Prions are the black holes of the neurodegeneration field—lethal, ubiquitous, and paradigm-challenging. Although these were originally discovered and characterized by Prusiner and his laboratory group as disease-causing agents 4, Jarosz and colleagues have described multiple salutary effects such as response to stress and non-Mendelian inheritance of protective traits 24, while discovering dozens of new prions in yeast. These protective traits could be passed from generation to generation via protein in a non-Mendelian fashion, and thus prions are also mediators of protein-inherited traits.

While some of these newly discovered prions form amyloid (as do amyloid-β, p-tau, and α-synuclein), others are DNA-interacting proteins (as is TDP-43). These proteins have the ability to spread protective properties from cell to cell and organism to organism. Just as bacteria may spread adaptive traits genetically via plasmids, prions may represent a mechanism for both horizontal and vertical, non-genetic transfer of adaptive traits. It is of more than passing interest that both neuroplasticity and prions are drivers of organismal evolution and adaptation, since it has been demonstrated that prions play a role in learning and memory, specifically in persistence of memory 25.

These findings expand the repertoire of prions far beyond causing disease, and in keeping with the Pr2 theory, shift focus from pathology to physiology. The prions associated with neurodegeneration function as anti-infectious agents, and the distribution of microbes against which each prion exhibits efficacy should help to reveal the upstream contributors to neurodegeneration for each individual. For example, as Soscia et al. reported, amyloid-β peptides exert antimicrobial effects against Streptococcus pneumoniae, Listeria monocytogenes, and the virus Herpes simplex, among others 14. Therefore, in those with evidence of amyloid-β deposition, these organisms should be sought and, if revealed, treated. Furthermore, these prions also possess metal-binding properties, and the combination of metal binding and antimicrobial effects make prions strong candidates to function as anti-biofilm agents.

Beyond their antimicrobial effects, prions serve as physiological response amplifiers of signals: classic homeostatic signaling occurs when single-target outcomes are the goal, such as a serum pH of 7.4, and amplification is not required; in contrast, prionic loop signaling occurs when switching is required in a bistable system such as blood clotting or neurite extension/retraction. APP signaling represents an example, in which two different outcomes—the α and β pathways—are mutually inhibitory 23.

Here it is proposed that prions also play a crucial role in brain protection by functioning as pre-inflammatory immune mediators (“priion” effect of prions), internal barriers analogous to the antimicrobial barrier substance cerumen, and that this function explains the decades-long presymptomatic stage that may occur in neurodegenerative diseases such as Alzheimer’s. Thus many cognitively normal adults are amyloid positive by PET or CSF 26, and while many of these avoid symptomatic Alzheimer’s disease throughout their lifetimes, there is nonetheless an increase in risk for cognitive decline in those who are amyloid positive 27. These findings are compatible with the Pr2 theory, in which prions may isolate and destroy pathogens without creating a pro-inflammatory, cognition-compromising state until such time as an inflammatory, degenerative state supervenes.

The conversion of the pre-inflammatory protective state afforded by prions to a pro-inflammatory neurodegenerative, symptomatic state depends on factors that include pro-inflammatory gene products such as ApoE4, continued exposure to inflammagens, a failure to contain pathogens with the prions, the concentration of anti-inflammatory balancers such as sAPPα and PrP^C, microglial activation, mast cell activation, astrocytic activation, and pro-inflammatory cytokine signaling, among others.

Antagonistic pleiotropy, a concept first developed in 1957 by Williams 28 following earlier suggestions by Medawar, argues that genes related to early-life fitness are selected at the cost of later-life decline. This concept is now the most well-accepted theory of the evolution of aging, and it dictates that performance be selected over durability. Major evolutionary needs include defense/aggression/predation (i.e., lethal force), and learning/adaptation/plasticity (i.e., optimization of behavior for a given set of circumstances); in other words, optimal movement and adaptability for survival and procreation, with neuroplasticity playing a role as organismal evolution in real time.

This has important consequences for highly evolved human neural subnetworks, implying that, since we have evolved to outthink and outmaneuver our competition, we shall be susceptible to neurodegeneration. Indeed, the vulnerabilities of each network are revealed by the drivers of the associated neurodegenerative process. This is perhaps most obvious for Parkinson’s disease, in which mitochondrial function, and specifically complex I (NADH:ubiquinone oxidoreductase), is compromised by the many different agents that increase risk for Parkinson’s disease 2930. This implies that the motor modulatory neural network that is insufficiently supported in Parkinson’s disease is rate-limited by mitochondrial complex I. However, it is clear that other functional members of this network may also play a role in support and disease-associated dysfunction, such as ubiquitin-mediated protein degradation, dopamine synthesis and neurotransmission, and lysosomal function, among others.

In the case of ALS, the Achilles heel of a system that provides several orders of magnitude of nearly instantaneous power amplification, from thought to maximal action, appears to be the removal of the excitotoxic neurotransmitter glutamate from the associated synaptic clefts 31. Compromise of this function by transporter mutations, toxins, reactive oxygen species, or by other mechanisms; or repeated excessive demand on this system, is associated with increased risk for ALS.

In the case of age-associated macular degeneration (AMD), the Achilles heel of a system that is highly demanding energetically, and allows humans to distinguish approximately one million different colors and shades, involves the energetic support of the retinal pigment epithelial cells (RPE) and photoreceptor cells, as well as the complement that is activated when the physiological RPE phagocytosis of the photoreceptor outer segments (POS) begins to fail to keep up with demand.

In the case of Alzheimer’s disease, perhaps in keeping with its high incidence, there are multiple paths that lead to the associated neuroplasticity network insufficiency: insulin resistance, sleep apnea, chronic infections such as Herpes family viruses, and air pollution, among many others, with the list of risk factors suggesting that the Achilles heel for this neuroplasticity network involves inflammation and energetic support.

The model proposes that each neural subnetwork possesses its own activity-dependent demand for energetic, trophic, and transmitter support, a demand that is augmented by inflammation, some toxins and toxicants, and stress. Chronic or repeated failure to provide the supply required for each network’s demand leads to degeneration, which may be programmatic (analogous to programmed cell death) or non-programmatic. Thus, for Alzheimer’s disease, the probability of developing AD is proportional to the integration, over time, of the sum of the many factors contributing to synaptoclastic signaling (such as various chronic infections), divided by the sum of the many factors contributing to synaptoblastic signaling (such as nutrients and trophic support). This probability, in turn, may be approximated by considering the six major groups of drivers noted above:

The constants associated with each of the inflammatory mechanisms, toxins, stressors, energetics, trophic species, and neurotransmitters should be calculable from patient data and artificial intelligence (AI), enhancing both prognostics and therapeutic outcomes. However, the various constants (k_i through k_n) will only be applicable to a given genome; nonetheless, it may be feasible to estimate these by grouping data into four different genetic backgrounds, and calculating approximate constants within each of the four groups: (1) ApoE4-negative, FAD-mutation-negative; (2) ApoE4 heterozygotes; (3) ApoE4 homozygotes; (4) familial Alzheimer’s disease mutations (APP, PS1, or PS2 mutations associated with familial Alzheimer’s disease).

The theory proposes that each of the major neurodegenerative diseases represents a network insufficiency driven by various combinations of the six groups of contributors shown in the equation above, and which degenerative disease results is dependent on three variables: (1) insult(s) (such as insulin resistance, specific infections, etc.); (2) insult entry point(s) (olfactory system, blood-borne, epidermal, etc.); and (3) prionic response(s) (amyloid-b, p-tau, a-synuclein, TDP-43, etc.). Thus an insufficiency due to metabolic syndrome, and HSV-1 entering via the trigeminal nerve, in the setting of enhanced inflammation due to ApoE4, with amyloid-b prionic response, would create Alzheimer’s pathophysiology, whereas an insufficiency due to mitochondrial complex I inhibition caused by systemic exposure to paraquat in the setting of compromised systemic detoxification due to a null mutation in glutathione-S-transferase M1, with an a-synuclein prionic response, would create Parkinson’s pathophysiology.

Energetic reduction is a very common contributor to Alzheimer’s disease, and likely to other neurodegenerative diseases as well, such as Parkinson’s (and related extrapyramidal diseases such as progressive supranuclear palsy and corticobasal degeneration) 32, Lewy body disease, and age-related macular degeneration. This reduction may stem from vascular compromise, oxygen reduction (e.g., due to sleep apnea), mitochondrial damage (e.g., from accumulated mitochondrial DNA mutations, reduced mitophagy, toxin exposure, or other causes), or metabolic dysfunction with insulin resistance and reduced ability to generate ketones.

Neuroinflammation has been well documented in Alzheimer’s and other neurodegenerative diseases, and microbiome studies of the brain have revealed oral flora as well as rhinosinal 33, somewhat surprisingly both in brains without neurodegeneration and in those with neurodegeneration. The causes of neuroinflammation range from chronic infections such as Herpes simplex or Porphyromonas gingivalis to dysbiosis (e.g., alterations in the oral microbiome) to gastrointestinal hyperpermeability to metabolic syndrome to biotoxins, as well as other causes.

Toxin or toxicant exposures may be inorganics (such as air pollution 34), organics (such as anesthetic agents), or biotoxins (such as trichothecenes) 3536.

Reduction in trophic support may involve reduced neurotrophins such as BDNF or NGF, reduced trophic hormones (or hormone ratios) such as estradiol or testosterone or thyroid 37, or reduced nutrients such as vitamin D or vitamin B12. Work from Cattaneo and his colleagues has shown that simply reducing the NGF levels in transgenic mice by about 40%, using a germline genomic insertion encoding a fragment of anti-NGF antibody, was sufficient to create Alzheimer/Fischer type pathology 38.

Reduced neurotransmitter support in Alzheimer’s disease involves multiple neurotransmitters, such as acetylcholine, glutamate, serotonin, somatostatin, norepinephrine, and GABA.

Increased stress, with increased cortisol levels, is associated with increased risk for Alzheimer’s 39. Furthermore, antagonism at the corticotropin releasing hormone receptor has been shown to reduce Alzheimer pathology and symptoms in a pre-clinical model 40.

These six major determinants of the synaptoclastic:synaptoblastic ratio determine the rapidity of functional, and ultimately structural, synapse loss, and thus the time to cognitive compromise (Figure 1).

1) As noted in the figure, the curve associated with patient 1 (1) begins with a greater number of functional synapses than for (2), in keeping with observations that those with higher educational levels are at reduced risk for Alzheimer’s 41.

2) In contrast, the curve associated with patient 2 (2) begins with fewer functional synapses.

3) The downward slope of remaining functional synapses over time is proportional to the synaptoclastic:synaptoblastic ratio at any given time. Anatomic synapse loss and neuronal loss lag behind functional synapse loss, and thus reversal of cognitive decline is more readily accomplished earlier in the course.

4) With enough synapse loss over time, symptoms appear, and initially the patient develops subjective cognitive decline (SCI), thus remaining functional, capable of independent living, and capable of scoring in the normal range on cognitive assessments.

5) This depicts the age at death, indicating that patient 1 died without enough functional synapse loss to develop mild cognitive impairment (MCI) or dementia.

6) In contrast, the timeline for patient 2 indicates that the initial number of functional synapses is fewer than for patient 1, and (6) depicts the finding that cognitive loss, as reflected by the functional synapse number, is not smooth, but rather displays increases and decreases as the variables in the equations above change, e.g., due to sleep apnea treatment or improved sleep or reduced stress or treatment of chronic infections, etc.

7) With enough functional synapse loss, the patient develops MCI. Spontaneous return to SCI (or possibly normal cognition) has been documented in some patients (5-10%). However, about 5-10% each year progress to dementia. The acceleration in the curve reflects the combination of the increasing insufficiency of support and the underlying prionic amplification (e.g., Aβ begets more Aβ by inhibiting APP cleavage by a-secretase, among other mechanisms).

8) The repeated observation that spontaneous reversals of MCI may occur 42, but only very rarely (if at all) spontaneous reversals of dementia 43, suggests that a prionic point of no spontaneous return occurs near the MCI-dementia transition point, i.e., that prionic amplification dominates the progression from that point forward.

9) With continued functional synapse loss, MCI progresses to dementia.

The network insufficiency that leads to synaptic loss and Alzheimer’s symptoms manifests in multiple different ways, but clinical presentations fall broadly into two groups: amnestic and non-amnestic. These distinct presentations are associated with different mechanistic drivers, neuroanatomical regions, genetic predispositions, and other parameters, summarized in Table 2.

The physicist and Nobelist Richard Feynman said, “Nature uses only the longest threads to weave her patterns, so that each piece of her fabric reveals the organization of the entire tapestry.” This insight is applicable to the many published observations in Alzheimer’s disease (Table 3): at the molecular level, amyloid precursor protein (APP) signaling comprises two distinct and mutually inhibitory pathways (i.e., a switch). In the presence of trophic support such as the APP ligand netrin-1 2344, APP is cleaved at a single site, the α site, yielding two synaptoblastic and anti-apoptotic peptides, sAPPα and αCTF 45. However, with loss of trophic support, APP is cleaved at three alternative sites, yielding four synaptoclastic and pro-apoptotic peptides, sAPPβ, Aβ, J_casp, and C₃₁464748. These observations indicate that APP is a dependence receptor, i.e., a receptor capable of activating apoptosis when induced by the withdrawal of its trophic ligand 49. Furthermore, these two alternative APP processing pathways are mutually inhibitory: sAPPa inhibits cleavage of APP at the β-site 50, in part by inhibiting BACE proteolytic activity 50, whereas the Aβ peptide inhibits α-site cleavage, thus inhibiting the production of sAPPa and aCTF 23. These antagonistic processing pathways allow APP to function as a molecular switch, one that is potentially sensitive to many different inputs 51, and provide a mechanism by which the Aβ peptide fulfills the criteria for a prion.

At the cellular level, switching between neurite outgrowth and dendritic spine collapse—i.e., synaptoblastic vs. synaptoclastic effects—is mediated in part by APP. At the tissue level, the protection mode is characterized by enhanced tendency to thrombosis, which is lacking in the connection mode. At the organ level, the connection mode features synaptic plasticity that results readily in learning and memory; whereas the protection mode features synaptic loss, neurite retraction, and brain atrophy that are associated with poor cognitive function.

The sum of these effects at the organismal level is to mediate a state switch from connection to protection (Table 3): from synaptoblastic, non-inflammatory, oxidatively metabolic, low-thrombotic tendency, parasympathetic-dominant to synaptoclastic, pro-inflammatory (initially pre-inflammatory due to priions), glycolytically metabolic, pro-thrombotic, sympathetic-dominant. The common genetic risk for Alzheimer’s, the apolipoprotein E e4 allele (ApoE4), triggers this switch more readily by altering cellular programming toward the protection mode, due at least in part to its transcriptionally repressive effects at multiple gene promoters (e.g., SirT1) that include anti-inflammatory genes 52. The systemic communication of this organismal response is presumably mediated by the nervous and immune systems, and may involve trained immunity. Over years, this state switch may lead to a reduction in functional synapses to the point of symptoms, ultimately leading to dementia.

The Pr2 theory explains the observed characteristics of Alzheimer’s disease, many of which are not explained by other theories:

· Why the origin of Alzheimer’s and other neurodegenerative conditions is in antagonistic pleiotropy.

· Why there are many disparate risk factors.

· Why age is the greatest risk factor.

· Why ApoE4 is such an important genetic risk for Alzheimer’s, Lewy body dementia, LATE, and Parkinson’s-associated dementia.

· Why many different infectious agents have been associated with Alzheimer’s disease, such as HSV-1, HHV-6, P. gingivalis, and Cytomegalovirus.

· How these various risk factors contribute to the neurodegenerative process.

· The physiological and pathophysiological roles of prions such as Aβ, a-synuclein, p-tau, and PrP^Sc.

· Why prions are key mediators, present in all of the major neurodegenerative diseases, yet in most patients with neurodegenerative diseases there is no evidence that these infectious agents were transmitted horizontally as an infection, and removing them (at least in the case of Aβ) has not led to sustained improvement in Alzheimer’s disease (and is sometimes associated with worsening of symptoms) 55.

· Why aggregated proteins are present in all of the major neurodegenerative diseases.

· Why Aβ-containing plaques may be present for years with minimal symptomatology 1156, yet over time, those with amyloid are more likely to develop cognitive decline 27.

· Why many die with normal cognition yet display Alzheimer’s pathology 26.

· Why Aβ exhibits multiple salutary properties—such as metal binding, effects as a sequestrant, patching of vascular damage 57, and anti-microbial effects—in addition to its association with neurodegeneration.

· Why monotherapeutics have been so unsuccessful in the treatment of neurodegenerative diseases.

· Why personalized, precision medicine protocols have led to the greatest improvements to date.

· Why systems biological models may be crucial for improved understanding and treatment of neurodegenerative diseases.

· Why previous models, such as misfolded proteins, type III diabetes, and reactive oxygen species damage, have failed to lead to successful treatments.

· Why prevention is likely to be much easier, more efficient, and more effective than reversal.

· Why some patients with Alzheimer’s have been able to achieve sustained improvement for over a decade 58.

· Why higher education levels are associated with lower risk for cognitive decline.

· Why some vaccines lower risk (e.g., Influenza A, Herpes zoster, and BCG), whereas others like COVID may be associated with decline.

The pathological elements of Alzheimer’s disease all fit well within the schema of the Pr2 theory: upstream insults (such as insulin resistance, pathogens, or toxins) impact the equations above, switching signaling from synaptoblastic to synaptoclastic, from connection to protection mode. The response to these insults is analogous to the recognition of PAMPs (pathogen-associated molecular pattens) by the innate immune system, but here represents insult-associated signaling patterns, and the response includes activation of pre-immune mediators, with production of anti-microbial peptides amyloid-b, p-tau, and a-synuclein 59, all of which accrete in Alzheimer’s disease. These are then stored in long-lived, relatively non-toxic and non-inflammatory accumulations—much as one would keep troops in a fort after a battle—thus offering a memory-like component to the pre-inflammatory innate immune response, one rapidly available if insults are repeated. These different anti-microbial peptides are likely to display different profiles of target pathogens and cellular localizations (e.g., extracellular vs. intracellular vs. intraorganellar).

Furthermore, the protective nature of the synaptoclastic response against insults such as pathogens fits very well with the remarkably long—decades long—period of high functionality during ongoing pathophysiology. The presence of pre-inflammatory immune mediators (“priion” effect of prions) such as amyloid-b and a-synuclein allows sequestration and destruction of small numbers of pathogens without generating a function-compromising inflammatory response. The presence of anti-inflammatory balance such as sAPPa and PrP^C helps to prolong the time until inflammatory pathways involving microglial activation, astrocytosis, and cytokine signaling are activated.

Another implication of the theory proposed here regards the many disparate risk factors for Alzheimer’s disease, from early oophorectomy 60 to sleep apnea to metabolic syndrome to hypertension to Herpes simplex to ApoE4 to smoking to air pollution to age to multiple others 61. Previous theories have not explained these seemingly unrelated factors, but all of these fit squarely in the equations above, and all impact the synaptoclastic:synaptoblastic ratio, either via effects on inflammation, toxic burden, energetics, stress, neurotransmitters, or trophic activity. Furthermore, genes associated with Alzheimer’s risk also impact this signaling ratio, either directly via the APP interactome or indirectly via processes such as trophic signaling or inflammatory response or neurite extension (Figure 2, Figure 3). While the theory does not offer new insight into the gender discrimination in Alzheimer’s disease, it is compatible with the notion that women, by virtue of their more rapid decline in trophic sex steroids during menopause than men suffer during andropause, are at increased risk.

Another implication is that, as shown in the equations above, although there are many potential contributors, some patients will have dominant contributors via inflammation, and thus develop an inflammatory type of Alzheimer’s; others will have dominant contributors via toxins, and thus develop a toxic type of Alzheimer’s; still others will have a dominant contribution from a reduction in trophic support, and thus develop an atrophic type of Alzheimer’s; and others will have their dominant contribution from a reduction in energetic support, and thus develop an energetic (often vascular) type of Alzheimer’s disease.

Another implication of the theory is that the standard transgenic mouse models 62 in use lack a critical feature, and thus it is not surprising that therapeutic success in these models has not predicted clinical success. The mice represent genetic models, and they lack the many upstream contributors (insulin resistance, chronic pathogens, toxin exposure, vascular disease, nocturnal hypoxemia, etc., as described above) that play central roles in Alzheimer’s disease in human patients. However, the theory predicts that mouse models employing the numerous features typical in sporadic AD should be more accurate in their guidance for therapeutic success.

Similarly, there is a critical distinction between the concept that an increase in amyloid-β cancause Alzheimer’s disease—which the rare APP mutations in familial Alzheimer’s disease support—and the concept of amyloid-β as the cause of Alzheimer’s in sporadic cases, as opposed to being a mediator of the pathophysiology in these much more common cases.

The immune system plays a critical role in the proposed model, just as it does for COVID-19: an optimal adaptive immune system would limit pathogen survival and avoid chronic innate system activation, thus limiting chronic inflammation. Poor adaptive system function, as well as ApoE4, increase chronic inflammation 63 and thus move the signaling toward synaptoclastic.

The theory also explains why some vaccines lower risk—such as influenza 64, Herpes zoster6566, and BCG 67—whereas others such as COVID may be associated with increased risk for decline 68. Vaccines result in a lower pathogen burden over time, but also may trigger inflammation. The theory predicts that the effect on cognitive risk depends on which effect is dominant: if the dominant effect is to reduce pathogen burden (especially for pathogens that commonly trigger cognitive decline), then protection will occur, whereas if the dominant effect is to trigger inflammation associated with innate immune system activation (especially chronic inflammation), then increased risk for decline will occur. The pro-inflammatory effect of ApoE4 skews this outcome toward the latter scenario.

The theory offers a physiological explanation for the observation that prions are present in all of the major neurodegenerative diseases. Instead of horizontal transmission of prions, as has occurred, for example, in cases of Creutzfeldt-Jakob disease due to cadaver-derived human growth hormone or depth electrodes 69, prions such as Aβ, p-tau, and a-synuclein are host-generated and subserve the antimicrobial and pre-inflammatory effects described above.

Furthermore, although neurodegenerative diseases have been associated with “aggregated proteins” or “misfolded proteins,” evidence is accumulating that, rather than misfolded and aggregated proteins, these accumulations may be physiologically relevant multi-protein and RNA complexes, assembly machines 70 that play a role in both viral capsid assembly and cellular processes such as the innate immune system response.

Another implication of the theory is the exponential increase in Alzheimer’s disease with advancing age, since aging affects multiple variables in the equation above: inflammation, toxic burden, energetics, and trophic activity, as well as the time-dependent integration of the ratio of synaptoclastic to synaptoblastic signaling based on these parameters. As p-tau studies are demonstrating, the underlying pathophysiological process is ongoing for years before enough functional synapses from the network are lost to produce symptoms 71.

The Pr2 theory is compatible with the observation that some patients with Alzheimer’s disease may not only improve cognition, but sustain improvement for years, sometimes over a decade 58. Addressing the identified drivers of cognitive decline may lead to synaptoblastic signaling rather than synaptoclastic signaling, and return functional synaptic transmission, especially for those patients who begin treatment early, prior to major neuronal cell loss.

Ramifications of an optimal theory would also place the conceptual framework in its proper position with respect to other diseases, such as cancer. Unlike in cancer—in which the original insult need not continue, since its impact is usually etched into the somatic genome (or epigenome), and even amplified in a Darwinian fashion, making the removal of the original insult (such as cigarette smoke) an ineffective treatment—in Alzheimer’s disease, the insults are typically ongoing. This offers the opportunity to intervene in an insult-dependent fashion, which is an advantage over cancer treatment, but also highlights the importance of identifying and addressing ongoing insults as part of an optimal therapy for Alzheimer’s at all stages: dementia, MCI (mild cognitive impairment), SCI (subjective cognitive impairment), and presymptomatic.

The theory also suggests that progressively more predictive risk models may be derived by combining age, genome, and the variables that contribute to the ratio described above: inflammatory status, toxic burden (including effects on biochemistry, such as reduction in glutathione), energetic status (which would include several variables, such as mitochondrial function, oxygen saturation, blood flow, and ketone concentration), trophic status (which would include neurotrophic factors, hormones, and specific nutrients), stress (including HPA axis status), and neurotransmitter status (especially acetylcholine, glutamate, GABA, and dopamine). This should ultimately lead to improved actuarial tables for Alzheimer’s, thus making dementia insurance (as an active rather than passive insurance) feasible and readily affordable for most.

The Pr2 theory offers clear predictions about what is required for prevention and effective treatment of each neurodegenerative disease. The rationale for Alzheimer’s disease treatment has been presented previously 72, and the main tenets are summarized here:

1) The contributors to reduced supply and/or increased demand for the affected network should be identified. The earlier these are identified and addressed, the more rapid and complete the return of function is. Therefore, methods of early detection, such as plasma p-tau 217 in Alzheimer’s, and mass spectrometric profiling of sebum in Parkinson’s 73), represent key advances in the reduction of neurodegenerative disease morbidity and mortality (Table 4).

2) The supply should be restored and demand reduced until supply once again exceeds demand, preventing further degeneration and allowing repair to occur. Only after the contributors have been identified and addressed, and the supply/demand ratio normalized, should the prions such as Aβ or a-synuclein be removed, and then only slowly, to avoid traumatic damage to vessels and other prion-associated structures.

3) The lost elements of the affected network—synapses, neurites, neurons, and other structures—should be restored, and the pathophysiological responses such as microglial and astroglial activation should be mitigated.

4) The support for the affected network should be continued and repeatedly optimized, so that sustained improvement occurs.

The Pr2 theory proposes that each of the major neurodegenerative diseases is the result of a network insufficiency determined by three major supplies and three major demands. However, within these groups, there are more common and less common contributors. For example, in Alzheimer’s disease, examples of the common contributors to this net insufficiency include insulin resistance, metabolic syndrome, sleep apnea, air pollution exposure, Herpes family infections, and gastrointestinal inflammation, among others. In contrast, Parkinson’s disease is often associated with organic toxicants such as trichloroethylene or paraquat, and has also been associated with a gut microbiome that includes Desulfovibrio species 75. For ALS, exposures such as lead, some pesticides, solvents, and glutamatergic agents such as beta-methylamino-L-alanine are likely contributors 76. And for age-related macular degeneration, cigarette smoking is a common contributor, along with high-fat diets, hypertension, and high-energy electromagnetic radiation such as ultraviolet light or blue light 77.

Therefore, the common contributors are quite distinct for these major neurodegenerative diseases, yet all contribute, by various mechanisms, to net insufficiencies in the support of these finely tuned and highly demanding neural subnetworks, all of which feature only modest reserves due to their evolutionary pressure. Thus the subnetwork involved and prionic response—i.e., the diagnosis—dictates the priorities of the evaluation and treatment, even though all six of the major groups of drivers listed above may contribute to the insufficiency.

For compromise of the subnetwork subserving neuroplasticity (at all levels, from molecular to cellular to synaptic to systemic), and leading to diseases including Alzheimer’s (including PCA, PPA (lv), and those cases of corticobasal syndrome associated with AD), diffuse Lewy body disease, LATE, and frontotemporal lobar degeneration, as noted above, the rationale for evaluation and treatment based on this theory has been published 72. As noted in Table 1 above, contributors to non-amnestic presentations of AD tend to be different from those of more classical, amnestic AD, and thus evaluation and treatment should take that into account, with focus on toxin exposures and chronic, tick-borne infections in addition to the typical metabolic and inflammatory contributors. For DLBD, the focus should include pesticides, heavy metals, and air pollutants.

From these evaluations, the various contributors to the degenerative process are identified and then addressed therapeutically. Infections such as HSV-1 are readily treated. Metabolic dysfunction may be corrected with a plant-rich, mildly ketogenic diet and lifestyle optimization (including both aerobic exercise and strength training), or treated with GLP-1 receptor agonists, among other approaches. Oral dysbiosis and gut dysbiosis are readily detected and treated (and it is noteworthy that studies of the brain microbiome have repeatedly revealed oral microbiome members 33). Sleep apnea, a common contributor to cognitive decline, is also readily detected and addressed. Toxins and toxicants from the three major groups—inorganics, organics, and biotoxins—are also detectable, typically in blood and urine samples, and detoxification methods are in widespread use.

Adapting this approach to the major neurodegenerative diseases affecting motor modulation—Parkinson’s, progressive supranuclear palsy, and corticobasal degeneration—shifts focus to prioritize organic toxicants such as trichloroethylene, dieldrin, and paraquat, while retaining the overall evaluation and treatment targeting the six groups of contributors listed above. Furthermore, in addition to the production of anti-microbial a-synuclein in patients with Parkinson’s (as well as diffuse Lewy body disease and multiple system atrophy), the surprising identification of a Parkinson’s-specific profile in sebum years before the appearance of motor symptoms supports the notion of a response to pathogens, since sebum possesses antibacterial and antifungal properties 78.). Furthermore, the recent association of bacteria of the genus Desulfovibrio with Parkinson’s disease 75, along with the discovery of the gut microbiome’s potential role in Parkinson’s 79, emphasizes the need for the evaluation of the gut microbiome in patients with possible Parkinson’s disease or risk for Parkinson’s disease. Optimal treatment would then require inclusion of detoxification, mitochondrial support and/or enhancement and/or transfusion 80, and optimization of the gut microbiome, along with addressing common parameters such as systemic inflammation and insulin resistance 81.

Modification to treat ALS requires the recognition that the system that has evolved to mediate massive amplification, from thought to maximal strength and movement almost instantaneously, possesses a vulnerability in glutamate excitotoxicity, for example in the clearance of the excitotoxic neurotransmitters glutamate and aspartate 82 or in the presence of an exogenous excitotoxin such as β-N-methylamino-L-alanine (BMAA) 83. Many potential contributors to this excitotoxic excess have been identified, such as Lyme disease 84, cyanobacteria 85, heavy metals (lead, mercury) 86, copper deficiency 87, retroviruses 88, and professional athleticism 89, among many others. A review of patients with definitively diagnosed ALS who nevertheless achieved sustained improvement in at least one objective measure (as compared to typical progressive ALS patients) revealed associations with numerous medications and supplements such as azathioprine, glutathione, curcumin, vitamin D, and luteolin, among others 90. Thus optimal treatment should identify the cause(s) of the reduced glutamate/aspartate clearance, as well as addressing the other parameters of trophic support, energetics, etc. listed above.

Modification to treat age-related macular degeneration increases the priority of evaluation to metabolic demand and the complement system: demand related to the requirement for the retinal pigment epithelial cells (RPE) to keep up with the exigency to phagocytose the photoreceptor outer segments (which increases with longer daily light exposure, especially blue light, and is compromised by living at high altitude, cigarette smoking, sleep apnea, and other factors), and complement to respond to the spillover (thus the common CFH mutations lead to enhanced inflammation and increase AMD risk, analogous to the ApoE e4 allele in Alzheimer’s risk).

Previous theories of Alzheimer’s disease have not led to effective treatment. However, the theory described here has received support both from anecdotal studies in over 100 patients 18192091, and from multiple clinical trials, in which over 70% of the patients showed improvement (not simply a slowing of decline), including some patients increasing their MoCA (Montreal Cognitive Assessment) scores (by external raters) from 19/30 to 30/30 16.

These patients were evaluated for the many potential contributors to synaptoclastic and synaptoblastic signaling, from infectious agents to other inflammagens to insulin resistance to sleep apnea to nutrient and hormonal deficiencies, as well as the other contributors listed above. These were then targeted with a personalized, precision medicine protocol, using a computer-based algorithm derived from the theory described here. It is noteworthy that some of the earlier reported patients have now been on this protocol for over 10 years, with sustained improvement, suggesting that the root cause contributors to the cognitive decline are indeed being impacted 58.

The model presented here also offers explanations for many of the otherwise unexplained results reported. For example, the model predicts that the removal of the anti-microbial amyloid-β with an antibody may do little if the upstream drivers of synaptoclastic signals are not identified and removed, and in fact may lead to decline if there is ongoing exposure to pathogens (which has been observed in several patients). However, these same antibodies may potentially be highly effective were they to be combined with the rest of the protocol, and administered after any identified pathogens had been treated and insults addressed. A similar argument can be made for removing another endogenous anti-microbial, p-tau.

In contrast, novel pharmaceuticals that target the various contributors to the equations above may prove useful as part of an overall protocol, especially if tested in those individuals shown to have the respective contributors, such as specific toxins associated with air pollution or specific insulin signaling or specific trophic factor reduction. This offers new avenues for therapeutic development, based on personalized, precision-medicine models.

Another implication of the theory is that a monotherapy—any monotherapy—is unlikely to be as effective for patients with Alzheimer’s or other neurodegenerative diseases as a targeted, comprehensive protocol. This provides some insight into why so many clinical trials of pharmaceutical agents have failed despite supportive pre-clinical data. However, these same agents may prove effective when used as part of a targeted, combination therapy.

Furthermore, the network that is being targeted therapeutically is highly unlikely to function as a linear system, and therefore, the idea that any combination therapy must include only agents that show a statistically significant effect when used alone (along with the related notion that one can simply add up the effects of different agents to determine the effect of a combination) is invalid. In other words, optimal combinations may indeed turn out to include agents that show no significant effect when used as monotherapies. This suggests that determining the contributions from each of the components of a combination therapy may achieve more success by subtracting a single agent from a combination instead of testing single agents alone, or by using AI (artificial intelligence)-based methods to evaluate many varied combinations.

Pr2 theory also suggests that Alzheimer’s disease should be readily preventable in a cost-effective, globally-scalable way, using a sieve approach: in a given population, adoption of a basic set of practices that increases synaptoblastic signals and reduces synaptoclastic signals—such as insulin-sensitizing diet and identifying and treating sleep apnea, among others—should reduce population risk; those at high risk would take a second step of evaluation for the manyparameters that affect the synaptoclastic ratio and address those specifically; the small minority that would go on to display symptoms would have these addressed early on with deeper studies on what failed in prevention, and be treated with a more extensive, targeted protocol; and those rare individuals failing even that would become in-patients briefly so that more extensive evaluation and optimal treatment could be undertaken. This approach would allow the direction of the majority of resources to those in greatest need, and render the overall program cost-effective. Given the results of proof-of-concept clinical trials, in which more than 70% of patients with MCI and early dementia demonstrated cognitive improvement 1617, early evaluation and treatment have the potential to reduce the societal burden of dementia markedly.

As noted above, previous, mono-etiological models of Alzheimer’s disease have failed in the translation to effective therapy. However, there are features of the various theories that are compatible with observations, and those features all fit within the context of the theory presented here. For example, the infectious theory of Alzheimer’s fits well with the finding that many patients with Alzheimer’s do indeed have pathogens within their brains (and associated inflammation), such as Herpes simplex 92. However, this does not explain the many who have these same pathogens but do not develop Alzheimer’s disease, nor does it explain the many other risk factors, nor does it explain the many who develop Alzheimer’s disease in the absence of those pathogens. Thus the case for infections as contributory to Alzheimer’s disease pathophysiology is strong, and fits well within the current theory, but there is no evidence that these are either necessary or sufficient.

Similarly, the notion that Alzheimer’s disease is “type 3 diabetes” 93 is only partially correct—just as for infections, insulin resistance is neither necessary nor sufficient for Alzheimer’s disease, although it is a common contributor, and fits well within the model proposed here.

Both amyloid-β peptide and tau have been shown to possess prionic properties, i.e., to replicate in a process that does not require them to carry a nucleic acid. This is an important feature of the plasticity of the network itself: biological systems with a single goal outcome and no requirement for amplification—such as serum pH—feature negative (homeostatic) feedback; in contrast, biological systems that feature alternating goal outcomes and require amplification as part of the switching—such as blood clotting—feature positive feedback, i.e., prionic loop feedback. The synaptoblastic-synaptoclastic signaling switching mediated by APP (part of the overall connection vs. protection mode switch), based on the many contributory factors listed above, features prionic loop feedback: amyloid-β peptide inhibits the a-secretase cleavage of APP 23, and conversely, sAPPa is a potent inhibitor of the β-secretase, BACE1 50. Thus this signaling tends toward one of two stable states—synaptoblastic signaling or synaptoclastic signaling—and there is a threshold for switching states. In other words, APP signaling acts essentially as a bistable threshold switch, leading to either synaptoblastic or synaptoclastic signaling/outcomes. Prionic signaling is thus a property of the system, and a mediator of the pathophysiology, but there is no evidence that horizontal spread is a common cause of Alzheimer’s disease.

Similarly, none of the mono-etiological, non-network theories—those noted above or the amyloid cascade hypothesis 94 or the tau hypothesis 95 or the oxidative/nitrosative stress theory 96 or the myelin theory 97 or others—offers adequate explanation for the remarkably diverse set of risk factors for Alzheimer’s, nor does it offer a clear program for prevention, nor is it supported by evidence of translation to clinical efficacy.

The theory is supported by patient outcome data in patients with MCI and dementia, in two proof-of-concept trials 1617, with an ongoing randomized controlled trial 98. Some illustrative examples include the following:

The Pr2 theory has a number of implications for the evaluation, treatment, prevention, and understanding of neurodegenerative diseases:

· Removing prions such as amyloid-β or a-synuclein may potentially lead to worsening of symptoms and the degenerative process if these are targeted therapeutically prior to addressing the upstream inducers of the prions, such as various pathogens. If prions are to be removed, the best timing would be after the inducers have been treated, and the best predicted rate of removal would be over months or years rather than weeks.

· Since neurodegenerative diseases represent complex network insufficiencies rather than simple diseases, adjusting multiple factors rather than complete suppression of one may lead to the best outcomes—multiple modest changes instead of one massive change.

· The Pr2 theory suggests multiple new and recent therapeutic targets, such as various infectious agents (e.g., Herpes virus family members, fungal species, and bacterial species such as P. gingivalis), toxins (inorganic toxicants, organic toxicants, and biotoxins), gastrointestinal hyperpermeability, sleep apnea, oral and CNS dysbiosis, innate immune system mechanisms (which may be treated with specialized pro-resolving mediators, mast cell inhibition, and other anti-inflammatory therapeutics), support for the adaptive immune system, energetic support (e.g., ketones, creatine, and EWOT (exercise with oxygen therapy)), glycotoxicity (e.g., GLP-1 agonists or ketogenic diets), neurotrophins (such as BDNF and NGF), and hypothalamic-pituitary-adrenal axis-related stress mediators.

· The prion paradox—the fact that prions are by definition proteinaceous infectious agents, and have been identified in all of the major neurodegenerative diseases (Aβ, p-tau, a-synuclein, TDP-43, and PrP^Sc), yet there is only evidence for horizontal transfer of prions (“prion infection”) in a tiny fraction of cases—is explained by the Pr2 theory, since prion function includes antimicrobial effects and pre-inflammatory mediation.

· The Pr2 theory predicts that most patients who develop neurodegenerative diseases will have more than one contributor, and that all of the contributors will compromise the implicated neural subnetwork’s energetic support, trophic support, or neurotransmitter(s), and/or increase the demand via inflammation, toxicity, or stress, either directly or indirectly. Therefore, identifying and addressing these contributors is important for optimal outcomes. The therapeutic method of simply removing the associated “misfolded” proteins is unlikely to offer optimal efficacy. Furthermore, the constants associated with each of the many potential contributors should be calculable from patient data and AI, leading to more exacting prognostics and improved outcomes.

· Which neurodegenerative disease develops in any given individual is specified by the inducing insults, their routes of entry, and the prionic response.

· Reducing the overall burden of neurodegenerative disease should be enhanced by early detection methods (such as sensitive measurements of p-tau 217 in plasma) and wearables (such as continuous glucose monitors, continuous ketone monitors, and sleep trackers), which reveal suboptimal parameters such as glucose spikes and troughs, loss of metabolic flexibility, reduction in deep sleep, suboptimal nocturnal oxygenation, and many other parameters critical for optimal neural protection. In the long run, it should be possible to develop a real-time analysis of cerebral supply/demand status, allowing earlier detection and ensuring prevention for virtually all.