Shear-Induced Amyloid Aggregation in the Brain: V. Are Alzheimer’s and Other Amyloid Diseases Initiated in the Lower Brain and Brainstem by Cerebrospinal Fluid Flow Stresses?

Flow stress-induced aggregation of tau and Aβ

Quiescent aqueous solutions of Aβ and other amyloid monomer molecules slowly aggregate naturally in the amyloid cascade [8]. These Aβ molecules aggregate first in low molecular weight oligomers, then protofibrils, fibrils, and plaques. Shear flow-induced aggregation to form Aβ amyloid fibrils was first reported by Dunstan et al. [27]. Cohen et al. [28] investigated the mechanism of this slow fibril formation in quiescent solutions, but also showed that increasing rates of solution agitation significantly increased the rate of Aβ fibril formation. Aβ oligomers have been identified as toxins responsible for AD neuron damage [29, 30].

The mechanisms for in vivo and in vitro Aβ and tau oligomer formation are still uncertain. A significant activation energy is required for conformation change from an Aβ random coil conformation into a cross-beta conformation that precedes Aβ self-aggregation to form oligomers [31]. Aβ oligomers extracted from AD patients and oligomers containing the same number of Aβ monomers but synthesized in the laboratory can have quite different chemical and biological characteristics [32, 33]. The author has proposed that the brain flow-stress and environment provide the activation energy needed for these conformation changes [15].

Surface- and flow-induced aggregation

Studies of many different non-amyloid proteins show that most resist aggregation even at very high shear rates [24]. However, in the presence of certain surfaces, solutions of these same proteins aggregate at much lower shear rates, some even without shear [25]. This property is a serious problem in the manufacture and shipment of proteins. Hydrophobic surfaces have been shown to be quite efficient in promoting agitated protein monomer aggregation [25]. Air-water surfaces, which behave as hydrophobic surfaces, are also very efficient in promoting protein aggregation. Surfaces containing electrostatic charges attract oppositely-charged protein surfaces resulting in protein aggregation. Thus, the combination of flow and surfaces would appear to be critical variables in amyloid monomer protein aggregation, which is, in turn, critical in initiating amyloid disease.

Because of its amino acid composition, Aβ is attracted to both hydrophobic and positive and negative surfaces [34]. Thus, in our work with small samples of Aβ injected into a mobile phase flowing through narrow bore stainless steel capillary tubing [14], nearly 100%of the Aβ was initially adsorbed on the tube’s inner walls. However, when flow was halted for a short period and flow was resumed, part of the adsorbed sample had been released back into the mobile phase, the amount depending on the flow rate. However, repeated sample injections demonstrated increasing retention of each subsequent injected sample, ultimately with no release of the Aβ wall aggregate upon flow cessation. The nature of the type of surface retention depended on both the shear rate and surface characteristics of the previously adsorbed Aβ.

Molecular dynamics studies of the interaction of Aβ with bilayer membranes without shear predict the formation of Aβ barrel structures with membrane pores that allow calcium ions to flow through the membrane [35]. This formation could inactivate a neuron or astrocyte of which that membrane is a part. These calculations did not take into account probable shear effects in CSF at the membrane wall that could alter the conformation of the Aβ just prior to adsorption, thus possibly affecting its subsequent fate on the surface.

Shear, shear rate, and strain

Any time that an aqueous liquid flows in a confined space, it will be subjected to shear flow. Since the liquid wets the wall, the velocity of the innermost water surface layer during flow is zero. However, with flow this generates a liquid velocity gradient as one moves away from the wall. A typical result is represented by the laminar flow profile shown in Fig. 1. As indicated in this figure, the action brought about by laminar flow interacting with a flexible molecule has two components. It is both rotated and distorted. The nature of the molecular distortion is determined by the relative velocities of the different water layers within this profile and the molecule’s resistance to distortion. In a complex system, such as an in vivo fluid like brain CSF, a particular velocity field can be very complex since the fluid flow path wall geometry may also be complex. Shear rates, generally given in sec^–1, are related to the fractional distortion of the molecule’s dimension per unit time. The local shear rate depends on the flow system’s geometry. Changes in any part of this geometry can change the shear rates at relatively distant points. Input flow rates help determine both the flow and shear rates at any given point of an entire closed flow system. Thus, sometimes shear rate information about complex flow systems can only be determined by experimental measurements and computer calculations.

Fig. 1

a) Laminar shear flow: flexible molecule undergoing stretching, distortion, chemical excitation, and molecular rotation. b) Later time than (a) showing direction of rotation.

Strain is defined as the dimensionless magnitude of an object’s deformation, equal to the change in the dimension of the object such as a molecule divided by its original dimension. The strain rate is the rate of change of strain per unit time, generally given in units sec^–1.

Extensional flow and protein aggregation

When fluid flow encounters a narrowing of the flow path or there is flow between flow-impeding objects, as indicated in Fig. 2, there is one particular region of interest. This is a region in which there is increasing acceleration as a dissolved solute molecule moves toward severe narrowing of the flow path. Such a flow situation produces a dominant extensional shear component and a molecule trapped in such a flow undergoes a non-uniform stretch designated as extensional flow of the type illustrated in Fig. 3. Such a flow-induced stretch can change the molecular conformation. This stretch is released after it passes through the AB “narrows” (Fig. 2). Both Aβ and tau are intrinsically disorganized proteins (IDPs) [15], which, when stretched and then relaxed, could well drastically change conformations under such dynamic stress conditions.

Fig. 2

In both cases, extensional shear is found on the left side. A lower extensional shear rate is found between B and C. Predominant laminar shear is found close to A, B, and C surfaces.

Fig. 3

Lines representing protein under extensional flow, with two sets of points, leading and lagging. The line is unevenly stretched, with the distance between the leading set of points accelerating faster than the distance between the lagging set of points. The molecule is moving in the direction of flow. This is the type of extensional strain found on the left side of Fig. 2 below between A and B objects, as well as between B and C.

Fluid flow regions with a dominant extensional stress component can induce molecular conformational transitions that are rarely, if ever, allowed under quiescent laboratory solution conditions. The type of transition depends on the amount and directions of kinetic energy transferred to the molecule from CSF kinetic flow energy. Increased chemical energy induced by flow stress is represented in this and previous papers of this author [10–15] by an asterisk (*), e.g., flow stress-excited amyloid beta is Aβ* in a low shear flow-stressed excited state.

Experiments reported in a paper by Dobson et al. [36] demonstrate that many dissolved proteins that do not have a tendency to self-aggregate under quiescent conditions aggregate to ultimately form fibrils when repeatedly exposed to extensional flow stresses. The experiment consisted of two syringes connected by a narrow bore capillary. The solution was passed through this capillary back and forth many times to induce this aggregation. Extensional flow is created when the fluid flow path is abruptly narrowed, thus increasingly accelerating the fluid flow as it approaches the entrance to the narrowed flow path. It is primarily in this small, suddenly accelerated flow region near the entry to the junction of the syringes and the capillary that extensional flow is generated. Thus, for every pass through the high extensional stress flow zone at these junctions, the times of exposure of protein molecules to extensional flow stresses were very short, on the order of tens of milliseconds. Calculations showed that the rate of energy transfer for extensional flow is orders of magnitude faster than for wall-generated laminar shear flows in the capillary in these experiments [36]. Shortening the length of the capillary did not change the results, indicating that the extensional flow-induced change was caused exclusively by extensional flow. The question arises, are there CSF flow situations within the brain that generate similar extensional flows?

The energy source for Aβ and tau conformation changes is CSF flow stress

Shear rates on the order of several hundred sec^–1 are needed to change the Aβ conformation from the highly flexible ensemble IDP state [15] (the “random coil”) to the much more rigid cross-beta non-IDP conformation. Shear experiments using Raman spectra to explore protein backbone alterations have found that similar order of magnitude shear rates reversibly alter quite different protein backbone conformations [43]. These low shear rates contrast with some reported experiments where many orders of magnitude higher shear rates could not denature proteins. However, similar shear experiments in the presence of surfaces have demonstrated that, depending on the nature of the surface, certain surfaces can readily promote protein aggregation at very low shear rates [24, 44]. Hydrophobic and polar surfaces were found to be quite effective in this regard.

Biological membranes have distinct hydrophobic as well as polar character. Cellular membranes line CSF flow channels. CSF flow through confining surfaces generates laminar as well as extensional flow depending on the complex geometries of these channels. The highest shear rate in pure laminar flow is at the confining wall [10]. Shear-stressed CSF solute molecules such as Aβ will likely bond, at least temporarily, to the wall if their surface intermolecular bonding potential is high. Flowing Aβ will be increasingly shear-stressed as it approaches the confining wall and may stay stressed during and after its wall adsorption. When flow rates slow or reverse direction, there may be a tendency for the stressed molecule to return to at least some intramolecular Aβ structure, unless there is very strong intermolecular surface bonding. The latter could be the case where there is very high flow shear exposing long lengths of the Aβ protein molecule to the membrane surface. The adsorbed Aβ can migrate on a protein island membrane surface and be stabilized and form barrel-shaped complexes [35]. These complexes can form channels that leak calcium ions and thus inactivate neuron and astrocyte cells.

However, if there is weak intramolecular bonding, the molecule could regain partial structure during periods of diminished shear and return to the flowing CSF stream. Thus, the membrane adsorption process would be quite sensitive to flow rate.

Aβ and tau AD pathology and major flow stress are in the same lower brain regions

The earliest Aβ pathology is located near a highly complex region of the glymphatic flow pathway surrounding the basal cistern through which Aβ- and tau-containing CSF travels (Fig. 8). The critical pathological regions are the basal temporal and orbitofrontal neocortex. This is in the same location as one of the two major regions of strong CSF chaotic flow rates shown in Fig. 5. This path encompasses two relatively large CSF basilar cisterns (interpeduncular and chiasmatic), a major brain vascular complex (the circle of Willis), and neuron axon bundles that connect the brainstem with many parts of the brain cortex. It is also CSF downstream from the intensely chaotic CSF motion shown in Fig. 5 in the pontomedullary cistern. The basal frontal temporal brain region laterally contains the EC. The EC region is the epicenter and origin of tau tangle pathology.

The major vascular components of the circle of Willis are found in the center of Fig. 10, just above the basilar artery in this figure. These are contained within each of the cisterns making up the basal cistern (Fig. 8). They generate strong heart pulse waves [19] especially because of the presence of the major carotid arteries. These waves are major generators of the CSF back-and-forth flow in this CSF-filled cistern. There also may be CSF flow in the narrow space between the two brain hemispheres in the interhemispheric cistern (Fig. 8) that could produce Aβ shear-induced products in the lower frontal and temporal lobes because of the presence of the surface cerebral arteries. This results from the tissue strain (Fig. 6), giving rise to the additional chaotic flow (Fig. 5) in the third ventricle. Such chaotic flows could also give rise to shear-induced activation of Aβ or tau molecules within that ventricle.

Arteries such as the basilar, carotid, and cerebral arteries and their branches (Fig. 10) are covered by the arachnoid membrane within the SAS in the basal part of the brain. The pulse waves from these arteries result in reciprocal CSF pumping action within these glymphatic CSF paths in a number of cisterns. Thus, this cistern region and that surrounding the pons surface containing the major basilar artery show maximum chaotic flow in the MRI studies (Fig. 5) [18]. The chaotic CSF motion observed in this glymphatic region of interest results from intense pressure waves. These waves emanating from major arteries have been reported in MRI studies [19]. Thus, the same type of chaotic CSF motion observed in the pontomedullary cistern (Fig. 8) from the pulsation of the major basilar artery is apparently generated by pulsations of the circle of Willis arteries. The pontomedullary cistern, which is upstream from the basilar cisterns, delivers a heavy dose of chaotic energy to cistern amyloid monomer molecules. Could there already possibly be flow-stress altered Aβ and tau conformations formed even before their next, perhaps even greater, dose of chaotic motion in the basil cisterns?

Aβ aggregation is initiated by cardiac pulse-driven, CSF laminar and extensional stress

One possible molecular consequence of all this chaotic energy exposure is that many small, random conformational changes are induced by chaotic CSF flow surroundings. Since Aβ and tau are IDP molecules, each of these changes could cause small changes in the composition of their IDP conformation ensembles [15]. In each newly altered ensemble, one or more of the IDP conformations may exceed the activation energy and entropy requirements for the ultimate formation of a cross-beta conformation that is necessary for Aβ or tau monomer aggregation. Once the transition to a cross-beta conformation is completed, the Aβ* or tau* protein is now quite prone to aggregate with other monomers. If it is a seed molecule, it will convert other Aβ molecules to another cross-beta conformation and likely bond with the seed molecule, initiating a self-propagating aggregation reaction.

Thus, it is proposed that the initiation of Aβ pathology in the basil cistern is due to the molecular distortions of Aβ caused by the chaotic motions revealed in Fig. 5. The chemical reactions that form oligomers may take place on the frontal lobe wall membranes on which the distorted Aβ* molecule is adsorbed. If so, the wall acts as a catalyst for the formation of these seeds because of the lower activation energy for dimer and possibly multimer formation, initiating further aggregation to form Aβ oligomers. If these are seed Aβ oligomers, they can then propagate through the frontal lobe membrane and then through the rest of the cortex and follow the numerous neural paths to critical locations involved in memory formation and storage as suggested by Chen and Mobley [17].

Laboratory flow experiments have reported that Aβ aggregation is both initiated [27] and accelerated [28] by shear. However, the membrane surfaces that generate shear stresses are likely to play a critical role in Aβ protein aggregation [10]. Because back-and-forth CSF flow is observed both around cortex pial arteries [45, 46] and through brain ventricles [39], it is assumed that the same kind of directional oscillation of CSF motion is taking place in other regions of the glymphatic path, e.g., in the cisterns. Because of this CSF flow reciprocation, solutes such as Aβ and tau have ample time to diffuse to and interact with the wall under these flow conditions. Our experiments [14] have shown that when Aβ interacts under steady laminar flow conditions with a stainless steel wall, it both reversibly and irreversibly aggregates, depending on both flow rates and repeated exposure of solutes to the walls already containing previously adsorbed Aβ. Experiments with proteins other than Aβ have shown that, under shear stress, many adsorb to surfaces in several layers, one of these being a surface layer, but that flow stress can remove outer layers as microparticles [24]. Could this be a possible mechanism for the initial formation of Aβ plaque?

Extensional flow-stressed molecules are especially vulnerable to conformational change

As indicated in Fig. 1, although one component of shear experienced in laminar flows is extensional (elongational), leading to stretching of CSF solute molecules, there is also a rotational tumbling component. When there is free rotation in bulk CSF undergoing laminar flow, this causes a cyclic reversal of the stretching motion. At or near the wall, such rotation can force more intimate contact of the rotating, stretching molecule with the wall, providing increased opportunity for adsorption. However, if CSF is in a part of the fluid flow field resulting in primarily extensional flow with a minimal rotational component, there is primarily molecular stretching. That is, until there is a reversal in reciprocating bulk CSF flow direction, at which point the stretching is released. Thus, there is a time window in which the stretched molecule can reorganize into a different conformation. This could represent one possible mechanism for the conversion of the IDP random coil to cross-beta Aβ or tau conformation.

In free shear flow, during the initial stretch and rotation part of the cycle of a solute, a stretched molecule may not have an opportunity to react with a nearby stretched molecule to form a stretched dimer because its stretch has been relaxed by the rotation. However, at higher solute concentrations, if there is mainly extensional stress for a long enough period with little shear rotational contribution, there is a greater chance for two elongated molecules to meet and form a stretched dimer or a conformationally altered monomer (Fig. 11).

Fig. 11

Extensional flow-induced transformation from an alpha to beta conformation via intramolecular and intermolecular mechanisms: (a) Two nearby alpha helical segments within the same molecule are exposed to extensional flow to such a point where their protein backbones are exposed to each other resulting in an alpha to beta conformation change within a single molecule; (b) When the molecular concentration is high enough, a concentration dependent-bimolecular collision between extensional flow-stretched molecules can produce a dimer that is much more stable than the monomer; (c) However, if the strain-induced stretch is not sufficient to allow close proximity of each alpha helix in a single molecule, then there is a lack of intramolecular product formation, resulting in heat release.

If high amounts of elongational energy are very rapidly delivered, extensively stretching the molecule quickly, there would be an opportunity to rapidly form a conformationally high energy transformed monomer (Fig. 11a or Fig. 12). The latter could be a monomer precursor that ultimately produces a seed oligomer.

Fig. 12

Proposed mechanism for the rapid strain-induced transformation of Aβ from an IDP alpha conformation to a beta hairpin conformation through extensional stretching distortion. This particular alpha state was selected from favored IDP states in [60]. This beta sheet structured molecule may then react quickly with another and possibly form a dimer seed. This aggregation reaction could take place over many stretch-and-release cycles.

Experiments with pulsating and reciprocating flows with varying energies are needed to identify possible fluid flow-induced Aβ and tau chemical products. Such experiments may require different flow geometries that mimic those found within the glymphatic system. Since Aβ, tau, and phosphorylated tau are all found in CSF, it will be important to also test the effects of shear- and extensional flow-rates on mixtures of these molecules.

With Aβ-tau mixtures, there is also a possibility for the formation of a complex between shear- or extensional flow-induced tau and one or more flow-activated Aβ molecules that is a catalytic intermediate prior to the formation of a cross-beta tau conformer, a precursor to a tau oligomer or tau seed. Figures 11, 12 demonstrate possible extensional flow effects on molecules containing alpha helical segments that are flow stress-induced to produce potential Aβ seeds, seed precursors, or Aβ dimers.

It is claimed by Dunstan et al. [27] that: “The change in time scales for amyloid (Aβ) formation from sheared to un-sheared may be used to interpret the magnitude of energy change that is produced by shear . . .  Assuming a single activated process, the ratio of these times gives us the difference in energy between the un-sheared and sheared samples, which is∼4.3 kT, or 1 weak hydrogen bond per entity in solution. Thus, we are studying a very weak perturbation on the protein structure.” The author has suggested [12] that there is a concerted shear-induced conformational change, similar to that shown in Figs. 11a or 12, involving the whole molecule that involves a relatively low energy, shear-induced transition state. This low energy event could possibly involve quantum tunneling in hydrogen bond exchanges [47]. That is, H atoms may tunnel through potential energy barriers at much lower energies than would be needed to actually break current H bonds and make new ones. The further suggestion is made here that it is not necessary to unfold the amyloid monomer in a reverse path of the protein folding process in order to convert it into a cross-beta structure. Instead, shear or extensional forces spread over the entire molecule such as that shown in Fig. 1 may require less energy because there are weak bonds being simultaneously broken at the same time as new ones are being formed as depicted in the transition states suggested in Figs. 11, 12.

Flow stress events in and around CA initiate tau pathology in the locus coeruleus (LC)

As mentioned previously by the author in a discussion of flow stress in the glymphatic path [15], the CA (Fig. 13) is the narrowest, most flow-restricting parts of the ventricular CSF path. The LC is a large bundle of neuron axon projections located within the CA, near its lower exit into the fourth ventricle, and which continues in two locations on or near the CA wall surface into the fourth ventricle. The LC is the principal site for brain synthesis of norepinephrine, which is a hormone and neurotransmitter that increases arousal and alertness, promotes vigilance, enhances formation and retrieval of memory and focuses attention. The LC nuclei send this hormone signal throughout the entire cerebral cortex as well as to a variety of other structures including the amygdala, hippocampus, cerebellum, and spinal cord. Thus, these LC neuron projections are connected to many different regions of the brain that deal with memory, arousal, sleep-wake cycle, attention, behavioral flexibility, behavioral inhibition, psychological stress, cognitive control, and emotions. These are all reflected in AD symptoms.

Fig. 13

Views of the brain ventricles, showing the narrowing of CSF motion channels in flowing from the third ventricle (“3V”) to the fourth ventricle (“4V”), with increased CSF extensional stress rates as the flow channel narrows in the CA (cerebral aqueduct; “Aq” in this illustration) and creating zones of dominant extensional stress. Source: [39] Licensed under Creative Commons CC-BY license.

The LC is also reported to reveal the earliest pre-tangle [48, 49] phosphorylated tau (p-tau) AD pathology. This p-tau was found in the proximal axons of the LC, leading to speculation that p-tau was an even earlier pathological marker than tau tangle AD pathology in the EC. Aggregated tau tangles were observed first in the EC, but the locus of the earliest tau pathology is closer to that of Aβ.

MRI observations [19] demonstrate that the reciprocal back-and-forth CSF motion in the CA is not symmetrical. A well-regulated laminar flow was observed in this aqueduct, and a high-velocity downward motion was noted after a slow upward motion. This bidirectional CSF motion was seen in all healthy individuals.

Extensional fluid stress is felt by Aβ and tau molecules when CSF in the third and fourth brain ventricles approaches its top and bottom CA entrances. CSF is cycling back-and-forth with a net flow direction through the CA from third to fourth ventricle. However, a persistent vortex is reported near the entrance to the CA within the third ventricle (see Fig. 7 and [39] for details). One must remember that molecules dissolved in this fluid have already been exposed to highly chaotic flow within the third ventricle (Fig. 5). Thus, Aβ and tau molecules and possibly their oligomers dissolved in the CSF that are transiting through the CA are subjected multiple times to both chaotic flow and extensional flow at junctions of the CA within the third and fourth ventricles, and laminar flow within the CA. It is suggested here that, under these energy rich conditions, such flow-stressed molecules may have an opportunity to collide with another similarly stressed molecule to form a more stable dimer, possibly leading to higher neurotoxic oligomers. However, it is also possible that neurotoxic oligomers can form on the LC-containing walls that have adsorbed flow-stressed amyloid monomers of Aβ and tau. These molecules can form oligomers on this surface that can interact with the LC causing neural damage.

In flow experiments reported by Dobson et al. [36], it was suggested that the passage of protein solutions back and forth through the abrupt flow path dimensional changes was responsible for incremental changes in protein conformations that finally produced one that was responsible for initiating rapid protein aggregation. The aggregation process was protein concentration dependent, but it was also protein-type dependent. For example, how flexible is the protein when it is flow-stressed?

The biological surroundings of the many different types of glymphatic environments will also affect the conformational makeup of IDP ensembles. For example, high extensional flow regions will probably favor stretched ensemble conformations that rarely exist in quiescent solution.

Oligomers extracted from AD patients have different characteristics than oligomers of the same composition synthesized in the laboratory. It is suggested here that the reason for this is that human brain-derived oligomers were formed under unusual brain fluid flow conditions described above and thus have different fluid flow-stressed conformations during and after their formation. Therefore, it is suggested that laboratory experiments be conducted under these same flow stress conditions and oligomers be compared with those synthesized under quiescent conditions.

Pulse energy is a critical pathological variable

Low bulk ISF flow rates in the upper cortex, if bulk flow indeed exists [50], are the result of the very high density of parenchymal neurons, neuroglia, basal lamina, and the very small extracellular spaces that all serve to constrict ISF fluid flow channels. Blood flow intensity as well as CSF and ISF intensities generally decrease with distance along the vascular tree connections from the heart to the brain [23]. However, low flow rates do not necessarily guarantee low flow stress rates, since there is a nonlinear inverse relationship between flow path width dimensions and shear rate. However, there is still little evidence of high energy chaotic flow in the cerebral cortex (Fig. 5).

It would appear that there are only low energy flow stress events in the cortex parenchyma that may first produce a mechanically distorted Aβ* molecule that initiates the formation of a relatively mild neurotoxic oligomers [10]. This low energy is generated by slower cortex parenchymal fluid flow processes than those found in the lower brain and brainstem region. Higher energy CSF motion in the lower brain region causes more energetic fluid stress-induced molecular distortion of Aβ and tau, producing higher energy dissolved Aβ** and possibly tau** molecules. Such energetic distortion could possibly break or at least weaken hydrogen bonds that stabilize whatever characteristic protein structure that these IDP molecules might temporarily contain. It is expected that this chaotic higher energy can create flow stress-induced products containing longer, comparatively unstructured stretches of amino acid chains (Fig. 14). During flow reversal periods, such long, highly activated molecules could refold into different prion-like oligomer seeds containing cross-beta structure. Such potential seed structures may only be available initially in the lower brain regions where higher flow stress energies are available. These seeds could then produce the types of oligomers that are the keys to seed formation suggested in the Chen-Mobley synthesis for producing tau pathology [17].

Fig. 14

Extensional flow-induced intramolecular processes during CSF passage through the cerebral aqueduct. Column A: low-energy cycle with no conformation-altered monomer. Column B: High-energy toxic monomer seed formed that can react with locus coeruleus. Modified from [39]. Licensed under Creative Commons CC-BY.

Cerebral aqueduct MRIs might possibly be amyloid biomarkers

As indicated in Fig. 15, MRI is a means of determining in vivo CA dimensions and LC locations under full CSF motion. Because of the extensive flow-induced Aβ wall deposits on the walls of laboratory capillaries with roughly similar dimensions and flow rates as those measured in CA structures [14], the contrast in CA cross sections noted in Fig. 16 were of great interest. If we describe the CA cross section as a crown with a flat floor, the CA floor of the Parkinson’s disease patient in Fig. 16 looks as though it has been filled in. This considerably narrows the flow dimensions for CSF, increasing the CSF flow resistance in the CA in addition to also greatly increasing the shear rates experienced by any Aβ and tau molecules dissolved in the oscillating CSF within the CA of the Parkinson’s disease patient. The reason for this narrowing of the CA dimensions could well be the result of flow stress-induced aggregation processes (Fig. 17). Further investigations are needed involving longitudinal MRI studies to see whether this effect is correlated with disease progression. If so, laboratory flow experiments should be conducted to explore a possible mechanism for the buildup and explore whether MRI measurements in the CA and other ventricle and cistern flow stress hot spots can act as amyloid disease biomarkers.

Fig. 15

Visualization of human LC as an MRI hyperintense signal in the brainstem using T1-weighted MRI optimized to enhance neuromelanin LC content (white arrows), seen in axial (B) and coronal (C) planes acquired as shown in (A). Source [51] with permission.

Fig. 16

Contrast in LC (white arrow) environment in cerebral aqueduct in (A) a healthy patient; (B) Parkinson’s disease patient with early symptoms.Source [52] with permission.

Fig. 17

Cross section of the cerebral aqueduct, which is surrounded above by the pons and that has a floor of the cerebellum. In addition to laminar shear flow, there may be secondary flow, indicated by the arrows causing flow toward the floor of flow-stressed molecules that aggregate on the floor.

A flow-stress energy continuum is responsible for a preclinical progression continuum to MCI

It is now known that decades before the first clinical AD symptoms appear, there are other preclinical and mild cognitive impairment (MCI) AD symptoms that can be observed, measured, and classified [53]. Among these are accumulation within CSF and ISF of Aβ in the form of soluble Aβ and tau monomers and oligomers, synaptic disfunction, tau mediated neuronal injury, and the initiation of the reduction of brain volume. MCI symptoms are codified indicators of mild cognitive dysfunction. It is within these first two periods in Fig. 18 that the author wishes to focus in an attempt to identify flow-stress events, locations, and mechanisms of initiating AD.

Fig. 18

Deterioration of cognitive function in AD with time correlated with postulated corresponding increases in flow-stress energized Aβ and tau molecules contained in brain CSF. Both of these are postulated to be smooth transitions. The dashed line is that for AD patients and the solid line for healthy individuals. Asterisks indicate the degree of flow-stress energy and aggregation in the amyloid monomers. Note that this diagram represents a hypothetical model for the pathological-clinical continuum of AD but does not imply that all individuals with biomarker evidence of AD-pathophysiological process will progress to the clinical phases of the illness. Aβ42* represents cortex plaque formation. Aβ40* represents low energy oligomers and CAA formation. Aβ** represents higher energy oligomer seed precursors. Tau** represents higher energy tau oligomer seed precursors. Figure modified (in red) from [53] with permission.

There are smooth, continuous transitions between three stages: preclinical, MCI, and AD dementia, as shown in Fig. 18 [53]. The symbols in red (Aβ*, Aβ**, and Tau**) represent the steadily increasing degree of flow-stress energy within the cortex and lower brain region, with resulting aggregation of Aβ and tau monomers, at various points in the AD timeline. At lower energies, Aβ₄₂ is readily aggregated in the form of senile plaque in the cortex. At higher energies Aβ₄₀, which escapes the cortex, is aggregated as perivascular CAA [10]. Note that proposed high amounts of stress energy are needed to induce tau to aggregate, whereas Aβ aggregates much more readily at lower energies and at earlier times. There would appear to be a higher energy Aβ** oligomer seed that spreads prion-like throughout the cortex, possibly formed within or on the walls of the basal cistern. Tau** apparently requires a more complex flow environment.

It is reported [54] that Aβ induces phosphorylation of tau (p-tau) that is released into ISF and CSF. p-tau is both water soluble and an IDP molecule. As such, it should be a relatively flexible molecule but probably with more alpha helical and possibly beta structure in the various members of its IDP p-tau conformational ensemble than that found in the Aβ ensemble.

Tau is a much larger IDP than Aβ and thus is a more complex molecule from a conformational change point of view. Flow-stressed conformational changes involving tau have more intricate energy and entropy demands. Increasing energy extensional flow should stretch tau into increasingly longer, relatively-linear ensemble conformations. This energy is probably available from chaotic processes in the lower brain regions. However, the conformational changes required to convert tau into a cross-beta seed conformation may require a number of sequential and focused molecular perturbation processes. These may only arise from cyclic extensional strain events initiated within the unique geometry of the combination of the third ventricle, CA, and fourth ventricle to accomplish the conversion of p-tau into a cross-beta conformation capable of ultimately forming a tau oligomer seed.

The smooth and continuous nature of these increasing energy transfer processes, according to this hypothesis, is made possible by the smooth and continuous hardening of the carotid and aorta arteries that produce smooth and continuous increases in systolic CSF pulse energies in the lower brain regions. It is acknowledged that some of these curves and data in Fig. 18 are hypothetical and oversimplify a much more complex situation, but nevertheless may offer a useful overall framework that is able to support the more complex theory outlined above.

Cystolic CSF spikes arising from hardened major arteries are the missing link to the origin of AD

Chaotic flow regions in the lower brain, cisterns, and around the brainstem are present in healthy adults, as indicated in Fig. 5. Thus, chaotic lower brain CSF flow per se does not cause AD. What, then, are the conditions that trigger AD and possibly other amyloid diseases? This paper suggests that the reason is, as the AD patient ages, the aorta and carotid arteries harden and lose their wall flexibility that helps even out sharp blood pressure pulses emanating from the heart [55–57]. As these arteries harden, the heart must work harder to make up for resulting lower total oxygen delivered to the brain. Thus, with increasing age, these hardened arteries pass along a sharper pulse intensity of the heart systole pressure pulse further down the vascular tree, shown symbolically in Fig. 19. Thus, the pulse becomes increasingly destructive, ultimately causing brain silent microbleeds in brain capillaries [55].

Fig. 19

Vascular pulsing that is primarily responsible for CSF flow –the effects of aging and artery hardening. The aorta and carotid harden and lose their flexibility that is responsible for steady capillary flow (upper figure). But artery hardening causes intense heart pulses to pass through these two major arteries and increases the pulse intensity motion in the circle of Willis, the minor arteries, and capillaries, causing microbleeds [55]. Thus, the increased pulse intensity in the circle of Willis with age should be responsible for increased chaotic CSF flow in the neighborhood of the circle of Willis, namely in the basal cistern and its direct CSF connections. This increased motion could lead to higher energy Aβ and tau molecular states that are precursors to neurotoxic seeds, thus possibly initiating Aβ seeds leading to AD.

It is assumed that perivascular CSF motion is generated in the cisterns and CA by these same heart pressure pulse waves. If so, as one ages, there should also be increasing energy pulsed CSF waves and therefore increasing chaotic flow in the lower brain region. In turn, these waves should cause increased CSF flow-induced Aβ and tau molecule distortion. Such an increase in CSF pulse energy could reach a critical energy level such that it exceeds the activation energy for conversion of the comparatively large, complex tau molecule into a conformation that subsequently converts other tau molecules into tau tangles in the EC.

Could age-elevated CSF pulse intensity initiate both Aβ and tau AD pathology elsewhere?

During the history of the search for the AD’s origins, many have noted parallels between cognitive decline and vascular dysfunction [26, 56]. “In humans, a chronic increase in pulse pressure is associated with cognitive decline and dementia. More specifically, increased PWV (pulse wave velocity) is associated with memory loss, MCI, cognitive decline, and dementia” [56]. MCI patients have been shown to have the sharpest, most intense pulse blood flow when compared with AD patients and healthy controls [58]. Could age-related excess blood and CSF pulse energy be the critical AD-vascular connection? That is, if vascular degeneration of the aorta and carotid artery flexibility lead to increased systolic pulse pressure and PWV, does this also cause increased CSF chaotic flow? These highly directional kinetic energy CSF motions could overcome activation energy barriers, allowing conformational transformations of each monomer that lead to the formation of a catalytic Aβ-tau** transient, possibly on a CA surface. This could be followed by the formation and release of either a tau** seed or an amyloid-beta tau** oligomer seed from this surface that can spread in a prion-like fashion throughout the brain to initiate AD symptoms.

The locations of CSF chaotic motion hot spots are the same as Aβ and tau pathology

Aβ oligomer seeds can spread from the medial temporal and orbitofrontal neocortex throughout the anterior cortex. The Chen-Mobley synthesis [17] suggests that these oligomers migrate through neuronal circuits from the cortex to regions near the EC, LC, and the hippocampus, all of which are critical to functioning memory. However, Aβ as well as tau oligomers could also be formed within the CA and third and fourth ventricle. It is here in a highly localized pons location that tau seeds are apparently initially formed, not in the upper cortex.

Proposed mechanism for the formation of fluid stress-induced tau seeds

The above mechanism still neglects one critical prerequisite, the need for both Aβ and tau to somehow be involved in the critical step or steps leading to initiation of AD tau pathology. Why are the origins of Aβ and tau seeds located in different parts of the brain? It is suggested that Aβ seed is formed by the lower-energy chaotic motions within the Circle of Willis region because of the pulse-induced CSF motions in the MCI phase toward the beginning of the major artery hardening process. The larger tau molecule is harder to convert into a seed oligomer. It takes some combination of a more-focused, more highly energetic third ventricle and, in particular, a CA extensional flow pattern to convert tau into the necessary cross-beta conformation needed to form a neurotoxic tau seed. The energy needed for this tau conversion could be achieved at a later time in the artery hardening process when systolic pulse energy reaches a high critical energy level. El Sankeri et al. show [58] that a maximum systolic pulse flow is achieved in MCI patients when compared with healthy controls and AD patients.