β2-subunit alternative splicing stabilizes Cav2.3 Ca2+ channel activity during continuous midbrain dopamine neuron-like activity

  1. Anita Siller
  2. Nadja T Hofer
  3. Giulia Tomagra
  4. Nicole Burkert
  5. Simon Hess
  6. Julia Benkert
  7. Aisylu Gaifullina
  8. Desiree Spaich
  9. Johanna Duda
  10. Christina Poetschke
  11. Kristina Vilusic
  12. Eva Maria Fritz
  13. Toni Schneider
  14. Peter Kloppenburg
  15. Birgit Liss
  16. Valentina Carabelli
  17. Emilio Carbone
  18. Nadine Jasmin Ortner
  19. Jörg Striessnig

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This paper suggests that assembly of CaV2.3 with b2a/b2e splice variants confers biophysical properties that enable these channels to contribute to calcium-dependent pacemaking in dopaminergic neurons. The findings could have implications for why these neurons are vulnerable to degeneration in Parkinson's disease. The work will be of interest to ion channel biophysicists and neuroscientists.

    (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. The reviewers remained anonymous to the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

In dopaminergic (DA) Substantia nigra (SN) neurons Cav2.3 R-type Ca 2+ -currents contribute to somatodendritic Ca 2+ -oscillations. This activity may contribute to the selective degeneration of these neurons in Parkinson’s disease (PD) since Cav2.3-knockout is neuroprotective in a PD mouse model. Here, we show that in tsA-201-cells the membrane-anchored β2-splice variants β2a and β2e are required to stabilize Cav2.3 gating properties allowing sustained Cav2.3 availability during simulated pacemaking and enhanced Ca 2+ -currents during bursts. We confirmed the expression of β2a- and β2e-subunit transcripts in the mouse SN and in identified SN DA neurons. Patch-clamp recordings of mouse DA midbrain neurons in culture and SN DA neurons in brain slices revealed SNX-482-sensitive R-type Ca 2+ -currents with voltage-dependent gating properties that suggest modulation by β2a- and/or β2e-subunits. Thus, β-subunit alternative splicing may prevent a fraction of Cav2.3 channels from inactivation in continuously active, highly vulnerable SN DA neurons, thereby also supporting Ca 2+ signals contributing to the (patho)physiological role of Cav2.3 channels in PD.

Article activity feed

  1. Version published to 10.7554/elife.67464 on eLife
  2. eLife

    Author Response

    Reviewer #1 (Public Review):

    The results are quite interesting and potentially have important therapeutic implications. Nevertheless, in the current form there are several weaknesses that diminish the strength of the findings.

    1. As the authors note, they do not provide direct evidence for the ultimate conclusion of the study that assembly with β2a and β2e subunits are necessary for CaV2.3 channels to contribute to pacemaking in SN DA neurons. The authors state siRNA knockdown experiments in SN DA neurons are technically challenging. Nevertheless, shRNA knockdown studies in SN neurons have been previously published. Such a study is critical to provide direct evidence for what would be a very important and impactful finding.

    Please refer to our detailed response to essential revision 1 above.

    1. Relative contribution of CaV1.3 (L‐type) and CaV2.3 channels to pacemaking in SN DA neurons. As the authors note, a phase III clinical trial for the L‐type channel blocker, isradipine, showed no efficacy for neuroprotection, even though some mice studies suggested this might be efficacious. On the other hand, the authors' previous work with CaV2.3 knockout mice suggest inhibition of this channel would be more appropriate for a neuroprotective response. It would be useful to get a direct comparison of the impact of isradipine and SNX‐482 on pacemaking in SN DA neurons (Figs. 1 and 2). If their impacts on pacemaking (and Ca2+ oscillations) are similar it would suggest something beyond the pacemaking Ca2+ influx could be responsible for neuroprotection (e.g. changes in NCS‐1 expression as previously suggested by the authors).

    The question about the relative contribution of Cav1.3 and Cav2.3 on pacemaking is complex due to the finding that different results have been obtained regarding the role of L‐type channels on pacemaking. In Cav1.3 knockout mice pacemaking frequency is normal (7, 8). Inhibition (of Cav1.2 and Cav1.3) by dihydropyridine Ca2+ channel inhibitors (e.g. isradipine, nimodipine) was found to inhibit pacemaking in some (e.g. 9‐11) but not in all (8, 12) reports. This seems to be dependent on experimental conditions, but the reasons for these discrepancies are currently unclear. Similarly, we find inhibition of pacemaking by SNX‐482 in cultured midbrain neurons (this paper) but, as previously reported, not in Cav2.3‐deficient mice (1). While this toxin is well suited to isolate Cav2.3‐mediated Ca2+ current components, effects on pacemaking in DA neurons have to be interpreted with more caution because (as clearly outlined in our original MS and our previous paper, 1), SNX‐482 is also a potent inhibitor of Kv4.3 channels. We consider this limitation even more in the discussion of SNX‐482 effects on pacemaking in cultured neurons (data now moved to Suppl Fig. 5) in the revised MS (end of page 15, top of page 16), although the SNX‐482 changes suggest an involvement of Cav2.3 for AP generation.

    Although we acknowledge the relevance of the question raised by the reviewer, based on our previous findings (1) the absence of an obvious role of Cav2.3 for pacemaking in SN DA neurons (despite their role for Ca2+ transients) as an experimental read‐out prevents a straightforward approach to study the contribution of different β‐subunits and their splice variants for this process.

    1. The slice recording data (Fig. 9) are confusing and raise concerns about adequacy of pharmacological isolation of CaV2.3 currents in this preparation. The accuracy of interpretation of the data in Fig. 9 rests critically on the idea that the cocktail of CaV channel blockers given successfully isolates CaV2.3 currents. Yet, the amplitudes of the exemplar currents shown for plus or minus the CaV channel blocker cocktail are almost the same. This cannot be due to CaV2.3 providing the dominant current in the slice preparation since addition of SNX‐482 only decreased Ca2+ current amplitude by 13% (Suppl Fig. 5). It is not clear to me why the steady‐state activation and inactivation curves experiments were not conducted in the cultured neuron preparation (Figs. 1 and 2) where there seems to be better control of pharmacological block of different Cav channel isoforms.

    We have now performed the isolation of SNX‐482sensitive currents not only in the cultured neuron preparation as suggested but, in addition, also in SN DA neurons. The latter experiments gave essentially identical steady‐state inactivation parameters as compared to our "R‐type" current (current remaining in the presence of all other channel blockers). This now also allows a direct comparison of SNX‐482‐sensitive current properties in cultured neurons and in slices (see response above). We now also specifically discuss previous reports of SNX‐482‐sensitive Rtype components in the introduction to allow comparison of these reports with our findings. Please also note that in our legend to Fig. 9A (original MS, now Fig. 6) we have explicitly stated that recordings of "similar amplitudes were chosen" to facilitate comparison of current kinetics. We still think that this makes sense and kept this part of the figure but now strengthened this point even more in the figure legend (Fig. 6).

    1. While the transcript data show that β2a and β2e are present in SN DA neurons, numerically they would still represent only a minority of the beta subunits present (<25%). I don't think sufficient thought has been given to this in the discussion of the results. Unless there is some preferential association of CaV2.3 with β2a and/or β2e, there would be a mix of channels with the majority incapable of supporting pacemaking in SN DA neurons. Given this, one would not necessarily expect that the gating characteristics of CaV2.3 would be the same as what is obtained with reconstituted channels in tsA201 cells where all the channels are assembled with β2a or β2e (see point #5 below).

    We now give this important point more thought in the discussion and mention that our data would imply such a preferential association of Cav2.3 with β2a and/or β2e and provide possible explanations. In addition, as in the original MS, we also provide alternative interpretations (Discussion, pg 14, 2nd and 3rd paragraph).

    1. The V0.5,inact of putative CaV2.3 channels in SN DA neurons of ‐52.4 mV was said to be 'very similar' to the value of ‐40 mV that was observed in tsA201 cells. A difference of +12 mV in voltage‐dependence gating of ion channels is substantial and should not be brushed off. A more nuanced interpretation would be that in SN DA neurons CaV2.3 likely associates with other beta subunits in addition to b2a and b2e and so one would not necessarily expect the V0.5,inact to be the same as what is observed in reconstituted channels in tsA201 cells.

    The V0.5,inact of ‐52.4 mV refers to the control current. We correctly stated that the V0.5,inact of R‐type current was ‐47.5 mV (as also shown in Table 3), i.e. only about 7 mV more negative than in tsA‐cells. We now rephrased this chapter because we also included the new data with inactivation data of SNX‐482sensitive currents in cultured neurons and in SN DA neurons recorded in slices (Discussion, page 13, 2nd paragraph). We do not refer to "'very similar" (difference ~5 mV) values anymore as suggested.

    Reviewer #2 (Public Review):

    This reviewer is very enthusiastic about the work but notes that most of the conclusions are based on data obtained by overexpressing Cav2.3 and accessory subunits in a heterologous expression system. The authors make a good argument for cross‐correlation between data in tsA‐201 cells and dopaminergic neurons, but it is unclear that the results will translate from one system to another. More data may be needed to do so (the reviewer does understand that these are challenging experiments), which the authors acknowledge in a section about the study's limitations. Based on this, it seems that the title is misleading without additional data supporting the role of Cav2.3 in dopaminergic neurons. Along the prior line, statements linking the study results to potential pathological implications seem a big stretch not supported by current data, and therefore should be eliminated.

    An issue with this manuscript is that the narrative and organization of the data are difficult to follow. The reviewer understands that the authors are weaving a complex story that involves using multiple techniques and approaches. Still, the way the data is organized and described makes the reader go back and forward to compare and contrast results constantly. This is further complicated by the fact that some experiments are done in dopaminergic neurons and others in tsA‐201 cells (the identity of the cell type used should be made clearer), the order of some figures is not appropriate (Supp Fig 1 for example) and some figure panels are not discussed (Supp Fig 5E to 5J).

    The MS has been completely rewritten, based on the additional SNX‐482experiments we have now performed both in the cultured DA neurons as well as in the midbrain slices. We therefore also moved data on effects on the spontaneous activity of cultured neurons by SNX‐482 into the supplement to make the key results easier to follow. The identity of neurons is indicated in all headers of table and figure legends to identify cell types. We also changed the title to “β2‐subunit alternative splicing stabilizes Cav2.3 Ca2+ channel activity during continuous midbrain dopamine neuronlike activity” to attenuate our previous statement regarding a role in dopaminergic midbrain neurons.

  3. eLife

    Evaluation Summary:

    This paper suggests that assembly of CaV2.3 with b2a/b2e splice variants confers biophysical properties that enable these channels to contribute to calcium-dependent pacemaking in dopaminergic neurons. The findings could have implications for why these neurons are vulnerable to degeneration in Parkinson's disease. The work will be of interest to ion channel biophysicists and neuroscientists.

    (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. The reviewers remained anonymous to the authors.)

  6. eLife

    Reviewer #1 (Public Review):

    This work follows up on a previous study which found that R-type (CaV2.3) channels contribute to pacemaking and somatodendritic Ca2+ oscillations in substantia nigra dopaminergic (SN DA) neurons. That study showed that SN neurons from CaV2.3 knockout mice were protected from neurodegeneration in a neurotoxin-induced Parkinson's disease model. Here the authors address how CaV2.3 channels are able to contribute to SN DA neuron pacemaking given that what is currently known about their gating suggests they would be mostly inactivated at pacemaking potentials. They first confirm using the blocker SNX that CaV2.3 contributes to pacemaking and whole-cell currents in cultured DA neurons. They further show that recombinant CaV2.3 channels reconstituted in tsA201 cells with auxiliary b2a and b2e, but not b3 or b4, subunits displayed biophysical properties (right-shifted voltage-dependence of steady-state inactivation, decreased rate of inactivation) that permitted significant CaV2.3 activity during pacemaking potentials, inferred from SN DA neuron-derived action potential clamp experiments. They find b2 and b4 transcripts dominate in SN DA neurons and that b2a and b2e splice variants are present. In slice recordings, they find that pharmacologically isolated CaV2.3 currents in SN DA neurons display window currents over the range of voltages at which pacemaking occurs.

    The results are quite interesting and potentially have important therapeutic implications. Nevertheless, in the current form there are several weaknesses that diminish the strength of the findings.

    1. As the authors note, they do not provide direct evidence for the ultimate conclusion of the study that assembly with b2a and b2e subunits are necessary for CaV2.3 channels to contribute to pacemaking in SN DA neurons. The authors state siRNA knockdown experiments in SN DA neurons are technically challenging. Nevertheless, shRNA knockdown studies in SN neurons have been previously published. Such a study is critical to provide direct evidence for what would be a very important and impactful finding.

    2. Relative contribution of CaV1.3 (L-type) and CaV2.3 channels to pacemaking in SN DA neurons. As the authors note, a phase III clinical trial for the L-type channel blocker, isradipine, showed no efficacy for neuroprotection, even though some mice studies suggested this might be efficacious. On the other hand, the authors' previous work with CaV2.3 knockout mice suggest inhibition of this channel would be more appropriate for a neuroprotective response. It would be useful to get a direct comparison of the impact of isradipine and SNX on pacemaking in SN DA neurons (Figs. 1 and 2). If their impacts on pacemaking (an Ca2+ oscillations) are similar it would suggest something beyond the pacemaking Ca2+ influx could be responsible for neuroprotection (e.g. changes in NCS-1 expression as previously suggested by the authors).

    3. The slice recording data (Fig. 9) are confusing and raise concerns about adequacy of pharmacological isolation of CaV2.3 currents in this preparation. The accuracy of interpretation of the data in Fig. 9 rests critically on the idea that the cocktail of CaV channel blockers given successfully isolates CaV2.3 currents. Yet, the amplitudes of the exemplar currents shown for plus or minus the CaV channel blocker cocktail are almost the same. This cannot be due to CaV2.3 providing the dominant current in the slice preparation since addition of SNX only decreased Ca2+ current amplitude by 13% (Suppl Fig. 5). It is not clear to me why the steady-state activation and inactivation curves experiments were not conducted in the cultured neuron preparation (Figs. 1 and 2) where there seems to be better control of pharmacological block of different Cav channel isoforms.

    4. While the transcript data show that b2a and b2e are present in SN DA neurons, numerically they would still represent only a minority of the beta subunits present (<25%). I don't think sufficient thought has been given to this in the discussion of the results. Unless there is some preferential association of CaV2.3 with b2a and/or b2e, there would be a mix of channels with the majority incapable of supporting pacemaking in SN DA neurons. Given this, one would not necessarily expect that the gating characteristics of CaV2.3 would be the same as what is obtained with reconstituted channels in tsA201 cells where all the channels are assembled with b2a or b2e (see point #5 below).

    5. The V0.5,inact of putative CaV2.3 channels in SN DA neurons of -52.4 mV was said to be 'very similar' to the value of -40 mV that was observed in tsA201 cells. A difference of +12 mV in voltage-dependence gating of ion channels is substantial and should not be brushed off. A more nuanced interpretation would be that in SN DA neurons CaV2.3 likely associates with other beta subunits in addition to b2a and b2e and so one would not necessarily expect the V0.5,inact to be the same as what is observed in reconstituted channels in tsA201 cells.

  10. eLife

    Reviewer #2 (Public Review):

    Siller et al seek to define underlying mechanisms by which the R-type calcium channel Cav2.3 may contribute to calcium-dependent pacemaking in dopaminergic neurons. The premise for the work is based on prior observations from this research team, particularly the finding that Cav2.3 is the most abundant Cav expressed in dopaminergic neurons and that it may contribute to activity-dependent calcium oscillations in these cells. Because of the known biophysical properties of Cav2.3, the authors propose a distinctive regulation of Cav2.3 by different accessory beta subunits that may contribute to calcium-dependent pacemaking in dopaminergic neurons. To test this, they first confirmed the sensitivity of spontaneous firing activity to the R-type calcium channel blocker SNX in midbrain dopaminergic neurons. They also recorded SNX-sensitive inward currents suspected to be R-type currents. Electrophysiology in tsA-201 cells expressing Cav2.3, a2d and different beta subunit isoforms and splice variants revealed unique biophysical properties. More importantly, it was found that the b2a and b2e subunits could slow the activity-dependent inactivation time course of the Cav2.3 channels and shift the voltage-dependent of inactivation toward more positive potentials. Subsequent molecular biology experiments revealed the presence of several b subunit splice variants in dopaminergic neurons. An unexpected but appreciated observation is that b2a subunits prevent/ameliorate the rundown of Cav2.3 currents in response to a train of square pulses and simulated neuronal pacemaking protocol in tsA-201 cells. An effort is made to record R-type currents from dopaminergic neurons, and the authors show that these currents activate at more positive voltages than total calcium currents. In general, the data seem of very high quality and support the conclusions that different splice variants of the beta subunit distinctive modulate Cav2.3 activity. Importantly, these changes could help explain a role for Cav2.3-containing b2a/b2e channels in calcium-dependent pacemaking in dopaminergic neurons. The problem examined is of wide interest and the conclusions may have implications that span from the regulation of Cav2.3 channels in dopaminergic neurons to the conceivable involvement of the Cav2.3-b2a-e complex in Parkinson's disease to therapeutic opportunities. Thus, it is likely that this work will be of interest to a broad audience.

    This reviewer is very enthusiastic about the work but notes that most of the conclusions are based on data obtained by overexpressing Cav2.3 and accessory subunits in a heterologous expression system. The authors make a good argument for cross-correlation between data in tsA-201 cells and dopaminergic neurons, but it is unclear that the results will translate from one system to another. More data may be needed to do so (the reviewer does understand that these are challenging experiments), which the authors acknowledge in a section about the study's limitations. Based on this, it seems that the title is misleading without additional data supporting the role of Cav2.3 in dopaminergic neurons. Along the prior line, statements linking the study results to potential pathological implications seem a big stretch not supported by current data, and therefore should be eliminated.

    An issue with this manuscript is that the narrative and organization of the data are difficult to follow. The reviewer understands that the authors are weaving a complex story that involves using multiple techniques and approaches. Still, the way the data is organized and described makes the reader go back and forward to compare and contrast results constantly. This is further complicated by the fact that some experiments are done in dopaminergic neurons and others in tsA-201 cells (the identity of the cell type used should be made clearer), the order of some figures is not appropriate (Supp Fig 1 for example) and some figure panels are not discussed (Supp Fig 5E to 5J).

  12. eLife

    Reviewer #3 (Public Review):

    In this manuscript, Siller et al describe the role of beta subunit variants in facilitating CaV2.3 current at voltages positive to -50 mV. They propose that this could be one of the mechanisms that allow CaV2.3 to contribute to Ca2+ influx during regular pacemaking and burst activities of dopaminergic substantia nigra neurons and thus potentially contribute to the pathophysiology of Parkinson's disease.

    Strengths:
    1. The novelty of this manuscript includes the authors' identification of an important and previously unrecognized role for CaV2.3 in dopaminergic substantia nigra neurons. This discovery could shed light on the pathophysiology of Parkinson's disease.
    2. The authors nicely correlated the expression pattern of beta subunit variants in mouse brain and their functional effects on CaV2.3 currents. In order to precisely measure the effect of beta subunit variants on CaV2.3 currents, they perform voltage clamp experiments in a heterologous system (tsA-201 cells) using physiologic solutions. Based on these results, they demonstrate that β2a and β2e shift CaV2.3 inactivation to more positive voltages, thus allowing channels to remain available to open in the voltage range of physiologic burst activities. To support this claim, they measured the relative levels of transcription of beta subunit variants in mouse substantia niagra and ventral tegmental area using RT-qPCR and demonstrated significant expression of β2a and β2e variants compared to the rest of the beta subunits.
    3. Authors have thoughtfully designed experimental methods and provided rigorous controls which make the results compelling. Moreover, the dosage of channel blockers (SNX-482, isradipine, and Cd2+) used was appropriately chosen to avoid off-target effects and was consistent with what is currently described in the literature.

    Weaknesses:
    1. Although the data is compelling and experimental designs were rigorous, the inference about the potential contribution of the role of beta subunits in Parkinson's disease is still limited due to use of primarily heterologous systems and wild-type neurons. This consideration is appropriately discussed by the authors.
    2. The description of how this study fits into what is already known about beta subunit function is somewhat limited. For example, a considerable amount of literature describes the structure and function of the β2a/β2e interactions with CaV and their impact on various channel subtypes. The paper could be strengthened by a discussion of how the results relate to these previously published studies.

  15. Version published to 10.1101/2021.02.10.430224v2 on bioRxiv
  16. Version published to 10.1101/2021.02.10.430224v1 on bioRxiv