LAUSR

GABAergic interneurons contribute to cortical function by regulating the balance of activity. Malfunction of inhibitory interneurons has been associated with neuropsychiatric disorders (Hashimoto et al. 2003; Marín 2012). The most striking feature of GABAergic interneurons is their diversity of cell types. Although fate-mapping studies have demonstrated the presence of spatially and temporally distinct progenitor domains for interneuron subclasses within the embryonic ventral telencephalon (Butt et al. 2005; Inan et al. 2012), the mechanisms of interneuron subtype-determination during development remain poorly understood. Considerable evidence supports the idea that cell fate determination throughout neurogenesis is intimately linked with cell-cycle length (Ohnuma and Harris 2003; Pilaz et al. 2016). In fact, disruption of normal cell-cycle length is known to alter neurogenesis (Hardwick et al. 2015; Boyd et al. 2015). However, due to technical limitations, we still lack detailed information about the specific impact of cell-cycle length on cell fate determination in vivo. An important unanswered question is whether there is a direct relationship between cell-cycle length of neural progenitors and interneuron fate determination. In this study, we developed a triple thymidine analog labeling method to label progenitors undergoing short cycles (PSC) or progenitors undergoing long cycles (PLC) in the medial ganglionic eminence (MGE) in vivo (Fig. 1A–C). In brief, we carried out an initial injection of IdU followed 2 h later by an injection of BrdU at embryonic day 13.5 (E13.5). Cells labeled by IdU, but not BrdU (IdUBrdU) had left S phase. After various intervals (4, 9, 12 and 14 h, defined as ΔT), we injected EdU to label cells that re-entered S phase (IdUBrdUEdU) (Fig. 1A). Using this labeling method, the cell-cycle lengths (TC) of IdU BrdUEdU cells can be roughly calculated for different ΔTs (Fig. 1B). We found that the percentage of IdUBrdUEdU cells in IdUBrdU cells in ΔT = 9 h group was around three times higher than in the other ΔT groups at E13.5 (Figs. 1D and S1A–H; Table S1); whereas, the peak of ΔT at E15.5 was 12 h (Figs. 1E and S1I–P; Table S1), longer than at E13.5, suggesting that the average cell-cycle length of MGE progenitors gradually increases during embryonic development. To further characterize the variability of cell-cycle length, we used timelapse microscopy to monitor cell-cycle progression in cultured MGE cells at E13.5 (Fig. S2A–C). Histogram analysis showed that the distribution of cell-cycle lengths of MGE progenitors occupied a wide range, between 8 and 22 h (Fig. S2D). The average cell-cycle length of cultured MGE cells at E13.5 was 12.20 ± 0.20 h, which was significantly shorter than at E15.5 (20.11 ± 1.16 h) (Fig. S2E; Table S1). In addition, we found that only 24.0% of lineages showed relatively stable cell-cycle length in 3-round divisions (relative variation less than 10%) (Fig. S2F and S2G). These results suggest that cell-cycle length in individual lineages of MGE progenitors over multiple divisions exhibits remarkable heterogeneity and instability. To determine whether IdUBrdUEdU cells do indeed undergo division, and if so, how many times they divide, we monitored the lineages of cultured MGE cells at E13.5, and sequentially applied IdU, BrdU and EdU into the culture medium to mimic the triple-labeling method in vivo (Fig. 1F and 1G). Indeed, for each ΔT, 100% of IdUBrdUEdU cells divided at least once during observation (Figs. 1G–I and S2H; Table S1). However, in ΔT = 14 h group, 34.05% ± 5.66% of IdUBrdUEdU cells divided twice (Fig. 1H and 1I; Table S1), suggesting that about one-third of IdUBrdUEdU cells in ΔT = 14 h group were progenitors undergoing short cycles. During measuring the cell-cycle length of MGE progenitors in the different ΔT groups, we found that the ranges of cell-cycle length for 10%–90% progenitors in the ΔT = 4 h, ΔT = 9 h and ΔT = 12 h groups were 6–12 h, 11–15 h and 13–18 h, respectively (Fig. S2I and S2J; Table S1). Taken together, these results indicate that, for ΔT ≤ 12 h, the triple-labeling method can specifically label MGE progenitors with different cell-cycle lengths at E13.5, and the progenitors labeled in the ΔT = 4 h and ΔT = 12 h have distinct cell-cycle lengths. As mentioned above, PSC and PLC in MGE at E13.5 can be individually labeled in vivo by the triple-labeling method when ΔT = 4 h and ΔT = 12 h. We divided the MGE in coronal brain slices into six 30° sectors (Fig. 1J), the density