This article is originally compiled by brain science world team of brainnews, please contact permission to reprint.
For more than 20 years, neuroimaging has become a powerful tool for understanding the macroscopic characteristics of the brain.
The study of MRI in neurological and psychiatric disorders has given us an understanding of the regional changes in the structure and function of the brain that occur in these diseases.
However, most MRI-based techniques are unable to reveal specific changes at the molecular and cellular levels, so we cannot know the biological mechanisms of brain diseases through neuroimaging, such as regional variation characteristics of neuroimaging markers, which biological pathways are involved.
The use of whole-brain gene expression maps, such as the Allen Human Brain Atlas (AHBA), allows us to begin studying the correlation between transcriptomics and neuroimaging.
Imaging transcriptomics is the identification of which biological pathways characterize neuroimaging markers by analyzing the spatial correlation between gene expression characteristics revealed by transcriptomics and the structural or functional features of the brain revealed by neuroimaging. However, the effectiveness of imaging transcriptomics has yet to be studied.
Researchers at King's College London and the University of Padua in Italy collaborated on a paper in Cell Reports for imaging transcriptomics analysis using imaging data from a series of PET tracers in healthy human brains.
The results of this study show that imaging transcriptomics cannot clearly reveal the transcriptome characteristics corresponding to the neuroinflammation axis-related tracers in healthy human brains, but in neurons, astrocytes, and oligodendrocyte-associated tracers, the method successfully identified the corresponding transcriptome characteristics.
Imaging transcriptomics may help to understand the biological pathways corresponding to multiple neuroimaging phenotypes.
As shown in Figure 1, the researchers plotted the gene expression data and neuroimaging data in AHBA into the same brain partition map, used multivariate statistical analysis methods to identify thousands of genes that were related to the neuroimaging measurements in each brain partition, and then sorted and enriched according to the size of the correlation, and finally obtained the cells or biological pathways corresponding to the genes that were significantly enriched. Thus, neuroimaging data can be thought to reflect changes in these biological pathways.
Figure 1: Imaging transcriptomics analysis steps
The researchers chose three classes of PET tracers that cover a variety of brain biological functions: receptors, synaptic proteins, and metabolism; astrocytes and myelin sheaths; transposed proteins and cyclooxygenases.
Based on the target of each tracer, the researchers pre-empted a conjecture of the corresponding biological or cellular pathways. If imaging transcriptomics is effective, then the enrichment pathways derived from the above analysis steps should be consistent with conjecture.
Table 1: PET tracers used and corresponding conjectures
Receptors, synaptic proteins, and metabolically associated tracers
[11C] Flumazenil targeting GABA-A receptors is significantly enriched into excitatory and inhibitory neurons. Although there is literature suggesting that astrocytes also express GABA-A receptors, the results of the study did not significantly enrich any non-neuronal cells.
[18F] GE179 targeting NMDA receptors is significantly enriched into excitatory, inhibitory neurons and oligodendrocytes, consistent with the literature-reported expression of NDMA receptors. [11C]UCB-J targeting SV2A is also enriched into neurons consistently with predictions.
[18F]FDG, which reflects glucose uptake, was significantly enriched to neurons and endothelial cells, but not to astrocytes.
Figure 2: Imaging results and cell type enrichment analysis of receptors, synaptic proteins, and metabolic-related tracers
Astrocyte and myelin-related tracers
Both tracers [11C]BU99008 and L-[11C]deprenyl-D2 are highly enriched into astrocytes and these cell-related biological pathways, such as RNA and ribonucleoprotein metabolism. In addition, some neurons were enriched, which is also consistent with the data reported in the literature.
Mt and myelin WCs, two markers that target oligodendrocytes, are also effectively enriched into oligodendrocytes and oligodendrocytes, but also to other non-neuronal cells, which is also consistent with literature skepticism about whether MT and WC are specific for oligodendrocytes.
Figure 3: Imaging results and cell type enrichment analysis of astrocytes and myelin-related tracers
Transposed protein (TSPO) and cyclooxygenase (Cox)-associated tracers
The enrichment results of four tracers targeting TSPO and one tracer targeting Cox-1 were inconsistent with the predictions. With the exception of [11C]PK11195 and [18F]DPA174, none of the other three tracers were enriched to microglia and astrocytes. Instead, these tracers are highly enriched into neurons.
Confusingly, in addition to [11C]PK11195, imaging measurements of tracers targeting TSPO were not correlated with the spatial distribution of TSPO messenger RNA expression.
Figure 4: Imaging results and cell type enrichment analysis of transposed proteins and cyclooxygenase-associated tracers
Summary
In summary, the study verifies that imaging transcriptomics analysis can successfully reveal the biology and cell pathways characterized by neuroimaging data related to neurons, astrocytes, and oligodendrocytes, while the imaging data related to the neuroinflammation axis of healthy human brains cannot be reliably analyzed by this method for biological pathways.
Original link:
https://doi.org/10.1016/j.celrep.2021.110173
bibliography
Martins D, Giacomel A, Williams SCR, Turkheimer F, Dipasquale O, Veronese M; PET Templates Working Group. Imaging transcriptomics: Convergent cellular, transcriptomic, and molecular neuroimaging signatures in the healthy adult human brain. Cell Rep. 2021 Dec 28;37(13):110173. doi: 10.1016/j.celrep.2021.110173. PMID: 34965413.
Compiled by Hong Chaoli (brainnews creative team)
Reviewer: Simon (Brainnews Editorial Board)