Diversity of transgenic mouse models for selective targeting of midbrain dopamine neurons

Abstract

Ventral tegmental area (VTA) dopamine (DA) neurons have been implicated in reward, aversion, salience, cognition, and several neuropsychiatric disorders. Optogenetic approaches involving transgenic Cre-driver mouse lines provide powerful tools for dissecting DA-specific functions. However, the emerging complexity of VTA circuits requires Cre-driver mouse lines that restrict transgene expression to a precisely defined cell population. Because of recent work reporting that VTA DA neurons projecting to the lateral habenula release GABA, but not DA, we performed an extensive anatomical, molecular, and functional characterization of prominent DA transgenic mouse driver lines. We find that transgenes under control of the tyrosine hydroxylase, but not the dopamine transporter, promoter exhibit dramatic non-DA cell-specific expression patterns within and around VTA nuclei. Our results demonstrate how Cre expression in unintentionally targeted cells in transgenic mouse lines can confound the interpretation of supposedly cell-type-specific experiments. This Matters Arising paper is in response to Stamatakis et al. (2013), published in Neuron. See also the Matters Arising Response paper by Stuber et al. (2015), published concurrently with this Matters Arising in Neuron.

Figure 1. Anatomical Characterization of Transgenic Cre Mouse Driver Lines

(A) Regions of the posterior midbrain in which eYFP/TH co-localization was analyzed (blue: midline VTA; green: lateral VTA, red: interpeduncular nucleus, IPN). (B) Fluorescence image of a coronal brain slice showing TH immunostaining (546 nm, red) of the same region depicted in (A) from a C57Bl6 mouse. Note the low number of TH-immunopositive cells in the midline VTA region (blue) and absence of TH-positive cells in the IPN (red). (C) Fluorescence image showing eYFP expression (488 nm, green) and TH immunostaining (546 nm, red) of the same region depicted in (A) from a TH-Cre mouse. (D) Mouse brain atlas showing areas in the anterior midbrain in which eYFP/TH co-localization has been analyzed (brown: lateral VTA; orange: midline VTA). (E) Fluorescence image of a coronal brain slice showing TH immunostaining (546 nm, red) of the same region depicted in (D) from a C57Bl6 mouse. Note the very low number of TH-immunopositive cells in the midline VTA region (orange). (F) Fluorescence image showing eYFP-expression (488 nm, green) and TH immunostaining (546 nm, red) of the same region depicted in (D) from a TH-Cre mouse. (G) Fluorescence images comparing the areas depicted in (A) and (D) for different transgenic mouse driver lines (eYFP, 488 nm, green; TH, 546 nm, red). (H) Pie charts showing the total percentage of co-localization of eYFP with TH-immunopositive (yellow) or TH-immunonegative (green) cells for the five color-coded areas depicted in (A) and (D) and for the mouse lines presented in (G). All scale bars: 200 μm.

Figure 2. Electrophysiological and Molecular Properties of eYFP-Expressing Cells in TH-Cre Mice

(A) Experimental approach for characterizing electrophysiological properties of eYFP-expressing cells in TH-Cre mice. Inset shows a neuron expressing eYFP prior to recording. Scale bar: 10 μm. (B) Anatomical positions of the recorded and neurobiotin-filled cells that are TH immunonegative (green) or TH immunopositive (yellow). (C) Sample confocal images of individual eYFP-expressing cells (488 nm, green) that were filled with neurobiotin (NB; 546 nm, red) via the recording pipette and immunostained for tyrosine-hydroxylase (TH; 647 nm, blue). All scale bars: 10 μm. (D) Sample whole-cell patch-clamp recording from an eYFP-expressing cell that has been NB-filled and posthoc stained for TH. Left: TH-immunonegative cell; right: TH-immunopositive cell. Corresponding membrane potentials in response to hyperpolarizing currents shows lack of rebound delay in TH-immunonegative cells. Note that both cell types show prominent sags during the hyperpolarizing steps. (E) Bar graphs showing mean sag amplitudes (left) and mean rebound delays (right) for eYFP+/TH− (green) and eYFP+/TH+ (yellow) cells. ***p < 0.001. Data represent means ± SEM. (F) Corresponding membrane-potential responses to depolarizing current pulses. Note that eYFP+/TH− (left) and eYFP+/TH+ (right) cells show depolarization block of spontaneous electrical activity after discharging at different maximal frequencies. (G) Bar graphs showing mean maximal firing frequency after reaching depolarizing block (left) and mean spontaneous firing frequencies (no current injections) from eYFP+/TH− (green) and eYFP+/TH+ (yellow) cells. ***p < 0.001. Data represent means ± SEM. (H) Single high-resolution action-potential (AP) waveforms from an eYFP+/TH− (left) or eYFP+/TH+ (right) cell. (I) Bar graphs showing mean AP duration (left) and mean afterhyperpolarization (AHP) amplitudes (right) for eYFP+/TH− (green) and eYFP+/TH+ (yellow) cells. **p < 0.01. Data represent means ± SEM. (J) Schematic drawing of the experimental approach to analyze gene expression profiles of eYFP-expressing cells in TH-Cre mice. (K) Bar graph comparing the percentage of eYFP-expressing cells that express TH, DAT, or VMAT2 genes in the substantia nigra pars compacta (SNc), lateral VTA, midline VTA, and interpeduncular nucleus (IPN). (L) Pie charts showing eYFP-expressing cells that co-express TH, DAT, and VMAT2 (yellow) or lack co-expression of TH, DAT, and VMAT2 (green) for SNc and lateral VTA as well as midline VTA and IPN regions. Striped area indicates cells expressing GAD1 and/or GAD2 genes.

Figure 3. Optogenetic Dissection of the Mesohabenular Pathway

(A) Schematic showing dual injections of fluorescent beads into the LHb and AAV into the VTA of a TH-Cre mouse (JAX:8601). (B) Confocal image demonstrating retrogradely labeled (i.e., bead-containing, 546 nm, red) and TH (647 nm, blue)-immunonegative VTA neurons that express eYFP (488 nm, green). Scale bar: 20 μm. (C) Bar graph showing that all retrogradely labeled cells that also express eYFP do not possess detectable TH immunosignals. (D) Schematic of AAV injection into the VTA of TH-Cre, DAT-Cre, GAD2-Cre, and VGlut2-Cre mice and ex vivo whole-cell patch-clamp recordings of LHb neurons. (E) Left: light-evoked IPSCs recorded from LHb neurons in TH-Cre mice are blocked by bath application of picrotoxin (red trace). Scale bar: 100 pA, 10 ms. Inset shows ChR2-eYFP (488 nm, green) expression in the LHb of a TH-Cre mouse. Scale bar: 200 μm. Right: bar graph showing mean IPSC amplitudes before (blue) and after (red) bath application of picrotoxin. Pie chart indicates the percentage of recorded LHb neurons that responded to light stimulation. **p < 0.01. Data represent means ± SEM. (F) Same experimental approach as in Figure 3E, but for DAT-Cre mice. (G) Same experimental approach as in Figure 3E, but for GAD2-Cre mice. *p < 0.05. Data represent means ± SEM. (H) Left: light-evoked EPSCs recorded from LHb neurons in VGlut2-Cre mice are blocked by bath application of CNQX (red trace). Scale bar: 100 pA, 10 ms. Inset shows ChR2-eYFP (488 nm, green) expression in the LHb of a VGlut2-Cre mouse. Scale bar: 200 μm. Right: bar graph showing mean EPSC amplitudes before (blue) and after (red) bath application of CNQX. Pie chart indicates the percentage of recorded LHb neurons that responded to light stimulation. **p < 0.01. Data represent means ± SEM. (I) Schematic of AAV injection into the VTA of TH-Cre, GAD2-Cre, and VGlut2-Cre mice and light activation of LHb terminals. (J) Location of a TH-Cre (upper panel) and a VGlut2-Cre (lower panel) animal that received stimulation in the compartment marked by the blue arrow. (K) TH-Cre (n = 4) and GAD2-Cre (n = 5) mice show a mild preference for the compartment paired with optical stimulation, while VGlut2-Cre mice (n = 4) robustly avoid the same compartment. A 2-way repeated-measures ANOVA revealed a main effect of compartment (F_2,20 = 7.171, p = 0.004) and a genotype × compartment interaction (F_4,20 = 4.448, p = 0.01). Student-Newman-Keuls post-hoc tests demonstrated that TH-Cre and GAD2-Cre mice did not differ from each other in the amount of time spent in stimulated or non-stimulated compartments (p values > 0.63), but both differed from VGlut2-Cre mice on the same measures (p values < 0.012). Inset graph shows that reversing the compartment in which VGlut2-Cre animals received stimulation caused robust avoidance of that compartment. Data represent means ± SEM, *p < 0.05, **p < 0.01.