WO2025217551A1

WO2025217551A1 - Microrna-mediated obstruction of stem-loop alternative splicing

Info

Publication number: WO2025217551A1
Application number: PCT/US2025/024324
Authority: WO
Inventors: Rong Grace ZHAI; Stefan WUCHTY; Kai RUAN
Original assignee: University of Miami
Current assignee: University of Miami
Priority date: 2024-04-12
Filing date: 2025-04-11
Publication date: 2025-10-16
Anticipated expiration: 2026-10-12

Abstract

The present disclosure provides synthetic nucleic acid compositions and methods of use thereof to generate a computational modeling platform to uncover noncoding RNAs that drive disease progression or intrinsic neuroprotection through regulation of splicing.

Description

MICRORNA-MEDIATED OBSTRUCTION OF STEM-LOOP ALTERNATIVE SPLICING

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of priority to U.S. Provisional Application No. 63/633,084, filed April 12, 2024, which is incorporated by reference herein in its entirety.

GOVERNMENT SUPPORT CLAUSE

This invention was made with government support under Grant No. AT010408 awarded by the National Institutes of Health. The Government has certain rights in the invention.

REFERENCE TO SEQUENCE LISTING

The sequence listing submitted on April 11, 2025, as an .XML file entitled “11348- 055WOl_ST26.xml” created on April 4, 2025, and having a file size of 40,492 bytes is hereby incorporated by reference pursuant to 37 C.F.R. § 1.52(e)(5).

BACKGROUND

Alzheimer’s disease (AD) is among the most intractable diseases, with escalating social and economic burden worldwide. Fundamental understanding of degenerative processes, as well as neuroprotective concepts, are urgently needed. In the past decade, significant investments from NIH and numerous other funding organizations in the research of Alzheimer's Disease Related Dementias (ADRD) have led to a tremendous advance in the collection of ADRD related data. Despite herculean financial and scientific efforts to tackle this devastating disease, to date no cures exist. Given the present limitations, there is a need to address the aforementioned problems by developing a modeling system for driving neuroprotection against ADRD.

SUMMARY

The present disclosure provides synthetic nucleic acid compositions and methods of use thereof to generate a computational modeling platform to uncover noncoding RNAs that drive disease progression or intrinsic neuroprotection through regulation of splicing. In one aspect, disclosed herein is a synthetic nucleic acid comprising a reporter gene, a first stem loop, a second stem loop, a nuclear export signal, and an exon 10 fragment from the human Tau gene, wherein the nuclear export signal and the exon 10 fragment from the human Tau gene are located between the first stem loop and the second stem loop.

In some embodiments, the first stem loop has a 3’ splice site. In some embodiments, the second stem loop has a 5’ splice site. In some embodiments, the reporter gene is eGFP, or a variant thereof.

In one aspect, disclosed herein is a method of screening for miRNAs that modulate the alternative splicing of exon 10 of the human Tau gene, comprising providing the synthetic nucleic acid of any preceding aspect, incubating the synthetic nucleic acid with a miRNA, and determining whether the miRNA modulates the alternative splicing of exon 10 of the human Tau gene.

In some embodiments, the increased expression of the reporter gene in the cytoplasm provides a pathological outcome. In some embodiments, the increased expression of the reporter gene in the nucleus provides a non-pathological outcome.

In one aspect, disclosed herein is a method of diagnosing Alzheimer’s Disease or an increased risk of developing Alzheimer’s Disease, comprising providing the synthetic nucleic acid of any preceding aspect, incubating the synthetic nucleic acid with a miRNA, and determining whether the miRNA modulates the alternative splicing of exon 10 of the human Tau gene, wherein increased expression of the reporter gene in the cytoplasm confers a pathological outcome and a diagnosis of Alzheimer’s Disease or an increased risk of developing Alzheimer’s Disease.

In some embodiments, the method further comprises administering a specific treatment for Alzheimer’s Disease when the miRNA increases expression of the reporter gene in the cytoplasm.

In some embodiments, the miRNA is selected from the group consisting of hsa-miR- 541-3p, hsa-miR-5O5-5p, hsa-miR-1910-3p, hsa-miR-135a-3p, hsa-miR-362-5p, hsa-miR-9- 5p, hsa-miR-103a-2-5p, hsa-miR-887-5p, hsa-miR575, hsa-miR-589-3p, hsa-miR-378c, hsa- miR-378a-3p, hsa-miR378d, hsa-miR-346, hsa-miR-378h, hsa-miR-557, hsa-miR-500a-3p, hsa-miR-502-3p, hsa-miR-1271-5p, hsa-miR-5001-3p, hsa-miR-664b-3p, hsa-miR-101-5p, hsa-miR-204-5p, hsa-miR-764, hsa-miR-370-3p, hsa-miR-1343-3p, hsa-miR-371a-3p, hsa- miR-4485-5p, hsa-miR-2682-3p, hsa-miR-331-3p, hsa-miR-660-3p, hsa-miR-1306-5p, hsa- miR-6756-3p, hsa-miR-1908-3p, hsa-miR-6894-3p, hsa-miR-6760-3p, hsa-miR-6726-3p, hsa- miR-23b-3p, hsa-miR-1178-3p, hsa-miR-149-5p, hsa-miR-23a-3p, hsa-miR-324-3p, hsa-miR- 342-3p, hsa-miR-20b-5p, hsa-miR-106a-5p, hsa-miR-525-3p, hsa-miR-106b-5p, hsa-miR- 18b-5p, hsa-miR-28-5p, hsa-miR-6775-5p, hsa-miR-764, hsa-miR-3689d, hsa-miR-3622b-5p, hsa-miR-1294, hsa-miR-92b-5p, hsa-miR-30b-3p, hsa-miR-516b-5p, hsa-miR-299-3p, hsa- miR-769-3p, and hsa-miR-7114-5p.

In one aspect, disclosed herein is a method of treating or preventing Alzheimer’s Disease, comprising providing the synthetic nucleic acid of any preceding aspect, incubating the synthetic nucleic acid with a miRNA, determining whether the miRNA increases expression of the reporter gene in the cytoplasm, and administering a specific treatment for Alzheimer’s Disease when the miRNA increases expression of the reporter gene in the cytoplasm.

In some embodiments, the specific treatment for Alzheimer’s Disease comprises several prescription drugs that are approved by the U.S. Food and Drug Administration (FDA) for Alzheimer’s disease to help either manage symptoms or treat the disease. In particular, current medications include cholinesterase inhibitors that prevent the breakdown of acetylcholine, a brain chemical important for memory and cognition, as well as N-methyl-D- aspartate (NMDA) antagonists that regulates glutamate, an important brain chemical.

In one aspect, disclosed herein is a method of treating or preventing Alzheimer’s Disease in a subject in need thereof, comprising administering to the subject a miRNA selected from the group consisting of hsa-miR-541-3p, hsa-miR-505-5p, hsa-miR-1910-3p, hsa-miR- 135a-3p, hsa-miR-362-5p, hsa-miR-9-5p, hsa-miR-103a-2-5p, hsa-miR-887-5p, hsa-miR575, hsa-miR-589-3p, hsa-miR-378c, hsa-miR-378a-3p, hsa-miR378d, hsa-miR-346, hsa-miR- 378h, hsa-miR-557, hsa-miR-500a-3p, hsa-miR-502-3p, hsa-miR-1271-5p, hsa-miR-5001-3p, hsa-miR-664b-3p, hsa-miR-101-5p, hsa-miR-204-5p, hsa-miR-764, hsa-miR-370-3p, hsa- miR-1343-3p, hsa-miR-371a-3p, hsa-miR-4485-5p, hsa-miR-2682-3p, hsa-miR-331-3p, hsa- miR-660-3p, hsa-miR-1306-5p, hsa-miR-6756-3p, hsa-miR-1908-3p, hsa-miR-6894-3p, hsa- miR-6760-3p, hsa-miR-6726-3p, hsa-miR-23b-3p, hsa-miR-1178-3p, hsa-miR-149-5p, hsa- miR-23a-3p, hsa-miR-324-3p, hsa-miR-342-3p, hsa-miR-20b-5p, hsa-miR-106a-5p, hsa-miR- 525-3p, hsa-miR-106b-5p, hsa-miR-18b-5p, hsa-miR-28-5p, hsa-miR-6775-5p, hsa-miR-764, hsa-miR-3689d, hsa-miR-3622b-5p, hsa-miR-1294, hsa-miR-92b-5p, hsa-miR-30b-3p, hsa- miR-516b-5p, hsa-miR-299-3p, hsa-miR-769-3p, and hsa-miR-7114-5p.

In one aspect, disclosed herein is a synthetic nucleic acid comprising a reporter gene, a fragment from an Nmnat gene, wherein the fragment from the Nmnat gene comprises a stem loop comprising a box 1 sequence and a box 2 sequence, wherein the box 1 sequence and the box 2 sequence hybridize to form a stem of the stem loop. In some embodiments, the box 1 sequence comprises SEQ ID NO: 1. In some embodiments, the box 2 sequence comprises SEQ ID NO: 2.

In one aspect, disclosed herein is a method of screening for miRNAs that modulate the alternative splicing of an Nmnat gene, comprising providing the synthetic nucleic acid of any preceding aspect, incubating the synthetic nucleic acid with a miRNA, and determining whether the miRNA modulates the alternative splicing of the Nmnat gene.

In some embodiments, modified expression of the reporter gene provides a neuroprotective outcome.

In one aspect, disclosed herein is a method for identifying microRNA mediated obstruction of stem-loop alternative splicing, comprising identifying a target sequence of interest, and identifying a microRNA sequence from a miRNA that can hybridize to the target sequence, wherein the binding of the microRNA sequence to the target sequence obstructs and/or modulates the alternative splicing of a gene comprising the target sequence.

In one aspect, disclosed herein is a method for modulating expression of a gene, comprising identifying a target sequence of interest in the gene, and identifying a microRNA sequence from a miRNA that can hybridize to the target sequence of interest in the gene, wherein the binding of the microRNA sequence to the target sequence modulates the expression of the gene.

In one aspect, disclosed herein is a method for monitoring and/or treating a disease caused by a splicing defect, comprising identifying a target sequence near the genetic mutation;, identifying a microRNA sequence from a miRNA that can hybridize to the target sequence near the genetic mutation, administering a nucleic acid comprising the microRNA sequence, and allowing the nucleic acid to bind to the target sequence near the genetic mutation, wherein the binding of the nucleic acid to the target sequence modulates the splicing defect to treat the genetic disease.

In some embodiments, the target sequence is an Nmnat gene. In some embodiments, the microRNA sequence is miR-iab-8-5p, miR-137-5p, miR-210-3p, miR-137-3p, miR-274- 3p, miR-307a-5p, miR-278-5p, miR-10-3p, miR-983-3p, miR-278-3p, miR-1002-5p, miR-9b- 5p, miR-252-5p, or miR-193-3p.

In some embodiments, the target sequence is a tau gene. In some embodiments, the microRNA sequence is hsa-miR-541-3p, hsa-miR-505-5p, hsa-miR-1910-3p, hsa-miR-135a- 3p, hsa-miR-362-5p, hsa-miR-9-5p, hsa-miR-103a-2-5p, hsa-miR-887-5p, hsa-miR-575, or hsa-miR-589-3p.

In some embodiments, the nucleic acid comprises 80% similarity or more to SEQ ID NO: 3. In some embodiments, the nucleic acid comprises 90% similarity or more to SEQ ID NO: 3. In some embodiments, the nucleic acid comprises SEQ ID NO: 3.

In some embodiments, the nucleic acid further comprises a fluorescent moiety. In some embodiments, the fluorescent moiety comprises mCherry or GFP.

In some embodiments, the disease is a neurodegenerative disease, including but not limited to Alzheimer’s disease. In some embodiments, the disease is brain cancer.

Other systems, methods, features and/or advantages will be or may become apparent to one with skill in the art upon examination of the following drawings and detailed description. It is intended that all such additional systems, methods, features and/or advantages be included within this description and be protected by the accompanying claims.

BRIEF DESCRIPTION OF DRAWINGS

FIGURES 1A-1C depict the computational pipeline to detect microRNA- mediated obstruction of stem-loop alternative splicing (MIMOSAS). FIG. 1A shows that non-spliced mRNAs form secondary structures in mRNA around splice- acceptor and splice-donor sites through complementary sequences - termed boxes - that influence the splicing outcome. FIG. IB is an illustration showing the chromosomal location of the genes predicted to contain splicerelevant stem-loop secondary structures in Drosophila melanogaster. The names of the genes are shown to the right of each chromosome. The numbers to the left of each gene represent its location based on cytogenetic map. FIG. 1C shows the determination of binding sites of microRNAs that overlap with boxes are based on the assumption that the energy gained by the microRNA-mRNA binding (Gdupiex) offsets the energy lost by opening the splice-relevant secondary structures AG„_r„u, = Go - Gi. If Gdupiex > AG_open binding of the microRNA that destroys the splice relevant secondary structure is energetically more favorable. Assuming that nascent mRNAs immediately produce secondary structures, predictions are refined through a sequence window upstream a given box, calculating the difference of the duplex with the underlying microRNA and opened structure AGMIMOSAS = Gdupiex - AG_open as well as the probability that a microRNA binds in the given box region.

FIGURE 2 depicts the determination of splice relevant boxes (SEQ ID NO: 22, SEQ ID NO: 23) in the mRNA of Fas3 (SEQ ID NO: 20, SEQ ID NO: 21). The matrix shows the nucleotide binding probabilities (P > 0.8) of the sequence of the Fas3 mRNA that captures the exon 5 and flanking introns. The inset points to a stem formed from box sequences that are near splice sites (residues 2-29 of SEQ ID NO: 22, residues 3-33 of SEQ ID NO: 23), overlapping with complementary sequences (residues 22-34 of SEQ ID NO: 23). To refine these boxes, the energetically most stable RNA structure of the underlying mRNA subsequence confirms the box sequences through the emergence of a stem in the underlying secondary structure, providing final box sequences.

FIGURES 3A-3F depict that miR-973 and miR-1000 drive splice variants of Fas3. FIG. 3A shows the structure of the mRNA of Fas3 indicating box I and II (SEQ ID NO: 22, SEQ ID NO: 23) that form a splice-relevant stem around exon 5. FIG. 3B is a table which indicates miRNAs that bind boxes with AGMIMOSAS- Energetically (un)favorable microRNAs for further downstream testing are highlighted. FIG. 3C shows splice variant RC that contains exon 5 emerges when highlighted microRNAs thwart stem formation. In the absence of MIMOSAS, Fas3 splice variants occur that miss exon 5 (RA/B/D/E/F/G). FIG. 3D shows AGMIMOSAS and corresponding microRNA binding probabilities of highlighted miR-973 and miR-976 in 200 nt sequence windows upstream of box I. FIG. 3E shows AGMIMOSAS and corresponding miRNA binding probabilities of miR-1000 and miR-999 in 200 nt sequence windows upstream of box II. FIG. 3F shows a plot of the ratio between the endogenous RA/B/D/E/F/G (STEM) and total mRNA variants measured by qRT-PCR in the brains overexpressing microRNAs by pan-neuronal driver elavC155-GAL4. The ratio of STEM vs Total in the miR-9c expressing group was set to 1, and the fold changes were displayed. All data were presented as mean ± s.d. ***P<0.001, **P<0.01, *P<0.05, unpaired Student’s t test; n=4, triplicate sampling.

FIGURE 4 depicts the determination of splice-relevant boxes (SEQ ID NO: 24, SEQ ID NO: 25) in the mRNA of Nmnat. The matrix indicates nucleotide binding probabilities (>0.8) in the subsequence of the pre-mRNA of Nmnat, that captures exon 5 and flanking introns. A stem forms from sequences near splice sites (boxes around nucleotides) that overlap with previously found complimentary box sequences.

FIGURES 5A-5F depict that Nmnat variants are driven by various microRNAs. FIG. 5A shows the gene structure of Nmnat harbors a pair of boxes, encapsulating exon 5, that are bound by several microRNAs. If these boxes form a stem, splice variant RA is produced, while MIMOSAS drives the RB splice variant. FIG. 5B shows microRNAs that were sorted according to their binding characteristics on boxes I and II. Highlighted microRNAs were chosen for further testing. FIG. 5C shows a plot of the ratio between the endogenous RA (STEM) and RB (MIMOSAS) measured by qRT-PCR in the brains overexpressing microRNAs by pan-neuronal driver elavC155-GAL4. The ratio of RA vs RB in the miR-9c expressing group was set to 1 , and the fold changes were displayed. All data were presented as mean ± s.d. ***P<0.001 , **P<0.01 , *P<0.05, unpaired Student’s t test; n>3, triplicate sampling. FIG. 5D shows an alternative testing method reflects the production of the RB splice variant by EGFP, while mCherry signals the presence of the RA splice variant. FIG. 5E shows brain morphology of 5 days-old (5 DAE) adult flies co-expressing the alternative-splicing reporter. microRNAs by pan-neuronal driver nysb-GAL4 were imaged by confocal microscopy. The nucleus was marked with DAPI (white) stains, and the intensities of both EGFP and mCherry were indicated with heat maps. The scale bar is 50 pm. FIG. 5F quantifies the ratio of EGFP and mCherry intensities in FIG. 5E. All data were presented as mean ± s.d. ***P<0.001, **P<0.01, *P<0.05, unpaired Student’s t test; n>4.

FIGURE 6 depicts determination of microRNA candidates for Nnrnat splicing. Focusing on box I and II, the study considered a sequence window of multiples of corresponding box lengths around a given box on the Nmnat mRNA and calculated IGMIMOSAS as well as the probability that microRNA candidates indeed bind a given box.

FIGURES 7A-7B depict that knockdown of the endogenous microRNAs by miRNA sponges affects MIMOSAS outcome. FIG. 7A shows a gene structure diagram of the microRNA sponge transgenic fly line. 20 microRNA binding sites with mismatches at positions 9-12 were inserted downstream of mCherry in a UAS-containing attB vector. The resulting transgenic animals can be crossed to specific Gal4 lines to achieve a tissue- or cell-specific expression. FIG. 7B shows the scatter plot of the ratio between the endogenous RA (STEM) and RB (MIMOSAS) of Drosophila Nmnat measured by qRT-PCR in the brains overexpressing miRNA sponges by pan-neuronal driver elavC155-GAL4. The ratio of RA vs RB in the scramble sponge expressing group was set to 1 , and the fold changes were displayed. All data were presented as mean ± s.d. ***P<0.001, **P<0.01, *P<0.05, unpaired Student’s t test; n>3, triplicate sampling.

FIGURE 8 depicts determination of nested boxes in RpL3. Considering the subsequence of RpL3 that captures exon 1, 2 and 3, the most stable RNA secondary structure features a formation of a set between box I and II (SEQ ID NO: 26, SEQ ID NO: 27). Focusing on the subsequence of RpL3 that captures exon 1, 2, 3 and 4, the stem between box I and II dissolved while box II (SEQ ID NO: 28) formed a wider stem with box III (SEQ ID NO: 29).

FIGURES 9A-9D depict that microRNAs drive MIMOSAS on RpL3 with three competing boxes. FIG. 9A shows a diagram of gene structure of RpL3 and predicted spliced mRNA variants. Color coded boxes I, II and III mark the sequences that form the stem- loop structures that are required for alternative splicing. Several microRNAs are indicated to bind box I and II, driving certain splice variants. Black arrow indicates the shared donor splice site. Adjacent and distal arrows mark the two acceptor splice sites. Three TaqMan real-time PCR probes are marked for the following experiments. FIGS. 9B-9D show total mRNA variants, RA/H and RG, and RG only. The pie diagrams indicate the percentage of endogenous spliced mRNA variants measured by RT-PCR of total RNA from head extracts of adult wild-type flies (yw) (FIGS. 10A-10B). FIGS. 9B-9D also show the ratio of RpL3 spliced mRNA variants measured by two-color multiplex real-time PCR assay with specific TaqMan probes in flies overexpressing microRNAs. The ratio level in miR-9c overexpressing group was set to 1 and fold change was displayed. All data were presented as mean ± s.d. ***P<0.001, **P<0.01, *P<0.05, unpaired Student’s t test; n=4, triplicate sampling.

FIGURES 10A-10B depict that Drosophila RpL3 is alternatively spliced into various mRNA variants. FIG. 10A shows a diagram of RpL3 gene structure and predicted spliced mRNA variants. Box I, II, and III mark the sequences that form the stem-loop structures that are required for alternative splicing. Black arrows indicate the sequence location of shared forward primer (F-Rpl3-E2/3). The other arrows (R-Rpl3-RA/H, R-Rpl3-RG, and R-Rpl3-RD) are reverse primers marking the sequence locations for each specific mRNA variant. FIG. 10B shows RT-PCR of total RNA from head extracts of wild-type flies (yw) using three primer sets as indicated in FIG. 10A.

FIGURE 11 depicts determination of microRNAs binding boxes I, II and III in RpL3. Further investigating the efficacy of miRs to interfere with splice-relevant secondary structures of the underlying RpL3 mRNA through box I, II and III we considered sequence windows of multiples of corresponding box lengths around a given box and calculated /IGMIMOSAS as well as the probability that a miR indeed binds a given box.

FIGURES 12A-12E depict that MIMOSAS is driven by microRNAs on the splicing reporter of RpL3 in vitro. FIG. 12A shows a design of the splicing reporter for RpL3 that includes crucial sequences involved in pre-mRNA STEM formation and self-complementing split fluorescent proteins (FPs), nuclear (NLS) and cytoplasmic (NES) localization signals. Three splicing outcomes were accounted for in parallel, each with its spliced mRNA variant and protein product, as well as the visualization mechanism of each reporter. FIG. 12B shows that Cos-7 cells were transfected with the splicing reporter and sfCherry 11-pre-miRNAs (miR- 9c, miR-304, miR-210, miR-988, or miR-992), and imaged 48 hours after transfection. The nucleus was marked with DAPI and Lamin A/C stains, while intensity was indicated with heat maps. Possible schematic diagrams of each microRNA on the splicing reporter are presented below by their corresponding column of cell images. FIG. 12C shows quantified fluorescent intensities, the ratio of spliced variant RA/H is represented by cytoplasmic GFP vs. total fluorescent (GFP + mCherry). FIG. 12D shows the ratio of RG is represented by nuclear GFP vs. total fluorescent. FIG. 12E shows the ratio of RD is represented by mCherry vs. total fluorescent (all n>12). Statistical significance through unpaired Student’s t test, relative to controls (miR-9).

FIGURES 13A-13E depict that nuclear localized microRNAs drive MIMOSAS in the presence of AG01. FIG. 13A shows the subcellular localizations of miR-210-3p in the salivary gland cells of the larvae with the expression of miR-210 pre-microRNA in the motor neurons (using OK371-GAL4). Scale bar was 20 pm for each row. The nucleus was marked with DAPI (gold) stains, while miR-210-3p was probed by its anti-sense FISH probe (magenta), and its intensity was indicated with heat maps. FIGS. 13B-13C show the fluorescence intensity of miR-210-3p measured by ImageJ. The quantification of Nuclear/Total ratio was shown in a scatter plot (FIG. 13B) and the average intensity of each salivary gland was shown in a box and whisker plot (FIG. 13C). 13-16 salivary gland cells per larva and at least six larvae per group were measured in each experiment. Statistical significance through unpaired Student’s t test. All data were presented as mean ± SD. ****P < 0.0001, ns: not significant. FIGS. 13D- 13E show a scatter plot of the ratio between the endogenous Nmnat-RA (STEM) and Nmnat- RB (MIMOSAS) (FIG. 13D), and RpL3 spliced mRNA variants (FIG. 13E) measured by two- color multiplex real-time PCR assay with specific TaqMan probes in the fly brains overexpressing microRNAs by pan-neuronal driver elavC155-GAL4 in both AGO1 wild-type (AGO1^+/+) and heterozygous mutant (AGO1^+/ ) genetic backgrounds. The ratio of RA vs RB as well as RG vs Total in the miR-9c expressing group with AGO1 wild-type background was set to 1, and the fold changes were displayed. All data were presented as mean ± SD. ***p<0.001, *P<0.05, unpaired Student’s t test; n=3, triplicate sampling.

FIGURE 14 depicts that nuclear localized microRNAs drive MIMOSAS in the presence of AGO1. The subcellular localizations of miR-210-3p were shown in the salivary gland cells of the larvae with the expression of miR-210 pri-microRNA in the motor neurons (using OK371 -GAL4). Scale bar was 20 pm for each row. The nucleus was marked with DAPI stains, while miR-210-3p was probed by its anti-sense FISH probe and sense probe is used as control.

FIGURES 15A-15C show a diagram of the overview of the present disclosure. FIG. ISA indicates the development of an in-silico approach to predict MIMOSAS-mediated alternative splicing regulation on a genome- wide level. FIG. 15B shows the combination of predictions with the analysis of ADRD RNAseq datasets and genomic profiles to identify ADRD relevant pairs of ncRNA and differential splice forms. FIG. ISC shows in vivo modeling to experimentally find MIMOSAS through innovative genetic splicing reporter systems using the splicing of exon 10 of Tau and exon 5 of NMNAT genes as examples.

FIGURES 16A-16C show the long-range stem-loop formations and MIMOSAS (MIcroRNA-Mediated Obstruction of Stem-loop Alternative Splicing). FIG. 16A shows MIMOSAS mechanism stipulates that nucleotide binding of Box 1 and 2 form a stem loop in the pre-mRNA that leads to the excision of exon 3. microRNA binding to either box disrupts the formation of the stem loop and leads to inclusion of exon 3. FIG. 16B shows the work identified a list of microRNAs that regulate the abundance of splice variants of NMNAT, Fas3 and Rpl3. Grey boxes mark the corresponding sequences that form stem-loop secondary structures in the pre-mRNA and targeting miRNA. Dashed box marks miR-137, miR-210, miR-278 and miR-1002 binding boxes 1 and 2 of NMNAT pre-mRNA, thereby disrupting stemloop and promoting spliceform RB. Figure 16C shows the experimental confirmation of the effect of microRNAs. Total mRNA was extracted from fly brains overexpressing microRNAs, and NMNAT RA and RB expression was determined with real-time PCR. Dashed boxes mark the microRNAs indicated in Figure 16B.

FIGURE 17 shows the microRNA mediated post-transcriptional regulation of Nmnat to enhance neuroprotection. Drosophila Nmnat transcription is upregulated directly through heat shock factor (HSF) or indirectly through hypoxia-inducible factor- la (HIF-la). Proteotoxic stress also induces transcriptional upregulation of DmNmnat. In neurons, the splicing of Nmnat is switched to Nmnat-RB under stress to promote the production of Nmnat- PD through miR-1002, therefore achieving neuronal protection.

FIGURE 18 shows the generation of alternative splicing reporter lines to monitor splicing events in vivo. The splicing reporter construct contains alternatively spliced exons 2- 4 with common N-terminal fragments of DsRed and AcGFP inserted in exon 2. The distinct sequences of DsRed and AcGFP are inserted into alternatively spliced exon 3 or 4. When stemloop structure forms, DsRed is expressed. When ncRNAs bind to either Box 1 or 2, MIMOSAS-based splicing expresses AcGFP. FIGURE 19 shows that the secondary structure reduces the donor- acceptor distance in secondary structure graph. The two Gs are 34 nucleotides apart along the sequence but separated by only 10 bonds in the secondary structure graph. An expected distance is obtained by averaging over all structures in the equilibrium ensemble.

FIGURES 20A-20B show the determination of splice-relevant binding between microRNA and mRNA. FIG. 20A shows the deep-learning model consists of an encoding module that extracts sequence from (non-)interacting pairs of microRNAs and mRNAs. One- hot encodings of microRNA and AGO-specific binding sites on mRNA are used. In the convolution/pooling layers, higher order features are first extracted from RNA sequence. Such representations are subjected to bi-directional LSTM layers that allow to extract sequential features to predict a probability of microRNA and mRNA (un-)pairing using a final soft-max layer. FIG. 20B shows the model assumes that the energy gained by the ncRNA-mRNA binding offsets the energy lost by opening the splice-relevant secondary structures. The difference of the energy of the mRNA structure with intact stem loop (Go) and absent structure (Gy) as DGopen = Go - Gi determines if the energy gained by binding of the ncRNA to the underlying mRNA DGdupiex- If DGdupiex > DGopen binding of the ncRNA is energetically more favorable.

FIGURE 21 shows the determination of splice-relevant SNPs. Genetic variations associated with ADRD are enriched in the vicinity of splice-relevant secondary structures.

FIGURES 22A-22C show the diagrams of experimental characterization of ADRD MIMOSAS candidates. FIG. 22A shows the MIMOSAS regulation of the pathogenic splice isoforms of Tau. FIG. 22B shows the MIMOSAS regulation of the neuroprotective isoforms of NMNAT. FIG. 22C shows the experimentally test top ADRD MIMOSAS candidates.

FIGURE 23 shows the altered splicing outcome of human Tau contributes to the pathogenesis of frontotemporal dementia. Local hairpin stem-loop formation of Tau facilitates the inclusion of exon 10. MicroRNA binding to the stem disrupts hairpin and allows exon 10 skipping. To measure such a molecular event, a splicing reporter, TauElOx, is designed that uses FP radio as a proxy for 4R/3R pathogenesis.

FIGURES 24A-24B show the human NMNAT genes are alternatively spliced. FIG. 24A shows the predicted alternative splicing of human NMNAT1, NMNAT2, and NMNAT3 are illustrated. FIG. 24B shows the RT-PCR analysis of total mRNA extracts from human embryonic brain at 87 and 110 days of gestation or HEK293T cells using variant-specific primer pairs indicated in FIG. 24A. Two variants of NMNAT3 have been experimentally identified. Arrowheads indicated unique fragments corresponding to variant specific transcripts.

FIGURE 25 shows the miRNAs of STEM 1 3’SS B0X1.

FIGURE 26 shows the miRNAs of STEM 1 3’SS B0X2.

FIGURE 27 shows the miRNAs of STEM 2 5’SS B0X1.

FIGURE 28 shows the miRNAs of STEM1 5’SS B0X2.

DETAILED DESCRIPTION

It is appreciated that certain features of the disclosure, which are, for clarity, described in the context of separate aspects, can also be provided in combination with a single aspect. Conversely, various features of the disclosure, which are, for brevity, described in the context of a single aspect, can also be provided separately or in any suitable subcombination. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. Methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present disclosure.

Definitions

In this specification and in the claims that follow, reference will be made to a number of terms, which shall be defined to have the following meanings:

As used herein, “comprising” is to be interpreted as specifying the presence of the stated features, integers, steps, or components as referred to, but does not preclude the presence or addition of one or more features, integers, steps, or components, or groups thereof. Moreover, each of the terms “by”, “comprising,” “comprises”, “comprised of,” “including,” “includes,” “included,” “involving,” “involves,” “involved,” and “such as” are used in their open, nonlimiting sense and may be used interchangeably. Further, the term “comprising” is intended to include examples and aspects encompassed by the terms “consisting essentially of’ and “consisting of.” Similarly, the term “consisting essentially of’ is intended to include examples encompassed by the term “consisting of.

As used in the specification and the appended claims, the singular forms “a,” “an” and “the” include plural referents unless the context clearly dictates otherwise. Thus, for example, reference to “a compound”, “a composition”, or “a cancer”, includes, but is not limited to, two or more such compounds, compositions, or cancers, and the like. It should be noted that ratios, concentrations, amounts, and other numerical data can be expressed herein in a range format. It can be further understood that the endpoints of each of the ranges are significant both in relation to the other endpoint, and independently of the other endpoint. It is also understood that there are a number of values disclosed herein, and that each value is also herein disclosed as “about” that particular value in addition to the value itself. For example, if the value “10” is disclosed, then “about 10” is also disclosed. Ranges can be expressed herein as from “about” one particular value, and/or to “about” another particular value. Similarly, when values are expressed as approximations, by use of the antecedent “about,” it can be understood that the particular value forms a further aspect. For example, if the value “about 10” is disclosed, then “10” is also disclosed.

When a range is expressed, a further aspect includes from the one particular value and/or to the other particular value. For example, where the stated range includes one or both of the limits, ranges excluding either or both of those included limits are also included in the disclosure, e.g. the phrase “x to y” includes the range from ‘x’ to ‘y ’ as well as the range greater than ‘x’ and less than ‘y’. The range can also be expressed as an upper limit, e.g. ‘about x, y, z, or less’ and should be interpreted to include the specific ranges of ‘about x’, ‘about y’, and ‘about z’ as well as the ranges of ‘less than x’, less than y’, and ‘less than z’. Likewise, the phrase ‘about x, y, z, or greater’ should be interpreted to include the specific ranges of ‘about x’, ‘about y’, and ‘about z’ as well as the ranges of ‘greater than x’, greater than y’, and ‘greater than z’. In addition, the phrase “about ‘x’ to ‘y’”, where ‘x’ and ‘y’ are numerical values, includes “about ‘x’ to about ‘y’”.

It is to be understood that such a range format is used for convenience and brevity, and thus, should be interpreted in a flexible manner to include not only the numerical values explicitly recited as the limits of the range, but also to include all the individual numerical values or sub-ranges encompassed within that range as if each numerical value and sub-range is explicitly recited. To illustrate, a numerical range of “about 0.1% to 5%” should be interpreted to include not only the explicitly recited values of about 0.1% to about 5%, but also include individual values (e.g., about 1%, about 2%, about 3%, and about 4%) and the subranges (e.g., about 0.5% to about 1.1%; about 5% to about 2.4%; about 0.5% to about 3.2%, and about 0.5% to about 4.4%, and other possible sub-ranges) within the indicated range.

As used herein, the terms “about,” “approximate,” “at or about,” and “substantially” mean that the amount or value in question can be the exact value or a value that provides equivalent results or effects as recited in the claims or taught herein. That is, it is understood that amounts, sizes, formulations, parameters, and other quantities and characteristics are not and need not be exact, but may be approximate and/or larger or smaller, as desired, reflecting tolerances, conversion factors, rounding off, measurement error and the like, and other factors known to those of skill in the art such that equivalent results or effects are obtained. In some circumstances, the value that provides equivalent results or effects cannot be reasonably determined. In such cases, it is generally understood, as used herein, that “about” and “at or about” mean the nominal value indicated ±10% variation unless otherwise indicated or inferred. In general, an amount, size, formulation, parameter or other quantity or characteristic is “about,” “approximate,” or “at or about” whether or not expressly stated to be such. It is understood that where “about,” “approximate,” or “at or about” is used before a quantitative value, the parameter also includes the specific quantitative value itself, unless specifically stated otherwise.

An "increase" can refer to any change that results in a greater amount of a symptom, disease, composition, condition, or activity. An increase can be any individual, median, or average increase in a condition, symptom, activity, composition in a statistically significant amount. Thus, the increase can be a 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100% or more increase so long as the increase is statistically significant.

A "decrease" can refer to any change that results in a smaller amount of a symptom, disease, composition, condition, or activity. A substance is also understood to decrease the genetic output of a gene when the genetic output of the gene product with the substance is less relative to the output of the gene product without the substance. Also, for example, a decrease can be a change in the symptoms of a disorder such that the symptoms are less than previously observed. A decrease can be any individual, median, or average decrease in a condition, symptom, activity, composition in a statistically significant amount. Thus, the decrease can be a 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, or 100% decrease so long as the decrease is statistically significant.

As used herein, the term “effective amount” refers to an amount that is sufficient to achieve the desired modification of a physical property of the composition or material. For example, an “effective amount” of a monomer refers to an amount that is sufficient to achieve the desired improvement in the property modulated by the formulation component, e.g. desired antioxidant release rate or viscoelasticity. The specific level in terms of wt% in a composition required as an effective amount will depend upon a variety of factors including the amount and type of monomer, amount and type of polymer, e.g., acrylamide, amount of antioxidant, and desired release kinetics. As used herein, the term “therapeutically effective amount” refers to an amount that is sufficient to achieve the desired therapeutic result or to have an effect on undesired symptoms but is generally insufficient to cause adverse side effects. The specific therapeutically effective dose level for any particular patient will depend upon a variety of factors including the disorder being treated and the severity of the disorder; the specific composition employed; the age, body weight, general health, sex and diet of the patient; the time of administration; the route of administration; the rate of excretion of the specific compound employed; the duration of the treatment; drugs used in combination or coincidental with the specific compound employed and like factors within the knowledge and expertise of the health practitioner and which may be well known in the medical arts. In the case of treating a particular disease or condition, in some instances, the desired response can be inhibiting the progression of the disease or condition. This may involve only slowing the progression of the disease temporarily. However, in other instances, it may be desirable to halt the progression of the disease permanently. This can be monitored by routine diagnostic methods known to one of ordinary skill in the art for any particular disease. The desired response to treatment of the disease or condition also can be delaying the onset or even preventing the onset of the disease or condition.

For example, it is well within the skill of the art to start doses of a compound at levels lower than those required to achieve the desired therapeutic effect and to gradually increase the dosage until the desired effect is achieved. If desired, the effective daily dose can be divided into multiple doses for purposes of administration. Consequently, single dose compositions can contain such amounts or submultiples thereof to make up the daily dose. The dosage can be adjusted by the individual physician in the event of any contraindications. It is generally preferred that a maximum dose of the pharmacological agents of the invention (alone or in combination with other therapeutic agents) be used, that is, the highest safe dose according to sound medical judgment. It will be understood by those of ordinary skill in the art however, that a patient may insist upon a lower dose or tolerable dose for medical reasons, psychological reasons or for virtually any other reasons.

A response to a therapeutically effective dose of a disclosed drug delivery composition can be measured by determining the physiological effects of the treatment or medication, such as the decrease or lack of disease symptoms following administration of the treatment or pharmacological agent. Other assays will be known to one of ordinary skill in the art and can be employed for measuring the level of the response. The amount of a treatment may be varied for example by increasing or decreasing the amount of a disclosed compound and/or pharmaceutical composition, by changing the disclosed compound and/or pharmaceutical composition administered, by changing the route of administration, by changing the dosage timing and so on. Dosage can vary, and can be administered in one or more dose administrations daily, for one or several days. Guidance can be found in the literature for appropriate dosages for given classes of pharmaceutical products.

As used herein, the term “prophylactically effective amount” refers to an amount effective for preventing onset or initiation of a disease or condition.

As used herein, the term “prevent” or “preventing” refers to precluding, averting, obviating, forestalling, stopping, or hindering something from happening, especially by advance action. It is understood that where reduce, inhibit or prevent are used herein, unless specifically indicated otherwise, the use of the other two words is also expressly disclosed.

As used herein, the terms “optional” or “optionally” means that the subsequently described event or circumstance can or cannot occur, and that the description includes instances where said event or circumstance occurs and instances where it does not.

As used interchangeably herein, “subject,” “individual,” or “patient” can refer to a vertebrate organism, such as a mammal (e.g. human). "Subject" can also refer to a cell, a population of cells, a tissue, an organ, or an organism, preferably to human and constituents thereof.

As used herein, the terms "treating" and "treatment" can refer generally to obtaining a desired pharmacological and/or physiological effect. The effect can be, but does not necessarily have to be, prophylactic in terms of preventing or partially preventing a disease, symptom or condition thereof. The effect can be therapeutic in terms of a partial or complete cure of a disease, condition, symptom or adverse effect attributed to the disease, disorder, or condition. The term "treatment" as used herein can include any treatment in a subject, particularly a human and can include any one or more of the following: (a) preventing the disease from occurring in a subject which may be predisposed to the disease but has not yet been diagnosed as having it; (b) inhibiting the disease, i.e., arresting its development; and (c) relieving the disease, i.e., mitigating or ameliorating the disease and/or its symptoms or conditions. The term "treatment" as used herein can refer to both therapeutic treatment alone, prophylactic treatment alone, or both therapeutic and prophylactic treatment. Those in need of treatment (subjects in need thereof) can include those already with the disorder and/or those in which the disorder is to be prevented. As used herein, the term "treating", can include inhibiting the disease, disorder or condition, e.g., impeding its progress; and relieving the disease, disorder, or condition, e.g., causing regression of the disease, disorder and/or condition. Treating the disease, disorder, or condition can include ameliorating at least one symptom of the particular disease, disorder, or condition, even if the underlying pathophysiology is not affected, e.g., such as treating the pain of a subject by administration of an analgesic agent even though such agent does not treat the cause of the pain.

The term “administer,” “administering”, or derivatives thereof refer to delivering a composition, substance, inhibitor, or medication to a subject or object by one or more the following routes: oral, topical, intravenous, subcutaneous, transcutaneous, transdermal, intramuscular, intra-joint, parenteral, intra- arteriole, intradermal, intraventricular, intracranial, intraperitoneal, intralesional, intranasal, rectal, vaginal, by inhalation or via an implanted reservoir. The term “parenteral” includes subcutaneous, intravenous, intramuscular, intraarticular, intra-synovial, intrasternal, intrathecal, intrahepatic, intralesional, and intracranial injections or infusion techniques.

As used herein, “dose,” “unit dose,” or “dosage” can refer to physically discrete units suitable for use in a subject, each unit containing a predetermined quantity of a disclosed compound and/or a pharmaceutical composition thereof calculated to produce the desired response or responses in association with its administration.

As used herein, “therapeutic” can refer to treating, healing, and/or ameliorating a disease, disorder, condition, or side effect, or to decreasing in the rate of advancement of a disease, disorder, condition, or side effect.

A “nucleic acid” refers to a chemical compound that serves as the primary informationcarrying molecules in cells and make up the cellular genetic material. Nucleic acids comprise nucleotides, which are the monomers made of a 5-carbon sugar (usually ribose or deoxyribose), a phosphate group, and a nitrogenous base. A nucleic acid can also be a deoxyribonucleic acid (DNA) or a ribonucleic acid (RNA), including but not limited to a non-coding (nc) RNA. A chimeric nucleic acid comprises two or more of the same kind of nucleic acid fused together to form one compound comprising genetic material.

The terms “percent identity” and “% identity,” as applied to polynucleotide sequences, refer to the percentage of residue matches between at least two polynucleotide sequences aligned using a standardized algorithm. Such an algorithm may insert, in a standardized and reproducible way, gaps in the sequences being compared in order to optimize alignment between two sequences, and therefore achieve a more meaningful comparison of the two sequences. Percent identity for a nucleic acid sequence may be determined as understood in the art. (See, e.g., U.S. Pat. No. 7,396,664, which is incorporated herein by reference in its entirety). A suite of commonly used and freely available sequence comparison algorithms is provided by the National Center for Biotechnology Information (NCBI) Basic Local Alignment Search Tool (BLAST) (Altschul, S. F. et al. (1990) J. Mol. Biol. 215:403 410), which is available from several sources, including the NCBI, Bethesda, Md., at its website. The BLAST software suite includes various sequence analysis programs including “blastn,” that is used to align a known polynucleotide sequence with other polynucleotide sequences from a variety of databases. Also available is a tool called “BLAST 2 Sequences” that is used for direct pairwise comparison of two nucleotide sequences. “BLAST 2 Sequences” can be accessed and used interactively at the NCBI website. The “BLAST 2 Sequences” tool can be used for both blastn and hlastp (discussed above).

Percent identity may be measured over the length of an entire defined polynucleotide sequence or may be measured over a shorter length, for example, over the length of a fragment taken from a larger, defined sequence, for instance, a fragment of at least 20, at least 30, at least 40, at least 50, at least 70, at least 100, or at least 200 contiguous nucleotides. Such lengths are exemplary only, and it is understood that any fragment length may be used to describe a length over which percentage identity may be measured.

A “full length” nucleic acid sequence is one containing at least a translation initiation codon (e.g., methionine) followed by an open reading frame and a translation termination codon. A “full length” polynucleotide sequence encodes a “full length” polypeptide sequence. A “variant,” “mutant,” or “derivative” of a particular nucleic acid sequence may be defined as a nucleic acid sequence having at least 50% sequence identity to the particular nucleic acid sequence over a certain length of one of the nucleic acid sequences using blastn with the “BLAST 2 Sequences” tool available at the National Center for Biotechnology Information's website. (See Tatiana A. Tatusova, Thomas L. Madden (1999), “Blast 2 sequences — a new tool for comparing protein and nucleotide sequences”, FEMS Microbiol Lett. 174:247-250). In some embodiments a variant polynucleotide may show, for example, at least 60%, at least 70%, at least 80%, at least 90%, at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, or at least 99% or greater sequence identity over a certain defined length relative to a reference polynucleotide.

Synthetic Nucleic Acid Compositions

In one aspect, disclosed herein is a synthetic nucleic acid comprising a reporter gene, a first stem loop, a second stem loop, a nuclear export signal, and an exon 10 fragment from the human Tau gene, wherein the nuclear export signal and the exon 10 fragment from the human Tau gene are located between the first stem loop and the second stem loop.

In one aspect, disclosed herein is a synthetic nucleic acid comprising a reporter gene, a fragment from an Nmnat gene, wherein the fragment from the Nmnat gene comprises a stem loop comprising a box 1 sequence and a box 2 sequence, wherein the box 1 sequence and the box 2 sequence hybridize to form a stem of the stem loop.

As used herein, a “composition” refers to any agent that has a beneficial biological effect. Beneficial biological effects include both therapeutic effects, e.g., treatment of a disorder or other undesirable physiological condition, and prophylactic effects, e.g., prevention of a disorder or other undesirable physiological condition. The terms also encompass pharmaceutically acceptable, pharmacologically active derivatives of beneficial agents specifically mentioned herein, including, but not limited to, a vector, polynucleotide, cells, salts, esters, amides, proagents, active metabolites, isomers, fragments, analogs, and the like. When the term “composition” is used, then, or when a particular composition is specifically identified, it is to be understood that the term includes the composition per se as well as pharmaceutically acceptable, pharmacologically active vector, polynucleotide, salts, esters, amides, proagents, conjugates, active metabolites, isomers, fragments, analogs, etc. In some aspects, the composition disclosed herein comprises the synthetic nucleic acid of any disclosed aspect.

In some embodiments, the box 1 sequence comprises SEQ ID NO: 1. In some embodiments, the box 2 sequence comprises SEQ ID NO: 2.

Methods

In some embodiments, the increased expression of the reporter gene in the cytoplasm provides a pathological outcome. In some embodiments, the increased expression of the reporter gene in the nucleus provides a non-pathological outcome. In one aspect, disclosed herein is a method of diagnosing Alzheimer’ s Disease or an increased risk of developing Alzheimer’s Disease, comprising providing the synthetic nucleic acid of any preceding aspect, incubating the synthetic nucleic acid with a miRNA, and determining whether the miRNA modulates the alternative splicing of exon 10 of the human Tau gene, wherein increased expression of the reporter gene in the cytoplasm confers a pathological outcome and a diagnosis of Alzheimer’s Disease or an increased risk of developing Alzheimer’s Disease.

In some embodiments, the miRNA is selected from the group consisting of hsa-miR- 541-3p, hsa-miR-505-5p, hsa-miR-1910-3p, hsa-miR-135a-3p, hsa-miR-362-5p, hsa-miR-9- 5p, hsa-miR-103a-2-5p, hsa-miR-887-5p, hsa-miR575, hsa-miR-589-3p, hsa-miR-378c, hsa- miR-378a-3p, hsa-miR378d, hsa-miR-346, hsa-miR-378h, hsa-miR-557, hsa-miR-500a-3p, hsa-miR-502-3p, hsa-miR-1271-5p, hsa-miR-5001-3p, hsa-miR-664b-3p, hsa-miR-101-5p, hsa-miR-204-5p, hsa-miR-764, hsa-miR-370-3p, hsa-miR-1343-3p, hsa-miR-371a-3p, hsa- miR-4485-5p, hsa-miR-2682-3p, hsa-miR-331-3p, hsa-miR-660-3p, hsa-miR-1306-5p, hsa- miR-6756-3p, hsa-miR-1908-3p, hsa-miR-6894-3p, hsa-miR-6760-3p, hsa-miR-6726-3p, hsa- miR-23b-3p, hsa-miR-1178-3p, hsa-miR-149-5p, hsa-miR-23a-3p, hsa-miR-324-3p, hsa-miR- 342-3p, hsa-miR-20b-5p, hsa-miR-106a-5p, hsa-miR-525-3p, hsa-miR-106b-5p, hsa-miR- 18b-5p, hsa-miR-28-5p, hsa-miR-6775-5p, hsa-miR-764, hsa-miR-3689d, hsa-miR-3622b-5p, hsa-miR-1294, hsa-miR-92b-5p, hsa-miR-30b-3p, hsa-miR-516b-5p, hsa-miR-299-3p, hsa- miR-769-3p, and hsa-miR-7114-5p.

In some embodiments, the specific treatment for Alzheimer’s Disease comprises drugs that are approved by the U.S. Food and Drug Administration (FDA) for Alzheimer’s disease to help either manage symptoms or treat the disease. In particular, current medications include cholinesterase inhibitors that prevent the breakdown of acetylcholine, a brain chemical important for memory and cognition, as well as N-methyl-D-aspartate (NMDA) antagonists that regulate glutamate, an important brain chemical.

In one aspect, disclosed herein is a method of treating or preventing Alzheimer’s Disease in a subject in need thereof, comprising administering to the subject a miRNA selected from the group consisting of hsa-miR-541-3p, hsa-miR-505-5p, hsa-miR-1910-3p, hsa-miR- 135a-3p, hsa-miR-362-5p, hsa-miR-9-5p, hsa-miR-103a-2-5p, hsa-miR-887-5p, hsa-miR575, hsa-miR-589-3p, hsa-miR-378c, hsa-miR-378a-3p, hsa-miR378d, hsa-miR-346, hsa-miR- 378h, hsa-miR-557, hsa-miR-500a-3p, hsa-miR-502-3p, hsa-miR-1271 -5p, hsa-miR-5001 -3p, hsa-miR-664b-3p, hsa-miR-101-5p, hsa-miR-204-5p, hsa-miR-764, hsa-miR-370-3p, hsa- miR-1343-3p, hsa-miR-371a-3p, hsa-miR-4485-5p, hsa-miR-2682-3p, hsa-miR-331-3p, hsa- miR-660-3p, hsa-miR-1306-5p, hsa-miR-6756-3p, hsa-miR-1908-3p, hsa-miR-6894-3p, hsa- miR-6760-3p, hsa-miR-6726-3p, hsa-miR-23b-3p, hsa-miR-1178-3p, hsa-miR-149-5p, hsa- miR-23a-3p, hsa-miR-324-3p, hsa-miR-342-3p, hsa-miR-20b-5p, hsa-miR-106a-5p, hsa-miR- 525-3p, hsa-miR-106b-5p, hsa-miR-18b-5p, hsa-miR-28-5p, hsa-miR-6775-5p, hsa-miR-764, hsa-miR-3689d, hsa-miR-3622b-5p, hsa-miR-1294, hsa-miR-92b-5p, hsa-miR-30b-3p, hsa- miR-516b-5p, hsa-miR-299-3p, hsa-miR-769-3p, and hsa-miR-7114-5p.

In one aspect, disclosed herein is a method for identifying microRNA mediated obstruction of stem-loop alternative splicing, comprising identifying a target sequence of interest, and identifying a microRNA sequence from a miRNA that can hybridize to the target sequence, wherein the binding of the microRNA sequence to the target sequence obstructs and/or modulates the alternative splicing of a gene comprising the target sequence. In one aspect, disclosed herein is a method for modulating expression of a gene, comprising identifying a target sequence of interest in the gene, and identifying a microRNA sequence from a miRNA that can hybridize to the target sequence of interest in the gene, wherein the binding of the microRNA sequence to the target sequence modulates the expression of the gene.

EXAMPLES

Example 1: MicroRNA-Mediated Obstruction of Stem-loop Alternative Splicing (MIMOSAS): a mechanism for the regulation of alternative splicing in Drosophila

Alternative splicing plays an important role in cellular differentiation and stress response in both physiological and pathological conditions. While the vast majority of human genes are alternatively spliced (Bush et al., 2017; Chen et al., 2014; Kim et al., 2018), 22% of disease-causing mutations are splicing sensitive (Lim et al., 2011), and 25% of disease-causing exonic mutations induce exon skipping (Sterne-Weiler et al., 2011). Recent evidence points to the active roles of noncoding RNAs in alternative splicing. For example, long non-coding RNA (IncRNA) can promote the retention of a particular exon in FGFR2 alternative splicing (Gonzalez et al., 2015). Furthermore, double stranded RNAs can affect the splicing of genes associated with muscular dystrophy, while a small nucleolar RNA was found to regulate the splicing of Serotonin Receptor IIC, that is associated with Prader-Willi Syndrome (Kishore and Stamm, 2006). MicroRNAs (miRNAs) as a highly conserved group of small non-coding RNA molecules, play key roles in gene regulation. While miRNA targeting rules are quite complex and flexible, miRNAs can bind to a wide variety of targets, including UTRs, coding sequences and non-coding RNA through both canonical and noncanonical base pairings (Helwak et al., 2013). Moreover, a study mapping Kaposi's sarcoma-associated herpes virus (KSHV) microRNA binding sites in B cell lymphoma-derived cells and epithelial cells showed that a majority of microRNA target sites were found within the coding sequence of mRNAs, instead of the 3’UTR (Gay et al., 2018). Such observations of microRNA target sites in exonic and intronic regions may indicate the microRNAs to regulate splicing and influence alternative splicing outcomes in addition to regulating transcript stability and translation.

MicroRNAs are short noncoding RNAs (~22nt) that have been well studied in mRNA regulation (Carthew et al., 2017; Chawla et al., 2016) typically binding complementary nucleotide sequences in the 3’ UTR of target mRNAs through the RNA-induced silencing complex (RISC) (O'Brien et al., 2018). While the vast majority of microRNA-target interactions result in decreased translation of the target (Gebert and MacRae, 2019), microRNAs have been reported to upregulate translation of their target mRNAs (Cordes et al., 2009; Ghosh et al., 2008; Jopling et al., 2005; Lu et al., 2010; Mortensen et al., 2011; Orom et al., 2008; Tsai et al., 2009; Vasudevan et al., 2007), often through binding the 5’UTR (Jopling et al., 2005; Orom et al., 2008; Tsai et al., 2009). Given the prevalence of microRNA-RNA interaction, a study was conducted which predicted that the binding of microRNAs can disrupt splice-relevant RNA secondary structures of targets, pointing to an exciting mechanism of splicing and RNA secondary structure regulation (Gueroussov et al., 2017; Raker et al., 2009), which was term MicroRNA-Mediated Obstruction of Stem-loop Alternative Splicing (MIMOSAS). Previously, splice-relevant secondary structures have been identified purely on the basis of sequence composition and computational RNA structural modelling of the NMNAT pre- mRNA (Lorenz et al., 2011). In particular, splice junctions were scanned for complementary sequences (i.e., boxes) and corroborated the emergence of a stem-based loop secondary structure that captured intermediate introns (Gruber et al., 2008). Based on the assumption that the duplex of the binding microRNA and mRNA is (i) located at the stem of a hairpin loop and is (ii) energetically more favorable than the formation of the stem loop itself, this study predicted that miR-1002 interfered with a splice-relevant stem loop forming on the Nmnat mRNA that regulated the alternative splice variants RA and RB. Most notably, the study corroborated the computational prediction by experimentally demonstrating the predicted effect of miR-1002 in shifting the splicing towards variant RB and away from RA (Park et al., 2019).

While initial consideration was primarily limited to a pair of one microRNA and mRNA, here this study shows that MIMOSAS is a genome-wide occurring mechanism of splicing regulation. In particular, the study designed a computational pipeline that predicts microRNAs that disrupt such secondary structures on a genome-wide scale in Drosophila. Most importantly, the study experimentally corroborated the computational predictions in vivo and in cultured mammalian cell models.

Materials and Methods

Plasmid construction'. pBlD-LJASC was a gift from Brian McCabe (Addgene plasmid #35200) and pcDNA3.1_sfCherry2 (1-10) was a gift from Bo Huang (Addgene plasmid # 82602; RRID:Addgene_82602). The alternative splicing reporters of Drosophila Nmnat and RpL3 were synthesized and purchased from VectorBuilder. Seven recombinant plasmids were generated for this study. They are pBID-UASC-Amnat_AltReport v2, pLV[Exp]-puro-EFl A- iReporl. and pcDNA3.1_sfCherry2(l-10)-miR-9c (miR-210, miR304, miR-988, miR-992). The primer sequences used in detection and cloning are listed in TABLE 1.

Fly stocks and culture'. Flies were maintained on a cornmeal-molasses-yeast medium at room temperature (22°C) with 60-65% humidity. All of the Drosophila lines were obtained from the Bloomington Stock Center, including elav-GAL4 and nysb-GAL4. as well as the overexpression and knockdown lines of the microRNAs: UAS-LUC-miRs and UAS- mCherry.miRs. sponge. V2.

RNA Extraction'. Total RNA was extracted from at least 40 fly heads per group by FavorPrep tissue total RNA purification kit (Favorgen), according to the manufacturer’s protocol. For each extraction, RNA concentration was measured spectrophotometrically at 260 nm, and 2 pg of RNA was used for reverse transcription reaction with a high-capacity cDNA reverse transcription kit (Applied Biosystems).

Florescence in situ hybridization (FISH): Dissected salivary glands from wandering third instar larvae were fixed with 4% paraformaldehyde (PFA) in Schneider’s Drosophila Medium at room temperature for 10 min, washed one time with lx PBS for 5 min at room temperature. Followed with Removing the PBS and washed the tissue for at least 10 min with 100 pl of in situ mix: (Combine 12.5 mL of 100% formamide, 12.5 ml of 20x SSC, 0.9 ml of 0.5 M citric acid, pH 6, 0.75 ml of 20% TritonTM X-100, and 23.35 ml nuclease-free H2O to make a final volume of 50 ml.) at room temperature. The probe mix was prepared by diluting the probe to 3 pg/sample and DAPI (to final concentration of 1 pg/ml) in the in situ mix. Specimen was incubated with the probe for overnight at 38°C. After the incubation, the probe was removed and the specimen was washed for at least 10 min with 100 pl of in situ mix with 1 pg/ml DAPI at room temperature 3.

Drosophila RpL3 Variant Detection-. Standard PCR was performed using the FailSafe PCR System (EpiCentre, Chicago, IL, USA) with the following amplification conditions: 25 cycles of 50 seconds at 95°C, 50 seconds at 60°C and 1 min at 72°C. The primers designed (TABLE 1) for variant detection are common forward primer spanning the exon-exon junction between exon 2 and 3, and three distinct reverse primers targeted unique coding sequences for each spliced mRNA variant located in exon 2 or exon 3, respectively (FIGS. 10A-10B).

Real-time PCR: Quantification of mRNA levels was performed using a CFX connect real-time detection system (Bio-Rad) and TaqMan probe-based gene expression analysis (Applied Biosystems). The amplification mix (20 pl) contained 100 ng of ssDNA reverse transcribed from total RNA, and 1 pl of gene-specific TaqMan probe-primer set. The samples were amplified by a two-color multiplex real-time PCR program of 40 cycles of 10 seconds at 95°C, 15 seconds at 55°C, and 1 min at 72°C. The quantification of mRNA levels was carried out by the 2(-Delta Delta C(T)) Method 4. Seven TaqMan probes were used in this study. They are FAM-DmO215O883_gl, VIC-DmO2144515_gl, FAM-Dm01810909_gl, VIC-

Dm01810910_ml, FAM-Dm02135667_gl, FAM-Dm02135669_gl, VIC-Dm02148683_gl.

In-cell reporter assay: Cos-7 cells were co-transfected with lipofectamine 2000 (Life Technologies). 1-2 pg of cDNA was diluted into 100 pl of Opti-MEM I Medium (Invitrogen) and mixed gently. Lipofectamine 2000 mixture was prepared by diluting 2-4 pl of Lipofectamine 2000 in 100 pl of Opti-MEM I Medium. The ratio of DNA to Lipofectamine 2000 used for transfection was 1:2 as indicated in the manual. The DNA-Lipofectamine 2000 mixture was mixed gently and incubated for 20 min at room temperature. Cells were directly added to the 200 pl of DNA-Lipofectamine 2000 mixture. After 48 h transfection, cells were fixed and stained with 4,6-diamidino-2-phenylindole (DAPI) and Lamin A/C conjugated Alexa Fluor® 647 (1:500, Cell Signaling Technology #41357). Samples were visualized with an Olympus 1X81 confocal microscope under x 60 magnification. In-Fly reporter assay. Adult brains were fixed in phosphate buffered saline (PBS) with 3.7% formaldehyde for 15 min and washed in phosphate buffered saline with 0.4% Triton X- 100. DAPI (1 : 1000, Invitrogen) staining was performed after 3 times wash. Before imaging, tissues were mounted on microscope slides in Vectashield Mounting Medium for Fluorescence (Vector Laboratories).

Confocal Image Acquisition and Processing: Confocal microscopy was performed with an Olympus 1X81 confocal microscope and processed using FluoView 10-ASW (Olympus) or ImageJ (NTH), and Adobe Photoshop 2023 (Adobe, USA).

Quantification and Statistical Analysis: Data were represented as mean ± s.d.. Statistical analysis was performed in GraphPad Prism using unpaired Student’s t-test. Values of P < 0.05 were considered statistically significant.

Selection of candidate stems: Starting point for the candidate selection was a list of evolutionary conserved complementary regions, termed boxes, identified in 5. For each pair of boxes in a gene of interest, the study checked whether the secondary structure induced by pairing of the boxes could influence isoform selection. Minimum free energy (MFE) secondary structures and base pairing probabilities were computed using the RNAfold program of the ViennaRNA package 6 for a region of the pre-mRNA encompassing both boxes and nearby exon-intron boundaries to ensure that the boxes indeed from a stem with high probability. In some cases, additional pairs of boxes were identified in this step.

Identification of MIMOSA miRNAs with RNAup: For each miRNA with a binding site overlapping the boxes, the secondary structure that would be formed after miR binding, was predicted using RNAfold with constraints that forbid the target site to form intra-molecular base pairs. The predicted structures were inspected to check whether the box interaction was indeed destroyed as well as to compute the opening energy for miR binding as the difference between the folding energies with and without constraints. Moreover, the study used RNAup 7 to compute the strength of interaction between the miRNA and mRNA. The free energy of binding computed by RNAup consists of the duplex energy, i.e. the free energy of the intermolecular duplex formed between miR and mRNA, as well as the opening energy necessary to make the binding sites accessible for interaction, AGMIMOSAS = Gdupiex - AG_Open- In principle, the binding free energy allows to compute the fraction of bound mRNAs if both miR and mRNA concentrations were known. In addition, RNAup computes the conditional probability that the interaction covers a given position. A value close to 1 ensures that no other favorable binding sites exist in the region considered. Since most splicing events occur co- transcrip tionally, the study mimicked the effects of co-transcriptional folding by performing RNAup computations for three different windows along the mRNA, roughly representing three different time points during transcription.

Results

In silico transcriptome-wide screen for stem-loop mediated alternative splicing: The splicing of long introns is likely facilitated by the formation of RNA secondary structures within the intronic region of the pre-mRNA. This process brings the 5’ and 3’ splice sites into relatively close spatial proximity, thereby promoting the excision of specific introns. While many of these long-range interactions are evolutionarily conserved and strongly associated with alternative splicing (Kalmykova et al., 2021 ; Pervouchine et al., 2012; Raker et al., 2009), a crucial molecular mechanism underlying this phenomenon involves the formation of stemloop RNA structures between complementary sequences, referred to as “boxes” (FIG. 1A). When these box sequences are situated near splice sites within introns or exons, the formation of the stem-loop structure brings distant splice sites together, facilitating long-range splicing (FIG. 1A). Using a previous screen of highly evolutionarily conserved complementary “box” sequences near splice sites in Drosophila, the study identified 49 genes containing such regions of splice-relevant stem-loop structures with the influence of alternative splicing outcomes (Kalmykova et al., 2021; Pervouchine et al., 2012). Notably, these genes are distributed across all four chromosomes showing the broad and abundant presence of splice-relevant RNA secondary structures (FIG. IB).

The first step of the bioinformatic pipeline involves the identification of sequences within the pre-mRNA of a gene that have formed secondary structures around splice-relevant sites (FIG. 1C). While this approach identified sequences capable of forming stems, the study also assessed the energetic aspects of stem formation. Using RNA secondary structure prediction algorithm RNAplex, the study confirmed the formation of stem structures, indicating the presence of "stem-loop" structures that influence alternative splicing outcomes. In particular, the study calculated the free energy G of the hybridization sites of the boxes including flanking regions 20 nucleotides downstream of boxes I and half the genetic distance to box II as well as the remaining distance to boxes II and 20 nucleotides upstream of boxes II (Tafer and Hofacker, 2008). (FIG. 1C). In TABLE 1, the study observed that the vast majority of box pairs indeed formed stems with G < -15kcal/mol. In the second step of the pipeline these splice-relevant secondary structures were subjected to predicting overlapping microRNA binding sites, assuming that the energy gained from microRNA-mRNA binding offsets the energy lost from opening the splice-relevant secondary structures. Specifically, the study calculated the energy difference between mRNA structures with intact (Go) and absent stem (AGopen - Go - Gi). A microRNA candidate was considered if the energy of the duplex binding structure between a microRNA and mRNA (Gdupiex) exceeded AGopen, indicating that microRNA binding, which disrupts the splice-relevant secondary structure, is energetically favorable (Lorenz et al., 2011 ; Lorenz et al., 2016) (FIG. 1C). It was observed that in almost all cases microRNAs were found that bind the underlying complementary box sequences, disrupting stem formation. To refine the predictions in the third step, the study considered that nascent mRNAs promptly form secondary structures upon emergence from RNA polymerase. Consequently, the study analyzed a sequence window upstream of the 5’ end of a box, calculating the energy difference between the duplex with the underlying microRNA (Gdupiex) and opened structure ( G_o;,_en), denoted as AGMIMOSAS = Gdupiex - AG_open, along with the probability of a microRNA binding to a given box (Lorenz et al., 2011; Muckstein et al., 2006) (FIG. 1C).

In vivo analysis of microRNAs driving Fas3 splice variants'. To experimentally test the predictions, the study started with the gene Fas3 as an example that is multi-exonic and is predicted to express numerous splice variants. While the initial screen pointed to complementary sequences around splice-donor and acceptor sites flanking exon 5, the study did not find a stem forming in the initial screen. However, sequence complementarity does not automatically lead to the formation of a stem, prompting the study to refine the prediction by considering the ensemble energy of secondary structures formed on the sequence encompassing exon 5 and its 3’ and 5’ flanking introns. Plotting probabilities of nucleotide interactions in the underlying mRNA subsequence (FIG. 2) highlighted alternative boxes overlapping with one box that was found through initial sequence complementarity. Additionally, the study identified a stem encompassing these boxes when determining the energetically most stable structure of the underlying mRNA subsequence (FIG. 2). Using the box sequences from FIG. 3A, the study identified microRNAs binding to box I and II (FIG. 3B), that were sorted based on their AGMIMOSAS- Specifically, the study selected two microRNAs targeting the underlying boxes with differing energetic favorability. It was anticipated that identified candidate microRNAs would disrupt the stem structure, favoring the RC splice variant, while other splice variants (RA/B/D/E/F/G) would predominate in the absence of MIMOSAS (FIG. 3C). Refining the predictions further, the study analyzed sequence windows of 200 nucleotides upstream of the 5’ end of a box, shifting upstream in increments of the length of the underlying box. By determining the energetics of microRNAs binding to the mRNA subsequence, it was notably found that miR-973 exhibited a high AGMIMOSAS and binding probability in the box I region, pointing to a prime candidate for driving the RC splice variant, while miR-976 showed no significant binding ability in the box region (FIG. 3D). Similarly, miR-1000 emerged as a prime MIMOSAS candidate binding to box II, contrasting with the negligible binding observed for miR-999 (FIG. 3E). To experimentally validate the predictions, the study conducted real-time PCR after over-expressing corresponding microRNAs in Drosophila brains, allowing observation of the emergence of different splice variants in vivo. These results demonstrated that miR-973 and miR-1000 significantly increased the preference of the RC (MIMOSAS) splice variant, while no significant changes were observed when miR-976 or miR-1000 were overexpressed (FIG. 3F).

In vivo analysis of microRNAs driving Nmnat splice variants: Previously, miR-1002 was identified as a binder that disrupted splice-relevant stem- loop formation in the Nmnat pre- mRNA, distinguishing between splice variants RA and RB (Park et al., 2019). Subsequently, by expressing miR-1002, the anticipated shift toward the predicted neuroprotective splice variant RB, away from RA, was observed. Initially, the study identified splice-relevant stems in Nmnat solely through complimentary subsequences located at the splice sites in exon 5 and the upstream intron. To enhance the detection of these secondary structures, the study analyzed the ensemble of secondary structures within exon 5 and its flanking introns in the pre-mRNA of Nmnat (FIG. 4), confirming substantial overlap between the identified stem and the boxes that were found through complementary sequences. The formation of the stem corresponds to the production of the RA splice variant, where exon 5 is spliced out, while the absence of stem formation leads to the emergence of the RB splice variant, retaining exon 5 (FIG. 5A). Interestingly, the identified pair of boxes encompassing exon 5 were targeted by numerous candidate microRNAs in addition to miR-1002 (FIG. SB). By focusing on sequence windows of 200 nucleotides upstream of the 5’ end of the underlying boxes and assessing microRNA binding energies, the study identified several microRNAs that bind to box I and II, ranging across a spectrum of binding strengths (FIG. 6). For instance, miR-137 exhibited strong binding probability with box I, while miR-307a showed no binding signal. Similarly, miR-278 and miR-983 demonstrated robust probabilities of binding to box II (FIG. 6). To validate the predictions, the study conducted real-time PCR experiments after expressing corresponding microRNAs or microRNA sponges to reduce endogenous microRNAs. MicroRNAs with favorable binding characteristics, such as miR-210, miR-137 (targeting box I) as well as miR- 983, miR-9b, miR-278 and miR-1002 (targeting box II) indeed promoted the emergence of the MIMOSAS-specific RB splice variant. In turn, miR-307a and miR-193, exhibiting minimal binding qualities, had no effect on RB production (FIG. 5C and FIGS. 7A-7B). To refine the ability to monitor splicing events with greater accuracy and to directly visualize MIMOSAS events in vivo, the study developed fluorescence-based genetic reporters engineered to track alternative splicing events in a cell type-specific manner. Improving a previous version of the alternative splicing reporter for Drosophila Nmnat (Ruan et al., 2015), modifications were made by moving N-terminal sequence encoding the first seven amino acids including the start codon that is shared between EGFP and mCherry to the 5 ’ of endogenous intron containing box I. Furthermore, the C-terminal unique sequence of EGFP was fused with the endogenous 3’UTR of Mww/-RB variant located in exon 5 containing box II that is connected by an endogenous intron 5, followed by the C-terminal unique sequence of mCherry (FIG. 5D). As a result of alternative splicing, the EGFP expression and fluorescence indicates MIMOSAS events, while mCherry expression requires alternative splicing mediated by stem formation between two box sequences, indicated as STEM (FIG. 5D). Importantly, this design allowed precise measurement of splice variant production, as an incomplete splice product would lack functional fluorescence. To visualize MIMOSAS events in vivo, the study generated a transgenic fly line carrying the UAS-AltSplicing Reporter and expressed both microRNAs and splicing reporter in neurons by nysb-GAL4. A significant increase in the fluorescent intensity ratio of EGFP (MIMOSAS) and mCherry (STEM) signals in the brains was observed upon expression of miR-210, miR-9b, miR-983, or miR-278 compared to the expression of non-targeting miRNA (miR-9c) or negative control (miR-307a) (FIGS. 5E-5F). MIMOSAS events in vivo in the brain visualized with the genetic fluorescence reporter further validated the computational predictions and underscored the robustness of the microRNAs in regulating alternative splicing.

In vivo analysis of microRNAs driving RpL3 splice variants'. Both Fas3 and Nmnat exemplified alternative splicing regulation through a single stem-loop based secondary structure. To explore more intricate secondary structures in pre-mRNAs, the study turned to RpL3, where a nested stem loop structure was discovered in addition to the previously noted complementary pair of boxes I and II near splice sites in exon 3 (Kalmykova et al., 2021 ; Pervouchine et al., 2012). Analyzing the energetically most stable secondary structure of exon 3, the study confirmed stem formation between the complementary sequences (FIG. 8). However, examining the most stable structure involving exons 3 and 4 revealed that box II forms a distinct stem with a newly identified box III in exon 4 (FIG. 8). Consequently, the formation of splice-relevant stems between boxes I and II, or boxes II and III, enable RpL3 pre- mRNA to generate additional splice variants. Specifically, stem formation between boxes 11/111 produces RA/H while stem formation between boxes Eli leads to RG, and an open confirmation without any stems results in RD (FIG. 9A). Remarkably, the study found splice variant RA/H (from stem II/III) to be most prevalent, constituting 94% of total transcripts, while RG (from stem I/II) accounts for 4%, and RD (from no stems open confirmation) accounts for 2% (FIG. 9A and FIGS. 10A-10B).

The nested stem loop structures would predict a highly dynamic process of stem formation, involving box sequences that compete for pairing to yield diverse alternative splicing outcomes. As a result, miRNAs binding to distinct boxes would influence the dynamic pairing of the boxes and hence alter the pattern of splicing. Specifically, the study identified several microRNAs that bind to the box regions in RpL3 and disrupt stem formation of its pre- mRNA (FIG. 9A). Based on energetically favorable and structurally feasible criteria, miR- 210-3p is predicted to interact with both box I and II, while the 5' and 3' arms of miR-988 are predicted to interact with box I and II, respectively. Additionally, the study found that miR- 992-3p may have a strong interaction with box II (FIG. 11). To experimentally assess the effect of predicted miRNAs on pre-mRNA splicing, the study employed a two-color multiplex realtime PCR assay with specific TaqMan probes to quantitatively assess the ratio of different RpL3 mR A variants after microR A overexpression. Specifically, the study designed probes to quantify total mRNA variants, RA/H and RG, and RG only (FIG. 9A). These results demonstrated that miR-210 and miR-988 expression significantly reduced the ratio of both RA/H and RG in all four mRNA variants compared to the negative control, miR-304, which had no effect as a non-targeting control (FIG. 9B). Furthermore, only miR-210, instead of miR- 988, reduced the portion of RG (FIG. 9C). However, miR-992 expression increased the portion of RG among all spliced mRNA variants (FIG. 9C). Notably, miR-210 targeting both boxes I and II disrupted both stems through MIMOSAS, leading to the open confirmation and splicing of RD (FIGS. 9B-9D). Conversely, both miR-988 and miR-992 significantly increased the preference for splicing the RG variant (FIG. 9D), showing that they disrupt the stem between boxes II/III and/or enhance the stem between I/II. While real-time PCR measurements of the endogenous variant expression validate the effects of microRNAs on splicing and experimentally confirmed predictions from the computation pipeline, this approach is limited by the availability of variant-specific probes, with the expression level of RD inferred from the difference of total transcript and other variants due to the lack of RD-specific probes.

To directly observe the impact of MIMOSAS on RpL3 pre-mRNA splicing, the study designed a split-fluorescent proteins (FPs) two color reporter construct that mirrored the complex secondary structures of the pre-mRNA. The study retained all the essential sequences involved in stem formation within the pre-mRNA and employed self-complementing split FP (Feng et al., 2017) to monitor spliced mRNA variants. Utilizing two distinct FPs, sfGFP and sfCherry, alongside nuclear (NLS) and cytoplasmic (NES) localization signals, the study differentiated each spliced mRNA variant based on its color and cellular location (FIG. 12A). Specifically, stem (II/III) favoring the RA/H variant was represented by the expression of cytoplasmic-localized (NES) super fold GFP (sfGFP). In particular this variant included two major parts of split GFP1-10 and GFP11 subunits, a cytoplasmic localization sequence (NES), and a 96bp linker sequence. Such a construct allows two split GFP subunits to form a functional GFP via self-complementation, enabling detection of green signals in the cytosol. Stem (I/TI) driving the splicing of RG variant was indicated by green-fluorescent signals in the nuclei. Conversely, when no stems were formed, the RD variant was transcribed, producing a chimera construct of sfGFPl-10 subunit linked by the sfCherry 11 subunit. As a consequence, a red fluorescent signal was created when combined with the pre-miRNA linked sfCherry 1-10 subunit, indicating a functional sfCherry formed by the complementation of two split subunits. The study co-transfected the reporter and pre-microRNA constructs into mammalian COS7 cells for visualization to assess the MIMOSAS effect of each candidate microRNA. Cells showing red fluorescent signals, indicating successful co-transfection of both constructs, were selected, and the fluorescence intensity of each FP was measured using confocal microscopy (FIG. 12B). Expression of miR-210, miR-988, or miR-992 significantly reduced the cytoplasmic GFP intensity (RA/H mRNA variant) compared to the non-target control miR-9c or negative control miR-304 (FIG. 12C). Expression of miR-210 decreased the intensity of nuclear GFP signals (RG mRNA variant) (FIG. 12D), consistent with real-time PCR results of endogenous RG (FIG. 12C). Interestingly, miR-992 expression increased RG (nuclear GFP intensity, FIG. 12D), while red fluorescent signal intensities significantly increased with both miR-210 and miR-988 overexpression (FIG. 12E), confirming the predictions. Collectively, the split-FP dual color splicing reporter allowed the in vivo visualization of MIMOSAS events and further validated the predicted effect of miRs-210, 992 and 988 in impeding the binding of box II and III and reducing the abundance of splice variants RA/H while increasing the splicing of variant RD. Notably, screening of sequence windows upstream of box I and II showed that miR-992 does not hinder stem formation but instead stabilizes box binding between box I and II, which is confirmed with the increased abundance of splice variant RG.

Nuclear enrichment of microRNAs and MIMOSAS events are dependent on Agol : During RISC formation, mature Drosophila microRNAs are typically loaded onto Argonaute 1 before binding to target transcripts (Forstemann et al., 2007; Pushpavalli et al., 2012). Drosophila Argonaute 1, the homolog of mammalian Argonaute 2, is the major RISC component onto which microRNAs are loaded (Y ang et al., 2014), prompting the determination of the role of Argonaute 1 in MIMOSAS. In particular, the study obtained Argonaute 1 heterozygous (AGO1^+/ ) mutant flies which had previously been shown to have significantly reduced Argonaute 1 levels (Pushpavalli et al., 2014). The study expressed miR-210, the candidate microRNA targeting both Nmnat and RpL3 in Drosophila salivary gland cells using OK371-GAL4 in either AGO1 wild-type (AGO1^+/+) or heterozygous mutant (AGO1^+/ ) flies, and carried out a fluorescent in situ hybridization (FISH) assay to visualize and quantitatively measure the subcellular localization of miR-210-3p. Intriguingly, miR-210-3p was highly enriched in the nuclei in the wildtype tissue, but significantly reduced in the loss of Agol (AGO1^+/ ) mutant tissue (FIGS. 13A-13B and FIG. 14). Notably, loss of AGO1 only affected the nuclear enrichment of miR-210 (FIG. 13B) but did not alter the total microRNA expression levels (FIG. 13C).

To further analyze the functional consequence of AGO1 loss of function, the study carried out real-time PCR analysis of Nmnat and RpL3 splice variants in miR-9c (control) or miR-210 expressing brains. Under normal condition, miR-210 can disrupt the stem formation in Nmnat (FIGS. 5A-5F) and in RpL3 (FIGS. 9A-9D). However, in AGO1 heterozygous mutant (AGO1^+/ ) brains, the MIMOSAS effects of miR-210 was significantly diminished in both Nmnat (FIG. 13D) and RpL3 (FIG. 13E), while miR-9c expression did not affect alternative splicing outcomes of endogenous Nmnat and RpL3 in either genetic background (FIGS. 13D-13E). Taken together, these results showed that microRNAs localize to the nucleus with the presence of AGO1, while MIMOSAS events are likely to occur in the nucleus and rely on the presence of AGO1.

Discussion

Although RNA secondary structures have been shown to regulate alternative splicing of long-range pre-mRNA, the driving factors that modulate RNA structure and interfere with the recognition of the splice sites are largely unknown. This study developed a computational pipeline to scan the genome and identified broad presence of splice-relevant microRNA-pre- mRNA binding sites based on sequence complementarity and energetics of RNA secondary structures. These in vivo experiments indicated the functional consequence of microRNA-pre- mRNA pairing where microRNAs can either disrupt or stabilize stem-loop RNA secondary structures to influence splicing outcomes. This study shows that MicroRNA-Mediated Obstruction of Stem-loop Alternative Splicing (MIMOSAS) is a regulatory mechanism for the transcriptome-wide regulation of alternative splicing, expanding the repertoire of microRNA function and further indicating cellular complexity of post-transcriptional regulation. The process of alternative splicing requires the recognition of certain components within a single-stranded pre-mRNA, such as splice sites, branch sites, and cis-acting elements, by trans-acting factors. The formation of RNA secondary structures can inhibit/activate the assembly of spliceosomes through splice site suppression, occlusion/exposure of cis-acting elements, “looping-out” mechanism, as well as the competition between RNA secondary structures (Jin et al., 2011). Notably, non-coding RNAs which are a type of regulatory noncoding RNA have previously been implicated in posttranscriptional regulation by influencing pre-mRNA splicing through altering chromatin, hybridizing with genomic loci or pre-mRNA molecules to create an RNA-DNA or RNA-RNA duplex, or regulating splicing factors (Pisignano and Ladomery, 2021). In turn, microRNAs, another type of regulatory non-coding RNA, are typically associated with mRNA degradation or translational inhibition, impacting alternative splicing indirectly by post-transcriptionally regulating splicing factors (Liu et al., 2021). This study here provides the evidence for a direct role of microRNAs in splicing regulation.

Given that the majority of splicing events occur co-transcriptionally (Herzel et al., 2017), it is plausible that microRNAs regulate splicing in the nucleus. Recent advances in imaging techniques have shown the translocation of endogenous mature microRNA from the cytoplasm to the nucleus, visualized by superquencher molecular probes through optoporation to selectively permeabilize single cells (Foldes-Papp et al., 2009). Several studies have showed that karyopherin XPO1 (Exportin- 1) or IOP8 (Importin- 8) facilitate the nuclear import of microRNAs-AGO complexes in human cells (Castanotto et al., 2009; Till et al., 2007; Weinmann et al., 2009). Although the detailed function of nuclear microRNAs remains incompletely understood, current evidence shows that nuclear RISC complexes can modulate chromatin compaction to achieve post-transcriptional gene silencing, as well as regulate the rate of RNA polymerase II procession to influence alternative splicing outcomes (Allo et al., 2009; Komblihtt, 2006). Notably, the modulation of the nuclear levels of AGO1, AGO2 or DICER 1 was shown to impact splicing decisions at certain alternatively spliced exons (Allo et al., 2009; Ameyar-Zazoua et al., 2012; Liu et al., 2012; Park et al., 2019).

This study identified candidate microRNAs that directly regulate alternative splicing by interfering with pre-mRNA stem-loop structures. After generating predictions through a computational pipeline, the study experimentally validated splicing predictions of three distinct long-range pre-mRNAs in both mammalian cell culture models and in vivo in Drosophila. These observations reveal that microRNAs can either disrupt or stabilize stem-loop structures, thereby influencing splicing outcomes, pointing to a regulatory mechanism for the transcriptome-wide control of alternative splicing, termed MicroRNA-Mediated Obstruction of Stem-loop Alternative Splicing (MIMOSAS). While microRNAs have traditionally been studied in the context of binding to the mature mRNA 3’UTR and downregulating target transcripts, MIMOSAS shows their binding to interior intronic or exonic regions in the pre- mRNA, thereby regulating mRNA splicing. The microRNA-pre-mRNA duplex binding modulates mRNA expression efficiently and in a splice variant- specific manner. Given the dynamic expression patterns and cell type/tissue-specificity of microRNAs, MIMOSAS presents a new regulatory layer through which cells in different tissues can specifically control the expression of transcript variants. Furthermore, MIMOSAS highlights the regulation of gene expression by noncoding RNAs in stress or disease-specific contexts, emphasizing the importance of considering alternative splicing modulation as a therapeutic means. As 60% of disease-associated point mutations have been linked to splicing defects (lensen et al., 2009), various methods have been developed to modify alternative splicing. For instance, duplex RNAs have been utilized to modulate alternative splicing in mammalian cells (Liu et al., 2012), and a method of employing single-stranded siRNAs has been developed to modify the splicing of a Dystrophin RNA, associated with Duchene muscular dystrophy (Liu et al., 2015). Given the relevance of RNA as splicing modulators demonstrated in these reports, this work shows the microRNAs as tools for modulating alternative splicing as a promising therapeutic intervention.

Example 2: MIMOSAS for the Degenerating Brain: Global mRNA Splicing Regulation by microRNAs.

One of the exciting unifying contributing factors of AD and other neurodegenerative diseases that recently emerged is the role of altered RNA metabolism in disease pathogenesis [3, 4], Specifically, there has been increasing recognition that neurological and neuromuscular diseases are caused by splicing errors [5, 6], mutations in RNA binding proteins (reviewed in [7, 8]), as well as dysregulated microRNAs (reviewed in [9, 10]) and long-noncoding RNAs (reviewed in [11-13]). Despite significant recent insights into the role of dysregulated splice variants [14], microRNAs [9] and to a lesser degree IncRNAs [12] in ADRD, the current understanding of the disease mechanisms is largely based on isolated, scattered, and limited characterizations of individual genes. Furthermore, the role of RNA metabolism in neuroprotection and the mechanisms of regulating disease-relevant splice-forms are largely unknown, leading to uncharted territory in ADRD. Recently, a mode of transcriptional regulation was discovered, where binding of microRNA miR-1002 thwarts the formation of secondary structures in the un-spliced mRNA at splice-junctions regulating and controlling the abundance of a neuroprotective isoform of the gene NMNAT [1]. This mode of transcriptional regulation is termed MIMOSAS: MIcroRNA-Mediated Obstruction of Stem-loop Alternative Splicing. Additional studies further identified additional neuronal genes with MIMOSAS mode of transcriptional regulation (see below, FIGS. 16A-16C). To examine the genome-wide impact of MIMOSAS, a computational platform is built to scan and predict binding sites of microRNAs to interior intronic or exonic regions in the pre-mRNA of targets on a genome-wide scale. Given that microRNA-based transcriptional regulation is an efficient mechanism to modulate expression of gene isoforms, it is contemplated that the identification of microRNAs in particular and ncRNAs in general that control the abundance of disease-specific and neuroprotective splice forms will fundamentally benefit the understanding of disease pathogenesis and the design of neuroprotective strategies.

The large-scale availability of RNAseq and genomic data sets from blood and brain tissue of ADRD patients and corresponding non-disease controls is the perfect starting point to investigate the role of ncRNAs such as microRNA in the dys-regulation of mRNA splice variants in ADRD patients, allowing ncRNA mediated dys-regulation of mRNA splice-forms (MIMOSAS) to be mechanisms of pathogenesis and therapeutic intervention on the map of ADRD. As depicted in FIGS. 15A-15C, the overview of the present disclosure provides: (1) Development of a computational pipeline to predict MIMOSAS-mediated alternative splicing regulation; (2) Aggregation and deep analysis of existing ADRD RNAseq and genomic profiles to extract global ncRNA and splice variant expression profiles to quantitatively corroborate predictions; (3) Determination of splice-relevant quantitative trait loci to validate and stratify ncRNA-splice mRNA binding candidates; and (4) Experimentally obtained proof-of-principle through biological/experimental confirmation of MIMOSA predictions in ADRD. The outcomes are, (i) the MIMOSAS platform, a computational platform that predicts MIMOSAS- mediated mode of regulation, (ii) a genome-wide experimentally testable dataset of ncRNA- splice-form pairs that drive the underlying phenotype, (iii) specific RNA metabolic mechanisms driving the abundance of Tau pathological splice species and NMNAT neuroprotective variants in ADRD, and (iv) identification of ncRNA biomarkers for disease progression and genetic modifiers for neuroprotective outcome.

Long range stem-loop mediated alternative splicing and MIMOSAS (MIcroRNA- Mediated Obstruction of Stem-loop Alternative Splicing): Alternative splicing plays important roles in cellular differentiation and stress response in both physiological and pathological conditions and has been a subject for disease association and for therapeutics. While 88% of human genes are alternatively spliced [15, 16], 22% of disease-causing mutations are splicing sensitive [17], and 25 % of disease-causing exonic mutations induce exon skipping [18]. Recently emerging evidence supported active roles of noncoding RNAs in alternative splicing. For example, it has been shown that a IncRNA can promote the retention of a particular exon in FGFR2 alternative splicing [21], and double stranded RNAs can affect the splicing of genes associated with muscular dystrophy, while a small nucleolar RNA has been found to regulate the splicing of Serotonin Receptor IIC, which is associated with Prader- Willi Syndrome [22]. Given that approximately 95% of multi-exonic human genes are alternatively spliced [23], and an estimated -40% of microRNA target sites were found in mRNA CDS regions in human cells [19], microRNA target sites in the exonic and intronic regions point to ample and opportunities for microRNAs to target and influence pre-mRNA splicing.

Splicing of long introns is likely facilitated by RNA secondary structure formation in the intron of the pre-mRNA, bringing 5’ and 3’ splice sites into relatively close proximity and promoting excision of a particular intron. It has been shown that many of such long-range interactions are evolutionary conserved, and are strongly associated with alternative splicing [24-26]. A key molecular mechanism facilitating long-range splicing is the stem-loop RNA structure formation between complementing sequences, termed boxes. The Box sequences are located near splice sites in introns or exons where the formation of the stem brings distant splice sites together to facilitate splicing (FIG. 16A). MIMOSAS presents a new regulatory layer through which cells can specifically regulate splice variant expression through microRNAs. While mostly studied in the context of mature mRNA 3’UTR binding and target transcript downregulation, MIMOSAS predicts that the binding of microRNAs can disrupt such splicerelevant RNA secondary structures of targets, pointing to a mechanism of splicing and RNA secondary structure regulation [26, 27]. Notably, this splice-relevant secondary structure was identified purely on the basis of sequence composition and computational folding algorithms [28] of the NMNAT pre-mRNA. In particular, splice junctions were scanned for complementary sequences (i.e. boxes) and corroborated the emergence of a stem-based loop secondary structure that captured intermediate introns, using folding algorithms that are based on energy optimization. Subsequently, binding events of microRNAs were considered, that interfere with stem loop formation, if the duplex of binding microRNA and mRNA is (i) located at the stem of a hairpin loop and is (ii) energetically more favorable than the formation of the stem loop itself.

Studies have been carried out on NMNAT, Fas and RPL3 with MIMOSAS-mediated splicing events (FIGS. 16A-16C). In all these examples, the un-spliced pre-mRNA contains two complimentary sub-sequences, termed Boxl and Box2, that form a stem loop, leading to the excision of an exon between the two boxes (FIG. 16A). As for NMNAT, microRNAs (e.g. miR-1002, FIG. 16B) were discovered to bind Box 1 and 2 through its seed sequences, disrupting the formation of the stem loop in the pre-mRNA [ 1] (FIG. 16B). As a result, exon 5 is kept and the neuroprotective splice isoform RB of NMNAT promoted, showing that miR- 1002 is a switch to enhance neuroprotection under stress [39] (FIG. 17). In addition to the originally found miR-1002 [1], miR-210, miR-278 and miR-137 were predicted to have such disruptive power as well. Indeed, overexpression of these microRNAs led to a significant shift in splicing to RB (FIG. 16C), showing that such microRNAs function to increase intrinsic neuroprotective capability (see below, FIG. 17). This approach was applied to genes Rpl3 and Fas3, allowing for corroboration of the experiments.

MIMOSAS mediated regulation of intrinsic neuroprotective stress response: NMNAT proteins are among the most robust neuroprotective factors ([36-38]). A previous study showed that the DmNMNAT RB variant is upregulated to produce the neuroprotective protein isoform (PD), showing that alternative splicing through microRNA miR-1002 is a switch to enhance neuroprotection under stress [39] (FIG. 17). This exciting discovery outlines a new process for microRNAs in regulating alternative splicing and modulating stress resistance that influences the disease etiology of ADRD. The role of splicing in neuroprotection has been implicated as several genes in addition to NMNAT have been demonstrated to protect against neurodegenerative diseases in a robust, splicing-dependent manner [93]. For example, the expression of a CD33 (Cluster of Differentiation 33) splice variant with a skipped exon is associated with reduced risk of Late-Onset Alzheimer’s disease. The soluble isoform of RAGE (Receptor for Advanced Glycation End-products) is produced via an alternative 3’ splice site and results in decreased Amyloid beta (A|3) uptake in Alzheimer’s. MGF (Mechano-Growth Factor), a variant of Insulin-like growth factor results from a frame shift coupled to the inclusion of an exon and has demonstrated protection in many models of neurodegenerative diseases, such as Amyotrophic lateral sclerosis (ALS), ischemia, and MPP+ (l-methyl-4-phenylpyridinium) treatment. With already existing reports of developing therapeutics to influence splice variants [11, 12], these genes present excellent opportunities to combat ADRD. Given the ample examples of alternative splicing in regulating either the process of neurodegeneration or the neuroprotective capability, MIMOSAS mediated transcriptional regulation of ADRD relevant genes presents promising strategies for new mechanisms of disease pathogenesis and new directions of neuroprotective therapies. This work integrates ADRD RNAseq, noncoding RNA transcription and genomic profiles to extract global noncoding RNA and splice variant expression profiles, determine splice-relevant quantitative trait loci to validate and stratify noncoding RNA-splice mRNA binding candidates, and discover ncRNA biomarkers and neuroprotective therapeutic interventions.

MicroRNA target sites and Argonaut footprints: It has been reported that about 42.6% of microRNA target sites in human HEK293 cells were found in mRNA CDS regions [19], and a majority of microRNA target sites are located within the CDS of mRNAs instead of the 3’UTR [20]. The probability of microRNA targeting the exonic and intronic regions and influencing pre-mRNA secondary structure and splicing is similar to if not higher than the canonical targeting of microRNAs to 3’UTR region. The canonical function of microRNA targeting 3’UTR is through Argonaut protein in the cytoplasm. While findings show that MIMOSAS through miR-1002 is mediated by Argonaut 1, the cellular location where miR- 1002 regulation of NMNAT occurs (i.e. in the nucleus, cytoplasm, or both) is important for understanding the mechanism of noncoding RNA-mediated splicing regulation. Since most splicing events occur co-transcriptionally [29], the nucleus is the most likely location. Specifically, RNALocate lists over 200 noncoding RNAs that were detected in the nucleus of neuronal or neuron-like cells [30]. In addition, Argonaut proteins have been detected in the nucleus, binding chromatin [31-35]. The results indicate that Argonautel indeed appears in the nucleus as well as the cytoplasm, further supporting the notion that noncoding RNA regulation of splicing may occur in both the nuclear and cytoplasmic cellular compartments. As a consequence of such observations, it is contemplated that the critical role that Argonaut plays for mediating MIMOSAS provides an opportunity for scanning Argonaut-footprints to identify global microRNA binding sites in the realm of ADRD relevant expression regulation.

Herein, the concept is to put a splicing mechanism that governs the abundance of disease specific splice forms of mRNAs through ncRNAs on the ADRD map. Although different splice forms of genes have been recently observed as playing a fundamental role in ADRD, the regulatory mechanisms that drive the abundance of disease specific splice forms are unknown. Based on existing algorithms that reliably detect mRNA local and long-range secondary structures that govern splice-forms of ADRD relevant genes, microRNA as well as long ncRNA binding sites are primarily identified that interfere with mRNA sequences with splice- specific secondary structures. Based on a large-scale data collection and reanalysis effort, it has been contemplated that this approach will indicate (i) numerous instances of regulation mechanisms that show switches between the function-specific splice isoforms of ADRD relevant genes, where (ii) ncRNAs modulate variant-specific expression of genes and regulate neuronal self-protective response in neurodegeneration. These studies identify the ncRNAs that govern splice forms, modulating neuroprotective response, uncovering the endogenous regulatory mechanisms of pathogenesis as well as neuroprotection in ADRD.

The methodological innovation stems from a combined computational and experimental data science approach. Bioinformatics and data science methods are employed to determine secondary structures of pre-mRNA that are splice-relevant as well as binding sites of ncRNA and full-length pre-mRNAs (including introns) to predict microRNAs that modulate splicing, a mode of action for microRNAs. Based on such predictions, ncRNA targeting and characterization of the mode of ncRNA action in vivo are experimentally confirmed. This powerful combination of data science with in vivo characterization led to discoveries of transcriptional regulatory mechanisms relevant to neuroprotection. The innovation concerning Drosophila models lies in established behavior paradigms 140, 411, quantitative morphological analysis methods to characterize phenotypes in the adult brain [42], and a series of biochemical analyses to enable dissection of causal relationships among genetic and molecular changes, protein homeostasis, and neuropathy in vivo [43, 44]. For this application, an in vivo alternative splicing reporter system is developed adapting a ‘split Fluorescent Protein’ design where the N-terminal shared sequence fragments of DsRed and AcGFP is inserted in commonly spliced exon, while the distinct sequences of DsRed (red) and AcGFP (green) are inserted into alternatively spliced exons (FIG. 18). The ratio metric measurement of DsRed and AcGFP fluorescence allows the in vivo live reporting of the alternative splicing events.

Development of a computational pipeline for the prediction of MIMOSAS- mediated splice regulation and ncRNA/microRNA biomarker discovery: Although significantly altered splice forms of genes are increasingly recognized to play a fundamental role in ADRD, underlying mechanisms that govern their over/under-representation are widely unknown. Based on previous results indicating that microRNAs can interfere with secondary structures in un-spliced mRNA, and drive the emergence of certain gene splice forms, a computational pipeline is designed that detects (a) splice-relevant secondary structures, and (b) ncRNAs binding sites that disrupt such secondary structures in differentially spliced mRNAs in ADRD (FIG. 15A). While such predictions are solely based on RNA sequence data, the candidate pool is further augmented by machine learning (ML)-based integration of experimental CLASH and CLIP data, which reveals binding sites of microRNAs and the underlying mRNA through the location of Argonaut 1. Finally, candidate interactions between ncRNAs and splice relevant sites on pre-mRNAs are evaluated by assessing their energetics, assuming that successful binding and disruption are energetically more advantageous than keeping a splice-relevant secondary structure intact.

Developing a computational pipeline that reliably detects mRNA local and long-range secondary structures that govern splice-forms of ADRD relevant genes: While the role of spliceosomal components in splice site selection has been the focus of intense study, the role of RNA structure has received little attention. Splicing of long introns is likely facilitated by RNA secondary structure formation in the intron that brings 5’ and 3’ splice sites into relatively close proximity, promoting excision of a particular intron. Previous work [24-26] showed that many such long-range interactions exist that are evolutionary conserved and are strongly associated with alternative splicing. Likewise, local RNA structures that occlude binding sites of small-nuclear RNAs or proteins can hinder splice site selection. A well-studied example of such a mechanism is the Tau gene in neural degenerative diseases [45, 46]. There, a small hairpin structure directly downstream of exon 10 competes with U 1 snRNP binding to the 5’ splice site, promoting skipping of exon 10. Mutations that destabilize this secondary structure lead to increased usage of the exon 10 splice site, sufficiently disturbing the balance between 3R-tau and 4R-tau isoforms and causing neurodegeneration and dementia. Herein, it is contemplated that MIMOSAS regulation of splicing can occur not only when a binding ncRNA perturbs long-range structures (as shown for NMNAT, Fas3 and Rpl3), but also when it affects local splice-relevant structures, such as the exon 10 hairpin of Tau. Therefore, a set of candidates for both long-range and local splice regulatory structures are compiled, that are bona-fide targets for MIMOSAS regulation upon ncRNA binding.

To detect local splice regulatory structures, small-nucleolar RNA (snRNA)-mRNA interactions are the focus. The first steps of intron recognition involve the formation of RNA- RNA interactions between U1 snRNA at the 5’ splice site and U2 snRNA with the branch point sequence of the mRNA. The corresponding binding free energies can be computed directly by in-house RNAup software [47] as part of the ViennaRNA package. Such energetical considerations not only depend on sequence complementarity, but also competing local RNA structure. In particular, this approach is based on determining the expended energies to locally unfold the binding site on the mRNA and the interaction energy between the mRNA and snRNA. As for the Tau gene example, this approach calculates the energy needed to dissolve the competing hairpin structure, allowing the snRNA to bind the underlying Tau mRNA. Contemplation that such an activity is supported by ncRNAs that bind in the vicinity of the hairpin structure, calculations are performed for all splice sites. Introns that are constitutively spliced are presumed to be free of splice inhibiting structures and serve as a baseline that allows for assignment of a z-score or P-value to each splice site through permuting the underlying RNA sequences.

A catalog of long-range structures in introns that are splice-relevant was recently published [24], which is based on genome-wide scans of complimentary sequences, forming stems of RNA hairpin loops capturing whole exons and introns. While solely based on sequence complementarity between short fragments this approach completely discounts energetics of RNA structure formation and the competition between local and long-range structures, \ missing numerous splice-relevant structures. Therefore, this catalog complements the present predictions, accounting for energetic parameters of the underlying RNA folding. However, prediction of the global, energetically most stable structure of a complete RNA intron is only biologically meaningful for short introns. Since the primary interest is towards the interactions that bring the ends of the introns in close proximity through stem formation, flanking regions around the 5 ’ and 3 ’ end of the intron are extracted and long-range structures are predicted as RNA-RNA interaction between the 5 ’ and 3 ’ fragments through the established RNAup and/or RNAplex software [48]. Simplifying the energy function to achieve predictions for two RNAs in quadratic time, RNAplex allows for ranking predicted interacting RNA segments through their binding energy as well as length and distance to the splice sites. Furthermore, the search for long-range structures is complemented by directly computing the expected distance between donor and acceptor sites in the secondary structure graph of the underlying pre-mRNA, where secondary structures can significantly reduce the donor- acceptor distance (FIG. 19). Specifically, smaller donor- acceptor distances correlate with preferentially spliced introns. While a dynamic programming algorithm to compute such end-to-end distances was proposed in [49], the approach is limited to distances from 5’ to 3’ end of the underlying pre-mRNA. Instead, the present disclosure approximates the distance through pair probabilities that are obtained through the application of the RNAfold routine in the Vienna RNA package [50]. Such an approach also provides the opportunity to account for binding of microRNAs or other ncRNAs, allowing for computation of the change in donor-acceptor distance upon binding.

Previous work has strongly indicated that splice regulatory structures are likely to be evolutionarily conserved [25, 26]. To further corroborate the relevance of predicted splicerelevant local and long-range structures, two approaches to quantify their evolutionary conservation are pursued: (i) In an alignment-based approach that starts from genome-wide alignments between mammalian genomes as provided by the UCSC genome browser and measure conservation at the base-pair level. In this case structure predictions are performed directly on the alignment using RNAalifold [51, 52], also allowing the analysis of compensatory mutations. Statistical significance of conserved individual base pairs in local and long-range structures are assessed using R-scape [53]. (ii) In turn, different species form long- range interactions where the details of the underlying structures differ. Such cases occurred in studies of bacteria] small RNA regulation, where the same sRNA-mRNA pair is observed in many species, although binding sites are slightly different [54]. While alignment-based approaches would fail in this case, it is still possible to include conservation on a structural level by performing predictions in each species separately. To assess the level of evolutionary conservation of local and long-range splice relevant structures, P-values are assigned to the individual structural element in in each species, determining an overall P-value using Hartung’s method [55].

Identifying microRNA binding sites that interfere with mRNA sequences that influence splice-specific secondary structures: To determine bona-fide cases of MIMOSAS regulation, microRNA binding sites are found that perturb local or long-range splicing regulatory structures (previously identified above). In a first step, microRNA target predictions are performed and binding sites in the vicinity of splice sites are identified. Binding sites that overlap with structure elements (previously identified above) are prime candidates for MIMOSAS regulation, but even binding sites that don’t overlap may affect structures in their vicinity. The ViennaRNA package provides a sophisticated constraints framework [50] used to predict the effect of microRNA binding on RNA structure.

To find splice-relevant microRNAs, complementary binding sites of the microRNA seed sequences are first scanned, allowing for (im)perfect seed matching with at most one mismatch and one ‘wobble’ nucleotide. In particular, perfect matching between nucleotide 2 to 7 of the microRNA specific seed sequence without any GU wobbles are considered, as well as perfect matching of nucleotide 2 to 8 of the microRNA specific seed sequence without one GU wobble and imperfect matching of between nucleotide 2 to 9, 1 mismatch and one GU wobble.

As the current method, herein, to find microRNA binding site is relative strict and does not account for experimentally observed interactions, currently available machine-learning approaches are adapted to predict microRNA binding sites to extend the set of candidate ncRNAs. In particular, large-scale human binding screens are considered using CLIP [62] and CLASH [19, 63] approaches that are the closest experimentally observed proxy to microRNA- mRNA binding events. Briefly, both methods map microRNA-RNA interactions through cross-linking RNA-protein complexes that are affinity purified. Ligating and reversetranscribing microRNA-mRNA hybrids into cDNA followed by deep sequencing provides the identity of the microRNA and the putative 40 nucleotide long binding area of the Argonaut 1 (AGO)-microRNA complex on the underlying mRNA. The model herein assumes that the binding of the microRNA/AGO complex on the un-spliced mRNA hampers the formation of splice-relevant secondary structures, disrupting the underlying splice-relevant secondary structures, characteristics of binding events that are not be captured with the strict matching approach.

While not many microRNA-binding prediction algorithms that are trained on CLIP/CLASH data exist [65-67], a popular deep-learning approach is used that already has been used for the prediction of microRNA-mRNA binding events [66], capturing the (a) sequence of the microRNA, and (b) the binding site on the mRNA where the microRNA putatively binds through the AGO protein (FIG. 20A). Currently, sequences of a microRNA and the binding site of the AGO protein on the mRNA are represented as one-hot encoding matrices in the encoding module of the architecture. In the next step of the architecture, convolutional and pooling layers are used to extract higher order features from the underlying input data. Finally, the architecture provides layers of bi-directional LSTMs [66], that allow to learn sequential features of the underlying local features that were determined in the previous layers. Finally, a soft-max layer allow for predictions if a given microRNA and mRNA sequences indeed interact. As a consequence of the trained model, microRNA/mRNA candidate pairs in the vicinity of splice relevant secondary structures are scanned to augment the candidate pool obtained with simple sequence matching.

While there exist many different method- specific ways to improve, classification accuracy heavily depends on the actual training data. As another project- specific consideration, contemporary microRNA binding prediction algorithms aim at the prediction of microRNA binding sites in the 3’ and 5’ un- transcribed regions of mRNAs. Notably, a recent in-depth analysis of 79 human large-scale CLASH/CLIP screens [64] reported roughly half of 50,000 microRNA-mRNA interactions that were completely annotated with the specific microRNA and microRNA binding site sequence were distributed across the coding regions and introns of human genes, pointing to a bona-fide training data set for our purposes. Furthermore, the choice of negative training data is fundamental for the ability to predicted splice-relevant microRNA-mRNA binding events, avoiding an avalanche of false-positive predictions. Previous approaches considered synthetic mock data derived from positive training data. However, such pairs lack biological meaning and provide no guarantee that randomly created pairs may show some functional binding. To avoid such problems, the observation that such screens provide target binding regions on the mRNA with an average length of 40 nucleotides is utilized. Screen data based on CLIP/CLASH methodology show that microRNAs usually align with the central nucleotides of the corresponding target site on the mRNA. In turn, binding of the microRNA and mRNA rarely happens near the boundaries of the target binding sites. Utilizing the set of CLIP/CLASH data that focuses on coding regions and introns, negative training data is generated through screening the target binding sites and binding energies of the duplex between microRNA and mRNA (DGdupiex) are determined. Interactions between microRNAs and corresponding binding sequences in the target sequence with the best energy is the most probable and vice versa. Assuming that the positive training set consists of microRNA sequences and subsequences of the corresponding target site in a mRNA that at least partially overlaps with a splice-relevant secondary structure, a set of negative training data of equal size is generated by selecting the subsequence of the observed target site that has the least favorable energetic characteristic, reliably indicating that a binding event at the corresponding mRNA location is improbable.

To classify if the matching of the underlying ncRNA obtained either through matching or machine learning indeed is a bona fide candidate to disrupt a splice relevant mRNA secondary structure, the energy of the complete ncRNA binding to the underlying mRNA (i.e. the duplex, DG) is calculated using the Vienna RNA software package [47] (FIG. 20B). To find relevant ncRNA candidates, the energy gained by the ncRNA binding to the mRNA needs to offset the energy lost by opening the splice-relevant secondary structures. Specifically, the difference of the energy of the mRNA structure with intact and perturbed structures is calculated as a function of the nucleotides that cannot participate in the stem loop as a result of ncRNA binding to the corresponding nucleotides, DGopen, using software of the Vienna RNA package. In turn, the energy that is gained by binding of the ncRNA to the underlying mRNA DGdupiex is calculated. As a consequence, DGdupiex > DGopen indicates that the binding of the ncRNA is energetically more favorable than keeping the splice-relevant structure intact, pointing to a splice relevant microRNA. By computing these free energies in the presence and absence of microRNA binding, the change in stability or equivalently the probability that the splice regulatory structure is formed or destroyed is calculated. The level of evolutionary conservation of pairs of mRNA structure and microRNA binding sites is determined to identify high confidence MIMOSAS candidates. To assign statistical significance to the predicted long- range interactions and donor- acceptor distances, a background model is employed through shuffling sequences and/or computing interactions between intron fragments that are not derived from the same gene as they are expected to form no interactions.

While the primary focus is on microRNAs, this approach can be applied to any ncRNA. In particular, cA-functioning IncRNAs that were knocked down modulated the nearby proteincoding gene’s transcriptional burst frequency, showing that such IncRNAs may have splice relevant functions as well [68]. Furthermore, in ADRD the ability of long ncRNAs to regulate associated protein -coding genes may contribute to disease if misexpression of a IncRNA deregulates a protein coding gene with clinical significance. For example, an antisense IncRNA that regulates the expression of the antisense BACE1 gene, a crucial enzyme in AD etiology, exhibits elevated expression in several regions of the brain in individuals with Alzheimer's disease [69]. Another example is IncRNA 51 A, which regulates APP processing and A[3 production and promotes A[> secretion [70]. As a consequence, this methodology can be extended to account for IncRNAs, matching their sequences to splice-relevant mRNAs that carry splice-relevant secondary structures.

Establish an ADRD ncRNA-splice mRNA cross-correlation dataset through extraction and aggregation of existing ADRD transcriptional RNAseq datasets: The present disclosure provides candidate pairs of ncRNAs and pre-mRNAs that are splicerelevant, such predictions were independent of ADRD. Herein, ADRD relevance is established (FIG. 15B). A deep reanalysis of RNAseq data will be performed to obtain a global ADRD splice isoform expression dataset. Furthermore, the dataset is harmonized with microRNA and ncRNA expression data of matching ADRD (non-)disease samples. Such a step allows for drilling down on ADRD-specific predictions of splice-relevant microRNA (ncRNA)-mRNA pairs that can be subjected to experimental confirmation. It is contemplated that pre-mRNAs with splice relevant RNA secondary structures are enriched with ADRD specific splicerelevant genomic variations. In particular, quantitative trait loci in ADRD as well as genomic variations already associated to ADRD are determined, that influence the MIMOSAS mechanism (FIG. 15B).

Construction of a global ADRD splice isoform expression dataset through extracting splice isoform expression abundance from large ADRD RNAseq and harmonizing ADRD microRNA expression dataset: Herein, the focus is on the collection and re-analysis of genomewide RNAseq data of ADRD patients and control cohorts. In particular, the present disclosure focuses on RNAseq data as of the Religious Orders Study (ROS) and Memory and Aging Project (MAP) that are available through the synapse portal [14, 71], In particular, ROS is a longitudinal clinical-pathologic cohort study of aging and Alzheimer's disease (AD) patients from religious communities. As a complement to the ROS study, MAP is a longitudinal study of a cohort of older individuals without any signs of dementia with common chronic conditions of aging [72]. Patients were categorized as having no cognitive impairment (NCI) if diagnosed without dementia or mild cognitive impairment (MCI), while diagnoses of dementia and AD conform to standard definitions [72]. RNA of ADRD patients and controls were sequenced from the dorsolateral prefrontal cortex (DLPFC) [14, 71].

Based on raw RNAseq data of ROSMAP patients the abundance of different splice forms of genes in each ADRD patient and non-disease control case are analyzed. To assess the quality of the data, FastQC (vO.11.9) and MultiQC (vl.9) [73] are used to assess the need to trim reads and/or remove adapters. To remove rRNA contamination, trimmed reads are aligned to the rRNA reference genes as of the UCSC genome browser by BWA. Datasets thus are obtained mapped to the human reference genome (GENCODE Release 19, GRCh37.pl 3) using STAR (v.2.7.3a) [74]. After performing read quantification with StringTie [75] the IsoformS witchAnalyzeR R package (v 1.11.3) [76] is used to quantify isoform abundance allowing for determination individual isoform expression levels E_IFk of isoform k of a given gene. Based on such results, the IF value, that quantifies the relative isoform expression level defined as IF_{iso forml} = ^E,F1 [77] in a given gene, is determines. To determine 'iso form switches’, DIF is calculated between sample measures to assess the effect size between ADRD patients and (non-)disease by A/F = 1F_ADRD — lF_controi. DIF is measured on a scale of 0 to 1, with 0 indicating no (0%) change in usage between conditions and 1 points to complete (100%) change in usage. Isoforms that experienced >30% switch in usage (DIF > 10.31) and had an FDR-corrected -value cutoff of <0.05 ( -value < 0.05) are defined as “significant isoform switches”. Such calculations are fine grained by calculating such iso-form switches in patients that suffer from AD or MCI compared to controls, where blood and brain tissue samples are distinguished.

To correlate the abundance of isoforms of mRNAs and predicted MIMOSAS specific microRNAs, microRNA expression profiles of matching ADRD patients and non-disease control samples are reanalyzed. As part of the ROSMAP project data sets, also microRNA expression profiles were collected from post-mortem DLPFC samples. The quality of such raw microRNA-Seq data is reassessed with mimaQC [78] and FastQC. Trimming adapters with Cutadapt [79], such sequence reads are aligned on the human reference genome hg38 [80] using STAR [74]. Subsequent transcript quantification is performed using featureCounts [81] against all miRNA genes in miRbase [82]. Furthermore, the mentioned RNA-seq data is used to quantify the expression of long ncRNAs as well. In particular, featureCounts [81] is used against all IncRNA genes in LncRNAdb a database of literature described JncRNAs [83] .

Employing the MIMOSAS algorithm platform to establish a microRNA-splice mRNA cross correlation dataset. Correlating microRNA expression with splice isoform abundance within ADRD patient cohorts: Applying the MIMOSAS computational platform, splicerelevant interactions are predicted between microRNAs and genes with “significant isoform switches”. Since the predictions are based on sequence and structural data of the involved mRNAs and microRNAs, a catalog of binding events of microRNAs are established that can drive the abundance of isoform switches in ADRD. As a proof of concept, it is contemplated that the relevance of the underlying predictions are reflected by the significant enrichment of differentially expressed ncRNAs that were predicted to drive significantly changed spliceforms. Furthermore, this allows for generating a relatively small bona-fide set of predicted pairs of significantly switched isoforms and differentially expressed microRNAs.

Identification of ADRD related SNPs influencing structures with altered splicing outcomes through splice-relevant QTL analysis: To further corroborate the ADRD relevance of the microRNA-based splice mechanism, it is contemplated that splice-relevant genomic variations are enriched at genomic locations in the underlying mRNA that harbor splicerelevant secondary structures (FIG. 21) as SNPs can significantly strengthen or weaken RNA structures. In particular, previous genome-wide studies revealed that splice-relevant QTLs indeed were enriched within introns they regulate [84], Furthermore, previous genome-wide association studies revealed numerous genomic loci that were associated with ADRD [57-60], prompting for the determination of the presence and influence of such ADRD associated variations in or nearby splice-relevant secondary structures. While it was investigated into the weakening/strengthening impact of such variants on the stability of splice-relevant secondary structures through in-house developed algorithm RNAsnp [56], such ADRD associated loci are collected from NIAGADS Alzheimer’s GenomicsDB database [61], and previous genomewide association studies that revealed numerous genomic loci associated with ADRD [57-60]. Such known ADRD-specific loci is tested for their enrichment nearby splice-relevant secondary structures using a set of matched control variants (i.e. random control SNPs that match the index SNP for a number of variants in LD, minor allele frequency and distance to nearest intron), allowing for providing a subset of known ADRD specific loci that are causal for splicing events in ADRD. isoform Quantitative Trait Loci (iQTL) is determined to influence genes that produce different isoforms. Genomic variations in the vicinity of splice-relevant secondary structures are significantly associated with isoform expression levels of the underlying genes. Therefore, matching genomic data is selected in the ROSMAP project data sets that match samples of RNAseq data. Briefly, samples in the ROSMAP dataset were genotyped, quality controlled and imputed as described in [85]. In particular, isoform expression levels E/rof each isoform IF is determined in each (non-)ADRD sample and linearly regress to test for cis-associations between SNP dosages (MAF> 0.01) within lOOkb of the predicted microRNA binding site and splice relevant secondary structures using contemporary tools such as fastQTL [86] and/or matrixeQTL [58].

To estimate the number of iQTLs at any given false discovery rate (FDR), an adaptive permutation scheme [86] is applied which maintains a reasonable computational load by tailoring the number of permutations to the significance of the association. The empirical isoform-level p-value is computed for the most significant iQTL for each isoform. The probability of the observed presence of such cis-iQTLs is estimated using a set of matched control variants utilizing Fisher’s exact test.

The present disclosure allows for full circle identification of bona-fide MIMOSAS pairs of microRNAs/ncRNAs and isoforms in ADRD patients. In particular, predictions are filtered through correlations between microRNA/ncRNA expression and isoform abundance. To estimate the number of genes that are differentially spliced and carry splice relevant pre-mRNA secondary structures, 236 genes are considered that had a differentially spliced intron in ADRD (FDR < 0.01) [14]. Considering a total of 11,813 genes with splice-relevant secondary structures [24], an overlap of 174 genes are found. Based on such statistics, roughly 75% of differentially spliced genes in ADRD also harbor at least one splice-relevant structure.

Further genomic variants and iQTL analysis supports the relevance of the predictions, pointing to genomic variants that play a role in the MIMOSAS mechanism. As alternative to the iQTL analysis, a different analysis method of splice-relevant QTLs can be employed as introduced in [14]. In contrast to iQTL analysis, quantitative trait loci are determined as a function of intron usage instead of complete isoforms. Utilizing the LeafCutter algorithm, intron usage ratios (DPSI) are determined for each intron or excised region when comparing ADRD to controls [87].

Experimental identification and characterization of MIMOSAS candidate microRNAs in models of ADRD in vivo'. The innovative MIMOSAS computational platform as described herein, is powerful to predict microRNA-splice-variant pairs. Experimental confirmation of these interactions is a critical step in unleashing the predictive power relevant to progression of neurodegeneration and neuroprotective outcomes in AD RD. In FIGS. 22A- 22C, two examples of alternative spliced genes, Tau and NMNAT, are first characterized and MIMOSAS algorithm is applied to predict and experimentally identify microRNAs that can modulate the expression of pathogenic splice isoforms of Tau or neuroprotective splice isoforms of NMNAT. Next, the established experimental approach is used to characterize the ADRD MIMOSAS candidates. Cultured human HEK293T cells and Drosophila are used as model systems to integrate the mammalian specificity in HEK293T and the in vivo expression in Drosophila brains. Furthermore, ‘humanized flies’ are generated where human microRNA and mRNA target sequences are expressed in the brain and the MIMOSAS mode of regulation is observed in vivo. The experimental results identifies microRNA biomarkers for AD progression and neuroprotection.

Alternative splicing regulation of the pathogenic splice isoforms of Tau: It has been shown that altered splicing outcomes of human Tau gene significantly contributes to the pathogenesis of frontotemporal dementia, as the inclusion of exon 10 in Tau leads to the severe form of FTDP-17 145, 461. In particular, a small hairpin structure directly downstream of exon 10 competes with U1 snRNP for the 5’ splice site, promoting alternative splicing through skipping of exon 10. Notably, numerous mutations in the intronic regions adjacent to exon 10 that lead to an increased inclusion of exon 10, disrupting the balance between 3R-tau and 4R- tau isoforms have been identified to be significantly associated with AD (FIG. 23).

Establishing a list of candidate microRNAs that target Tau through MIMOSAS mechanism: A list of microRNAs that target the human Tau pre-mRNA sequence are generated (as previously described) based on the MIMOSAS pipeline. This list of microRNAs are overlayed and the sub-list of microRNAs that correlated to ADRD disease pathogenesis are established. Furthermore, the pipeline allows for identification iQTLs that further influence the abundance of isoforms.

Design a Tau splicing reporter TauEl Ox: Using the design of an in vivo reporter system (FIG. 18), the TauElOx reporter is designed where the N-term fragment of FP is inserted into exon 9, while the distinctive sequences of the C-terminal of GFP and DsRed are inserted into exon 10, and 1 1, respectively. The exon 10 inclusion events and the splice isoform 3R-tau are marked by GFP while 4R-Tau is marked by DsRed. The ratio metric measurement of GFP/DsRed within each cell indicates the 3R/4R splicing outcome and the toxicity of the Tau expression. Experimentally test the effect of candidate microRNAs in Tan splicing using the TauElOx reporter system: Human HEK293T cells and Drosophila are used as model systems to integrate the mammalian specificity in HEK293T and the in vivo expression in Drosophila brains. A stable cell line and transgenic Drosophila lines expressing the TauElOx reporter construct are established. Pre-miR sequences of candidate microRNAs are cloned into Luciferase-tagged expression vectors and expressed in TaulOx reporter cells and transgenic fly lines. Luciferase is used as an indicator of microRNA expression efficiency. The expression of splice forms are quantified by the fluorescence, and the ratio of 3R-tau/4R-tau isoforms with individual cells or brains are used as the splicing index.

Alternative splicing regulation of the neuroprotective splice isoforms of NMNAT: Humans have three NMNAT genes, encoding HsNMNATl, -2, and -3, with distinct enzyme kinetics and localizations [88, 89]. It was previously shown that HsNMNATl and -2 have robust neuroprotective activity in a mouse model of Tauopathy [90], while HsNMNAT3 has chaperone activity, similar to Drosophila NMNAT [91]. Work on axonal degeneration from several groups has shown protective effects by all three mammalian NMNAT proteins ([36- 38]). Notably, only one isoform of each gene has been included in all studies so far. Gene structure analyses show that all three human Nmnat genes are alternatively spliced (FIG. 24A). In particular, two splice variants of HsNMNAT3 (vl and v3-FKSG76) have been reported [92]. RT-PCR analysis was performed with variant-specific primer sets and the expression of splice variants was confirmed from total mRNA extracts from human embryonic brains and human HEK293T cells (FIG. 24B).

Establishing a list of candidate microRNAs that target human NMNAT genes through MIMOSAS mechanism: A list of microRNAs that target the human NMNAT pre-mRNA sequence is generated based on the MIMOSAS pipeline (previously described). This list of microRNAs is overlayed and the sub-list of microRNAs that correlated to ADRD disease is established.

Experimental testing of the effect of candidate microRNAs in NMNAT splicing using the splicing reporter system: A splicing reporter is designed for each of the NMNAT genes, and both cultured human HEK293T cells and Drosophila are used as models to determine the efficiency of microRNA in regulating the expression of the neuroprotective isoforms of NMNAT using the reporters (FIG. 18). The expression of splice form is quantified by fluorescence and the ratio of GFP/DsRed isoforms are used as the splicing index.

Experimental confirmation of ADRD MIMOSAS genes: Establishing a priority list of ADRD candidate microRNA-mRNA pairs: 174 genes were found that had at least one spliced intron [14] and a splice-relevant secondary structures [24], showing that roughly 75% of differentially spliced genes in ADRD also harbor at least one splice-relevant structure. Considering such an upper bound, a manageable number of candidate microRNA-gene pairs is obtained where the binding of a microRNA indeed influence splice-relevant secondary structures and drive splicing of differentially expressed genes in ADRD. To stratify such a candidate list and find a subset of roughly 10 pairs for further experimental validation, microRNA-gene pairs are scored based on predicted DDG values as well as correlation of their expression values.

Experimental testing of the effect of candidate ADRD MIMOSAS splicing using the splicing reporter system: A FP-based splicing reporter is designed for each of the 10 top ranked genes, and cultured human HEK293T cells are used as primary screening model to determine the efficiency of microRNA in regulating the expression of the ADRD relevant splice isoforms. Out of the 10 genes screened in HEK cells, the top 5 genes are selected to generate expression models in Drosophila brains to identify ADRD relevant mechanisms.

The candidate microRNAs serve as biomarkers for AD disease severity and progression, or, more importantly, to reveal therapeutic venues to enhance neuroprotection.

The following patents, applications and publications as listed below and throughout this document are hereby incorporated by reference in their entirety herein.

Reference List for Example 1

1. Allo, M., et al. (2009). Control of alternative splicing through siRNA-mediated transcriptional gene silencing. Nat Struct Mol Biol 16, 717-724.

2. Ameyar-Zazoua, M., et al. (2012). Argonaute proteins couple chromatin silencing to alternative splicing. Nat Struct Mol Biol 19, 998-1004.

3. Bush, S.J., et al. (2017). Alternative splicing and the evolution of phenotypic novelty. Philos Trans R Soc Lond B Biol Sci 372.

4. Carthew, R.W., et al. (2017). MicroRNA function in Drosophila melanogaster. Semin Cell Dev Biol 65, 29-37.

5. Castanotto, D., et al. (2009). CRM1 mediates nuclear-cytoplasmic shuttling of mature microRNAs. Proc Natl Acad Sci U S A 106, 21655-21659.

6. Chawla, G., et al. (2016). Analysis of MicroRNA Function in Drosophila. Methods Mol Biol 1478, 79-94. 7. Chen, L., et al. (2014). Correcting for differential transcript coverage reveals a strong relationship between alternative splicing and organism complexity. Mol Biol Evol 31, 1402- 1413.

8. Cordes, K.R., et al. (2009). miR- 145 and miR- 143 regulate smooth muscle cell fate and plasticity. Nature 460, 705-710.

9. Feng, S., et al. (2017). Improved split fluorescent proteins for endogenous protein labeling. Nat Commun 8, 370.

10. Foldes-Papp, Z., et al. (2009). Trafficking of mature miRNA-122 into the nucleus of live liver cells. Curr Pharm Biotechnol 10, 569-578.

11. Forstemann, K., et al. (2007). Drosophila micro RNAs are sorted into functionally distinct argonaute complexes after production by dicer- 1. Cell 130, 287-297.

12. Gay, L.A., et al. (2018). Modified Cross-Linking, Ligation, and Sequencing of Hybrids (qCLASH) Identifies Kaposi's Sarcoma-Associated Herpesvirus MicroRNA Targets in Endothelial Cells. J Virol 92.

13. Gebert, L.F.R., and MacRae, I.J. (2019). Regulation of microRNA function in animals. Nat Rev Mol Cell Biol 20, 21-37.

14. Ghosh, T., et al. (2008). MicroRNA-mediated up-regulation of an alternatively polyadenylated variant of the mouse cytoplasmic beta-actin gene. Nucleic Acids Res 36, 6318- 6332.

15. Gonzalez, I., et al. (2015). A IncRNA regulates alternative splicing via establishment of a splicing-specific chromatin signature. Nat Struct Mol Biol 22, 370-376.

16. Gruber, A.R., et al. (2008). The Vienna RNA websuite. Nucleic Acids Res 36, W70- 74.

17. Gueroussov, S., et al. (2017). Regulatory Expansion in Mammals of Multivalent hnRNP Assemblies that Globally Control Alternative Splicing. Cell 170, 324-339 e323.

18. Helwak, A., et al. (2013). Mapping the human miRNA interactome by CLASH reveals frequent noncanonical binding. Cell 153, 654-665.

19. Herzel, L., et al. (2017). Splicing and transcription touch base: co-transcriptional spliceosome assembly and function. Nat Rev Mol Cell Biol 18, 637-650. 20. Jensen, C.J., et al. (2009). Splicing, cis genetic variation and disease. Biochem Soc Trans 37, 1311-1315.

21. Jin, Y., et al. (2011). New insights into RNA secondary structure in the alternative splicing of pre-mRNAs. RNA Biol 8, 450-457.

22. Jopling, C.L., et al. (2005). Modulation of hepatitis C virus RNA abundance by a liverspecific MicroRNA. Science 309, 1577-1581.

23. Kalmykova, S., et al. (2021). Conserved long-range base pairings are associated with pre-mRNA processing of human genes. Nature communications 12, 2300.

24. Kim, H.K., et al. (2018). Alternative splicing isoforms in health and disease. Pflugers Arch 470, 995-1016.

25. Kishore, S., and Stamm, S. (2006). The snoRNA HBII-52 regulates alternative splicing of the serotonin receptor 2C. Science 311, 230-232.

26. Kornblihtt, A.R. (2006). Chromatin, transcript elongation and alternative splicing. Nat Struct Mol Biol 13, 5-7.

27. Lim, K.H., et al. (2011). Using positional distribution to identify splicing elements and predict pre-mRNA processing defects in human genes. Proc Natl Acad Sci U S A 108, 11093- 11098.

28. Liu, J., et al. (2012). Expanding the action of duplex RNAs into the nucleus: redirecting alternative splicing. Nucleic Acids Res 40, 1240-1250.

29. Liu, J., et al. (2015). Modulation of Splicing by Single-Stranded Silencing RNAs. Nucleic Acid Ther 25, 113-120.

30. Liu, Y., et al. (2021). Noncoding RNAs regulate alternative splicing in Cancer. J Exp Clin Cancer Res 40, 11 .

31. Lorenz, R., et al. (2011). ViennaRNA Package 2.0. Algorithms Mol Biol 6, 26.

32. Lorenz, R., et al. (2016). RNA folding with hard and soft constraints. Algorithms Mol Biol 11, 8.

33. Lu, H., et al. (2010). MicroRNA-223 regulates Glut4 expression and cardiomyocyte glucose metabolism. Cardiovasc Res 86, 410-420. 34. Mortensen, R.D., et al. (2011). Posttranscriptional activation of gene expression in Xenopus laevis oocytes by microRNA-protein complexes (microRNPs). Proc Natl Acad Sci U S A 108, 8281-8286.

35. Muckstein, U., et al. (2006). Thermodynamics of RNA-RNA binding. Bioinformatics 22, 1177-1182.

36. O'Brien, J., et al. (2018). Overview of MicroRNA Biogenesis, Mechanisms of Actions, and Circulation. Front Endocrinol (Lausanne) 9, 402.

37. Orom, U.A., et al. (2008). MicroRNA-lOa binds the 5'UTR of ribosomal protein mR As and enhances their translation. Mol Cell 30, 460-471.

38. Park, J., et al. (2019). MicroRNA miR-1002 Enhances NMNAT-Mediated Stress Response by Modulating Alternative Splicing. iScience 19, 1048-1064.

39. Pervouchine, D.D., et al. (2012). Evidence for widespread association of mammalian splicing and conserved long-range RNA structures. Rna 18, 1-15.

40. Pisignano, G., and Ladomery, M. (2021). Epigenetic Regulation of Alternative Splicing: How LncRNAs Tailor the Message. Noncoding RNA 7.

41. Pushpavalli, S.N., et al. (2012). Drosophila Argonaute-1 is critical for transcriptional cosuppression and heterochromatin formation. Chromosome Res 20, 333-351.

42. Pushpavalli, S.N., et al. (2014). Argonaute-1 functions as a mitotic regulator by controlling Cyclin B during Drosophila early embryogenesis. FASEB J 28, 655-666.

43. Raker, V.A., et al. (2009). Modulation of alternative splicing by long-range RNA structures in Drosophila. Nucleic Acids Res 37, 4533-4544.

44. Ruan, K., et al. (2015). Alternative splicing of Drosophila Nmnat functions as a switch to enhance neuroprotection under stress. Nat Commun 6, 10057.

45. Sterne-Weiler, T., et al. (2011). Loss of exon identity is a common mechanism of human inherited disease. Genome Res 21, 1563-1571.

46. Tafer, H., and Hofacker, I.L. (2008). RNAplex: a fast tool for RNA-RNA interaction search. Bioinformatics 24, 2657-2663.

47. Till, S., et al. (2007). A conserved motif in Argonaute-interacting proteins mediates functional interactions through the Argonaute PIWI domain. Nat Struct Mol Biol 14, 897-903. 48. Tsai, N.P., et al. (2009). MicroRNA mir-346 targets the 5 '-untranslated region of receptor-interacting protein 140 (RIP140) mRNA and up-regulates its protein expression. Biochem J 424, 411-418.

49. Vasudevan, S., et al. (2007). Switching from repression to activation: microRNAs can up-regulate translation. Science 318, 1931-1934.

50. Weinmann, L., et al. (2009). Importin 8 is a gene silencing factor that targets argonaute proteins to distinct mRNAs. Cell 136, 496-507.

51. Yang, J.S., et al. (2014). Intertwined pathways for Argonaute-mediated microRNA biogenesis in Drosophila. Nucleic Acids Res 42, 1987-2002.

Reference List for Example 2

1. Park, J., et al., MicroRNA miR-1002 Enhances NMNAT-Mediated Stress Response by Modulating Alternative Splicing. iScience, 2019. 19: p. 1048-1064.

2. Collaborators, U.S.B.o.D., The state of US health, 1990-2010: burden of diseases, injuries, and risk factors. JAMA, 2013. 310(6): p. 591-608.

3. Liu, E.Y., et al., RNA metabolism in neurodegenerative disease. Dis Model Meeh, 2017. 10(5): p. 509-518.

4. Idda, M.L., et al., Noncoding RNAs in Alzheimer's disease. Wiley Interdiscip Rev RNA, 2018. 9(2).

5. Ranum, L.P. and T.A. Cooper, RNA-mediated neuromuscular disorders. Annu Rev Neurosci, 2006. 29: p. 259-77.

6. Licatalosi, D.D. and R.B. Darnell, Splicing regulation in neurologic disease. Neuron, 2006. 52(1): p. 93-101.

7. Ramaswami, M., et al., Altered Ribostasis: RNA-Protein Granules in Degenerative Disorders. Cell, 2013. 154(4): p. 727-36.

8. Li, Y.R., et al., Stress granules as crucibles of ALS pathogenesis. J Cell Biol, 2013. 201(3): p. 361-72.

9. Miya Shaik, M., et al., The Role of microRNAs in Alzheimer's Disease and Their Therapeutic Potentials. Genes (Basel), 2018. 9(4). 10. Saraiva, C., et al., MicroRNA: Basic concepts and implications for regeneration and repair of neurodegenerative diseases. Biochem Pharmacol, 2017. 141: p. 118-131.

11. Luo, Q. and Y. Chen, Long noncoding RNAs and Alzheimer's disease. Clin Interv Aging, 2016. 1 1: p. 867-72.

12. Li, D., et al., Insights into IncRNAs in Alzheimer’s disease mechanisms. RNA Biology, 2021. 18(7): p. 1037-1047.

13. Liu, Y., et al., LncRNAs as the Regulators of Brain Function and Therapeutic Targets for Alzheimer’ s Disease. Aging and disease, 2022. 13(3): p. 837-851.

14. Raj, T., et al., Integrative transcriptome analyses of the aging brain implicate altered splicing in Alzheimer's disease susceptibility. Nat Genet, 2018. 50(11): p. 1584-1592.

15. Bush, S.J., et al., Alternative splicing and the evolution of phenotypic novelty. Philos Trans R Soc Lond B Biol Sci, 2017. 372(1713).

16. Chen, L., et al., Correcting for differential transcript coverage reveals a strong relationship between alternative splicing and organism complexity. Mol Biol Evol, 2014. 31(6): p. 1402-13.

17. Lim, K.H., et al., Using positional distribution to identify splicing elements and predict pre-mRNA processing defects in human genes. Proc Natl Acad Sci U S A, 2011. 108(27): p. 11093-8.

18. Steme-Weiler, T., et al., Loss of exon identity is a common mechanism of human inherited disease. Genome Res, 2011. 21(10): p. 1563-71.

19. Helwak, A., et al., Mapping the human miRNA interactome by CLASH reveals frequent noncanonical binding. Cell, 2013. 153(3): p. 654-65.

20. Gay, L.A., et al., Modified Cross-Linking, Ligation, and Sequencing of Hybrids (qCLASH) Identifies Kaposi's Sarcoma-Associated Herpesvirus MicroRNA Targets in Endothelial Cells. J Virol, 2018. 92(8).

21. Gonzalez, I., et al., A IncRNA regulates alternative splicing via establishment of a splicing-specific chromatin signature. Nat Struct Mol Biol, 2015. 22(5): p. 370-6.

22. Kishore, S. and S. Stamm, The snoRNA HBII-52 regulates alternative splicing of the serotonin receptor 2C. Science, 2006. 311(5758): p. 230-2. 23. Kim, H.K., et al., Alternative splicing isoforms in health and disease. Pflugers Arch, 2018. 470(7): p. 995-1016.

24. Kalmykova, S., et al., Conserved long-range base pairings are associated with pre- mRNA processing of human genes. Nat Commun, 2021. 12(1): p. 2300.

25. Pervouchine, D.D., et al., Evidence for widespread association of mammalian splicing and conserved long-range RNA structures. RNA, 2012. 18(1): p. 1-15.

26. Raker, V.A., et al., Modulation of alternative splicing by long-range RNA structures in Drosophila. Nucleic Acids Res, 2009. 37(14): p. 4533-44.

27. Gueroussov, S., et al., Regulatory Expansion in Mammals of Multivalent hnRNP Assemblies that Globally Control Alternative Splicing. Cell, 2017. 170(2): p. 324-339 e23.

28. Lorenz, R., et al., ViennaRNA Package 2.0. Algorithms Mol Biol, 2011. 6: p. 26.

29. Oesterreich, F.C., et al., Splicing of Nascent RNA Coincides with Intron Exit from RNA Polymerase II. Cell, 2016. 165(2): p. 372-381.

30. Zhang, T., et al., RNALocate: a resource for RNA subcellular localizations. Nucleic Acids Res, 2017. 45(D1): p. D135-D138.

31. Cemilogar, F.M., et al., Chromatin- associated RNA interference components contribute to transcriptional regulation in Drosophila. Nature, 2011. 480(7377): p. 391-5.

32. Huang, V. and L.C. Li, Demystifying the nuclear function of Argonaute proteins. RNA Biol, 2014. 11(1): p. 18-24.

33. Huang, V., et al., Agol Interacts with RNA polymerase II and binds to the promoters of actively transcribed genes in human cancer cells. PLoS Genet, 2013. 9(9): p. el003821.

34. Pushpavalli, S.N., et al., Drosophila Argonaute-1 is critical for transcriptional cosuppression and heterochromatin formation. Chromosome Res, 2012. 20(3): p. 333-51.

35. Taliaferro, J.M., et al., Two new and distinct roles for Drosophila Argonaute-2 in the nucleus: alternative pre-mRNA splicing and transcriptional repression. Genes Dev, 2013. 27(4): p. 378-89.

36. Coleman, M.P., The challenges of axon survival: introduction to the special issue on axonal degeneration. Exp Neurol, 2013. 246: p. 1-5. 37. Ali, Y.O., et al., NMNATs, evolutionarily conserved neuronal maintenance factors.

Trends Neurosci, 2013. 36(11): p. 632-40.

38. Coleman, M.P. and M.R. Freeman, Wallerian degeneration, wld(s), and nmnat. Annu Rev Neurosci, 2010. 33: p. 245-67.

39. Ruan, K., et al., Alternative splicing of Drosophila Nmnat functions as a switch to enhance neuroprotection under stress. Nat Commun, 2015. 6: p. 10057.

40. Ali, Y.O., et al., Assaying locomotor, learning, and memory deficits in Drosophila models of neurodegeneration. J Vis Exp, 2011(49).

41. Brazill, J.M., et al., Nmnat mitigates sensory dysfunction in a Drosophila model of paclitaxel-induced peripheral neuropathy. Dis Model Meeh, 2018. 11(6).

42. Brazill, J.M., et al., Quantitative cell biology of neurodegeneration in Drosophila through unbiased analysis of fluorescently tagged proteins using ImageJ. J Vis Exp, 2018. In Press.

43. Ali, Y.O., et al., Nicotinamide mononucleotide adenylyltransferase is a stress response protein regulated by the heat shock factor/hypoxia-inducible factor 1 alpha pathway. J. Biol. Chem., 2011. 286(21): p. 19089-99.

44. Ali, Y.O., et al., NMNAT suppresses tau-induced neurodegeneration by promoting clearance of hyperphosphorylated tau oligomers in a Drosophila model of tauopathy. Hum. Mol. Genet., 2012. 21(2): p. 237-50.

45. Liu, F. and C.X. Gong, Tau exon 10 alternative splicing and tauopathies. Mol Neurodegener, 2008. 3: p. 8.

46. Grover, A., et al., 5' splice site mutations in tau associated with the inherited dementia FTDP-17 affect a stem-loop structure that regulates alternative splicing of exon 10. J Biol Chem, 1999. 274(21): p. 15134-43.

47. Muckstein, U., et al., Thermodynamics of RNA-RNA binding. Bioinformatics, 2006. 22(10): p. 1177-82.

48. Tafer, H., et al., Fast accessibility-based prediction of RNA-RNA interactions. Bioinformatics, 2011. 27(14): p. 1934-40. 49. Clote, P., et al., Expected distance between terminal nucleotides of RNA secondary structures. J Math Biol, 2012. 65(3): p. 581-99.

50. Lorenz, R., et al., RNA folding with hard and soft constraints. Algorithms Mol Biol, 2016. 11: p. 8.

51. Bemhart, S.H., et al., RNAalifold: improved consensus structure prediction for RNA alignments. BMC Bioinformatics, 2008. 9: p. 474.

52. Hofacker, I.L., et al., Secondary structure prediction for aligned RNA sequences. J Mol Biol, 2002. 319(5): p. 1059-66.

53. Rivas, E., J. Clements, and S.R. Eddy, A statistical test for conserved RNA structure shows lack of evidence for structure in IncRNAs. Nat Methods, 2017. 14(1): p. 45-48.

54. Wright, P.R., et al., Comparative genomics boosts target prediction for bacterial small RNAs. Proc Natl Acad Sci U S A, 2013. 110(37): p. E3487-96.

55. Hartung, J., A Note on Combining Dependent Tests of Significance. Biometrical Journal, 1999. 41(7): p. 849-855.

56. Wegener, W.A., On an exact starting expression for macromolecular hydrodynamic models. Biopolymers, 1986. 25(4): p. 627-37.

57. Schwartzentruber, J., et al., Genome-wide meta-analysis, fine-mapping and integrative prioritization implicate new Alzheimer's disease risk genes. Nat Genet, 2021. 53(3): p. 392- 402.

58. Kunkle, B.W., et al., Genetic meta-analysis of diagnosed Alzheimer’s disease identifies new risk loci and implicates Abeta, tau, immunity and lipid processing. Nat Genet, 2019. 51(3): p. 414-430.

59. Jansen, T.E., et al., Genome-wide meta-analysis identifies new loci and functional pathways influencing Alzheimer's disease risk. Nat Genet, 2019. 51(3): p. 404-413.

60. Bellenguez, C., et al., New insights on the genetic etiology of Alzheimer’s and related dementia. medRxiv, 2020: p. 2020.10.01.20200659.

61. Greenfest-Allen, E., et al., NIAGADS Alzheimer’s GenomicsDB: A resource for exploring Alzheimer’s Disease genetic and genomic knowledge. bioRxiv, 2020: p. 2020.09.23.310276. 62. Grosswendt, S., et al., Unambiguous identification of mi RN A: target site interactions by different types of ligation reactions. Mol Cell, 2014. 54(6): p. 1042-1054.

63. Kudla, G., et al., Cross-linking, ligation, and sequencing of hybrids reveals RNA-RNA interactions in yeast. Proc Natl Acad Sci U S A, 2011. 108(24): p. 10010-5.

64. Plotnikova, O., et al., Comprehensive Analysis of Human microRNA-mRNA Interactome. Front Genet, 2019. 10: p. 933.

65. Wen, M., et al., DeepMirTar: a deep-learning approach for predicting human miRNA targets. Bioinformatics, 2018. 34(22): p. 3781-3787.

66. Gu, T., et al., miTAR: a hybrid deep learning-based approach for predicting miRNA targets. BMC Bioinformatics, 2021. 22(1): p. 96.

67. Pan, X., et al., Prediction of RNA-protein sequence and structure binding preferences using deep convolutional and recurrent neural networks. BMC Genomics, 2018. 19(1): p. 511.

68. Johnsson, P., et al., Transcriptional kinetics and molecular functions of long noncoding RNAs. Nature Genetics, 2022. 54(3): p. 306-317.

69. Faghihi, M.A., et al., Expression of a noncoding RNA is elevated in Alzheimer's disease and drives rapid feed-forward regulation of beta-secretase. Nat Med, 2008. 14(7): p. 723-30.

70. Ciarlo, E., et al., An intronic ncRNA-dependent regulation of SORL1 expression affecting A formation is upregulated in post-mortem Alzheimer's disease brain samples. Disease Models & Mechanisms, 2013. 6(2): p. 424-433.

71. De Jager, P.L., et al., A multi-omic atlas of the human frontal cortex for aging and Alzheimer’s disease research. Sci Data, 2018. 5: p. 180142.

72. Bennett, D.A., et al., Overview and findings from the religious orders study. Curr Alzheimer Res, 2012. 9(6): p. 628-45.

73. Ewels, P., et al., MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics, 2016. 32(19): p. 3047-8.

74. Dobin, A., et al., STAR: ultrafast universal RNA-seq aligner. Bioinformatics, 2013. 29(1): p. 15-21.

75. Pertea, M., et al., StringTie enables improved reconstruction of a transcriptome from RNA-seq reads. Nat Biotechnol, 2015. 33(3): p. 290-5. 76. Vitting-Seerup, K. and A. Sandelin, IsoformS witchAnalyzeR: analysis of changes in genome-wide patterns of alternative splicing and its functional consequences. Bioinformatics, 2019. 35(21): p. 4469-4471.

77. Ferrarini, M.G., et al., Genome-wide bioinformatic analyses predict key host and viral factors in SARS-CoV-2 pathogenesis. Commun Biol, 2021. 4(1): p. 590.

78. Aparicio-Puerta, E., et al., mirnaQC: a webserver for comparative quality control of miRNA-seq data. Nucleic Acids Res, 2020. 48(W1): p. W262-W267.

79. Martin, M., Cutadapt Removes Adapter Sequences From High-Throughput Sequencing Reads. EMBnet Journal, 201 1. 17(1): p. 10-12.

80. Langmead, B., et al., Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol, 2009. 10(3): p. R25.

81. Liao, Y., et al., featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics, 2014. 30(7): p. 923-30.

82. Kozomara, A. and S. Griffiths-Jones, miRBase: annotating high confidence microRNAs using deep sequencing data. Nucleic Acids Res, 2014. 42(Database issue): p. D68- 73.

83. Quek, X.C., et al., IncRNAdb v2.0: expanding the reference database for functional long noncoding RNAs. Nucleic Acids Res, 2015. 43(Database issue): p. D168-73.

84. Li, Y.I., et al., RNA splicing is a primary link between genetic variation and disease. Science, 2016. 352(6285): p. 600-4.

85. De Jager, P.L., et al., A genome-wide scan for common variants affecting the rate of age-related cognitive decline. Neurobiol Aging, 2012. 33(5): p. 1017 el-15.

86. Ongen, H., et al., Fast and efficient QTL mapper for thousands of molecular phenotypes. Bioinformatics, 2016. 32(10): p. 1479-85.

87. Li, Y.I., et al., Annotation-free quantification of RNA splicing using LeafCutter. Nat Genet, 2018. 50(1): p. 151-158.

88. Nikiforov, A., et al., Pathways and subcellular compartmentation of NAD biosynthesis in human cells: from entry of extracellular precursors to mitochondrial NAD generation. J Biol Chem, 2011. 286(24): p. 21767-78. 89. Berger, F., et al., Subcellular compartmentation and differential catalytic properties of the three human nicotinamide mononucleotide adenylyltransferase isoforms. J Biol Chem, 2005. 280(43): p. 36334-41.

90. Ljungberg, M.C., et al., CREB-activity and nmnat2 transcription are down-regulated prior to neurodegeneration, while NMNAT2 over-expression is neuroprotective, in a mouse model of human tauopathy. Hum Mol Genet, 2012. 21(2): p. 251-67.

91. Zhai, R.G., et al., NAD synthase NMNAT acts as a chaperone to protect against neurodegeneration. Nature, 2008. 452(7189): p. 887-91.

92. Felici, R., et al., Insight into molecular and functional properties of NMNAT3 reveals new hints of NAD homeostasis within human mitochondria. PLoS One, 2013. 8(10): p. e76938.

93. Park, J., et al., Knight in Splicing Armor: Alternative Splicing as a Neuroprotective Mechanism. OBM Genetics, 2020. 4(1): p. 21.

TABLES

TABLE 1. A list of primers used in Example 1.

TABLE 2. miRNAs of STEM 1 3’SS B0X1.

TABLE 3. miRNAs of STEM 1 3’SS B0X2.

TABLE 4. miRNAs of STEM 2 5’SS B0X1.

TABLE 5. miRNAs of STEM1 5’SS B0X2.

SEQUENCES

SEQ ID NO: 1 - BOX 1

GATACGTACACTGA

SEQ ID NO: 2 - BOX 2

TCAGTGTGCGTATC

SEQ ID NO: 3 - hTau exlO dual colors splicing reporter

ATGGTGAGCAAGGGCGAGGAGGTGAGAGTGGCTGGCTGCGGATGTGACTCAACC

TCCCGTCACTCCCCAGACTGCCTCTGCCAAGTCCGAAAGTGGAGGCATCCTTGCG

AGCAAGTAGGCGGGTCCAGGGTGGCGCATGTCACTCATCGAAAGTGGAGGCGTC

CTTGCGAGCAAGCAGGCGGGTCCAGGGTGGCGTGTCACTCATCCTTTTTTCTGGC

TACCAAAGGTGCAGATAATTTTAGCCTTGAAATTAGCAGGTCTTGATATCCTGTT

CACCGGGGTGGTGCCCATCCTGGTCGAGCTGGACGGCGACGTAAACGGCCACAA

GTTCAGCGTGTCCGGCGAGGGCGAGGGCGATGCCACCTACGGCAAGCTGACCCT

GAAGTTCATCTGCACCACCGGCAAGCTGCCCGTGCCCTGGCCCACCCTCGTGACC

ACCCTGACCTACGGCGTGCAGTGCTTCAGCCGCTACCCCGACCACATGAAGCAG

CACGACTTCTTCAAGTCCGCCATGCCCGAAGGCTACGTCCAGGAGCGCACCATCT

TCTTCAAGGACGACGGCAACTACAAGACCCGCGCCGAGGTGAAGTTCGAGGGCG

ACACCCTGGTGAACCGCATCGAGCTGAAGGGCATCGACTTCAAGGAGGACGGCA

ACATCCTGGGGCACAAGCTGGAGTACAACTACAACAGCCACAACGTCTATATCA

TGGCCGACAAGCAGAAGAACGGCATCAAGGTGAACTTCAAGATCCGCCACAACA

TCGAGGACGGCAGCGTGCAGCTCGCCGACCACTACCAGCAGAACACCCCCATCG

GCGACGGCCCCGTGCTGCTGCCCGACAACCACTACCTGAGCACCCAGTCCGCCCT

GAGCAAAGACCCCAACGAGAAGCGCGATCACATGGTCCTGCTGGAGTTCGTGAC

CGCCGCCGGGATCACTCTCGGCATGGACGAGCTGTACAAGTAAAATAAGAAGCT

GGATCTTAGCAACGTCCAGTCCAAGTGTGGCTCAAAGGATAATATCAAACACGT

CCCGGGAGGCGGCAGTGTGAGTACCTTCACACGTCCCATGCGCCGTGCTGTGGCT

TGAATTATTAGGAAGTGGTGTGAGTGCGTACACTTGCGAGACACTGCATAGAAT

AAATCCTTCTTGGGCTCTCAGGATCTGGCTGCGACCTCTGGGTGAATGTAGCCCG

GCTCCCCACATTCCCCCACACGGTCCACTGTTCCCAGAAGCCCCTTTCTCATTGA

GTTACACCCTTAATAAAATAATCCCATTTTATCCTTTTTGTCTCTCTGTCTTCCTCT

CTCTCTGCCTTTCCTCTTCTCTCTCCTCCTCTCTCATCTCCAGGATAACATGGCCAT

CATCAAGGAGTTCATGCGCTTCAAGGTGCACATGGAGGGCTCCGTGAACGGCCA CGAGTTCGAGATCGAGGGCGAGGGCGAGGGCCGCCCCTACGAGGGCACCCAGA

CCGCCAAGCTGAAGGTGACCAAGGGTGGCCCCCTGCCCTTCGCCTGGGACATCCT

GTCCCCTCAGTTCATGTACGGCTCCAAGGCCTACGTGAAGCACCCCGCCGACATC

CCCGACTACTTGAAGCTGTCCTTCCCCGAGGGCTTCAAGTGGGAGCGCGTGATGA

ACTTCGAGGACGGCGGCGTGGTGACCGTGACCCAGGACTCCTCCCTGCAGGACG

GCGAGTTCATCTACAAGGTGAAGCTGCGCGGCACCAACTTCCCCTCCGACGGCCC

CGTAATGCAGAAGAAGACCATGGGCTGGGAGGCCTCCTCCGAGCGGATGTACCC

CGAGGACGGCGCCCTGAAGGGCGAGATCAAGCAGAGGCTGAAGCTGAAGGACG

GCGGCCACTACGACGCTGAGGTCAAGACCACCTACAAGGCCAAGAAGCCCGTGC

AGCTGCCCGGCGCCTACAACGTCAACATCAAGTTGGACATCACCTCCCACAACG

AGGACTACACCATCGTGGAACAGTACGAACGCGCCGAGGGCCGCCACTCCACCG

GCGGCATGGACGAGCTGTACAAGggcggatccCCTGCTGCCAAGAGGGTCAAGTTGG

ACTGA

SEQ ID NO: 4 - F-Rpl3-E2/3

CCTGGCTCCAAGATCAACAAG

SEQ ID NO: 5 - R-Rpl3-RD

CGTTTCATCATCCAGACTCGAG

SEQ ID NO: 6 - R-Rpl3-RG

GCTCACACAACGTTTAGCGATTTG

SEQ ID NO: 7 - R-Rpl3-RA/H

CTGCGAGTGGGCAATCACAC

SEQ ID NO: 8 - F-9c

AATTCATTTTTGCTGTTTCTTTGGTATTCTAGCTGTAGATTGTTTCACGCACATTGT

ATATCATCTAAAGCTTTTATACCAAAGCTCCAGCTTAAATC

SEQ ID NO: 9 - R-9c

TCGAGATTTAAGCTGGAGCTTTGGTATAAAAGCTTTAGATGATATACAATGTGCG

TGAAACAATCTACAGCTAGAATACCAAAGAAACAGCAAAAATG SEQIDNO: 10-F-210

AATTCAAAGGTGCTTATTGCAGCTGCTGGCCACTGCACAAGATTAGACTTAAGAC

TCTTGTGCGTGTGACAGCGGCTATTGTAAGAGGCCATAGAAGCAACAGCCC

SEQIDNO: 11 -R-210

TCGAGGGCTGTTGCTTCTATGGCCTCTTACAATAGCCGCTGTCACACGCACAAGA

GTCTTAAGTCTAATCTTGTGCAGTGGCCAGCAGCTGCAATAAGCACCTTTG

SEQIDNO: 12-F-304

AATTCGCAGCATTGAATAATCTCAATTTGTAAATGTGAGCGGTTTAAGCCATTTG

ACGCACTCACTTTGCAATTGGAGATTGCTCGAGACTGCC

SEQIDNO: 13-R-304

TCGAGGCAGTCTCGAGCAATCTCCAATTGCAAAGTGAGTGCGTCAAATGGCTTA

AACCGCTCACATTTACAAATTGAGATTATTCAATGCTGCG

SEQIDNO: 14-F-988

AATTCGACGGCGGTACCGGGCATTTTGGGTGTGTGATTTGTAGCAAAGTGATATG

TATTTGATCATCCCCTTGTTGCAAACCTCACGCCAAAGATGATCTGCGAC

SEQIDNO: 15-R-988

TCGAGTCGCAGATCATCTTTGGCGTGAGGTTTGCAACAAGGGGATGATCAAATA

CATATCACTTTGCTACAAATCACACACCCAAAATGCCCGGTACCGCCGTCG

SEQIDNO: 16-F-992

AATTCATTTTCCCAAGTGCCTGGTATCAGCAAAGTGTTATTTTTTATGTTTATGTA

AAGTACACGTTTCTGGTACTAAGTACTTCGAGAAAGTTACCC

SEQIDNO: 17-R-992

TCGAGGGTAACTTTCTCGAAGTACTTAGTACCAGAAACGTGTACTTTACATAAAC

ATAAAAAATAACACTTTGCTGATACCAGGCACTTGGGAAAATG

SEQ ID NO: 18 - miR-210-3p FISH probe cloning Sense

CTTGTGCGTGTGACAGCGGCTAT SEQ ID NO: 19 - miR-210-3p FISH probe cloning anti-sense

ATAGCCGCTGTCACACGCACAAG

SEQ ID NO: 20

GGGUGAAACAUCGUCAUUAGCUUCUCAGCCACAUCAACAAAAAC

SEQ ID NO: 21

GAUAAAGUUUUGAUGUGGCAGACAGUAAGCUGAUGAUGGUGUUGUCAGGUGG

GAUACCUUUUCCCG

SEQ ID NO: 22 - Box I

AACAUCGUCAUUAGCUUCUCAGCCACAUCAA

SEQ ID NO: 23 - Box II

UUGAUGUGGCAGACAGUAAGCUGAUGAUGGUGUU

SEQ ID NO: 24 - Box I

GAUACGUACACUGA

SEQ ID NO: 25 - Box II

UCAGUGUGCGUAUC

SEQ ID NO: 26 - Box I

CACACACACAAAAU

SEQ ID NO: 27 - Box II

GUUUUGUGUGUGUG

SEQ ID NO: 28 - Box II

GCGCUGUUUUGUGUGUGUGCGA

SEQ ID NO: 29 - Box III

UCGCCCCACAUAAACAGCGC

Claims

1. A synthetic nucleic acid comprising: a reporter gene; a first stem loop; a second stem loop; a nuclear export signal; and an exon 10 fragment from the human Tau gene; wherein the nuclear export signal and the exon 10 fragment from the human Tau gene are located between the first stem loop and the second stem loop.

2. The synthetic nucleic acid of claim 1 , wherein the first stem loop has a 3 ’ splice site.

3. The synthetic nucleic acid of claim 1 or 2, wherein the second stem loop has a 5’ splice site.

4. The synthetic nucleic acid of any one of claims 1-3, wherein the reporter gene is eGFP, or a variant thereof.

5. A method of screening for miRNAs that modulate the alternative splicing of exon 10 of the human Tau gene, comprising: providing the synthetic nucleic acid of any one of claims 1-4; incubating the synthetic nucleic acid with a miRNA; determining whether the miRNA modulates the alternative splicing of exon 10 of the human Tau gene.

6. The method of claim 5, wherein increased expression of the reporter gene in the cytoplasm provides a pathological outcome.

7. The method of claim 5 or 6, wherein increased expression of the reporter gene in the nucleus provides a non-pathological outcome.

8. A method of diagnosing Alzheimer’s Disease or an increased risk of developing Alzheimer’s Disease, comprising; providing the synthetic nucleic acid of any one of claims 1-4; incubating the synthetic nucleic acid with a miRNA; and determining whether the miRNA modulates the alternative splicing of exon 10 of the human Tau gene; wherein increased expression of the reporter gene in the cytoplasm confers a pathological outcome and a diagnosis of Alzheimer’s Disease or an increased risk of developing Alzheimer’s Disease.

9. The method of claim 8, further comprising administering a specific treatment for Alzheimer’s Disease when the miRNA increases expression of the reporter gene in the cytoplasm.

10. The method of claim 8 or 9, wherein the miRNA is selected from the group consisting of hsa-miR-541-3p, hsa-miR-505-5p, hsa-miR-1910-3p, hsa-miR-135a-3p, hsa-miR-362-5p, hsa-miR-9-5p, hsa-miR-103a-2-5p, hsa-miR-887-5p, hsa-miR575, hsa-miR-589-3p, hsa-miR- 378c, hsa-miR-378a-3p, hsa-miR378d, hsa-miR-346, hsa-miR-378h, hsa-miR-557, hsa-miR- 500a-3p, hsa-miR-502-3p, hsa-miR-1271-5p, hsa-miR-5001-3p, hsa-miR-664b-3p, hsa-miR- 101-5p, hsa-miR-204-5p, hsa-miR-764, hsa-miR-370-3p, hsa-miR-1343-3p, hsa-miR-371a- 3p, hsa-miR-4485-5p, hsa-miR-2682-3p, hsa-miR-331-3p, hsa-miR-660-3p, hsa-miR-1306- 5p, hsa-miR-6756-3p, hsa-miR-1908-3p, hsa-miR-6894-3p, hsa-miR-6760-3p, hsa-miR-6726- 3p, hsa-miR-23b-3p, hsa-miR-1178-3p, hsa-miR-149-5p, hsa-miR-23a-3p, hsa-miR-324-3p, hsa-miR-342-3p, hsa-miR-20b-5p, hsa-miR-106a-5p, hsa-miR-525-3p, hsa-miR-106b-5p, hsa- miR-18b-5p, hsa-miR-28-5p, hsa-miR-6775-5p, hsa-miR-764, hsa-miR-3689d, hsa-miR- 3622b-5p, hsa-miR-1294, hsa-miR-92b-5p, hsa-miR-30b-3p, hsa-miR-516b-5p, hsa-miR-299- 3p, hsa-miR-769-3p, and hsa-miR-7114-5p.

11. A method of treating or preventing Alzheimer’s Disease, comprising: providing the synthetic nucleic acid of any one of claims 1-4; incubating the synthetic nucleic acid with a miRNA; determining whether the miRNA increases expression of the reporter gene in the cytoplasm; and administering a specific treatment for Alzheimer’s Disease when the miRNA increases expression of the reporter gene in the cytoplasm.

12. The method of claim 14, wherein the specific treatment for Alzheimer’s Disease comprises a cholinesterase inhibitor and/or a N-methyl-D-aspartate (NMD A) antagonist.

13. A method of treating or preventing Alzheimer’s Disease in a subject in need thereof, comprising: administering to the subject a miRNA selected from the group consisting of hsa-miR- 541-3p, hsa-miR-5O5-5p, hsa-miR-1910-3p, hsa-miR-135a-3p, hsa-miR-362-5p, hsa-miR-9- 5p, hsa-miR-103a-2-5p, hsa-miR-887-5p, hsa-miR575, hsa-miR-589-3p, hsa-miR-378c, hsa- miR-378a-3p, hsa-miR378d, hsa-miR-346, hsa-miR-378h, hsa-miR-557, hsa-miR-500a-3p, hsa-miR-502-3p, hsa-miR-1271-5p, hsa-miR-5001-3p, hsa-miR-664b-3p, hsa-miR-101-5p, hsa-miR-204-5p, hsa-miR-764, hsa-miR-370-3p, hsa-miR-1343-3p, hsa-miR-371a-3p, hsa- miR-4485-5p, hsa-miR-2682-3p, hsa-miR-3 1-3p, hsa-miR-660-3p, hsa-miR-1306-5p, hsa- miR-6756-3p, hsa-miR-1908-3p, hsa-miR-6894-3p, hsa-miR-6760-3p, hsa-miR-6726-3p, hsa- miR-23b-3p, hsa-miR-1178-3p, hsa-miR-149-5p, hsa-miR-23a-3p, hsa-miR-324-3p, hsa-miR- 342-3p, hsa-miR-20b-5p, hsa-miR-106a-5p, hsa-miR-525-3p, hsa-miR-106b-5p, hsa-miR- 18b-5p, hsa-miR-28-5p, hsa-miR-6775-5p, hsa-miR-764, hsa-miR-3689d, hsa-miR-3622b-5p, hsa-miR-1294, hsa-miR-92b-5p, hsa-miR-30b-3p, hsa-miR-516b-5p, hsa-miR-299-3p, hsa- miR-769-3p, and hsa-miR-7114-5p.

14. A synthetic nucleic acid comprising: a reporter gene; a fragment from an Nmnat gene, wherein the fragment from the Nmnat gene comprises: a stem loop comprising a box 1 sequence and a box 2 sequence; wherein the box 1 sequence and the box 2 sequence hybridize to form a stem of the stem loop.

15. The synthetic nucleic acid of claim 14, wherein the box 1 sequence comprises SEQ ID NO: 1.

16. The synthetic nucleic acid of claim 14, wherein the box 2 sequence comprises SEQ ID NO: 2.

17. A method of screening for miRNAs that modulate the alternative splicing of an Nmnat gene, comprising: providing the synthetic nucleic acid of any one of claims 1-4; incubating the synthetic nucleic acid with a miRNA; and determining whether the miRNA modulates the alternative splicing of the Nmnat gene.

18. The method of claim 17 wherein modified expression of the reporter gene provides a neuroprotective outcome.

19. A method of screening for miRNAs that modulate the alternative splicing of an Nmnat gene, comprising: providing the synthetic nucleic acid of any one of claims 1-4; incubating the synthetic nucleic acid with a miRNA; determining whether the miRNA modulates the alternative splicing of the Nmnat gene.

20. The method of claim 19, wherein modified expression of the reporter gene provides a neuroprotective outcome.

21. A method for identifying microRNA mediated obstruction of stem-loop alternative splicing, comprising: identifying a target sequence of interest; identifying a microRNA sequence from a miRNA that can hybridize to the target sequence; wherein the binding of the microRNA sequence to the target sequence obstructs and/or modulates the alternative splicing of a gene comprising the target sequence.

22. The method of claim 21 , wherein the target sequence is an Nmnat gene.

23. The method of claim 21 or 22, wherein the microRNA sequence is miR-iab-8-5p, miR- 137-5p, miR-210-3p, miR-137-3p, miR-274-3p, miR-307a-5p, miR-278-5p, miR-10-3p, miR- 983-3p, miR-278-3p, miR-1002-5p, miR-9b-5p, miR-252-5p, or miR-193-3p.

24. The method of claim 21 , wherein the target sequence is a tau gene.

25. The method of claim 24, wherein the microRNA sequence is hsa-miR-541-3p, hsa- miR-505-5p, hsa-miR-1910-3p, hsa-miR-135a-3p, hsa-miR-362-5p, hsa-miR-9-5p, hsa-miR- 103a-2-5p, hsa-miR-887-5p, hsa-miR-575, or hsa-miR-589-3p.

26. A method for modulating expression of a gene, comprising: identifying a target sequence of interest in the gene; identifying a microRNA sequence from a miRNA that can hybridize to the target sequence of interest in the gene; wherein the binding of the microRNA sequence to the target sequence modulates the expression of the gene.

27. The method of claim 26, wherein the target sequence is an Nmnat gene.

28. The method of claim 27, wherein the microRNA sequence is miR-iab-8-5p, miR-137- 5p, miR-210-3p, miR-137-3p, miR-274-3p, miR-307a-5p, miR-278-5p, miR-10-3p, miR-983- 3p, miR-278-3p, miR-1002-5p, miR-9b-5p, miR-252-5p, or miR-193-3p.

29. The method of claim 26, wherein the target sequence is a tau gene.

30. The method of claim 29, wherein the microRNA sequence is hsa-miR-541-3p, hsa- miR-505-5p, hsa-miR-1910-3p, hsa-miR-135a-3p, hsa-miR-362-5p, hsa-miR-9-5p, hsa-miR- 103a-2-5p, hsa-miR-887-5p, hsa-miR-575, or hsa-miR-589-3p.

31. A method of monitoring and/or treating a disease caused by a splicing defect, comprising: identifying a target sequence near the genetic mutation; identifying a microRNA sequence from a miRNA that can hybridize to the target sequence near the genetic mutation; administering a nucleic acid comprising the microRNA sequence; allowing the nucleic acid to bind to the target sequence near the genetic mutation; and wherein the binding of the nucleic acid to the target sequence modulates the splicing defect to treat the genetic disease.

32. The method of claim 31 , wherein the target sequence is an Nmnat gene.

33. The method of claim 32, wherein the microRNA sequence is miR-iab-8-5p, miR-137- 5p, miR-210-3p, miR-137-3p, miR-274-3p, miR-307a-5p, miR-278-5p, miR-10-3p, miR-983- 3p, miR-278-3p, miR-1002-5p, miR-9b-5p, miR-252-5p, or miR-193-3p.

34. The method of claim 31 , wherein the target sequence is a tau gene.

35. The method of claim 34, wherein the microRNA sequence is hsa-miR-541-3p, hsa- miR-505-5p, hsa-miR-1910-3p, hsa-miR-135a-3p, hsa-miR-362-5p, hsa-miR-9-5p, hsa-miR- 103a-2-5p, hsa-miR-887-5p, hsa-miR-575, or hsa-miR-589-3p.

36. The method of claim 35, wherein the nucleic acid comprises 80% similarity or more to SEQ ID NO: 3.

37. The method of claim 36, wherein the nucleic acid comprises 90% similarity or more to SEQ ID NO: 3.

38. The method of claim 37, wherein the nucleic acid comprises SEQ ID NO: 3.

39. The method of claim 31, wherein the nucleic acid further comprises a fluorescent moiety.

40. The method of claim 39, wherein the fluorescent moiety comprises mCherry or GFP.

41. The method of claim 31 , wherein the disease is a neurodegenerative disease.

42. The method of claim 41, wherein the disease is Alzheimer’s disease.

43. The method of claim 31 , wherein the disease is brain cancer.