Alzheimer’s Disease-Related Proteins Targeted by Secondary Metabolite Compounds from Streptomyces: A Scoping Review

Literature search strategy

A collection of articles on Streptomyces sp. secondary metabolites related to neuroprotection in AD was obtained by conducting an extensive literature search using different keywords and subject headings. Five electronic databases (PubMed, Ovid Medline, ScienceDirect, Embase, and Scopus) were used to search for relevant original research articles published within the last 10 years (January 2010– December 2021). The search terms included “Alzheimer’s disease”, “Alzheimer’s”, “Streptomyces”, “Streptomyces sp.”, “Actinomycete”, “Neuroprotection”, “Neuroprotective”, “Proteins”, “Genes”, “Secondary metabolites”, and “metabolites”. The terms were searched in titles, abstracts, and Medical Subject Headings (MeSH) keywords using Boolean operators for defined search results.

Study eligibility and selection

The eligibility of the compiled articles was determined by screening the title and abstract using a set of inclusion and exclusion criteria (Table 1). The shortlisted research articles were then subjected to full-text review, and only research articles that fulfilled the inclusion criteria were selected for data extraction.

Screening and data extraction

A preliminary literature search was carried out using specific search terms. After removing duplicate articles, the articles were imported into Covidence (https://www.covidence.org/). Then, the articles were subjected to title and abstract review. Any papers aside from original research articles or those unrelated to AD studies were also removed. Subsequently, the shortlisted articles were subjected to full-text review based on this study’s inclusion and exclusion criteria (Table 1). The review and selection process were assessed by two independent reviewers. If any disagreements or conflicts arose, a third independent reviewer would assist in resolving the decision-making. The selection and exclusion processes are outlined in the PRISMA 2020 flow chart (Fig. 1).

Fig. 1

PRISMA flow chart (2020) PRISMA, Preferred Reported Items for Systematic Review and Meta-analysis [12] outlining the step-by-step process involved in the selection of studies included for this scoping review. *n, number.

Bioinformatics analysis

Three different bioinformatics tools were employed to investigate and analyze the potential AD-related proteins: 1) Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) data (http://string-db.org) (version 11.5), 2) PANTHER Classification System (PANTHER17.0) (http://www.pantherdb.org), and 3) Kyoto Encyclopedia of Genes and Genomes (KEGG) Mapper.

Literature search and selection

Using several search terms (Streptomyces AND Neuroprotection AND Alzheimer’s disease AND metabolites), 256 articles were first retrieved from the five databases. After 46 duplicates were removed, the inclusion and exclusion criteria were used to screen the remaining articles, producing 135 full-text articles. These articles underwent further independent screening by two reviewers, of which 83 articles were deemed credible for further screening based on the AD model. Finally, 12 articles were shortlisted for detailed data extraction (Fig. 1).

Characteristics of included studies

The scoping review focused on past studies pertaining to neuroprotection (n = 83) based on initial assessment studies (Fig. 1). Approximately 14.5% of the research articles (12 out of 83 studies) were selected based on the AD model. In addition, data were extensively extracted from these selected studies, primarily focusing on protein regulation by Streptomyces compounds (Table 2). The summary of the compiled studies showed that cumulatively, the highest number of articles on Streptomyces compounds in AD was published in Asia (70%), while the remaining 30% were published equally from Europe, South America, and North America (10% each) (Fig. 2A). Spain has the highest number of published articles with 25% (3 out of 12 research articles), followed by China, India, and South Korea (16.7% each), and the United States of America, Iran, and Chile (8.3% each) (Fig. 2B).

Fig. 2

Publication on Streptomyces compounds in Alzheimer’s disease models in the last decade. A) Articles published (%) on implications of Streptomyces compounds in Alzheimer’s disease model in the last 10 years (2010–2021) based on continents. B) Breakdown of published articles (%) by countries in the last decade.

Replicability of proteins

From the 12 shortlisted articles, 53 proteins were extracted, and their replicability was assessed. Only 38% of the proteins (20 out of 53) were reported in two or more research articles and chosen for further analysis (Fig. 3). Notably, 2 proteins (Aβ and BACE1) were reported in 5 articles, while the tau protein was described in 4 articles. Furthermore, 5 proteins (ERK, JNK, IL-1β, BECN1, and GSK3β) were mentioned in 3 articles. Another 2 articles described 13 proteins, and the remaining proteins were stated in a single article (Fig. 4).

Fig. 3

The summary of selected research articles and identified proteins. A) 14.5 % of the total articles (12 out of 83 articles) were selected based on the Alzheimer’s disease model. B) From the 12 independent research articles around 53 proteins were identified to play a role in Alzheimer’s disease. C) Only 38% of total proteins (20 out of 53 proteins) were shortlisted as they were present in more than two research articles. The list of shortlisted proteins used for further STRING, PANTHER, and KEGG analysis.

Fig. 4

Replicability of proteins identified across 12 independent Streptomyces-treated studies.(Created in BioRender.com).

PPI network analysis using STRING

The Protein-Protein Interaction (PPI) network was constructed by uploading the 20 shortlisted proteins to the STRING database (https://string-db.org) to evaluate their differential expression and alteration of protein interactions. The PPI network successfully mapped 20 nodes (interactions) with 63 edges and 14 expected edges, with the PPI enrichment p-value (p < 1.0 e-16) analyzed at high confidence (0.700) minimum required interaction score. In addition, three distinctive protein clusters involved in the Aβ and tau metabolism were generated using the K-means clustering, which were; GSK3β (GSK3B), PS1 (PSEN1), ERK (MAPK1), JNK (MAPK8), BACE1 (BACE1), Aβ(APBB1) and tau (MAPT), neuroinflammation; iNOS(NOS2), IL-6 (IL6), IL-1β (IL1β), p62 (SQSTM1), p65 (RELA), TNF-α (TNF), IL-10 (IL10), p38 (MAPK14); and autophagy: mTOR (MTOR), Beclin-1 (BECN1), ATG5 (ATG5), AKT (AKT1), and Nrf-2 (NFE2L2), as shown in Fig. 5.

Fig. 5

The STRING plot depicts the functional interrelationships of proteins analysed using the STRING 11.5 database. It shows 20 nodes (interactions) and 63 edges represent the protein and protein-protein interaction with an enrichment p-value (p < 1.0e–16). Three distinctive clusters were generated using K-means clustering shown in blue (MTOR, BECN1, ATG5, AKT1, NFE2L2), red (NOS2, IL6, IL1β, SQSTM1, RELA, TNF, IL10, MAPK14) and green (GSK3B, PSEN1, MAPK1, MAPK8, BACE1, MAPT, APBB1) clusters involved autophagy, neuroinflammation and beta-amyloid and tau metabolism respectively. *The genes of proteins are italicized.

Functional annotation analysis via PANTHER

The 20 shortlisted proteins represented as genes were classified using PANTHER GO-Slim based on the gene ontology domain: PANTHER pathway. Accordingly, nearly 30% of the genes (6 out of 20) were present in the AD-amyloid secretase pathway consisting of MAPK1, PS1, APBB1, MAPK14, BACE1, and MAPK8. On the contrary, 20% of the genes (4 out of 20) comprise the AD-presenilin pathway, including PS1, APBB1, BACE1, and GSK3B. Interestingly, both pathways show three common genes; PS1, APBB1, and BACE1. Figure 6 shows the pathway classification of the 20 proteins in the AD-amyloid secretase pathway (P00003) and AD-presenilin pathway (P00004).

Fig. 6

PANTHER analysis of shortlisted proteins from 12 independent Streptomyces-treated studies. The classification of proteins present in more than two research articles were analysed using PANTHER gene ontology of 20 selected proteins based on PANTHER Pathway; Alzheimer’s disease-amyloid secretase pathway (P00003) and Alzheimer’s disease-presenilin pathway (P00004). The Venn diagram shows overlapping between two pathways producing three (3) specific proteins; PSEN1 (PSEN1), amyloid-β(APBB1) and BACE1(BACE1). (Created in BioRender.com).

KEGG mapper

The functional role of these 20 shortlisted proteins in AD pathogenesis was assessed using the KEGG pathway. Around 75% of the total proteins (15 out of 20) were successfully mapped in the AD pathway map (hsa05010) in Homo sapiens, as highlighted in green (Fig. 7) indicating these proteins are involved directly in the pathogenesis of AD. A total of 15 proteins, which consisted of AKT (AKT1), BACE1 (BACE1), mTOR (MTOR), GSK3β (GSK3B), Aβ (APBB1), IL-6 (IL6), IL-1β (IL1β), tau (MAPT), iNOS (NOS2), ERK (MAPK1), JNK (MAPK8), PS1 (PSEN1), p65 (RELA), TNF-α (TNF), and Beclin-1 (BECN1), were involved in multiple pathways, including axonal transport defects, impaired autophagy, impaired neuronal insulin signaling, long-term reduction in potentiation, microglial inflammation, Aβ and tau processing, apoptosis, and dysfunctional mitochondria. Noteworthy, 25% of the total proteins (5 out of 20) namely Nrf-2 (NFE2L2), ATG5 (ATG5), IL-10 (IL10), p38 (MAPK14), and p62 (SQSTM1) may play an in-direct role in AD pathogenesis.

Fig. 7

Only 15 proteins (75%) consisted of AKT (AKT1), BACE1 (BACE1), mTOR (MTOR), GSK3β (GSK3B), amyloid-β (APBB1), IL-6 (IL6) IL-1β (IL1β), tau (MAPT), iNOS (NOS2), ERK (MAPK1), JNK (MAPK8), PS1 (PSEN1), p65 (RELA), TNF-α (TNF), and beclin-1 (BECN1) were successfully mapped highlighted in green in the Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway of Alzheimer’s disease in Homo sapiens (hsa05010). *The genes of respective proteins are italicized. Published with permission from Kanehisa Laboratories.

Differentially expressed AD-related proteins

In AD pathogenesis, 70% of the proteins (14 out of 20) are upregulated, while the remaining 30% (6 out of 20) are downregulated. In comparison, it also shows that 35% of the proteins (7 out of 20) were inconsistently regulated when treated with Streptomyces sp. compounds or extract (Table 3). The remaining 45% (9 out of 20) proteins were consistently downregulated while the remaining 20% (4 out of 20) of the proteins were consistently upregulated.

Amyloidosis in Alzheimer’s disease

Five articles showed that Aβ was consistently downregulated when treated with Streptomycessp. compounds, including rhizolutin, rapamycin, enterocin, bonnevillamides D and E, and chloroform extract. In contrast, their expressions were upregulated in AD pathogenesis [14, 15, 18, 20, 23]. In general, Aβ is released from the amyloid-β protein precursor (AβPP) by β-secretase (BACE1) and γ-secretase, producing neurotoxic Aβ fragments. In the first rate-limiting step, the N-terminus of the Aβ is cleaved by β-secretase, followed by γ-secretase cleaving the C-terminus of Aβ. The resulting formation of Aβ oligomers is then polymerized into aggregated Aβ before being converted into plaques. This leads to kinase activation, such as GSK3β, ERK2, and CDK5, leading to the hyperphosphorylation of microtubule-associated tau proteins, which are polymerized into insoluble NFTs. Subsequently, these proteins induce microglia recruitment, triggering a local inflammatory response that causes neurotoxicity and neuronal death, giving rise to AD [56]. Hence, targeting the earlier stages by inhibiting the activity of BACE1 before the formation of Aβ oligomers may provide a novel treatment against AD.

Although BACE1 is an ideal target candidate for treatment as it initiates the amyloidogenic pathway, however, past studies reported that BACE1 was inconsistently regulated [13, 14, 16, 17, 24]. When treated with a well-known and established Streptomyces hygroscopicus-derived compound known as rapamycin, showed that BACE1 expression was insignificant in murine neuroblastoma cells overexpressing the “Swedish” mutant APP6965aa isoform (SweAPP N2a), as reflected by the increased amyloid production. This was also corroborated by the enhanced β-Carboxyl-Terminal Fragment (β-CTF) level of AβPP and the decreased soluble AβPP alpha (sAβPPα), a neuroprotective cleavage product. The findings suggest that the α-secretase activity was inhibited by rapamycin, like the inhibition of α-secretase disintegrin and metallopeptidase domain-10 (ADAM-10) [24]. The suppressed ADAM-10 pathway initiates the amyloidogenic pathways typically reported in AD [57]. However, Chen et al. (2019) refuted the suggestion with a contradicting finding, revealing that the Aβ production in AβPP/PSEN1 transgenic mice was reduced following treatment with rapamycin, which downregulated the expression of AβPP-cleaving enzymes, including BACE1 and PS1.

It should be noted that rapamycin is an established mTOR inhibiting drug, although its downregulation reduces Aβ₄₂, Aβ plaques, and NFT levels and improves AD-like cognitive deficits. It also serves a multifaceted role in normal cellular functions, such as gene translation and cellular growth. However, it was agreed that prolonged intake of rapamycin leads to severe side effects due to the inhibition of the mTOR pathway [28]. Noteworthily, two clinical trials (NCT04200911 and NCT04629495) are being conducted to investigate the long-term use of rapamycin in older adults with MCI and early-stage AD. Although NCT04200911 has reported adverse side effects including altered mental status and headache in participants, however, both studies are still in its preliminary stages and incomplete, as neither could provide supporting evidence on the feasibility of rapamycin as long-term usage in treating AD thus far [58, 59]. Therefore, rapamycin may not be an ideal Streptomyces sp. compound for long-term AD treatment.

The contradicting BACE1 regulation from the two studies supports the notion that rapamycin is unsuitable for AD treatment. Therefore, other Streptomyces sp. compounds with prospective targets against BACE1 should be identified and considered. Examples of Streptomyces sp. compounds that were shown to downregulate BACE1 expression include caniferolide A, anhydroexfoliamycin, and streptocyclinone B isolated from Streptomyces caniferus, Streptomyces sp. Lt 005, and Streptomyces sp. CA-237351, respectively [13, 16, 17]. Therefore, Aβ and BACE1 remain the prospective target proteins to inhibit the amyloidogenic pathway in AD.

Autophagic improvement

The accumulation of autophagosomes containing amyloid-beta and tau proteins contributes to the intracellular build-up of toxic peptides in dystrophic neurites, leading to AD progression [60–62]. Previously, five proteins, consisting of ATG5, AKT, Beclin-1, Nrf-2, and mTOR, were identified to play a role in autophagy through downregulation in AD, except for mTOR. Interestingly, all proteins showed counteracting results when treated with different Streptomyces sp. compounds (Fig. 9), indicating that these proteins are potential targets. Given this, targeting these proteins involved in autophagy is advantageous as an alternative treatment approach.

BECN1 is involved in autophagy by promoting the removal of toxic peptides and preventing neuronal death by maintaining basal autophagy and protein turnover. Additionally, BECN1 plays a crucial role in several membrane transports responsible for autophagy, phagocytosis, and endocytosis [36]. The reduced BECN1 gene and protein expression in AD patients’ brain tissue results in autophagy defects, similar to the abnormal accumulation of subcellular vesicles, producing toxic AβPP metabolites. Meanwhile, BECN1 deletion in human APP (hAPP)-transgenic mice in vivo model increased Aβ plaque deposition, neuronal loss, and the accumulation of AβPP and its metabolites, particularly AβPP C-terminal fragment peptides [27]. On the contrary, mRNA and BECN1 protein expressions were upregulated in mice brains with AD when treated with rapamycin [14, 21], although the latter reported an insignificant Beclin-1 protein expression. Additionally, 24-hour treatment with alborixin, a known Streptomyces scabrisporus compound, enhanced the BECN1 expression in both N9 cells and primary neuronal cells (C57BL/6) [22].

On top of that, ATG5 plays a significant role in autophagy, as its downregulation leads to either partial or total autophagy inhibition [63]. The knockdown effect of the ATG5 gene in AD patients results in the accumulation of ubiquitinated proteins, triggering neuronal loss [26]. When microglial N9 cells were treated with alborixin, the ATG5 expression was upregulated, as reflected by the enhanced autophagic activities in the lysosomal removal of Aβ_1–42. Such an occasion was due to the PTEN suppression of the P13K-AKT pathway [22]. In another study, rapamycin treatment in an in vivo Wistar rats model increased the ATG5 mRNA expression in the hippocampal region when treated with Aβ_1–42 via autophagic induction [21].

Mitigating neuroinflammation

The accumulation of Aβ and tau proteins caused by autophagic dysregulation activities stimulates microglial cells to remove these proteins. However, the removal process is impaired as the system becomes overwhelmed, leading to prolonged activation and enhanced inflammatory response [41]. Various pro-inflammatory cytokines, including TNF-α, IL-6, IL-1β, and small-molecule messengers, such as nitric oxide further damage the brain cells as they become more susceptible to oxidative stress-mediated cell death [41, 64]. For instance, Nrf-2 was shown to play a dual role in the fine-tuning feedback loop resulting from an elevated level of oxidative stress in combination with the autophagic process through the AMP-activated protein kinase upregulation and mTOR downregulation [65]. Nrf-2 also activates numerous antioxidative enzymes, including catalase, superoxide dismutase, and glutathione peroxide, to mitigate reactive oxygen species and protect from neuroinflammation [16]. Furthermore, Nrf-2 upregulation prevents the transcription of NFkB, which comprises p65, c-Rel, p50, RelB, and p52. Consequently, the nuclear expressions of p65/p50 are downregulated, which prevents the induction of pro-inflammatory genes, suppressing neuroinflammation [13].

Recently, caniferolide A and streptocyclinone A and B were isolated from Streptomyces caniferus and Streptomyces sp. CA-237351, respectively. Interestingly, the nuclear Nrf-2 levels in BV2 microglial cells were significantly enhanced when pre-treated with these compounds. Comparatively, caniferolide A produced the highest effect on its expression at the lowest concentration of 0.001 μM, implying that it is a potent Nrf-2 activator compared to streptocyclinone A and B. The authors also identified that all three compounds downregulated the nuclear NFkB-p65 expression in BV2 microglial cells due to their relation with the NFkB in oxidative stress [13, 17]. This evidence supports the association between the upregulation of Nrf-2 and the suppression of the NFkB expression, subsequently mitigating inflammatory response.

The upregulation of Nrf-2 also provides sufficient neuroprotection against oxidative stresses through increasing antioxidant activities, as reflected by the increased glutathione (GSH) levels when treated using all three compounds. This indicates that these compounds could improve the GSH enzyme expressions, including glutathione peroxidase and glutathione transferase, in tackling oxidative stress. Furthermore, these compounds diminished the release of pro-inflammatory cytokines, including IL-1B, TNF-α, and IL-6. However, IL-6 downregulation was only detected when BV2 microglial cells were treated with caniferolide A. The iNOS expression was also downregulated, reducing the release of nitric oxide in the BV microglial cells [13, 17].