For ETS design, we selected the binding sequence found in the mouse Gria2 mRNA (at position 172–192 in the UTR). This choice was made based on our previous experience (Gascon et al., 2014) as well as the observations from other groups (Dutta et al., 2013; Namkung et al., 2023) having demonstrated that Gria2 is a relevant target for miR-124 in rodents and humans. For PCS, we used the complementary miR-124 sequences. For mutated versions, we replaced three nucleotides of the binding sequence from CUU to GGG. These nucleotides base pair within the seed region (position 2–4) and seve. All sequences were shown in Table 1 as well as in Supplementary Figure S1A. All sequences were cloned into the 3’ UTR of a nuclear RFP (H2B-RFP) for easy quantification using cell cytometry.

Perfectly complementary sequences or endogenous sequences were cloned into a plasmid backbone containing a H2B-RFP using appropriate restriction enzymes. For miR-124 overexpression, 2 kb of the primary miR-124 were amplified from mouse genomic DNA using PCR and then cloned into a plasmid allowing the co-expression of H2B-GFP. All plasmids used here contained Tol2 sequences for the generation of stable cell lines (see below). The constructs encompassing the H2B-RFP and the different 3’UTR were amplified using PCR and cloned in an inverted orientation into a pAAV-DIO vector (Addgene #35507). Plasmids maps are shown in Supplementary Figure S4.

Conditional reporters AAVs were generated using pAAV-DIO plasmids described before. Synapsin-Cre AAVs were produced from Addgene plasmid #105540. Viral productions were carried out using standard protocols with slight modifications (Su et al., 2020). Briefly, HEK293T cells were transfected using the calcium phosphate method (pHelper, pRC2, pAAV; ratio 2:1:1). After 60–72 h, cells were harvested and AAV particles purified and concentrated using Takara dedicated kit (Takara, France). Viral productions were tittered using quantitative PCR (Takara, France). AAV9 serotypes of a titer around 10¹² viral genomes/mL were used in this report. The different vectors are listed in Table 2.

HEK293T cells were obtained from ATCC and cultured under standard conditions (DMEM-10% Fetal Calf serum). They were used to generate cell lines stably expressing miR-124 as well as the different reporters. For that, we used a previously described procedure (Kawakami and Noda, 2004). Briefly, the different constructs were cloned into a vector containing the Tol2 sequences for the medaka fish transposase. Then, these plasmids were co-transfected with a plasmid containing the transposase into HEK cells. After a clonal dilution, cells having stably incorporated the transgene were selected by fluorescence expression. The expression of the proper construct was verified using quantitative PCR.

Cell cultures were prepared from newborn CD1 mice (postnatal day 0). Briefly, animals were decapitated, the brains rapidly transferred into ice-cold Hank’s balanced salt solution (HBSS, Gibco, France) and the different brain dissected out. Then, they are dissociated mechanically, trypsinized and purified using percoll gradient centrifugation. Cells were plated onto poly-ornithine (Sigma, France) coated 6-wells plates and allowed to grow in Neurobasal medium (Gibco) with 2% B27 supplement (Gibco), 2 mm glutamate (Gibco), 1 mm sodium pyruvate (Gibco) and antibiotics (Gibco). Cell density at plating was 3 × 10⁵ cells/well Half of the medium was replaced twice a week.

In all experiments with cell lines, we used Lipofectamine 3000 (Thermo) as transfection reagent following manufacturer’s recommendations. For primary cultures, AAVs particles (2 μL/well, 1 μL from the reporter AAV and 1 from the Cre AAV, representing around 10¹⁰ genomes/culture) were diluted in 50 μL of culture medium and then added to the cells 1 day after plating (DIV1). Neurons were analyzed 10 days after viral infection.

Cells were gently detached using trypsin and rinsed in PBS. After removal of cell clumps (70 μm cell strainer), individual cells were resuspended into 500 μL of PBS-0.1% BSA. The cells were analyzed on a cytoFlex device (Beckman Coulter) and data were analyzed with the CytoExpert software (Beckman Coulter). For FACS, infected neurons were sorted using a MoFlo Astrios EQ device (Beckman Coulter). Negative cells (either not transfected or transduced) were used to set up the gates.

Total RNA containing miRNAs and mRNAs was then extracted from these samples using NucleoSpin miRNA kit (Machery Nagel) following manufacturer’s instructions. RNA extraction protocol includes an on-column DNase treatment (30 min). Total RNA was quantified using a NanoDrop (Fisher Scientific).

For miRNA quantification, we first performed reverse transcription using TaqMan advanced miRNA cDNA synthesis kit (Applied Biosystems, France) according to the protocol provided by the manufacturer. Abundance of miRNAs in the samples was measured using custom TaqMan probes (Applied Biosystems, United States) and a QuantStudio 7 PCR thermocycler (Applied Biosystems, France). Relative quantification was carried using the ΔΔCt method. References genes were GAPDH (Figure 1C), miR-92 (Supplementary Figure S2), miR-16 and miR-101 (Figure 3A), and miR-9 (Figure 3D). For quantification of miR-124 in FACS sorted cerebellar neurons, we did not use miR-16 or miR-101 because their expression is much lower than miR-124 and could not be consistently amplified in these samples. miR-9, a neuronal enriched miRNA, was preferred as it was the miRNA whose levels were closer to miR-124 in these cultures. All qPCR assays (including for AAV tittering) are commercially available and their references indicated in Table 3.

Because of the limited sample size in our experiments (n = 3 or 4), we could not perform normality checks. To avoid using low-power non-parametric tests (i.e., Kruskal Wallis), we transformed our data using a relative normalization method. Instead of comparing the absolute fluorescence intensity, we calculated the percentage of fluorescence in each experiment taking the RFP control sample as reference (whose intensity was aribitrarily set to 100). For the detection of significant differences among groups, we used therefore ANOVA analysis followed by post-hoc tests. Data is presented as means ± SEM unless indicated. All statistical analysis was performed using Prism GraphPad software (version 9). Details of tests and n for each experiment are provided in figure legends.

We generated and systematically tested in vitro different fluorescent reporter designs for miR-124. Our constructs encompass a nuclear RFP and different numbers (1 or 2) and types (perfectly complementary or sequences found in endogenous targets) of miR-124 target sequences (Figure 1A; Supplementary Figure S4). We first generated stable cell lines containing such reporters and evaluated how miR-124 expression influenced their activity. For that, we transiently transfected these different lines with a plasmid driving the expression of miR-124 and GFP and use untransfected cells as a reference for RFP intensity. As expected, in cells transfected with this plasmid, we did not observe any change in RFP fluorescence in GFP⁻ cells (Supplementary Figure S1A). A significant reduction of RFP fluorescence was already observed 2 days after transfection in cells expressing the perfectly complementary sequences (PCS). Silencing was exacerbated at 6 days post-transfection leading to a near complete extinction of fluorescence and confirming previous reports. The effect was similar in cells containing either one or two PCS.

In sharp contrast, reporters bearing endogenous sequences (ETS) exhibit more modest decreases in fluorescence and a slightly additive effect of increasing the number of target sequences (Figure 1B). This moderate effect might reflect a regulation comparable to that of endogenous targets.

To further understand the effect of miR-124 on the different reporters, we FACS sorted transfected cells and analyzed reporter expression at the mRNA level. Our analysis (Figure 1C; Supplementary Figure S1B) revealed striking differences of miR-124 effect on the reporters bearing ETS and PCS. Levels of RFP transcript were similar in GFP⁻ and GFP⁺ cells in the lines containing ETS reporters. Conversely, a significant reduction of RFP mRNA was found in lines bearing PCS. Together with our observations on fluorescence intensity, those results suggest that translation is inhibited for ETS reporters whereas both transcript degradation and translation blocking operate to silence PCS reporters. Interestingly, despite the fact that miR-124 levels in GFP⁺ cells increased along with time (Supplementary Figure S1C), there is no change in the abundance of RFP transcripts neither in ETS nor in PCS-bearing reporters from 2 to 6 days indicating a different time course between mRNA and protein regulation. These results indicate that ETS and PCS can be useful to estimate miR-124 expression at the protein and the mRNA level, respectively.

Since the final aim of reporters is uncovering different levels of miR-124 across different cell types in vivo, we explored whether our different designs can detect different miR-124 contents in vitro (Figure 2A). We engineered HEK cells, endogenously devoid of miR-124 (Visvanathan et al., 2007; Gascon et al., 2014; Xue et al., 2016), to stably express different amounts of this miRNA. We selected two cell lines displaying moderate but different levels of miR-124 to approach more physiological conditions (Supplementary Figure S2A).

We then analyzed the reporter activity of our different constructs in these cell lines. To avoid the intrinsic variability associated with transient transfection (in terms of copy number in each cell and plasmid dilution), we constructed dual reporter plasmids (Supplementary Figure S4) containing not only a miR-124 reporter but also a reference reporter (CFP) that serves as internal reference (Supplementary Figure S4). Using flow cytometry, we quantified the intensity of both RFP and CFP fluorescence. The RFP/CFP ratio was calculated as a reliable output of miR-124 effects on the RFP reporter. We observe that PCS containing reporters are subjected to strong inhibition (Figure 2B). As before, we found an almost identical silencing independently of the cell line and the number of binding sequences. This suggest that PCS reporter are extremely sensitive even to moderate miR-124 levels. ETS-containing reporters, in turn, exhibit a more tunable regulation and a silencing proportional to the number of sequences present in the reporter (Figure 2C). As a control, transfection of all these constructs in regular HEK cells resulted in no modification of RFP/CFP ratio (Supplementary Figure S2). Additionally, mutating 3 nucleotides of the seed region in 1× ETS reporter abrogate the regulatory effect of miR-124 arguing for the specificity of our reporters (Supplementary Figure S2). These observations together indicate that reporters containing ETS might be a more appropriate strategy to measure miR-124 activity in vivo.

We next tested one of our ETS designs in a more physiological setting, primary neuronal cultures from the mouse hippocampus, cortex, and cerebellum. We selected them because, as shown in Figure 3A, cultured neurons from these brain regions express different endogenous levels of miR-124 but not of other miRNAs.

To introduce different reporters in primary neurons, we used adeno-associated vectors (AAVs) as previous work have shown their efficiency and lack of toxicity (Malik et al., 2012). We constructed three Flex conditional AAVs containing, respectively, a control reporter (no miR-124 binding sequence, AAV-RFP), a reporter bearing a single ETS (AAV-RFP-miR-124) and a mutated version (3 nucleotide mismatch in the seed region) of the reporter (AAV-RFP-miR-124mut). We transduced different neuronal cultures with the reporter AAVs along with a synapsin-Cre-GFP AAV. This strategy enables not only to drive the reporter express in similar neuronal subsets in the cultures but also to have a reference fluorescent protein from the Cre virus. We observe that AAVs transduce primary cultures with a high efficiency (ranging from 30–60% of hippocampal cells, 15–30% of cortical neurons, and 50–75% of cerebellar culture, Supplementary Figure S3A). As expected, the GFP fluorescence and the proportion of GFP+ cells does not differ among the different conditions confirming that reporter proteins are not subjected to any unspecific silencing (Supplementary Figures S3A,B). In contrast, compared to the control AAV, neurons transduced with miR-124 reporter show a marked reduction in the proportion of RFP⁺ cells (Supplementary Figure S3A) as well as in the intensity fluorescence (Supplementary Figure S3C). The intensity of such decrease varies among the brain structures analyzed. Thus, cerebellar neurons exhibit the highest silencing and hippocampal cultures are less inhibited (Figure 3B). Mutation in the miR-124 binding sequence abolishes such effect arguing for the specificity of our results. To further support these observations, we also calculated the RFP/GFP ration as a more robust quantitative measure of reporter fluorescence (Figure 3B). In the same line, we detected significant changes only in neurons transduced with the miR-124 reporter AAV.

To confirm these observations and to avoid the variability among different experiments, we sought to analyze the activity of our reporters in neurons from the same culture. We reasoned that, if our reporter fluorescence is negatively correlated to miR-124 contents, neurons displaying high RFP levels (RFP^high) might contain lower amounts of miR-124 relative to those in which RFP silencing is stronger (RFP^low). Conversely, in neurons transduced with the control or mutated reporter, miR-124 levels should be no correlated to fluorescence intensity. We therefore transduced cerebellar neurons (those showing highest miR-124 expression and the highest transduction efficiency), carried out FACS sorting to isolate RFP^high and RFP^low neurons and measured miR-124 in these different populations. We first confirmed that, in neurons transduced with the miR-124 reporter, there was an obvious left-shift in the fluorescence distribution that resulted in a net decrease of RFP^high cells (Supplementary Figure S3C). Regarding miR-124 contents, neurons having received the control or mutated construct show similar miR-124 contents irrespective of the fluorescence intensity (Figure 3C). In contrast, in those transduced with miR-124 reporter, RFP^low neurons exhibit a significant enrichment in miR-124 levels compared to RFP^high neurons suggesting that our reporter is subjected to miRNA silencing and its levels inversely correlated to miR-124 (Figure 3B). Finally, using linear regression, we sought to confirm the inverse correlation of reporter fluorescence and miR-124 contents in FACS-sorted cerebellar neurons. For that, we plotted the mean RFP fluorescence and relative miR-124 levels in each condition. As shown in Figure 3D, linear correlation was remarkably good for the miR-124 reporter (R² = 0.6584) but rather poor for the control RFP reporter as well as in the mutated reporter (R² = 0.01877 and R² = 0.004265, respectively). More importantly, the slope analysis revealed a significant difference from 0 only for miR-124 reporter (p = 0.01255) but not the control reporters (p = 0.8955 and 0.9022). These findings provide further support for the inverse relationship between miR-124 abundance and reporter fluorescence. Overall, our in vitro experiments support the notion that optimized miR-124 reporters could be useful for in vivo applications.