New MiniPromoter Ple389 (ADORA2A) drives selective expression in medium spiny neurons in mice and non-human primates

Abstract

Compact cell type-specific promoters are important tools for basic and preclinical research and clinical delivery of gene therapy. In this work, we designed novel MiniPromoters to target D1 and D2 type dopaminoceptive medium spiny neurons in the striatum by manually identifying candidate regulatory regions or employing the OnTarget webserver. We then empirically tested the designs in rAAV-PHP.B for specificity and robustness in three systems: intravenous injection in mice, intracerebroventricular injection in mice, and intracerebroventricular injection in non-human primates. Twelve MiniPromoters were designed from eight genes: seven manually and five using OnTarget. When delivered intravenously in mice, three MiniPromoters demonstrated highly selective expression in the striatum, with Ple389 (ADORA2A) showing high levels of dopamine D2-receptor cell co-localization. The same three MiniPromoters also displayed enriched expression in the striatum when delivered intracerebroventricularly in mice with high levels of DARPP32 co-localization. Finally, Ple389 (ADORA2A) was intracerebroventricularly injected in non-human primates and showed enriched expression in the striatum as in the mouse. Ple389 (ADORA2A) demonstrated expression in the medium spiny neurons in all three systems tested and exhibited the highest level of D2-MSNs and DARPP32 co-labeling in mice, demonstrating its potential as a tool for gene therapy approaches for Parkinson and Huntington disease treatment.

Introduction

Driving cell- and tissue-specific expression of transgenes is one of the primary interests for targeting and studying cellular populations in neurodegenerative diseases, and also an important area of research for promoting safe and specific therapeutic approaches. In gene therapy, the promoter is one of the essential components of the expression cassette for controlling the region and level of expression of the therapeutic gene¹. A cell- or tissue-specific promoter can restrict unwanted transgene expression and facilitate persistent expression. Therefore, choosing a strong and specific promoter is an important step toward achieving successful therapeutic benefits. Additionally, selecting a suitable vector for expression allows robust transduction of desired target cells and may improve clinical outcomes.

Recombinant adeno-associated virus (rAAV) vectors are currently the leading gene delivery platform for the treatment of various human diseases². However, rAAV vectors are small, with a packaging capacity of only ~ 4.9 kb³, and human promoters are generally large and complex. Therefore, there is an emergent focus on developing short or compact promoters that can be used to safely deliver therapies in vivo and to improve the targeting efficacy of specific cell types that play an important role in various neurological disorders and behavioral addictions, such as the medium spiny neurons (MSNs) of the striatum⁴.

The striatum is a key component of the basal ganglia that plays a major role in motor, cognitive, and limbic functions^5,6. It receives inputs from most cortical areas and provides output to other basal ganglia components⁵. The GABAergic MSNs are the most abundant striatal neurons, representing 95% of the total neuronal population within the rodent dorsal striatum^7,8. The MSNs can be differentiated into two major subtypes based on their expression in D1 or D2 dopamine receptors^9–13. MSNs expressing dopamine D1-receptors (D1Rs) project primarily to the substantia nigra pars reticulata and entopeduncular nucleus (direct striatonigral pathway), while MSNs expressing dopamine D2-receptors (D2Rs) project to the lateral part of the globus pallidus (indirect striatopallidal pathway)¹⁴. Alterations of the abundance or functionality of the two populations of MSNs can contribute to the pathophysiology of several neurodegenerative diseases, including Parkinson Disease^12,15,16 and Huntington disease^17,18. Hence, targeting striatal cells is of interest for the study of cellular specificity in the development of these diseases and to target and deliver therapeutic approaches.

Previous studies have shown promising striatal-selective promoters which drove expression in the MSNs, specifically in the striatonigral pathway¹⁹ and cholinergic interneurons (1–3% of abundance in the striatum)²⁰. However, in both studies, the transduction spread to other brain regions^19,20. An additional study demonstrated robust and exclusive D1-MSNs expression using short promoters designed from the GPR88 gene and their application as promising tools for targeted circuit modulation for Parkinson Disease treatment²¹. Still, additional novel promoters are needed and will be beneficial for specific targets, including the D2-MSNs, especially for the study and treatment of neurodegenerative diseases in which D2-MSNs are affected at earlier stages, such as Huntington disease^19–22.

We have been designing and refining Pleaides (Ple) human-DNA MiniPromoters (minimal human promoter element(s) designed for expression in restricted cell types) based on short cis-regulatory regions (CRRs) that control gene expression in specific cells and tissues through the Pleiades Promoter Project^23–27. The Pleiades Promoter Project was established to advance research and therapies for brain disorders by bioinformatically producing high-throughput, brain-directed human MiniPromoters to drive DNA expression in defined brain regions of therapeutic interest²³. The project generated various human MiniPromoters, which are available on Addgene.org, using a three-step strategy: (1) identifying genes specifically expressed in adult brain regions or cell types of therapeutic interest; (2) predicting, computationally, the human regulatory regions (RRs) responsible for the specific expression, and (3) testing the MiniPromoters in vivo²³. Recently, we introduced the webserver OnTarget (http://ontarget.cmmt.ubc.ca) that semi-automates the MiniPromoter design process and improves the development of promoters for cell and tissue-specific through the combination of CRRs, such as promoters and enhancers²⁸. OnTarget provides compact promoters for gene therapy and can be useful for identifying CRRs in specific genomic loci²⁸.

In this work, we designed and characterized MiniPromoters to target striatal MSNs. We characterized these MiniPromoters combined with the PHP.B capsid to evaluate their specificity and off-target expression. The PHP.B capsid is a variant that efficiently transduces at least 40-folder greater than AAV9 in the adult mouse central nervous system (CNS) after intravenous injection²⁹. This variant allowed us to assess for maximal off-target expression in our system. Furthermore, we demonstrated strong transduction using the PHP.B capsid when combined with a ubiquitous promoter and delivered via intracerebroventricular injection in the mouse and monkey brain³⁰.

Using both mouse and non-human primate models, we characterized MiniPromoters for specificity and robustness in vivo, with the aim of producing novel tools for the study of therapeutic approaches against neurodegenerative diseases, such as Parkinson and Huntington.

Results

Twelve MiniPromoters designed from eight D1- and D2-MSN specific genes

D1- and D2-MSNs, the two main neuronal subtypes of the mammalian striatum, were chosen as cell types of interest in the present study because they are important targets for the development of therapeutic approaches to various neurodegenerative disorders such as Parkinson and Huntington disease^{9–13,15,17,31}. Candidate genes were chosen from commonly used D1- and D2-MSN markers in the literature^8,32–34 and are listed in Table 1. A total of twelve D1- and D2-MSN MiniPromoters were designed from eight genes: adenosine A2a receptor (ADORA2A; Ple355 and Ple389), dopamine receptor D1 (DRD1; Ple357 and Ple390), dopamine receptor D2 (DRD2; Ple358, Ple359, and Ple391), G protein-coupled receptors 6 (GPR6; Ple385) and 88 (GPR88; Ple379), prodynorphin (PDYN; Ple392), proenkephalin (PENK; Ple393), and tachykinin precursor 1 (TAC1; Ple386). Seven MiniPromoters were designed manually, while the other five were designed using the OnTarget²⁸. MiniPromoters varied in size, with one very compact at 516 bp (Ple379), six having an intermediate size between 1 and 2 kb (Ple358, Ple385, Ple386, Ple389, Ple392, and Ple391), and five larger at over 2 kb (Ple355, Ple357, Ple359, Ple390, and Ple393). The CRRs used to design each MiniPromoter are provided in Supplementary File S1 online. Figure 1 details the design of six MiniPromoters for the most successful genes: ADORA2A, DRD2 and GPR6. For each gene, a schematic representing the region surrounding ADORA2A (Fig. 1A), DRD2 (Fig. 1B), and GPR6 (Fig. 1C) is shown, highlighting the CRRs used in the final designs (Fig. 1D).

Table 1.

Summary of MiniPromoters designed to target striatum.

Gene	MiniPromoter	MiniPromoter size (bp)	Target	Components of viruses genome	MiniPromoter plasmid (pEMS)	rAAV plasmid (pEMS)	rAAV serotype	Species/route tested	Figure showing expression
ADORA2A	Ple355	2666	Striatum (MSN D2)	ssAAV-Ple355-EmGFP-WPRE	2290	2319	rAAV9	Mouse/IV and IP injection	Suppl. Figures 2, 5
ADORA2A	Ple355	2666	Striatum (MSN D2)	ssAAV-Ple355-EmGFP-WPRE	2290	2319	rAAV-PHP.B	Mouse/IV and ICV injection	Suppl. Figures 2, 5
ADORA2A	Ple389	1344	Striatum (MSN D2)	ssAAV-Ple389-EmGFP-WPRE	2372*	2387*	rAAV-PHP.B	Mouse/IV and ICV injection; Monkey /ICV injection	Figures 1, 2, 3, 4, 5, 6, 7
DRD1	Ple357	2200	Striatum (MSN D1)	ssAAV-Ple357-EmGFP-WPRE	2292	2321	rAAV9	Mouse/IV injection	Suppl. Figure 2
DRD1	Ple390	2312	Striatum (MSN D1)	ssAAV-Ple390-EmGFP-WPRE	2373	2388	rAAV-PHP.B	Mouse/IV injection	Suppl. Figure 2
DRD2	Ple358	1659	Striatum (MSN D2)	ssAAV-Ple358-EmGFP-WPRE	2293	2322	rAAV9	Mouse/IV injection	Suppl. Figure 2
DRD2	Ple359	2680	Striatum (MSN D2)	ssAAV-Ple359-EmGFP-WPRE	2294	2323	rAAV9	Mouse/IV injection	Suppl. Figure 2
DRD2	Ple391	1723	Striatum (MSN D2)	ssAAV-Ple391-EmGFP-WPRE	2374*	2389*	rAAV-PHP.B	Mouse/IV and ICV injection	Figures 1, 2, 3, 4, 5
GPR6	Ple385	1362	Striatum (MSN D2)	ssAAV-Ple385-EmGFP-WPRE	2368*	2383*	rAAV-PHP.B	Mouse/IV and ICV injection	Figures 1, 2, 3, 4, 5
GPR88	Ple379	516	Striatum (MSN D1 & D2)	ssAAV-Ple379-EmGFP-WPRE	2362	2377	rAAV-PHP.B	Mouse/IV and ICV injection	Suppl. Figures 2, 5
PDYN	Ple392	1112	Striatum (MSN D1)	ssAAV-Ple392-EmGFP-WPRE	2375	2390	rAAV-PHP.B	Mouse/IV and ICV injection	Suppl. Figures 2, 5
PENK	Ple393	2209	Striatum (MSN D2)	ssAAV-Ple393-EmGFP-WPRE	2376	2391	rAAV-PHP.B	Mouse/IV injection	Suppl. Figure 2
TAC1	Ple386	1199	Striatum (MSN D1)	ssAAV-Ple386-EmGFP-WPRE	2369	2384	rAAV-PHP.B	Mouse/IV injection	Suppl. Figure 2

MSN, Medium spiny neurons; D1, Dopamine 1 receptor; D2, Dopamine 2 receptor; IP, Intraparenchymal; ICV, Intracerebroventricular; IV, Intravenous; *, Plasmids available at Addgene (www.addgene.org).

Fig. 1.

Bioinformatics design of six MiniPromoters targeting medium spiny neurons in the striatum. (A–C) Identification of promoter and enhancer elements potentially regulating the expression of ADORA2A (A), DRD2 (B) and GPR6 (C). Vertically highlighted colored regions correspond to their color-matched element included in (D) the final MiniPromoters designs. (D) Final MiniPromoter designs for Ple389 (ADORA2A), Ple355 (ADORA2A), Ple391 (DRD2), Ple358 (DRD2), Ple359 (DRD2), and Ple385 (GPR6). ADORA2A. Adenosine A2a Receptor; DRD2, Dopamine D2 Receptor; GPR6, G protein-coupled receptor 6; Ple, Pleiades MiniPromoter; P2, the second promoter region studied by these authors; RR, regulatory regions; E, enhance; SINE, Short Interspersed Nuclear Element; TSS, Transcription Start Sites; CTCF, CCCTC-binding factor.

A total of twelve MiniPromoters were designed, sent for DNA synthesis, cloned into an rAAV genome plasmid driving EmGFP expression, and packaged into the rAAV9 and rAAV-PHP.B capsids (Table 1). MiniPromoter-EmGFP rAAV were then characterized in vivo first by intravenous injection in mouse, following which the most specific were tested by ICV injection in mouse, and finally the most promising MiniPromoter was selected for characterization by ICV injection in monkey.

Three MiniPromoters expressed in the striatum, with variable D1- and D2-MSN specificity, when delivered intravenously to mice

First, we evaluated the ability of the rAAV-PHP.B capsid to transduce neurons in multiple brain regions, using both a ubiquitous promotor, smCBA, and a pan-neuronal promoter, SYN1, following IV injection (Supplementary Fig. S1 online). smCBA positive control promoter resulted in cellular expression across the whole brain (cortex, forebrain, brainstem, cerebellum), with the strongest expression in cortical and thalamic regions. Similarly, despite some structure-specific differences in the level of expression compared with smCBA, the SYN1 promoter also led to widespread labeling in cortical, forebrain, and brainstem neurons (Supplementary Fig. S1 online). Together, these data illustrate the ability of the rAAV-PHP.B capsid to promote widespread expression of EmGFP throughout the brain.

Twelve MiniPromoters driving EmGFP (Table 1) were delivered through IV injection in neonatal and/or adult mice. Reporter expression was assessed in the brain (Fig. 2 and Supplementary Fig. S2 online) and peripheral tissues (Supplementary Fig. S3 online) using an antibody against GFP. Ple385 (GPR6), Ple389 (ADORA2A), and Ple391 (DRD2) in the rAAV-PHP.B capsid demonstrated highly selective EmGFP expression in the striatum and low or absent expression in other regions such as cortex and thalamus (Fig. 2), compared to the other nine MiniPromoters which resulted in overall non-specific expression (Supplementary Fig. S2 online). Ple385 (GPR6) resulted in a high level of cellular expression across the basal ganglia: cell bodies expression in the caudate putamen (CP) and nucleus accumbens (ACB); axonal expression in the globus pallidus external (GPe) and internal (GPi), spreading to substantia nigra (SN); low level of pyramidal expression observed in the cortex, and an absence of expression in the thalamus. Ple389 (ADORA2A) resulted in highly cell-type-specific expression, confined exclusively to the cell bodies in the CP and ACB, and axons in the GPe and GPi, with a low level of cellular expression in the cortex, and expression in the thalamus almost absent. Ple391 (DRD2) resulted in high levels of expression in the striatum, specifically in the CP. However, cellular expression was also observed at a low level in the cortex, thalamus, and substantia nigra. Brain regions were identified and labeled using the Allen Mouse Brain Atlas (atlas.brain-map.org)³⁵ as a reference source.

Fig. 2.

Three MiniPromoters delivered by intravenous injection in mice showed expression in the striatum. MiniPromoters driving EmGFP in rAAV-PHP.B were delivered by intravenous injection in adult mice and harvested 4 weeks later. Ple385 (GPR6), Ple389 (ADORA2A), and Ple391 (DRD2) demonstrated enhanced EmGFP expression in the striatum and low to no expression in the cortex and thalamus. Considerable EmGFP labeling is also seen in striatofugal axons traveling through the globus pallidus (GP) and substantia nigra (SN) in the case of Ple385 and Ple391 or exclusively in the GP in animals injected with Ple389. Left to right: Low-power views of GFP immunostaining in sagittal sections of the mouse brain at the level of striatum and other basal ganglia nuclei (GP, SN), and medium—and high-power views of labeling in key on-target (striatum) and off-target regions (cortex, thalamus). CP, caudate putamen; ACB, nucleus accumbens; GPe, globus pallidus external; GPi, globus pallidus internal; SN, substantia nigra; EmGFP, emerald-green fluorescent protein; rAAV-PHP.B, recombinant adeno-associated virus packaged in capsid 9 variant PHB-B. Blue, DAPI; green, anti-GFP. Scale bars, 100 µm.

At the peripheral tissues, Ple385 (GPR6) resulted in low levels of expression in the retina, spinal cord, and liver, but none in the pancreas and heart (Supplementary Fig. S3 online). Ple389 (ADORA2A) was highly expressed in the liver, and a few cells were seen in the pancreas as a side-effect of the delivery method, which was also observed previously by other authors and by us following the delivery of different AAV serotypes^26,27,36. A low expression level was observed in the spinal cord, and no expression was found in the retina and heart (Supplementary Fig. S3 online). Finally, Ple391 (DRD2) resulted in a low level of expression in the retina, spinal cord and liver without substantial expression in the pancreas and heart (Supplementary Fig. S3 online).

To characterize D1- and D2-MSN specificity, we performed RNA in situ hybridization on the three most specific MiniPromoters: Ple385 (GPR6), Ple389 (ADORA2A), and Ple391 (DRD2) (Fig. 3), using RNA-probes for GFP, Drd1, and Drd2. A high level of co-labeling expression (70.25 ± 3.68%) of GFP and Drd1 mRNA was observed in Ple385 (GPR6) in comparison to the Drd2 probe (34.25 ± 2.76%), demonstrating greater specificity for D1Rs. In contrast, Ple389 (ADORA2A) and Ple391 (DRD2) resulted in higher co-localization of GFP neurons with the Drd2 probe, indicating more specificity for D2Rs. Ple389 (ADORA2A) resulted in 86.67 ± 2.51% co-labeling of GFP and Drd2 probe and 28.75 ± 2.31% co-labeling of GFP and Drd1. While Ple391 (DRD2) resulted in 71.13 ± 2.62% co-labeling GFP and Drd2 probe, and 44.75 ± 1.89% of co-labeling of GFP and Drd1. Based on these observations, we considered Ple389 (ADORA2A) the most promising MiniPromoter to preferentially target D2-MSNs.

Fig. 3.

Three MiniPromoters delivered by intravenous injection in mice are preferentially expressed in D1- and D2-MSN-specific markers in the striatum. MiniPromoters driving EmGFP in rAAV-PHP.B were delivered by intravenous injection in adult mice and were harvested 4 weeks later. High-power confocal images are shown for Ple385 (GPR6), Ple389 (ADORA2A), and Ple391 (DRD2) following RNA in situ hybridization analysis (left) and quantification (right) of EmGFP and Dopamine receptors D1 (Drd1) and D2 (Drd2) probes in the striatum. A high degree of co-labeling of EmGFP and Drd1 is seen with Ple385 (GPR6), while Ple389 (ADORA2A) and Ple391 (DRD2) demonstrated higher co-localization with Drd2 cells. Ple389 (ADORA2A)-positive striatal cells show the highest degree of Drd2 positive cells. EmGFP, emerald-green fluorescent protein; rAAV-PHP.B, recombinant adeno-associated virus packaged in capsid 9 variant PHB-B. ****p < 0.0001; Blue, DAPI; green, GFP probe; red, Drd1 probe or Drd2 probe. Scale bars, 100 µm.

Three MiniPromoters demonstrated enriched neuronal expression in the striatum when delivered intracerebroventricularly to mice

First, we evaluated the ability of the rAAV-PHP.B capsid to transduce neurons by ICV injection using the ubiquitous smCBA and pan-neuronal SYN1 promoters driving EmGFP (Supplementary Fig. S4 online). As observed after the IV injections, the smCBA and SYN1 promoters drove expression across the whole brain, including cortical, striatal, and hypothalamic regions, with the highest expression level in the cortex for both promoters (Supplementary Fig. S4 online)³⁰. As expected, EmGFP expression was also observed along the membrane of the lateral ventricles (Supplementary Fig. S4 online) at the level of the injection site, confirming the ICV delivery of the viral vectors. This pattern was also observed before by Galvan et al.³⁰, with a single unilateral ICV injection of AAV9-PHP.B SYN1-EmGFP in the mouse lateral ventricle. Overall, these data confirm the ability of the rAAV-PHP.B capsid to promote widespread transduction in the brain following ICV injection.

The transduction efficacy and specificity of seven MiniPromoters (Table 1, Fig. 4 and Supplementary Fig. S5 online) were then evaluated in adult mice by ICV injection. Figure 4 highlights the results of the top three striatal-specific MiniPromoters: Ple385 (GPR6), Ple389 (ADORA2A) and Ple391 (DRD2). Both Ple385 (GPR6) and Ple389 (ADORA2A) were highly expressed in the striatum, especially in the caudate putamen (CP) in both hemispheres. Low to absent EmGFP expression driven by these promoters was observed in the cortex and hypothalamus. Ple391 (DRD2) was also strongly expressed in the striatum, although some EmGFP labeling was also observed in the cortex and hypothalamus (Fig. 4). For all MiniPromoters, it was possible to observe expression along the injection tract in the cortex and ventricles. The four remaining MiniPromoters tested resulted in very low expression in the striatum and high off-target expression (Supplementary Fig. S5 online).

Fig. 4.

Three MiniPromoters delivered by intracerebroventricular injection in mice showed expression in the striatum. MiniPromoters, driving EmGFP in rAAV-PHP.B, were delivered by ICV injection into the lateral ventricle of adult mice and harvested 4 weeks later. Ple385 (GPR6), Ple389 (ADORA2A), and Ple391 (DRD2) demonstrated enhanced EmGFP expression in the striatum and low to no expression in the cortex and hypothalamus. Left to right: Low-power view of GFP immunostaining in a coronal section of the mouse brain at the level of the striatum and medium—and high-power views of key on-target (striatum) and off-target regions (cortex, hypothalamus). CP, Caudate putamen; HY, Hypothalamus; EmGFP, emerald-green fluorescent protein; rAAV-PHP.B, recombinant adeno-associated virus packaged in capsid 9 variant PHB-B. Green, anti-GFP. Scale bars, 1000 µm.

To further assess the positive expression in striatal MSNs, EmGFP-positive cells were co-labeled with a DARPP32 antibody (Fig. 5). DARPP32 is a regulatory protein found in the cytoplasm and nuclei of MSNs and colocalizes with D1R and D2R in the striatum³⁷. All three MiniPromoters showed similarly high levels of GFP-positive striatal cells that co-labeled with DARPP32: Ple385 (GPR6) 79.08 ± 3.39%; Ple389 (ADORA2A) 72.35 ± 3.24%; and Ple391 (DRD2) 80.14 ± 1.29% (Fig. 5). Together, these results show that all three MiniPromoters can transduce efficiently D1- and D2-MSNs of the striatum when delivered intracerebroventricularly in mice.

Fig. 5.

Three MiniPromoters delivered by ICV injection in mice showed co-labeling with a medium spiny neuron-specific marker in the striatum. MiniPromoters, driving EmGFP in rAAV-PHP.B, were delivered by ICV injection into the lateral ventricle of adult mice, and harvested 4 weeks later. High-power view (left) and quantification of co-localization (right) are shown for Ple385 (GPR6), Ple389 (ADORA2A), and Ple391 (DRD2) following double immunostaining of striatal cells for GFP and DARPP32, a marker for striatal MSNs. A considerable number of GFP-labelled cells in the striatum co-expressed DARPP32 immunoreactivity after ICV injections of the three MiniPromoters, confirming specific MSNs expression. EmGFP, emerald-green fluorescent protein; rAAV9-PHP.B, recombinant adeno-associated virus packaged in capsid 9 variant PHB-B. Green, anti-GFP; red, anti-DARPP32. Scale bars, 100 µm.

Ple389 (ADORA2A) expression is enriched in the striatum after intracerebroventricular injection in non-human primate

Based on the IV and ICV results in mice highlighted above, and the importance of ADORA2A for therapeutic approaches for brain disorders^16,18,38 and addictive drug-related behaviors³⁹ in the D2-MSNs, Ple389 (ADORA2A) was selected for evaluation by ICV in adult rhesus macaque monkeys. Although two monkeys received ICV injection of Ple389 (ADORA2A), only one animal exhibited substantial transduction of the rAAV-PHP.B and expression of GFP. The reason for the lack of transduction in the second animal is unclear. Therefore, data from a single monkey are presented.

We have previously demonstrated that ICV injections of rAAV-PHP.B SYN1-EmGFP in monkey, following the same protocol as in this study, results in pan-neuronal GFP expression throughout cortical and subcortical regions³⁰, indicating the capacity of the viral vector to reach large brain areas. As shown in Fig. 6, there was a considerable difference in the density of transduced pyramidal cortical neurons between Ple389 (ADORA2A) in comparison with the pan-neuronal control SYN1³⁰.

Fig. 6.

Ple389 (ADORA2A) MiniPromoter and SYN1 pan-neuronal promoter delivered by ICV injection in non-human primate showed differing expression in the cortex. Ple389 (ADORA2A) and SYN1, driving EmGFP in rAAV-PHP.B, were delivered by ICV injection to the left lateral ventricle of the monkey’s brain, and harvested 4 weeks later. (A–D). Immunoperoxidase-stained sections that compare the overall pattern of GFP labeling throughout the anterior part of the telencephalon and the cerebral cortex between an animal injected with the Ple389 (ADORA2A) (A, C) or the pan-neuronal marker SYN1 (B, D). EmGFP, emerald-green fluorescent protein; CD, Caudate nucleus; CTX, cortex. Scale bars: A: 10 mm (same for B), C: 1 mm.

Figure 7 shows the expression of Ple389 (ADORA2A), as revealed with GFP immunoperoxidase staining, in the monkey basal ganglia after ICV injection of rAAV-PHP.B Ple389(ADORA2A)-EmGFP. Note the dense neuropil and cellular labeling in both the left (Fig. 7A, A’) and right (Fig. 7B, B’) caudate nucleus (CD). When examined at high magnification, the density and size of the GFP-positive cell bodies were reminiscent of striatal MSNs (Fig. 7A’). In some cases, the staining is heterogeneous and includes patches of lower immunoreactivity (box in Fig. 7B’ and higher magnification in Fig. 7B’’). Ple389 expression at a more posterior level of the left caudate is shown in Fig. 7C, while axonal labeling in the globus pallidus (Fig. 7D, D’) and the substantia nigra (Fig. 7E, E’, and E’’), which likely originates from transduced CD neurons, is also depicted. These observations suggest that Ple389 (ADORA2A) also labels MSNs in the monkey striatum.

Fig. 7.

Ple389 (ADORA2A) MiniPromoter delivered by ICV injection in non-human primate showed expression in the striatum. Ple389 (ADORA2A), driving EmGFP in rAAV-PHP.B, was delivered by ICV injection to the left lateral ventricle of a monkey, and harvested 4 weeks later. Immunoperoxidase labeling at various levels of the telencephalon and midbrain. (A–B) Low-power views of GFP expression in the head of the caudate nucleus (CD) from the left and right hemispheres, respectively. (A’–B’) High- and medium-power views of the CD from A and B, respectively. (B”) Higher magnification of the outlined box in B’. (C–E) Low-power views showing cellular and neuropil GFP expression in more posterior levels of the left CD (C, D, E) and axonal labeling in the internal and external globus pallidus (GPi, GPe, D), and substantia nigra (E). (D’) Medium magnification view of the area of GPi framed in D. (E’ and E”) Medium- and high-power views of fibers labeling in the substantia nigra (SN). The areas within the black rectangle in E are depicted at higher magnification in E’ (higher magnification of E’ shown in E’’). EmGFP, emerald-green fluorescent protein; AC, Anterior commissure; Amy, Amygdala; CD, Caudate nucleus; GPe, Globus pallidus, external; GPi, Globus pallidus, internal; Hipp, Hippocampus; IC, internal capsule; Pu, Putamen; SN, substantia nigra; Th, thalamus. Scale bars: 7A: 10 mm (valid for A, B, C, D, E), A’: 200 µm, B’: 5 mm, B’’: 1 mm, D’: 0.5 mm (valid for E’’), E’: 2.5 mm.

Discussion

Cell- or tissue-specific targeting of transgene expression or drug delivery is crucial for the safety and efficacy of many therapeutic approaches. In this study, we developed a novel human-DNA MiniPromoter, Ple389 (ADORA2A), which, in combination with rAAV-PHP.B capsid delivered intravenously and intracerebroventricularly, demonstrated specific and robust expression in striatal MSNs of mice and non-human primates.

Co-localization of DARPP32 and EmGFP in the striatum of mice following ICV injection of Ple389 (ADORA2A) demonstrated that most labeled neurons were, indeed, MSNs³⁷. In addition, the in situ labeling for Drd1 or Drd2 in mice indicated that Ple389 (ADORA2A) mainly labeled D2-MSNs after IV injection, which is consistent with endogenous co-localization of ADORA2A and D2R in the striatopallidal MSNs⁴⁰.

In the direct pathway, D1-MSNs project preferentially to the substantia nigra pars reticulata (SNr), whereas in the indirect pathway, D2-MSNs project preferentially to the lateral part of the globus pallidus (LGP)¹⁴. Consistent with that, it has been reported that striatonigral neurons express mostly D1Rs and at very low levels in D2Rs^7,11, while D2Rs are expressed in the striatopallidal neurons¹¹. These observations are consistent with Ple389 (ADORA2A) labeling being preferentially expressed in D2-MSNs and suggest that Ple389 (ADORA2A) can be employed to target D2-MSNs in a proof-of-concept study and may be useful as a tool for targeted therapeutic approaches.

Expression of Ple389 (ADORA2A) in non-human primates also showed robust labeling in the striatum, especially in the caudate nucleus, with axonal labeling in the globus pallidus and the substantia nigra. Although co-localization studies could not be performed in the monkey striatal tissue, the morphology of transduced neurons in the caudate nucleus and the axonal labeling in GPe, GPi, and SNr support expression in striatal MSNs. The relative prevalence of expression in D2 vs D1 MSNs remains to be established. Regardless, Ple389 (ADORA2A) is the best MiniPromoter designed in size to target MSNs in the striatum for non-human primates. A future study to confirm and strengthen the specificity and efficacy of the promoter in a larger cohort of animals is desirable.

Given that ADORA2A plays an important role in various neurodegenerative diseases, such as Parkinson^12,15,16, Huntington^17,18, and Alzheimer⁴¹ disease, selective targeting of these cells is useful for basic, pre-clinical, and clinical research. ADORA2A receptors are highly expressed in the striatopallidal MSNs of the indirect pathway, in which they antagonistically interact with D2Rs and modulate glutamatergic signaling^40,42,43. Stimulation of ADORA2A receptors inhibits D2 receptor-mediated locomotor activation, while blockade of ADORA2A receptors increases locomotor activation via D2R^42,43. Therefore, interactions between ADORA2A and D2R may be employed in novel strategies for the treatment of dopamine-related diseases, such as Parkinson Disease, schizophrenia, drug addiction, and attention deficit hyperactivity disorder⁴³. Moreover, in 2019, the U.S. Food and Drug Administration approved the ADORA2A antagonist Nourianz® (istradefylline) to treat Parkinson Disease⁴⁴, demonstrating the importance of this gene as a therapeutic target for PD and related disorders. Gene therapy approaches employing a MiniPromoter to specifically target ADORA2A neurons may improve efficiency and decrease or abolish unwanted off-target side effects.

Although Ple389 (ADORA2A) was the MiniPromoter that drove the highest expression in D2-MSN intravenously (86.67 ± 2.51%), Ple385 (GPR6) and Ple391 (DRD2) also exhibited a specific and strong expression in the MSNs following IV and ICV injection in mouse. Ple391 (DRD2) demonstrated substantial co-localization of EmGFP and D2-MSNs (71.13 ± 2.62%). In contrast, Ple385 (GPR6) demonstrated more specificity for D1-MSNs (70.25 ± 3.68%), despite previous evidence that GPR6 is predominantly expressed in striatopallidal neurons and co-localizes with D2R MSNs⁴⁵ with very low expression in the striatonigral neurons and other central and peripheral nervous system regions⁴⁶. Therefore, it was unexpected to find a higher level of transduced Ple385 (GPR6) labeling in D1-MSNs than in D2-MSNs. We hypothesize that the CRRs identified manually increased the expression of GPR6 in D1- over D2-MSNs, suggesting that Ple385 (GPR6) may potentially be considered a MiniPromoter to preferentially target D1-MSNs.

Regarding the other MiniPromoters designed in this study, they failed to induce expression in the brain or show either an off-target or widespread endogenous expression, suggesting that either the parameters chosen to identify the candidate CRRs or the genomic coordinates selected require additional bioinformatic refinement to make these MiniPromoters strong and specific.

An example of refinement of the design of MiniPromoters is shown in the comparison between Ple355 (ADORA2A) and Ple389 (ADORA2A). While the manually designed Ple355 (ADORA2A) displayed widespread brain expression after both IV and ICV injections in mice, the On-Target designed Ple389 (ADORA2A) exhibited enriched expression in the striatum. These observations were also described previously in a preliminary characterization of both MiniPromoters²⁸.

The generation of human MiniPromoters is an important development compared to traditional studies that use ubiquitous promoters, such as the cytomegalovirus enhancer and promoter (CMV), the chicken beta-actin (CBA), and the CAG, which is a fusion of the CMV enhancer and CBA promoter⁴⁷. In a recent study, about 45% of rAAVs in clinical studies used these three promoters to drive gene expression of interest, and few of them used tissue-specific but still wide promoters such as albumin and synapsin⁴⁸, showing solid evidence that compact cell- tissue-specific promoters could be advantageous for rAAV gene therapies. Besides addressing the limited DNA payload capacity of the rAAV vector³, our tools also allow the generation of cell- and tissue-specific small constructs, potentially making these more portable and usable in many applications, especially therapeutic approaches. A study showed that smaller promoters increase the chance of delivering large cargo DNA and improve packaging efficiency and vector titers⁴⁹.

Compared to the GENSAT project, a Gene Expression Nervous System Atlas that generates reporter gene expression using engineered bacterial artificial chromosome (BACs) of about 100–200 kb driving EGEP regionally in CNS cell types⁵⁰, our MiniPromoters are generated based on CRRs that control gene expression in specific cells and tissues. Such specific expression patterns within the brain tend to be conserved between species, especially humans and mice⁵¹, which makes our tool more specific and comparable between species. Although BAC-Cre transgene lines can give cell-specific expression, the regulation of Cre recombinase expression may suffer from chromosomal effects due to the random site of integration⁵². Moreover, unexpected or incomplete expression of Cre can be anticipated due to transient genes that are expressed during development but are limited in the adult brain and the variable numbers of gene copies^52,53. Our tool identifies genes specifically expressed in the adult brain regions or cell types of therapeutic interest and could improve Cre-dependent transgenic line generation by producing more specificity of the floxed constructs into the brain and strengthening recombinase.

The MiniPromoters characterized in this study could have broad utility, such as brain-directed delivery of molecules such as siRNA, cre recombinase, fluorescent reports, and research proteins, besides being used in flow-sorting experiments to enrich or exclude cells of specific neural types. Ultimately, the greatest purpose of generating specific MiniPromoters is to provide a tool that is specific for therapeutic gene delivery into the human brain.

Still, some of the MiniPromoters designed in this study, such as Ple392 (PDYN) and Ple393 (PENK), showed off-target GFP expression when delivered intravenously and intracerebroventricularly in mice and, therefore, require computational refinement to be comparable to the Cre molecular system available in the literature⁵⁴.

Other groups have also developed small promoters for cell type-specific targeting in the striatum. In a recent study, the short promoters, G88P2 (1395 bp) and G88P7 (896 bp), derived from the GPR88 gene, were enveloped in mutant AAV8 capsid, induced robust gene expression in D1-MSNs in mice and monkey, and modulated the direct pathway neurons when combined with a chemogenetic effector in monkey²¹. In contrast, our MiniPromoter, Ple379 (GPR88) with 516 bp, enveloped into the rAAV9-PHP.B capsid to ensure widespread transduction of brain cells⁵⁵, exhibited widespread brain labeling, including the striatum after IV injection in mice, but very little expression after ICV injections except for off-target expression in the Lateral Septal nucleus (LS). In comparison with the two short promoters designed by Chen et al.²¹, Ple379 (GPR88) is smaller and was designed manually by identification of one promoter downstream in the genomic coordinates, while G88P2 and G88P7 were designed more upstream in the genomic coordinates²¹, which may be one of the factors that contribute to the success of the two short promoters over our Ple379 (GPR88).

Moreover, selecting an rAAV capsid can also determine the efficiency of the transduction and may vary according to the delivery method and the target cell surface receptor². Here, we chose the rAAV-PHP.B capsid, a variant of rAAV9 capsid, due to its enhanced transduction efficiency for systemic CNS transduction over rAAV9 in adult C57BL/6J mice^29,55. As seen in previous studies, the rAAV-PHP.B capsid is suitable for either IV²⁸ or ICV injections³⁰ and promotes efficient transduction of neurons in the brain. Additionally, due to its wide transduction, PHP.B allowed us to evaluate the maximal off-target expression of our system, demonstrating that the three MiniPromoters (Ple385 (GPR6), Ple389 (ADORA2A), and Ple391 (DRD2)) are specific. This finding suggests that using these MiniPromoters combined with a capsid that further narrows the transduction would achieve even greater specificity.

Regarding off-target gene expression observed in areas of the body that are not expected based on the endogenous gene, it is worth considering the liver off-target GFP expression after intravenous injection observed for us in this study and described in previous ones^26,27. Of the three main MiniPromoters evaluated here, we observed a non-homogenous GFP expression in the liver, which was higher when delivering Ple389 (ADORA2A). Liver side effects of intravenously delivered rAAV serotypes are reported in the literature³⁶. However, GFP expression is significantly lower with the AAV9-PHP.B capsid than other AAV serotypes³⁶. Moreover, ADORA2A is expressed in the liver and plays an important role in metabolic regulation in mice⁵⁶. Thus, we find that the liver expression observed in this study could be promoter dependent. A further investigation of liver expression driven by the MiniPromoters would be desirable. However, since this assessment is beyond the scope of this study, we did not pursue this further. Despite our goal to design and empirically test MiniPromoters in vivo for brain disorders, this tool could also benefit other diseases, such as hepatic cancer and inflammation.

In summary, our data describe three MiniPromoters that were efficiently expressed in striatal MSNs, with emphasis on Ple389 (ADORA2A), which displayed the highest MSNs labeling in mice and non-human primates, with a preferential co-localization into D2-MSN after IV injection in mice. These data suggest that Ple389 (ADORA2A) may be a robust candidate for developing targeted therapeutic approaches for neurologic disorders with causes related to an imbalance in activity between direct and indirect pathway MSNs. As proposed by Chen et al.²¹ our Ple389 (ADORA2A) has the potential to be studied in combination with a chemogenetic effector to modulate the indirect striatopallidal pathway in both Parkinson and Huntington disease patients to ameliorate the core motor symptoms of both diseases. Decreasing the activity of the abnormally overactive indirect pathway MSNs could help reduce the increased GABAergic basal ganglia outflow to the thalamus and promote movements in parkinsonian patients, while increasing the basal ganglia output by activation of indirect pathway MSNs could mediate thalamic inhibition and, hence, suppress abnormal movements in individuals with Huntington disease. Moreover, Ple389 (ADORA2A) could also be employed in electrophysiology and optogenetic techniques to modulate the activity of specific neurons. Inspired by Pettibone et al.⁵⁷, Ple389 (ADORA2A) could be used in combination with an A2a-Cre mouse line to drive specific Cre-dependent expression of opsin and monitor neuronal responses to light pulses in awake behaving animals. More specifically, our MiniPromoters could be used in optogenetic investigation of graft function and graft-to-host connectivity in Parkinson disease cellular models rather than usual human SYN1 promoter⁵⁸ and to evaluate GABA synaptic activity in MSNs to assess inhibitory activity in Huntington mouse models⁵⁹.

Conclusions

We designed twelve Pleiades MiniPromoters to selectively target striatal MSNs and demonstrated the efficiency of three of them (Ple385 (GPR6), Ple389 (ADORA2A), and Ple391 (DRD2)) to express EmGFP in striatal neurons in mice and showed the effectiveness of Ple389 (ADORA2A) in drive MSN expression in non-human primate. Additionally, we showed the specificity of Ple389 (ADORA2A) for Dopamine D2-Receptors and DARPP32 labeling in mice, which are important targets for studying and developing novel strategies for the treatment of dopamine-related diseases. Overall, these findings demonstrate the potential of MiniPromoters as a tool for gene therapy approaches for both Parkinson and Huntington disease treatment. We also suggest the employment of Ple389 (ADORA2A) as a specific promoter with a focus on modulating the indirect striatopallidal pathway in a pre-clinical study.

Methods

MiniPromoter design

MiniPromoters Ple355 (ADORA2A), Ple357 (DRD1), Ple358 and Ple359 (DRD2), Ple379 (GPR88), Ple385 (GPR6), and Ple386 (TAC1) were designed manually, as previously described²⁷. First, a genomic window was defined around each gene to focus on the identification of candidate CRRs. For example, for Ple385 (GPR6), the window spanned ~ 80 kb: upstream from the last relatively conserved block of nucleotides in mouse and rhesus macaque to a downstream insulator (i.e. hg19:chr6:110279239-110361652). Then, the promoters and enhancers of each gene were identified within their respective genomic windows by integrating various types of genomic evidence associated with CRRs and visualizing them in the UCSC Genome Browser⁶⁰. These included nascent transcription from CAGE^61,62 and GRO-seq⁶³, as well as native tracks available at the UCSC Genome Browser, such as DNaseI hypersensitivity clusters in 125 cell types⁶⁴, layered H3K27ac in 7 cell lines⁶⁴, transcription factor (TF) ChIP-seq clusters from 161 TFs^65,66, genome segmentations from ENCODE⁶⁷, and multi-species conservation⁶⁸. Next, the MiniPromoters were designed by placing the promoter of each gene at the 3’ end and subsequently adding enhancers. Upstream enhancers were added first, from more proximal (i.e. closer to the promoter) to more distal (i.e. farther from the promoter), followed by downstream enhancers (also from more proximal to more distal). However, when the gene was located on the reverse strand, downstream enhancers were added first, followed by upstream enhancers (in both cases, from more proximal to more distal). Finally, the MiniPromoter sequences were analyzed to detect the presence of undesired restriction sites using NEBcutter⁶⁹; however, none were detected.

MiniPromoters Ple389 (ADORA2A), Ple390 (DRD1), Ple391 (DRD2), Ple392 (PDYN), and Ple393 (PENK) were designed using OnTarget²⁸. First, ATAC-seq data (in bigWig format) of post-mortem human neurons of the accumbens nucleus and putamen⁷⁰, mapped to the hg19 genome assembly, was downloaded from GEO (accessions: GSM2546440, GSM2546463, GSM2546465, GSM2546489, GSM2546535). The files were reformatted from bigWig to bedGraph using bigWigToBedGraph⁷¹ and then processed one-by-one using MACS2 bdgpeakcall⁷² (version 2.1.4) with default parameters. The resulting peaks were pooled into a single file and remapped to the mm10 genome assembly using liftOver with the option “-minMatch” set to 0.1⁷³. Moreover, ChIP-seq peaks of histone marks H3K4me1, H3K4me3, H3K36me3, and of RNA polymerase II (in bigWig format) from 8 weeks-old mice (mm9 genome assembly) saline-treated accumbens nucleus samples⁷⁴ were downloaded from GEO (accessions: GSM1050343, GSM1050344, GSM1050345, GSM1050349, GSM1050350, GSM1050351, GSM1050355, GSM1050356, GSM1050357, GSM1050368, GSM1050369, GSM1050370). Similar to the ATAC-seq data, the bigWig files were reformatted into bedGraph format and processed using MACS2. The resulting peaks were pooled into individual files, one for each factor, and remapped to mm10. In addition, ChIP-seq peaks of histone mark H3K27ac of medium spiny neurons of the mouse striatum⁷⁵, mapped to mm9, were also downloaded from GEO (accession: GSM2230267) and remapped to mm10 using liftOver with default parameters. Then, OnTarget was employed to identify candidate CRRs by: (1) setting the genome assembly to mm10; (2) specifying the remapping of identified CRRs to hg19; (3) defining the genomic window for CRR identification based on distance (e.g. ± 100 kb around the mouse gene for ADORA2A, resulting in the genomic coordinates mm10:chr10:75216877-75434792); (4) uploading the processed ATAC-seq and ChIP-seq data from GEO as genomic evidence; and (5) setting the score threshold for CRR identification to a value that would yield the top 5% of nucleotides (e.g. ~ 0.555 for ADORA2A). The full details for Ple389 (ADORA2A) can be found in the OnTarget publication²⁸ and are available on the webserver through a dedicated page (http://ontarget.cmmt.ubc.ca/adora2example)²⁸.

Cloning and virus production

The smCBA promoter was used as a ubiquitous expression control, and the human synapsin (SYN1) promoter was used as a pan-neuronal expression control^26,76,77. The sequences for smCBA and SYN1 controls and all MiniPromoters were synthesized and incorporated in a plasmid backbone (GenScript, Inc., Piscataway, NJ). The synthetic smCBA and SYN1 containing plasmids were pEMS2075²⁷ and pEMS2156³⁰, respectively. Table 1 summarizes all MiniPromoters used in this work and their respective plasmid identifiers.

Synthetic plasmids were then cloned into the multiple cloning sites (AvrII, FseI, MluI, and AscI) of our ‘Plug and Play’ rAAV genome plasmid backbone (pEMS2131)^26,27. The plasmid features an intron (optimized chimeric; 173 bp) (Promega, Madison, MI, USA)⁷⁸; NotI flanked emerald green fluorescent protein (EmGFP) (720 bp)⁷⁹; AsiSI flanked removable woodchuck hepatitis virus posttranscriptional regulatory element (WPRE) mut6 (587 bp)⁸⁰; and SV40 polyA (222 bp) (Promega, Madison, MI, USA). Plasmids were propagated in E.coli SURE cells (Agilent Technologies, Santa Clara, CA, USA). DNA was prepared by QIAGEN Spin MiniPrep Kit (QIAGEN, Germantown, MD, USA); plasmids were demonstrated free of rearrangements by AhdI digest, inverted terminal repeats were verified via SmaI digest, and cloning sites verified by sequencing. Confirmed plasmids were sent to the University of Pennsylvania Vector Core (Philadelphia, PA, USA) for large-scale DNA amplification using the EndoFree Plasmid Mega Kit (QIAGEN, Hilden, Germany). Quality control on the plasmid preparation was done by SmaI, PvuII, and SnaBI digests, and confirmed plasmids were subsequently packaged into rAAV-PHP.B and/or rAAV9 capsids (See Table 1)^29,55,81.

The main viruses used in this study are: rAAV-PHP.B smCBA-EmGFP-WPRE (pEMS2143)^26,27,30, rAAV-PHP.B SYN1-EmGFP-WPRE (pEMS2155)³⁰, rAAV-PHP.B Ple389(ADORA2A)-EmGFP-WPRE (pEMS2387), rAAV-PHP.B Ple391(DRD2)-EmGFP-WPRE (pEMS2388), and rAAV-PHP.B Ple385(GPR6)-EmGFP-WPRE (pEMS2383), and will be referred to by their promoter or Pleiades (Ple) MiniPromoter names: smCBA, SYN1, Ple389 (ADORA2A), Ple391 (DRD2), and Ple385 (GPR6), respectively.

Animals

A total of 81 mice (45 for intravenous (IV), 3 for intraparenchymal (IP), and 33 for intracerebroventricular (ICV) injections) and two rhesus macaque monkeys were used in this work. Adult mice (mixed sex C57BL/6J, Stock No:000664, The Jackson Laboratory (JAX), Bar Harbour, Maine, USA) and neonatal mice (B6129F1 hybrids produced as the first generation of crossing C57BL/6J (JAX stock No:000664) to 129S1/SvImJ (JAX stock No:002448), were housed in the pathogen-free Centre for Molecular Medicine and Therapeutics facility on a 7 am–7 pm light cycle, 20 ± 2 °C with 50 ± 5% relative humidity, food and water ad libitum. All procedures involving mice were in accordance with the guidelines of the Canadian Council on Animal Care for the Use of Experimental Animals and approved by the University of British Columbia Animal Care Committee (Protocols A17-0204, A17-0205, and A18-0117). Two adult rhesus macaque monkeys were housed in the Emory National Primate Research Center, with food and water ad libitum. Monkeys were screened for AAV9 seronegativity before being assigned to the project. All procedures followed the guidelines for animal use and welfare set by the National Institutes of Health (National Research Council) and were approved by the Institutional Animal Care and Use Committee (IACUC) of Emory University (Protocol no. 201700768). The authors complied with the ARRIVE guidelines.

Intravenous injections in neonatal and adult mice

Adult mice (C57BL/6J) were injected intravenously (IV) with 100 µl of rAAV-PHP.B virus (3 × 10¹³ GC/mL) at 6 weeks of age via the tail vein, using a 31-gauge needle on a 0.33 cc syringe (320440, BD, Franklin Lakes, NJ). The rAAV-PHP.B virus solution was diluted using phosphate-buffered saline (PBS) and 0.001% Pluronic acid (PBS + P). A minimum of three animals were injected per MiniPromoter. Neonatal mice were produced by timed pregnancies using crowded females, experienced studs, and plug-checking of females such that the day of birth could be accurately predicted. Postnatal (P) injections were performed into the superficial temporal vein of neonatal mice; a 31-gauge needle on a 0.33 cc syringe was inserted under the skin approximately 1–2 mm parallel to the vessel, and then advanced into the vein, 50 µl of rAAV9 virus (3 × 10¹³ GC/mL) was slowly injected. Table 1 details the virus information delivered through intravenous injection. Following injection, the pups were tattooed for identification and returned to their home cage with the dam and a nanny (companion female).

Intraparenchymal and intracerebroventricular injections in mice

Adult mice (C57BL/6J; three animals per MiniPromoter) were anesthetized with 4% isoflurane with 1% oxygen and then fixed in a stereotaxic apparatus. A single injection of buprenorphine (0.1 mg/kg s.c.) and bupivacaine (1 mg/kg s.c. infiltrated under the skin at the injection site) were administered for prophylactic analgesia. The hair on the head was shaved, and the area was cleaned with hibitane and alcohol prior to making a small incision in the skin to expose the skull. For intraparenchymal injection (IP), mice received bilateral injections of rAAV9 in PBS + P into the striatum using the coordinates of anterior–posterior + 0.8 mm, medial–lateral+/− 1.8 mm, and dorsal–ventral − 3.5 mm from Bregma⁸². The virus solution (3. × 10¹³ GC/mouse, in a total volume of 1 μl) was injected through small burr holes in the skull using a 10 μl Hamilton syringe coupled to a 32-gauge glass microneedle at a rate of 250 nL/min. For intracerebroventricular injection (ICV), mice received unilateral injections of virus solution into the left lateral ventricle using coordinates of -0.3 mm posterior, + 1.1 mm lateral and -3.5 mm ventral from Bregma⁸². The rAAV-PHP.B virus solutions (3. × 10¹¹ GC/mouse, in a total volume of 10 μl) were injected through small burr holes in the skull using a 10 μl Hamilton syringe coupled to a 32-gauge glass microneedle at a rate of 5 µL/min. Table 1 details the virus information delivered through IP and ICV injection. For both IP and ICV, the microneedle was left in position for 2 min before initiating the injection and left in place for an additional 5 min to allow diffusion of the virus solution. The needle was then slowly removed from the brain, and the incision was sutured. During the surgical procedure and recovery from anesthesia, the body temperature of the mice was monitored and controlled using a heating pad. The animals were monitored every 15 min until they woke up and were then transferred to a housing room. Mice were monitored for body weight and general health twice daily for 72 h post-surgery.

Intracerebroventricular injection in monkey

Each monkey underwent an MRI scan on the day of the surgical procedure before being brought to the surgery suite. The animal was initially sedated with ketamine (10 mg/kg, intra-muscular) before being transported to the imaging facility, where it was intubated and anesthetized with isoflurane (1–3%). The head of the animal was then fixed in a stereotaxic frame and brought to the scanner. During the MRI scan (~ 1 h), the isoflurane anesthesia was administered and monitored by the MRI anesthesia technician with the assistance of the veterinary staff, as needed. Once the MRI had been completed, the animal remained anesthetized with isoflurane in the stereotaxic frame and brought to the preoperative room in the surgery suite. The head of the animal was shaved, and the skin overlying the site of the operation was cleaned with betadine and alcohol. After retracting the skin and exposing the skull, a small hole (3–5 mm in diameter) was drilled in the skull, and a Hamilton microsyringe was lowered into the brain under stereotaxic guidance to the left lateral ventricle (18 mm anterior to the interaural line, 2 mm from midline and 17.5 mm ventral from the cortical surface). Once on target, the syringe was left in place for 5 min before beginning the injection. A total volume of 3 mL of virus was delivered in the left lateral ventricle over a period of 3 min. Each rhesus macaque monkey received 1 × 10¹³ GC (0.8 × 10¹² and 1.6 × 10¹² GC/kg for each monkey, respectively). The virus solution was diluted to the final titer using artificial cerebrospinal fluid (CMA Microdialysis, Kista, Sweden). The injection was conducted manually at an approximate rate of 1 mL/min. The needle was left in situ for 10 min before being slowly withdrawn. Once the injections had been completed, the surgical site was cleaned with sterile saline, and the skin was sutured. At the end of the procedure, the animals received post-surgical analgesics (buprenorphine or banamine, continued up to three days post-surgery) before being brought to their home cage, where their behavior and health status were monitored daily by the veterinarian staff.

Immunohistochemistry on mouse tissue

Mice from IV: Four weeks following injection, mice were given a lethal dose of avertin (2,2,2 tribromoethanol; Millipore Sigma, Burlington, MA, USA) injected intraperitoneally and perfused transcardially with 1 × PBS for 2 min and 4% paraformaldehyde for 10 min. The brain, spinal cord, eyes, heart, liver, pancreas, kidneys, and adrenal glands were harvested and post-fixed for 2 h at 4 °C.

Mice from ICV: Four weeks following injection, animals were injected with 100 µL heparin, anesthetized by intraperitoneal injection of a lethal dose of avertin, and terminally perfused through the ascending aorta with cold 3% paraformaldehyde in 0.1 M phosphate buffer at flow rate 5 mL/min for 10 min. After perfusion, the brains were kept within the skull for 24 h in a perfusion buffer (3% paraformaldehyde in 0.1 M phosphate buffer). Subsequently, brains were removed from the skull and kept at 4 °C in PBS until further processing.

Immunostaining: Prior to sectioning, the tissue was equilibrated in 30% sucrose in PBS for 24–48 h at 4 °C, after which it was embedded in optimal cutting temperature compound (Fisher, Pittsburgh, PA, USA), frozen on dry ice and then cut on a cryostat into 20 µm sagittal sections (IV) and 25 µm coronal sections (ICV). ICV sections were collected floating into PBS supplemented with 0.01% sodium azide and kept at 4 °C until further processing. Sections were permeabilized in 0.1% Triton X-100 in PBS, blocked with 5% normal donkey serum and 5% bovine serum albumin (BSA) at room temperature (RT), then incubated at 4 °C for 12–20 h in rabbit anti-GFP primary antibody (1:1000). All antibodies used are shown in Table 2. Immunofluorescent detection was achieved by 1 h incubation at RT with secondary antibodies appropriate to the primary antibodies used (Table 2).

Table 2.

Source and concentrations of antibodies and probes.

Antibody	Source	Catalog no.	Studies	Concentration used
Primary
Chicken anti-GFP	Aves Labs Inc. (Tigard, OR)	GFP-1020	Mice (IV)	1:500
Rabbit anti-GFP	ThermoFisher Scientific (Waltham, MA)	A11122	Mice (ICV); Monkey (ICV)	1:1000; 1:5000
Rat anti-DARPP32	R&D System (Minneapolis, MN)	MAB4230	Mice (ICV)	1:1000
Drd1 Probe	Bio-Techne (Minneapolis, MN)	Mm-Drd1-C3 (461901)	Mice (IV)
Drd2 Probe	Bio-Techne (Minneapolis, MN)	Mm-Drd2-C3 (406501)	Mice (IV)
GFP Probe	Bio-Techne (Minneapolis, MN)	GFP-O2-C2 (556431)	Mice (IV)
Secondary
Donkey anti-rabbit Alexa 488	ThermoFisher Scientific (Waltham, MA)	A21206	Mice (ICV)	1:1000
Donkey anti-rat Alexa 594	ThermoFisher Scientific (Waltham, MA)	A21209	Mice (ICV)	1:1000
Goat anti-chicken Alexa 488	ThermoFisher Scientific (Waltham, MA)	A11039	Mice (IV)	1:1000
Goat anti-rabbit Biotinylated	Vector Laboratories (Burlingame, CA)	BA-1000	Monkey	1:200

GFP, Green fluorescent protein; ICV, Intracerebroventricular; IV, intravenous; DARPP32, Dopamine- and cAMP-regulated neuronal phosphoprotein; Drd1, Dopamine receptor D1; Drd2, Dopamine receptor D2.

Double immunofluorescence procedures were used to characterize the proportion of GFP-immunoreactive neurons that were also DARPP32-positive in the ICV sections. Sections were incubated for 12–20 h in a mixture of anti-GFP (as indicated above) and rat anti-DARPP32 (1:1000) (Table 2) antibodies, which was followed by exposure to the secondary fluorescent antibodies (Table 2). DARPP32 antibody was chosen as a known marker for striatal MSNs⁸³.

All sections were then incubated at RT for 5 min in 4′,6-diamidino-2-phenylindole (DAPI; Millipore Sigma; 1:10,000). Sections were coverslipped using Depex fluorescence mounting medium (Electron Microscopy Sciences, Hatfield, PA, USA). Images were captured using an Olympus BX61 Fluorescence and Transmittance Wide Field Microscope (Olympus Life Sciences, Tokyo, Japan). The images were compiled using Adobe Illustrator software (Adobe Systems, San Jose, CA, USA).

RNA in situ hybridization (RISH)

To evaluate the RNA expression for Drd1 (mouse Drd1, ID: 461901), Drd2 (mouse Drd2, ID: 406501), and GFP (ID: 556431), Bio-techne ACD’s RNAscope Multiplex Fluorescent V2 Assay kit (Catalogue# 323100; Bio-techne ACD, Newark, CA, USA) was used according to the manufacturer’s instructions. Briefly, slides were post-fixed in 4% PFA for 90 min at RT, followed by dehydration in ethanol (ranging from 50 to 100%), target retrieval, and protease treatment. Slides were then hybridized with probes for 2 h at 40ºC, and then the signal amplified, and opal dyes developed by incubation at 40ºC. Following the final development, slides were incubated with DAPI, and coverslipped as above.

Immunohistochemistry of monkey tissue

After 29–34 survival days, animals were deeply anesthetized with pentobarbital (100 mg/kg) and perfusion-fixed with cold oxygenated Ringer solution (~ 300 ml) followed by 2.5 L of a mixture of 4% paraformaldehyde and 0.1% glutaraldehyde in phosphate buffer (0.01 M, pH 7.4). The brains were then removed from the skull, blocked in the coronal plane, and cut with a freezing microtome in 50 µm-thick serial coronal sections that were stored in an anti-freeze solution at − 20 °C until further processing.

Immunoperoxidase: Immunoperoxidase methodology was chosen in contrast to immunofluorescence because (1) this is a routine method that we use for GFP immunostaining in monkey tissue, and (2) because of the signal amplification steps, the avidin–biotin complex immunoperoxidase method is considered as one of the most sensitive immunostaining approaches which assure detection of low levels of antigen expression.

One out of 10 sections collected throughout the whole brain of each monkey were processed for GFP immunostaining. The tissue was first placed in a sodium borohydride (1% PBS) solution for 20 min. After thorough washes in PBS, sections were put into a pre-incubation solution (10% normal goat serum, 1% bovine serum albumin (BSA), and 0.3% Triton X-100) for 1 h at RT before being incubated with rabbit anti-GFP antibody (1:5000) (Table 2) for 24 h at RT. Sections were then washed in PBS and placed for 90 min at RT in a secondary antibody solution consisting of 1% normal goat serum, 1% BSA, biotinylated goat anti-rabbit IgGs (BA-1000, Vector Laboratories, 1:200) (Table 2), and 0.3% TritonX-100. Sections were washed thoroughly in PBS before incubating in an Avidin-Biotinylated-Complex (ABC) solution (Vector Laboratories, Burlingame, CA, USA) for 90 min. After this incubation, the tissue was rinsed twice with PBS and once with tris(hydroxymethyl)aminomethane (TRIS, 0.05 M, pH 7.6). Thereafter, the sections were placed in 0.025% 3-3-diaminobenzidine tetrahydrochloride (DAB; Millipore Sigma, Burlington, MA, USA), 0.01 M Imidazole (ThermoFisher Scientific, Waltham, MA, USA), and 0.006% H₂O₂ for 10 min. The sections were then mounted on slides, covered, and digitized with an Aperio Scanscope CS system (Leica, Buffalo Grove, IL, USA). Control sections were incubated in solutions from which the primary antibody was omitted.

Image processing

Images were processed using ImageJ, Photoshop (Adobe, San Jose, CA, USA), and Illustrator (Adobe, San Jose, CA, USA). For the mouse tissues, brightness, contrast, color balancing, and scaling adjustments were made uniformly across each image in ImageJ where necessary to improve visibility.

Data analysis of immunohistochemical material

For co-localization of EmGFP and Drd1 and Drd2 probes in the mouse striatum after IV injection, sections were stained with anti-GFP and anti-Drd1 and Drd2 probes, as described in the RISH methodology. For quantification of EmGFP and DARPP-32 co-labeling in mouse striatum after ICV injection, sections were stained with anti-GFP and anti-DARPP32 antibodies as described above. Using Stereo Investigator software (version 2017.1.01, MicroBrightField, Williston, VT, USA), the region of interest (ROI) for both IV and ICV sections was outlined at 10 × magnification on three to six sections from each of two to three injected mice using either Drd1, Drd2 or DARPP32 immunoreactivity to outline the striatum. A sampling grid (100 × 100 μm) and counting frame (100 × 100 μm) were set over the ROI. Using 10× magnification, GFP-positive cells at each sampling site were marked with a digital marker. The filter was subsequently switched to the red channel, and a second marker was placed on cells that were also positive for either Drd1, Drd2 or DARPP32. The percentage of double-labeled cells was calculated as the total number of GFP-positive cells that also expressed immunoreactivity for either Drd1, Drd2 or DARPP32 over the total counts of GFP-immunoreactive cells in the ROI.

Statistical analysis

Results are reported as mean ± SEM. Analysis and statistics were performed using GraphPad Prism software, Version 8 (GraphPad Software Inc., San Diego, CA, USA). An unpaired Student’s t-test was used to compare the mean percentage of GFP-positive cells that co-localized with either Drd1 or Drd2 for each MiniPromoter evaluated in the RISH analysis.

Author contributions

Conceptualization and Funding acquisition: E.M.S., B.R.L.; Supervision: E.M.S., B.R.L., Y.S., W.W.W.; Data curation, Formal Analysis and Software: O.F., R.A.F.; Investigation: A.M.G., A.J.K., T.L.P., A.G., A.M.H., S.L.L., G.L., A.Y., Y.S.; Methodology: O.F., A.Y.; Project administration and Resources: B.R.L., Y.S.; Visualization: A.M.G., A.J.K., T.L.P., A.Y., B.R.L., Y.S.; Writing—original draft: A.M.G., O.F.; Writing—review & editing: A.M.G., A.J.K., T.L.P., A.G., O.F., A.M.H., G.L., B.R.L., Y.S., W.W.W.

Data availability

Plasmids for cloning, controls, recommended MiniPromoters, and virus genomes, are available to the research community through the non-profit distributor Addgene (Cambridge, MA, USA) (www.addgene.org). The viral genome plasmid for cloning was rAAV2 backbone version 2, pEMS2131^26,27. For the smCBA control, the promoter and viral genome plasmids were pEMS2075²⁷ and pEMS2143^26,27,84, respectively. For the SYN1 control, the promoter and viral genome plasmids were pEMS2156 and pEMS2155, respectively³⁰. For MiniPromoter plasmid numbers, see Table 1.

Declarations

Competing interests

The authors declare no competing interests.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-024-79004-y.