10.84) in the PD patients (see Figure 6). The Mann-Whitney U test showed no difference between the mean intensity of the staining of PD patients and non-demented elderly (p = .07, W = 43, CI = -.082 to 3.78). For the percentage stained, also no difference was found between the two groups (p = .11, W = 41, CI = -3.75 to 22.15).
15
Figure 5. Boxplot of the mean intensity of the CD34 staining in the STN of non-demented elderly and elderly
with Parkinson’s disease. CTRL = non-demented elderly, PD = elderly with Parkinson’s disease.
Figure 6. Boxplot of the percentage stained of the STN with the CD34 staining of non-demented elderly and
16
Discussion
This study aimed to investigate the vasculature of the STN with immunohistochemistry in human post-mortem material of PD patients and non-demented elderly. Our results showed no heterogeneous distribution of the structural vasculature marker within the STN: the
ventromedial and dorsolateral part of the STN did not show different staining intensities. Further, no difference was found when comparing the staining intensity of a central section of each STN between PD patients and non-demented elderly.
The finding that the staining intensity did not differ between the ventromedial and dorsolateral part of the STN is contrary to our first hypothesis, and not in line with the few studies that pointed out a higher density of markers in the ventromedial part of the STN (Alkemade, Balesar, Zhao, Swaab, & Forstmann, n.d.; Foix & Nicolesco, 1925; Fussenich, 1967; Kodama, 1928). However, the markers used in these previous studies were no specific structural vasculature markers, but protein markers that for example stain neuronal cell bodies (e.g. Alkemade et al., n.d.). Thus, the different findings might be explained by the different markers used. Our results are also different from one study that showed increased blood flow in the ventromedial part (Wårdell, Zsigmond, Richter, & Hemm, 2013). It has to be noted that Wårdell et al. (2013) used 1.5T and CT scans to locate the STN. The spatial resolution of 1.5T is too low a field strength to precisely delineate subcortical structures such as the STN
(Forstmann et al., 2017; Turner, 2016). Therefore, we cannot be certain whether their reported increased blood flow in the direction of the SN is truly located in the ventromedial tip of the STN. Further, the microvascular blood flow measured by Wårdell et al. (2013) is a different method from our structural vasculature staining. Both are relevant for and might be related to the BOLD signal (Bagher & Segal, 2011; Hall et al., 2014), but measure different
mechanisms that are probably not directly related. For example, pericytes might regulate local blood flow by dilating capillaries, in which the pericytes directly control the amount of blood flow, and the amount of capillaries only indirectly by providing the structure through which the blood flow can occur (Hall et al., 2014). For these reasons, our findings might not have been in line with those of Wårdell et al. (2013). Still, our results could indicate that the ventromedial and dorsolateral part of the STN might be able to provide equally strong BOLD signals. This is relevant for fMRI studies on PD, since our results do not invalidate fMRI BOLD as a method to study neural activity in the STN. Then, fMRI might be a valid method to study activity of possible subdivisions of the STN. However, studies with larger samples are recommended to establish this, and our results should be viewed as preliminary.
17 Second, since the intensity of the vasculature staining in the STN did not differ
between PD patients and non-demented elderly, our second hypothesis could not be
confirmed. However, this hypothesis was based on only one previous study in the SN, and the influence of PD on the SN might be different from the STN, considering their different roles in the neuropathology of PD (Guan et al., 2013; Temel et al., 2005). Our results do not enable us to associate differences in structural vascular pathology of the STN between PD patients and non-demented elderly. It was already suggested that it is still unknown whether and which changes to the vasculature are a characteristic of PD (Guan et al., 2013). It might also be that our staining method cannot show the damage to the vasculature caused by PD. For instance, PD might shorten the length of capillaries or decrease capillary connections (Guan et al., 2013), or could damage the microvasculature in another way that our structural marker might not have been able to capture. Further, it might be that the structural vasculature is not one of the main biomarkers affected by PD. Then, studying the structural vasculature of the STN might not be favourable for understanding the pathological mechanisms of PD. In sum, we cannot conclude that there is lower immunoreactivity to the vasculature marker in PD patients than in non-demented elderly.
Since our results do not show a difference in immunoreactivity to the vasculature marker between the dorsolateral and ventromedial part of the STN, this suggests the vasculature of the STN to be homogeneous. Then, the vasculature might not show clear boundaries within the STN according to our results, which could imply that the STN has no clear subdivisions, in line with the idea that the tripartite division of the STN should be revised (Alkemade & Forstmann, 2014; de Hollander et al., 2014; Keuken et al., 2012). Thus, since our results demonstrate no rigid boundaries, the STN might rather have a gradual topographical organisation than a distinct organisation, making prevention of side effects of STN DBS by accidentally stimulating putative untargeted subparts of the STN in PD difficult (Temel et al., 2005). Additionally, due to the increasing availability of UHF fMRI and thus improved visualization of the subcortex (Forstmann et al., 2017; Turner, 2016), it is of importance to know whether the measured BOLD signal originating from, for example, the STN can actually differentiate between subparts of the STN, this not being due to differences in vasculature between subparts of the STN. Since our results indicate the subparts to in theory be able to provide similar BOLD signals, this study supports the use of UHF fMRI to study neural activity in subparts of small subcortical structures.
Some limitations of this study should be mentioned. First of all, the small sample size limits the reach of our conclusions. Unfortunately, in post-mortem research, large sample
18 sizes are difficult to obtain and this limitation is rather inherent to this type of study.
However, a solution for future research might be combining UHF MRI research and post-mortem research by registering acquired images through these methods to the same common space, yielding the MRI and post-mortem images comparable (Forstmann et al., 2017). Consequently, since MRI images are relatively easier to collect than post-mortem images since they can be obtained in living persons, this would make the low sample sizes of post-mortem research less problematic, and furthermore, the benefits of both methods could then be combined by their integration (Forstmann et al., 2017). Second, some of the participants showed comorbid Lewy Body Dementia which was confirmed by post-mortem pathological examination. Clinically, PD and variants of Lewy Body Dementia are difficult to differentiate (Gomperts, 2016). Due to the limited availability of post-mortem material and similar clinical presentations of PD and variants of Lewy Body Dementia in vivo we did not regard patients with Lewy Body Dementia as a separate group, which might have influenced our results. For instance, the STN hyperactivity typical of PD could be lower in patients with comorbid Lewy Body Dementia than in patients with only PD. Furthermore, all participants in this study were elderly with high medical comorbidity for which cannot be controlled. Therefore, it is difficult to ascertain whether our findings comparing the PD patients and non-demented elderly are due to their neurological state or whether other medical conditions were of significant influence. Additionally, the PD patients in our sample were slightly younger than the non-demented elderly, which might have influenced our findings in comparing the two groups. Last, since the microvascular mechanisms behind the BOLD response are not understood yet (Hall et al., 2014; Hillman, 2014), we cannot be certain that the CD34 structural vasculature marker used in this study, present in the endothelium of small blood vessels, is directly related to the vasculature mechanisms behind the BOLD response. It was suggested that the vascular endothelium mediates vasodilation and plays a role in the BOLD response (Bagher & Segal, 2011; Figueroa & Duling, 2009; Wölfle et al., 2011), but several other coupling mechanisms have also been suggested (Cauli, 2004; Hall et al., 2014; Hillman, 2014; Nizar et al., 2013; Piché et al., 2010). Thus, even though our results indicate that the ventromedial and
dorsolateral part of the STN could provide equally strong BOLD signals, other studies using different vasculature markers and methods should confirm these results. In short, our results should be seen as a first indication that the structural vasculature of the STN might be homogeneous.
This study brought up some suggestions for further research. First, the vasculature of the STN could be measured in more than the two subparts we investigated. In our images, we
19 only measured the intensity of the staining in the ventromedial and dorsolateral part, and not in for example, the central part of the STN. Measuring the intensity of the staining in the ventromedial and dorsolateral part is namely more evident than measuring another part of the STN, because the ventromedial and dorsolateral part are at the STN’s extremes. Since there is no consensus on the amount of anatomical subparts of the STN (Alkemade & Forstmann, 2014; Keuken et al., 2012), studying several subparts could support our finding that the vasculature appears homogeneous across the STN, indicating no rigid boundaries in the vasculature of the STN. Secondly, since we only measured the immunoreactivity in the middle section of the STN for each participant we recommend future studies to measure across the entire rostro-caudal axis. Studying the STN across its rostro-caudal axis would allow for measuring the length and change of shape of blood vessels throughout the STN. In addition, different structural vasculature markers could be used to support the current
findings.
Conclusions
In conclusion, this study did not find a difference in intensity of a structural vasculature staining when comparing the dorsolateral and ventromedial part of the STN of PD patients and non-demented elderly. This might indicate that all putative STN parts can provide equally strong BOLD signals. Also, this could imply that at least in the structural vasculature the STN has no clear anatomical subdivisions, possibly making prevention of side effects to STN DBS challenging. Also, our results do not invalidate fMRI BOLD as a method to study neural activity in the STN. Secondly, no difference in vasculature staining was found between PD patients and non-demented elderly. We suggest future studies to expand upon our findings to further validate the use of BOLD fMRI to study neural activity in subcortical structures such as the STN and the processes behind the pathology and side effects of treatment in
20
References
Abràmoff, M. D., Magalhães, P. J., & Ram, S. J. (2004). Image processing with imageJ. Biophotonics International, 11(7), 36–41. doi:10.1117/1.3589100
Alkemade, A. (2013). Subdivisions and anatomical boundaries of the subthalamic nucleus. Journal of Neuroscience, 33(22), 9233–9234. doi:10.1523/JNEUROSCI.1266-13.2013
Alkemade, A., Balesar, R., Zhao, J., Swaab, D. F., & Forstmann, B. U. (2017).
Immunocytochemical characterization of the human subthalamic nucleus. Manuscript submitted for publication.
Alkemade, A., & Forstmann, B. U. (2014). Do we need to revise the tripartite subdivision hypothesis of the human subthalamic nucleus (STN)? NeuroImage, 95, 326–329. doi: 10.1016/j.neuroimage.2014.03.010.
Attwell, D., Buchan, A., Charpak, S., Lauritzen, M., MacVicar, B., & Newman, E. (2010). Glial and neuronal control of brain blood flow. Nature, 468, 232–243.
doi:10.1038/nature09613
Bagher, P., & Segal, S. S. (2011). Regulation of blood flow in the microcirculation: role of conducted vasodilation. Acta Physiologica, 202(3), 271–284. doi:10.1111/j.1748-1716.2010.02244.x
Benabid, A. L. (2003). Deep brain stimulation for Parkinson’s disease. Current Opinion in Neurobiology, 13(6), 696–706. doi:10.1016/j.conb.2003.11.001
Beurrier, C., Bioulac, B., Audin, J., & Hammond, C. (2001). High-frequency stimulation produces a transient blockade of voltage-gated currents in subthalamic neurons. Journal of Neurophysiology, 85(4), 1351–1356.
Boertien, T., Zrinzo, L., Kahan, J., Jahanshahi, M., Hariz, M., Mancini, L., … Foltynie, T. (2011). Functional imaging of subthalamic nucleus deep brain stimulation in Parkinson’s disease. Movement Disorders, 26(10), 1835–1843. doi:10.1002/mds.23788
Bronstein, J. M., Tagliati, M., Alterman, R. L., Lozano, A. M., Volkmann, J., Stefani, A., … DeLong, M. R. (2011). Deep brain stimulation for Parkinson disease. Archives of Neurology, 68(2), 165. doi:10.1001/archneurol.2010.260
21 Platel, B. (2012). Structural and resting state functional connectivity of the subthalamic nucleus: Identification of motor STN parts and the hyperdirect pathway. PLoS ONE, 7(6). doi:10.1371/journal.pone.0039061
Buxton, R. B., & Frank, L. R. (1997). A model for the coupling between cerebral blood flow and oxygen metabolism during neural stimulation. Journal of Cerebral Blood Flow and Metabolism, 17, 64–72. doi:10.1097/00004647-199701000-00009
Cacabelos, R. (2017). Parkinson’s disease: From pathogenesis to pharmacogenomics. International Journal of Molecular Sciences, 18(3), 551. doi:10.3390/ijms18030551
Cauli, B. (2004). Cortical GABA interneurons in neurovascular coupling: Relays for subcortical vasoactive pathways. Journal of Neuroscience, 24(41), 8940–8949. doi:10.1523/JNEUROSCI.3065-04.2004
de Hollander, G., Keuken, M. C., Bazin, P. L., Weiss, M., Neumann, J., Reimann, K., … Schäfer, A. (2014). A gradual increase of iron toward the medial-inferior tip of the subthalamic nucleus. Human Brain Mapping, 35(9), 4440–4449. doi:10.1002/hbm.22485 Fernández-Seara, M. A., Mengual, E., Vidorreta, M., Castellanos, G., Irigoyen, J., Erro, E., &
Pastor, M. A. (2015). Resting state functional connectivity of the subthalamic nucleus in Parkinson’s disease assessed using arterial spin-labeled perfusion fMRI. Human Brain Mapping, 36(5), 1937–1950. doi:10.1002/hbm.22747
Ferreira, T., & Rasband, W. (2012). ImageJ User Guide. Retrieved from https://imagej.nih.gov/ij/docs/guide/user-guide.pdf
Figueroa, X. F., & Duling, B. R. (2009). Gap junctions in the control of vascular function. Antioxidants & Redox Signaling, 11(2), 251–266. doi:10.1089/ars.2008.2117
Foix, C., & Nicolesco, J. (1925). Les noyaux gris centraux et la région mesencéphalo-sous-optique. Paris: Masson.
Forstmann, B. U., de Hollander, G., van Maanen, L., Alkemade, A., & Keuken, M. C. (2017). Towards a mechanistic understanding of the human subcortex. Nature Reviews
Neuroscience, 18(1), 57–65. doi:10.1038/nrn.2016.163
Fussenich, M. (1967). Vergleichend anatomische studien uber den nucleus subthalamicus (corpus Luys) bei primaten. Ph.D. Thesis, University of Freiburg.
22 Gagnon, L., Sakad i, S., Lesage, F., Musacchia, J. J., Lefebvre, J., Fang, Q., … Boas, D. a.
(2015). Quantifying the microvascular origin of BOLD-fMRI from first principles with two-photon microscopy and an oxygen-sensitive nanoprobe. Journal of Neuroscience, 35(8), 3663–3675. doi:10.1523/JNEUROSCI.3555-14.2015
Gomperts, S. N. (2016). Lewy body dementias: Dementia with Lewy bodies and Parkinson disease dementia. Continuum: Lifelong Learning in Neurology, 22(2), 435–463. doi:10.1212/CON.0000000000000309
Grill, W. M., & McIntrye, C. C. (2001). Extracellular excitation of central neurons:
implications for the mechanisms of deep brain stimulation. Thalamus & Related Systems, 1, 269–277. doi:10.1002/mds.10145
Grill, W. M., Snyder, A. N., & Miocinovic, S. (2004). Deep brain stimulation creates an informational lesion of the stimulated nucleus. Neuroreport, 15(7), 1137–1140. doi:10.1097/00001756-200405190-00011
Guan, J., Pavlovic, D., Dalkie, N., Waldvogel, H. J., O’Carroll, S. J., Green, C. R., & Nicholson, L. F. B. (2013). Vascular degeneration in Parkinson’s disease. Brain Pathology, 23(2), 154–164. doi:10.1111/j.1750-3639.2012.00628.x
Hall, C. N., Reynell, C., Gesslein, B., Hamilton, N. B., Mishra, A., Sutherland, B. A., … Attwell, D. (2014). Capillary pericytes regulate cerebral blood flow in health and disease. Nature, 508(7494), 55–60. doi:10.1038/nature13165
Hamani, C., Saint-Cyr, J. A., Fraser, J., Kaplitt, M., & Lozano, A. M. (2004). The subthalamic nucleus in the context of movement disorders. Brain, 127(1), 4–20. doi:10.1093/brain/awh029
Hashimoto, T., Elder, C. M., Okun, M. S., Patrick, S. K., & Vitek, J. L. (2003). Stimulation of the subthalamic nucleus changes the firing pattern of pallidal neurons. Journal of
Neuroscience, 23(5), 1916–1923.
Hesselmann, V., Sorger, B., Girnus, R., Lasek, K., Maarouf, M., Wedekind, C., … Sturm, V. (2004). Intraoperative functional MRI as a new approach to monitor deep brain
stimulation in Parkinson’s disease. European Radiology, 14(4), 686–690. doi:10.1007/s00330-003-2051-0
23 Hillman, E. M. C. C. (2014). Coupling mechanism and significance of the BOLD Signal: A
status report. Annual Review of Neuroscience, 37(1), 161–181. doi:10.1146/annurev-neuro-071013-014111
Hirsch, L., Jette, N., Frolkis, A., Steeves, T., & Pringsheim, T. (2016). The incidence of Parkinson’s disease: A systematic review and meta-analysis. Neuroepidemiology, 46(4), 292–300. doi:10.1159/000445751
Jech, R., Urgosík, D., Tintera, J., Nebuzelský, A., Krásenský, J., Liscák, R., … Růzicka, E. (2001). Functional magnetic resonance imaging during deep brain stimulation: a pilot study in four patients with Parkinson’s disease. Movement Disorders, 16(6), 1126–1132. doi:10.1002/mds.1217
Kahan, J., Mancini, L., Urner, M., Friston, K., Hariz, M., Holl, E., … Foltynie, T. (2012). Therapeutic subthalamic nucleus deep brain stimulation reverses cortico-thalamic coupling during voluntary movements in Parkinson’s disease. PLoS ONE, 7(12). doi:10.1371/journal.pone.0050270
Keuken, M. C., Uylings, H. B. M., Geyer, S., Schäfer, A., Turner, R., & Forstmann, B. U. (2012). Are there three subdivisions in the primate subthalamic nucleus? Frontiers in Neuroanatomy, 6, 14. doi:10.3389/fnana.2012.00014
Kodama, S. (1928). Beitrage zur normalen anatomie des corpus luysi beim menschen. Arbeit. Anat. Inst. Sendai, 13, 221–254.
Logothetis, N. K. (2008). What we can do and what we cannot do with fMRI. Nature Reviews Neuroscience, 453(June), 869–878. doi:10.1038/nature06976
Mai, J. K., Majtanik, M., &, & Paxinos, G. (2015). Atlas of the human brain. Academic Press. McIntyre, C. C., Grill, W. M., Sherman, D. L., Thakor, N. V, Mcintyre, C. C., Grill, W. M.,
… Thakor, N. V. (2004). Cellular effects of deep brain stimulation : Model-based analysis of activation and inhibition cellular effects of deep brain stimulation : Model-based analysis of activation and inhibition. Journal of Neurophysiology, 91, 1457–1469. doi:10.1152/jn.00989.2003
Min, H. K., Ross, E. K., Lee, K. H., Dennis, K., Han, S. R., Jeong, J. H., … Bennet, K. E. (2014). Subthalamic nucleus deep brain stimulation induces motor network BOLD
24 activation: Use of a high precision MRI guided stereotactic system for nonhuman
primates. Brain Stimulation, 7(4), 603–607. doi:10.1016/j.brs.2014.04.007
Moser, A., Gieselberg, A., Ro, B., Keller, C., & Qadri, F. (2003). Deep brain stimulation: Response to neuronal high frequency stimulation is mediated through GABAA receptor activation in rats. Neuroscience Letters, 341(1), 57–60.
doi:10.1016/S0304-3940(03)00133-2
Nizar, K., Uhlirova, H., Tian, P., Saisan, P. A., Cheng, Q., Reznichenko, L., … Devor, A. (2013). In vivo stimulus-induced vasodilation occurs without IP3 receptor activation and may precede astrocytic calcium increase. Journal of Neuroscience, 33(19), 8411–8422. doi:10.1523/JNEUROSCI.3285-12.2013
Parent, A., & Hazrati, L. N. (1997). Functional anatomy of the basal ganglia. Revista de Neurologia, 25, 121–128. doi:10.1002/mds.10138
Phillips, M. D., Baker, K. B., Lowe, M. J., Tkach, J. A., Cooper, S. E., Kopell, B. H., & Rezai, A. R. (2006). Parkinson disease: Pattern of functional MR imaging activation during deep brain stimulation of subthalamic nucleus - initial experience. Radiology, 239(1), 209–216. doi:10.1148/radiol.2391041990
Piché, M., Uchida, S., Hara, S., Aikawa, Y., & Hotta, H. (2010). Modulation of
somatosensory-evoked cortical blood flow changes by GABAergic inhibition of the nucleus basalis of Meynert in urethane-anaesthetized rats. The Journal of Physiology, 588, 2163–71. doi:10.1113/jphysiol.2010.187633
Plantinga, B. R., Temel, Y., Duchin, Y., Uludag, K., Patriat, R., Roebroeck, A., … Harel, N. (2016). Individualized parcellation of the subthalamic nucleus in patients with
Parkinson’s disease with 7T MRI. NeuroImage. doi:10.1016/j.neuroimage.2016.09.023 Prodoehl, J., Spraker, M., Corcos, D., Comella, C., & Vaillancourt, D. (2010). Blood
oxygenation level-dependent activation in basal ganglia nuclei relates to specific symptoms in de novo Parkinson’s disease. Movement Disorders, 25(13), 2035–2043. doi:10.1002/mds.23360
Royston, P. (1995). Remark AS R94: A remark on Algorithm AS 181: The W test for normality. Applied Statistics, 44, 547–551.
25 Satterthwaite, A. B., Burn, T. C., Le Beau, M. M., & Tenen, D. G. (1992). Structure of the
gene encoding CD34, a human hematopoietic stem cell antigen. Genomics, 12(4), 788– 794.
Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., … Cardona, A. (2012). Fiji: an open-source platform for biological-image analysis. Nature Methods, 9(7), 676–682. doi:10.1038/nmeth.2019
Simmons, D. L., Satterthwaite, B., Tenen, D. G., & Seed, B. (1992). Molecular cloning of a cDNA encoding CD34 , a sialomucin of human hematopoietic stem cells. The Journal of Immunology, 148, 267–271.
Stefurak, T., Mikulis, D., Mayberg, H., Lang, A. E., Hevenor, S., Pahapill, P., … Lozano, A. (2003). Deep brain stimulation for Parkinson’s disease dissociates mood and motor circuits: A functional MRI case study. Movement Disorders, 18(12), 1508–1541. doi:10.1002/mds.10593
Temel, Y., Blokland, A., Steinbusch, H. W. M., & Visser-Vandewalle, V. (2005). The functional role of the subthalamic nucleus in cognitive and limbic circuits. Progress in Neurobiology, 76(6), 393–413. doi:10.1016/j.pneurobio.2005.09.005
Turner, R. (2016). Uses, misuses, new uses and fundamental limitations of magnetic
resonance imaging in cognitive science. Philosophical Transactions of the Royal Society B: Biological Sciences, 371(1705). doi:10.1098/rstb.2015.0349
Von Campenhausen, S., Bornschein, B., Wick, R., Bötzel, K., Sampaio, C., Poewe, W., … Dodel, R. (2005). Prevalence and incidence of Parkinson’s disease in Europe. European Neuropsychopharmacology, 15(4), 473–490. doi:10.1016/j.euroneuro.2005.04.007
Wårdell, K., Hemm-Ode, S., Rejmstad, P., & Zsigmond, P. (2016). High-resolution laser Doppler measurements of microcirculation in the deep brain structures: A method for potential vessel tracking. Stereotactic and Functional Neurosurgery, 94(1), 1–9. doi:10.1159/000442894
Wårdell, K., Zsigmond, P., Richter, J., & Hemm, S. (2013). Relationship between laser
doppler signals and anatomy during deep brain stimulation electrode implantation toward the ventral intermediate nucleus and subthalamic nucleus. Neurosurgery, 72, 127–140. doi:10.1227/NEU.0b013e31827e5821
26 Weiss, M., Alkemade, A., Keuken, M. C., Muller-Axt, C., Geyer, S., Turner, R., &
Forstmann, B. U. (2014). Spatial normalization of ultrahigh resolution 7 T magnetic resonance imaging data of the postmortem human subthalamic nucleus: a multistage approach. Brain Structure and Function, 1695–1703. doi:10.1007/s00429-014-0754-4 Windels, F., Bruet, N., Poupard, A., Feuerstein, C., Bertrand, A., & Savasta, M. (2003).
Influence of the frequency parameter on extracellular glutamate and α-aminobutyric acid in substantia nigra and globus pallidus during electrical stimulation of subthalamic nucleus in rats. Journal of Neuroscience Research, 72(2), 259–267.
doi:10.1002/jnr.10577
Wölfle, S. E., Chaston, D. J., Goto, K., Sandow, S. L., Edwards, F. R., & Hill, C. E. (2011). Non-linear relationship between hyperpolarisation and relaxation enables long distance propagation of vasodilatation. The Journal of Physiology, 589, 2607–23.
