Fragile X protein controls neural stem cell proliferation in the Drosophila brain

Abstract

Fragile X syndrome (FXS) is the most common form of inherited mental retardation and is caused by the loss of function for Fragile X protein (FMRP), an RNA-binding protein thought to regulate synaptic plasticity by controlling the localization and translation of specific mRNAs. We have recently shown that FMRP is required to control the proliferation of the germline in Drosophila. To determine whether FMRP is also required for proliferation during brain development, we examined the distribution of cell cycle markers in dFmr1 brains compared with wild-type throughout larval development. Our results indicate that the loss of dFmr1 leads to a significant increase in the number of mitotic neuroblasts (NB) and BrdU incorporation in the brain, consistent with the notion that FMRP controls proliferation during neurogenesis. Developmental studies suggest that FMRP also inhibits neuroblast exit from quiescence in early larval brains, as indicated by misexpression of Cyclin E. Live imaging experiments indicate that by the third instar larval stage, the length of the cell cycle is unaffected, although more cells are found in S and G2/M in dFmr1 brains compared with wild-type. To determine the role of FMRP in neuroblast division and differentiation, we used Mosaic Analysis with a Repressible Marker (MARCM) approaches in the developing larval brain and found that single dFmr1 NB generate significantly more neurons than controls. Our results demonstrate that FMRP is required during brain development to control the exit from quiescence and proliferative capacity of NB as well as neuron production, which may provide insights into the autistic component of FXS.

Figure 1.

Loss of dFmr1 alters cell cycle progression in the larval brain. (A–F) Staining for the M-phase marker PH3 in wild-type (A) and dFmr1³ mutant (B) larval brains. Dashed lines delineate the CB areas. The stem cell-specific protein Miranda marks NBs (C–F). Higher magnification indicates that the dFmr1 mutant brains contain more PH3 positive neuroblasts (F, arrowheads) compared with wild-type (E, arrowheads). (K) Total number of mitotic neuroblasts (NBs) are significantly increased in both dFmr1³ homozygotes (P = 0.04) and dFmr1³/50M (P < 0.001) brains compared with the genetic rescue, P[dFmr1];dFmr1³. There is no statistical significance (n.s.) between the w¹¹¹⁸ and genomic rescue controls. Student's t-test was used to calculate statistical significance. (G–J) Staining for the S-phase marker BrdU in wild-type (G and I) and dFmr1 mutant (H and J) larval brains. The cortical protein Lgl was used to delineate the CB (see also dashed areas) region, where the neuroblasts and associated cells display higher levels of BrdU incorporation in dFmr1 mutants (H and J) compared with wild-type (G and I). Stainings and genotypes as indicated. Brains were dissected from age matched late third instar larvae. Scale bar in (A) 120 µm, in (E) 30 µm. All images are projections of three confocal slices, slice size = 2 µm.

Figure 2.

FMRP regulates neural stem cell proliferation in the larval brain. (A–H) Single CB neuroblast clones (marked with GFP) in late third instar larval brains. Control clones [wild-type (A, B, E, F)] have fewer cells than dFmr1 mutant clones (C, D, G, H) after induction at first instar stage (4–10 h ALH). The type I neuroblast-specific marker, Asense, was used to differentiate type I (A–D) from type II (E–H) clones. Individual clone sizes were quantified by counting cells within each clone, grouped by genotype for type I (M) and type II (N) clones. Statistical significance (see text for P-values) was calculated using Student's t-test. Error bars indicate standard error of the mean. (I–L) Late clone induction (48–54 h ALH) results in single cell clones in both wild-type (I, J) and mutant (K, L) backgrounds in the ventral brain. Stainings, genotypes and clone induction regimes as indicated. Scale bar in (A) 25 µm, (I) 60 µm. Panels shown represent projections of two to three individual confocal slices, as needed, to include the entire clone.

Figure 3.

Loss of dFmr1 causes an increase in the number of Elav positive neurons with no effect on Prospero expression. (A, B, G, H) Low magnification images of larval brains containing control (A and B) and dFmr1 (G and H) clones (see arrowheads). (C–F) High magnification views of representative clones show that Elav is expressed in all cells within the clone but the neuroblast (NB) and a few surrounding cells, which are presumably GMCs (see arrows). Quantification of the Elav positive cells/per clone results in a significant increase in both dFmr1³ and dFmr1^50M clones compared with control (M). Student's t-test was used to calculate statistical significance. (G–L) Prospero stainings in control and dFmr1 clones show that a comparable number of cells surrounding the NB exhibit cytoplasmic Prospero corresponding to GMCs (arrows) with the remainder showing nuclear localization characteristic to differentiated neurons. Genotypes and stainings as indicated. Scale bar in (A) 100 µm. Panels shown represent projections of two to three individual confocal slices.

Figure 4.

Increased neurons in mutant clones survive to adulthood. (A and B) Low magnification images of adult brains, clones induced at 4–10 h ALH. (C and D) High magnification images indicate GFP positive cells also stain positive for neuronal marker Elav. (E) Quantification of adult clone numbers. Increased number of cells per clone is statistically significant in dFmr1³ mutants compared with control clones. Scale bar in (A) 275 µm, (C) 35 µm. (A) and (B) are projections of the entire brain. (C) and (D) are projections of two to three individual confocal slices to capture entire clone.

Figure 5.

Live imaging of dividing neuroblasts indicates no change in the duration of the cell cycle in third instar dFmr1 mutant brains compared to wild-type. (A–D) Miranda-GFP (shown in green) and Jupiter-mCherry (shown in red) label neuroblasts, GMC daughters and mitotic spindles, respectively. Several mitoses were imaged live in cultured explanted brains. Two mitoses are shown at the time of GMC pinching off (arrows). Genotypes and division times as indicated. Note: dFmr1 neuroblasts shown were located deep within the brain, hence the lower quality image. (E) Quantification of cell cycle time shows no significant difference between wild-type (1.48 ± 0.11 h) and mutant brains (1.70 ± 0.26 h). Imaging was performed using early third instar brains. Student's t-test was used to calculate statistical significance (see text for P-value). Scale bar in (A) 10 µm.

Figure 6.

Loss of dFmr1 leads to premature exit from quiescence in the developing larval brain. Wild-type (w¹¹¹⁸) and dFmr1^3/50M larval brains analyzed at different time points ALH, as shown. Brains were immunostained for Miranda and CycE (A–N′). DAPI was used to label nuclear DNA. (A–B′) At 0–6 h ALH, both wild-type (A, B, see arrows) and dFmr1^3/50M mutant brains (A′, B′, see arrows) display large, Miranda/CycE positive mb neuroblasts (NBs) only. (C) Quantification of Miranda/CycE double positive cells shows no significant change between genotypes at this timepoint. (D–E′) At 6–12 h ALH, wild-type brains contain four large mb NBs (D, D′, arrows). Note: only three out of the four mb NBs clearly visible in the single confocal slice shown. Mutant brains show an increased number of smaller, Miranda/CycE positive NBs (E, E′, arrows) compared with controls. (F) The total number of Miranda/CycE positive NBs at this time point is significantly increased in the dFmr1^3/50M mutant brains. (G–H′) At 12–18 h ALH, the number of Miranda/CycE positive cells continues to be significantly increased in dFmr1^3/50M mutant brains (H, H′), quantified in (I). (J–K′) At 18–24 h ALH, the number of Miranda/CycE positive cells remains significantly increased in dFmr1^3/50M mutant brains (K, K′), quantified in (L). (M–N′) At 24–30 h ALH, the number of Miranda/CycE positive cells is now significantly increased in w¹¹¹⁸ brains (M, M′), compared with the mutant brains (N, N′), quantified in (O). (P) Line plot of average CycE/Miranda positive neuroblasts at each 6 h interval ALH. Student's t-test was used to calculate statistical significance (see text for P-values). Scale bars: (A) 30 µm, (D) 30 µm, (G) 30 µm, (J) 40 µm, (M) 40 µm. Panels (A–B′) represent projections of two to three individual confocal slices to capture all mb neuroblasts. All remaining images represent single confocal slices (2 µm thick).

Figure 7.

Proposed model for FMRP function in the developing brain. FMRP controls the dynamic proliferative activity of neuroblasts during development. See text for details.