Intrinsically determined cell death of developing cortical interneurons

Abstract

Cortical inhibitory circuits are formed by γ-aminobutyric acid (GABA)-secreting interneurons, a cell population that originates far from the cerebral cortex in the embryonic ventral forebrain. Given their distant developmental origins, it is intriguing how the number of cortical interneurons is ultimately determined. One possibility, suggested by the neurotrophic hypothesis, is that cortical interneurons are overproduced, and then after their migration into cortex the excess interneurons are eliminated through a competition for extrinsically derived trophic signals. Here we characterize the developmental cell death of mouse cortical interneurons in vivo, in vitro and after transplantation. We found that 40% of developing cortical interneurons were eliminated through Bax (Bcl-2-associated X)-dependent apoptosis during postnatal life. When cultured in vitro or transplanted into the cortex, interneuron precursors died at a cellular age similar to that at which endogenous interneurons died during normal development. Over transplant sizes that varied 200-fold, a constant fraction of the transplanted population underwent cell death. The death of transplanted neurons was not affected by the cell-autonomous disruption of TrkB (tropomyosin kinase receptor B), the main neurotrophin receptor expressed by neurons of the central nervous system. Transplantation expanded the cortical interneuron population by up to 35%, but the frequency of inhibitory synaptic events did not scale with the number of transplanted interneurons. Taken together, our findings indicate that interneuron cell death is determined intrinsically, either cell-autonomously or through a population-autonomous competition for survival signals derived from other interneurons.

Figure 1. Bax-dependent programmed cell death eliminates 40% of developing interneurons from the postnatal mouse neocortex

(a) Cleaved caspase-3 expression (red) observed in GAD67-GFP neurons (green; arrowheads) and other cell types (arrow) of the P7 neocortex. Scale bar, 100 μm (left) and 50 μm (right). (b) Temporal profile of cleaved caspase-3 expression in the neocortex of GAD67-GFP mice. Cleaved caspase-3 expression is highest at P7, and declines to nearly undetectable levels by P15 (ANOVA, F = 84.00 and P < 0.0001; n = 3 per timepoint). (c) Temporal profile of the neocortical GAD67-GFP population size. Between P5 and P20, the neocortical GAD67-GFP population decreases by approximately 40% (ANOVA, F = 32.10 and P < 0.0001; n = 5 per timepoint). (d) The Bax mutation disrupts the developmental cell death of cortical interneurons. Bax^−/− mice exhibit a 99.8% reduction in the number of cells double labeled by cleaved caspase-3 and GAD67-GFP, as compared to Bax^+/+;GAD67-GFP littermates (Student’s t-test, ** P < 0.01; n = 3 per genotype). (e) The neocortical GAD67-GFP population does not decrease in Bax^−/− mice (ANOVA, F = 2.28 and P = 0.18). At P120, the neocortical GAD67-GFP population is approximately 33% smaller in wild type mice (Student’s t-test, ** P < 0.01; n = 3 per genotype at each timepoint). In all figures, error bars represent the standard error of the mean.

Figure 2. In vitro, and following heterochronic transplantation, interneuron precursors undergo programmed cell death during a period defined by their intrinsic cellular age

(a) Primary feeder layers prepared from P0-P2 neocortex. At 14 days in vitro (DIV), the feeder layer contains neurons (Tuj-1, green), astrocytes (GFAP, red), and oligodendrocytes (Olig-2, white). All cells are labeled by DAPI (blue). Scale bar, 50 μm. (b) At 14 DIV, double-labeled cells expressing cleaved caspase-3 (red) and GAD67-GFP (green; arrowheads) are observed along with cells singly labeled by cleaved caspase-3 (arrows). Scale bars, 200 μm. (c) Temporal profile of cleaved caspase-3 expression in GAD67-GFP neuronal cultures. Cleaved caspase-3 expression is highest at 13 DIV (ANOVA, F = 9.12 and P < 0.0001). (d) Temporal profile of the GAD67-GFP population size in vitro. The GAD67-GFP population increases in number between 4 and 9 DIV, likely due to cell proliferation (see Methods), reaches a maximum size around 9 to 11 DIV, and then declines approximately 30% before reaching a stable size around 17 to 22 DIV (ANOVA, F = 4.53 and P < 0.01). (e) A transplanted interneuron precursor expressing cleaved caspase-3 (red) and beta-actin:GFP (green; arrowhead) at 15 DAT. Scale bars, 50 μm (left) and 25 μm (right). (f) Temporal profile of cleaved caspase-3 expression in transplanted interneuron precursors. Cleaved caspase-3 is highest at 15 DAT, when the transplanted population reaches an intrinsic cellular age similar to that of endogenous interneurons during the peak of normal developmental cell death (Figure 1b; ANOVA, F = 17.79 and P < 0.0001; n = 5 per timepoint). In all figures, error bars represent the standard error of the mean.

Figure 3. Transplanted interneuron cell death is not governed by competition for survival signals derived from other cell types in the recipient neocortex

(a) Over a broad range of transplant sizes (from 5 × 10³ to 1000 × 10³ cells) nearly constant fractions of the transplanted populations survive at 60 DAT (approximately 15 to 22%; ANOVA, F = 0.34 and P = 0.12; n = 6, 7, 3, 3 per transplant size, respectively). When the initial transplant size is increased to 2 × 10⁶ cells, a smaller fraction of transplanted cells survives in the recipient neocortex (approximately 8%; n = 3). (b) Equal numbers of transplanted TrkB^−/−;beta-actin:GFP interneurons and TrkB^+/+;beta-actin:GFP interneurons survive in the recipient neocortex at 60 DAT (Student’s t-test, P = 0.75; n = 5 per genotype). (c) Transplanted cortical interneuron cell death occurs through a Bax-dependent mechanism. Greater numbers of Bax^−/−;beta-actin:GFP cortical interneurons survive in the recipient cortex at 60 DAT, compared to transplanted Bax^+/+;beta-actin:GFP and Bax^+/−;beta-actin:GFP cortical interneurons (Student’s t-test, * P < 0.05; n = 5 wild-type and Bax^+/−; n = 6 Bax^−/−). (d) Transplanted beta-actin:DsRed interneurons (red) and endogenous GAD67-GFP neurons (green) at 60 DAT (initial transplant size = 10⁶ cells; Scale bar = 150 μm). (e) At 60 DAT, transplanted DsRed-labeled interneurons increase the cortical interneuron population size by 34% (Red) without affecting the endogenous GAD67-GFP population (green; Student’s t-test, P = 0.28; n = 3). In all figures, error bars represent the standard error of the mean.

Figure 4. Interneuron population size is not a primary determinant of the level of functional cortical inhibition

(a) Representative traces of spontaneous inhibitory postsynaptic currents (sIPSCs) recorded from endogenous neocortical pyramidal neurons in vitro (media vehicle (Con), top; interneuron transplant recipient (Int), bottom). Vertical scale bar, 40 pA; horizontal scale bar, 200 ms. (b) Transplanted interneurons increase the frequency (top), but not the amplitude (bottom) of sIPSCs recorded at 30 to 40 DAT (Wilcoxon rank-sum test, * P < 0.05 and P = 0.22, respectively; n = 23 recorded cells from control animals, n = 37 recorded cells from interneuron transplant recipients). Mean transplanted cell density for transplant recipient group = 23.3 ± 3.8 cells/mm². Error bars represent the standard error of the mean. (c) The frequency of sIPSCs onto host pyramidal neurons does not increase with the density of transplanted interneurons (linear regression analysis, slope = 0.0003, r² = 0.0003).