Satellite glia modulate sympathetic neuron survival, activity, and autonomic function

Curation statements for this article:
  • Curated by eLife

    eLife logo

    Evaluation Summary:

    The role of satellite glial cells in the sympathetic nervous system has not been extensively investigated. Using targeted ablation of SGCs, the authors demonstrate that satellite glia has a profound effect on neuronal activity and the survival of sympathetic neurons. The peripheral sympathetic system is responsible for a wide spectrum of activities, including blood flow, heart rate, respiration, and digestion.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #3 agreed to share their name with the authors.)

This article has been Reviewed by the following groups

Read the full article See related articles

Abstract

Satellite glia are the major glial cells in sympathetic ganglia, enveloping neuronal cell bodies. Despite this intimate association, the extent to which sympathetic functions are influenced by satellite glia in vivo remains unclear. Here, we show that satellite glia are critical for metabolism, survival, and activity of sympathetic neurons and modulate autonomic behaviors in mice. Adult ablation of satellite glia results in impaired mTOR signaling, soma atrophy, reduced noradrenergic enzymes, and loss of sympathetic neurons. However, persisting neurons have elevated activity, and satellite glia-ablated mice show increased pupil dilation and heart rate, indicative of enhanced sympathetic tone. Satellite glia-specific deletion of Kir4.1, an inward-rectifying potassium channel, largely recapitulates the cellular defects observed in glia-ablated mice, suggesting that satellite glia act in part via K + -dependent mechanisms. These findings highlight neuron–satellite glia as functional units in regulating sympathetic output, with implications for disorders linked to sympathetic hyper-activity such as cardiovascular disease and hypertension.

Article activity feed

  1. Author Response

    Reviewer #1 (Public Review):

    “Strengths of the paper include the use of the novel promoter (which is stated to have ~50-fold higher abundance in SGCs than astrocytes) and the dataset itself, which is for the most part thorough and convincing.”

    We thank the Reviewer for appreciating the novelty of the study, and that in general, our findings are well-supported and convincing.

    “Concerning specificity, CNS involvement through effects on other cell types is not totally ruled out in these studies, and effects on the same cell type but in other ganglia (parasympathetic and sensory) might be expected to impact sympathetic function. For example, as Vit (2008) reported that following shRNA knockdown of Kir4.1 in trigeminal ganglia hypersensitivity to mechanical stimulation could affect autonomic activity. The authors tested for the influence of parasympathetic using pupillary constriction, and it is somewhat surprising that there is no deficit if neuronal death and dysfunction are as profound in parasympathetic ganglia as shown here for the superior cervical ganglia.”

    We include new results to show that the elevated heart rate in BLBP:iDTA mice is prevented by chemically ablating sympathetic nerves using 6-OHDA (Figures 3E-F, revised manuscript). Since 6-OHDA does not cross the blood-brain barrier in adult mice after i.p injections (Kostrzewa and Jacobowitz, 1974), we conclude that these results point to a peripheral locus for the cardiovascular defect in BLBP:iDTA mice. Since 6-OHDA is selective for sympathetic nerves, we also reason that the potential loss of sensory or parasympathetic satellite glia do not contribute to the increased heart rate in BLBP:iDTA mice. As the Reviewer notes, we also found that BLBP:iDTA mice fully constrict their pupils in response to light, indicative of normal parasympathetic function.

    We do not exclude the possibility of defects in sensory or parasympathetic ganglia in BLBP:iDTA mice. A comprehensive analysis of these two systems will warrant significant effort, which we respectfully state is outside the scope of this initial study where we have focused on satellite glia in the sympathetic nervous system. However, our results in the revised manuscript and in the original submission provide evidence that enhanced sympathetic tone is responsible for driving the autonomic defects (elevated heart rate and pupil dilation) observed in BLBP:iDTA mice.

    To the Reviewer’s point about CNS involvement in the autonomic dysfunction in BLBP:iDTA mice, we cannot completely exclude the possibility since BLBP is also expressed in astrocytes, albeit to much lower levels (45-fold less) compared to satellite glial cells. However, as discussed in our original submission, astrocyte ablation results in severe motor deficits in mice including limb paralysis, ataxia, as well as smaller body weights (Schreiner et. al, 2015), none of which were observed in BLBP:iDTA mice. These results suggest that astrocytes are minimally perturbed in BLBP:iDTA mice.

    “Physiological effects of DTX but not Kir4.1 deletion increased sympathetic activity, whereas increased heart rate was also observed following chemical activation of SGCs using DREADD ligands (Xie et al., 2017). This opposite action is not discussed at length but is attributed to "context-dependence." Inconsistent results with stimuli believed to target the same substrate are worthy of additional consideration by the authors.”

    The Reviewer asks about the differences between our findings and a previous study (Xie et. al., 2017) where the authors reported increased heart rate with chemo-genetic manipulation in Gfap-hM3Dq mice. In contrast, we observe increased heart rate with satellite glia ablation in BLBP:iDTA mice. We are unable to reconcile the apparent differences because of the following reasons:

    (i) The experimental manipulations in the two studies are very different; acute chemo-genetic manipulation (over a time-scale of minutes) versus genetic ablation of satellite glial cells (over two weeks), making it difficult to directly compare behavioral outcomes (heart rate) at the whole animal level.

    (ii) The Reviewer does not distinguish between activation of the Gq-GPCR signaling pathway in satellite glial cells using DREADD ligands in the Xie et. al., study versus “activation” of satellite glia. It remains unknown how activation of this signaling pathway affects satellite glia physiology and functions. Indeed, it remains unclear what “activation” or “silencing” even mean for satellite glial cells. For satellite glial cells, it remains unknown as to how calcium mobilization affects these glial cells, and how this in turn, affects neuronal activity.

    We have read the manuscript by Xie et. al., carefully and could not find any direct evidence for exactly how DREADD-based activation of the Gq-GPCR signaling pathway in satellite glial cells could activate sympathetic neurons. The authors speculate that activation of the Gq-GPCR signaling pathway in satellite glia could activate sympathetic neurons via modulating glutamate transporters or inwardly rectifying K+ channels expressed in satellite glia. However, there are no glutamatergic neurons in sympathetic ganglia, and whether glutamate transporters in peripheral satellite glia have a role in glutamate uptake analogous to CNS astrocytes remains to be established. Further, if activation of the Gq-GPCR signaling pathway in satellite glia would lead to increased K+ uptake, as occurs in astrocytes, then this would result in reduced sympathetic neuron activity. However, Xie et. al., observed an increase in sympathetic tone in Gfap-hM3Dq mice.

    (iii) The cellular specificity of the Cre driver lines used in the two studies are different. We, and others, have shown that BLBP is a highly specific satellite glial cell marker (this study; Mapps et. al., 2022; Avraham et. al., 2020; 2021; 2022). Notably, using single-cell sequencing, we, and others do not detect GFAP in mouse satellite glial cell under normal or reactive conditions (Mapps et. al., 2022; Mohr et. al., 2021; Jager et. al., 2020), although it is found in rat satellite glial cell (Mohr et. al., 2021), raising the question of the suitability of GFAP as a satellite glia-specific marker in mice.

    We have included a modified version of this Discussion (pages 18-19, revised manuscript).

    “An alternative conclusion from the finding that the similar cellular level changes in sympathetic neurons induced by DTX and Kir4.1 cKO led to distinct changes in autonomic tone is that the neuronal phenotype does not dictate whole animal physiology”

    We have made this same point in our original submission that while BLBP:iDTA and Kir4.1 cKO mice show similar neuronal phenotypes at the cellular level, the loss of a single gene, Kcnj10 (for Kir4.1), from satellite glial cells is not sufficient to drive behavioral changes (pupil size and heart rate) at the whole animal level as seen with satellite glial cell ablation. We reason that there are Kir4.1-independent mechanisms that contribute to neuronal excitability to drive network-level changes.

    “Spatial buffering is given as the proposed benefit of Kir4.1 channels to the sympathetic neurons. However, this concept arose from studies in which clearance of local extracellular space was limited, and astrocytes were appreciated to be connected to a vast syncytium allowing siphoning away from the high levels near active neurons. The organization in peripheral ganglia differs in three major respects: Despite narrow extracellular space, there is no true barrier to diffusion of K ions from the neurons (one factor that makes drug targeting peripheral neurons appealing), SGCs are very thin (and thus without spatial consequence to uptake), and the coupling among the SGCs is local to those surrounding individual neurons, with very little coupling under normal conditions to other distal SGC-neuron units.”

    Briefly, previous studies indicate that satellite glial cells are capable of influencing neuronal excitability by dissipating extracellular K+ increases, largely by acting via Kir4.1 channels (Tang et. al., 2010). In addition, we include in the revised Discussion (page 16, revised manuscript) other ways by which Kir.4.1 loss in satellite glial cells could indirectly influence neuronal excitability, notably by promoting membrane depolarization of satellite glia or through regulation of diffusible signals such as BDNF.

    Reviewer #2 (Public Review):

    “Mapps and colleagues' potentially very interesting and important work investigates the biological function of satellite glial cells (SGCs) in the nervous system. SGCs are extremely understudied, even compared to other glia, which are themselves generally understudied compared to neurons. Thus, discoveries concerning what SGCs do would be of very high importance.”

    We thank the Reviewer for appreciating the novelty and importance of the study.

    “Major Concerns:”

    “I have major concerns related to the experimental approach to ablate SGCs, specifically in sympathetic ganglia, the successful accomplishment of which underlies the entire study.”

    • Most troublingly, in Figure 1E, Sox2(+) cells do not appear to be gone 14 days post-tamoxifen injection, just dimmer. The cell bodies in the treated panel also appear much dimmer than controls (on a separate note, these cell bodies do not appear to be atrophied, as shown in Figure 2A, which is also confusing). Although Figure 1B shows decreased Blbp levels by IF, there is no quantification. Coupled with the data that tamoxifen administration does not cause any change in phagocytic immune cells, as one might expect if a population of cells was ablated, this raises concerns about whether the experimental paradigm is working as expected. The authors need to convincingly show that SGCs are ablated from sympathetic ganglia for any subsequent critical claims to be supported.”

    We include new data to demonstrate the loss of satellite glial cells in BLBP:iDTA mice. Briefly;

    (i) There is a significant increase in the number of TUNEL+; Sox10+ satellite glial cells in BLBP:iDTA sympathetic ganglia at 5 days post-tamoxifen injection (Figures 1E, G, revised manuscript)

    (ii) Expression of several satellite glia-specific transcripts, including BLBP, is markedly decreased in BLBP:iDTA sympathetic ganglia, revealed by q-PCR analyses (Figure 1-figure supplement 2H, revised manuscript).

    (iii) We re-analyzed the images of Sox2-labeling by generating binary images, which removes the dependence on pixel values and simply records the presence/absence of a signal. Quantification revealed a substantial decrease in Sox2-labeled cells (33% decrease) in the BLBP:iDTA ganglia compared to controls at 14 days post-tamoxifen injections (Figure 1-figure supplement 2D-G in the revised manuscript). The 33% decrease in satellite glial cells, when analyzed in this manner, is lower than the 54% loss we had initially reported, suggesting that we may have included some proportion of cells that had down-regulated Sox2 expression but had not been ablated. We have clarified this point in the Results section (page 7, revised manuscript).

    Together, our results indicate that a significant proportion of SGCs are being ablated in BLBP:iDTA sympathetic ganglia.

    “Given the lack of changes in phagocytes in the experimental approach, it would also be important to show what happens to this large population of dead SGCs to understand the environment of the ganglia more fully and to interpret cellular and behavioral phenotypes better”

    Using IBA-1 labeling, we show that macrophage density is unaffected in BLBP:iDTA sympathetic ganglia. However, we do not make any conclusions about the reactivity of the macrophages or their ability to engulf dying satellite glia. It is also possible that persisting satellite glia in BLBP:iDTA ganglia are involved in clearing apoptotic cells, as reported by Wu et. al., 2009. We respectfully state that addressing the precise mechanisms by which dying satellite glial cells are cleared from the mutant ganglia is outside the scope of the current study.

    “• If the transgenic approach can be shown to be working to ablate SGCs as expected, it would be important to demonstrate that Blbp is not driving diphtheria toxin in any other cell type. The authors rule out a role for Schwann cells based on prior RNAseq and reporter mouse studies, but they do not verify these findings in their system. RNAseq can lack sufficient depth, and reporter mice do not always faithfully recapitulate endogenous expression. Similarly, the authors rule out astrocytic contributions based on lack of phenocopy but do not directly examine CNS tissue to support this claim.”

    We have provided additional evidence in the revised manuscript to support the cellular specificity of BLBP expression in satellite glial cells using single-cell RNA sequencing data from our lab and others (Mapps et. al., 2022; Avraham et. al., 2020; 2021; 2022) as well as the use of genetic reporter mice. Our single-cell RNA sequencing analyses also revealed that Schwann cells are scarce in sympathetic ganglia compared to sensory ganglia (Mapps et. al., 2022).

    As discussed in our response to Reviewer #1, DTA expression is restricted to BLBP-positive satellite glial cells and cannot be taken up by other cell types since this would require the DT receptor, which is not endogenously expressed in mice.

    In response to Reviewer #1 above and in the revised manuscript, we also discuss that although we cannot exclude the involvement of astrocytes in the behavioral defects observed in BLBP:iDTA mice, BLBP is 45-fold enriched in satellite glia compared to astrocytes. Genetic ablation of astrocytes results in severe motor deficits in adult mice including limb paralysis, ataxia, and smaller body weights (Schreiner et. al, 2015), which are not present in BLBP:iDTA mice. Importantly, we provide new data to show that chemical ablation of sympathetic nerves using 6-OHDA, which does not cross the blood-brain barrier in adult mice, prevents the cardiac dysfunction in BLBP:iDTA mice, indicating a peripheral locus. Together, with the evidence that we have provided for sympathetic neuronal defects at the morphological and cellular levels in BLBP:iDTA mice, we conclude that behavioral defects arise primarily from dysfunction of peripheral sympathetic neurons.

    “Additionally, I have some other concerns related to the data/data interpretation that should be clarified:

    • In Figure 1C, the authors note that TUNEL-labeled cells have large ovoid nuclei and are likely neuronal. A double-label would demonstrate this claim with more certainty than cell shape.”

    We have performed these additional experiments. We show that the majority of apoptotic cells at 5 days post-tamoxifen injections are satellite glial cells (Figure 1E, G, revised manuscript). Although, there was a trend toward enhanced neuronal apoptosis at this early stage, the number of apoptotic neurons in BLBP:iDTA ganglia was not statistically different from that in controls (Figure 1F, H, revised manuscript). We also did not observe a significant loss of neurons at 5 days post-tamoxifen injections (Figure 2-figure supplement 1B, revised manuscript). However, by 14 days post-tamoxifen injections, there is a significant loss (24% decrease) of sympathetic neurons (Figure 2G, revised manuscript). Together, these results suggest that sympathetic neuron loss occurs secondarily to the loss of satellite glial cells in BLBP:iDTA mice.

    “Related to this experiment, the quantification in Figure 1D does not appear to match the image shown in Figure 1C. Many TUNEL(+) cells are shown in the Blbp:iDTA image compared to control outside of the ganglionic borders, but this was not mentioned in the manuscript.”

    As discussed above in response to Reviewer #1, the quantifications in Figure 1D represents the total number of TUNEL-positive cells in the entire superior cervical ganglia (approximately 24-32 tissue sections of 12 m thickness each), while images in Figure 1C show a single tissue section from the ganglia.

    The Reviewer also notes that we observe TUNEL-positive cells outside the ganglia. However, this is evident in both control and mutant ganglia. This may represent a normal turnover of cells in tissues outside sympathetic ganglia (fat deposits and arteries). In the revised manuscript, we provide new images that show TUNEL labeling outside both the control and mutant ganglia (Figures 1C and 4D, revised manuscript), and also clarify this point in the text (page 6, revised manuscript).

    “• It is unclear from the Materials and Methods section if the mice are all on a congenic background. For example, how standard is pupil size from mouse to mouse, and is this more variable if mice are not congenic? This may be an issue given that the pupil measurements used as a read-out of sympathetic function have no baseline comparison, just control, and BLBP:iDTA animals.”

    We compared pupil size in BLBP:iDTA mice and their litter-mate controls, which are of the same genetic background. Our values of basal pupil sizes in mice after dark adaptation are consistent with previously reported results (Keenan et. al., 2016). Pupil size tends to be similar in darkness across mice of different genetic backgrounds (Keenan et. al., 2016).

    Reviewer #3 (Public Review):

    “In this manuscript, Mapps et al. report on the very interesting finding that satellite glia deletion significantly impacts sympathetic neuron function and survival. .. This is a very novel finding that reveals an important role for satellite glia in sympathetic physiology. It is comprehensive and well controlled. There are just a few issues that the authors should consider.”

    We thank the Reviewer for finding the study to be novel, comprehensive, and well-controlled.

    “In Fig. 1C-D, how many dpi was the TUNEL assay performed? It would be helpful to know how quickly the neurons die after glial depletion and if the cell death continues or plateaus. The authors should also co-label using neuronal and glial markers to evaluate whether the apoptotic cells are primarily neurons or glia. They report a loss of neurons, but how much of that is reflected in the TUNEL labeling is not clear.”

    Figures 1C-D represent the results of TUNEL labeling done at 5 days post-tamoxifen injections. As discussed above in responses to Reviewers #1 and 2, we include new data in the revised manuscript, using co-labeling with TUNEL and sympathetic neuron/satellite glia markers, to show early apoptosis of satellite glia, but not of neurons, at 5 days post-tamoxifen injections in BLBP;iDTA mice (Figures 1E-H, revised manuscript). We also did not observe a significant decrease in neuronal numbers at 5 days post-tamoxifen injections (Figure 2-figure supplement 1B, revised manuscript). However, by 14 days post-tamoxifen injections, there is a significant loss (24% decrease) of sympathetic neurons (Figure 2G, revised manuscript). These results indicate that satellite glia apoptosis occurs first, followed by the loss of sympathetic neurons in BLBP:iDTA mice. At the moment, we do not know if the neuronal death continues and/or reaches a plateau at later stages. This is a good point, which we will investigate in future analyses.

    “In Figs. 1C and 5C TUNEK analysis, there are quite a few TUNEL+ puncta outside of the ganglia, suggesting that there may be apoptosis in other adjacent tissues when the glia removed or Kir4.1 is deleted. The authors should comment on that if it were something consistently observed.”

    We observe TUNEL-positive cells immediately adjacent to the ganglia in both control and mutant mice (BLBP:iDTA or Kir4.1 cKO mice). This may represent a normal turnover of cells in tissues outside sympathetic ganglia (fat deposits and arteries). In the revised manuscript, we provide new images that show TUNEL labeling outside both the control and mutant ganglia (Figures 1C and 4D, revised manuscript), and also clarify this point in the text (page 6, revised manuscript).

    “The loss of neurons upon glial cell loss or Kir4.1 deletion is interesting. The authors discuss how neuron death could occur, but did they observe TUNEL+ cells in regions where the glia had been deleted? Given that the diphtheria toxin did not ablate all glia, were the neurons left with little or no surrounding glia more likely to die? This may be difficult to tell, but from the images in 1E, it looks like some neurons lack nearby glia. This would be a potential explanation for why only a fraction of the neurons died; those neurons with associated glia may be more protected.”

    The Reviewer makes an interesting point that the neurons without attached satellite glial cells are the most vulnerable to apoptosis. We were unable to conclusively make this correlation after looking through multiple images of TUNEL labeling in BLBP:iDTA ganglia. Interestingly, we found a similar (22% loss) of sympathetic neurons in Kir4.1 cKO mice in the absence of obvious disruptions in satellite glia association with sympathetic neurons (see Figure 4C and Figure 4-figure supplement 1C, revised manuscript). Thus, while the loss of neuron-satellite glia contacts may contribute to the neuronal death, there are also other mechanisms that could be involved in neuronal apoptosis.

    “It would be helpful to clarify a bit more what the control mice used for comparison were. From the text, it seems as if they were the same mice but not treated with tamoxifen. Were they given diphtheria toxin?”

    The Reviewer is correct that we used Fabp7-CreER2;ROSA26eGFP-DTA mice that were injected with either vehicle (corn oil) to serve as controls or injected with tamoxifen for satellite glia ablation. Mice did not have to be injected with diphtheria toxin, since tamoxifen injections would drive CRE-mediated DTA expression in BLBP-positive satellite glia.

    “In addition, did the authors check for any effects of tamoxifen alone? Given that estrogen can affect many physiological parameters, including cardiac function, tamoxifen alone could have some effect, e.g., Kuo et al., PMID: 20392827.”

    We thank the Reviewer for this point. We measured heart rate by electrocardiogram recordings in adult (P45 day old) wild-type C57Bl/6J mice or control (ROSA26eGFP-DTA) mice that did not express Cre that were either injected with corn oil or tamoxifen using the same paradigm as for BLBP:iDTA mice and litter-mate controls (sub-cutaneous injections, 180 mg/kg body weight for 5 consecutive days). We show that tamoxifen injection alone does not elicit any effects on heart rate in wild-type or control mice in the absence of Cre (Figure 3-figure supplement 1C-D, revised manuscript).

    “Interestingly, TH levels in BLBP:iDTA mutant axons appeared to be similar to that in controls, despite the marked reduction in TH mRNA and protein levels in neuronal cell bodies (Figure S2A). The Kaplan lab (PMC7164330) showed that TH mRNA trafficking and local synthesis play an important role in synthesizing catecholamines in the axon and presynaptic terminal. Although a bit beyond the scope of this study, it would be interesting to determine whether TH mRNA transport is altered by deletion of the glia. The authors might check to see if TH transcripts are reduced in axons by something like RNAscope.”

    We thank the Reviewer for this interesting point. The Reviewer likely meant that TH levels are up-regulated in axons since BLBP:iDTA mutant axons maintain TH expression despite the reduction of TH in neuronal soma. We tried assessing Th mRNA in axons in vivo using single molecule fluorescence in situ hybridization (smFISH). However, despite several attempts, we were not successful in getting the TH RNAscope probe to work. In the revised manuscript, we discuss enhanced Th mRNA trafficking and local translation as a possible mechanism for maintenance of axonal TH levels in BLBP:iDTA sympathetic neurons (Discussion, pages 19- 20, revised manuscript), and cite the Gervasi et. al., 2016 study.

  2. Evaluation Summary:

    The role of satellite glial cells in the sympathetic nervous system has not been extensively investigated. Using targeted ablation of SGCs, the authors demonstrate that satellite glia has a profound effect on neuronal activity and the survival of sympathetic neurons. The peripheral sympathetic system is responsible for a wide spectrum of activities, including blood flow, heart rate, respiration, and digestion.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #3 agreed to share their name with the authors.)

  3. Reviewer #1 (Public Review):

    The authors use a novel Satellite Glial Cell (SGC)-enriched promoter Blbp to target diphtheria toxin (DTX) killing, then analyze changes in sympathetic ganglia and autonomic function. These changes are compared to those resulting from similarly targeted deletion of Kir4.1 channels. To summarize, tamoxifen induction of DTX in adult mice led to >50% reduction in cervical sympathetic SGCs, a substantial decrease in adrenergic enzymes (~90-99% loss of TH and DBH), smaller neurons, and decreased pS6 (from which impaired mTOR is inferred), loss of ~25% neurons but 8-fold cFOS activation, and maintained axons and 60% increased circulating NE. Expression of certain adrenergic receptor subtypes was also found to be decreased. Conditional knockdown of Kir4.1 (by ~75% in RT-PCR) led to no apparent decrease in SGC numbers (judged by Sox2 and Blbp staining), ~60% decrease in TH and DHB, increased numbers of smaller neurons and impaired mTOR signaling, loss of about 20% neurons and increased cFOS.

    Although cellular effects of DTX ablation and Kir4.1 deletion in SGCs overlap considerably, the overlap does not include changes in autonomic function, where the DTX and Kir4.1-targeted deletion mice were quite different. DTX led to increased sympathetic activity (increased pupil size without apparent parasympathetic change in constriction and increased rate but reduced variability in heart rate). However, none of these changes were observed in the Kir4.1-targeted mice. The authors conclude that satellite glia is important for sympathetic neurons, partly through the provision of Kir4.1 channels and spatial buffering of potassium.

    Strengths of the paper include the use of the novel promoter (which is stated to have ~50-fold higher abundance in SGCs than astrocytes) and the dataset itself, which is for the most part thorough and convincing Issues include specificity of the targeting, opposite effects on sympathetic function reported from studies using DREADD activation of SGCs, and conclusions regarding Kir4.1 effects and mechanism.

    Concerning specificity, CNS involvement through effects on other cell types is not totally ruled out in these studies, and effects on the same cell type but in other ganglia (parasympathetic and sensory) might be expected to impact sympathetic function. For example, as Vit (2008) reported that following shRNA knockdown of Kir4.1 in trigeminal ganglia hypersensitivity to mechanical stimulation could affect autonomic activity. The authors tested for the influence of parasympathetic using pupillary constriction, and it is somewhat surprising that there is no deficit if neuronal death and dysfunction are as profound in parasympathetic ganglia as shown here for the superior cervical ganglia.

    Physiological effects of DTX but not Kir4.1 deletion increased sympathetic activity, whereas increased heart rate was also observed following chemical activation of SGCs using DREADD ligands (Xie et al., 2017). This opposite action is not discussed at length but is attributed to "context-dependence." Inconsistent results with stimuli believed to target the same substrate are worthy of additional consideration by the authors. An alternative conclusion from the finding that the similar cellular level changes in sympathetic neurons induced by DTX and Kir4.1 cKO led to distinct changes in autonomic tone is that the neuronal phenotype does not dictate whole animal physiology.

    Spatial buffering is given as the proposed benefit of Kir4.1 channels to the sympathetic neurons. However, this concept arose from studies in which clearance of local extracellular space was limited, and astrocytes were appreciated to be connected to a vast syncytium allowing siphoning away from the high levels near active neurons. The organization in peripheral ganglia differs in three major respects: Despite narrow extracellular space, there is no true barrier to diffusion of K ions from the neurons (one factor that makes drug targeting peripheral neurons appealing), SGCs are very thin (and thus without spatial consequence to uptake), and the coupling among the SGCs is local to those surrounding individual neurons, with very little coupling under normal conditions to other distal SGC-neuron units.

  4. Reviewer #2 (Public Review):

    Mapps and colleagues' potentially very interesting and important work investigates the biological function of satellite glial cells (SGCs) in the nervous system. SGCs are extremely understudied, even compared to other glia, which are themselves generally understudied compared to neurons. Thus, discoveries concerning what SGCs do would be of very high importance. The authors use a Blbp conditional approach to ablate SGCs with diphtheria toxin and examine subsequent phenotypes and behavioral outputs. The authors also begin to dissect some mechanisms by examining mTOR signaling and loss of Kir4.1 in SGCs. Unfortunately, despite great enthusiasm for the ideas underlying the study, in its present form, there are significant issues with the data as presented, which do not fully support the authors' conclusions.

    Major Concerns:
    I have major concerns related to the experimental approach to ablate SGCs, specifically in sympathetic ganglia, the successful accomplishment of which underlies the entire study.

    • Most troublingly, in Figure 1E, Sox2(+) cells do not appear to be gone 14 days post-tamoxifen injection, just dimmer. The cell bodies in the treated panel also appear much dimmer than controls (on a separate note, these cell bodies do not appear to be atrophied, as shown in Figure 2A, which is also confusing). Although Figure 1B shows decreased Blbp levels by IF, there is no quantification. Coupled with the data that tamoxifen administration does not cause any change in phagocytic immune cells, as one might expect if a population of cells was ablated, this raises concerns about whether the experimental paradigm is working as expected. The authors need to convincingly show that SGCs are ablated from sympathetic ganglia for any subsequent critical claims to be supported. Given the lack of changes in phagocytes in the experimental approach, it would also be important to show what happens to this large population of dead SGCs to understand the environment of the ganglia more fully and to interpret cellular and behavioral phenotypes better

    • If the transgenic approach can be shown to be working to ablate SGCs as expected, it would be important to demonstrate that Blbp is not driving diphtheria toxin in any other cell type. The authors rule out a role for Schwann cells based on prior RNAseq and reporter mouse studies, but they do not verify these findings in their system. RNAseq can lack sufficient depth, and reporter mice do not always faithfully recapitulate endogenous expression. Similarly, the authors rule out astrocytic contributions based on lack of phenocopy but do not directly examine CNS tissue to support this claim.

    Additionally, I have some other concerns related to the data/data interpretation that should be clarified:
    • In Figure 1C, the authors note that TUNEL-labeled cells have large ovoid nuclei and are likely neuronal. A double-label would demonstrate this claim with more certainty than cell shape. Related to this experiment, the quantification in Figure 1D does not appear to match the image shown in Figure 1C. Many TUNEL(+) cells are shown in the Blbp:iDTA image compared to control outside of the ganglionic borders, but this was not mentioned in the manuscript. This raises further concerns about the ability of the approach to only drive toxin in SGCs in sympathetic ganglia.

    • It is unclear from the Materials and Methods section if the mice are all on a congenic background. For example, how standard is pupil size from mouse to mouse, and is this more variable if mice are not congenic? This may be an issue given that the pupil measurements used as a read-out of sympathetic function have no baseline comparison, just control, and BLBP:iDTA animals.

  5. Reviewer #3 (Public Review):

    In this manuscript, Mapps et al. report on the very interesting finding that satellite glia deletion significantly impacts sympathetic neuron function and survival. Specifically, loss of the glia results in reduced mTOR signaling, norepinephrine production, and a loss of neurons. Surprisingly, there was an increase in neuronal activity, leading to increased physiological effects such as increased heart rate and pupil dilation. The authors also demonstrate that many of these effects can be mimicked by glial K+ channel, Kir4.1, deletion, indicating that loss of the glia disrupts K+ buffering around the neurons. This is a very novel finding that reveals an important role for satellite glia in sympathetic physiology. It is comprehensive and well controlled. There are just a few issues that the authors should consider.

    In Fig. 1C-D, how many dpi was the TUNEL assay performed? It would be helpful to know how quickly the neurons die after glial depletion and if the cell death continues or plateaus. The authors should also co-label using neuronal and glial markers to evaluate whether the apoptotic cells are primarily neurons or glia. They report a loss of neurons, but how much of that is reflected in the TUNEL labeling is not clear.

    In Figs. 1C and 5C TUNEK analysis, there are quite a few TUNEL+ puncta outside of the ganglia, suggesting that there may be apoptosis in other adjacent tissues when the glia removed or Kir4.1 is deleted. The authors should comment on that if it were something consistently observed.

    The loss of neurons upon glial cell loss or Kir4.1 deletion is interesting. The authors discuss how neuron death could occur, but did they observe TUNEL+ cells in regions where the glia had been deleted? Given that the diphtheria toxin did not ablate all glia, were the neurons left with little or no surrounding glia more likely to die? This may be difficult to tell, but from the images in 1E, it looks like some neurons lack nearby glia. This would be a potential explanation for why only a fraction of the neurons died; those neurons with associated glia may be more protected.

    It would be helpful to clarify a bit more what the control mice used for comparison were. From the text, it seems as if they were the same mice but not treated with tamoxifen. Were they given diphtheria toxin? In addition, did the authors check for any effects of tamoxifen alone? Given that estrogen can affect many physiological parameters, including cardiac function, tamoxifen alone could have some effect, e.g., Kuo et al., PMID: 20392827.

    Interestingly, TH levels in BLBP:iDTA mutant axons appeared to be similar to that in controls, despite the marked reduction in TH mRNA and protein levels in neuronal cell bodies (Figure S2A). The Kaplan lab (PMC7164330) showed that TH mRNA trafficking and local synthesis play an important role in synthesizing catecholamines in the axon and presynaptic terminal. Although a bit beyond the scope of this study, it would be interesting to determine whether TH mRNA transport is altered by deletion of the glia. The authors might check to see if TH transcripts are reduced in axons by something like RNAscope.