Spontaneous activity of striatal projection neurons supports maturation of striatal inputs to substantia nigra dopaminergic neurons

Reviewer #1 (Public Review):

Summary:

This study offers a comprehensive examination of the early postnatal development of the patch and matrix compartments within the striatum. These are segregated circuits within the striatum circuits with distinct embryonic origins and functional roles in mature brain physiology. Despite the recognized significance of these circuits, a comprehensive understanding of their postnatal maturation remains elusive.

Strengths:

The authors undertake a thorough investigation, characterizing the intrinsic properties of direct pathway spiny projection neurons (dSPNs) and indirect pathway spiny projection neurons (iSPNs) across both matrix and striosome compartments throughout development. The authors identify the regulatory role of M1 receptors in modulating spontaneous activity in SPNs, and demonstrate the impact of chemogenetic inhibition of MOR-positive neurons during development on GABAergic synapses in substantia nigra pars compacta (SNc) dopamine (DA) neurons. These findings significantly advance our understanding of striatal development and function.

Weaknesses:

Certain methodological considerations warrant attention. Notably, the reliance on TdTomato expression for the identification of striosomes raises concerns, particularly regarding the substantial difference in slice thickness between the immunohistochemistry (IHC) images (50um) shown in Figure 2 and those utilized for whole-cell recordings (300um).

Enhanced clarification regarding the identification of cell patches is possible in the electrophysiology rig conditions. Using a widefield microscope rather than a confocal would strengthen the reliability of this methodology.

In the Ca2+ imaging experiments of Figure 2, striosomes were defined as the regions of brighter GCaMP fluorescence. This presents a potential limitation because it presupposes higher activity levels within patch cells, which is what the experiment is designed to test. Based on this criteria, neurons of this region will necessarily have more activity than in others.

There is also no information on how Ca2+ imaging traces were analyzed. In the examples provided, putative matrix neurons seem to exhibit different Ca2+ dynamics compared to striosome neurons. The plateau responses might reflect even higher activity than the transient signals observed in striosome neurons. It'll be important to know how the data was quantified. For example, calculations of F0 based on rolling functions tend to underestimate dF/F in traces like this. Calculations of the area under the curve can also provide valuable information in these cases.

There is no description of the 8mM KCl treatment in the methods. Was this only used for the Ca2+ imaging experiments? The percentage of active cells in Figures 2C-D is similar to or lower than that described in Figure 2B, which is confusing. Were recordings always performed in 8mM KCl?

Lastly, while the findings of Figure 6 suggest a deficit in striosomal inputs to SNc DA neurons, they do not conclusively demonstrate this point (DA neurons receive many sources of inhibition, and local interneurons in SNc are highly plastic). Given the availability of Opmr1-Cre mice and the utilization of multiple viruses in Figure 6 experiments, the inclusion of experiments employing ChR2 to directly assess striatal/striosome inputs would substantially strengthen this claim. This is the main claim stated in the manuscript title, so it is important to provide evidence of specific striatonigral deficits.

Reviewer #2 (Public Review):

Summary:

The manuscript by Kokinovic et al. presents evidence that a significant portion of striatal projection neurons (SPNs) are spontaneously active early in development. This spontaneous activity (as measured in ex vivo brain slices) is due to intrinsic mechanisms, and subsides over the course of the first few postnatal weeks in a cell-type specific way: striosome direct and indirect pathway SPNs (dSPNs and iSPNs, respectively) remain spontaneously active until postnatal days 10-14, by which time matrix dSPNs and iSPNs have become entirely silent. The authors suggest that this early spontaneous activity may be in part due to M1 muscarinic receptor signaling. Through chemogenetic inhibition of striosome SPNs (of which dSPNs target dopaminergic neurons of the SNc), the authors present evidence that critical postnatal windows of SPN activity shape the strength of GABAergic innervation of the SNc (measured in adults). This study provides a useful and solid characterization of the functional, postnatal compartmental development of the striatum. However, some weaknesses in the experimental design should be addressed before definitively concluding that postnatal striosome SPN activity determines its functional innervation of dopaminergic SNc neurons.

Specific Comments:

(1) While certainly interesting and possibly true, evidence for the necessity of early striosome dSPN activity in shaping their functional innervation of dopaminergic SNc neurons is not entirely convincing. The functional measure of GABAergic innervation of dopamine neurons is inferred from mIPSCs. As the authors state, dopaminergic neurons have numerous other sources of GABAergic inputs in addition to striosome dSPNs. So while manipulating striosome activity may ultimately alter the overall GABAergic innervation of SNc dopamine neurons, the specificity of this to striosome dSPN inputs is not known. Optogenetic stimulation of striosome->SNc neurons after chemogenetic silencing would help support the authors' interpretation. Related to this point, while striatonigral projections form embryonically, is there evidence that striosome->SNc synapses are indeed functional by P6-14 when CNO is delivered?

(2) One big caveat that needs to be addressed is that all measures of early postnatal spontaneous SPN activity were performed in ex vivo slices. Are SPNs active (in pathway/compartmental specific ways) in vivo during this time? If it is unknown, is there other evidence (e.g. immediate early gene expression, etc...) that may suggest this is indeed the case in vivo?

(3) It appears that 8mM KCl (external) was only used while measuring spontaneous calcium oscillations, not spontaneous spiking (Figure 2). Was there any evidence of spontaneous calcium activity in the lower KCl concentration (3mM?) used for cell-attached recordings? One caveat is that experiments demonstrating that SPNs fire spontaneously in the presence of AMPA receptor blockers (Figure S1) were presumably performed in 3mM KCl. Does elevated KCl increases spontaneous EPSPs during the ages examined? If so, are the calcium oscillations shown in Figure 2 synaptically driven or intrinsically generated? Somewhat related, speculation on why M1 receptor blockade reduces calcium oscillations but not spontaneous spikes in striosome dSPNs would be useful.

(4) Several statements in the introduction could use references.

Reviewer #3 (Public Review):

Summary:

Kokinovic et al. presents an interesting paper that addresses an important gap in knowledge about the differences in the development of direct and indirect pathway striatal neurons in the striosome and matrix compartments. The division of the striatum into 4 distinct populations, striosome-dSPNs, striosome-iSPNs, matrix-dSPNs, and matrix-iSPNs is important, but rarely done. This study records all four populations across early development and shows differences in action potential characteristics and intrinsic properties. They also suppress striosome activity during postnatal development and evaluate the characteristics of adult dopaminergic neurons in control and previously striosome-quieted conditions.

Strengths:

The striatal electrophysiology is beautifully and carefully done and shows important developmental differences between neural subtypes.

The idea to test the striatonigral connection is a good idea.

Weaknesses:

The authors didn't actually test the striatonigral connection. The experiments they do instead don't convincingly show that the striosomal or even striatal connection to the dopaminergic neurons is altered after postnatal striosome suppression.

Major concerns:

(1) mIPSCs are measured and are reduced after chemogenetic suppression of striosomal neurons during development. This is an interesting finding, but these mIPSCs could be coming from any inhibitory input onto the SNc neurons. It is unlikely that most of the mIPSCs are coming from the striosomal inputs. The GPe is much more likely to be the source of these mIPSCs than the striatum because the GPe inputs form synapses nearer the soma and have a higher probability of release (Evans et al., 2020). dSPNs inhibit GPe neurons through a non-canonical pathway (Cui et al., 2021; Spix et al., 2021) and striosomes also inhibit the SNr (McGregor et al., 2019). The striatum has the potential to disinhibit SNc neurons through both the SNr or the GPe (Evans, 2022), and modification of the striosome-SNr or striosome-GPe connections during development could be what is causing the mIPSC changes. To claim that the striosome-SNc connection is altered, a direct test of this connection is necessary.

(2) The dopaminergic neurons recorded seem to be randomly selected, but the striosomes do not inhibit all SNc dopamine neurons. They selectively inhibit the ventral tier SNc neurons (Evans et al., 2020). In the present manuscript, it is impossible to know which subpopulation of SNc neurons was recorded, so it is impossible to tell whether the dopaminergic neurons recorded are the ones expected to receive striosomal input.

(3) Very similarly, the striosomes selectively wrap around the "SNr dendrite" of SNc neurons that participate in striosome-dendron bouquets (Crittenden et al., 2016). However, not all SNc neurons have prominent SNr dendrites (Henny et al., 2012). In the morphological images of Supplemental Figure 3, it looks like the recorded cells sometimes have an SNr dendrite and sometimes don't (but it is hard to tell because the medial-lateral rostral-caudal axis is not labeled in the images). The presence or absence of the "SNr dendrite" is a strong determinant of whether an individual dopaminergic neuron receives striosomal inhibition or not (Evans et al., 2020). As above, not knowing whether the neurons recorded have SNr dendrites makes it impossible to know whether they should be receiving striosomal input at all.

(4) It's quite interesting that the dendron-bouquet structure is intact even after striosomal activity suppression, as cannabinoid receptor knockout greatly disrupts the structural integrity of bouquets (Crittenden et al., 2022). However, going along with point 3, the gephyrin puncta analysis only at the somas is very limiting. The striosome-SNc relevant puncta would be primarily on the SNr dendrite. Gephyrin density on the SNr dendrites or in bouquets would be much more informative than density on the soma.

(5) The authors claim that "CNO didn't affect the shape of the DA neuron dendritic tree", but more information about the morphological analysis should be added. It is not clear how the sholl analysis was conducted or whether a full 3D reconstruction was made. This claim seems to be based on only one dendritic measurement (sholl analysis), but many other dendritic or morphological features could be altered.

Crittenden, J.R., Tillberg, P.W., Riad, M.H., Shima, Y., Gerfen, C.R., Curry, J., Housman, D.E., Nelson, S.B., Boyden, E.S., & Graybiel, A.M. (2016) Striosome-dendron bouquets highlight a unique striatonigral circuit targeting dopamine-containing neurons. Proc. Natl. Acad. Sci. U.S.A., 113, 11318-11323.
Crittenden, J.R., Yoshida, T., Venu, S., Mahar, A., & Graybiel, A.M. (2022) Cannabinoid Receptor 1 Is Required for Neurodevelopment of Striosome-Dendron Bouquets. eNeuro, 9, ENEURO.0318-21.2022.
Cui, Q., Du, X., Chang, I.Y.M., Pamukcu, A., Lilascharoen, V., Berceau, B.L., García, D., Hong, D., Chon, U., Narayanan, A., Kim, Y., Lim, B.K., & Chan, C.S. (2021) Striatal Direct Pathway Targets Npas1+ Pallidal Neurons. J Neurosci, 41, 3966-3987.
Evans, R.C. (2022) Dendritic involvement in inhibition and disinhibition of vulnerable dopaminergic neurons in healthy and pathological conditions. Neurobiol Dis, 172, 105815.
Evans, R.C., Twedell, E.L., Zhu, M., Ascencio, J., Zhang, R., & Khaliq, Z.M. (2020) Functional Dissection of Basal Ganglia Inhibitory Inputs onto Substantia Nigra Dopaminergic Neurons. Cell Rep, 32, 108156.
Henny, P., Brown, M.T.C., Northrop, A., Faunes, M., Ungless, M.A., Magill, P.J., & Bolam, J.P. (2012) Structural correlates of heterogeneous in vivo activity of midbrain dopaminergic neurons. Nat. Neurosci., 15, 613-619.
McGregor, M.M., McKinsey, G.L., Girasole, A.E., Bair-Marshall, C.J., Rubenstein, J.L.R., & Nelson, A.B. (2019) Functionally Distinct Connectivity of Developmentally Targeted Striosome Neurons. Cell Rep, 29, 1419-1428.e5.
Spix, T.A., Nanivadekar, S., Toong, N., Kaplow, I.M., Isett, B.R., Goksen, Y., Pfenning, A.R., & Gittis, A.H. (2021) Population-specific neuromodulation prolongs therapeutic benefits of deep brain stimulation. Science, 374, 201-206.

Author Response

We are deeply grateful for the highly professional analysis of our work by the Journal Editor and Reviewers. Here is our provisional response to some of the reviewer comments. In our response, we would like to address two comments that were common to all Reviewers' responses. We will thoroughly address all of the Reviewers' comments in the final version of the paper.

Incomplete analysis of maturational changes of striato-nigral connections.

In the initial study, we showed that chronic inhibition of striosomal neurons with the DREADD approach during early postnatal development leads to decreased functional innervation of dopaminergic cells by striosomes in adulthood. We have shown that by two approaches: (1) analysis of miniature inhibitory post-synaptic currents (mIPSCs) and (2) analysis of GFP and gephyrin puncta densities around dopaminergic cells. The results from these experiments strongly suggest a decrease in inhibitory drive to dopaminergic neurons of substantia nigra pars compacta, yet we agree that only GFP puncta density can be considered as a direct evidence for weakened striatonigral connections. Reviewers indicated that additional direct measurements of striatonigral synaptic efficacy would be needed to strengthen our conclusions. We completely agree with this statement and will evaluate the possibility of doing the suggested experiments, using optogenetic stimulation of striosomal inputs to dopaminergic neurons.

Inconsistent description of Ca2+ imaging experiments.

Unfortunately, there was a general misunderstanding in interpreting the Ca2+ imaging methods description. All our experiments were done so that baseline Ca2+ oscillations and oscillations in the presence of a drug were recorded in the usual ACSF (containing 3 mM KCl) at the patch-clamp setup chamber. So, conditions were exactly the same as for cell-attached and whole-cell recordings. At the end of each experiment, ACSF containing 8 mM KCl was applied. This high-KCl condition was used to calculate the total number of viable cells reacting to elevated potassium concentrations, and this number was taken as 100 %. Therefore, the percents displayed in the paper represent the actively oscillating cells in common ACSF (3 mM KCl), counted as a percent of the total number of cells that responded to the following high potassium stimulation (8 mM KCl). The formula was: (Number of active cells in 3 mM KCl / number of viable cells active at 8 mM KCl)*100.