(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .
Regulation of chromatin accessibility and gene expression in the developing hippocampal primordium by LIM-HD transcription factor LHX2 [1]
['Varun Suresh', 'Department Of Biological Sciences', 'Tata Institute Of Fundamental Research', 'Mumbai', 'Bhavana Muralidharan', 'Institute For Stem Cell Science', 'Regenerative Medicine', 'Bangalore', 'Saurabh J. Pradhan', 'Chromatin Biology']
Date: 2023-10
In the mammalian cerebral cortex, the hippocampal primordium (Hcp) occupies a discrete position in the dorsal telencephalic neuroepithelium adjacent to the neocortical primordium (Ncp). We examined transcriptomic and chromatin-level features that distinguish the Hcp from the Ncp in the mouse during the early neurogenic period, embryonic day (E)12.5. ATAC-seq revealed that the Hcp was more accessible than the Ncp at this stage. Motif analysis of the differentially accessible loci in these tissues revealed LHX2 as a candidate transcription factor for modulating gene regulatory networks (GRNs). We analyzed LHX2 occupancy profiles and compared these with transcriptomic data from control and Lhx2 mutant Hcp and Ncp at E12.5. Our results revealed that LHX2 directly regulates distinct genes in the Hcp and Ncp within a set of common pathways that control fundamental aspects of development namely pluripotency, axon pathfinding, Wnt, and Hippo signaling. Loss of Lhx2 caused a decrease in accessibility, specifically in hippocampal chromatin, suggesting that this factor may play a unique role in hippocampal development. We identified 14 genes that were preferentially enriched in the Hcp, for which LHX2 regulates both chromatin accessibility and mRNA expression, which have not thus far been examined in hippocampal development. Together, these results provide mechanistic insight into how LHX2 function in the Hcp may contribute to the process by which the hippocampus acquires features distinct from the neocortex.
The brain performs an array of functions via functionally specialized structures. We examined how the hippocampus, a structure that is critical for learning and memory, acquires distinct molecular features from an adjacent structure, the neocortex, which processes different functions. As development proceeds, nuclear DNA, which is packaged into DNA-protein complexes called chromatin, is acted upon by regulatory factors that modulate its accessibility to the molecular machinery that produces messenger RNA (mRNA). Measurements of chromatin accessibility and quantification of gene-specific mRNA production, called gene expression profiling, are key tools to uncover early steps in how stem cells give rise to daughter cells with diverse functional capabilities. We found that the early embryonic hippocampal chromatin was significantly more accessible than that of the neocortex. We identified a regulatory protein, LHX2 that maintains this accessibility selectively in the hippocampus, and also controls the expression of distinct sets of genes in the hippocampus and the neocortex. Finally, we identified 14 LHX2 target genes that are important candidates for regulating hippocampal development. Our integrated multiomics approach offers an insight into how the hippocampus and the neocortex develop their unique neuronal compositions which underlie their specialized functions.
Funding: This work was funded by a Wellcome Trust- Department of Biotechnology India Alliance ( www.indiaalliance.org ) Early Career Fellowship (500197/Z/11/Z) to BM; a Centre of Excellence in Epigenetics program of the Department of Biotechnology Government of India
https://dbtindia.gov.in (BT/COE/34/SP17426/2016) to SG; a JC Bose Fellowship (JCB/2019/000013) from the Science and Engineering Research Board, Government of India
http://www.serbonline.in/ to SG; a grant from the Canadian Institutes of Health Research, a Canada-Israel Health Research Initiative jointly funded with the Israel Science Foundation, the International Development Research Centre, Canada
https://idrc.ca/en and the Azrieli Foundation to ST; a grant from the Department of Atomic Energy Government of India
https://dae.gov.in/ - Tata Institute of Fundamental Research www.tifr.res.in (12-R&D-TFR-5.10-0100RTI2001) to ST; a grant from the Department of Science and Technology, Govt. of India (DST/CSRI/2017/202) to ST. The funders had no role in study design, data collection and analysis, decision to publish, or the preparation of the manuscript.
LHX2 is a well-established regulator of distinct phenomena in early Hcp and Ncp development (reviewed in [ 4 ]). Loss of Lhx2 prior to E10.5 causes the Hcp and Ncp to transform into the hem and antihem, respectively [ 5 – 7 ]. Between E10.5 and E11.5, loss of Lhx2 causes the Ncp to acquire characteristics of the paleocortical primordium [ 8 ]. Loss of Lhx2 from E11.5 causes a range of phenotypes that have been well-characterized: early cell cycle exit of Ncp and Hcp progenitors leading to thinning of the superficial layers of the neocortex [ 9 ]; loss of the corpus callosum [ 10 ]; profoundly deficient thalamocortical innervation of the neocortex accompanied by the reduced electrical activity of subplate neurons [ 11 , 12 ]; reduction in layer 6 TBR1+ neurons and increase in layer 5 FEZF2+/CTIP2+ neurons [ 13 ]; drastic shrinkage of the hippocampus [ 14 ]. These studies motivated a chromatin-level analysis of LHX2 function in the Ncp and Hcp. We hypothesized that Lhx2 may participate in unique gene regulatory networks (GRNs) in the Hcp and Ncp by comparing chromatin accessibility, histone modifications, and transcriptomic changes in the cortex-specific Lhx2 conditional mutant Ncp and Hcp. We report that Lhx2 regulates unique genes in each tissue that map to four major developmental/signaling pathways. Furthermore, we found loss of Lhx2 leads to a decrease in chromatin accessibility, specifically in the Hcp, suggesting it is a major regulator of the Hcp chromatin state, and may control a cascade of processes that promote a distinct identity to the developing hippocampus.
The mammalian hippocampus arises from the dorsal telencephalic neuroepithelium that lies adjacent to that of the neocortical primordium (Ncp). The hippocampal primordium (Hcp) contains apical progenitors in the ventricular zone, intermediate progenitors in the subventricular zone, postmitotic neurons in the overlying cortical plate, and Cajal-Retzius cells in the marginal zone, similar to the cellular composition of the Ncp. The proliferation and differentiation of these cell types are regulated by a common set of transcription factors (TFs) e.g. PAX6, SOX2, and FOXG1 [ 1 – 3 ] in both structures. The acquisition of Hcp regional identity is expected to involve the regulation of chromatin both in terms of accessibility and histone modifications. Therefore, we analyzed these features in chromatin obtained from the Hcp and the Ncp at embryonic day (E) 12.5 in the mouse, when the neuroepithelium is predominantly proliferative and neurogenesis is in its early stages. The binding motif of transcription factor LHX2 emerged as a candidate in differentially accessible loci in the Hcp and Ncp chromatin.
Results
We isolated the Hcp and Ncp from E12.5 brains, as shown in the diagram, from intact hemispheres for all analyses (Fig 1A) [15]. Both tissues contain apical+basal progenitors, postmitotic neurons, and Cajal-Retzius cells at this stage. The progenitors in each tissue are actively producing neurons [15,16], although the Ncp tissue is likely to contain more postmitotic neurons than the Hcp tissue at its lateral extreme (Fig 1H). Fundamental regulators of progenitor proliferation, PAX6, FOXG1, SOX2, and TBR2, are expressed in both tissues at this stage [17–19]. To examine transcriptomic differences between the Ncp and Hcp, we performed RNA-seq (Fig 1B and S1 Table). We identified 1248 Ncp (Ncp>Hcp; p < 0.05) enriched and 1364 Hcp enriched genes (Hcp>Ncp; p < 0.05, Fig 1B and 1C).
PPT PowerPoint slide
PNG larger image
TIFF original image Download: Fig 1. Transcriptomic analysis of the wild type E12.5 Ncp and Hcp. (A) Schematic representation of the E12.5 mouse brain. (B) A volcano plot comparing Ncp and Hcp mRNA expression identifies preferentially enriched genes in the Ncp (1248) and the Hcp (1364). (C) A heatmap of the top 100 enriched genes in each tissue type; color bar: blue (low expression), red (high expression), clustering method: K means. (D, E) A tree plot depicting the top GO Biological Processes (GO: BPs) from (B). (F) Bar plot of selected differentially expressed genes between Ncp and Hcp. (G, H). In situ hybridization for selected genes identified in (B).
https://doi.org/10.1371/journal.pgen.1010874.g001
As expected, the top enriched Biological Processes (BPs) in both primordia span a range of neurodevelopmental phenomena, including neurogenesis, generation of neurons, and plasma membrane-bounded cell projection organization. However, each tissue is enriched with a distinct set of genes for each process, suggesting that although the Ncp and Hcp share similar cellular compositions consisting of apical and basal progenitors, newborn neurons, and Cajal-Retzius cells, the regulatory process that govern development in these tissues may be different (Fig 1C and 1D and S1 Table). Based on our RNA-seq results, we performed RNA in-situ hybridizations of select genes. Among the Hcp > Ncp genes were Wnt signaling components such as Fzd1, Lef1, and Axin2, as well as previously reported Hcp markers such as Lhx9 and Ephb1 (Fig 1F–1H [20]). Several TFs were also identified to be differentially enriched (111 Ncp > Hcp; 94 Ncp < Hcp) that are known or putative regulators of forebrain development (Fig 1F–1H1 and S1 Table). We also identified BPs enriched in only one tissue, such as cell-cell signaling and synaptic transmission in the Ncp and cell adhesion, extracellular matrix organization, and cilium organization in the Hcp (Fig 1D and 1E). Analysis of the top KEGG pathways displayed a similar pattern, with the Wnt signaling pathway being common to both primordia, but axon guidance, synaptic vesicle cycle, cAMP and MAPK signaling pathway being enriched in the Ncp, whereas TGF-beta signaling, Hippo pathway, and pathways regulating pluripotency of stem cells being enriched in the Hcp (S2A and S2B Fig). Overall our results revealed major transcriptomic differences between the Ncp and Hcp, suggesting that distinct GRNs are operational in each of these two tissue types.
The distinct transcriptomic profile of the E12.5 Ncp and Hcp motivated a comparative analysis of chromatin accessibility using ATAC-seq (assay for transposase-accessible chromatin sequencing) and identified similar numbers of accessible loci in both tissues (>100,000; Fig 2A). However, the Hcp was more accessible in 14804 loci (differentially accessible regions; DARs) which mapped to 9580 DAR-associated genes (DAGs), while 70 DARs (64 DAGs) were enriched in the Ncp (Fig 2B; FDR < 0.05, fold change > 1.5, [21]). Consistent with this, active histone marks H3K27Ac, H3K4Me3, and H3K4Me1 displayed greater occupancy in the 14804 Hcp>Ncp DARs (Fig 2D). To identify potential regulators of change in chromatin state, we investigated transcription factor-binding motifs in the DARs. Of the top 10 motifs, LIM-HD transcription factor LHX2 emerged as a factor of interest (Figs 2C and S2C-S2E) because it was the only one expressed at E12.5 in both tissues ([5], Fig 3B), and known to have stage-specific and cell type-specific roles in the early development of the neocortex and hippocampus [4]. Examples of Hcp-enriched genes, Lef1 and Wif1 [20,22], display greater chromatin accessibility in the Hcp than in the Ncp, and these chromatin regions also display greater occupancy of active histone marks (Fig 2E).
PPT PowerPoint slide
PNG larger image
TIFF original image Download: Fig 2. Chromatin accessibility comparison of the E12.5 wild type Ncp and Hcp. (A) A heatmap comparing open chromatin in the Ncp and Hcp. (B) Differential accessibility analysis shows 14804 loci (9508 genes) to be preferentially open in the Hcp and 70 loci (64 genes) to be more open in the Ncp. (C) Motif analysis of the differentially open loci identified in (B) reveals LHX2 among the top candidates. (D) Heat maps display greater active histone modifications on the 14804 loci identified as more open in the Hcp. (E) Genomic loci corresponding to the Lef1 and Wif1 loci demonstrating the correspondence between the open chromatin and activating histone marks in the Ncp (red) and Hcp (green). Black boxes mark regions enriched in open chromatin in the Hcp that align with one or more histone modifications. The numbers indicate the maximum peak height for each pair of (Hcp/Ncp) tracks.
https://doi.org/10.1371/journal.pgen.1010874.g002
PPT PowerPoint slide
PNG larger image
TIFF original image Download: Fig 3. LHX2 occupancy in the E12.5 mouse neocortical and hippocampal primordia (Ncp and Hcp, respectively). (A) Schematic representation of the E12.5 mouse brain. (B) Progenitor markers PAX6, LHX2, and TBR2 immunostaining/in situ hybridization (Lhx2) in the Ncp and Hcp. (C-G) LHX2 ChIP-seq data in the Ncp and Hcp. Plots of PePr peak-called regions in the Ncp and Hcp show TF LHX2 occupancy in each tissue. Only statistically significant peaks were used for further analysis (p-value 0.0001 and fold change over input: cut off >10 fold) (C); The number of LHX2 occupancy peaks and associated number of genes (D); common genes occupied by LHX2 between the Ncp and Hcp (E); Percentage of LHX2 occupancy peaks categorized by type of genomic region (F); LHX2-occupied genes enriched in different cortical cell types identified by gene enrichment profiles in [28](G). The scale bars in B are 100 μm.
https://doi.org/10.1371/journal.pgen.1010874.g003
Since LHX2 emerged as a common factor in the motif analysis of the DARs in the Ncp and Hcp, we investigated its genomic occupancy in these tissues. We performed ChIP-seq to examine LHX2 occupancy in these tissues and identified 2222 binding sites mapping to 1870 genes in the Ncp and 5166 binding sites mapping to 3758 genes in the Hcp. Of these, 1018 genes were common to the Ncp and Hcp (Fig 3C–3E). Binding on promoters was limited to 9–12% in both tissues, whereas regions such as introns, putative enhancers, and intergenic regions accounted for the majority of the occupancy loci (Figs 3F and S3 and S3 Table). These results suggest that global LHX2 function may be linked to occupancy not only at the TSS/promoter but also in intronic and intergenic regions as well as 5’ or 3’ UTRs as shown across multiple systems (S3H Fig [13,23–25]).
At E12.5, both tissues contain a mix of progenitors (apical and basal) and newly produced postmitotic neurons. Lhx2 is expressed in each of these populations (Fig 3B, [26–28]. We compared the LHX2-occupied genes with those known to be enriched in apical progenitors/basal progenitors/postmitotic neurons in a single-cell RNA-seq (scRNA-seq) dataset of Ncp tissue [28]. Although no such dataset is available for Hcp tissue, the similarities in cell type composition permitted a comparison and revealed that LHX2-bound genes corresponded to those enriched in progenitors (31% of LHX2-occupied genes in Ncp and 35% in Hcp) as well as those enriched in postmitotic neurons (55% of LHX2-occupied genes in Ncp and 50% in Hcp, Fig 3G), consistent with its diverse roles in multiple aspects of telencephalic development [4].
For subsequent analysis, we individually focused on the Ncp and Hcp, comparing wild-type and Lhx2 mutant tissue in each case. The LHX2 occupancy profile motivated an examination of how this factor may regulate chromatin accessibility and histone modifications associated with active or repressed loci. Therefore, we induced conditional loss of Lhx2 using a dorsal telencephalon-specific driver, Emx1Cre [29]. This Cre line acts from E11.5 [9], ideally suited to examine the potentially immediate effects of loss of Lhx2 in the Ncp and Hcp by E12.5. A comparison of accessibility of wild-type (wt) and Lhx2 mutant (mut) Ncp and Hcp respectively, was performed using the DESeq2 package [21]. Loss of Lhx2 did not alter global chromatin accessibility significantly in Ncp (Fig 4A). In contrast, the loss of Lhx2 caused a striking reduction in accessibility in the Hcp at 463 DARs (405 Differentially Accessible Genes, DAGs) and an increase in accessibility in 1 DAR (1 DAG; Fig 4B). Since these data were a result of pairwise locus comparisons across the genome, we classified the comparisons based on TSS, DARs, and LHX2 binding region (LHX2BR) and compared data for these categories (Fig 4C). In the Hcp, all three categories displayed greater accessibility in the wild type than in the mutant tissue. Consistent with this, ChIP-seq for the repressive H3K27Me3 mark revealed a sharp increase in occupancy in the mutant. The active H3K27Ac and H3K4Me3 marks displayed little or no change in occupancy upon the loss of Lhx2 (Fig 4D). We correlated the DARs (463 up +1 down) that exhibit a change in accessibility upon loss of Lhx2 with LHX2 genomic occupancy in the Hcp. 360 of these DARs mapped to 311 genes that showed LHX2 occupancy. The majority of these are in intronic or intergenic regions (S3I Fig). IGV tracks of some examples of these loci, Fezf2, Robo1, and Hopx, revealed loss of open chromatin at or near the site of LHX2 occupancy (Fig 4E and 4F), suggesting a role for this factor in maintaining specific loci open in the Hcp. In summary, the loss of Lhx2 in the Hcp renders the chromatin less accessible. This feature is not seen in the Ncp, suggesting that the regulation of chromatin accessibility may be an Hcp-specific function of LHX2.
PPT PowerPoint slide
PNG larger image
TIFF original image Download: Fig 4. Chromatin accessibility changes upon loss of Lhx2 in the Ncp and Hcp. (A, B) Scatter plots comparing control versus Lhx2 mutant chromatin identify loci for which the global accessibility has changed upon loss of Lhx2 in the Ncp (0) and Hcp (463+1 DARs which map to 405+1 DAGs). (C) Plot profile comparisons of wild type and mutant chromatin in the Hcp showing that the mutant chromatin is less accessible at LHX2 binding sites, TSS, and regions identified in Fig 2B to be differentially accessible (DARs) between wtNcp and wtHcp. (D) Histone modification profiles in the Hcp focusing on the TSS reveal that the loss of Lhx2 appears to be associated with an increase in the repressive mark H3K27Me3, a reduction in the mark H3K4Me3, and no apparent change in H3K27Ac. (E) A Venn diagram illustrates the majority of the down-regulated DARs are associated with an LHX2 binding peak in the Hcp. (F) Examples of genomic loci showing LHX2 binding regions at which chromatin accessibility is decreased upon loss of Lhx2 (also see S6D Fig).
https://doi.org/10.1371/journal.pgen.1010874.g004
A major functional consequence of changes in chromatin accessibility is the alteration of gene expression [30,31]. The Lhx2 mutant phenotype has been extensively characterized in the neocortex and hippocampus after inducing conditional loss of function at different stages (reviewed in [4]; Fig 5I–5O). We sought to identify the GRNs that were affected by the loss of Lhx2 in the Ncp and the Hcp by analyzing RNA-seq data. In the Ncp, 2372 differentially expressed genes (DEGs) were identified by comparing wild-type Ncp and mutant Ncp datasets. These consisted of 1150 DEGs that were downregulated (wtNcp > mutNcp) and 1222 that were upregulated (mutNcp > wtNcp) in the mutant Ncp (FDR <0.05; Fig 5B, 5G and 5H). In the Hcp, 1217 DEGs were identified by comparing wild-type Hcp and mutant Hcp datasets. These consisted of 401 DEGs that were downregulated (wtHcp > mutHcp) and 816 that were upregulated (mutHcp > wtHcp) in the mutant Hcp (FDR <0.05; Fig 5A, 5G and 5H). For each tissue, downregulated and upregulated genes were analyzed by Over Representation Analysis (ORA) and Gene Set Enrichment Analysis (GSEA) for Gene Ontology Biological Processes (GO: BPs). In the Hcp, downregulated genes affected the BPs corresponding to DNA conformation, chromosome organization, and DNA replication by both methods of analysis (Fig 5C and 5E), which are consistent with the reduced accessibility at 463 loci in this tissue upon loss of Lhx2 (Fig 4B). In the Ncp, the downregulated genes affected the BPs corresponding to pathways such as Wnt signaling and Hippo signaling (Fig 5D and 5F). The genes that are upregulated upon the loss of Lhx2 are associated with a common set of BPs in the Ncp and Hcp, such as Neurogenesis, Neuron projection development, and related processes (S4 Fig), which correspond well with the phenotype of precocious neurogenesis upon the loss of Lhx2 in both tissues [13,32].
PPT PowerPoint slide
PNG larger image
TIFF original image Download: Fig 5. Loss of Lhx2 causes distinct patterns of transcriptomic dysregulation in the Ncp and Hcp. (A, B) Volcano plots displaying genes dysregulated in each tissue upon loss of Lhx2. (C-F) Gene Ontology Biological Processes (GO: BP) corresponding to genes downregulated in Hcp (C, E) and Ncp (D, F). (C, D) shows the GSEA analysis and (E, F) shows the overrepresentation test of the GO: BPs. The corresponding upregulated gene analysis is in S4 Fig. (G) mRNA in situ hybridization at E12.5 for some dysregulated genes in the Emx1Cre::Lhx2cko. (H) Corresponding bar plots of the mRNA fold changes from the RNA-seq data. (I-O) Schematics summarizing loss of Lhx2 phenotypes. (I-K, O) are partially modified from[4]. (I) Disruption before E10.5 causes the dorsal telencephalic primordium to take on the fate of the hem and the antihem [5,6,39]. (J) Disruption from E10.5 results in shrinkage of the neocortical primordium (Ncp) and expansion of the paleocortical primordium [8]. (K-N) Disruption at E11.5 causes the Ncp and Hcp progenitors to exit the cell cycle early, resulting in the dramatic shrinking of both structures (K, L; [13,32]) due to premature neurogenesis; a perturbation of cell fate such that TBR1+ layer 6 neurons are reduced in number, and CTIP2+ layer 5 neurons are increased in number (M; [13]); thalamocortical axons (green fibers) prematurely grow into the cortical plate due to a deficit in the subplate (N; [11]). (O) Lhx2 disruption at E15.5 in hippocampal progenitors results in premature gliogenesis during the neurogenic period [60].
https://doi.org/10.1371/journal.pgen.1010874.g005
We compared the genes dysregulated upon loss of Lhx2 with LHX2 occupancy to arrive at a set of potential direct targets in the Ncp (Fig 6A) and the Hcp (Fig 6C). These “direct” targets as well as “all” DEGs, were then curated using the Ncp scRNA-seq dataset of [28] (Fig 6B and 6D), to identify genes expressed in progenitors (apical/basal) or neurons. In the Ncp, the majority of downregulated DEGs were enriched in progenitors, whereas the majority of upregulated DEGs were enriched in neurons (Fig 6B). These data are consistent with the established role of Lhx2 in maintaining progenitor proliferation and premature cell-cycle exit and depletion of the progenitor population upon loss of Lhx2 (Fig 5K, [9,14,32]). Such a pattern was not obvious in the Hcp DEGs. We sought to define the core GRNs regulated by LHX2 in the Ncp and Hcp, first examining only the potential “direct” targets. Comparing potential direct targets of LHX2 in both tissues revealed DEGs unique to the Ncp and the Hcp and some common genes (Fig 6E). Four common signaling pathways emerged from these datasets in both tissues: Wnt signaling, Hippo signaling, Signaling pathways related to pluripotency of stem cells, and Axon guidance (S5A and S5B Fig). In each pathway, there were LHX2 targets unique to each tissue as well as some common targets, the fold changes for which are displayed in Fig 6I–6L). Since these dysregulated pathways are fundamental to developmental processes, we examined whether LHX2 occupied the corresponding genes at E10.5. At this stage, the dorsal telencephalon (dtel) is almost entirely composed of apical progenitors proliferating in self-expansion mode, and the medial primordium that will later form the Hcp has not invaginated substantially. Therefore, we compared LHX2 occupancy in E10.5 dtel tissue (Fig 6H) with the E12.5 Ncp and Hcp occupancy data (Fig 3). Approximately 50% of the E10.5 dtel-occupied genes (2254) overlapped with genes occupied by LHX2 at E12.5. 286 of these overlapped with the E12.5 Ncp, 1252 with the E12.5 Hcp, and 716 were occupied in both primordia (Fig 6F). Several of these genes were present in the four signaling pathways identified to be dysregulated at E12.5 (black gene names, Fig 6I–6L), and only a handful were occupied only at E12.5 (blue gene names, Fig 6I–6L). We extended our analysis to indirect targets of LHX2 that belong to the same KEGG pathways identified for the direct targets, resulting in common as well as Ncp/Hcp-specific GRNs. Interestingly, these indirect targets included several ligands and receptors known to participate in these same four pathways, such as Rspo1,3, Wnt2b, Fzd3, Tgfbr2, Bmpr2, Slit2,3, Sfrp4, and Cxcl12 (S5 Fig).
PPT PowerPoint slide
PNG larger image
TIFF original image Download: Fig 6. Gene Regulatory Networks modulated by LHX2 in the Ncp and Hcp. (A, C) Venn diagrams depicting the number of genes occupied by LHX2 and dysregulated (blue: downregulated, red: upregulated) upon loss of Lhx2 in the Ncp (A) and Hcp (C) respectively to identify direct targets of LHX2 in the Ncp and Hcp. (B, D) Genes dysregulated upon loss of Lhx2 in the Ncp (B) and Hcp (D) respectively, categorized by “Direct” or (direct + indirect) = “All” targets, mapped to the cell-type specific gene enrichment profiles in [28]) to identify progenitor-enriched (grey) and neuron-enriched genes (black). (E) Venn diagram comparing the direct targets of LHX2 that are dysregulated upon loss of Lhx2 in the Ncp (112 downregulated; 118 upregulated) and Hcp (70 downregulated; 153 upregulated), and in both tissues (43 downregulated; 35 upregulated). (F) Comparison of LHX2 occupancy in the E10.5 dorsal telencephalon (dtel; blue circle) with that in the E12.5 Ncp (red) and Hcp (green) results in genes occupied in all these three tissues (716, yellow), in the E10.5 dtel and the E12.5 Ncp (286, red) or the E12.5 Hcp (1252, green). (G) Venn diagram comparing the genes in E (LHX2 direct targets) that are also occupied by LHX2 at E10.5. In the Ncp, there are 62 downregulated 59 upregulated genes. In the Hcp there are 33 downregulated upregulated 94 upregulated. 37 downregulated and 25 upregulated are common to both tissues. (H) Heatmaps displaying genes occupied by LHX2. Cluster 1: Occupancy at both E10.5 (dtel) and E12.5 (Ncp and Hcp). Cluster 2: Occupancy at only E10.5. (I-L) KEGG pathway analysis (GO: BP) of genes identified in (E, G) reveals 4 pathways dysregulated upon loss of Lhx2 in the Ncp (red bars) and Hcp (green bars). Individual fold changes are plotted from the RNA-seq data (black: genes occupied by LHX2 at E12.5 and E10.5; blue: occupied only at E12.5).
https://doi.org/10.1371/journal.pgen.1010874.g006
As a final step, we focused on LHX2-regulated genes that were unique to the Hcp across multiple datasets. We progressively filtered data from each approach, beginning with the 3758 genes occupied by LHX2 in the Hcp (Fig 3C and 3E). Of these, 308 were also enriched in the wtHcp (Fig 1B: Hcp > Ncp genes). Of these, 39 genes were downregulated in the Hcp upon loss of Lhx2 (Fig 5A). Finally, of these, 14 genes displayed decreased chromatin accessibility in the mutHcp (Fig 7A). This group contains genes that encode molecules with established roles in development: DNA-binding protein Atxn7; transcription factors Bach2, Fezf2, Hopx; membrane-associated molecules Flrt3, Lrrn1, Tenm2, Slc39a10; ligand/secreted molecule Lrrc4c, Frzb; enzyme Dct; cytoplasmic protein Mtss1; kinase binding partner Rab11fip2, and a previously uncharacterized Gm14015 (Fig 7B). These 14 genes represent a unique set that is occupied and regulated by LHX2 at the level of chromatin accessibility and mRNA expression in the Hcp (IGV tracks in Figs 4F, 7C and S6D). None of these genes has thus far been examined for a role in hippocampal development and offer new avenues for understanding the ontology of this structure.
[END]
---
[1] Url:
https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1010874
Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.
via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/