Anatomical characterization of Cre driver mice for neural circuit mapping and manipulation

Abstract

Significant advances in circuit-level analyses of the brain require tools that allow for labeling, modulation of gene expression, and monitoring and manipulation of cellular activity in specific cell types and/or anatomical regions. Large-scale projects and individual laboratories have produced hundreds of gene-specific promoter-driven Cre mouse lines invaluable for enabling genetic access to subpopulations of cells in the brain. However, the potential utility of each line may not be fully realized without systematic whole brain characterization of transgene expression patterns. We established a high-throughput in situ hybridization (ISH), imaging and data processing pipeline to describe whole brain gene expression patterns in Cre driver mice. Currently, anatomical data from over 100 Cre driver lines are publicly available via the Allen Institute's Transgenic Characterization database, which can be used to assist researchers in choosing the appropriate Cre drivers for functional, molecular, or connectional studies of different regions and/or cell types in the brain.

Keywords: Cre driver mice; anatomical characterization; genetic tools; in situ hybridization; neuronal cell types.

Figure 1

Cre characterization data types and processing pipeline. (A) An overview of the high-throughput pipeline starting with crosses of mice from every characterized Cre line to a Cre reporter strain (e.g., Ai9, Ai14) to generate double positive mice for brain collection at up to four postnatal ages. The entire brain is sectioned and tissue subjected to in situ hybridization (ISH) before informatics data processing and web presentation. (B) Types of ISH data collected. For every Cre/reporter line, ISH was used to detect the tdTomato reporter gene at P56. For most Cre lines, additional data types include ISH to detect the Cre gene itself at P56, and double-fluorescent ISH (DFISH) using probes for the tdTomato gene and corresponding endogenous gene to the Cre line promoter or to another cell type marker (e.g., Gad1 for inhibitory neurons). Image series throughout the rostral-caudal extent of the brain were collected to gather whole brain expression patterns. (C) Informatics data processing includes signal detection, quantification, and registration of each image series into the 3D Allen Reference Atlas (ARA) space to assign signal (expression energy) to all annotated brain regions within the ARA. A whole brain expression summary is presented online at a coarse level across 12 major brain divisions along with the image series for each brain. Abbreviations are Iso, isocortex; OLF, olfactory areas; HPF, hippocampal formation; CTsp, cortical subplate; STR, striatum; PAL, pallidum; TH, thalamus; HY, hypothalamus; MB, midbrain; P, pons; MY, medulla; and CB, cerebellum.

Figure 2

Manual analysis of Cre reporter expression patterns across the brain. To aid in interpretation and further mining of the informatics-derived expression energy values, expression patterns were visually classified into one of six categories and recorded in Supplemental Tables 2, 3; (1) widespread, (2) scattered, (3) sparse, (4) enriched, (5) laminar or restricted, and (6) restricted but sparse. Examples from the cortex (top) and thalamus (bottom) are shown for each category.

Figure 3

Overall correlations between Cre expression and corresponding endogenous gene expression across all Cre lines. (A) Expression energy values across 295 annotated structures covering the entire brain were correlated between tdTomato or Cre ISH datasets, and each endogenous gene profile corresponding to the unique promoter driving Cre expression from the Allen Brain Atlas (ABA). The average correlation between ABA and tdTomato datasets was positive (n = 119 lines, r² = 0.32, Spearman correlation). The average correlation between ABA and Cre ISH datasets was significantly higher (n = 83 paired lines, r² = 0.41, paired t-test ^***p < 0.001). (B) Cre reporter (tdTomato) expression values across the whole brain were significantly more similar to the corresponding ABA gene expression values in knock-in compared with transgenic Cre lines (unpaired t-test, n = 52 knock-in and 66 transgenic, ^*p < 0.05). (C) Scatter plot showing the ABA vs. tdTomato whole brain r²-value for each of the 119 Cre lines grouped by percentile rank. The point on the far right shows the average r²-value between biological replicates within a Cre line for comparison. Red circles indicate examples shown in Figure 4.

Figure 4

Representative examples of the correlations between individual Cre lines and the corresponding endogenous gene expression. Images of ABA, Cre, tdTomato ISH, and tdTomato/endogenous gene DFISH are shown at two positions in the brain (top and bottom panels for each line). For each Cre line, a scatterplot comparing the expression energy in each of 295 brain structures from ABA vs. tdTomato data is shown on the far right. The examples shown here are representative of lines with high (A) Erbb4-CreERT2, mid (B) Drd2-Cre_ER44, low (C) Slc6a3-Cre, and very low (D) Cdhr1-Cre_KG66 overall correlations with the corresponding ABA gene expression profiles. In each case shown, DFISH demonstrates that for highly correlated areal gene expression, cell-specific gene expression is also mostly overlapping, whereas when expression energy at the areal level is not correlated, tdTomato and endogenous gene expression are not usually expressed in the same cells. Abbreviations are SSp-tr, primary somatosensory area, trunk region; MEA, medial amygdalar nucleus; CP, caudoputamen; VISp, primary visual area; SNc, substantia nigra, compact part; VTA, ventral tegmental area; LRN, lateral reticular nucleus; MOB, main olfactory bulb; and SCs, superior colliculus, sensory related.

Figure 5

Cre lines provide genetic access to specific brain areas through enriched or restricted gene expression patterns. Within this collection of 135 characterized Cre lines, enriched expression patterns can be seen in nuclei from all major brain regions. Representative examples of Cre lines with region-selective expression patterns of the tdTomato reporter are shown within olfactory areas (Rbp4-Cre_KL100, Sim1-Cre_KJ18), hippocampal formation (Pomc-Cre (ST), Gpr26-Cre_KO250), cortical subplate (Syt17-Cre_NO14, Grik4-Cre), striatum (Penk-2A-CreERT2, Sst-Cre, Prkcd-GluCla-CFP-IRES-Cre), hypothalamus [Lypd6-Cre_KL156, Nr5a1-Cre, Pomc-Cre (BL)], pallidum (Slc6a3-Cre, Gnrh-Cre), thalamus (Ntrk1-IRES-Cre, Gal-Cre_KI87), midbrain (Rorb-IRES2-Cre, Pvalb-IRES-Cre), pons (Scnn1a-Tg2-Cre, Th-Cre_FI172), medulla (Crh-IRES-Cre(ZJH), Pnmt-Cre), and cerebellum (Pcp2-Cre_GN135, Ntsr1-Cre_GN220). Abbreviations are: AOB, accessory olfactory bulb; NLOT, nucleus of the lateral olfactory tract; DG, dentate gyrus; field CA1; EPd, endopiriform nucleus, dorsal part; BLA, basolateral amygdalar nucleus; CP, caudoputamen; LS, lateral septal nucleus; CEA, central amygdalar nucleus; SCH, suprachiasmatic nucleus; VMH, ventromedial hypothalamic nucleus; ARH, arcuate nucleus; BST, bed nuclei of the stria terminalis; MS, medial septal nucleus; PVT, paraventricular nucleus of the thalamus; AD, anterodorsal; and AV, anteroventral nucleus of the thalamus; SCs, superior colliculus, sensory related; III, oculomotor nucleus; V, motor nucleus of trigeminal; PG, pontine gray; IO, inferior olivary complex; DCO, dorsal cochlear nucleus; CUL, culmen; and IP, interposed nucleus.

Figure 6

Cre lines provide genetic access to specific cortical layers and cortical interneurons for circuit analyses. (A,B) Many Cre lines have been generated with selective expression in distinct cortical layers. Representative tdTomato ISH images are shown from a subset of Cre lines with preferential expression in neurons of layers 2/3 (Cux2-CreERT2, Rasgrf2-2A-dCre), layer 4 (Nr5a1-Cre, Rorb-IRES2-Cre, Scnn1a-Tg3-Cre), layer 5 (Rbp4-Cre_KL100, Etv1-CreERT2, Chrna2-Cre_OE25), and layer 6 (Ntsr1-Cre_GN220, Syt6-Cre_KI14). (C) Representative images from Cre lines with tdTomato expression driven by cortical interneuron marker promoters (Gad2-IRES-Cre, Pvalb-IRES-Cre, Sst-IRES-Cre, VIP-IRES-Cre, Calb2-IRES-Cre). Density and level of restriction to specific layers varies and is often dependent on the specific cortical region of interest, thus researchers should use the entire whole brain image series to confirm expression in other areas.

Figure 7

Cre lines provide genetic access to specific neuromodulatory cell types and brain regions. Endogenous gene expression and tdTomato reporter ISH images show the faithful recapitulation of regional expression patterns between Cre line reporter expression at P56 and the corresponding endogenous gene expression patterns from the ABA for four major neuromodulatory systems. DFISH data from each line also demonstrates high levels of correspondence between the endogenous gene and reporter expression within the same cell. (A) Cholinergic neurons in the SI, basal forebrain; CP, caudoputamen; LDT, laterodorsal tegmental nucleus; and V, motor nucleus of the trigeminal; that express choline acetyltransferase mRNA (Chat), among other regions not shown, are also specifically labeled in the Chat-IRES-Cre line. (B) Noradrenergic neurons labeled for dopamine beta-hydroxylase mRNA (Dbh) are specifically located with the LDT, SLD, sublaterodorsal nucleus; and LC, locus ceruleus. The same regions have enriched expression of the tdTomato reporter in the Dbh-Cre_KH212 transgenic line. (C) Dopaminergic cells expressing dopamine transporter mRNA (Slc6a3) and Cre reporter mRNA in the Slc6a3-Cre line are found within the VTA, ventral tegmental area; SNc, substantia nigra, pars compacta; CLI, central linear nucleus raphe; and RR, retrorubral area. (D) Serotonergic neurons, shown labeled with a probe against the serotonin transporter gene, Slc6a4, are located with the dorsal raphe nucleus (DR), and the brainstem raphe nuclei, obscurus (RO) and pallidus (RPA). The same expression pattern is observed in the transgenic Slc6a4-Cre_ET33 line.