The emerging role of the microbiome in Alzheimer’s disease

Abstract

Alzheimer’s disease (AD) is the most prevalent form of dementia and can be influenced by genetic and environmental factors. Recent studies suggest that the intestinal microbiota is altered in AD patients when compared to healthy individuals and may play a role in disease onset and progression. Aging is the greatest risk factor for AD, and age-related changes in the microbiota can affect processes that contribute to cognitive decline. The microbiota may affect AD by modulating peripheral and central immunity or by secreting factors that influence neurogenesis or neuronal cell death. Finally, probiotic and dietary interventions that target the microbiome may have therapeutic potential to prevent or treat AD.

1. Alzheimer’s disease

Alzheimer’s disease (AD) is a neurodegenerative disease characterized by the accumulation of misfolded proteins, neuronal cell death, and cognitive impairment (Hebert, Weuve, Scherr, & Evans, 2013; Alzheimer’s’s, 2013; Masters et al., 2015). Mutations associated with familial AD provide evidence that Aβ and tau contribute to the disease, however, 90% of AD is sporadic (Alzheimer’s’s, 2013), suggesting that there may be important environmental drivers of the disease. While altered genetics can increase the risk of developing AD (Tanzi, 2012), the largest risk factor is advanced age (Alzheimer’s Association, 2013). Recent studies suggest that age-related changes in the gut microbiota may contribute to immunologic and neurologic decline, raising the question of whether the aging microbiome contributes to AD (Fig. 1). Furthermore, studies have found that carrying genetic variants that increase the risk of developing AD may also lead to an altered gut microbiota, for example, the presence of the apolipoprotein E4 (Apoe4) allele (Tran et al., 2019), suggesting that there may be important environmental and genetic interactions between the gut microbiota and AD.

Fig. 1.

Many factors influence the gut microbiota that interact with host immunologic and metabolic processes and play a role in Alzheimer’s disease. These factors include aging, diet, use of antibiotics, geographical location, lifestyle, and sex.

2. The gut microbiota influence health and disease in aging

The intestinal microbiota encodes a metabolic capacity that can influence the host physiology (O’Hara & Shanahan, 2006). It is environmentally acquired, making it a modifiable therapeutic target. Its diverse functions may benefit the host, such as maintaining gastrointestinal health, producing vitamins, and promoting neuronal health (O’Hara & Shanahan, 2006), but disruption can result in disease (Cho et al., 2012). Altered microbiota, whether induced by diet (Turnbaugh, Backhed, Fulton, & Gordon, 2008), medication (Cox et al., 2014), hormone status (Markle et al., 2013), or host immune and metabolic status (Vijay-Kumar et al., 2010) can play an active role in disease. The microbiota influence the brain (Ochoa-Reparaz, Mielcarz, Begum- Haque, & Kasper, 2011) and interact with AD (Bekkering, Jafri, van Overveld, & Rijkers, 2013; Bhattacharjee & Lukiw, 2013), although precise mechanisms are still being actively investigated. Identification of key microbial changes could potentially have diagnostic or therapeutic value if successfully translated to humans.

In the elderly, age-related changes in host physiology and immunosenescence can destabilize the microbiota (Tiihonen, Ouwehand, & Rautonen, 2010). Because the incidence of AD increases with increasing age, these changes to the gut microbiota are believed to contribute to the risk of developing AD. Notably, there is an increase in intestinal permeability in aging (Tran & Greenwood-Van Meerveld, 2013) that leads to a higher exposure of the host to microbial products. This exposure may disrupt immune signaling pathways that protect against AD pathology. Certain lifestyle changes may prevent age-related changes to the gut. Long-term calorie-restriction substantially alters the gut microbiota and decreases gut permeability leading to lower circulating lipopolysaccharide-binding protein that indicates less translocation of pro-inflammatory bacterial products from the gut to the circulation (Zhang et al., 2013). This suggests that lifestyle modifications that improve AD may be partially mediated by controlling microbe-induced systemic inflammation.

3. The gut microbiota of patients with Alzheimer’s disease

Several studies have investigated the gut microbiome in patients with AD. It is important to note that the AD risk and the gut microbiota composition have a high degree of geographic variation, which may be shaped by the host genetics, lifestyle factors, diet, and indigenous microbes to that location (Gupta, Paul, & Dutta, 2017). Furthermore, lifestyle factors within a population, including mobility and living independently vs. living in a long-term care facility can affect microbiota composition. Thus, for human studies of the microbiome in AD, it is critical to obtain well-matched controls for both geography and lifestyle.

One of the earliest studies of the microbiome in AD investigated changes in 6 microbial groups by qPCR in an Italian cohort of 40 amyloid positive subjects with cognitive impairments, 33 amyloid negative subjects with cognitive impairment, and 10 healthy controls. They found that amyloid positive subjects had higher Escherichia/Shigella, and a lower abundance of Bacteroides fragilis and Eubacterium rectale compared to amyloid negative subjects (Cattaneo et al., 2017). They also found elevated levels of proinflammatory cytokines, including elevated IL-6, IL-1β, CXCL-2, and NLRP3 compared to both amyloid negative subjects and healthy controls, suggesting that microbial alterations were associated with increased markers of inflammation. In 2017, Vogt et al. investigated differences in the gut microbiome of 25 patients with AD and 25 age- and sex-matched healthy controls. The study participants were recruited in Wisconsin, USA, and were all home-dwelling with no significant differences in diet. According to 16S rRNA sequencing of stool samples, AD patients had reduced Firmicutes and Actinobacteria, and increased Bacteroidetes at the phylum level compared to the healthy controls. On the family level, bacteria that were increased in AD patient included Bacteroidaceae and Rikenellaceae and families less abundant included Ruminococcaceae and Bifidobacteriaceae. Furthermore, the levels of Bacteroides and Blautia correlated with a lower Aβ-42/Aβ-40 ratio, and higher pTau in cerebrospinal fluid of AD patients and controls (Vogt et al., 2017).

Haran et al. investigated the gut microbiome of elderly individuals residing in nursing homes in Massachusetts, USA. The study included a total of 108 participants diagnosed with AD, other dementia or no dementia and aimed to use a statistical model to predict AD diagnosis based on the gut microbial composition of the participants. Increased levels of Bacteroides ssp., together with Alistipes ssp., Odoribacter ssp., Barnesiella ssp. and lower Lachnoclostridium ssp. were found in AD patients compared to controls. A random forest classification algorithm was used to classify study participants into AD patients and healthy controls based on clinical variables and the abundance of bacterial species in stool. The 30 most important predictors for AD included the abundance of B. fragilis and other bacterial species measured in the stool samples, many of them identified as butyrate producers, in addition to frailty and malnutrition (Haran et al., 2019).

4. Antibiotics and Alzheimer’s disease

Antibiotics can have a profound effect on the gut microbiota and may have beneficial effects when targeting pathogens in a setting of infectious diseases or may have detrimental off-target effects. Disruption of the gut microbiota with antibiotics can reduce colonization resistance and increase the risk of infection with gastrointestinal pathogens (Kim, Covington, & Pamer, 2017) and has been linked to several chronic conditions, including obesity and metabolic syndrome (Cox & Blaser, 2015). The Nurses’ Health Study II examined changes in cognition with previous antibiotic use in a population of 14,542 women and found that long-term (> 2 months) antibiotic use in midlife (mean age of 55) was associated with decreases in cognition 7 years later even while controlling for diet, socioeconomic status, body mass index, smoking, exercise, use of anti-inflammatory medications and comorbidities (Mehta et al., 2022). This data suggests that microbiota disruptions in midlife, when AD pathogenesis is estimated to start, could be linked to later cognitive decline. While the mechanisms are unknown, long-term antibiotic use may increase the risk of cognitive decline by permanently depleting beneficial microbiota, decreasing neuroprotective functions of the gut microbiota including the production of neurotransmitters and modulating immune responses, or by favoring colonization by detrimental microbiota post antibiotics. One limitation that was acknowledged by the Nurses’ Health Study was that the data on cognitive decline could be stratified by antibiotic class, which can have markedly different effects on the gut microbiota.

An important consideration in the influence of antibiotics in cognitive decline is the reason for antibiotic use and the timing in relation to diagnosis. In the Nurses’ Health Study, all antibiotics were given to cognitively normal individuals the most common reasons were respiratory infection, urinary tract infection, acne/rosacea, dental, and chronic bronchitis. It is possible that an acute treatment for a specific infection may have a beneficial effect. A study of 6628 individuals in Taiwan found that individuals hospitalized for infections with Chlamydia pneumoniae had an increased risk of developing AD, while those with C. pneumoniae infections receiving a macrolide or fluoroquinolone for more than 15 days had a decreased risk of developing AD (Ou et al., 2021). Interestingly, in a pilot study of 10 individuals with mild to moderate probable AD dementia (mini-mental state examination (MMSE) score 17 ±3), 3-month antibiotic treatment with rifaximin reduced serum levels of phosphorylated tau and neurofilament light chain and increased levels of Erysipeloclostridium (Suhocki, Ronald, Diehl, Murdoch, & Doraiswamy, 2022). Rifaximin is a broad-spectrum antibiotic that is poorly absorbed from the gastrointestinal tract; thus, effects are likely to be limited to modulating the gut microbiota. On the other hand, a multicenter trial of 406 patients with mild to moderate AD (MMSE score 14–26) found that 12 months of treatment with rifampin or doxycycline was associated with significant decreases in cognition, as measured by the Standardized Alzheimer’s Disease Assessment Scale-Cognitive Subscale. Both rifampin and doxycycline are well absorbed from the GI tract, thus may influence microbial colonization at multiple sites, as well as directly affect host physiology.

The use of antibiotics for the treatment of infections in patients with severe AD and dementia may vary by country and clinical practice. For example, in the Netherlands, antibiotic treatment is commonly withheld from end-stage dementia patients (van der Steen, Ooms, Adèr, Ribbe, & van der Wal, 2002). In Canada, a survey of attitudes toward withholding antibiotics from individuals with advanced dementia from stakeholders, including senior citizens, family caregivers, nurses, and physicians, found that 75% of individuals supported withholding antibiotics from patients with advanced stage dementia and 98% of physicians supported withholding antibiotics from terminal stage dementia patients (Bravo, Van den Block, Downie, Arcand, & Trottier, 2021). Altogether, these studies highlight the need for a better understanding of antibiotics in several stages of AD, including assessing the degree to which chronic midlife antibiotics may be involved in the initiation of processes that lead to AD dementia, whether specific classes of antibiotics may have benefit for early- and mid-stage AD, and whether antibiotics can or should be used for end-stage AD.

5. The gut microbiota of murine models of Alzheimer’s disease

Microbiota differences have been reported in several animal models of AD. Many of these animal models carry transgenes coding for proteins involved in amyloid processing which lead to the overproduction of amyloid peptides and formation of Aβ plaques. Other models carry transgenes with mutated forms of tau leading to the formation of neurofibrillary tangles. Depending on the mutations and expression levels of the transgenes, the animal models may exhibit pathology at different time points. The experimental timeline of different models may be a critical consideration for modeling age-dependent changes in the gut microbiota, for that reason each model is discussed separately in the following description of changes to the gut microbiota to mice with experimental AD.

The Tg2576 model overexpresses a mutant form of APP with the Swedish mutation KM670/671NL and exhibits an age-dependent accumulation of Aβ plaques by 12 months of age (Hsiao et al., 1996). A study by Cox et al., investigated longitudinal changes in Tg2576 mice which typically develops plaques between 11 and 13 months of age (Cox et al., 2019). Because mice were singly housed, it was possible to measure microbiota drift over time in individual animals. The microbiota in aging Tg2576 females showed a substantial change in β-diversity from 5 to 15 months of age, which was not observed in littermate controls or in male mice. This suggests that the microbiota showed pronounced age-related changes in female AD mice. The largest age-related changes in female Tg2576 mice were an age-related decrease in Faecalibaculum rodentium, a member of the Erysipleotrichaceae family, and an age-related increase of Bacteroides, which correlated with the Aβ40 and AB42 levels found in the brain at 15 months of age, male mice did not have this association (Cox et al., 2019). However, another study by Honarpisheh et al. found lower levels of Bacteroidetes and higher levels of Firmicutes at the phylum level in Tg2576 mice at 15 months compared to 6 months of age and compared to wildtype controls (Honarpisheh et al., 2020). Differences in age-related microbiota changes in the same model may be due to sex-specific differences, as both sexes were used in a combined analysis in Honarpisheh et al., or may be due to animal facility dependent effects.

The APP/PS1-ΔE9 model carries a transgene inserted into chromosome 9, which has a chimeric mouse/human APP Mo/HuAPP695swe and PSEN1 with a deletion of exon 9 (Goodwin et al., 2019) and shows the first signs of Aβ deposition at 6 months of age (Jankowsky et al., 2004). The APP/PS1-21, also known as the APP/PS1-Jucker mouse model of Alzheimer’s disease expresses human transgenes for both APP with the Swedish mutation and PSEN1 with an L166P mutation under the control of the Thy1 promoter. The expression of APP is approximately three times higher than the endogenous APP expression in these mice leading to the formation of amyloid plaques starting around 1.5 months of age (Radde et al., 2006). Harach et al. analyzed the microbial composition of fecal pellets from APP/PS1 mice at 1 month (before amyloid pathology), 3.5 months (early pathology) and 8 months (late pathology) of age by sequencing the 16S rRNA gene. They reported that while the 1-month-old APP/PS1 mice had a microbial composition of their gut that was similar to wildtype C57BL/6 mice, aging APP/PS1 mice had an increasing fraction of Bacteroidetes and decreasing levels of Firmicutes in the stool (Harach et al., 2017). Furthermore, Cox et al. found that administering Bacteroides to APP/PS1-21 female mice by weekly oral gavages between 2.5 and 5 months of age, increased plaque burden in the cortex compared to controls (Cox et al., 2019). Chen et al., on the other hand, found an age-related decrease Bacteroidetes in APP/PS1 mice vs. non-transgenic controls 1 and 9 months of age. APP/PS1 mice also had higher levels of Verrucomicrobia, Actinobacteria, and Proteobacteria. Of note, several of the changes in the gut microbiota were detected at early time points before changes in AD pathology of changes in microglia were observed (Chen et al., 2020). Bäuerl et al. analyzed the gut microbiome of APP/PS1 mice at 3, 6 and 24 months of age and reported that Turicibacteriaceae and Rikenellaceae were decreased in aging wildtype. In addition, APP/PS1 mice had increased Sutterella and Erysipelotrichaceae in aging and also compared to wildtype mice (Bauerl, Collado, Diaz Cuevas, Vina, & Perez Martinez, 2018). Sutterella is a bacteria associated with autism disorder and Down Syndrome (Biagi et al., 2014; De Angelis et al., 2013) and the Erysipelotrichaceae family is highly immunogenic and linked to inflammatory bowel disease (Schaubeck et al., 2016), colorectal cancer (Chen, Liu, Ling, Tong, & Xiang, 2012), and Th17 responses in an animal model of multiple sclerosis (Miyauchi et al., 2020). Shen et al. reported higher levels of Prevotella in APP/PS1 mice compared to wildtype mice, and increasing levels of Prevotellaceae in 6- and 8-month-old mice compared to APP/PS1 mice at 3 months of age (Shen, Liu, & Ji, 2017). Zhang et al. found no difference in Bacteroides between APP/PS1 mice that were 3 and 9 months of age, differences on phylum level were not analyzed in this study (Zhang et al., 2021). Zhang et al. found that aging APP/PS1-ΔE9 mice had an increased ratio of Firmicutes/Bacteroidetes between 3 and 12 months of age (Zhang et al., 2017). Furthermore, the level of Parabacteroides distasonis was lower in aging wildtype and APP/PS1-ΔE9 mice. Cuervo-Zanatta et al. found that in APP/PS1-ΔE9 mice, female mice had higher Bacteroidetes than male mice (Cuervo-Zanatta, Garcia-Mena, & Perez-Cruz, 2021). Another study investigating the APP/PS1-ΔE9 strain concluded that the abundance of Bacteroides and Lactobacillus correlated with the age of the mice and were mostly found in mice that were 12 months of age (Li, Zhu, Guo, Du, & Qin, 2020). Furthermore, Harach et al. also demonstrated that APP/PS1 mice raised in germ-free conditions had decreased amyloid peptides levels in their brains and fewer amyloid plaques.

The 5xFAD mice express human APP with the Swedish, Florida and London mutations and human PS1 with the M146L and L286V mutations, also under the Thy1 promoter, and develop plaques around 2 months of age (Oakley et al., 2006). Brandscheid et al. reported higher levels of Bacteroidetes between 1.5 and 4 months of age (Brandscheid et al., 2017), but Shukla et al. reported lower levels between 5 and 15 months of age although levels of Bacteroidetes was higher than wildtype controls at both timepoints (Shukla et al., 2021). Shukla et al. also reported increasing levels of Firmicutes and Bifidobacterium bifidum in older mice, although levels were lower in 5xFAD mice compared to wildtype mice. Furthermore, Tran et al. compared the gut microbiome of 5xFAD mice that has been crossed with APOE target replacement mice to express human APOE3 and APOE4, called E3FAD and E4FAD, respectively. EFAD mice develop cognitive impairment at 1–2 months and amyloid plaques around 4 months of age. At 4 months of age, E4FAD mice had lower levels of Bacteroides than E3FAD mice and a similar trend was seen in mice at 18 months. The E4FAD mice also had lower levels of Alistipes and Johnsonella at 4 months and higher levels of six other genera at either 4 or 18 months including Desulfovibrio and Butyricicoccus (Tran et al., 2019). Maldonado et al. reported lower levels of Prevotella Sutterella and Ruminococcus, higher levels of Anaeroplasma and no difference in Bacteroides between 4-month-old E3FAD and E4FAD mice (Maldonado Weng et al., 2019). Parikh et al. found differences in the levels of Erysipelotrichaceae, Lactobacillaceae, Ruminococcaceae and Rikencellaceae between mice with APOE2, −3 or −4 (Parikh et al., 2020).

The 3xTg-AD mouse model harbor human transgenes of APP with the Swedish mutation and PS1 with the M146V mutation but also tau with the P301L mutation, meaning that these mice develop both amyloid and tau pathology (Oddo et al., 2003). This model develops amyloid plaques at 6 months of age and tau neurofibrillary tangles at 12 months. Bello-Medina et al. investigated changes in the microbiota of wild type B6129SF1/J mice at 3- vs. 5-months of age and found that both male and female transgenic mice showed increased levels of minor gut microbial populations, including phyla Fusobacteria and Cyanobacteria in comparison to non-transgenic mice (Bello-Medina et al., 2021). Transgenic female mice showed a decrease in the phylum Actinobacteria in aging which was not observed in wild type mice, and this bacterium has been linked to anti-inflammatory properties as well as a decreased gut permeability. Non-transgenic male mice showed an increase in Firmicutes and Actinobacteria, while also showing a decrease in Proteobacteria, Fusobacteria, Cyanobacteria, and Bacteroidetes. Overall, this study suggested that changes in the gut microbiota could be observed in models of AD that precede the development of cognitive symptoms (Bello-Medina et al., 2021).

Taken together, no specific species or group of bacteria has been consistently reported to have a higher or lower abundance in the gut of AD mouse models. As described above, a number of studies has reported higher levels of Bacteroidetes or Bacteroides in aging AD mice, but there are also reports indicating the opposite direction. However, a detrimental role of Bacteroides was supported by the report of higher levels of amyloid plaques found in mice administered with B. fragilis, supporting a direct relationship between these bacteria and amyloid pathology (Cox et al., 2019). There can be many reasons why studies come to different conclusions with regard to which bacteria are associated with amyloid and tau pathology in mice. We have discussed that the timing for microbiome collection could be an important factor as these mice develop pathology at different ages. In addition, the gut microbiota is sensitive to the environment, and may be affected by the facility that the mice were bred and raised in. Another important factor is that most studies are relying on sequencing of specific regions of the 16S rRNA gene, and there might be important genetic differences in these bacteria on species and strain levels that this methodology is not designed to find. To support this notion, Cox et al. reported that while the abundance of Akkermansia is generally associated with multiple sclerosis (MS), administration of MS-derived strains of Akkermansia to mice with experimental MS led to ameliorated disease in a strain-specific manner (Cox et al., 2021). Therefore, it is important to focus future research on functional aspects of the gut microbiota that may be linked to AD.

5.1. The role of the microbiota in modifying Aβ pathology in animal models

While several studies provide evidence that the gut microbiota shows age- and AD-dependent alterations in several animal models, microbiota changes may be a consequence of disease-specific changes in host physiology. To test whether there is a causal relationship between the microbiota and AD, several groups have investigated the effect of microbiota interventions, including depletion of the microbiota, addition of whole communities or specific bacteria, or dietary interventions that may shape the gut microbiota.

5.2. Depleting the gut microbiome alleviates amyloid and tau pathology

To test whether the gut microbiota was necessary for the development of Aβ plaques, researchers have utilized germ-free mice and antibiotics to deplete the gut microbiota. Germ-free APP/PS1 mice showed reduced amyloid plaque deposition, less neuroinflammation, and increased levels of Aβ-degrading enzymes compared to conventionally raised mice (Harach et al., 2017). Furthermore, the colonization of 4-month-old germ-free APP/PS1 mice with gut microbiota from 12-week-old conventionally raised wildtype or APP/PS1 mice for 8 weeks resulted in increased Aβ pathology (Harach et al., 2017), suggesting that the gut microbiota is both necessary and sufficient for the promotion of amyloid deposition. However, germ-free mice have substantially altered development and physiology, including immune and neurologic functions.

To address whether depleting the microbiota after development could affect AD pathology, several studies have investigated the effects of antibiotics in AD animal models. Minter et al. administered a combination of eight antibiotics given at a moderate to low dose to modulate the microbiome without inducing substantial host toxicity from long-term use of antibiotics. Strikingly, only male mice showed reduced Aβ plaque accumulation, which was linked to a reduction in proinflammatory cytokines in the serum (Minter et al., 2016). This low-dose antibiotic treatment did not entirely deplete the microbiota but was associated with a shift in microbiota composition over time. Further work from the same group found that antibiotics limited to early-life was sufficient to decrease plaque development in adult mice (Minter et al., 2017), and that the full combination of eight antibiotics was necessary for the effect, rather than single antibiotics (Dodiya et al., 2020). Altogether, these studies suggest that the microbiome can contribute to amyloid deposition, which may be influenced by antibiotic choice, the timing of administration, and sex-specific effects.

5.3. Fecal microbiota transfers reveal an important link between the gut microbial composition and memory function

There is a bidirectional connection between the gut microbiota and AD-like pathology and the influence of the gut microbiota on the brain has been directly investigated by transferring fecal matter between wildtype healthy mice and AD mouse models. A transfer of wildtype fecal microbiome to 6-month-old APPswe/PS1-ΔE9 mice decreased amyloid levels in the brain and improved spatial learning and recognition memory (Sun et al., 2019). The reduced plaque levels were associated with an increased abundance of Bacteroidetes, reduced abundance of Verrucomicrobia and increased butyrate production (Sun et al., 2019). Transfer of wildtype gut microbiome to ADLP^APT also led to reduced levels of amyloid plaques in addition to phosphorylated tau and improved spatial memory (Kim et al., 2020). Interestingly, when the microbiota of 9-month-old 5xFAD mice were transferred to wildtype mice, the 5XFAD microbiota recipient wildtype mice demonstrated reduced spatial memory and reduced neurogenesis in the hippocampus (Kim et al., 2021). Furthermore, colonizing germ-free APP/PS1 with the microbiome of aged conventionally raised wildtype and APP/PS1 mice increased the accumulation of amyloid peptides in the brains of the mice regardless of genotype, although mice colonized with the microbiome from APP/PS1 mice had higher levels than those colonized with the microbiome of wildtype mice (Harach et al., 2017), suggesting that the overproduction of Aβ could select for a microbiota that promotes plaque deposition.

6. The gut microbiota and the immune system in Alzheimer’s disease

6.1. The gut microbiome modulates peripheral immunity

There is substantial evidence that both peripheral immune activation and immunosenescence can play a role in AD (Fig. 2). Studies have found that patients with AD have increased levels of circulating proinflammatory cytokines (Licastro et al., 2000). These proinflammatory cytokines may be produced by peripheral immune cells or microglia and may lead to disruptions in the blood-brain barrier, leading to reduced expression of tight junction proteins and increased permeability to serum proteins, reviewed in (Desai, Monahan, Carvey, & Hendey, 2007).

Fig. 2.

The microbiome modulates peripheral and central immunity and immunosenescence to exacerbate AD pathogenesis. Three potential mechanisms by which the microbiome is able to modulate AD is through peripheral immunity, central immunity, and immunosenescence. The microbiome interacts with the peripheral and central immune systems through the secretion of metabolites including short-chain fatty acids (SCFAs), bile acids, polysaccharides, and bacterial components such as toxins. In the periphery, these microbial products can act to enhance or hinder the recruitment of monocytes to the brain to participate in phagocytosis of Aβ plaques. These microbial products can also promote systemic pro-inflammatory cytokine release, including tumor necrosis factor a (TNFa) and interleukins (IL), and promote the senescence-associated secretory phenotype (SASP) in immune cells. Microbial derived products, monocytes, and cytokines can then interact with brain immunity. The overall effect is the immune system’s inability to clear plaques and resolve neuroinflammatory insults leading to exacerbated AD pathology.

Peripheral immune cells may also have protective functions in AD. Monocytes can be recruited from the periphery to the central nervous system and play an important role in phagocytosing Aβ plaques. Peripheral monocytes from AD subjects have been reported to have impaired Aβ phagocytosis, suggesting a loss of innate immune function involved in Aβ clearance (Fiala et al., 2005). Nasally administering a proteosome adjuvant, derived from Gram-negative bacterial cell walls, leads to the activation of the peripheral innate immune system and improved clearance of Aβ in mice (Weiner & Frenkel, 2006). This is also associated with upregulating CCL2 and CCR2, critical factors that aid in monocyte trafficking to the brain (El Khoury et al., 2007). Thus, immunostimulatory products from the microbiota may play an important role in maintaining innate immune function that can prevent processes associated with AD. Furthermore, regulatory T cells have been shown to support microglia clearance of amyloid plaques (Dansokho et al., 2016).

Age- and AD-related immune changes may be affected by intestinal bacteria. Several studies demonstrate that the gut microbiota can modulate immune function during development and throughout the lifespan. Furthermore, gastrointestinal infections have been associated with an increased risk for dementia (Fink, Doblhammer, & Tamgüney, 2021).

6.2. The gut microbiome is important for microglia function

Microglia are the primary immune cell of the brain and play an important role in AD. In both mice and humans, microglia function diminishes with age due to changes in phagocytotic ability, morphology, activity, and even the production of toxic chemicals. In AD, microglia undergo functional change that contribute to neuroinflammation, which results in increased Aβ plaque deposition and neural atrophy.

Studies using germ-free or antibiotic-treated mice demonstrate that the gut microbiota affect microglia in both wildtype animals and models of AD. Germ-free mice have fewer microglia and fewer stromal choroid plexus macrophages than specific pathogen-free (SPF) mice (Erny et al., 2015; Sankowski et al., 2021). The depletion of the gut microbiota by administration of antibiotics to young male APP/PS1 mice did not reduce microglia numbers, but caused morphological and transcriptional changes associated with an inactive and immature state of microglia that could be reversed by fecal microbiota transfers from either wildtype or APP/PS1 mice (Dodiya et al., 2022). The mechanism by which the fecal matter transfer restores amyloidosis in mice with a depleted microbiota is unclear, but administration of short-chain fatty acids (SCFAs) resulted in higher microglia numbers in germ-free mice (Colombo et al., 2021), suggesting that microbial metabolites play an important role in supporting the microglia population in the brain. Furthermore, it was determined that the reduction in Aβ-plaques observed in antibiotics-treated APP/PS1 mice was diminished when microglia was depleted by an CSF-1R inhibitor (Dodiya et al., 2022), demonstrating that microglia play a critical role in mediating the effect of gut microbiota interventions on AD pathology. Similar observations were made in antibiotic-treated or germ-free 5xFAD mice. Germ-free mice, but not antibiotics-treated mice, had reduced microglia numbers (Mezö et al., 2020). Furthermore, microglia isolated from germ-free 5xFAD mice had higher expression of genes involved in phagosome maturation and were more efficient at phagocytosis ex vivo, while no difference was observed in microglia from antibiotics-treated mice. This observation suggests that lower plaque burden in germ-free 5xFAD mice may be the result of improved Aβ phagocytosis by microglia. Interestingly, these differences were only observed in microglia from relatively young animals (4 months old). At an older age (10 months old), no differences were observed in total microglia numbers or numbers of microglia that phagocyte Aβ, suggesting that long-term exposure to Aβ makes the microglia less responsive to gut microbiota modulations.

6.3. The gut microbiome modulates metabolic pathways in microglia

Cellular metabolic pathways regulate immune cell function (O’Neill, Kishton, & Rathmell, 2016). The gut microbiome and immune system are inextricably linked and one source of their connection is through microbial metabolites altering the metabolism and function of immune cells (Michaudel & Sokol, 2020).

Microglia, which have been widely implicated in their role in AD, are metabolically plastic and dependent on metabolic reprogramming for their function (Bernier, York, & MacVicar, 2020). A switch from mitochondrial oxidative phosphorylation to glycolysis in microglia contributes to AD pathology by impairing phagocytosis of Aβ (Pan, Ma, & Kong, 2019). Studies have shown that inflammatory stimuli from bacterial product may alter metabolic function in microglia. For example, stimulation of microglia with Aβ in combination with LPS or IFNγ led to increased glycolysis and a reduced ability to phagocytize Aβ in vitro (McIntosh et al., 2019; Rubio-Araiz, Finucane, Keogh, & Lynch, 2018). Another study found that Aβ alone was sufficient to induce reprogramming of microglia from oxidative phosphorylation to glycolysis in an mTOR-HIF1α dependent pathway (Baik et al., 2019). Increased glycolysis was also found in microglia isolated from APP/PS1 mice compared to wildtype mice (Holland et al., 2018).

The gut microbiome derived SCFA butyrate polarizes macrophages to an M2 phenotype by promoting a shift of the metabolism toward oxidative phosphorylation (Scott et al., 2018). According to a network-based ranking algorithm that were developed to investigate interactions between gut microbiome metabolites, microglia and AD, SCFAs was identified as the top microbial metabolites modulating microglia in the context of AD (Wang et al., 2021). Indeed, microbially derived SCFAs supplementation is sufficient to reverse the immature phenotype of microglia observed in germ-free mice (Erny et al., 2015). The mechanism by which SCFAs modulate microglia was determined to be through metabolic reprogramming of microglia metabolism. In the absence of the gut microbiome-derived SCFAs, there is modulations of histone acetylation and methylation in the proximity of promoter regions of genes that regulate metabolic functions in microglia. Bacteria-derived acetate was also shown to participate in the tricarboxylic acid cycle of microglia, and acetate-treated and SPF mice had lower levels of oxaloacetate leading to higher activity of complex II of the electron transport chain (Erny et al., 2021). It was further shown that acetate, but not butyrate or propionate, promoted microglia maturation in germ-free mice, and acetate-treated mice had impaired microglia phagocytosis of Aβ leading to higher plaque burden in germ-free 5xFAD mice.

Immunometabolism and its connection with the gut microbiome offer insight into potential explanations for sex dimorphism in AD. It is known that the gut microbiome and immune axis is different between males and females, especially in the context of AD (Cox et al., 2019). Immune cells and their metabolism are sex dimorphic in numerous metabolic pathways (Lee, Profant, & Wang, 2022). In APP/PS1 mice female and male microglia are distinct in their transcriptome and these differences can be mapped to metabolic genes. In functional assays, female microglia showed increased glycolytic flux and reduced phagocytosis suggesting that host sex is determinant in microglial metabolism and subsequent function (Guillot-Sestier et al., 2021). Several studies have begun to elucidate parts of this pathway. Thion et al. found that the microbiome contributes to the sex dimorphic development of microglia beginning in the prenatal stage (Thion et al., 2018). This establishes that microglial development is sex-specific and regulated by the gut microbiome. Rothhammer et al. found that type I interferons along with microbial metabolites modulate neuroinflammation (Rothhammer et al., 2016). Collectively, these findings suggest that sexual dimorphism in AD is present in the gut microbiome and contributes to altered immune cell metabolism.

7. Interactions between the diet, the microbiota, and Alzheimer’s disease

The intestinal microbiota is shaped by diet among other factors and in turn, these microbes and their metabolites significantly influence gut health. Dietary patterns have been shown to influence AD pathology. There is a plethora of evidence of the neurocognitive improving actions of various diets including ketogenic and Mediterranean, as well as the consumption of prebiotics such as polyphenols in individuals and mouse models. However, there is a need to investigate the effects of these interventions on neurodegeneration biomarkers, especially in AD and dementia.

Studies have found that calorie restriction may have beneficial effects in aging and Alzheimer’s disease. In mice, a 30% reduction in calorie intake in Tg2576 prior to plaque development led to improved glucose tolerance, reduced Aβ peptides and reduced amyloid plaque burden (Cox et al., 2019; Wang et al., 2005). As the mice aged, calorie restriction counteracted the increase in Bacteroides and the reduction in Faecalibaculum that were observed in Tg2576 mice fed ad libitum. Furthermore, the abundance of Bacteroides in feces correlated with higher levels of Aβ peptides in the brain and Faecalibaculum correlated with lower levels (Cox et al., 2019). In APPswe/PS1ΔE9 mice, calorie restriction led to improved glucose metabolism, improved spatial learning and lower amyloid plaque levels, which was associated with improved autophagy (Müller et al., 2021).

The Mediterranean diet plan supports the high intake of meals rich in vegetables, legumes, fruits, cereals, fish, and unsaturated fatty acids with low consumption of dairy products, and processed red meat. Most of the supported foods are rich in dietary fiber and polyphenols which are known for their numerous health benefits, including anti-inflammatory, anti-apoptotic and antioxidant action. Interestingly, adherence to the Mediterranean diet attenuated disease phenotype, characterized by decreased amyloidosis, decreased mediotemporal atrophy, and improved memory in participants of the German Longitudinal Cognitive Impairment and Dementia Study (Ballarini et al., 2021). A modified Mediterranean-ketogenic diet has been shown to modulate the gut microbiome and mycobiome in human subjects with MCI (Nagpal et al., 2020; Nagpal, Neth, Wang, Craft, & Yadav, 2019).

AD is often characterized by impaired brain metabolism and cognitive function. A positive correlation between physiological ketosis and improved cognitive function in the AD population has been suggested. The consumption of high-fat and low-carbohydrate food, called the ketogenic diet (KD), promotes the formation of ketone bodies such as acetone, beta-hydroxybutyrate, and acetoacetate which are excellent fuel sources (Altayyar, Nasser, Thomopoulos, & Bruneau Jr., 2022). The use of these substrates can alleviate impaired glucose metabolism often associated with AD (Phillips et al., 2021). Although the mechanisms through which KD may improve AD are not fully understood, the anti-inflammatory and antioxidant potential of ketones have been reported (McDonald & Cervenka, 2018).

The ketogenic diet has been shown to modify the gut microbiota in humans and mice (Park, Zhang, Wu, & Yi Qiu, 2020; Ang et al., 2020; Olson et al., 2018). The shift in microbial profiles in KD-fed over-weight/class I obese men was attributed to ketone bodies produced by the host. FMT from KD-fed human donors decreased the population of Th17 cells in the gut and b-hydroxybutyrate selectively repressed the growth of Bifidobacteria in germ-free mice (Ang et al., 2020). The neuroprotective role and cognitive benefits of KD in AD including clinical trials have been extensively reviewed here (Hersant & Grossberg, 2022; Rusek, Pluta, Ulamek-Koziol, & Czuczwar, 2019). However, these studies did not assess the microbial profiles of the intervened population. In mice, KD led to worsened disease phenotype characterized by gut dysbiosis with an elevated abundance of the phylum Proteobacteria, increased neuroinflammation, and glucose intolerance (Park et al., 2020). Other studies have demonstrated that the antiseizure effect of KD in mice is gut microbiota-dependent and increased the abundance of Akkermansia muciniphila and Parabacteroides spp (Olson et al., 2018).

Polyphenols are bioactive compounds produced by various plants such as fruits, legumes, herbs, and teas. Phenolic compounds are known for their evident health benefits including autoinflammatory, antioxidant, anti-microbial, and radical scavenging action, and have been considered potential therapeutics for diverse health conditions. Phenolic compounds often possess low bioavailability and are transformed into metabolites by the intestinal microbiota. These metabolites are released into the systemic circulation. In patients with mild cognitive decline, grape juice formulation, rich in phenols offered protection against cerebral metabolic decline and improved cognitive features such as attention and working memory (Lee, Torosyan, & Silverman, 2017) Similar improvements were reported for chlorogenic acids richly found in coffee beans administered to MCI patients in a randomized controlled crossover trial (Ochiai, Saitou, Suzukamo, Osaki, & Asada, 2019). Further, in a placebo-controlled, double-blind, randomized clinical trial, a homogenized mixture of fruits and vegetables improved neurocognitive patterns in a healthy non-elderly population (Carrillo, Arcusa, Zafrilla, & Marhuenda, 2021).

Trimethylamine (TMA) is a gut microbiota metabolite derived from dietary choline and carnitine and can be oxidized to Trimethylamine-N-oxide (TMAO) in the liver. The latter has been implicated in the pathophysiology of AD. In a recent study to investigate the connection between AD pathology and TMAO levels in the cerebrospinal fluid (CSF) of individuals with Alzheimer’s clinical syndrome and those with mild cognitive impairment, elevated CSF TMAO levels were reported in both groups, compared to healthy control (Vogt et al., 2018). Additionally, elevated TMAO levels correlated with disease phenotype as characterized by AD biomarkers(Vogt et al., 2018). Similarly, elevated levels of TMAO were observed in urine samples from human AD and patients with mild cognitive impairment compared to healthy controls (Yilmaz et al., 2020). 3xTg-AD mice exhibited elevated levels of TMAO in an age-dependent manner and these mice showed less synaptic strength (Govindarajulu et al., 2020). Importantly, TMAO impaired synaptic transmission in hippocampal tissues retrieved from healthy mice and treated with TMAO and upregulated PKR-like endoplasmic reticulum kinase (PERK) pathway and consequently endoplasmic reticulum stress signaling (Govindarajulu et al., 2020). Overall, the mechanisms through which TMAO may influence AD pathology include protein aggregation (Aβ, tau), endoplasmic reticulum stress as well as microglial activation (Arrona Cardoza, Spillane, & Morales Marroquin, 2022). Intriguingly, attempts to lower TMAO production have shown promising therapeutic effects in mice. Coadministration of Lactobacillus plantarum and memantine decreased Aβ plaques in AD mice and cognitive decline, this was associated with decreased plasma TMAO levels (Wang et al., 2020). Similar interventions have been demonstrated here (Gao et al., 2019; Li et al., 2019; Liu et al., 2020).

8. Prebiotics

Prebiotics are dietary compounds that are not fully digested by the host and serve as growth substrates for beneficial microbiota. These can be a relatively low-risk intervention to modulate the gut microbiota, but the efficacy depends on having beneficial microbes already present in the microbiota, albeit at low levels.

8.1. Sodium oligomannate

Oligosaccharides, indigestible carbohydrates, are classical examples of prebiotics and are fermented by the gut microbiota. Sodium oligomannate, (GV-971) a derivative of marine algae has shown great therapeutic potential (Xiao et al., 2021). In a recently concluded phase III trial for mild for moderate AD/dementia, with 818 participants (placebo n = 410, GV-971 n = 408), GV-971 improved cognitive performance (Xiao et al., 2021). Elevated levels of amino acids such as phenylalanine and isoleucine have been associated with gut dysbiosis in AD patients and consequently proinflammatory responses in the brain (Wang et al., 2019). In AD mice, GV-971 countered cognitive impairment and neuroinflammation, by reconstituting the gut microbiota composition and decreasing Th1 cells in the brain. They also reported a reduction in the levels of phenylalanine and isoleucine in the blood and stool post-treatment (Wang et al., 2019).

8.2. Dietary inulin

Dietary inulin is another prebiotic with great health benefits. In APOE4 mouse models, E3FAD and E4FAD mice, inulin decreased hippocampal inflammatory gene expression, and increased metabolism characterized by increased abundance of SCFAs, tryptophan-derived metabolites and bile acids (Hoffman et al., 2019; Yanckello et al., 2021). This enhanced metabolism could be attributed to gut microbial alterations as inulin increased the abundance of taxa such as Prevotella and Lactobacillus but decreased the relative abundance of Escherichia and Turicibacter spp (Hoffman et al., 2019; Yanckello et al., 2021).

9. Microbial mediators associated with Alzheimer’s disease

The gut microbiota is estimated to contain as many cells as the human body and is estimated to carry 100 times the number of genes as the human genome. Along with this, the species within the gut microbiota can secrete a wide variety of neuromodulatory and immunomodulatory substances, may influence AD by direct or indirect mechanisms (Fig. 2).

9.1. Short-chain fatty acids (SCFAs)

We previously stated that germ-free mice have reduced plaque burden and it has been suggested that reduced plasma levels of SCFAs are responsible for this effect (Colombo et al., 2021). Supplementing germ-free mice with SCFA through drinking water significantly increased Aβ pathology in APP/PS1 mice. This difference was determined not to be due to a change in Aβ production, but instead to the deposition and clearance of the plaques, which is linked to the function of microglia. The introduction of SCFAs to germ-free and SPF mice was noted to induct a more active shape of microglia; however, this was not associated with an increased microglial phagocytic capacity toward Aβ, but rather promoted microglia activation marked by increased expression of TREM2. Other studies have reported an immunosuppressive role of SCFAs. Administration of acetate to APP/PS1 mice led to improved spatial learning by binding to the Free Fatty Acid Receptor 3 (FFAR3), inhibiting the ERK/JNK/NF-κB pathway and suppressing the production of pro-inflammatory cytokines (Liu et al., 2020). This anti-inflammatory effect of SCFA was also demonstrated in a cell culture model of human microglia-like cells, in which it also suppressed phagocytic activity (Wenzel, Gates, Ranger, & Klegeris, 2020). It was further demonstrated that propionate can protect blood-brain barrier integrity from LPS insults using an in vitro system to simulate the blood-brain barrier (Hoyles et al., 2018).

9.2. Bile acids

Bile acids (BAs) are the main metabolic end products of cholesterol metabolism. Primary bile acids such as cholic acid and chenodeoxycholic acid are produced in the liver while secondary bile acids such as deoxycholic acid and lithocholic acid are synthesized by colonic gut resident communities (Zeng, Umar, Rust, Lazarova, & Bordonaro, 2019). Both primary and secondary bile acids have been implicated as important contributors to AD pathophysiology. Importantly, altered BAs synthesis and metabolism in AD patients have been highlighted. Transcriptomic analysis of 2114 postmortem brains of AD and HC subjects revealed prevalent expression of BA synthesis pathways genes (Baloni et al., 2020). Targeted metabolomic profiling of these brain samples showed higher ratios of primary and secondary BAs in AD subjects compared to controls, suggesting a correlation between BAs and cognitive deficits in AD (Baloni et al., 2020). Additionally, the detection of secondary BAs in the brain suggests a potential influence of the gut microbiota(Baloni et al., 2020). Similar profiling of sera from subjects enrolled in the AD Neuroimaging Initiative revealed a significant association between altered serum-based BA with the Aβ, tau and neurodegeneration biomarkers for AD (Nho et al., 2019). Also, impaired glucose metabolism and larger atrophy of the brain were associated with the altered BA profiles (Nho et al., 2019). In a separate study, the bile acid profiles of AD patients indicated that increased bacterial 7α-dehydroxylation of citric acid into deoxycholic acid was associated with cognitive decline and immune-related AD-risk gene variants (MahmoudianDehkordi et al., 2019). Mass spectrometry profiling of the livers of APP/PS1 and App^NL-G-F male and female mice showed decreased cholesterol metabolism and primary bile acid synthesis compared to controls, but no difference in secondary bile acid production was observed (Kaur, Seeger, Golovko, Golovko, & Combs, 2021). Sex-specific differences were also reported, with females showing higher levels of cholesterol (Kaur et al., 2021). Importantly, genotype-associated bile acid alteration was reported as male APP/PS1 unlike App^NL-G-F, showed a selective reduction of the secondary bile acid taurourosdeoxycholic compared to controls (Kaur et al., 2021).

9.3. Polysaccharides

Some B. fragilis strains produce factors that are beneficial for non-AD neurologic diseases, including polysaccharide A (PSA), a cell wall glycan that induces T-regulatory cells (Tregs) (Mazmanian, Round, & Kasper, 2008). B. fragilis packages PSA and other immunomodulatory cargo in outer membrane vesicles (OMVs) that enter the circulation, travel to the brain and decrease inflammation in the EAE animal model of multiple sclerosis (Ochoa-Reparaz et al., 2010). Although Tregs are beneficial in MS, they may play a detrimental role in AD by inhibiting Aβ phagocytosis by infiltrating monocytes and microglia (Baruch et al., 2015), highlighting the importance of understanding disease-specific interactions.

9.4. Toxins

Some Bacteroides strains carry the pathogenicity factor Bacteroides fragilysin toxin (BFT), a matrix metalloprotease that increases gut barrier permeability and leads to colitis (Wu, Dreyfus, Tzianabos, Hayashi, & Sears, 2002). BFT interacts with Psen1 (Wu, Rhee, Zhang, Franco, & Sears, 2007), a γ-secretase that cleaves amyloid-precursor protein (APP). Further studies are needed to determine if BFT affects γ-secretase activity in the brain. In addition to the gut microbiota, the oral microbiota may also contribute to AD. One study has found the Gingipains, toxins from the oral microbe Porphyromonas gingivalis in the brain of AD patients (Dominy et al., 2019), suggesting that there may be multiple important routes for microbial toxin infiltration.

10. Potential use of probiotics for the treatment of AD

10.1. Probiotic interventions can improve memory in models of Alzheimer’s disease

A formulation of lactic acid bacteria and Bifidobacterium species (SLAB51) was given to triple-transgenic mice (3xTg-AD) for 4 months starting at 2 months of age. Administration of SLAB51 increased the abundance of Bifidobacterium spp in the gut of the mice and a reduction in Campylobacterales, which led to reduced accumulation of Aβ aggregates, improved recognition memory, and reduced cortical degradation. The reduced plaques were associated with higher plasma concentrations of gut hormones ghrelin, leptin, GLP1 and GIP, which have been shown to aid in learning, memory, and neuroprotection against Aβ plaques. Furthermore, mice treated with SLAB51 had reduced plasma levels of several pro-inflammatory cytokines including IL-1β and IFNγ, but higher levels of G-CSF, GM-CSF and IL-4. Within the brain, SLAB51 led to improved proteasome functionality and autophagy (Bonfili et al., 2017).

The probiotic Bifidobacterium breve strain A1 (B. breve A1) was given to 10-week-old male wildtype mouse (ddY strain) that were intra-cerebroventricularly given a 2 mM solution of Aβ protein 25–35 to induce the effects of AD. ddY (Deutschland[Germany], Denken[Tokyo], Yoken [Japan]) mice are an outbred mouse strain derived from “dd” stock mice in the pre 1920s and then fully raised in the National Institute of Infectious Diseases in Yoken, Japan (NIBIOHN JLARB, 2022). The ddY strain has been described as a general-use mouse model and are extensively employed in diverse biomedical research to investigate therapeutic interventions in various pathologies, including Alzheimer’s disease (The Jackson Laboratory, 2022; Farid, Yang, Kuboyama, & Tohda, 2020; Min et al., 2017; Tohda et al., 2021; Tomino, 2010). B. breve A1 was administered daily starting 2 days before Aβ injection. 6 days after Aβ injection, B. breve A1 led to improved spatial memory. Hippocampal gene expression was also profiled and indicated a reduced immune response following B. breve A1 treatment in AD mice only. Mice treated with B. breve A1 had higher levels of SCFAs in serum, and administration of acetate also improved memory deficits, but not to the same extent as mice treated with live bacteria. Transcriptional analysis of brain tissue indicated that differentially expressed genes were involved in immune responses (Kobayashi et al., 2017).

However, not all probiotic formulation has demonstrated beneficial effects. The probiotic VSL#3, which is a mixture of lactic-acid producing bacteria was administered to 6-month-old female AppNL-G-F mice for 8 weeks, which was not observed to reduce amyloid plaques nor improve memory as measured by a cross-maze rodent behavior test (Kaur et al., 2020).

10.2. Human trials of probiotics for Alzheimer’s disease

As the medical field has gained knowledge about the role of the gut microbiome in AD, there have been several attempts to leverage the gut microbiome to limit the disease. Most notably is the use of probiotics which involve directly colonizing the gut microbiome with beneficial microbes (Naomi et al., 2021). In humans, there have been clinical trials to assess the effects of probiotics on cognitive function (Table 1).

Table 1.

Probiotic trials for Alzheimer’s disease.

Author and study design	Intervention	Probiotic species	Study population	Results
Akbari et al. (2016): Double blind randomized control clinical trial Two study groups: control (milk) and treatment (milk + probiotic mix)	200 mL/day of probiotic milk containing 2 × 109CFU/g of each bacterial species for 12 weeks	Lactobacillus acidophilus, Lactobacillus casei, Bifidobacterium bifidum, and Lactobacillus fermentum	n = 30 per group Diagnosed AD patients Mean Age: 82.0 (ctrl) and 77.67 (treatment) Mean BMI: 22.73 (ctrl) and 23.77 (treatment) Sex: 6 males and 24 females per group	Improved cognition in probiotic treated group (measured by mini mental state examination, MMSE) Decreased CRP levels, serum triglycerides, insulin resistance in the probiotic treated group Increased beta cell function and insulin sensitivity in the probiotic treated group
Agahi et al. (2018): Double blind randomized control clinical trial Two study groups: placebo and probiotics Participants were matched for disease severity, age, sex, and BMI	500 mg capsules of either placebo (maltodextrin) or 3 × 109 CFU/g probiotics every other day for 12 weeks	Lactobacillus fermentum, Lactobacillus plantarum, and Bifidobacterium lactis Or Lactobacillus acidophilus, Bifidobacterium bifidum, and Bifidobacterium longum	n = 23 (ctrl) and 25 (probiotic) Participants were tested using the TYM cognitive test and patients with scores indicative of severe AD were included in the study Mean Age: 80.57 (ctrl) and 79.7 (treatment) Mean BMI: 24.44 (ctrl) and 24.05 (treatment) Sex:10Mand13F(ctrl)and 7M and 18F (treatment)	No significant change in cognitive scoring between probiotic treated and control groups No significant change in serum cytokine levels or markers of oxidative stress
Leblhuber, Steiner, Schuetz, Fuchs, & Gostner, (2018): Explorative longitudinal study in which a cohort of moderate AD patients were assessed for cognitive abilities and inflammatory biomarkers before and after prebiotic treatment One study group, two time points (before treatment and after)	Probiotic treatment was given daily for 4 weeks	Lactobacillus casei W56, Lactococcus lactis W19, Lactobacillus acidophilus W22, Bifidobacterium lactis W52, Lactobacillus paracasei W20, Lactobacillus plantarum W62, Bifidobacterium lactis W51, Bifidobacterium bifidum W23 and Lactobacillus salivarius W24	n = 19 Participants had moderate AD based on MMSE scores Mean Age: 76.7 Sex: 9 males and 11 females	Increased fecal zonulin levels and Faecalibacterium prausnitzii concentrations Increased serum kynurenine concentrations Increased nitrite and neopterin levels Increased RNA content in fecal bacteria
Tamtaji et al. (2019): Double blind randomized control clinical trial Three study groups: selenium alone, selenium +probiotic, or placebo	200μg/day of selenium and 2 × 109 CFU/day of probiotic for 12 weeks	Lactobacillus acidophilus, Bifidobacterium bifidum, and Bifidobacterium longum	n = 26 (Se alone), n = 27 probiotic+Se, n = 26 (placebo) Patients diagnosed with AD Mean Age: ~78 Mean BMI: ~21	Significant reduction in CRP, insulin, and insulin resistance, LDL-cholesterol, and total cholesterol:HDL ratio Significant increase in glutathione, insulin sensitivity, and anti-oxidant capacity Significant increase in MMSE
(Ton et al., 2020): Uncontrolled clinical investigation to assess effects of probiotics on AD patients One study group, two time points (before treatment and after)	4% kefir including the probiotic species was added to pasteurized milk and incubated for 24h. Then the fermented product was blended with 500 g of strawberries and 2L of milk. Treatment duration 3 months	Acetobacter aceti, Acetobacter sp., Lactobacillus delbrueckii delbrueckii, Lactobacillus fermentum, Lactobacillus fructivorans, Enterococcus faecium, Leuconostoc spp., Lactobacillus kefiranofaciens, Candida famata, and Candida krusei	n = 13 AD patients displaying some kind of cognitive deficit and no comorbidities Mean age: ~78 Mean BMI: ~26 Sex: 2 males and 11 females	Significant improvement in cognitive abilities such as memory, abstraction, and language as well as higher MMSE scores with treatment Significant reduction in TNFa, IL-8 (CXCL8), and IL-12p70 levels in the serum, as well as reduced pro- to anti-inflammatory cytokine levels with treatment Significant reduction in ROS (superoxide, hydrogen peroxide, and hydroxyl radical) in the serum with treatment Significant increase in mitochondrial membrane potential, DNA content and reduction in DNA fragmentation and apoptosis

A double-blind randomized controlled trial in 2016 treated a cohort of 60 AD patients with either probiotics containing Lactobacillus acidophilus, Lactobacillus casei, Bifidobacterium bifidum, and Lactobacillus fermentum, or milk. These patients were treated for 12 weeks and after the trial, it was found that patients treated with probiotics showed improvement on the MMSE. These probiotics also had an improved effect on the metabolic health as determined by insulin sensitivity test, serum triglycerides, and beta cell function. Although no other biomarkers for inflammation or oxidative stress were altered with treatment, there was a significant decrease in serum C reactive protein in the probiotic-treated patients (Akbari et al., 2016). The same group followed up this study with one in which severe AD patients were given a similar cocktail of bacteria. In this trial, the AD patients showed some elevation in serum glutathione levels, but no improvement in memory function and no inflammatory or oxidative stress markers were altered between the treated and untreated groups including proinflammatory cytokines (Agahi et al., 2018).

In another trial, moderate AD patients (average MMSE of 18.5) were treated with either a probiotic mix of Lactobacillus strains and Bifidobacterium strains for 28 days. The major finding from the trial showed a significant increase the serum kynurenine levels and increased kynurenine/tryptophan ratio as well as neopterin. The authors claim that this change in kynurenine levels is suggestive of myeloid cell activation. However, the authors did not assess the effect of probiotic treatment on cognitive impairment (Leblhuber, et al., 2018).

In a relatively large study, 79 patients were randomly assigned to three groups: placebo or selenium alone, or selenium plus probiotics. The probiotics contained Lactobacillus acidophilus, Bifidobacterium bifidum, and Bifidobacterium longum. The treatment course was 12 weeks, after which the results showed that the selenium plus probiotic group significantly out-performed the other groups on the MMSE test. The probiotic group also showed significant reductions in C reactive protein, increased glutathione and total antioxidant capacity, as well as improved metabolic health such as lowered serum insulin, cholesterol, and triglycerides (Tamtaji et al., 2019).

Another approach to manipulate the microbiome for other acute and chronic diseases is the use of a fecal microbiota transplant (FMT) from a healthy donor to an affected individual via an enema, a nasal gastric infusion, or from encapsulated pills, and may have therapeutic potential to combat infectious and non-infectious diseases (Xu et al., 2021). FMT has been most widely used in the treatment of Clostridiodes difficile diarrhea, which is an opportunistic infection that can occur when the normal microbiota is diminished by antibiotic treatment. In cases of recurrent C. difficile infection, FMT efficacy is over 90% (van Nood et al., 2013). This high degree of efficacy with multiple microbiota donor sources may be mechanistically linked to restoring colonization from the microbiome, which then inhibits the growth of C. difficile in the setting of an acute infection. Whether these same principals can be applied to chronic diseases is an active area of investigation and may have relevance to Alzheimer’s disease.

Because elderly individuals are at a greater risk of C. difficile infection (CDI), there have been a few case reports in which AD subject have received FMT from a cognitively normal subject. In one case, an 82-year old man with AD and an MMSE of 20 (mild/moderate dementia) was initially hospitalized for a case of methicillin resistant Staphylococcus aureus, then acquired a GI infection with C. difficile (Hazan, 2020). Following antibiotic treatment with vancomycin, metronidazole, fidaxomicin, and bezlotoxumab and recurrence of CDI, the patient received an FMT infusion of a stool sample from his 85-year-old intellectually acute wife, which cured the C. difficile infection. Over the next 2 and 6 months, the subject’s MMSE improved to 26 and 29, respectively, indicating normal cognition.

In a second case report (Park et al., 2021), a 90-year-old woman had a diagnosis of AD and an MMSE of 15 (moderate dementia) recorded 30 days prior to a hospitalization event. Later, she acquired severe and antibiotic-refractory CDI. She was treated with an FMT from 27-year-old male donor who underwent detailed health screening, which initially increased her MMSE to 18 at 1-week post FMT and 20 at 3-months post FMT. In addition, at 3-months post FMT, her CDI symptoms returned, including fever, diarrhea, and continuous abdominal pain. A second FMT was performed and 1 week later, her MMSE was stable at 20. Her microbiota was sequenced at pre-FMT and post first FMT. FMT increased the levels of Tannerellaceae and decreased the levels of Negativicutes and Lachnospira.

The same group then evaluated cognitive changes in 10 subjects with dementia and severe CDI that received FMT therapy vs. 10 subjects with dementia and severe CDI that received antibiotic therapy (Park et al., 2022). The group receiving FMT had a baseline mean MMSE of 10.0 (range of 7.0–14.0) that increased to a mean of 16 (range of 13–18) following treatment, whereas the group receiving antibiotics had a baseline MMSE of 14.0 (range 12.8–16.0) that decreased to a mean of 10.0 (range of 9.8–15.3) after treatment. In the pre-FMT microbiota, the phylum Proteobacteria was approximately 44% of the relative abundance. This phylum is typically a minor component of the microbiota compared to the phyla Bacteroidetes and Firmicutes, and high levels detected pre-FMT may be linked to antibiotic-mediated disruption in the gut. FMT lowered Proteobacteria levels (including Enterobacteriaceae, Sutterella, and Klebsiella) to 22%, increased microbiota diversity, and increased prevalence of bacteria more typical of a normal gut microbiota including Bacteroides, Alistipes, Blautia, and Bifidobacterium.

These limited FMT case studies in the context of CDI and AD raise the question of whether this could be a promising new therapy for AD not suffering from CDI. One advantage of an FMT vs. delivery of a probiotic is that it delivers of a complex community of bacteria that may be more stable and may retain beneficial functions supported by microbial cross-feeding. In addition, this approach does not rely on the identification of a specific bacteria for treatment efficacy. However, there are considerable disadvantages including variability of microbes delivered from different donors and the potential for rare serious adverse effects, including infections with antibiotic resistance bacteria (Gupta, Mullish, & Allegretti, 2021). While regulatory measures have been established by the FDA and are continuing to be improved (Gupta et al., 2021), AD subjects represent an elderly population that may be more vulnerable to infection than younger subjects. Nonetheless, more work may be warranted in this area to evaluate potential safety and efficacy. Because of the considerable health and ethical risk of performing clinical trials in elderly and cognitively impaired individuals, there may be substantial barriers for large scale clinical trials. Thus, small case reports in which AD patients receive FMT used for non-AD clinical indications can provide valuable information for the field.

The collective data from both probiotic and FMT trials suggest that while manipulating the microbiome has shown some promise and should undergo further investigation, the studies conducted to this day are relatively small and show some variation. Future trials should aim to improve the assays performed on the AD patients and include the measurement of biomarkers in cerebrospinal fluid to detect improvements in early AD pathology. Additionally, future trials should attempt to corroborate cognitive function data by using multiple types of tests. Basic research should continue to explore new strains of gut bacteria that could potentially mediate a stronger and more specific effect. The precise mechanisms of prebiotics in AD are still unknown, but clinical trials suggest mechanisms involved in improving body metabolism, immune activation, and serum metabolites and more studies are needed in both preclinical research and in AD patients to determine new targets and refine existing therapies.