Abstract
Heightened juvenile cortical plasticity declines into adulthood, posing a challenge for functional recovery following brain injury or disease. A network of inhibition is critical for regulating plasticity in adulthood, yet contributions of interneuron types other than parvalbumin (PV) interneurons have been underexplored. Here we show Lypd6, an endogenous positive modulator of nicotinic acetylcholine receptors (nAChRs), as a specific molecular target in somatostatin (SST) interneurons for reactivating cortical plasticity in adulthood. Selective overexpression of Lypd6 in adult SST interneurons reactivates plasticity through the α2 subtype of nAChR by rapidly activating SST interneurons and in turn inhibiting sub-population of PV interneurons, a key early trigger of the juvenile form of plasticity. Chemogenetic activation of SST interneurons confirms the causal role of SST interneuron activity in reactivating plasticity. Collectively, we identify the Lypd6-nAChRα2 system and associated SST-PV disinhibitory microcircuit as the novel SST interneuron-specific targets for reactivating plasticity in adulthood. This provides potential insights into therapeutic strategies against disorders where recovery is limited due to diminished plasticity in adulthood, such as amblyopia and psychiatric disorders characterized by deficits in SST interneurons.
Introduction
Experience-dependent brain plasticity is heightened during juvenile critical periods but this declines into adulthood, which poses a major challenge to functional recovery following injury or disease later in life 1,2. A prevailing concept of therapeutic strategy to support restoration of function from such enduring debilitating conditions is to reactivate juvenile-like levels of heightened plasticity in the adult brain. One of the best-studied models of critical period plasticity is ocular dominance plasticity – the enduring loss of responsiveness in primary visual cortex (V1) to an eye deprived of vision 3 that results in amblyopia, a disorder of sight with limited treatment options in adulthood that affects 2–4% of the human population. The ocular dominance plasticity model has long served to not only to advance our understanding of the fundamental regulatory mechanisms of critical periods, but also importantly to facilitate the discovery of novel targets to achieve reactivation of visual plasticity in adulthood.
Over the past decade, with the advent of gene targeting in mice 4-7, the network of cortical inhibition was elucidated as one of the critical mechanisms for regulation of visual plasticity. The developmental initiation of the critical period can be accelerated or delayed by genetically or pharmacologically altering GABAergic inhibition in visual cortex 8-16. Following critical period closure in the adult cortex, plasticity can be induced through pharmacological suppression of GABAA receptors 17 or through the transplantation of GABAergic precursors derived from the medial ganglionic eminence (MGE) 18-20. To date, the role of GABAergic signaling in cortical plasticity has largely focused on parvalbumin (PV) interneurons 15,18,21-23, where a number of known molecular mediators of plasticity appear to converge including orthodenticle homeobox protein 2 (otx2) 15,24 and perineuronal nets 23. A recent report has shown that PV interneuron activity is reduced during the initial 24 hours following visual deprivation only during the juvenile critical period but not in adulthood. Mimicking this early reduction in firing using chemogenetic means in adult animals led to reactivation of plasticity, suggesting that PV interneuron activity may be instructive for plasticity in visual cortex 22. While their importance for plasticity in the visual cortex is undeniable, PV interneurons represent only one member of a diverse group of cortical GABAergic interneurons 25. The role of other major subclasses of GABAergic interneurons, such as those expressing the peptide somatostatin (SST) that represent nearly a third of the cortical GABAergic interneurons, had largely been unquestioned until recently 19,26. Moreover, the existence of molecular and circuit mechanisms underlying SST interneuron-mediated regulation of cortical plasticity remains completely unknown.
SST interneurons are promising targets for reactivating plasticity in the adult brain, as they are ideally situated to integrate multiple inputs including bottom-up sensory signals 27, or neuromodulation 28 induced by locomotion 29-33 or by top-down regulation 34-36. Moreover, SST interneurons robustly innervate local PV interneurons, which places them in an ideal position for driving the rapid inhibition of PV interneurons that is causal to the juvenile form of plasticity. Neuromodulation of SST interneurons, by nicotinic signaling that occurs in several brain regions 28,37-39, may represent one potential means of restoring plasticity in the adult brain. Our previous study showed that an endogenous inhibitor of nicotinic Acetylcholine Receptors (nAChRs) known as Lynx1 is enriched in PV interneurons and actively limits plasticity in V1 following the critical period 40. The Lynx family of proteins are GPI-anchored membrane proteins and a have unique toxin-like protein structure, which enables them to bind to the extracellular face of nAChRs expressed in the same cells and regulate their signaling 41. Our recent study further showed that another closely related member of the Lynx family known as Lypd6 is enriched in V1 SST interneurons in 42. Interestingly, in direct contrast to the action of Lynx1, Lypd6 has been shown to potentiate calcium currents through nicotinic receptors 43,44. Here we sought to elucidate the potential for nicotinic modulation by Lypd6 in SST interneurons as regulators of ocular dominance plasticity. This work provides the first molecular mechanism specifically associated with SST interneurons for reactivating plasticity in adult visual cortex.
Results
Neuronal overexpression of Lypd6 prolongs ocular dominance plasticity into adulthood
First, we compared the expression profiles of Lypd6 between adult V1, where plasticity is limited, versus that of juvenile critical period. We performed in situ hybridization to label Lypd6 mRNA in V1 of mice at P28 (critical period) and >P60 (adult). We found that Lypd6 expression is significantly lower in the adulthood compared to juvenile period, which inversely correlates with the extent of ocular dominance plasticity (Fig. 1a, 1b). While Lypd6 expression was even higher at P18 (pre-critical period) (p18: 244.91 +/-19.28 normalized to p60 data in Fig. 1a, p<0.05 vs P28 or adult), interpretation of Lypd6 contribution to nicotinic signaling during the pre-critical period is confounded by various additional developmental changes that could impact nicotinic signaling beyond Lypd6 (e.g. nAChRs, cholinergic projections) 45.
Based on this declined Lypd6 expression in adult V1, we tested if overexpression of Lypd6 in adulthood could prolong ocular dominance plasticity beyond the V1 critical period, by utilizing a transgenic mouse line with pan-neuronal overexpression of Lypd6 (Lypd6Tg)43. Following a 4-day short-term monocular deprivation of the contralateral eye (4d MD), ocular dominance plasticity was assessed in anesthetized mice using a 16-channel linear silicone probe46 (Fig. 1c). Each recorded V1 neuron was assigned an ocular dominance (OD) score based on the relative balance of the neuron’s visually evoked responses to the visual stimulus independently presented to the contralateral and ipsilateral eye. Adult Lypd6Tg mice that underwent 4d MD displayed an overall significant shift in eye preference of the V1 (OD shift), away from the deprived contralateral eye and towards the non-deprived ipsilateral eye compared to adult Lypd6Tg mice that did not undergo any visual deprivation (no MD) or adult wild type (WT) mice that typically display no significant levels of plasticity regardless of visual deprivation (Fig. 1d). Cumulative distributions of ocular dominance index scores (ODI: -1=most ipsilateral dominant, 0=equal/binocular, 1=most contralateral dominant) for each recorded neuron from each cohort show a significantly higher percentage of single neuronal responses that are relatively ipsilateral dominant in the Lypd6Tg 4d MD mice compared to control groups (Fig. 1e). Lastly, individual animal level comparisons of contralateral bias index (CBI) scores – the relative strength to which the visually evoked activity of V1 neurons from the contralateral eye dominates over that from the ipsilateral eye, show a significant decrease only in the Lypd6Tg 4d MD mice (Fig 1f). Altogether the results suggest that Lypd6 is a novel positive regulator of ocular dominance plasticity.
Lypd6 expression specifically in SST interneurons reactivates ocular dominance plasticity in adult V1
While the aforementioned experiments outline the sufficiency of neuronal overexpression of Lypd6 for prolonging ocular dominance plasticity into adulthood, the transgenic line does not provide temporal, spatial, or cell type information regarding this effect. In order to address these questions, we developed a Cre-dependent AAV vector for overexpressing Lypd6 (AAV8-DIO-EGFP-2A-Lypd6: AAV-Lypd6) (Fig. 2a). Our previous study established that Lypd6 is most highly expressed in SST interneurons in V1 with lower expression in a subset of glutamatergic neurons 42. Therefore, we characterized Lypd6 overexpression in V1 SST interneurons through viral injection of AAV-Lypd6 into adult SST-cre mice, and in V1 glutamatergic neurons through dual viral injection of AAV-Lypd6 and an AAV with a Cre expression vector driven by a CamKII promoter (AAV8-CamKIIa-mCherry-Cre: AAV-CamKII-cre) into adult WT mice as it was done in previous reports 29,47. The viral injections were adjusted to target the deep cortical layers in the area where the endogenous expression of Lypd6 has been previously reported to be localized42. The two viral injection conditions resulted in GFP expression restricted to either SST or glutamatergic populations, but overall both conditions presented robustly increased expression of Lypd6 in V1 (Fig. 2b, c, d, and Supplementary Fig. 1).
Next, we assessed the ability of SST or glutamatergic overexpression of Lypd6 in V1 of adult animals to induce ocular dominance plasticity. We injected AAV-Lypd6 or control AAV vector (AAV8-hSyn-DIO-EGFP: AAV-GFP) into adult SST-cre mice and, after >3 weeks of viral incorporation, subjected them to 4d MD (Fig. 2e). In adult mice with SST-specific Lypd6 overexpression that underwent 4d MD (SST-Lypd6 4d MD) there was a significant OD shift compared to non-deprived (no MD) counterparts (SST-Lypd6 no MD) or control 4dMD mice injected with AAV-GFP (SST-GFP 4d MD) (Fig. 2f). To test whether glutamatergic overexpression of Lypd6 would also induce plasticity in adult mice, we injected a viral cocktail of AAV-CamKII-cre with either AAV-Lypd6 or AAV-GFP into V1 of adult WT mice and, after >3 weeks, subjected them to 4d MD (Fig. 2e). The viral cocktail approach resulted in 93.2% of successfully co-transfected (GFP-positive) cells as positive for the excitatory cell marker vGlut1 (Supplementary Fig. 1c.), and 52.2% of total vGlut1-positive population as GFP-positive (Supplementary Fig. 1d). We did not observe a significant shift in OD in mice with glutamatergic overexpression of Lypd6 that underwent 4d MD (CamKII-Lypd6 4d MD) or in either of control counterparts (CamKII-Lypd6 no MD and CamKII-GFP 4dMD) (Fig. 2f). The cumulative distributions of ODI also show only a significant shift in the SST-Lypd6 4d MD mice (Fig. 2g). Lastly, comparisons of CBI scores show a significant decrease only in the SST-Lypd6 4d MD mice (Fig. 2h). Altogether, robust plasticity was measured only in the deprived mice that were overexpressing Lypd6 specifically in SST interneurons. These results suggest that SST interneurons are critical for Lypd6-mediated reactivation of plasticity in adult V1. Furthermore, the temporal and spatial restrictions of viral injections imply that overexpression of Lypd6 directly in V1 after the closure of critical period plasticity is sufficient for the induction of robust plasticity in adulthood.
α2 nicotinic acetylcholine receptor subunit is required for reactivation of ocular dominance plasticity
We next aimed to determine the nAChR subtype required for reactivation of V1 plasticity mediated by Lypd6 in SST interneurons. In the hippocampus the SST-positive oriens-lacunosum moleculare (O-LM) interneurons, which are physiologically similar to a subpopulation of cortical SST-expressing Martinotti interneurons 48, express Lypd6 as well as α2-subunit-containing nAChR (nAChRα2), which is the most sparsely expressed nAChR subtype in the brain and features a unique non-desensitizing characteristic 39,42. In the cortex, the punctate expression pattern of nAChRα2 subunit concentrated in deep cortical layers 49 highly resembles that of cortical Lypd6 42. Indeed, in situ hybridization for Lypd6 and Chrna2 mRNA revealed that 79% of nAChRα2-positive cells also express Lypd6, and that nAChRα2 subunit is expressed on more than 53% of Lypd6-positive cells (Fig. 3a-c). Furthermore, while 23% of SST-positive cells overlapped with cells expressing nAChRα2, 95% of nAChRα2-positive cells overlapped with SST-positive cells (Fig. 3 d-f). Consistent with a recent single cell RNA sequencing study in adult V1 50, no overlaps were observed between nAChRα2 and VIP or vGlut1 – the markers of vasoactive intestinal peptide-expressing interneurons and pyramidal cells (data not shown). Together this confirms strong preferential co-expression of Lypd6 and nAChRα2 in a subpopulation of SST interneurons.
Due to this high degree of co-localization between Lypd6, a modulator of non-α7 nAChRs 43, and nAChRα2 we examined whether ablation of nAChRα2 could eliminate ocular dominance plasticity induced by SST-specific overexpression of Lypd6. To accomplish this, we created a bigenic SST-cre/Chrna2KO mouse line to allow SST-specific Cre-recombinase expression on a background of Chrna2 knockout (Chrna2KO: α2KO) 51. We then injected AAV-Lypd6 into V1 of adult SST-cre/α2KO bigenic mice, allowed >3 weeks for viral incorporation, and assessed ocular dominance plasticity after 4d MD (Fig. 3d). Intriguingly, ablating nAChRα2 (SST-Lypd6/α2KO 4d MD) eliminated the robust plasticity induced through SST-specific Lypd6 overexpression in adult V1. Unlike the adult SST-Lypd6 4d MD mice, adult SST-Lypd6/α2KO 4d MD mice as well as deprived and non-deprived adult α2KO mice (α2KO 4d MD and α2KO no MD) did not show significant shifts in OD (Fig. 3e), cumulative distributions of ODI (Fig. 3f), or significant decreases in CBI (Fig. 3g). These findings strongly suggest the requirement of nAChRα2 in the induction of plasticity by Lypd6 in SST interneurons.
To further investigate the physiological role of nAChRα2 in regulating plasticity in adult V1, we turned to a condition known for modulating SST interneuron activity as well as reactivating V1 plasticity. Voluntary physical exercise through the use of a running wheel was recently shown to induce juvenile form of plasticity in adult mice 52. Interestingly, locomotion is also known to modulate SST interneuron activity 29-33. We therefore examined the functional contribution of nAChRα2 in this model of adult reactivation of V1 plasticity. When allowed voluntary exercise (+Running) on a running dish during the 4d MD (Fig. 3h), adult WT mice (WT 4d MD, +Running) displayed a significant shift in ocular dominance, consistent with the aforementioned published study 52. However, adult α2KO mice subjected to the same condition (α2KO 4d MD, +Running) did not show significant shifts in OD (Fig. 3i), cumulative distributions of ODI (Fig. 3j), or significant decreases in CBI (Fig. 3k). In addition, SST-specific viral knockdown of Lypd6 in the adult V1 of did not block voluntary-exercise-induced plasticity and was no different from adult WT 4d MD+Running mice (Supplementary Fig. 2). We postulate that other members of the Lynx family played a compensational role in mediating voluntary-exercise-induced plasticity in the Lypd6 knockdown context. Alternatively, as Lypd6 is a positive modulator of non-α7 nAChRs, and given that knocking out Chrna2 eliminates the voluntary-exercise-induced plasticity, simply knocking down Lypd6 may be insufficient for blocking exercise induced-plasticity in the presence of nAChRα2. Finally, we performed qPCR on V1 extractions of adult WT 4d MD+Running mice (n=4) and control adult WT 4d MD (n=4) mice in order to examine whether voluntary exercise changes the expression of Chrna2 or Lypd6, but did not find any significant differences (Chrna2: P=0.7908, Lypd6: P=0.7111 Student’s t-test). We speculate that perhaps increased cholinergic activity is sufficient to enhance nAChRα2 signaling and encourage voluntary-exercise-induced plasticity in the adult brain. We also compared Chrna2 expression between adult SST-Lypd6 no MD mice (n=2), and adult SST-GFP no MD mice (n=2), and did not find any significant differences (P=0.20, Student’s t-test). Collectively, these results suggest that the voluntary-exercise-induced model of V1 plasticity requires nAChRα2, further signifying the critical role of nicotinic signaling through nAChRα2 in reactivating plasticity in adulthood.
Lypd6 expressed in SST interneurons increases their activity to drive ocular dominance plasticity
Having ascertained that the effect of Lypd6 in reactivating V1 plasticity is regulated through SST interneurons, we investigated the plasticity mechanisms in these cells as well as downstream circuits. Cortical SST interneurons are known to provide inhibitory local inputs onto PV interneurons 28,53-57 and inhibit them twice as potently they do pyramidal neurons in mouse V158. As rapid reduction in PV interneuron firing during the first day of MD is a unique trigger of V1 plasticity during the critical period 22, we examined whether adult overexpression of Lypd6 in SST interneurons can increase their activity and in turn suppress PV activity during the first day of MD.
First, to directly evaluate how overexpression of Lypd6 in SST interneurons affects their activity, we used channelrhodopsin-2 (ChR2) as an optogenetic tag to allow identification of visually evoked cells as SST interneurons based on whether they can also be activated by optogenetic stimulation (Fig. 4a, b). We targeted expression of ChR2 specifically in SST interneurons by injecting AAV with Cre-dependent expression of ChR2 (AAV1-dflox-hChR2(h134R)-mCherry-WPRE-hGH: AAV-ChR2) into adult SST–Cre mice along with either AAV-Lypd6 (SST-Lypd6) or AAV-GFP (SST-GFP) and, after >3 weeks of viral incorporation, subjected the mice to 24 hours of monocular deprivation of the contralateral eye (1d MD) (Fig. 4c). Considering the physiological context of endogenous Lypd6 expression in deep cortical layers, we restricted our analyses to only the SST interneurons recorded from the lower half of the 16-channel linear silicone probe in an effort to focus on the SST interneurons of lower layer IV, layers V, and VI. In adult SST-Lypd6 mice that underwent 1d MD (SST-Lypd6 1d MD), there was a significant increase in the visually evoked firing rate of SST interneurons compared to 1d MD adult SST-GFP mice (SST-GFP 1d MD) or non-deprived adult SST-Lypd6 (SST-Lypd6 no MD) mice when statistically analyzed using a linear mixed model (LMM; for details See Methods – Statistical analysis) that considers “animal” as a variable with random effect, and “genetic manipulation” (-GFP or -Lypd6) and “experience” (1d MD or no MD) as variables with fixed effects (Fig. 4d, e). In comparing baseline firing rates without visual stimulation, only a trending increase in firing rate was observed in SST-Lypd6 1d MD mice (4.47 spikes/sec) relative to SST-GFP 1d MD mice (2.87 spikes/sec) (SST-Lypd6 1d MD vs. SST-GFP 1d MD: P=0.057), and no significant difference was found when assessing experience-dependent effects (SST-Lypd6 1d MD vs. SST-Lypd6 no MD (4.77 spikes/sec): P=0.97, LMM). This result suggests an interactive effect of Lypd6 overexpression and sensory experience in rapidly increasing SST interneuron activity.
To then determine the causal role of rapid increase in SST interneuron activity itself in mediating plasticity, we examined if chemogenetic activation of SST interneurons strictly during the first day of monocular deprivation alone can reactivate plasticity (Fig. 4f and Supplementary Fig. 3). Adult SST-cre mice were injected with an AAV with Cre-dependent expression of excitatory Gq protein-Designer Receptors Exclusively Activated by Designer Drugs (AAV8-hSyn-DIO-hM3D(Gq)-mCherry: AAV-GqDREADD)59. In mice given Clozapine-N-Oxide (CNO), to activate SST interneurons during the first day of the 4d MD (CNO+), there was a significant OD shift compared to mice that received control saline injections (CNO-) (Fig. 4g). The cumulative distributions of ODI show a significant shift in CNO+ mice compared to CNO-mice (Fig 4h). Comparisons of CBI scores show a significant decrease only when SST interneurons were activated by injection of CNO (Fig. 4i). Together, these results imply a novel role for SST interneuron activity in reactivating a juvenile form of plasticity in adult V1.
Lypd6 expressed in SST interneurons suppresses PV interneuron activity to drive ocular dominance plasticity
Finally, we aimed to elucidate the consequence of Lypd6-mediated elevation in SST activity upon PV interneuron activity. In adult SST-cre mice, we injected AAV-Lypd6 (SST-Lypd6) or AAV-GFP (SST-GFP) and, after >3 weeks of viral incorporation, subjected the mice to 1d MD (Fig. 5a). We putatively determined the cell type identity of visually evoked PV interneurons using spike shape-based classification. Visually evoked cells were sorted into two categories: putative fast-spiking PV (pFS) neurons defined by a narrow spike waveform that is an electrophysiological characteristic unique to PV interneurons, or regular-spiking neurons (RS) (Supplementary Fig. 4d, e).
In adult SST-Lypd6 mice that underwent 1d MD (SST-Lypd6 1d MD), LMM showed a trending reduction in pFS neuron firing rates with no statistical significance compared to all control groups (Fig. 5b, c), likely reflecting a broad distribution of firing rates among pFS neurons. As it appeared that less pFS neurons of higher visually evoked firing rates were observed after 1d MD in adult SST-Lypd6 mice, we systematically binned normalized firing rates to determine the largest separation of firing rates between adult SST-Lypd6 and SST-GFP groups and compared the fractions of pFS neurons with firing rates higher than the separation. Using one-way analysis of variance at animal level, we found a significant reduction in the fraction of pFS neurons of higher firing frequency in adult SST-Lypd6 1d MD mice (Fig. 5d). This suggests an experience-dependent drop in firing rate of PV interneurons when Lypd6 is overexpressed in adult SST interneurons. Furthermore, the distributions of pFS neuron firing rates that are broader than those of SST interneurons in our data suggest that sub-populations of PV interneurons may particularly be impacted by SST-specific Lypd6 overexpression. With respect to baseline firing rates, no significant differences were observed (SST-Lypd6 1d MD (2.06 spikes/sec) vs. SST-GFP 1d MD (2.55 spikes/sec): P=0.86, vs. SST-Lypd6 no MD (3.08 spikes/sec): P=0.49, LMM). Altogether these results may suggest an interactive effect of overexpression of Lypd6 and sensory experience in rapidly decreasing PV interneuron activity.
To determine the causal role of PV interneuron activity in robust plasticity regulated by Lypd6, we examined if chemogenetic restoration of PV interneuron activity during the first day of deprivation can prevent robust plasticity regulated by Lypd6 overexpression (Fig. 5d and Supplementary Fig. 4). We created a bigenic Lypd6Tg/PV-cre mouse line to allow PV-specific Cre-recombinase expression on a background of pan-neuronal Lypd6 overexpression. Adult Lypd6Tg/PV-cre mice were injected with AAV-GqDREADD. In mice injected with CNO to activate PV interneurons during the first day of the 4d MD (CNO+), there was a significant blockage of ocular dominance plasticity compared with mice that received control saline injections (CNO-) (Fig. 5g). The cumulative distributions of ODI show that the shift is significantly eliminated in CNO+ mice compared to CNO-mice (Fig. 5h). Comparisons of CBI scores show a blockage of decrease only in CNO-injected mice where PV interneurons were activated (Fig. 5i). While we cannot fully rule out the possibility that the suppression of V1 responses by chemogenetic activation of PV interneurons partly contributed to the blockage of plasticity, together these results implicate the role of Lypd6 in potentiating SST activity to engage an SST interneuron-based inhibition of PV interneurons – a novel inhibitory-inhibitory microcircuit, to induce a juvenile form of plasticity in the adult V1.
Discussion
Our study identifies an endogenous nAChR modulator, Lypd6, and its action through nAChRα2 as the first targetable SST interneuron-specific molecular mechanism for reactivating plasticity in the adult cortex. As SST interneurons are known to be modulated by alteration of visual inputs 27, neuromodulation induced by locomotion 29-33, and top-down regulation 34-36, the Lypd6-nAChRα2 system is ideally situated to gate these multiple modulatory inputs onto SST interneurons to regulate plasticity. Having demonstrated the Lypd6-nAChRα2 system as a molecular target for restoring plasticity in the adult brain, it would be necessary for future studies to first investigate the mechanistic details of the action of Lypd6 on nAChRα2 in SST interneurons, and then determine to what extent this system is relevant in regulating plasticity in the juvenile cortex. Finally, our study provides critical knowledge underlying the molecular and circuit means for the induction of plasticity in the adult visual cortex, which can unveil insights into establishing new therapeutic targets and strategies for treating disorders where potential for recovery is limited due to diminished plasticity. Therefore it is imperative for future long-term efforts to aim towards the development of agonists with high specificity for nAChRα2 to test as drug candidates in disease models such as one for amblyopia.
Lypd6, like several members of the Lynx family 41, is membrane tethered through a glycophosphatidylinositol (GPI)-anchor and is known to positively modulate nicotinic signaling 43. Our results from SST interneuron-specific Lypd6 overexpression on a background of Chrna2 knockout suggest that the primary mechanism for Lypd6-based initiation of plasticity is mediated specifically through nAChRα2s. However, Lypd6 may also mediate additional mechanisms unrelated to nicotinic signaling 60. While the subcellular localization of Lynx proteins has not been fully elucidated, there are evidences that show they potentially localize to synaptic compartments to modulate nAChRs 61,62. Lypd6 may facilitate the activation of SST interneurons by impacting postsynaptic nicotinic modulation of excitatory synaptic input onto SST interneurons, such as in the case of hippocampal CA1 O-LM interneurons where nAChRα2 and their non-desensitizing features are critical for gating LTP through receptor-mediated calcium influx 39. Alternatively, Lypd6 may modulate nicotinic signaling at presynaptic SST terminals to facilitate the release of GABA 63. Nevertheless, either potential mechanism may achieve the facilitation of SST interneuron activity – physiological condition that we confirmed, through chemogenetic activation of SST-interneurons, to be a trigger of robust plasticity (Fig. 4). Proper antibodies suited for immunohistochemical analyses of Lypd6 and nAChRα2 are both currently unavailable. However, future studies may benefit by applying recently developed technologies, such one that uses in vivo genome editing 64 to label endogenous Lypd6 or nAChRα2 protein and visualize them at resolutions as high as the synaptic level. Such method would reveal finite details regarding their interaction as well as any other endogenous mechanisms governed by Lypd6.
Our study highlights Lypd6- and nAChRα2-expressing SST interneurons located in deep layers of V1 as a key cellular target for reactivating plasticity in adulthood through nicotinic neuromodulation. Previous studies have reported mixed results regarding the question of whether nicotinic signaling directly affects SST interneurons. While some studies reported clear nicotinic responses in neocortical SST interneurons 65,66,28, some have reported otherwise 67,68. This discrepancy may reflect the observation that expression of Lypd6 and nAChRα2 is selectively expressed in a subpopulation of SST interneurons located in deep cortical layers (layers 5 and 6) but not of those in layer 2/3 50,42. It is possible that such a minor subpopulation of deep layer Lypd6-expressing SST interneurons was not well represented in the previous negative reports. For example Gulledge et al.66 reported that 30% of SST interneurons in layer 5 rat V1 responded to nicotine, which is consistent with the percentage of SST cells expressing Lypd6 and nAChRα2 (Fig. 3). While the expression of Lypd6 and nAChRα2 are limited to a subpopulation of deep layer cortical SST interneurons, the unique non-desensitizing nature of nAChRα2 allows continual activation and calcium entry in the presence of Ach 39, making the nAChRα2 highly suited for driving cortical plasticity through elevation of nicotinic tone with Lypd6 or by receiving cholinergic inputs through locomotion. The SST-interneurons in deep cortical layers are also reported to project widely to almost all cortical layers and types of neurons 69, which can potentially contribute to widespread distribution of the drive for cortical plasticity exerted through nAChRα2. Overall our study highlights the SST interneuron that expresses Lypd6 and nAChRα2 as a cellular target of nicotinic neuromodulation for reactivating cortical plasticity in the adult V1. More mechanistic insight into the physiological action of ACh and Lypd6 on nAChRα2 signaling can be expected in future studies by taking full advantage of the recently developed mouse lines with selective expression of Cre recombinase in subpopulations of SST interneurons 38,70. This should be combined with future development of a new ChAT-ChR2 transgenic mouse line that avoids abnormally high endogenous cholinergic tone 71, to allow proper optogenetic modulation of endogenous Ach release. Finally, as our study resonates with the current understanding of the profound functional diversity of SST interneurons that has been steadily developing in recent years, we anticipate an impetus for development of selective nAChRα2-selective antagonists for acute pharmacological modulation, for they would become crucial tools for further investigating the functional role of the subpopulation of nAChRα2-expressing SST interneurons in the cortical circuit.
Our findings establish a novel circuit mechanism – SST interneuron-mediated inhibition of PV interneurons, for triggering a juvenile form of plasticity in adulthood. To our knowledge, this is the first evidence implicating the role of the SST-PV disinhibitory microcircuit in modulating cortical plasticity. We demonstrated that the overexpression of Lypd6 in SST interneurons leads to reduced activity in sub-population of PV interneurons, similar to what is typically observed specifically in juvenile mice after 1 day of visual deprivation 22. Interestingly, a recent study showed that silencing SST interneurons for the entirety of 5 days of visual deprivation could induce some level of cortical plasticity in adulthood. However the nature of plasticity measured in that study was that of an adult form, characterized by an elevation of open ipsilateral eye response but no reduction of the deprived contralateral eye response - the latter is required to define a juvenile form of plasticity 26. Nevertheless, collectively these studies highlight the capabilities of SST interneurons in engaging both a homeostatic adult form of plasticity 26 and a rapid form of juvenile like-plasticity 22, depending on the level of SST interneuron activity (reduced or enhanced) and the timing of activity change (during the initial phase within 1 day or later homeostatic phase after 5 days). This reflects the placement of the SST interneuron in the cortical circuit as a versatile integrator of a variety of inputs. Consistently, recent studies demonstrated that SST interneurons in V1 could either be activated or suppressed by neuromodulation triggered by locomotion depending on the context of the visual environment 29-33. Specifically, locomotion enhances overall SST interneuron activity when the mouse is presented with large visual input29,30 including a full-field gray screen 31. On the other hand, SST interneurons are largely non-responsive to small visual inputs 30 or are suppressed in darkness by the inhibitory action of the VIP interneurons 29,32. Here in our study we propose a model whereby the Lypd6-nAChRα2 system may function as a molecular switchboard for SST interneurons to shift modes and activate the SST-PV disinhibitory circuit 38,72. This mechanism then serves to trigger a juvenile form of plasticity 22 while suppressing an adult form of plasticity mainly through reduction of inhibitory action of SST interneurons upon the distal dendrites of pyramidal neurons 26,73,74 that would normally be engaged through an nAChRα2-independent VIP-SST disinhibitory circuit 26. In this model, both SST-PV and VIP-SST disinhibitory circuits can co-exist to play complementary roles in the modulation of plasticity, depending on the context and states of the adult brain. While SST interneurons are known to exhibit maturation towards the end of the critical period in V175, the differential contributions of SST-PV and VIP-SST circuits in regulating plasticity between critical period and adulthood remains unknown. It also remains to be determined how neuromodulation of SST interneurons change from critical period to adulthood to regulate plasticity.
The potential for signaling through nAChRα2 and its modulation by Lypd6 to induce plasticity in the adult cortex opens up the possibility of targeting this receptor for therapeutic interventions. The selective expression of both Lypd6 and nAChRα2 in SST interneurons make them attractive novel therapeutic targets with the potentially fewer off target effects 76-78 than interventions directed against more non-specific targets (e.g. more abundant nicotinic subunits), which are ideal for treating conditions such as amblyopia, stroke, and traumatic brain injury where hope for recovery is thin due to diminished plasticity. There have been recent reports of positive allosteric modulators with significantly higher specificity for nAChRα2 76-78 that could serve in future studies as promising drug candidates to test in animal models of amblyopia. Combinations of such potential pharmacological interventions with behavioral interventions like physical exercise 52 known to impact SST interneuron activity 29-32,79 (Fig. 3) may also be fruitful. Finally, our study highlights the disruption of cortical plasticity as a key pathophysiological consequence of mutations of the nAChRα2 gene and its non-coding regulatory elements 80,81,82,83, or deficits in SST interneurons84,85,86,87,88,89,90, which have been commonly reported among multiple neurodevelopmental, neurodegenerative, and psychiatric disorders including epilepsy, addiction, depression, and schizophrenia.
Methods
Animals
All mice were housed in groups of 2–5 together with the sibling groups of the same sex in standard and uniform cage sizes (199 mm x 391 mm x 160 mm: width x depth x height, GM500, Tecniplast) under a 12hr light:dark cycle (lights on at 7:00AM: lights off at 7:00PM) in a facility with constant temperature (23°C) and ad libitum access to food and water. Wild-type C57Bl/6 mice were obtained from Jackson laboratory and Charles River. Both male and female were used. Lypd6Tg mice were originally generated by A.Z.43, and transferred to H.M. and back-crossed to C57Bl/6. Somatostatin-ires-cre (SST-cre: Jackson laboratory #013044), Parvalbumin-ires-cre (PV-cre: Jackson laboratory #008069), and Chrna2KO (MMRRC #30508) were purchased and bred in-house. Bigenic lines were created through targeting breeding of above strains. All animal protocols were approved by the Institutional Care and Use Committee (IACUC) at Icahn School of Medicine at Mount Sinai.
Monocular deprivation
Monocular deprivation procedure for animals that underwent either 4 day or 1 day of monocular deprivation was conducted on the contralateral eye and with ipsilateral eye left open. The mice were anesthetized with isoflurane during the entire procedure. Eyelid margins were trimmed using an iris scissor and one eye was sutured closed for one or four days. Following MD, mice were returned to their home cage prior to extracellular recording and subsequent euthanasia.
Generation and validation of AAV-Lypd6 and AAV-miR-Lypd6
Lypd6 was amplified from a cDNA library derived from whole mouse V1, subcloned into a pcDNA3.1 (-) vector. Two oligonucleotides including shRNA sequence for Lypd6 (TGCTGAATGACAGAGATGCTGTTCCCGTTTTGGCCACTGACTGACGGGAACAGTCTCTGTCATT,CCTGAATGACAGAGACTGTTCCCGTCAGTCAGTGGCCAAAACGGGAACAGCATCTCTGTCATTC, life technologies) previously reported to knockdown Lypd6 expression 43 were subcloned into pcDNA6.2-GW/EmGFP-miR using BLOCK-iT Pol II miR RNAi Expression Vector Kit with EmGFP (life technologies). Subclonings were performed using an isothermal DNA assembly method (Gibson Assembly; New England Biolabs) and transformed into E. Coli. Colonies with correct insert were identified through DNA sequencing (Genewiz), cultured and then isolated using a Hi-speed Midiprep kit (Qiagen). Expression was first examined by transfection of N2A cells in vitro. To create pAAV vector, an inverted bicistronic 2A sequence was inserted into pAAV-Ef1α-DIO-EGFP-WPRE-pA (Addgene#37084) upstream of EGFP by PCR linearization and overhang production on pAAV vector. The pcDNA3.1(-)-Lypd6 vector was used as a template for the Lypd6 insert which was subsequently inserted into the pAAV-DIO-EGFP-2A vector as described above to create a pAAV-DIO-EGFP-2A-Lypd6-WPRE-pA vector. pcDNA6.2-mGFP-miR-Lypd6 vector was subcloned into pAAV-DIO-EGFP-2A-Lypd6-WPRE-pA vector to generate pAAV-DIO- mGFP-miR-Lypd6 (pAAV-miR-Lypd6) vector. After sequence verification, a large culture and Maxiprep isolation produced a purified vector that was sent to the UNC viral core for viral packaging using an AAV8 serotype. To confirm viral efficiency, AAVs were stereotaxically injected (see below for injection methods) into V1 of SST-cre mice. After perfusion, sections were labeled with rabbit anti-somatostatin antibodies (1:1000; Peninsula Laboratories) and the SST interneuron-specific GFP expression was confirmed. V1 was micro-dissected from an additional cohort of mice to assay overexpression or knockdown of Lypd6 through qPCR. RNA was isolated using an RNeasy lipid tissue mini kit (Qiagen) and cDNA was produced. The cDNA was subjected to qPCR analysis using a Taqman assay (Life Technologies) at the Icahn School of Medicine at Mount Sinai Quantitative PCR CORE facility.
Stereotaxic injection
Mice were isoflurane anesthetized and head-fixed in a mouse stereotaxic apparatus (Narishige). A mid-line incision was made in the scalp and a micro-drill was used to drill a small hole in the skull over the right visual cortex. Three injections (0.5 μl each) were made into the deep layers of V1 binocular zone (from lambda: AP: 0.0, ML: 3.1, DV: 0.6; AP: 0.0, ML: 2.85, DV: 0.6; AP: 0.3, ML: 3.0, DV: 0.6) using a 2.5 μl syringe (Hamilton Company) with a 30-gauge needle and microsyringe pump controller (World Precision Instruments) set to inject at 200nl/minute. The syringe remained in place for one minute following injection to reduce backflow of virus. After injections, the skull hole was sealed using Kwik-Sil (World Precision Instruments), and the scalp was sutured. The mice were allowed to recover from anesthesia in an empty cage over a warming pad. Following recovery, mice were returned to their home cage where they remained for >3 weeks to allow for viral incorporation prior to any additional procedures or testing. The following viral constructs were used in this study: AAV-Lypd6 (AAV8-EF1α-DIO-EGFP-2A-Lypd6), AAV-GFP (AAV8-hSyn-DIO-EGFP) (University of North Carolina (UNC) Vector Core), AAV-CamKII-Cre (AAV8-CamKIIa-mCherry-Cre: UNC Vector Core), AAV-GqDREADD (AAV8-hSyn-DIO-hM3D(Gq)-mCherry: UNC Vector Core), AAV-ChR2 (AAV1-Ef1α-dflox-hChR2(H134R)-mCherry-WPRE-hGH: Penn Vector Core), AAV-miRLypd6 (AAV8-Ef1α-DIO-EGFP-2A- miRLypd6), and AAV-hSyn-Cre (AAV8-hSyn-mCherry-Cre: UNC Vector Core).
Chemogenetic activation of hM3d(Gq) Designer Receptors Exclusively Activated by Designer Drugs (DREADD)
Clozapine-N-Oxide (CNO; Sigma-Aldrich), a normally inert compound that specifically activates DREADD receptors, was prepared in 0.9% saline and injected i.p. into adult SST-Cre (Fig. 4f-i) or Lypd6Tg/PV-cre bigenic mice (Fig. 5d-g) at a concentration of 3 mg/kg. CNO or saline was injected immediately following MD and again 12 hours later for DREADD induced activation of SST or PV interneurons during the first day of MD following the previously validated protocol in mouse visual cortex22(Supplementary Fig. 3, 4).
In vivo extracellular recording
As previous described, preparatory surgery leading to, and recording itself was performed initially under nembutal/ chlorprothixene anesthesia and then maintained with isoflurane 46. For animals that underwent either 4 days or 1 day of monocular deprivation, the deprived (contralateral) eye was reopened just prior to the start of recordings. Visually evoked single-unit activity were recorded in response to a high contrast single bar, generated by Visage System (Cambridge Research Systems), that laterally traveled across the monitor in a 4 second interval followed by a 6 second interstimulus interval per trial. Through the use of an eye patch, the stimulus was separately presented to the contralateral eye and then to the ipsilateral eye for at least 12 trials each. For each animal, 3 to 10 single units were recorded in each of the 4 to 6 vertical penetrations spaced evenly (250 μm intervals) across mediolateral extent of V1 to map the monocular and binocular zones to avoid sampling bias. A linear 16-channel electrode was used to record each vertical penetration, which allowed recording from upper (channels 1∼8) and lower (channels 1∼9) cortical layers with similar weight. By analyzing 7 of the experimental groups that were used in our data, we found that the fraction of neurons recorded from upper layers (0.5235 ± 0.02350) did not differ from that of neurons recorded from lower layers (0.4765 ± 0.02350) (n=35 animals: P=0.3247, Student’s t-test for paired samples). The signal was amplified and thresholded (OmniPlex, Plexon). To ensure single-unit isolation, the waveforms of recorded units were further examined offline (Offline Sorter, Plexon). To analyze the electrophysiology data, normalized OD index of single neuron was computed by custom made MATLAB program by peristimulus time histogram analysis of peak to baseline spiking activity in response to each eye: {[Peak(ipsi)-baseline (ipsi)]–[Peak (contra)-baseline(contra)]}/{[Peak (ipsi) – base line(ipsi)]+ [Peak(contra)-baseline(contra)]}.). OD scores were converted from OD index using a 7-point classification scheme as follows: -1 to -0.5 = 1, -0.5 to -0.3 = 2, -0.3 to -0.1 = 3, -0.1 to 0.1 = 4, 0.1 to 0.3 =5, 0.3 to 0.5 =6, 0.5 to 1 =7. For each binocular zone, contralateral bias index (CBI) is calculated according to the formula: [(n1-n7)+2/3(n2-n6)+1/3(n3-n5)+N]/ 2N, where N=total number of cells and nx=# of cells corresponding to OD score of x.
For optical tagging experiments, SST interneurons were optogenetically tagged by injecting a Cre-dependent virus expressing channelrhodopsin-2 (AAV1-Ef1α-dflox-hChR2(H134R)-mCherry-WPRE-hGH: Penn Vector Core) into the V1 binocular zone of SST-cre mice. The expression of channelrhodopsin in a cell type of interest allowed the use of optogenetic stimulation with a blue light as a search stimulus to simultaneously sort stimulus-locked responses from cell types of interest, and then record their visually evoked responses, by using an optic fiber-coupled 16-channel linear silicone probe (Neuronexus). This allowed accurate and high-throughput recordings of specific cell types, even against broad baseline noise and activities of other neuronal populations. Optogenetically tagged SST interneurons of the V1 binocular zone were identified using a 473 nm (blue) laser search stimulus emitted and delivered through the optic fiber coupled to the silicone probe and oriented immediately above the V1 cortical surface. After sorting for optogenetically responsive units (SST interneurons expressing ChR2), the optogenetic stimulus was switched off. Then a high-contrast single bar visual stimulus was presented to each eye to record the visual evoked responses (spike firing rate) of the sorted units. Analysis of visually evoked activities of sorted SST interneurons were restricted to only those recorded from the lower half of the 16-channel linear silicone probe in an effort to focus on the SST interneurons of lower layer IV, layers V, and VI. All animals with less than 10 analyzable lower layer SST interneurons recorded were excluded from the analysis. Normalized firing rate of each lower layer SST interneuron, in response to contralateral eye (or deprived eye in animals in 1 day monocular deprivation groups) stimulation, was computed by first using a custom made MATLAB program by generating a peristimulus time histogram-based analysis of peak and baseline spiking activity in response to visual stimulus. The peak visually evoked activity through contralateral/deprived eye stimulation was then subtracted by baseline spiking activity in order to obtain the normalized visually evoked firing rate.
To isolate visually evoked responses from putative fast-spiking (pFS) neurons, visual stimulus responsive cells were distinguished as pFS neurons or regular-spiking (RS) neurons by means of spike width- (trough-to-peak time) based classification. The spike-width criterion for separating into pFS and RS neurons was established by measuring the spike width of optogenetically tagged PV interneurons. PV interneurons were optogenetically tagged by injecting a Cre-dependent virus expressing channelrhodopsin-2 (AAV1-Ef1α-dflox-hChR2(H134R)-mCherry-WPRE-hGH, Penn Vector Core) into V1 binocular zone of PV-cre mice. During recording, pFS neurons were first identified and sorted, using a 473 nm laser search stimulus delivered through the optical fiber coupled to the 16-channel silicone probe and oriented immediately above the V1 cortical surface, depending on their responsiveness to blue light stimulation within 3 msec post blue light emission. The search stimulus was then exchanged to a visual stimulus. Visually evoked spike widths of the pFS neurons identified by optical tagging were then pooled to establish the official criterion for pFS neurons as having visually evoked spike width time (trough-to-peak time) of less than 412 μsec (Supplementary Fig. 4). Averaged spike waveform data for each unit was analyzed using a custom made MATLAB program to obtain a spike width time for each unit. All pFS neurons sorted and recorded were included in the analysis of their visually evoked activity. All animals with less than 10 analyzable pFS neurons recorded were excluded. Using the same method for analyzing SST interneurons, normalized firing rates of pFS neurons were calculated from responses to contralateral/deprived eye stimulation.
Following recordings, all AAV injected mice were transcardially perfused and the extent of GFP or mCherry signal was utilized to assess the viral transduction. Only mice that exhibited GFP or mCherry signal in the recorded V1 area were included for the analysis of ocular dominance plasticity or SST/pFS neuron activity. The experimenters performed while blind to experimental conditions such as viral constructs or CNO/saline administration, but not monocular deprivation.
In situ hybridization
The production of probes and methodology for in situ hybridization has been previously described42. Briefly, RNA probes including a fluorescein or digoxinogen (DIG) tag were generated and utilized to label Lypd6, SST, VIP vGlut1, GFP, and Chrna2 mRNA in 7 μm sections of V1 from fresh frozen brains of animals at P28 (CP) and >P60 (adult). To fluorescently label mRNA, anti-fluorescein/DIG-POD (1:2000; Roche) and anti-fluorescein/DIG-Alkaline phosphatase (AP) (1:1000; Roche) antibodies were used. POD-conjugated antibody labeling of mRNA was proceeded with with TSA Plus DNP signal amplification (Perkin Elmer) and a subsequently labeling with anti-DNP-KLH-488 antibodies (1:1000; Life Technologies). AP-conjugated-antibody-labeled mRNA was stained using HNPP/Fast Red (Roche). Imaging was performed using an LSM780 confocal microscope (Zeiss). ImageJ was used to quantify the density of labeled pixels from each image or to examine co-localization using a color based thresholding method. For the quantification of Lypd6 mRNA expression across age in V1, pixel density (>2 standard deviations above mean intensity of full image field) was determined from low magnification images of V1 binocular zone using ImageJ software. For the quantification of co-localization of Lypd6 or SST or VIP with nAChRα2, the number of cells positive for Lypd6/SST/VIP or nAChRα2 in each image was determined using ImageJ by automated counting using a threshold of >2 standard deviations above background and limiting to particles of >40 μm cell diameter. To calculate the co-localization percentage, first color based thresholding was utilized in ImageJ to isolate and quantify the co-localized cells, then percentage was calculated by dividing the number of co-localized cells by the number of Lypd6/SST/VIP or nAChRα2 positive cells in each image. For the quantification of co-localization of GFP and vGlut1 mRNA, thresholding was first utilized to isolate and calculate number of particles representing GFP labeled cells (>40 μm cell diameter). Then the corresponding vGlut1 labeled image was redirected to the mask retained from the analysis of GFP labeled image to calculate total number and then percentage of GFP labeled cells with co-localized vGlut1 labeling. For the quantification of Lypd6 mRNA expression in V1 binocular zone of mice injected with a cocktail of AAV-Lypd6 and AAV-CamKII-Cre, low magnification images of binocular V1 from the injected or the corresponding naïve hemisphere were analyzed for particles after thresholding. Absolute intensity for each particle was then measured as the product of mean intensity and area of particle. All absolute intensity values were summed to obtain total binocular V1 intensity, which was subsequently normalized by dividing by the μm2 area of the V1 binocular zone. The V1 binocular zone of each image was assessed using Paxinos and Franklin’s The Mouse Brain in Stereotaxic Coordinates (1997) as reference.
Quantitative PCR (qPCR)
Adult mice were deeply anesthetized with isoflurane, checked for response with paw pinch, and decapitated. The brain was extracted, and under RNAse-free conditions, briefly washed in ice cold 0.1 M phosphate buffer, and then an estimated area of cortical tissue representing V1 – a 2mm-by-2mm area of the cortex most posterior and 2mm lateral from the median, was cut out, white matter removed, immediately frozen on dry ice, and stored under -80°C. Total RNA was extracted from V1 using the RNeasy Mini Kit (Qiagen) and stored at -80°C. concentrations of total V1 RNA yielded ranged from 0.22 to 0.3 μg/μl. Total V1 RNA was converted to cDNA using a High-Capacity cDNA Reverse Transcription Kit (Life Technologies). qPCR was performed at the Mount Sinai Quantitative PCR core facility using 7900HT Real-Time PCR instrument (ABI/Life Technologies), TaqMan probes (catalog numbers: Lypd6 Mm00622636_01, Chrna2 Mm00460630_m1, ABI/Life Technologies) and TaqMan Universal Master Mix II, no UNG (ABI/Life Technologies). Quantification of the fold change was calculated via the –ΔΔCT method (equivalent to a log2 fold change), using ΔCT values derived by normalization of CT values to mouse beta actin as reference housekeeping gene.
Immunohistochemistry
Anesthetized mice were transcardially perfused with cold 4% paraformaldehyde (PFA) dissolved in 0.1M phosphate buffer. The brains were post-fixed in 4% PFA at 4 °C, and cryoprotected in 30% sucrose solution. The frozen brains were sectioned into 30-μm-thick coronal sections using a cryostat (CM3050, Leica). Free-floating sections were washed in tris-buffered saline (TBS), pH 7.5, and then blocked in 1% bovine serum albumin in TBST (0.25% Triton X-100 in TBS) for 1 h. The sections were incubated with mouse anti-parvalbumin (1:500; Swant), rabbit anti-somatostatin (1:1000; Peninsula Laboratories), or rabbit anti-c-Fos (1:500; Calbiochem) antibodies overnight at room temperature. After primary antibody incubation the slices were washed in TBST, followed by secondary antibody incubation with Alexa fluor dyes (Thermo Fisher Scientific). Imaging was performed using a Zeiss LSM780 confocal microscope at 20 or 40X magnification. The investigator performing the analysis was blind to the animal genotype.
Statistical analysis
The following statistical approaches were utilized for experiments assessing ocular dominance plasticity. χ2 test was used to compare at cell level distributions of ocular dominance scores between two groups and assess for ocular dominance shift. For readability, the histogram figures for distribution of ocular dominance scores (OD Scores) represent percentage of cells rather than actual cell number. However the χ2 statistics are results of the tests conducted on actual cell number. Cumulative distributions of ocular dominance index (ODI) were compared at cell level by using the Kolmogorov-Smirnov (K-S) test. One-way analysis of variance (ANOVA) was used to compare contralateral bias index (CBI) at animal levels and Tukey’s multiple comparisons test was used for post hoc analyses. As justification of the use of a parametric test, normality of the CBI data was tested by using the D’Agostino-Pearson Test and Shapiro-Wilk Test on any groups where n=mice exceeding the criteria for the tests (n>8, and n>7 respectively). Normality was confirmed in all groups where tests were applicable: WT MD, WT no MD, Lypd6Tg 4d MD, and SST-Lypd6 4d MD. In addition, we applied the Shapiro-Wilk Test to the CBI data of a key experimental group from our previous study (Smith et al., 2016) that utilized the same experimental design and statistical analyses for testing ocular dominance plasticity, and also confirmed normality. Initial comparison of pFS cell (PV interneuron) and SST interneuron firing rates were conducted using one-way ANOVA with Bonferroni corrected multiple comparison tests for post hoc analyses. To officially determine the statistical significance of differences in SST and PV interneuron firing rates, a linear mixed modeling (LMM) approach was used, from packages LmerTest (v. 2.0.32), lme4 (v. 1.1.12), and lsmeans (v. 2.25) in the R programing language (v. 3.2.2). This model considered “animal” as a variable with random effect, and “genetic manipulation” (-GFP or -Lypd6) and “experience” (1d MD or no MD) as variables with fixed effects. For further analysis of pFS neuron activity, the normalized firing rates were systematically binned into binary “high” and “low” firing rate categories by determining the firing rate value where the largest separation between the groups (SST-Lypd6 and SST-GFP) occurs. The fractions of pFS neurons of firing rate higher than the systematically determined separation point were then compared between groups at animal level using one-way ANOVA. Percentages were transformed using the classical arcsin transformation method to better-fit assumptions of normality and homogeneity of variance. Differences (diff) reported are from un-transformed data and reflect the mean difference in proportions of cells in the established firing rate bin, (i.e. -0.18 and -0.16, or 18% and 16% for SST-Lypd6 1d MD and SST-Lypd6 no MD respectively). Student’s t-test was used for all other analyses that compare mean differences between two groups. A minimum P value of 0.05 was accepted as statistically significant. Other than LMM, all statistical analyses were performed using Prism 6.0h (GraphPad Software). All CBI, quantified expression, quantified co-localization, and visually evoked firing rate data in figures are presented as mean±SEM. No method of randomization was used to determine allocation of samples. No statistical methods were used to predetermine sample sizes. However, all cell and animal-level sample sizes for all experiments including quantified expression/co-localization experiments and in vivo extracellular electrophysiological experiments were selected based on previously published studies.
Data availability
The data that support the findings of this study are available from the corresponding author (H.M.) upon reasonable request.
Author Contributions
M.S., M.P.D., and H.M. designed research; M.S., M.P.D, P.N.B., P.Y., and H.M. performed research; A.Z. contributed unpublished reagents and provided expertise; M.S., M.P.D, P.N.B., P.Y., M.R.S., and H.M. analyzed data; M.S., M.P.D, and H.M. wrote the manuscript. The authors declare no conflicts of interest. H.M.
Acknowledgments
This work was supported by National Eye Institute R01EY024918, R01EY 026053, R21EY026702 to H.M., National Institute on Drug Abuse T32 DA007135 to M.P.D., National Institute of Mental Health T32MH096678 to M.S., R21MH106919 to H.M., Knights Templar Eye Foundation to H.M., March of Dimes to H.M., Whitehall Foundation to H.M., and Brain and Behavior Research Foundation to H.M. AZ is a member of the Excellence Cluster Immunosensation. We thank Dr. Ming-Hu Han (Icahn School of Medicine at Mount Sinai) for providing technical expertise on optogenetics and Dr. Yasmin Hurd and Dr. Michael Michaelides (Icahn School of Medicine at Mount Sinai) for their expertise on chemogenetics.
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.
- 6.
- 7.↵
- 8.↵
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.
- 55.
- 56.
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵