Culling Connection From Chaos, Alzheimer’s Genetic Network Study Pins PLXNB1 and INPPL1
Apparently, it takes a village to cause late-onset Alzheimer’s disease. A village of genes, that is. Drawing on postmortem gene-expression data from the brains of more than 500 people whose cognition was closely tracked during life, researchers led by Philip De Jager at Columbia University in New York and David Bennett of Rush University Medical Center in Chicago exposed cliques of co-expressed genes that correlated with cognitive decline and/or plaques and tangles in the brain. Diving deeper into one of these groups, the researchers pulled out two genes—INPPL1 and PLXNB1—that boosted the release of Aβ in cell culture.
- Pulling clinical, neuropathological, and gene-expression data from 500-plus participants, scientists found modules of co-expressed genes that correlated with disease traits.
- A person’s rate of cognitive decline associated with his or her gene-expression patterns more than with any other trait.
- Genetic module 109 tracked with cognition, and two of its genes boosted Aβ release from astrocytes.
The study used an unbiased network approach. Its beauty is that it can identify whole biological pathways ripe for therapeutic targeting, not just individual genes, commented Eliezer Masliah of the National Institute on Aging in Bethesda, Maryland. Masliah was not directly involved in the work, but the study is part of the Accelerating Medicines Partnership (AMP-AD) at the National Institutes of Health, a collaborative search for new therapeutic targets among government, industry, and nonprofit organizations.
The complexity of both the human genome and neurodegenerative disease challenges researchers hunting for therapeutic targets, but it also offers opportunities. Genome-wide association studies are one way to sift through the chaos, but exactly how susceptibility loci identified in GWAS influence disease risk is difficult to ascertain (Oct 2013 news; Apr 2018 news). Most reside in noncoding regions of the genome, suggesting they play a role in regulating gene expression. One way to get at this is to measure gene expression directly, by sequencing RNA isolated from brain tissue. Using this approach, previous studies have correlated groups of co-expressed genes with AD (Zhang et al., 2013; Alzforum Research Timeline; Miller et al., 2013). However, scant data exist to pick out genes that correlate directly with cognitive decline or other measurable AD endophenotypes (Boyle et al., 2013; Jack et al., 2014).
The authors of this study have a data set that brings these factors together. The Religious Orders Study (ROS) and Rush Memory and Aging Project (MAP)—collectively known as ROSMAP—are prospective clinical and pathological studies of aging and dementia. Participants enter the studies free of dementia, are tracked for cognitive changes during life, and donate their brains to research after they die. This means the scientists can calculate each participant’s slope of cognitive decline, thus far spanning up to 20 years, and then analyze the postmortem brain samples for gene expression and neuropathology.
For the current study, co-first authors Sara Mostafavi of the University of British Columbia in Vancouver, Canada, and Chris Gaiteri of Rush University used data from 478 participants, whose average age at death was 88.7. About a third passed away with their cognition still normal, another third had mild cognitive impairment, and about 40 percent had been clinically diagnosed with AD dementia. The brains of nearly 60 percent of participants harbored AD pathology, though only half of those pathological cases of AD had memory loss at death, in line with earlier findings of this cohort.
The researchers sequenced RNA from the dorsolateral prefrontal cortex of each participant. This highly connected network hub succumbs to AD pathology later in the course of disease, making gene-expression changes there more likely to represent changes that take place prior to the onslaught of neurodegeneration, De Jager told Alzforum. They correlated gene-expression patterns with five phenotypic traits: clinical diagnosis, rate of cognitive decline, amyloid plaque burden, tau tangle density, and neuropathological AD diagnosis. Of these, cognitive decline correlated with expression of the most genes, more than 3,000, followed by Aβ, clinical AD, tau tangles, and pathological AD.
To uncover cellular processes related to these phenotypes, the researchers grouped co-expressed genes into modules. They identified 47 modules, consisting of 20 to 556 genes each. The genes within each module had diverse functions, although a majority of the modules were enriched for at least one functional category. Thirteen were enriched for cell-type-specific genes, including those expressed in microglia, astrocytes, and different neuronal subtypes. Finally, the researchers employed Bayesian statistics to find functional connections between the modules, and to select genes and pathways likely to lie upstream of disease traits.
Expression of 11 modules correlated with at least one disease phenotype. Modules that correlated positively with disease phenotypes included those containing genes involved in immunity, mitochondria, cell cycle, and transcriptional regulation. On the flip side, expression of modules chock-full of genes involved in neuronal or synaptic function correlated negatively with these traits. The researchers confirmed the modules’ links to disease in a separate group of 82 ROSMAP participants.
Of all the disease-associated modules, one called 109 jibed best with cognitive decline, plus to a lesser degree with Aβ burden and clinical AD diagnosis. M109 comprises 390 genes of different functions, predominantly cell-cycle control and chromatin modification. This link appeared largely independent of ApoE genotype, although ApoE4 was modestly associated with higher expression of this module. The researchers found no relationship between 21 other susceptibility loci identified in GWAS and expression of genes in m109. Another module, m116, contained a bounty of microglial genes that resided near GWAS hits; however, m116 associated with none of the defined phenotypes.
Trying to validate that their approach is relevant, the researchers dug deeper into m109. Its link to cognitive decline appeared to go through both Aβ-dependent and independent mechanisms. Lacking a cellular model for cognitive decline, the researchers decided to test whether individual genes in m109 influenced release of Aβ, either from cultured astrocytes or from neurons derived from induced pluripotent stem cells (iPSCs). The scientists sifted through m109’s nearly 400 member genes in search of those that connected most strongly with other genes in the module, were expressed in neurons and/or astrocytes, and associated individually with the disease phenotypes.
They identified 21 genes for experimentation; against these, they were able to deploy shRNA to knock down 14 in astrocytes and 11 in neurons, 11 of which overlapped. They found two genes—INPPL1 and PLXNB1—whose expression associated with Aβ levels in the culture medium of astrocytes, but not neurons. Knocking down either reduced Aβ levels by 25 percent. Why not in neurons? De Jager told Alzforum that the shRNA constructs did a better job knocking down their target gene in astrocytes, so this could explain the difference. Returning to the postmortem samples, the researchers detected INPPL1 and PLXNB1 expression in both neurons and astrocytes, and they saw astrocytes expressing both genes mingling near extracellular Aβ plaques.
How much did INPPL1 and PLXNB1 affect the overall Aβ burden? Not much. Their expression accounted for 2.8 and 3.1 percent of the variance separately, and 3.1 percent when combined, suggesting a redundant influence. However, this compares to less than 1 percent variance in Aβ explained by any AD risk variant besides ApoE. INPPL1 and PLXNB1 expression influenced cognitive decline somewhat more, explaining 5.4 percent of the variance when combined.
What do these genes do? PLXNB1 belongs to a protein family that mediates semaphorin signaling, which in turn influences neurite outgrowth and synaptic plasticity (Perälä et al., 2012). No links to AD or cognitive decline are known.
INPPL1, on the other hand, has a rap sheet. Also known as SHIP2, the protein is a phosphatase that regulates levels of PI(3,4,5)P3, a regulator of multiple signaling pathways. Polymorphisms in INPPL1 associate with Type 2 diabetes, overexpressing INPPL1 disrupts insulin signaling in mice, and treatment with an INPPL1 inhibitor relieves memory problems in diabetic mice (Suwa et al., 2010; Soeda et al., 2010).
Dimitrios Kapogiannis of the NIA commented that INPPL1’s connection to cognitive decline and AD pathology aligns with the known link between insulin resistance and neurodegenerative diseases. Kapogiannis hypothesized that insulin resistance renders the brain more vulnerable to neurodegenerative pathologies, including Aβ and tau, and perhaps to cognitive decline. Studies indicating that disruption in insulin signaling can cause both Aβ and tau accumulation, which in turn aggravate insulin resistance, suggest a feed-forward loop, he added. More validation is needed to sort out whether INPPL1 represents a therapeutic target, Kapogiannis said, but the findings broadly support the idea that improving insulin signaling could benefit neurodegenerative diseases.
De Jager emphasized that the real promise of the work lies in discovering biological pathways that influence disease. The unbiased approach allows the scientists to focus on targets other than Aβ and tau; along those lines, they are developing more complex cell culture models to explore the role of other genes in neuronal survival and synaptic function. The modules are publicly available for researchers to scour for ideas.
“What is important is not the individual genes, but that each of them connects to a wider network,” Masliah concurred. The goal is to find therapeutic points of entry into these networks. To that end, data from this study, as well as from other AMP-AD studies and input from industry, are being gathered together to come up with 100 potential therapeutic targets that AMP-AD consortium members will present at a workshop preceding the AAIC conference in July, Masliah said.—Jessica Shugart
- Paper Alert: New Alzheimer’s Genes Published
- GWAS, GWAX: bioRχiv Hosts Bonanza of Alzheimer’s Genetics
- Zhang B, Gaiteri C, Bodea LG, Wang Z, McElwee J, Podtelezhnikov AA, Zhang C, Xie T, Tran L, Dobrin R, Fluder E, Clurman B, Melquist S, Narayanan M, Suver C, Shah H, Mahajan M, Gillis T, Mysore J, MacDonald ME, Lamb JR, Bennett DA, Molony C, Stone DJ, Gudnason V, Myers AJ, Schadt EE, Neumann H, Zhu J, Emilsson V. Integrated systems approach identifies genetic nodes and networks in late-onset Alzheimer's disease. Cell. 2013 Apr 25;153(3):707-20. PubMed.
- Miller JA, Woltjer RL, Goodenbour JM, Horvath S, Geschwind DH. Genes and pathways underlying regional and cell type changes in Alzheimer's disease. Genome Med. 2013 May 25;5(5):48. PubMed.
- Boyle PA, Wilson RS, Yu L, Barr AM, Honer WG, Schneider JA, Bennett DA. Much of late life cognitive decline is not due to common neurodegenerative pathologies. Ann Neurol. 2013 Jun 24; PubMed.
- Jack CR Jr, Wiste HJ, Weigand SD, Rocca WA, Knopman DS, Mielke MM, Lowe VJ, Senjem ML, Gunter JL, Preboske GM, Pankratz VS, Vemuri P, Petersen RC. Age-specific population frequencies of cerebral β-amyloidosis and neurodegeneration among people with normal cognitive function aged 50-89 years: a cross-sectional study. Lancet Neurol. 2014 Oct;13(10):997-1005. Epub 2014 Sep 4 PubMed.
- Perälä N, Sariola H, Immonen T. More than nervous: the emerging roles of plexins. Differentiation. 2012 Jan;83(1):77-91. Epub 2011 Oct 14 PubMed.
- Suwa A, Kurama T, Shimokawa T. SHIP2 and its involvement in various diseases. Expert Opin Ther Targets. 2010 Jul;14(7):727-37. PubMed.
- Soeda Y, Tsuneki H, Muranaka H, Mori N, Hosoh S, Ichihara Y, Kagawa S, Wang X, Toyooka N, Takamura Y, Uwano T, Nishijo H, Wada T, Sasaoka T. The inositol phosphatase SHIP2 negatively regulates insulin/IGF-I actions implicated in neuroprotection and memory function in mouse brain. Mol Endocrinol. 2010 Oct;24(10):1965-77. PubMed.
- Gaiteri C, Ding Y, French B, Tseng GC, Sibille E. Beyond modules and hubs: the potential of gene coexpression networks for investigating molecular mechanisms of complex brain disorders. Genes Brain Behav. 2014 Jan;13(1):13-24. Epub 2013 Dec 10 PubMed.
- Mostafavi S, Gaiteri C, Sullivan SE, White CC, Tasaki S, Xu J, Taga M, Klein HU, Patrick E, Komashko V, McCabe C, Smith R, Bradshaw EM, Root DE, Regev A, Yu L, Chibnik LB, Schneider JA, Young-Pearse TL, Bennett DA, De Jager PL. A molecular network of the aging human brain provides insights into the pathology and cognitive decline of Alzheimer's disease. Nat Neurosci. 2018 Jun;21(6):811-819. Epub 2018 May 25 PubMed.
To make an annotation you must Login or Register.
No Available Comments
Make a Comment
To make a comment you must login or register.