Gene expression-based modeling of human cortical synaptic density

Abstract

Postnatal cortical synaptic development is characterized by stages of exuberant growth, pruning, and stabilization during adulthood. How gene expression orchestrates these stages of synaptic development is poorly understood. Here we report that synaptic growth-related gene expression alone does not determine cortical synaptic density changes across the human lifespan, but instead, the dynamics of cortical synaptic density can be accurately simulated by a first-order kinetic model of synaptic growth and elimination that incorporates two separate gene expression patterns. Surprisingly, modeling of cortical synaptic density is optimized when genes related to oligodendrocytes are used to determine synaptic elimination rates. Expression of synaptic growth and oligodendrocyte genes varies regionally, resulting in different predictions of synaptic density among cortical regions that concur with previous regional data in humans. Our analysis suggests that modest rates of synaptic growth persist in adulthood, but that this is counterbalanced by increasing rates of synaptic elimination, resulting in stable synaptic number and ongoing synaptic turnover in the human adult cortex. Our approach provides a promising avenue for exploring how complex interactions among genes may contribute to neurobiological phenomena across the human lifespan.

Fig. 1.

Expression of mitochondrial genes, but not of synaptic growth or elimination genes, follow the course of spine density changes across the lifespan. (A) Cortical expression of each of the 40 eigengenes and spine density data (11) were constant-normalized to a common arbitrary scale. The normalized spine density is plotted in orange, and overlying plots represent each of the eigengenes. The gray lines represent eigengenes that visually do not match the course of spine density changes. The red and blue lines represent the two eigengenes, B16 and B21, respectively, that most closely follow the spine density course. Clusters B16 and B21 are both highly enriched for mitochondrial genes. (B) Individual gene expression (gray lines, smoothed to individual gene expression values) and average gene expression (blue circles and line, smoothed to average expression values) of synaptic growth-related genes (HOMER1, MAP1A, CAMK2A, SYP, SYN1, and BAIAP2) across all cortical regions do not mirror changes in synaptic or spine density after birth (orange line). (C) Individual expression (gray lines, smoothed to individual gene expression values) of synaptic elimination-related genes (MEF2A, MEF2B, MEF2D, C1q, PARK2, FMR1, and PCDH10) are very similar to the expression of synaptic growth genes (blue line, smoothed to average expression values), with the notable exceptions of FMR1 and MEF2B, which demonstrate a somewhat delayed onset of expression.

Fig. 2.

Potential models of gene expression regulating synaptic density. (A) The “classic” model of synaptic development describes an initial period of exuberant growth soon after birth (blue line) followed by a period of pruning (orange line). Theoretically, it is possible that each period is defined by the isolated expression of a set of genes. Examination of the eigengenes in Fig. S2 show that such isolated patterns of gene expression are not present, at least among the 40 modes of expression identified by WGCNA. (B) Alternatively, growth may occur during a period of unopposed expression of growth-related genes (blue line), followed by delayed onset of synaptic elimination-related genes (orange), which balance the former to produce a period of pruning and then stabilization. (C) A final model suggests that although growth and elimination genes (blue and orange lines, respectively) are expressed concurrently, a third factor (red dashed line) can alter the balance by potentiating elimination, inhibiting growth, or both.

Fig. 3.

Lifespan dynamics of synaptic density can be accurately modeled by incorporating cluster B9 or B35 eigengenes to determine synaptic elimination rate. A model was constructed that relates synaptic density to zero-order growth rates established by eigengene B1 and first-order elimination rates established by a second unknown cluster. Each of the 40 gene clusters was then tested in the model as the unknown cluster, and the cumulative residual sum of squares error between predicted and actual synaptic density data were calculated. This shows that clusters B9 and B35 (green circles, indicated by the arrow) minimize the residual sum of squares. Similar results are seen when using the average expression of the six synaptic growth genes (Fig. S2).

Fig. 4.

Modeling synaptic density with gene expression patterns predicts differences among association areas, primary sensorimotor regions, and limbic system of the cortex. (A) In neocortical association areas, which include prefrontal, medial frontal, and parietal regions of the cortex, eigengene expression of synaptic growth gene cluster B1 (Left, blue circles and line) and the oligodendrocyte-related cluster B9 (Left, red circles and line) accurately predict spine density changes in the prefrontal cortex of humans (Right, blue line prediction); red circles represent actual spine density measurements, replotted from Petanjek et al. (11). (B) Similar analysis in primary sensorimotor cortical regions reveals a close fit to actual synaptic density changes in the auditory cortex, replotted from Huttenlocher and Dabholkar (9). (C) Similar analysis in limbic regions, including the hippocampus and amygdala, shows marked differences in the eigengene expression related to synaptic growth (Left, blue) and oligodendrocytes (Left, red) compared with other cortical regions in A and B. The predicted synaptic density (Right, orange line) in the limbic system is also different from other cortical regions in that it does not demonstrate a distinct period of pruning during adolescence. Comparative data from human hippocampus or amygdala are currently not available.