Oh my gut! Is the microbial origin of neurodegenerative diseases real?

ABSTRACT There is no cure or effective treatment for neurodegenerative protein conformational diseases (PCDs), such as Alzheimer’s or Parkinson’s diseases, mainly because the etiology of these diseases remains elusive. Recent data suggest that unique changes in the gut microbial composition are associated with these ailments; however, our current understanding of the bacterial role in the pathogenesis of PCDs is hindered by the complexity of the microbial communities associated with specific microbiomes, such as the gut, oral, or vaginal microbiota. The composition of these specific microbiomes is regarded as a unique fingerprint affected by factors such as infections, diet, lifestyle, and antibiotics. All of these factors also affect the severity of neurodegenerative diseases. The majority of studies that reveal microbial contribution are correlational, and various models, including worm, fly, and mouse, are being utilized to decipher the role of individual microbes that may affect disease onset and progression. Recent evidence from across model organisms and humans shows a positive correlation between the presence of gram-negative enteropathogenic bacteria and the pathogenesis of PCDs. While these correlational studies do not provide a mechanistic explanation, they do reveal contributing bacterial species and provide an important basis for further investigation. One of the lurking concerns related to the microbial contribution to PCDs is the increasing prevalence of antibiotic resistance and poor antibiotic stewardship, which ultimately select for proteotoxic bacteria, especially the gram-negative species that are known for intrinsic resistance. In this review, we summarize what is known about individual microbial contribution to PCDs and the potential impact of increasing antimicrobial resistance.

T he early broad therapeutic applications of antimicrobials, mainly heavy metals and other toxic chemicals, were influenced by the Germ Theory of Disease during the early 19th century (1).Interestingly, the initial diagnoses of neurodegenerative protein conformational diseases (PCDs), such as Alzheimer's disease (AD), Parkinson's disease (PD), Huntington's disease (HD), or amyotrophic lateral sclerosis (ALS) coincide with the same period (2)(3)(4)(5), suggesting that, in addition to heavy metal toxicity, antimicrobials could have contributed to the pathogenesis of these diseases.Although, the initial diagnoses of PCDs could have been a result of the rapid population expansion or increase in medical practice that led to better diagnosis.However, what is more striking is that the incidence of each of these diseases began to rapidly increase following the discovery and commercialization of antibiotics, especially after the second half of the 20th century post the "Golden Age" of antibiotics (6).There is stronger evidence suggesting that microbes contribute to the pathogenesis of PCDs, but could antimicro bials also contribute to these diseases by eradicating the protective microbiota?If so, do antibiotics eliminate protective microbes and enrich for detrimental bacteria?How will the growing antimicrobial resistance (AMR) among pathogenic bacteria affect the incidence of PCDs?In this review, we will attempt to address these intriguing questions by summarizing the current literature.
The first suggestion that bacteria may be affecting the pathogenesis of PCDs, or more specifically Parkinson's disease, was postulated by Braak (7); however, the evidence of the microbial origin of PCDs can be seen throughout the literature much earlier.For example, prior to the discovery of antibiotics, syphilis, a sexually transmitted infection caused by Treponema pallidum, a spirochaete bacterium, was the leading cause of paralytic dementia, a neurological condition that manifests in rapid cognitive decline, which accounted for a significant rise in asylum admissions (8,9).Spirochetal infections were later linked to AD (10,11).One of the first successful treatments at the begin ning of the 20th century was possible due to the discovery and commercialization of Salvarsan, an arsenic-based antimicrobial, followed by the discovery of penicillin, which provided effective remedies and consequently eliminated neurological complications associated with T. pallidium infection (12,13).Further evidence of microbial contribu tion to neurodegenerative diseases was observed in Japanese leprosy patients treated with dapsone, an antimicrobial sulfa drug.The patients who were administered the drug were more than twice less likely to develop dementia (14).It remains unknown whether the protective property of dapsone was the result of any off-target effects or whether the drug indirectly affected the disease pathogenicity by targeting Mycobac terium leprae.Another study has linked mycoplasma presence to ALS.Nicolson et al. found that 83% (30/36) of ALS patients had mycoplasma detected in blood cultures compared to only 2.8% in the control group (2/70) (15).Mycoplasma pneumonia has also been linked to acute parkinsonism (16).More recently, antibiotics have been shown to alleviate disease phenotypes in PCD patients suffering from chronic infection.For example, 88% of AD patients were diagnosed with Helicobacter pylori infection compared to 46.7% of controls, and eradication of the pathogen was associated with decreased disease progression in both a clinical trial and a population-based study (17)(18)(19).H. pylori eradication with an antibiotic cocktail consisting of omeprazole, clarithromycin, and amoxicillin improved AD-associated symptoms, enhancing cognitive and functional skills (18).Furthermore, a population-based study revealed that eliminating H. pylori decreases dementia progression (19).Enteric infections caused by H. pylori have been associated with increased intestinal permeability (20), and Parkinson's disease pathoge nicity (21,22).An increase in intestinal permeability can lead to the dissemination of the intestinal content to other parts of the body.In fact, bacteria and their products have been found in the brains of people with PCDs, including AD (23-25).Other bacteria, such as group B Streptococcus (GBS), are part of the commensal microbiota that asymptomatically colonize various sites of the human microbiome (26).GBS is known to disrupt intercellular junctions of intestinal epithelial cells (27), which likely contributes to changes in intestinal permeability, leading to a displacement of bacteria and possibly other members of the gut microbiota.Interestingly, acute parkinsonism has been observed in post-streptococcal infections in young individuals (28)(29)(30).The link between intestinal permeability and brain disorders is evident (31), as elevated levels of circulating lipopolysaccharide (LPS), a bacterial endotoxin that contributes to PCD pathogenicity (32,33), have been detected in elderly patients, patients with chronic infection (34), and in AD and ALS patients (23,35).LPS has also been shown to contribute to the pathology of PD in mouse models (36,37).The collective evidence suggests that antibiotics provide a level of protection against neurodegenerative diseases, but likely in the context of the infection.
Although antibiotics seem to have an inhibitory effect on the pathogenicity of PCDs, the benefits are not evident in the absence of infection.Certain antibiotics, for example, tetracyclines, were promising therapeutic candidates as they were found to directly interact with different protein amyloids and interfere with their aggregation in vitro (38)(39)(40)(41); however, they completely failed to provide any protection in clinical trials (42).An increasing body of evidence supports the proteotoxic role of antibiotics on the pathogenicity of neurodegenerative diseases.For instance, antibiotics were found to increase the risk of PCDs in two nationwide case studies (43,44).Sun et al. analyzed antibiotic use among 2,484 Swedish ALS patients and found an increased risk associated with all antibiotics, especially β-lactamase-sensitive penicillin (44).While these results suggest that the disruption of the protective microbiota might contribute to the disease pathogenesis, other studies showed a protective effect of penicillin (45).Such discrep ancy in the influence of antibiotics on ALS pathogenicity is likely attributed to two factors: disruption of the protective microbiota and elimination of proteotoxic bacteria.As such, in the absence of any underlying infection, penicillin administration would be either detrimental, as observed by Sun et al., or have no effect on ALS pathogenicity-a result that was observed in a small clinical trial (46).The detrimental effect of antibiotics on ALS was also seen in patients treated with a tetracycline-class antibiotic, minocycline, which accelerated disease progression and increased mortality (47).Tetracyclines are known to inhibit bacterial growth by reversibly binding to the prokaryotic ribosome and ultimately blocking translation (48); however, recent data suggest that tetracyclines can also interact with the eukaryotic ribosome (49), which could affect eukaryotic translation rate and ultimately alleviate the expression of destabilized aggregation-prone proteins that are most sensitive to proteotoxicity.While numerous studies have demonstrated the neuroprotective effect of minocycline in animal models, little evidence exists to support its efficacy in humans (50,51).Minocycline's lack of efficacy in humans may not be surprising given a 100% failure rate of potential therapeutics that were initially tested in around 170 transgenic murine models of AD (52).Such discrepancy between tetracycline efficacy in vitro and in vivo animal models vs humans may be mediated by the differences in human and rodent gut microbiota, which only share a mere 10% of bacterial species (53).In agreement with the detrimental effect of antibiotics in the absence of infection, a large Finnish study analyzed the effect of antibiotic usage on PD risk.The analysis of nearly 14,000 PD patients revealed an elevated risk of developing the disease after exposure to different antimicrobials, including macrolides, lincosamides, tetracyclines, sulfonamide, and trimethoprim (43).In a recent study that included 14,000 women, Mehta et al. noted a detectable cognitive decline 7 years post mild antibiotic use (54).Furthermore, a recent analysis of early-and mid-life antibiotic usage within the Swedish population revealed a significant association with AD and PD (55).The detrimental effect of antibiotics is further supported by another study that linked the duration of antibiotic treatment with increased risk for dementia in a cohort of 313,161 participants (56).Collectively, these findings indicate that the protective effect of antibiotics is likely associated with the eradication of proteotoxic bacteria that contribute to the pathogene sis of PCDs, and general antibiotic use could increase the risk of PCDs later in life, possibly by altering the protective microbiome.
Antibiotics might increase the risk for neurodegenerative diseases by depleting bacteria that confer neuroprotection to the host.For example, clinical data reveal that the abundance of bacteria from the Prevotella genus negatively correlates with the pathogenicity and prevalence of different PCDs, including ALS (57), PD (58)(59)(60)(61)(62)(63)(64)(65)(66), and a mouse AD model (67).A decreased abundance of Prevotella spp. is also associated with constipation, a condition that precedes motor symptoms in PD patients by up to 20 years (46,68).These results suggest that changes in the gut microbiota are the cause and not the effect of the disease.According to the Global Biodiversity Information Facility, there are 61 species of Prevotella (69), but only about 20 are associated with human diseases (70).As such, reporting results at the genus level hinders our true understand ing of bacterial contribution.While the mechanisms underlying the neuroprotective role of Prevotella spp.remain unknown, recent work revealed that specific species (i.e., disiens and corporis) decrease toxic polyglutamine (polyQ) aggregation in Caenorhabditis elegans (71,72).Studies in pigs and rats have shown that antibiotic use reduces the relative abundance of Prevotella spp.(73,74), suggesting that a similar effect can be expected in humans, which may support the link between antibiotics and neurodege nerative diseases (43,44).Ampicillin was shown to significantly alter the composition of the microbiota in rats and was associated with physiological and psychological changes, including two hallmarks of AD: memory defects and elevated glucocorticoid levels (75).The aforementioned issues, as well as gut dysbiosis, were rescued with probiotic supplementation of Lactobacillus fermentum, a commensal strain that is known for its potent antioxidative property (76).Several studies have demonstrated the beneficial effect of Lactobacillus spp.For example, a probiotic cocktail that included L. fermentum and L. plantarum delayed the onset and progression of AD in Drosophila (77), whereas another probiotic cocktail containing Lactobacillus spp.alleviated PD symptoms in mice (78) and reduced AD symptoms in rats (79).What is interesting to note is that while 100% of therapeutics that effectively treated AD in animal models failed in human clinical trials, this is not the case with bacteria-mediated approaches.For example, a double-blind, randomized study demonstrated that a probiotic cocktail containing Lactobacillus spp., including fermentum, improved cognition in people with AD after 12 weeks of supple mentation (80).Together, these results accentuate the risk of antibiotic-induced gut dysbiosis, particularly in the context of neurodegenerative PCDs.
Our recent work revealed that host proteostasis is affected by changes in the bacterial proteome.We showed that colonization of the C. elegans intestine with bacteria that have an increased abundance of protein aggregates leads to disruption of the host proteostasis (72).Walker et al. demonstrated that gentamicin, an antibiotic from the aminoglycoside class known to kill bacteria by inducing mistranslation and protein misfolding (81), not only increased the abundance of bacteria-derived protein aggre gates (BDPAs) but also enhanced polyQ aggregation in C. elegans colonized with bacteria exposed to this antibiotic (72).Since antibiotics are specifically targeting bacteria, these results suggest that they may also indirectly affect the stability of the host proteome.In fact, antibiotics can induce uncharacterized BDPAs that play a physiological role in bacterial virulence and persistence (82).Together, these results raise the concern that antibiotics may not be as inert on the host as is generally accepted, and they can contribute to the pathogenesis of PCDs by affecting the composition of the human microbiome, as well as enhancing BDPAs.
While it is estimated that the human gut harbors at least as many bacteria as there are cells in the human body, the number of bacterial genes far exceeds that of humans by a staggering projection of 100-or more-fold (83,84).Therefore, it should not be surprising that there are bacterial genes that affect host proteostasis and ultimately contribute to the pathogenesis of PCDs.Gram-negative bacteria, mainly from the Enterobacteriaceae family (85), produce curli which are functional amyloids that aggregate to form an extracellular matrix that closely resembles aggregates associated with PCDs, such as AD or PD (Fig. 1).These bacterial functional amyloids play a role in cell adhesion, biofilm formation, and antibiotic resistance, and recent studies suggest that these proteins may also be implicated in the pathogenesis of PCDs (86)(87)(88)(89)(90)(91).
For example, FapC, the amyloid produced by Pseudomonas, affected in vitro aggrega tion of α-synuclein and amyloid beta (Aβ), and enhanced AD pathogenesis in zebrafish (94,95).Interestingly, about 80% of cystic fibrosis (CF) patients are colonized by Pseudomonas aeruginosa by the age of 18 (96).Despite CF being a non-neurodegenera tive PCD, recent studies show that CF patients exhibit cognitive dysfunction and brain tissue changes (97,98).The life expectancy of those affected with CF ranges between 30 and 40 years, which is about 20-30 years less than the average diagnosis age of PCDs such as PD or AD (99,100).Therefore, it is possible that this age difference is the reason for no obvious correlation between CF and neurodegenerative PCDs; however, it is striking that CF patients develop secondary amyloidosis due to aggregation of serum amyloid A protein (SAA) (101).Interestingly, the levels of SAA were shown to correlate with P. aeruginosa presence in the lungs of CF patients and were attenuated with antibiotics, further supporting our hypothesis that bacteria, in this case, P. aeruginosa, contribute to the disruption of host proteostasis (102).
Additional indirect evidence suggesting that antibiotics are linked to neurodegenera tive diseases comes from their usage within the general population, where women are more commonly prescribed antibiotics, and in some studies, the difference accounts for nearly twice as many prescriptions when compared to men (103,104).Surprisingly, the lifetime risk for developing Alzheimer's dementia at the ages 45 and 65 is nearly twofold higher for women compared to men (105,106).These results support the finding by Mehta et al., who reported that antibiotic use in women during midlife is associated with decreased cognition 7 years later (54).Furthermore, women have a lower abundance of Prevotella spp. in their gut, which, given the suspected neuroprotective property of these bacteria, may contribute to the difference in the risk for AD between women and men (107)(108)(109).Based on the correlation between Prevotella abundance and antibiotic use in animals (73,74), the increased use of antibiotics may be linked to a lower abundance of Prevotella and an increased risk for AD and dementia.Other evidence linking antibiotics with AD is based on the correlation between AD prevalence and antibiotic consumption.Interestingly, based on 2018 surveillance data, Italy and Greece had the highest dementia prevalence among all European countries (110).In 2010, these two countries were also one of the largest consumers of antibiotics (111), and in 2010, Italy was also the largest user of antibiotics in livestock (112).These data reveal a surprising correlation between global antibiotic consumption and dementia and may support previous reports that link antibiotic usage to elevated risk for several neurodegenerative diseases, including AD and dementia (43,44,47,56).In addition to global antibiotic consumption and sex-specificity, the recent coronavirus disease 2019 (COVID-19) pandemic has increased the risk for dementia and dementia-related deaths, likely due to social isolation (113), reduced accessibility to social activities due to lockdowns and social distancing (114,115), and most importantly, the overuse of antibiotics (116).Moreover, it was recently estimated that COVID-19 increased the risk for AD twofold, and AD-related deaths increased by 16% (117,118).While it is not clear whether the virus directly affects the pathology of AD or whether the effect comes from the abovementioned consequences of the pandemic, the evidence points toward the inflammatory effect of the virus, secondary infections, and poor antibiotic steward ship-all of which are known to contribute to the pathogenesis of PCDs.A survey of 10,403 pneumonia patients revealed that 3% of the affected individuals who contracted COVID-19-related pneumonia developed new-onset dementia, an incidence that is much higher than in non-COVID-19 patients (119).Pneumonia was previously linked to PCDs in critically ill patients (120).In one study, it was noted that hospitalization with COVID-19 was associated with a cognitive decline, where the incidence reached over 26% in patients with severe COVID-19 (121).A recent pre-print also reported the presence of amyloid deposits in the brains of patients with severe COVID-19 (122).It is important to note that severe COVID-19 patients are subjected to empiric antibiotic treatment (123).Furthermore, in a Danish population-based study, Zarifkar et al. found that COVID-19 increased the frequency of AD and PD by 3.4 and 2.2 times, respectively (124).The study found that most neurological disorders were no more frequent after COVID-19 than after other respiratory infections, which suggests that the detrimental effect could be due to antibiotic treatment or general inflammation.
While the link between COVID-19 and AD could be explained by the disrupted antibiotic stewardship and underlying secondary bacterial infections, the potential role of other viral infections in AD pathogenesis has also been established, though the viral contribution to PCDs is another extensive topic that is outside the scope of this review.Nonetheless, it is important to emphasize that in addition to coronavirus other viruses including herpes simplex, human cytomegalovirus, Epstein-Barr virus, hepatitis C, picornavirus, Borna disease virus, and influenza were associated with cognitive decline, and in some instances their presence correlated with AD and other PCDs (125)(126)(127).Unlike common extracellular bacteria, viruses are obligate intracellular parasites and as a result, hijack the cellular protein homeostasis machinery, often sequestering heat shock proteins away from their endogenous roles (128).As such, it is not surprising that many of the aforementioned viruses activate the heat shock response and could potentially lead to misfolding and aggregation of any destabilized disease-associated host proteins (129).Additionally, viruses are also known to affect neurons leading to their degeneration (130,131).
Antibiotics directly target bacteria and can rapidly alter the composition of the gut microbiota, but they are not the only factor that influences the microbial balance (132).Factors such as alcohol consumption (133,134), smoking (135), physical activity (136), and diet are known to affect the gut microbiota (137).Lourida et al. looked at the association between these factors and dementia in a cohort of nearly 200,000 partici pants.Not surprisingly, a healthy lifestyle associated with moderate alcohol consump tion, no smoking, regular physical activity, and a healthy diet resulted in a lower risk of developing dementia (138).
In summary, numerous bacterial infections have been associated with the pathogen esis of neurodegenerative PCDs, and it does not seem that specific bacteria are linked to any particular PCD, but rather disrupt host proteostasis in general and lead to toxic protein aggregation that manifests in disease pathogenesis.Infection prevention and stricter antibiotic stewardship programs at any age need to be implemented in order to decrease the prevalence of PCDs.

ASYMPTOMATIC COLONIZATION, ANTIMICROBIAL RESISTANCE, AND PROTEIN CONFORMATIONAL DISEASES
Antibiotics are essential medicines known as the wonder drugs, but due to increasing AMR, their efficacy is slowly fading.In fact, AMR is emerging as one of the largest global healthcare problems that is already associated with an estimated 4.95 million deaths annually (139).The number of deaths is projected to increase 10-fold by 2050 if no immediate actions are taken (140).AMR is a dire crisis that likely affects the prevalence of PCDs, as proposed in this review based on the current interpretation of the literature that sits at the interface of microbiology and neuroscience.Many of the bacterial species, especially the gram-negative, which are associated with PCDs are known for intrinsic and acquired antibiotic resistance.Often, these resistant species are linked to PCDs and are known to asymptomatically colonize the human gut.Administra tion of antibiotics eliminates the protective microbiota but enriches for the abundance of the resistant strains.Recent studies show that even a short duration of antibiotic treatment will select for resistant genes within the gut microbiome (141).An increase in the abundance of such asymptomatic colonizers driven by antibiotics can silently contribute to the pathogenesis of PCDs years before the onset of any symptoms.For instance, it was shown that antibiotic treatment could enrich for P. aeruginosa in vitro as well as Klebsiella pneumoniae and Proteus mirabilis in mice (142,143).An increased presence of P. aeruginosa and K. pneumoniae in the gut has been associated with PCDs, and multidrug-resistant (MDR) strains of these bacteria are commonly found across hospitals and community (144)(145)(146).P. mirabilis is an opportunistic pathogen that is otherwise considered commensal, and its abundance was fivefold higher in septic PD patients compared to non-PD controls (147,148).Furthermore, oral administration of these bacteria induced α-synuclein aggregation in mouse brains and enhanced PD-like symptoms (149) (Table 1 summarizes all bacteria-animal models).Interestingly, all three of the aforementioned pathogens were shown to increase toxic polyQ aggregation and the associated toxicity upon colonization of the C. elegans intestine (71).
Over 50% of infections among nursing home residents were due to MDR bacte ria, indicating a high prevalence of antibiotic resistance present within the aged and predisposed population (173).It is estimated that nearly 70% of Americans with dementia will die in nursing homes (174).Recurrent infections and, consequently, antibiotic use among nursing home dementia patients are also frequent (175).In fact, long-term care residents are commonly colonized by MDR gram-negative bacte ria, which are known to be proteotoxic, and such colonization is strongly associated with advanced dementia (176).Therefore, with nursing homes being hotspots for MDR bacteria, the overuse of antibiotics likely fuels the enrichment of the resistant gram-neg ative proteotoxic bacteria that further contribute to dementia and other neurological conditions.In support of this statement, a recent study used metagenomic sequencing to investigate the presence of AMR genes in the microbiota of a group of patients with advanced dementia and found a link between resistance gene density and the relative abundance of P. mirabilis, Enterococcus faecalis, and Escherichia coli (177).The authors further demonstrated that these strains had more resistance genes than many commen sal bacteria in the gut.Perhaps not surprisingly, these bacteria are also associated with an increased risk of PCDs (Table 1).For example, a study found that people with PD have an increased presence of Enterococcus (178), and E. faecalis increased toxic polyQ aggrega tion in C. elegans (162) (Table 1).LPS from E. coli was found to be abundant in brains of AD patients (179), and Escherichia-Shigella abundance was found to be significantly increased in PD patients (178).Pathogenic strains of E. coli have also been shown to impair mobility in a polyQ Drosophila model and cognition in AD mice (164,165) (Table 1).As previously mentioned, P. mirabilis was also associated with PD in humans (148).
Porphyromonas gingivalis is commonly associated with oral infections but can also colonize asymptomatically, as it was found in up to 25% of people with no oral disease (180).P. gingivalis products were found in the brains of AD patients, and the bacterium was shown to enhance the pathogenesis of the disease in mice (181) (Table 1).Injections of encapsulated strains of P. gingivalis in the palatal mucosa of rats led to memory defects, increased Aβ levels, and tau hypophosphorylation in the hippocampus, and intracerebral injection of P. gingivalis increased Aβ deposition in mice (169,170).Similar findings were also present in another mouse study where oral inoculation of P. gingivalis impaired cognitive function and increased levels of Aβ deposition in the hippocampus and cortex of AD mice (168) (Table 1).Broad-spectrum antibiotics rarely eradicate this bacterium, which may lead to the emergence of and selection for MDR strains (182), increasing the concern that antibiotic use can enrich for its presence.The enrichment of these disease-associated bacteria becomes particularly concerning for people who might be asymptomatically colonized and are already at risk for PCDs.
Accumulating evidence from patient data suggests that Chlamydia pneumoniae might play a role in AD pathogenesis (183).Among perhaps the most convincing evidence is that C. pneumoniae DNA was found in ~90% of AD brains compared to ~5% of otherwise healthy controls (184).Intranasal inoculation of the bacterium led to Aβ deposition in the brains of infected mice (160) (Table 1).C. pneumoniae is a causative agent of pneumonia but likely exists asymptomatically in the majority of cases, as 50% of (Continued on next page) 20-year-olds have antibodies to the bacterium, which increases to 75% in 60-70-yearolds, and due to a 3-5-year limited antibody response, it is thought to commonly infect and reinfect people throughout their lifetime (185).Hence, antibiotic use could fuel resistance in those who are asymptomatically colonized with the pathogen and enrich its abundance (Fig. 2), as in vitro evidence supports the emergence of AMR after exposure to subinhibitory antibiotic concentrations (186).

MODEL ORGANISMS FOR NEURODEGENERATIVE PCDS
Common organisms employed to model neurodegenerative PCDs include the nematode Caenorhabditis elegans, the fruit-fly Drosophila melanogaster, and rodents such as rats and mice.The fact that C. elegans can easily be colonized by a single bacterial species has made the model an attractive tool for studying host-bacteria interactions.Additionally, its neuronal network has been fully mapped, and its evolutionary conserved protein homeostasis networks have afforded the production of conclusive literature on the physiological processes that relate to general eukaryotic protein folding (187).Hence, C. elegans is arguably one of the most suitable model organisms to study the effect of bacteria on protein-folding diseases.Relative to C. elegans, D. melanogaster is a higherorder organism and has a brain and is thus capable of more sophisticated behavior, allowing for brain histopathological analyses as well as rudimentary assessments of cognitive and motor function.The higher complexity of D. melanogaster compared to C. elegans can also be a disadvantage; for example, C. elegans is transparent and allows for the detection of fluorescent reporters and sensors in real-time, and similar experiments in C. elegans and D. melanogaster are likely less labor-intensive in the former.Due to their invertebrate classification, neither flies nor nematodes have animal welfare restrictions.Rodent models are advantageous in that their physiology is closely related to humans; however, they are expensive, and experiments are lengthy and labor-intensive, particu larly when studying age-associated diseases as they have a much longer lifespan.Furthermore, rodents do not seem to be a suitable model to identify PCD therapeutics given a 100% failure rate in humans (52).However, numerous studies, including those using rodents, have used animal models to link different bacteria to PCD pathogenicity.This section summarizes some of the results of the literature listed in Table 1, though not all of the information in Table 1 is elaborated here.Some of these bacteria and their connections to PCDs are discussed in more detail in previous sections.One of the characteristics of all neurodegenerative PCDs is a deficiency in the cellular ability to maintain proteins that have a high propensity for aggregation in their folded, non-toxic state; as such, PCDs can be collectively referred to as disorders of proteostasis.Transgenic C. elegans that express fluorescent polyQ reporters exhibit quantifiable aggregates in response to proteotoxic stress (71,162,188,189).Using these strains as sensors capable of detecting changes in polyQ aggregation has allowed for the discovery of microbes and microbial products that affect cellular proteostasis (71,72,190).P. aeruginosa was among the microbes that robustly disrupted host proteostasis and led to polyQ aggregation (71).In addition to Pseudomonas being linked to PCDs in humans and a Dopaminergic-specific absence of TFAM results in recapitulation of PD phenotypes.MitoPark is a genetic model for PD in which dopaminergic neurons selectively lack the transcription factor TFAM (mitochondrial transcription factor A), culminating in the symptoms and phenotypes present in PD.
b PolyQ was used as a sensor of proteostasis but also models CAG repeat diseases such as HD, which are characterized by abnormally long stretches of glutamine "Q" repeats.c Bacterial strain was part of a probiotic cocktail that consisted of six bacterial strains: Bifidobacterium bifidum, Bifidobacterium longum, Lactobacillus rhamnosis, Lactobacillus rhamnosus, Lactobacillus plantarum, and Lactococcus lactis.d Bacterial strain was part of a probiotic cocktail that consisted of three bacterial strains: Lactobacillus reuteri, Lactobacillus rhamnosus, and Bifidobacterium infantis.e Bacterial strain was part of a probiotic cocktail that consisted of three bacterial strains: Lactobacillus plantarum, Lactobacillus fermentum, and Bifidobacteria longum.f Alzheimer's disease (AD), Parkinson's disease (PD), Huntington's disease (HD), or amyotrophic lateral sclerosis (ALS).Aβ, amyloid-beta; APP, amyloid precursor protein; BACE-1, beta-site amyloid precursor protein cleaving enzyme 1; FAD, familial Alzheimer disease; N/A, not applicable; polyQ, polyglutamine; PS1, presenilin-1; TFAM, mitochondrial transcription factor A. "Worm" is used in place of Caenorhabditis elegans; "Fly," Drosophila melanogaster; "Mouse," Mus musculus; and "Rat," Rattus norvegicus.
in a mouse model (145,171,(191)(192)(193)(194), Pseudomonas entomophila also aggravated disease-associated phenotypes in an AD fly model (163).Also consistent with the results obtained in worms demonstrating that Salmonella enterica Typhimurium disrupts host proteostasis, it was found that this microbe drove amyloid deposition in AD mice (172).
The agreement in the microbial contribution to PCDs between humans and animal models and within models is also evident for other bacteria.For example, in addition to the aforementioned role of P. mirabilis in PD patients, the organism also contributed to the pathogenicity of PCDs in animal models, including worms and mice (71,149).Pathogenic E. coli disrupted C. elegans proteostasis and caused disease-associated phenotypes in an HD fly and AD mouse models (71,164,165).In another study, a C. elegans polyQ model was used to show that E. faecalis disrupted host proteostasis as assessed by the enhancement of the polyQ aggregation (162).E. faecalis is normally part of the commensal microbiota, but in the event of gut dysbiosis, it can become enriched, compromise the intestinal barrier, and manifest in infection (195).As mentioned in the previous section, damage to the intestinal epithelium can enhance bacterial displace ment, which could explain a significant increase in microbial presence in AD patients' brains (24,196).Gut dysbiosis is implicated in a variety of diseases and is widely documented in people with PCDs (197).Numerous studies in humans and various animal models have used bacteria to induce or reverse gut dysbiosis to further examine the effect of microbial imbalance on neurodegenerative disease.For example, Erwinia carotovora was used to induce gut dysbiosis in AD flies to examine the effects of gut dysbiosis on AD phenotypes (163).The authors chose this bacterium due to the fact that it is non-pathogenic but can still persist in and colonize the gut.As such, they argued that the microbial dysbiosis caused by the bacterium aggravated AD phenotypes such that similar results were found using the pathogenic P. entomophila.Studies have also used bacteria to reverse gut dysbiosis, as consumption of probiotic bacteria is thought to help restore gut eubiosis, and Bifidobacteria and Lactobacillus are among common examples that are also known to affect cognitive function in humans (198,199).Lee et al. have shown that Bifidobacterium longum can alleviate cognitive defects in AD mice, and this positive effect was associated with its restoration of microbial symbiosis in the gut (159).B. longum has been used in a probiotic cocktail in two separate studies that demonstrated positive effects on AD and PD pathology in flies and mice, respectively (77,78).Additional evidence for the protective role of Bifidobacterium spp.comes from studies showing their suppressive effect against PD in flies and AD in rodents (77,(156)(157)(158)200).Other beneficial bacteria come from the Lactobacillus genus and are associated with positive outcomes in a PD mouse model (78), AD fly models (166,167), and an AD rat model (79).
While L. plantarum has the capacity to produce butyrate, a short-chain fatty acid that was shown to suppress protein aggregation, it is also known to enrich for butyrogenic bacteria in the swine (71,201,202).Butyrate was shown to restore gut eubiosis and convey neuroprotection in ALS mice (203).Walker et al. demonstrated that exogenous butyrate supplementation, as well as bacteria engineered to overproduce butyrate, can be beneficial to C. elegans proteostasis as assessed by reduced polyQ aggregation and the associated toxicity (71).In addition to L. plantarum, Bacillus subtilis was also shown to increase butyrate-producing bacteria (204).However, its role in protecting against neurotoxicity is likely more direct as its colonization of the C. elegans intestine increased lifespan, inhibited and reversed α-synuclein aggregation, and alleviated motility defects and other AD-associated phenotypes (154,155,205).Interestingly, B. subtilis is used to ferment soy products, such as Japanese natto, which is known for its neuroprotective properties (206,207).Akkermansia muciniphila has emerged as the sentinel microbe that promotes intestinal permeability and growth of butyrogenic bacteria (208).While the bacterium is known for its numerous health benefits, including attenuation of cognitive defects and Aβ plaques in AD rodent models, its presence in PCD patients, predomi nantly PD, correlates with the pathogenicity (152,153,209,210).Microbial communities rich in butyrate-producing bacteria, such as Agathobaculum butyriciproducens, have been associated with a diet that consists of healthy, plant-based foods (211).A. butyriciprodu cens has been shown to improve AD-and PD-associated phenotypes in mice (150,151).In general, butyrogenic bacteria, including Clostridium butyricum, have been associated with neuroprotective effects in animal models and humans (161,212,213).

POSSIBLE MECHANISMS BY WHICH BACTERIA CONTRIBUTE TO NEURODEGE NERATIVE PCDS
The precise mechanisms by which bacteria impact neurodegenerative PCDs are not known; however, based on well-studied host-microbe interactions, the affected pathways likely involve inflammatory responses and all aspects of host proteostasis ranging from protein synthesis to folding, trafficking, and degradation.The cross talk between inflammation and proteostasis can further exaggerate proteotoxicity where inflammation is known to disrupt proteostasis which in turn triggers inflammation (214).Furthermore, bacteria promote the production and release of host reactive oxygen and nitrogen species (RONS), both of which also trigger inflammation and disrupt proteo stasis (Fig. 2) (215,216).While the abovementioned responses are mainly triggered as part of the innate immune reaction to bacteria and, therefore, contribute to the disruption of host proteostasis non-specifically, other responses are microbe specific.For example, P. aeruginosa inhibits USP10, a host deubiquitinating enzyme involved in the clearance of cystic fibrosis transmembrane receptor, thus, affecting protein concentra tion and trafficking (217).P. aeruginosa modulates other host processes, including translation and protein folding (218,219).The influence of P. aeruginosa on broad host processes demonstrates multiple potential mechanisms by which bacteria can affect host proteostasis and ultimately the stability of proteins associated with PCDs.Other bacteria were also found to directly regulate stress responses in the host.For example, Chlamydia activates IRE1α, a major mediator of the endoplasmic reticulum unfolded protein response, which in turn activates a cascade of key inflammatory molecules, including two pathogen recognition receptors (PRRs), NOD1 and NOD2, and NF-κB that are known to affect PCDs (220)(221)(222).Various types of PRRs that respond to bacterial pathogens, including toll-like receptors, nucleotide-binding oligomerization domain-like receptors, and absent in melanoma-2-like receptors are also implemented in inflammation, likely contributing to the pathogenicity of PCDs (223).

CONCLUDING REMARKS
In this review, we summarized current research on the microbial contribution to neurodegenerative PCDs and propose that bacteria indiscriminately affect destabilized proteins that are encoded within its host proteome via direct and indirect mechanisms.As such, bacteria or their products will affect host proteins that are more prone to misfolding and aggregation.While certain bacterial species may more robustly affect protein folding homeostasis leading to proteotoxicity, the disease they contribute to solely depends on the extent of the metastable proteome and the capacity of the proteostasis network to buffer any destabilizing mutations.
Antibiotic stewardship is increasingly becoming the top priority for all One Health sectors (healthcare, agriculture, and the environment) because of the antibiotics' role in the emergence of AMR.However, due to limited data and only an emerging understand ing of the microbial pathogenesis in PCDs, no emphasis is put on antibiotic steward ship in terms of prevention and management strategies for these debilitating diseases.Increasing evidence suggests that antibiotics are not inert in terms of their role in the pathogenesis of PCDs, and prescribers should be aware of that when recommending or implementing antimicrobial treatment strategies, not only in PCD patients but also in elders, and the remaining population from all ages.
With the increasing prevalence of AMR, more bacteria become resistant, which, upon antibiotic treatment, will lead to increased enrichment of proteotoxic gram-negative species while at the same time eradicating the protective microbiota (Fig. 2).Therefore, AMR will inevitably affect the prevalence of PCDs.
The vast amount of the described published literature concentrates mainly on AD and PD.As such, we have a much better understanding of the microbial contribution to these two diseases than other PCDs.Based on our analysis of the literature, we believe that the sporadic forms of AD and PD, and perhaps other PCDs, could be preventable and possibly manageable by targeting the gut microbiota.

FIG 2
FIG 2 The effect of antibiotics on neurodegenerative protein conformational diseases.Antimicrobial agents kill commensal antibiotic-susceptible bacteria (AbS) that suppress proteotoxic antibiotic-resistant bacteria (AbR) that are thought to secrete products or signals that disrupt protein folding via the gut-brain axis.Bacteria can affect the pathogenesis of PCDs through direct or indirect interactions.Direct:Bacteria and their products can directly modulate host proteostasis (protein synthesis, folding, trafficking, or degradation).Indirect: Bacteria non-specifically induce inflammation and the generation of reactive oxygen and nitrogen species (RONS), which further disrupt host proteostasis.

TABLE 1
Summary of strain-specific effects on animal models of neurodegenerative diseases f (Continued on next page) October 2023 Volume 91 Issue 10 10.1128/iai.00437-228

TABLE 1
Summary of strain-specific effects on animal models of neurodegenerative diseases f (Continued)

TABLE 1
Summary of strain-specific effects on animal models of neurodegenerative diseases f (Continued)