Alzheimer’s disease (AD) is the most common form of dementia worldwide, with the vast majority of cases being sporadic and occurring at age 65 years and over. Population-based studies have shown that AD and vascular pathologies commonly coexist in the brains of elderly individuals (Kapasi et al., 2017; Matthews et al., 2009; Neuropathology Group. Medical Research Council Cognitive and Aging, 2001; Rahimi and Kovacs, 2014). A major cardiovascular pathology that affects as many as up to 60% of all individuals after the age of 55 is atherosclerosis. Atherosclerosis is the progressive thickening, hardening and narrowing of major arteries, including those that supply the brain, such as the carotids (Lusis, 2000). Intracranial atherosclerosis does not occur until much later in life, around 75 years and above. As such, Alzheimer’s disease that begins around the 8th decade of life is usually present with other comorbidities such as atherosclerosis (Napoli et al., 1999). There is also evidence that, not only do these often exist as comorbidities, but they may interact pathogenically with vascular disease and neurovascular unit changes contributing to AD (Iadecola, 2017; Kapasi and Schneider, 2016). To date, there are very limited models of comorbidity with respect to preclinical studies, and instead models have been very specific and ‘pure’, and not reflective of the clinical pathology in humans. Atherosclerosis is known to be a major risk factor for the development of dementia. The progressive atheromatous plaque build-up within cerebral arteries that supply the cortex over time can lead to stenosis producing insufficient oxygen delivery to the brain parenchyma, potentially resulting in neuronal death and symptoms of dementia (Shabir et al., 2018). Indeed, the vascular cognitive impairment (VCI) which precedes the onset of dementia may be attributed to a variety of different vascular pathologies affecting either systemic or intracranial vasculature (both large or small vessels) (Iadecola et al., 2019). Due to the complexity of atherosclerosis and dementia pathogenesis, understanding the mechanisms of their mutual interactions is necessary if efforts to develop therapeutics to prevent VCI and vascular dementia, which currently has no disease-modifying cure, are to succeed.

The breakdown of NVC is thought to be an important and early pathogenic mechanism in the onset and progression of a range of neurological conditions (Zlokovic, 2011). In the present study, we aimed to investigate neurovascular function in mid-aged (9–12 m old) mice where atherosclerosis was a comorbidity. We used a novel model of atherosclerosis that utilises a single adeno-associated virus (AAV) i.v. injection of a gain-of-function mutation (D377Y) to proprotein convertase subtilisin/kexin type 9 (rAAV8-mPCSK9-D377Y), combined with a high-fat Western diet to induce atherosclerosis in most adult mouse strains (Bjørklund et al., 2014; Roche-Molina et al., 2015). This leads to the constitutively active inhibition of the LDL-receptor preventing cholesterol internalisation and degradation by hepatocytes, leading to hypercholesterolaemia to occur and the development of robust atherosclerotic lesions within 6–8 weeks (Bjørklund et al., 2014). Furthermore, in order to address the effect atherosclerosis could have on mild Alzheimer’s pathology, we combined the atherosclerosis with the mild J20-hAPP mouse model of familial Alzheimer’s disease (fAD) to create a mixed comorbid mouse model (J20-PCSK9-MIX). The J20-hAPP mouse model of fAD over-expresses human amyloid precursor protein (hAPP) with the Swedish (K670N and M671L) and the Indiana (V7171F) familial mutations (Mucke et al., 2000), which begin to develop amyloid-beta (Aβ) plaques around 5–6 months of age, and show signs of cognitive impairments from 4 months (Ameen-Ali et al., 2019). We hypothesised that atherosclerosis would exacerbate Alzheimer’s disease pathology in the brain and that neurovascular function would be further worsened compared to AD or ATH models alone. We have previously reported no significant alterations to evoked-haemodynamics in the J20-AD model of the same age (9–12 m); however, under acute imaging sessions where an electrode was inserted into the brain, we found significantly perturbed haemodynamics (Sharp et al., 2020). We hypothesised that electrode insertion causes cortical spreading depression (CSD). Based on recent data linking migraine with aura with cardiovascular disease (Kurth et al., 2020), we hypothesised that experimental CSD might be heightened in all disease models. We report that experimentally induced atherosclerosis in the J20-AD model increased the number of Aβ plaques by 300%. Furthermore, experimental CSD is more severe in all diseased groups compared to WT controls.

The present study investigated neurovascular function in a novel experimental model of atherosclerosis (PCSK9-ATH) and for the first time, in a comorbid setting whereby atherosclerosis was experimentally induced in a well characterised model of AD; J20-hAPP, to create a mixed comorbid model (J20-PCSK9-MIX). These mice were compared to age-matched (9–12 m) WT C57BL/6 J controls, and J20-AD mice. Given that systemic atherosclerosis is a major risk factor for dementia, the mechanisms underpinning the relationship between atherosclerosis, neurovascular decline and dementia are still largely unclear.

In the first part of the study, we characterised evoked-haemodynamic responses using a chronic skull-intact and surgery-recovered mouse preparation. We found that PCSK9-ATH mice displayed significantly reduced evoked blood volume (HbT) responses, in addition to reduced levels of oxyhaemoglobin (HbO) and notably, an impaired washout of deoxyhaemoglobin (HbR) across all stimulations and conditions. The J20-PCSK9-MIX mice did not display a significant reduction in HbT, nor in HbO or HbR levels but did trend towards a reduction in both. Interestingly, J20-PCSK9-MIX mice displayed a triphasic HbR profile compared to the typical biphasic shape seen in the other groups, with an initial inverted response. This could reflect worse overall metabolic deficits despite functional neurovascular coupling. With respect to J20-AD mice, we did not observe any significant alterations to HbT as previously published (Sharp et al., 2020). Another important finding from the present study was that 10%-hypercapnia responses were not significantly different in any of the mice compared to WT controls (Figure 2—figure supplement 4), thus suggesting that vascular reactivity was not impaired in any of the mice, indicating that cerebral arterioles were unaffected by atherosclerosis at this time-point (9–12 m). Thus, the basis of reduced HbT in PCSK9-ATH mice cannot be attributed to intracranial atherosclerosis. A recent study corroborated our findings showing that cerebrovascular reactivity and cerebral blood flow was preserved in the Tg-SwDI model of Alzheimer’s (Munting et al., 2021), similar to our J20-model.

In the second part of the study, we obtained neural multi-unit activity (MUA) by inserting a multichannel electrode into the active region defined from the chronic imaging experiments. As we showed in our previous reports (Shabir et al., 2020; Sharp et al., 2020), the technical procedure of electrode insertion causes a cortical spreading depression (CSD) to occur in all animals. Here, we describe the CSD and its recovery on the different disease groups. CSD has two distinct phases: (1) a wave of depolarisation within the grey matter characterised by neuronal distortion leading to a large change of the membrane potential whereby neuronal activity is silenced (spreading depression) and (2) haemodynamic changes that accompany neuronal spreading depolarisation which typically result in a wave of prolonged reduced perfusion that persists for some time (Ayata and Lauritzen, 2015; Dreier, 2011). CSD does not typically occur in healthy brain tissue, however, it is a common neurophysiological occurrence in certain pathological conditions including migraine, epilepsy, brain injury, hyperthermia, hypoxia, and ischaemia (Dreier, 2011). In WT mice, the initial reduction in HbT is small, and a robust haemodynamic recovery occurs which allows for neurovascular coupling to occur to sustain neurons metabolically. This is a marked difference to the diseased animals, which upon electrode insertion to cause a CSD, exhibit a large reduction in HbT with an extremely limited haemodynamic recovery resulting in a prolonged constriction, severe reductions to blood volume and HbO and HbR levels indicating hypoxia and ischaemia. We would also expect to see increased basal levels of ischaemia and hypoxia to some extent in the disease models, particularly the mixed model. Investigation of such markers will form the basis of future studies. Thus, the haemodynamic response to CSD in diseased mice is severely inappropriate and can lead to long lasting devastating effects such as widespread cortical pannecrosis of neurons and astrocytes (Dreier, 2011). As our data show, baseline blood volumes do not recover in the diseased animals for a much longer period compared to WT animals, with the most profound CSD occurring in J20-AD mice, followed by PCSK9-ATH mice.

CSD may be the neuropathological link between migraine, stroke, cardiovascular disease and dementia in which cardiovascular risk factors, genetics and other lifestyle factors which prime the onset of migraine to occur lead to vascular vulnerability within the brain predisposing affected individuals to an increased risk of cerebral ischaemia and haemorrhagic stroke (Ripa et al., 2015). There is accumulating evidence to suggest that shared genetic and associated clinical features observed in migraine patients are involved in the increased vulnerability to cerebral ischaemia, therefore, predisposing affected individuals to stroke and white matter lesions associated with dementia (Yemisci and Eikermann-Haerter, 2019). The underlying mechanism being CSD; the neurophysiological feature of aura in migraines, whose induction threshold can be reduced by genetic mutations and systemic comorbidities that contribute to vascular dysfunction and neuroinflammation (Yemisci and Eikermann-Haerter, 2019). Indeed, mouse models of cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy syndrome (CADASIL); a genetic cerebrovascular disease caused by NOTCH3 mutations that has a high frequency of migraines with aura, have enhanced CSD linking a dysfunctional neurovascular unit with migraine with aura (Eikermann-Haerter et al., 2011). Furthermore, a recent study examined women who suffered from migraines with and without aura and found that those that suffered migraines with aura had a higher incidence rate of cardiovascular disease compared to women without aura or any migraines (Kurth et al., 2020). In addition, another recent study found that migraine history was positively associated with an increased risk of developing both all-cause dementia and AD, but not VaD (Morton et al., 2019). Our study, along with the previously discussed studies provide an explanation for these recent findings and highlights how systemic disease can prime the brain to allow profound CSDs to occur in the context of migraine, and as such, migraine frequency and intensity may be related to the onset of neurological disease by later in life including dementia.

Neural MUA data; generated by local neuronal action potentials, was not significantly altered across any of the stimulations or conditions for any of the disease groups compared to WT controls (Figure 4A), although it does appear that there may be a trend toward slightly higher MUA for WT mice. A consistent finding irrespective of stimulation and condition was that PCSK9-ATH mice display consistently reduced evoked-HbT responses against evoked MUA (observed in chronic and acute experiments respectively) compared to WT controls, which suggests an advanced level of neurovascular breakdown and inefficiency (Figure 4B). This is further supported by the whisker stimulation data while animals breathed air (21% oxygen) where HbR is inverted in PCSK9-ATH and J20-PCSK9-ATH mice (Figure 2); as such, when breathing air the mice could be hypoxic and have a metabolic deficit. Other groups have found similarly reduced blood flow in the ApoE^-/- model without altered cortical activation (Ayata et al., 2013). A recent study found decreased tissue oxygenation in the LDLR^-/- mouse model of atherosclerosis (Li et al., 2019), and this is most likely to be the case in the PCSK9 model. Initially, the effect of CSD confounded the neurovascular assessment for both J20-AD and J20-PCSK9-MIX mice. Haemodynamic responses to the first 2 s and 16 s stimulations were altered in these mice, as compared to WT. However, haemodynamic responses to later stimulations matched those of WT mice, confirming the non-significant effect of disease. Throughout the whole experimental period, neurovascular measures in PCSK9-ATH mice remained consistently impaired compared to WT (Figure 4B). Together, these data indicate that whilst CSD has a profound effect on stimulation-evoked haemodynamic responses in all disease groups, time allows for recovery for WT, J20-AD and J20-PCSK9-MIX groups, but PCSK9-ATH; despite being recovered, are still affected due to genuine long-lasting neurovascular deficits.

A question that arises is why the J20-PCSK9-MIX mice HbT responses are not more severely impaired than J20-AD and PCSK9-ATH? There may be redundancies that occur physiologically to compensate for mild hypoxia in the brain, such as the possible angiogenesis within the brain. Angiogenesis is known to be triggered in cerebral microvessels in AD in response to increased Aβ and neuroinflammation and may initially reflect a compensatory mechanism to increase perfusion (Jefferies et al., 2013). In addition, the levels of neuroinflammation seen in these mice may be due to an altered disease-course and examining temporospatial expression may reveal much higher levels of inflammation in this mixed model at an earlier time-point. Other markers of inflammation may be upregulated compared to those that we assessed, and future studies would incorporate transcriptomic approaches to identify other mechanisms or markers. Nevertheless, a key translational finding from our study was that J20-PCSK9-MIX mice displayed a significant increase in the number of hippocampal plaques.

There are several notable limitations with the present study. Firstly, all imaging was performed on lightly anaesthetised animals, which is known to compromise neurovascular function (Gao et al., 2017). However, previous research from our laboratory has developed an anaesthetic regimen that is comparable to awake imaging in terms of the haemodynamic responses to physiological whisker stimulation with little effect on vascular reactivity (Sharp et al., 2015). The benefits of lightly anaesthetised preparations over awake preparations is that we can avoid the multiple considerations of behavioural state in which the animals may be whisking, grooming as well as their arousal and stress states which may be present in awake animals. Furthermore, we report the stability and robustness of our imaging preparation in this study. We present the average of all the raw stimulation trials from each animal across the whole experimental session (Figure 2—figure supplement 3), showing the stability and robustness of our preparation, as well as easily identifying any changes. Future studies directly comparing stimulation-evoked haemodynamic responses in anaesthetised and awake PCSK9-ATH and J20-PCSK9-MIX mice would be useful in determining if such responses were still comparable under these diseased states. We have ensured that the level of anaesthesia is minimal and by examining vascular reactivity are able to ensure all animals are responsive to the same extent. Secondly, our imaging analysis assumes O₂ saturation to be 70% with a baseline haemoglobin concentration of 100 µM. This may be important if the assumed baselines are different in the diseased animals compared to WT controls; however, our recent study (Sharp et al., 2020) using the same J20-AD mouse model discussed this issue in detail, in which we showed that regardless of the baseline blood volume estimation used, our percentage change was scaled by it (i.e. always the same change). Therefore, the observations in this paper with respect to the different diseased animals are robust.

In conclusion, we report novel findings of impaired neurovascular function in a novel experimental model of atherosclerosis (PCSK9-ATH) characterised by reduced stimulus-evoked blood volume without any significant alterations to evoked neural activity showing evidence of neurovascular uncoupling. We induced atherosclerosis in a mild fAD model (J20-AD) to create a mixed comorbid model (J20-PCSK9-ATH) in which we report a significant increase in the number of hippocampal Aβ plaques, however, without any significant changes to evoked haemodynamic or neural responses compared to WT or J20-AD mice. A key finding from this study was CSD was more severe in diseased animals. This may reflect the global inflammatory state of the brain and could also serve to be an effective preclinical and human clinical biomarker for baseline state and to assess therapies. Future studies should include assessment of other inflammatory markers and cellular pathway changes by a genome wide transcriptomics approach from single cell populations. It would also be prudent to induce atherosclerosis in a more severe fAD model to provide a severity continuum of mixed models that reflect clinical presentations of dementia.

Data files including figure generating MATLAB code and numerical values used for statistical analyses (Excel and SPSS) have been uploaded to Dyrad (doi: https://doi.org/10.5061/dryad.z08kprrcj).

1. 1. Waggitt J

   (2021) Dryad Digital Repository

   Data from: Assessment of Neurovascular Coupling & Cortical Spreading Depression in Mixed Mouse Models of Atherosclerosis and Alzheimer's Disease.

   https://doi.org/10.5061/dryad.z08kprrcj

1. 1. Ameen-Ali KE
   2. Simpson JE
   3. Wharton SB
   4. Heath PR
   5. Sharp PS
   6. Brezzo G
   7. Berwick J

   (2019) The Time Course of Recognition Memory Impairment and Glial Pathology in the hAPP-J20 Mouse Model of Alzheimer’s Disease

   Journal of Alzheimer’s Disease 68:609–624.

   https://doi.org/10.3233/JAD-181238

   • PubMed
   • Google Scholar
2. 1. Ayata Cenk
   2. Shin HK
   3. Dileköz E
   4. Atochin DN
   5. Kashiwagi S
   6. Eikermann-Haerter K
   7. Huang PL

   (2013) Hyperlipidemia disrupts cerebrovascular reflexes and worsens ischemic perfusion defect

   Journal of Cerebral Blood Flow and Metabolism 33:954–962.

   https://doi.org/10.1038/jcbfm.2013.38

   • PubMed
   • Google Scholar
3. 1. Ayata C.
   2. Lauritzen M

   (2015) Spreading Depression, Spreading Depolarizations, and the Cerebral Vasculature

   Physiological Reviews 95:953–993.

   https://doi.org/10.1152/physrev.00027.2014

   • PubMed
   • Google Scholar
4. 1. Berwick J
   2. Johnston D
   3. Jones M
   4. Martindale J
   5. Redgrave P
   6. McLoughlin N
   7. Schiessl I
   8. Mayhew JEW

   (2005) Neurovascular coupling investigated with two-dimensional optical imaging spectroscopy in rat whisker barrel cortex

   The European Journal of Neuroscience 22:1655–1666.

   https://doi.org/10.1111/j.1460-9568.2005.04347.x

   • PubMed
   • Google Scholar
5. 1. Bjørklund MM
   2. Hollensen AK
   3. Hagensen MK
   4. Dagnaes-Hansen F
   5. Christoffersen C
   6. Mikkelsen JG
   7. Bentzon JF

   (2014) Induction of atherosclerosis in mice and hamsters without germline genetic engineering

   Circulation Research 114:1684–1689.

   https://doi.org/10.1161/CIRCRESAHA.114.302937

   • PubMed
   • Google Scholar
6. 1. Dreier JP

   (2011) The role of spreading depression, spreading depolarization and spreading ischemia in neurological disease

   Nature Medicine 17:439–447.

   https://doi.org/10.1038/nm.2333

   • PubMed
   • Google Scholar
7. 1. Eikermann-Haerter K
   2. Yuzawa I
   3. Dilekoz E
   4. Joutel A
   5. Moskowitz MA
   6. Ayata C

   (2011) Cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy syndrome mutations increase susceptibility to spreading depression

   Annals of Neurology 69:413–418.

   https://doi.org/10.1002/ana.22281

   • PubMed
   • Google Scholar
8. 1. Gao YR
   2. Ma Y
   3. Zhang Q
   4. Winder AT
   5. Liang Z
   6. Antinori L
   7. Drew PJ
   8. Zhang N

   (2017) Time to wake up: Studying neurovascular coupling and brain-wide circuit function in the un-anesthetized animal

   NeuroImage 153:382–398.

   https://doi.org/10.1016/j.neuroimage.2016.11.069

   • PubMed
   • Google Scholar
9. 1. Iadecola C

   (2017) The Neurovascular Unit Coming of Age: A Journey through Neurovascular Coupling in Health and Disease

   Neuron 96:17–42.

   https://doi.org/10.1016/j.neuron.2017.07.030

   • PubMed
   • Google Scholar
10. 1. Iadecola C
    2. Duering M
    3. Hachinski V
    4. Joutel A
    5. Pendlebury ST
    6. Schneider JA
    7. Dichgans M

    (2019) Vascular Cognitive Impairment and Dementia

    Journal of the American College of Cardiology 73:3326–3344.

    https://doi.org/10.1016/j.jacc.2019.04.034

    • Google Scholar
11. 1. Jefferies WA
    2. Price KA
    3. Biron KE
    4. Fenninger F
    5. Pfeifer CG
    6. Dickstein DL

    (2013) Adjusting the compass: new insights into the role of angiogenesis in Alzheimer’s disease

    Alzheimer’s Research & Therapy 5:64.

    https://doi.org/10.1186/alzrt230

    • PubMed
    • Google Scholar
12. 1. Kapasi A
    2. Schneider JA

    (2016) Vascular contributions to cognitive impairment, clinical Alzheimer’s disease, and dementia in older persons

    Biochimica et Biophysica Acta 1862:878–886.

    https://doi.org/10.1016/j.bbadis.2015.12.023

    • PubMed
    • Google Scholar
13. 1. Kapasi A
    2. DeCarli C
    3. Schneider JA

    (2017) Impact of multiple pathologies on the threshold for clinically overt dementia

    Acta Neuropathologica 134:171–186.

    https://doi.org/10.1007/s00401-017-1717-7

    • PubMed
    • Google Scholar
14. 1. Kurth T
    2. Rist PM
    3. Ridker PM
    4. Kotler G
    5. Bubes V
    6. Buring JE

    (2020) Association of Migraine With Aura and Other Risk Factors With Incident Cardiovascular Disease in Women

    JAMA 323:2281–2289.

    https://doi.org/10.1001/jama.2020.7172

    • PubMed
    • Google Scholar
15. 1. Li B
    2. Lu X
    3. Moeini M
    4. Sakadžić S
    5. Thorin E
    6. Lesage F

    (2019) Atherosclerosis is associated with a decrease in cerebral microvascular blood flow and tissue oxygenation

    PLOS ONE 14:e0221547.

    https://doi.org/10.1371/journal.pone.0221547

    • PubMed
    • Google Scholar
16. 1. Livak KJ
    2. Schmittgen TD

    (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method

    Methods 25:402–408.

    https://doi.org/10.1006/meth.2001.1262

    • PubMed
    • Google Scholar
17. 1. Lusis AJ

    (2000) Atherosclerosis

    Nature 407:233–241.

    https://doi.org/10.1038/35025203

    • PubMed
    • Google Scholar
18. 1. Madisen L
    2. Mao T
    3. Koch H
    4. Zhuo J
    5. Berenyi A
    6. Fujisawa S
    7. Hsu YWA
    8. Garcia AJ
    9. Gu X
    10. Zanella S
    11. Kidney J
    12. Gu H
    13. Mao Y
    14. Hooks BM
    15. Boyden ES
    16. Buzsáki G
    17. Ramirez JM
    18. Jones AR
    19. Svoboda K
    20. Han X
    21. Turner EE
    22. Zeng H

    (2012) A toolbox of Cre-dependent optogenetic transgenic mice for light-induced activation and silencing

    Nature Neuroscience 15:793–802.

    https://doi.org/10.1038/nn.3078

    • PubMed
    • Google Scholar
19. 1. Matthews FE
    2. Brayne C
    3. Lowe J
    4. McKeith I
    5. Wharton SB
    6. Ince P

    (2009) Epidemiological pathology of dementia: attributable-risks at death in the Medical Research Council Cognitive Function and Ageing Study

    PLOS Medicine 6:e1000180.

    https://doi.org/10.1371/journal.pmed.1000180

    • PubMed
    • Google Scholar
20. 1. Mayhew J
    2. Zheng Y
    3. Hou Y
    4. Vuksanovic B
    5. Berwick J
    6. Askew S
    7. Coffey P

    (1999) Spectroscopic analysis of changes in remitted illumination: the response to increased neural activity in brain

    NeuroImage 10:304–326.

    https://doi.org/10.1006/nimg.1999.0460

    • PubMed
    • Google Scholar
21. 1. Morton RE
    2. St John PD
    3. Tyas SL

    (2019) Migraine and the risk of all-cause dementia, Alzheimer’s disease, and vascular dementia: A prospective cohort study in community-dwelling older adults

    International Journal of Geriatric Psychiatry 34:1667–1676.

    https://doi.org/10.1002/gps.5180

    • PubMed
    • Google Scholar
22. 1. Mucke L
    2. Masliah E
    3. Yu GQ
    4. Mallory M
    5. Rockenstein EM
    6. Tatsuno G
    7. Hu K
    8. Kholodenko D
    9. Johnson-Wood K
    10. McConlogue L

    (2000) High-level neuronal expression of abeta 1-42 in wild-type human amyloid protein precursor transgenic mice: synaptotoxicity without plaque formation

    The Journal of Neuroscience 20:4050–4058.

    https://doi.org/10.1523/JNEUROSCI.20-11-04050.2000

    • PubMed
    • Google Scholar
23. 1. Munting LP
    2. Derieppe M
    3. Suidgeest E
    4. Hirschler L
    5. van Osch MJ
    6. Denis de Senneville B
    7. van der Weerd L

    (2021) Cerebral blood flow and cerebrovascular reactivity are preserved in a mouse model of cerebral microvascular amyloidosis

    eLife 10:e61279.

    https://doi.org/10.7554/eLife.61279

    • PubMed
    • Google Scholar
24. 1. Napoli C
    2. Witztum JL
    3. de Nigris F
    4. Palumbo G
    5. D’Armiento FP
    6. Palinski W

    (1999) Intracranial arteries of human fetuses are more resistant to hypercholesterolemia-induced fatty streak formation than extracranial arteries

    Circulation 99:2003–2010.

    https://doi.org/10.1161/01.cir.99.15.2003

    • PubMed
    • Google Scholar
25. 1. Neuropathology Group. Medical Research Council Cognitive F
    2. Aging S

    (2001) Pathological correlates of late-onset dementia in a multicentre, community-based population in England and Wales. Neuropathology Group of the Medical Research Council Cognitive Function and Ageing Study (MRC CFAS)

    Lancet 357:169–175.

    https://doi.org/10.1016/s0140-6736(00)03589-3

    • PubMed
    • Google Scholar
26. 1. Rahimi J
    2. Kovacs GG

    (2014) Prevalence of mixed pathologies in the aging brain

    Alzheimer’s Research & Therapy 6:82.

    https://doi.org/10.1186/s13195-014-0082-1

    • PubMed
    • Google Scholar
27. 1. Ripa P
    2. Ornello R
    3. Pistoia F
    4. Carolei A
    5. Sacco S

    (2015) Spreading depolarization may link migraine, stroke, and other cardiovascular disease

    Headache 55:180–182.

    https://doi.org/10.1111/head.12436

    • PubMed
    • Google Scholar
28. 1. Roche-Molina M
    2. Sanz-Rosa D
    3. Cruz FM
    4. García-Prieto J
    5. López S
    6. Abia R
    7. Muriana FJG
    8. Fuster V
    9. Ibáñez B
    10. Bernal JA

    (2015) Induction of sustained hypercholesterolemia by single adeno-associated virus-mediated gene transfer of mutant hPCSK9

    Arteriosclerosis, Thrombosis, and Vascular Biology 35:50–59.

    https://doi.org/10.1161/ATVBAHA.114.303617

    • PubMed
    • Google Scholar
29. 1. Shabir O
    2. Berwick J
    3. Francis SE

    (2018) Neurovascular dysfunction in vascular dementia, Alzheimer’s and atherosclerosis

    BMC Neuroscience 19:62.

    https://doi.org/10.1186/s12868-018-0465-5

    • PubMed
    • Google Scholar
30. 1. Shabir O
    2. Sharp P
    3. Rebollar MA
    4. Boorman L
    5. Howarth C
    6. Wharton SB
    7. Francis SE
    8. Berwick J

    (2020) Enhanced Cerebral Blood Volume under Normobaric Hyperoxia in the J20-hAPP Mouse Model of Alzheimer’s Disease

    Scientific Reports 10:7518.

    https://doi.org/10.1038/s41598-020-64334-4

    • PubMed
    • Google Scholar
31. 1. Sharp PS
    2. Shaw K
    3. Boorman L
    4. Harris S
    5. Kennerley AJ
    6. Azzouz M
    7. Berwick J

    (2015) Comparison of stimulus-evoked cerebral hemodynamics in the awake mouse and under a novel anesthetic regime

    Scientific Reports 5:12621.

    https://doi.org/10.1038/srep12621

    • PubMed
    • Google Scholar
32. 1. Sharp PS
    2. Ameen-Ali KE
    3. Boorman L
    4. Harris S
    5. Wharton S
    6. Howarth C
    7. Shabir O
    8. Redgrave P
    9. Berwick J

    (2020) Neurovascular coupling preserved in a chronic mouse model of Alzheimer’s disease: Methodology is critical

    Journal of Cerebral Blood Flow and Metabolism 40:2289–2303.

    https://doi.org/10.1177/0271678X19890830

    • PubMed
    • Google Scholar
33. 1. Taniguchi H
    2. He M
    3. Wu P
    4. Kim S
    5. Paik R
    6. Sugino K
    7. Kvitsiani D
    8. Kvitsani D
    9. Fu Y
    10. Lu J
    11. Lin Y
    12. Miyoshi G
    13. Shima Y
    14. Fishell G
    15. Nelson SB
    16. Huang ZJ

    (2011) A resource of Cre driver lines for genetic targeting of GABAergic neurons in cerebral cortex

    Neuron 71:995–1013.

    https://doi.org/10.1016/j.neuron.2011.07.026

    • PubMed
    • Google Scholar
34. 1. Yemisci M
    2. Eikermann-Haerter K

    (2019) Aura and Stroke: relationship and what we have learnt from preclinical models

    The Journal of Headache and Pain 20:63.

    https://doi.org/10.1186/s10194-019-1016-x

    • PubMed
    • Google Scholar
35. 1. Zlokovic BV

    (2011) Neurovascular pathways to neurodegeneration in Alzheimer’s disease and other disorders

    Nature Reviews. Neuroscience 12:723–738.

    https://doi.org/10.1038/nrn3114

    • PubMed
    • Google Scholar

Decision letter after peer review:

Thank you for submitting your article "Assessment of Neurovascular Coupling and Cortical Spreading Depression in Mixed Models of Atherosclerosis and Alzheimer's Disease" for consideration by eLife. Your article has been reviewed by 2 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this letter to help you prepare a revised submission.

Essential revisions:

The Reviewers pointed out that all of the points raised in their comments are considered essential revisions and to be addressed with additional data and analyses.Reviewer #1:

Strengths:

The data presented in the paper is of pristine quality and the experimental setting has the power to uncover the changes in neurovascular coupling in a novel model of Alzheimer's disease and arteriosclerosis comorbidity; making this work particularly attractive and interesting.

Weaknesses:

1. Neurovascular "coupling" is not directly assessed and more detailed analysis of the hemodynamics is required to contrast between groups.

2. The paper is largely descriptive and no mechanistic insight is provided.

3. Some interpretation of the data ignores trends in the data (even if not statistically significant), that could lead to different conclusion if group sizes were larger or data analyzed differently.

4. The authors claim that the exaggerated HbT response to CSD in diseased animals results in increased hypoxia and ischemia yet no direct evidence is provided to support this claim.

5. Potential differential interactions between anesthetic agent with neuronal and vascular responses between experimental groups is not addressed.

Discussed here below are the points raised by this reviewer that need to be addresses to strengthen the claims made in the manuscript; they are elaborated in the same order as presented in the "Weakness" section above.

1. The most fundamental message that the authors aim to deliver in the paper is the abnormal neurovascular coupling, particularly in the MIX model. The conclusion is largely based on the fact that the WT mice had a significantly larger peak HbT amplitude (Figure 2A and B), this is just one metric that can be extracted from these time series. For example, there seems to be no differences in onset time or time to peak (obviously a rigorous analysis is required). On the flip side, a close-up look at the initial phase of the HbR shows that, for the WT, there is an almost immediate decrease in the fractional change while the other models show a more delayed (even opposite) directional change, particularly for the 2s stimulation under 100% and 21% oxygen. It should be notice that the shorter the stim, the more realistic it becomes given the nature of the whisker system. Even if one focuses solely on the HbT amplitude changes, how do these correlate with the underlying neuronal activity? Looking at the data presented in Figure 4, one can observe a clear trend (not statistically significant by the current analysis) where the activity in the WT has also a larger amplitude. A scatter plot of the peak MUA change vs peak fractional HbT for each trial could prove informative while exploring this questions. Next one would like to test and see if the regression coefficient between these variable changes between the different models and if so, what is the measured difference (i.e. changes in regression slope between groups). Not all spectral bands are equally correlated with haemodynamics so one would need to repeat this analysis for representative bands. Bottom line, there is a good chance to find some hidden gems in this high-quality data and it worths to dive into a more detailed analysis.

2. Alas, by "mechanistic" this reviewer does not implies finding a specific molecular pathway. On the contrary, in the context of the paper, a mechanistic explanation of the results could have been focused on differential responses across specific vascular compartments. For example, uncovering the vascular compartment behind the difference in HbT; one would hypothesize that pial arteries and the penetrating arteries covered by smooth-muscle cells might be the culprit as those are expected to loose functional responses as they become stiffer in arteriosclerosis ,similar to hypertension. Measuring changes in vessel diameter across pial and penetrating vessels (arteries and veins) would have pointed into a potential explanation for the observed phenomena. The authors should also measure changes in capillary structure (i.e. diameter, partial volume etc). Further, a detailed histological survey of the vascular compartments (arterial lumen vs wall diameter), pericyte/smooth-muscle coverage, could have been much more relevant than neuropathology and neuroinflammation data (Figure 5) which do not contribute directly to the neurovascular narrative of the manuscript. Moreover, the vascular and neuronal data are recorded from the cortex where the authors find no pathology hence the hippocampal data is harder to connect to functional findings presented in the manuscript. That said, the increased plaque density in the MIX model is intriguing raising questions regarding aBeta clearing mechanism; testing if this model recapitulates impaired clearing such in hypertension if of much interest albeit might be set aside for another manuscript.

3. Although rigorous statistics should be our guidance wile navigating the obscure waters of hypothesis testing, one cannot avoid but observing trends present in the data. Such observation might have been even used in a grant application as preliminary data, in the sake of asking for additional fundings to hone into the study of an intriguing phenomenon. I found that there are several of these cases that call for mentioning, as the conclusion of this manuscript might have differed if found to be statistically significant. First, I would like to point to the hypercapnia challenge presented in S3. In the chronic conditions (left) the response in the WT appears larger than in the other groups. While there where no statistical differences in peak values, it will be more representative to look at Area Under the Curve (AUC) as it might provide a more robust metric for comparison. A similar trend (mentioned above) can be seen for the multi-unit recording (Figure 4), WT has a larger amplitude in three out of four conditions. It would have been important to discuss, at least, what was the minimal effect size that could have been discovered given the group sizes and measured variance, or how many more animals would have been needed to detect a statistically significant difference for the observed means. It would have been instructive to have additional panels with summary statistics (such as peak amplitude) showing trial-by-trial variability.

4. As pointed by the authors, it is indeed expected that reduced HbT might lead to more pronounced ischemia and hypoxia and the tools to demonstrate are available calling to, at least, a preliminary demonstration that this is indeed the case. It should be noted that one should also expect to see already an increased basal level, at least for hypoxia markers, in the AD and MIX models.

5. The authors point in the discussion to the unavoidable limitations imposed by performing this line of research in anesthetized animals, their relay on previous work from their laboratory showing the specific anesthesia protocol is comparable to awake imaging in terms of vascular responses to whisker stimulation. It is imperative to show that this assumed lack of difference between awake and anesthetized conditions holds true for the AD and MIX models. This reviewer is aware of the complexity of such demonstration as vascular reactivity is expected to be impaired in arteriosclerosis conditions. Similarly, it is fundamental to rule out a potential interaction between the different mouse models and the response to the anesthetic agent in terms of neuronal activity. There is certainly much to debate regarding the optimal experimental setup (light anesthesia vs awake), pros and cons exist for both. We (as experimenters) should aim to make a clear statement about the impact of the experimental conditions on the metrics we report to better interpret our results. While reproducing the same experiment under awake conditions is beyond the scope of this work, we need to make sure vascular physiology and neuronal responses are not differentially affected in each experimental group.

Reviewer #2:

Summary:

Impairments in vascular structure/function likely play a causal role in the cognitive decline associated with aging and neurodegenerative disease. The purpose of this study (by the Berwick group) was to examine changes in cerebral neurovascular coupling in animal models of familial Alzheimer disease (J20 mouse), atherosclerosis (mPCSK9) and a combined comorbidity model. The authors make the novel discovery that both sensory driven and injured based changes in cerebral blood flow (CBF) are perturbed in the atherosclerosis model.

Strengths:

The authors are recognized experts in the methodologies employed, and not surprisingly, the experiments were conducted with rigor. The study incorporates multiple middle aged animal models to study disease related changes in vascular function. Of particular significance is the inclusion of age/disease related co-morbidities such as the atherosclerosis model. As a result, they made the novel discovery that atherosclerosis impairs the regulation of CBF. Overall, I think this is a well executed study whose primary conclusions were mostly justified by the data presented. Although I think this study adds important new knowledge, some additional analyses and changes to data/figure presentation would really enhance the readability of the manuscript.

Weaknesses:

One potential weakness could be that since mice were assessed under anesthesia in both 100% and 21% oxygen (hyperoxia and normoxia) , one wonders if there were baseline differences in cardiovascular parameters (blood oxygenation, BP, HR etc) that could partially explain why cerebral blood flow responses were altered. If baseline blood oxygenation levels were different between animal models (perhaps elevated under normoxic conditions), there may be less need to recruit blood flow to active cortical areas. The authors describe this limitation in the discussion but there are ways to address this issue in atherosclerotic mice, pulse oximetry for example.

1) Regarding the analysis. I found the paper makes the reader work really hard to understand the comparisons, analyses and effects. The data presented in figure 2, 3 and 4 show complex waveforms for each component of HbO, HbT and HbR in each animal model and each experimental condition (eg. Figure 2: 16 specific conditions x 3 waveforms per condition = 48 waveforms presented). However, it was not completely clear to me how these waveforms were analysed and compared across groups. It appears that just peak value was assessed but it left me wondering if other metrics such as total area under the curve, latency, or slope should also considered (maybe they don't have value?). In some cases the polarity of the waveform is bi or tri-phasic, is this aspect accounted for in the analyses? For the readers' sake, could the authors create graphs in Figure 2-4 that effectively captures how these waveforms were compared across groups and statistically analysed (eg. bar graph with individual data points or whisker box plot of peak amplitudes for each group)?

2) One result which confused me was the statement that ATH mice show a significant reduction in HbT, whereas the figure legend 2 starts with the statement that there was no significant overall effect of disease. Since it is clear that the different disease groups were going to be compared against WT mice anyways (a priori planned comparisons), I didn't understand why one would bother with the non-significant omnibus test to begin with.

3) Figure 3, how fast was the electrode inserted relative to the times shown in Figure 3A? Was the electrode still being pushed into the brain 20-60 seconds after touching the pial surface? Is this what the sharp deflection in values represent in Figure 3B? One small request would be to please note this aspect in the Figure. In addition, the main group differences appear to be from ~150-2000 seconds, perhaps the images shown in 3A could be extended to include more of the data beyond 150s in the experiment.

4) Does the mechanical deformation of the brain caused by the electrode insertion (I would expect the dura would first indent, then give way as the electrode pierces it) artificially change the reflectance of different wavelengths off the surface? A control could be to image during electrode insertion in a dead animal or one with brain activity silenced. Secondly, how do the authors know that the electrode induced a spreading depression wave, since that is typically shown with DC recordings? Perhaps this is already known from previous studies, but it should be mentioned in the results with an appropriate citation.

5) I think the authors should spend more text space describing the finding that sensory evoked responses were not different across any disease model 35 min after the electrode was inserted (Supp Figure 1). To help the reader, what does this finding mean in the broader context? Does it conflict at all with the findings presented in Figure 2? Does it reflect the fact that several metrics had not returned to baseline even 35min after electrode insertion?

6) Figure 4, the authors state that neural activity wasn't significantly different between groups but based on area under the curve for fractional changes in activity, there appears to be noticeably more activity in WT mice across the 2 or 16 sec stimulation duration (first 2 panels). Is this aspect not considered in the analysis? If not, it would be helpful to provide a rationale for why.

7) For mRNA analysis of inflammation, it was not clear under what conditions tissue was extracted. Was tissue extracted after an experiment (sticking an electrode in the brain) or in naïve animals?

Essential revisions:

The Reviewers pointed out that all of the points raised in their comments are considered essential revisions and to be addressed with additional data and analyses.

Reviewer #1:

Strengths:

The data presented in the paper is of pristine quality and the experimental setting has the power to uncover the changes in neurovascular coupling in a novel model of Alzheimer's disease and arteriosclerosis comorbidity; making this work particularly attractive and interesting.

We would like to thank the reviewer for their kind comments and recognition that this is an important issue, and the experimental work is a high standard.

Weaknesses:

1. Neurovascular "coupling" is not directly assessed and more detailed analysis of the hemodynamics is required to contrast between groups.

2. The paper is largely descriptive and no mechanistic insight is provided.

3. Some interpretation of the data ignores trends in the data (even if not statistically significant), that could lead to different conclusion if group sizes were larger or data analyzed differently.

4. The authors claim that the exaggerated HbT response to CSD in diseased animals results in increased hypoxia and ischemia yet no direct evidence is provided to support this claim.

5. Potential differential interactions between anesthetic agent with neuronal and vascular responses between experimental groups is not addressed.

We would like to thank the reviewer for highlighting some of the weaknesses and in response to the weaknesses highlighted, we have addressed all the issues (points 1-4) which are detailed below.

Discussed here below are the points raised by this reviewer that need to be addresses to strengthen the claims made in the manuscript; they are elaborated in the same order as presented in the "Weakness" section above.

1. The most fundamental message that the authors aim to deliver in the paper is the abnormal neurovascular coupling, particularly in the MIX model. The conclusion is largely based on the fact that the WT mice had a significantly larger peak HbT amplitude (Figure 2A and B), this is just one metric that can be extracted from these time series. For example, there seems to be no differences in onset time or time to peak (obviously a rigorous analysis is required). On the flip side, a close-up look at the initial phase of the HbR shows that, for the WT, there is an almost immediate decrease in the fractional change while the other models show a more delayed (even opposite) directional change, particularly for the 2s stimulation under 100% and 21% oxygen. It should be notice that the shorter the stim, the more realistic it becomes given the nature of the whisker system. Even if one focuses solely on the HbT amplitude changes, how do these correlate with the underlying neuronal activity? Looking at the data presented in Figure 4, one can observe a clear trend (not statistically significant by the current analysis) where the activity in the WT has also a larger amplitude. A scatter plot of the peak MUA change vs peak fractional HbT for each trial could prove informative while exploring this questions. Next one would like to test and see if the regression coefficient between these variable changes between the different models and if so, what is the measured difference (i.e. changes in regression slope between groups). Not all spectral bands are equally correlated with haemodynamics so one would need to repeat this analysis for representative bands. Bottom line, there is a good chance to find some hidden gems in this high-quality data and it worths to dive into a more detailed analysis.

We thank the reviewer for their kind comments and recognition that this is an important area of research. We agree

We have undertaken more data analysis as suggested. We have analysed the data for the arterial region as well as for the “area under curve” and included that in the supplemental data.

We have included rise times and times to peak data in the supplemental data, confirming there are no differences in these metrics across any of the groups or conditions.

In response to both reviewers’ comments concerning comparison of HbR between groups, we have added the following sentence in the manuscript:

“Examining the shape of the HbR washout in the some of the disease models shows key differences compared to WT mice. Most notably, J20-PCSK9-MIX mice show a triphasic response with an initial increase in HbR followed by a decrease then return to baseline (Figure 2). This contrasts with the biphasic waveform seen in WT, J20-AD and PCSK9-ATH mice; albeit lower in disease mice compared to WT controls.”

We agree that neurovascular coupling was not directly assessed despite having measures for both evoked haemodynamics and neural activity. We have therefore produced a neurovascular correlation scatter plot (shown in Figure 4; 4B) comparing the evoked MUA vs evoked HbT in the 4 different mouse groups in both the first 2 stimulations under oxygen and the second 2 stimulations under air. We find that the regression coefficient is poorer for atherosclerotic mice compared to WT as for the same amount of MUA there is less HbT suggesting neurovascular uncoupling. This finding confirms our fundamental message that atherosclerotic mice have significantly impaired haemodynamics compared to WT mice.

2. Alas, by "mechanistic" this reviewer does not implies finding a specific molecular pathway. On the contrary, in the context of the paper, a mechanistic explanation of the results could have been focused on differential responses across specific vascular compartments. For example, uncovering the vascular compartment behind the difference in HbT; one would hypothesize that pial arteries and the penetrating arteries covered by smooth-muscle cells might be the culprit as those are expected to loose functional responses as they become stiffer in arteriosclerosis ,similar to hypertension. Measuring changes in vessel diameter across pial and penetrating vessels (arteries and veins) would have pointed into a potential explanation for the observed phenomena. The authors should also measure changes in capillary structure (i.e. diameter, partial volume etc). Further, a detailed histological survey of the vascular compartments (arterial lumen vs wall diameter), pericyte/smooth-muscle coverage, could have been much more relevant than neuropathology and neuroinflammation data (Figure 5) which do not contribute directly to the neurovascular narrative of the manuscript. Moreover, the vascular and neuronal data are recorded from the cortex where the authors find no pathology hence the hippocampal data is harder to connect to functional findings presented in the manuscript. That said, the increased plaque density in the MIX model is intriguing raising questions regarding aBeta clearing mechanism; testing if this model recapitulates impaired clearing such in hypertension if of much interest albeit might be set aside for another manuscript.

We appreciate that we do not provide a specific molecular mechanism aside from neuroinflammation and work on this aspect will be the focus of future papers. The interesting mechanistic aspects mentioned by the reviewer are a programme of work all by themselves. At this stage, we wished to present the models and the exciting and detailed neurophysiological findings associated with them to the research community. The focus of our study was to characterise the ATH and MIX models, as compared to both WT and AD, with detailed analysis of the underlying mechanisms, molecular and cellular changes, being beyond the scope of the current study. Such studies (e.g. collagen IV staining of vessel thickness and density will form the basis of future work). We agree with the reviewer that looking at capillary structure in vivo is an interesting area of research, but it is not possible with our optical imaging setup used in this study but should from the basis for future studies also. We have, however, now included data from specific arterial ROIs (as pointed out in response to the above point) in the supplemental information. With respect to the neuroinflammation data, this is an area we have published on previously (Denes et al., JAHA 2012 J Am Heart Assoc 2012 Jun;1(3):e002006. doi: 10.1161/JAHA.112.002006). and we wanted to extend our findings to these new models. Much more detailed analysis of molecular and cellular changes will be the basis of future manuscripts.

With respect to amyloid β staining, we provided the significant differences (hippocampal) as there was not a clear increase within the cortex at this timepoint. However, if you combine whole brain amyloid β then there is a significant increase overall and this has now been reported in a Table 1 in the supplemental information.

3. Although rigorous statistics should be our guidance wile navigating the obscure waters of hypothesis testing, one cannot avoid but observing trends present in the data. Such observation might have been even used in a grant application as preliminary data, in the sake of asking for additional fundings to hone into the study of an intriguing phenomenon. I found that there are several of these cases that call for mentioning, as the conclusion of this manuscript might have differed if found to be statistically significant. First, I would like to point to the hypercapnia challenge presented in S3. In the chronic conditions (left) the response in the WT appears larger than in the other groups. While there where no statistical differences in peak values, it will be more representative to look at Area Under the Curve (AUC) as it might provide a more robust metric for comparison. A similar trend (mentioned above) can be seen for the multi-unit recording (Figure 4), WT has a larger amplitude in three out of four conditions. It would have been important to discuss, at least, what was the minimal effect size that could have been discovered given the group sizes and measured variance, or how many more animals would have been needed to detect a statistically significant difference for the observed means. It would have been instructive to have additional panels with summary statistics (such as peak amplitude) showing trial-by-trial variability.

We thank the reviewer for their suggestion and have investigated the area under curve data for chronic hypercapnia (see supplemental information). Using this metric, we found no significant differences between groups, confirming our conclusion that there are no differences in vascular reactivity between disease models. These findings support our suggestion that all differences are due to functional, rather than vascular deficits.

We have added summary statistics in a table in the supplemental information (Table 1).

4. As pointed by the authors, it is indeed expected that reduced HbT might lead to more pronounced ischemia and hypoxia and the tools to demonstrate are available calling to, at least, a preliminary demonstration that this is indeed the case. It should be noted that one should also expect to see already an increased basal level, at least for hypoxia markers, in the AD and MIX models.

We do not currently have measures of hypoxia directly, but these will be explored on a cellular and molecular level in future publications. We have added the following sentence to the manuscript:

“We would also expect to see increased basal levels of ischaemia and hypoxia to some extent in the disease models, particularly the mixed model. Investigation of such markers will form the basis of future studies.”

5. The authors point in the discussion to the unavoidable limitations imposed by performing this line of research in anesthetized animals, their relay on previous work from their laboratory showing the specific anesthesia protocol is comparable to awake imaging in terms of vascular responses to whisker stimulation. It is imperative to show that this assumed lack of difference between awake and anesthetized conditions holds true for the AD and MIX models. This reviewer is aware of the complexity of such demonstration as vascular reactivity is expected to be impaired in arteriosclerosis conditions. Similarly, it is fundamental to rule out a potential interaction between the different mouse models and the response to the anesthetic agent in terms of neuronal activity. There is certainly much to debate regarding the optimal experimental setup (light anesthesia vs awake), pros and cons exist for both. We (as experimenters) should aim to make a clear statement about the impact of the experimental conditions on the metrics we report to better interpret our results. While reproducing the same experiment under awake conditions is beyond the scope of this work, we need to make sure vascular physiology and neuronal responses are not differentially affected in each experimental group.

We completely agree with the reviewer about their comments regarding anaesthesia and we take the aspects mentioned very seriously indeed. Although these suggestions are not in the scope of the present study we have added the following sentences in the manuscript:

“Additional studies comparing anaesthetised and awake responses would be useful in determining if such responses were still comparable under diseased states. We have ensured that the level of anaesthesia is minimal and by examining vascular reactivity are able to ensure all animals are responsive to the same extent”.

Reviewer #2:

Summary:

Impairments in vascular structure/function likely play a causal role in the cognitive decline associated with aging and neurodegenerative disease. The purpose of this study (by the Berwick group) was to examine changes in cerebral neurovascular coupling in animal models of familial Alzheimer disease (J20 mouse), atherosclerosis (mPCSK9) and a combined comorbidity model. The authors make the novel discovery that both sensory driven and injured based changes in cerebral blood flow (CBF) are perturbed in the atherosclerosis model.

Strengths:

The authors are recognized experts in the methodologies employed, and not surprisingly, the experiments were conducted with rigor. The study incorporates multiple middle aged animal models to study disease related changes in vascular function. Of particular significance is the inclusion of age/disease related co-morbidities such as the atherosclerosis model. As a result, they made the novel discovery that atherosclerosis impairs the regulation of CBF. Overall, I think this is a well executed study whose primary conclusions were mostly justified by the data presented. Although I think this study adds important new knowledge, some additional analyses and changes to data/figure presentation would really enhance the readability of the manuscript.

We thank the reviewer for the very kind comments especially about the rigor of the conducted experiments as we take this very seriously.

Weaknesses:

One potential weakness could be that since mice were assessed under anesthesia in both 100% and 21% oxygen (hyperoxia and normoxia) , one wonders if there were baseline differences in cardiovascular parameters (blood oxygenation, BP, HR etc) that could partially explain why cerebral blood flow responses were altered. If baseline blood oxygenation levels were different between animal models (perhaps elevated under normoxic conditions), there may be less need to recruit blood flow to active cortical areas. The authors describe this limitation in the discussion but there are ways to address this issue in atherosclerotic mice, pulse oximetry for example.

We would like to thank the reviewer for highlighting some potential issues in our analysis and discussion. We have addressed these comments in detail below and we hope to increase the clarity of the analysis undertaken.

1) Regarding the analysis. I found the paper makes the reader work really hard to understand the comparisons, analyses and effects. The data presented in figure 2, 3 and 4 show complex waveforms for each component of HbO, HbT and HbR in each animal model and each experimental condition (eg. Figure 2: 16 specific conditions x 3 waveforms per condition = 48 waveforms presented). However, it was not completely clear to me how these waveforms were analysed and compared across groups. It appears that just peak value was assessed but it left me wondering if other metrics such as total area under the curve, latency, or slope should also considered (maybe they don't have value?). In some cases the polarity of the waveform is bi or tri-phasic, is this aspect accounted for in the analyses? For the readers' sake, could the authors create graphs in Figure 2-4 that effectively captures how these waveforms were compared across groups and statistically analysed (eg. bar graph with individual data points or whisker box plot of peak amplitudes for each group)?

We have produced bar graphs of the peak HbT, HbO and HbR for each of the conditions alongside Figure 2. Similarly, we have added bar graphs showing MUA peak in Figure 4.

We have investigated area under curve values for both haemodynamic and neural MUA data and presented these in the supplemental data, as mentioned in our response to Reviewer 1, above.

With regards to response latency, we have investigated time to peak and rise times and have included graphs for these in the supplemental information. We found no differences in these metrics across any of the groups or conditons.

We have added the following sentence describing the waveforms for HbR in the manuscript: “Examining the shape of the HbR washout in the some of the disease models shows key differences compared to WT mice. Most notably, J20-PCSK9-MIX mice show a triphasic response with an initial increase in HbR followed by a decrease then return to baseline (Figure 2). This contrasts with the biphasic waveform seen in WT, J20-AD and PCSK9-ATH mice; albeit lower in disease mice compared to WT controls.”

2) One result which confused me was the statement that ATH mice show a significant reduction in HbT, whereas the figure legend 2 starts with the statement that there was no significant overall effect of disease. Since it is clear that the different disease groups were going to be compared against WT mice anyways (a priori planned comparisons), I didn't understand why one would bother with the non-significant omnibus test to begin with.

Our main comparison of interest was to compare disease groups to WT, as pointed out by the reviewer. However, for HbO and HbR, the omnibus test revealed a main effect of disease, therefore, we have included the omnibus test results for completeness. For HbT, Dunnett’s test results are sufficient to show the differences in ATH mice compared to WT as the research question was are any of the disease groups different to WT.

3) Figure 3, how fast was the electrode inserted relative to the times shown in Figure 3A? Was the electrode still being pushed into the brain 20-60 seconds after touching the pial surface? Is this what the sharp deflection in values represent in Figure 3B? One small request would be to please note this aspect in the Figure. In addition, the main group differences appear to be from ~150-2000 seconds, perhaps the images shown in 3A could be extended to include more of the data beyond 150s in the experiment.

Insertion of the electrode into the brain occurs at the 0s timepoint, depicted in Figure 3A and is completed within 5s. Therefore, the electrode was not being pushed into the brain 20-60s after touching the pial surface and so the sharp deflection in Figure 3B reflects genuine decreases in HbT caused by CSD (characteristic profile, discussed in more detail in point 4 below). We thank the reviewer for raising this point and have added further information in the figure legend of Figure 3A to clarify this point.

We have now extended the image montage to include spatial figures beyond 150s and also extended this figure to include representative images from ATH and MIX mice.

4) Does the mechanical deformation of the brain caused by the electrode insertion (I would expect the dura would first indent, then give way as the electrode pierces it) artificially change the reflectance of different wavelengths off the surface? A control could be to image during electrode insertion in a dead animal or one with brain activity silenced. Secondly, how do the authors know that the electrode induced a spreading depression wave, since that is typically shown with DC recordings? Perhaps this is already known from previous studies, but it should be mentioned in the results with an appropriate citation.

We do not make a region of interest where the electrode site is. Our ROIs do not include the electrode site itself as such all data is not artificially changed by the electrode being present or any artificial changes in reflectance due to deformation of brain tissue.

The haemodynamic profile which we observe on electrode insertion is characteristic of CSD. The characteristic haemodynamic profiles that occur during CSD are well known and the following citations describe these further. We have added these citations to the text.

Østergaard et al., 2015. Neurovascular coupling during cortical spreading depolarization and -depression. Stroke. 46(5):1392-401.

Sanchez-Porras et al., 2017. Ketamine modulation of the haemodynamic response to spreading depolarization in the gyrencephalic swine brain. J Cereb Blood Flow Metab. 37(5):1720–1734.

Sun et al., 2011. Simultaneous monitoring of intracellular pH changes and hemodynamic response during cortical spreading depression by fluorescence-corrected multimodal optical imaging. Neuroimage. 57(3):873-84.

Chang et al., 2010. Biphasic direct current shift, haemoglobin desaturation and neurovascular uncoupling in cortical spreading depression. Brain. 133(Pt 4):996-1012.

5) I think the authors should spend more text space describing the finding that sensory evoked responses were not different across any disease model 35 min after the electrode was inserted (Supp Figure 1). To help the reader, what does this finding mean in the broader context? Does it conflict at all with the findings presented in Figure 2? Does it reflect the fact that several metrics had not returned to baseline even 35min after electrode insertion?

Confounds of CSD mean that haemodynamics are altered and thus may be different in acute imaging sessions as compared to chronic imaging sessions. As such, haemodynamic responses in chronic and acute experiments are different initially where chronic better reflects the “true” reflection of disease-related haemodynamics, whereas acute sessions reflect the addition of mild trauma caused by electrode insertion in combination with disease where the CSD is bigger. We have added the following sentences to the manuscript:

“This is due to the effect CSD has on baseline and evoked haemodynamics that persist for some time up to and during the experimental stimulations. Thus, the acute data presents a different outlook compared to chronic-only imaging sessions that do not have the added confound of CSD.”

6) Figure 4, the authors state that neural activity wasn't significantly different between groups but based on area under the curve for fractional changes in activity, there appears to be noticeably more activity in WT mice across the 2 or 16 sec stimulation duration (first 2 panels). Is this aspect not considered in the analysis? If not, it would be helpful to provide a rationale for why.

We thank the reviewer for their suggestion and we have now compared area under the curve for fractional changes in multi-unit activity. We do not find any significant differences in MUA using this metric across any of the groups across any of the stimulations. We have reported this finding in Table 1 of the supplemental information.

7) For mRNA analysis of inflammation, it was not clear under what conditions tissue was extracted. Was tissue extracted after an experiment (sticking an electrode in the brain) or in naïve animals?

We have now clarified in the methods that for mRNA analysis, brains were dissected after terminal experiments. Thus, all brains are from the experimental mice rather than naïve ones, allowing neurophysiological and neuropathological findings to be related directly within the same mice.

Author details

1. Osman Shabir

   1. Department of Infection, Immunity and Cardiovascular Disease (IICD), University of Sheffield Medical School, Royal Hallamshire Hospital, Sheffield, United Kingdom
   2. Healthy Lifespan Institute (HELSI), University of Sheffield, Sheffield, United Kingdom
   3. Neuroscience Institute, University of Sheffield, Sheffield, United Kingdom

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing – original draft

   For correspondence

   o.shabir@sheffield.ac.uk

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-7412-6966

2. Ben Pendry

   Sheffield Institute for Translational Neuroscience (SITraN), University of Sheffield, Sheffield, United Kingdom

   Contribution

   Formal analysis, Investigation

   Competing interests

   No competing interests declared

3. Llywelyn Lee

   1. Neuroscience Institute, University of Sheffield, Sheffield, United Kingdom
   2. Sheffield Neurovascular Lab, Department of Psychology, University of Sheffield, Sheffield, United Kingdom

   Contribution

   Formal analysis, Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-3449-9797

4. Beth Eyre

   1. Neuroscience Institute, University of Sheffield, Sheffield, United Kingdom
   2. Sheffield Neurovascular Lab, Department of Psychology, University of Sheffield, Sheffield, United Kingdom

   Contribution

   Investigation, Writing – review and editing

   Competing interests

   No competing interests declared

5. Paul S Sharp

   Medicines Discovery Catapult, Alderley Edge, United Kingdom

   Contribution

   Investigation, Writing – review and editing

   Competing interests

   No competing interests declared

6. Monica A Rebollar

   1. Neuroscience Institute, University of Sheffield, Sheffield, United Kingdom
   2. Sheffield Institute for Translational Neuroscience (SITraN), University of Sheffield, Sheffield, United Kingdom

   Contribution

   Investigation

   Competing interests

   No competing interests declared

7. David Drew

   Department of Infection, Immunity and Cardiovascular Disease (IICD), University of Sheffield Medical School, Royal Hallamshire Hospital, Sheffield, United Kingdom

   Contribution

   Investigation, Resources

   Competing interests

   No competing interests declared

8. Clare Howarth

   1. Healthy Lifespan Institute (HELSI), University of Sheffield, Sheffield, United Kingdom
   2. Neuroscience Institute, University of Sheffield, Sheffield, United Kingdom
   3. Sheffield Neurovascular Lab, Department of Psychology, University of Sheffield, Sheffield, United Kingdom

   Contribution

   Data curation, Formal analysis, Funding acquisition, Investigation, Supervision, Writing – review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-6660-9770

9. Paul R Heath

   Sheffield Institute for Translational Neuroscience (SITraN), University of Sheffield, Sheffield, United Kingdom

   Contribution

   Methodology, Supervision, Writing – review and editing

   Competing interests

   No competing interests declared

10. Stephen B Wharton

    1. Neuroscience Institute, University of Sheffield, Sheffield, United Kingdom
    2. Sheffield Institute for Translational Neuroscience (SITraN), University of Sheffield, Sheffield, United Kingdom

    Contribution

    Conceptualization, Funding acquisition, Writing – review and editing

    Competing interests

    No competing interests declared

11. Sheila E Francis

    1. Department of Infection, Immunity and Cardiovascular Disease (IICD), University of Sheffield Medical School, Royal Hallamshire Hospital, Sheffield, United Kingdom
    2. Healthy Lifespan Institute (HELSI), University of Sheffield, Sheffield, United Kingdom
    3. Neuroscience Institute, University of Sheffield, Sheffield, United Kingdom

    Contribution

    Conceptualization, Funding acquisition, Methodology, Project administration, Supervision, Writing – review and editing

    Contributed equally with

    Jason Berwick

    Competing interests

    No competing interests declared

12. Jason Berwick

    1. Healthy Lifespan Institute (HELSI), University of Sheffield, Sheffield, United Kingdom
    2. Neuroscience Institute, University of Sheffield, Sheffield, United Kingdom
    3. Sheffield Neurovascular Lab, Department of Psychology, University of Sheffield, Sheffield, United Kingdom

    Contribution

    Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Project administration, Software, Supervision, Validation, Writing – review and editing

    Contributed equally with

    Sheila E Francis

    For correspondence

    j.berwick@sheffield.ac.uk

    Competing interests

    No competing interests declared

• 2,577

  Page views

• 306

  Downloads

• 5

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.