We are extremely grateful to both reviewers for their insightful comments. We attempted to address all of their concerns during our revision of the manuscript, as detailed below (the original comments of the reviewers are included in italics).
The is an excellent and scholarly article from Freund. It provides a wide diversity of “high points” from the field of interneuron research as well as being an accurate primer on the role(s) played by these cells. It does a great job and the author is to be congratulated. I have only a handful of modest comments that I hope the author will take into consideration as the article is revised.
The title “Interneurons” probably needs to be changed. Although much of what is written could apply to generic interneurons it is definitely a cortico-centric view of inhibition. The author does a great job at pointing out points of confluence from different areas of the brain however, I am not sure that spinal cord or most other brain areas are adequately covered in this article to warrant the catch-all title of “interneurons”. Having said that the lions share of research is derived from interneurons of the the hippocampus and cortex so it makes perfect sense to focus on these cell types.
We agree entirely that the article as it stands now is overly "corticocentric" - while we attempted to point out connections to other parts of the brain, this was necessarily limited by our expertise. However, as the nature of this medium allows - and even encourages - contributions from other experts besides the original authors, we do not see this as a fundamental flaw, and explicitly invite corrections, additions, or even full subsections on topics we might have missed, from all qualified scholars, and especially from experts on areas of the nervous system which are currently underrepresented in the paper.
First para, last sentence- “The dominant excitatory drive of interneurons is either local or of extrinsic origin……” I think extrinsic origin needs a little better explanation. As the author is well aware many drivers/providers of feedforward inhibition can have their origins inside the same central structure so a better definition of what is meant by “extrinsic origin” is warranted.
The first two paragraphs of the introduction have been thoroughly revised to address this and other issues.
Section: The postsynaptic targets of cortical interneurons are in most cases principal cells…….. I think this section needs a little clarification. As written, I interpret the take home message being that interneurons target predominantly principal cells and that interneurons only receive a miminum of inhibitory input derived from CR or VIP interneurons. I am sure that is not what you are trying to imply here. Interneurons receive inhibitory input from a variety of other interneurons input to provide a steady barrage of inhibition that is surely not only derived from CR or VIP containing cells????
This part has been rewritten and is now hopefully clearer.
Figure 2 and related text. It may be useful for the passive reader to include some text about the length and area typically covered by interneuron axons. Figure 2 panel B shows nicely the extent of the dendritic and axonal arbor of a number of example cells but it may be useful to give an idea of the physical space typically occupied by some of these interneurons. An example is given later in the section discussing the O-LM/MOPP cells whose axon is described as “rather focused” at 400 x 800um some comparison with other cell types would be helpful for perspective.
The measured extent of the axonal arbors of several other cell types has been included for comparison.
Page 5 Interneurons also differ in the molecular identity and density of voltage-gated ion channels……. It may be relevant in this section to explicitly state that many (but not all ) glutaamtergic synapses onto interneurons signal via GluR2-lacking Ca permeable AMPA receptors; a feature that not only imparts a novel route for Ca entry but more importantly is an important determinant of the rapid synaptic signaling associated with many glutamate synapses onto ints.
A sentence on this feature has been added to the paragraph cited.
Page 6. In the discussion of PV basket cells it may be better to explicitly state what you mean by “driven very efficiently” here, size of synaptic input, epsp-spike coupling etc etc?
By “driven very efficiently”, we mean mainly that changes in the feedforward excitatory input (reflecting the activity of, say, CA3 pyramidal neurons) are mirrored faithfully by the activity of the basket cell population (in CA1). This is supported by, e.g., data on gamma phase coupling, activation during sharp waves, and (perhaps more directly) by the data of Pouille and Scanziani (2001).
Page 7. The sentence “Recent data demonstrate that CCK massively enhances the firing rate of PV basket cells……” seems a little incongruous here. What is the intent of including this sentence here? I would suggest deleting it or clarifying whether you mean endogenous or exogenous applied CCK
When discussing the O-LM cell there is something missing from the sentence that starts “They have a cell body with horizontally oriented dendritic tree confined to the stratum oriens……” please clarify.
The sentence has been rewritten; it is now hopefully clearer.
Some discussion of GABAB receptor targeting interneurons (i.e. the data of Scanziani and Tamas) may be useful.
We have added a section on Ivy cells, which may also preferentially use GABA-B receptors, and also refer to neocortical neurogliaform cells in this context.
This is beautiful overview article on interneurons, which nicely highlights how different interneurons and how they contribute to neuronal networks. I have only minor comments and suggestions:
- Regarding axo-axonic cells: their unique firing during ripples consists of their depression during and after the ripple episodes, which can be observed during almost all episodes; the increased firing before the ripples is not seen for all ripple episodes.
Thank you for the correction, which has been incorporated into the article.
- Regarding CA1 bistratified cells: Although calbindin staining of axon terminals is mainly present in str. oriens and radiatum suggesting an expression of calbindin in bistratified cells, neurobiotin or biocytin-filled bistratified cells have been tested negative for calbindin in their soma and dendrites, but positive for parvalbumin (Pawelzik et al., 2002; Klausberger et al. 2004). This could be explained by a possible expression of calbindin by these cells exclusively in their axons; alternatively, the calbindin axonal labelling might originate from other interneuron classes known to express calbindin, including hippocampal-septal cells, CCK-expressing Schaffer-collateral associated cells, or so far little-investigated calbindin-expressing and CCK-negative cells with large cell bodies. Maybe it is worth mentioning this open question rather then implying that all bistratified cells are calbindin-expressing. Anyway, the in vivo firing patterns that you describe for bistratified cells, have been recorded from calbindin-negative (tested on soma and dendrites), parvalbumin-expressing bistratified cells.
This is now listed as a controversial issue.
- In the legend of Figure 4, I suggest to state that the respective synapses are “enriched” in a1 or a2 subunit containing receptors, as data on any other alpha subunits are still missing.
This is now stated both in the figure caption and in the main text.
- Under “Dendritic inhibitory cells” and describing O-LM cells, I suggest not to use the wording “distinctive marker” for somatostatin, as other interneurons, including hippocampo-septal cells, also express somatostatin.
The word "distinctive" is not used any more, and the absence of calbindin (as a feature that distinguishes O-LM cells from HS cells) is also mentioned.
- Regarding long-range interneurons: It might be worth mentioning that - at least some of - the CA1 hippocampal septal cells also project with myelinated axon to the subiculum (Jinno et al., 2007), which is the main target for CA1 pyramidal cells.
A sentence to this effect has been added to the appropriate section.
- I fully agree that on several occasions incompletely filled hippocampal-septal cells have wrongly been identified as trilaminar cells. However, in two papers (Sik et al., 1995; Ferragutti et al., 2005) cells with very distinctive firing (high frequency bursts), molecular specialisation (receive mGluR8 expressing input) and axonal arborisation (innervating high amount of interneurons, but also pyramidal cell somata and projecting to subiculum) have been described. Because only two (!) such cells have been investigated in detail so far, maybe you want to leave out the controversial issue of trilaminar cells from this general article?
The reference to trilaminar cells has been deleted.