Molecular characteristics and laminar distribution of prefrontal neurons projecting to the mesolimbic system

  1. Ákos Babiczky
  2. Ferenc Matyas

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    This study examined the nature of projections from the prefrontal cortex (PFC) to the nucleus accumbens (NAc) and to the ventral tegmental area (VTA). The authors show that PFC projections to NAc and VTA are largely non-overlapping, originate in different layers of PFC, and express different molecular markers. This study provides high-quality data to the long-standing question.

    (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

Prefrontal cortical influence over the mesolimbic system – including the nucleus accumbens (NAc) and the ventral tegmental area (VTA) – is implicated in various cognitive processes and behavioral malfunctions. The functional versatility of this system could be explained by an underlying anatomical complexity; however, the detailed characterization of the medial prefrontal cortical (mPFC) innervation of the NAc and VTA is still lacking. Therefore, combining classical retrograde and conditional viral tracing techniques with multiple fluorescent immunohistochemistry, we sought to deliver a precise, cell- and layer-specific anatomical description of the cortico-mesolimbic pathways in mice. We demonstrated that NAc- (mPFC NAc ) and VTA-projecting mPFC (mPFC VTA ) populations show different laminar distribution (layers 2/3–5a and 5b–6, respectively) and express different molecular markers. Specifically, calbindin and Ntsr1 are specific to mPFC NAc neurons, while mPFC VTA neurons express high levels of Ctip2 and FoxP2, indicating that these populations are mostly separated at the cellular level. We directly tested this with double retrograde tracing and Canine adenovirus type 2 -mediated viral labeling and found that there is indeed minimal overlap between the two populations. Furthermore, whole-brain analysis revealed that the projection pattern of these populations is also different throughout the brain. Taken together, we demonstrated that the NAc and the VTA are innervated by two, mostly nonoverlapping mPFC populations with different laminar distribution and molecular profile. These results can contribute to the advancement in our understanding of mesocorticolimbic functions and its disorders in future studies.

Article activity feed

  1. Version published to 10.7554/elife.78813 on eLife
  2. eLife

    Author Response

    Reviewer #1: (Public Review):

    “My main quibble is with the framing. There are many places throughout the manuscript where the authors claim that there is a great deal of controversy about the extent of the branching of these neurons.”

    We agree with our reviewer that some of our statements were misleading. Thus, we have rephrased the Introduction (Pages 3-5; Lines 54-99), Results (Page 20; Lines 356-363) and Discussion (Pages 26-32; lines 472-478, 492-498, 525, 530532, 551-578, 586, 596-602) sections to focus on the controversial issues of the simultaneous projection to NAc and VTA by the same prefrontal cortical population.

    Reviewer #2 (Public Review):

    “The scope of this work is somewhat smaller than the recent reconstruction and molecular subtyping of ~6300 neurons performed by others: a comprehensive paper on this very topic ("Single-neuron projectome of mouse prefrontal cortex", Nat Neurosci 25:515”

    The Gao et al. (2022) study published during our reviewing process, indeed, confirms some of our findings. A common finding of ours and Gao et al (2022) is that mPFCNAc and mPFCVTA neurons form distinct classes within the mPFC projecting neuronal population. In addition, there is a small PT-like mPFC population, located rather in the L5b (showing RBP4 expression in our study), which sends branching axons to both NAc and VTA.

    Still, we think that the novelty of our work has remained significant.

    1. We have provided easy-to-use, widely available molecular approaches to investigate mPFC territories and laminar organization. Using the detailed expression pattern of neurochemical markers revealed via multiple immunohistochemical technique and confocal microscopy, we were able to delineate borders between mPFC regions and layers with considerably high precision. For this purpose, mostly brain atlases are used. However, in a cortical region, like the mPFC, territory borders and shapes, as well as laminar thickness and depth are greatly changing at antero-posterior as well as dorso-ventral axes. Therefore, experiment-to-experiment, ‘stable’ markers are necessary to identify the exact location of neurons, recording sites, optic fibre positions, etc; that we, in our opinion, provided in the present study.

    2. Using the presented direct molecular composition, we have identified genetic markers for selective examination of layer (and at some extent, region) -specific mPFCNAc and mPFCVTA populations.

    3. We have also provided evidence for the utility of this characterization using Cre mouse lines. The use of Calb1-, Rbp4-, Ntsr1- and FoxP2-Cre, which strains are widely used in cortical studies, in an intersectional approach, allow scientist to selectively investigate each of these mPFC populations, even in a target-selective manner via an intersectional approach.

    Major points:

    ”…But what I think is lacking here is a characterization in a region-by-region manner of the laminar organization of the cell types you either identify by retrograde label (CAV-Cre anatomy, for example) or by molecular approaches (how the lamination of Ntsr1+ neurons vary between the areas you lump together here as PFC).”

    ”…I think this subdivision might help by defining these areas in stereotaxic coordinates and giving some idea of how defined cell types (defined by Cre driver or retrograde label or other marker) might vary in their laminar distribution across these areas. Maybe I am wrong, but my perception of Fig 1 and 2 is mainly that the laminar pattern of cortical labeling from VTA and NAc varies somewhat depending on where you assess it in cortex...”

    We have plotted the location of the retrogradely labeled mPFCNAc and mPFCVTA cells which clearly shows their characteristic laminar and regional distributions. These panels are added to Figure 1-2.

    The lamination of NTSR1 neurons is remarkable, indeed. As it can also be seen on the GENSAT website (http://www.gensat.org/imagenavigator.jsp?imageID=48699), the L6 of PrL, IL and the more ventral regions lack NTSR1-expressing cells, while L6 of the cingulate and motor cortices contain a moderate density of this cell type.
    In our experiment, in which we aimed to target Prl-IL-MO (as majority of the retrogradely labeled mPFCNAc and mPFCVTA cells were located there), we could not detect any viral labeling in any of the layers ventral to PrL cortex (Figure 3). The labeling in the cingulate cortex is missing, probably, due to the lack of AVV diffusion into its deep layer, L6. However, as it can be seen in Figure 3–figure supplement 2H, the NTSR1-expressing the L6 neurons in M1 are also present in our trials. Altogether, it means that NTSR1 is only present in L5a of PrL cortex among the cortical regions which contain mesolimbic-projecting neurons. The shift in the distribution of Ntsr1expressing cells is present between PrL and Cg cortices.

  3. eLife

    Evaluation Summary:

    This study examined the nature of projections from the prefrontal cortex (PFC) to the nucleus accumbens (NAc) and to the ventral tegmental area (VTA). The authors show that PFC projections to NAc and VTA are largely non-overlapping, originate in different layers of PFC, and express different molecular markers. This study provides high-quality data to the long-standing question.

    (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.)

  4. eLife

    Reviewer #1 (Public Review):

    This is a lovely study focusing on the anatomical connectivity between the medial prefrontal cortex (mPFC) and the nucleus accumbens (NAc) and ventral tegmental area (VTA). The case for focusing on this circuitry is strong, with these regions' central role in motivation and reward processing. Prior work had looked at whether mPFC->NAc and mPFC->VTA neurons are separable populations, and what their other features may be. But here, the authors bring to bear a host of converging evidence, often replicating main results multiple times with different techniques, to establish the pattern: These are largely (but not wholly) separable populations, with distinct laminar and molecular profiles. I think this paper has substantial value to the field.

    My main quibble is with the framing. There are many places throughout the manuscript where the authors claim that there is a great deal of controversy about the extent of the branching of these neurons. That is true, if you ask about *all* mPFC neurons. However, the vast majority of their citations either do not look at mPFC->NAc or do not look at mPFC->VTA. Instead, they look at mPFC projections to other cortical regions, thalamus, amgydala, etc. So I don't think the contradiction is really as deep as what they suggest. It would be best to reframe those portions of the paper just about prior evidence for mPFC->NAc/mPFC->VTA neurons, not all of the others. Then, there will be plenty of space to do a deep dive into the few papers on this topic, which do actually contradict each other (particularly Gao et al., 2020 vs Pinto & Sesack 2000). Then, the broader discussion of IT vs PT cell types and other projections (amygdala, thalamus, and so on) can be shortened and mainly live in the Discussion.

  5. eLife

    Reviewer #2 (Public Review):

    Here the authors address the connections by which medial prefrontal cortex (PFC), a frontal brain area that provides input to the limbic system, targets nucleus accumbens (NAc, a ventral striatal region) and the ventral tegmental area (VTA, a midbrain dopaminergic region involves in reward). They combine chemical retrograde tracers and conditional viruses to study connectivity. The data suggest that PFC projections to NAc and VTA are mostly from separate cell types, since these outputs (a) originate in different layers of PFC, (b) express different biochemical markers (making these cell types molecularly distinct), and (c) have minimal overlap (in, for example double retrograde label experiments). Thus, the authors show that PFC outputs to two different limbic system components come from two different parts of the PFC circuitry, thus potentially conveying different information to subcortical brain areas.

    The study is done with high technical precision. The figures convey the findings and the clarity of thought effectively. Overall, I am convinced that the data support the conclusion that NAc and VTA projecting cells within PFC have "different laminar distribution (layers 2/3-5a and 5b-6, respectively) and ... different molecular markers". The larger claim that the authors "deliver a precise, cell- and layer-specific anatomical description of the cortico-mesolimbic pathways" is mostly accomplished. I feel this would be stronger if the different regions of the PFC were treating as distinct instead of one entity, as the output to VTA and NAc in each of the (potentially different) areas (PrL, IL, Cg, MO, &c) might differ in cell type and layer and such regional differences across cortex are addressed in other studies. This study contributes to understanding the anatomy underpinning earlier work that addresses the distinct functional and behavioral roles of NAc and VTA-projecting neurons.

    The scope of this work is somewhat smaller than the recent reconstruction and molecular subtyping of ~6300 neurons performed by others: a comprehensive paper on this very topic ("Single-neuron projectome of mouse prefrontal cortex", Nat Neurosci 25:515):
    https://www.nature.com/articles/s41593-022-01041-5
    These reconstructed individual axons originate across a range of mouse PFC areas, and the paper quantifies their targets and classifies them into 64 cell types defined by projection class. To some extent, this covers the IT types that project to striatum/nucleus accumbens (Fig 3) and provides a distribution of where they reside within the different PFC areas (Al, MO, M2, etc ... ). (See the top and bottom of figures 3a and 4a for many PT-types). Furthermore, the full transcriptome of all these cell types is examined and compared to projection pattern (Fig 7+). For what it's worth, playing with the visualization tool confirms the main points in the current manuscript:
    https://mouse.braindatacenter.cn/
    Displaying cells in a given IT-type projection group (group 21, I tried the first 20 cells) that project to NAc, confirms they don't project to VTA. Displaying cells in a given PT-type projection group (group 57, I tried the first 20 cells) that project to VTA, confirms they don't project to NAc. Just blown away by this, didn't even take 15 minutes to use it. I am not sure how to suggest that the current paper address this (e.g. how can they differentiate what they're showing from this somewhat complete projectome of IT and PT-type cells?), but this work should at least be pointed to in terms of addressing many of the same issues. If a counter-example to the current work is desired, look at cell group 59 (Fig 4a suggests this population projects to both ACB and VTA; examination of these cells with the visualization tool suggests there are ~122 examples of cells that project to both to some small extent.)

    Major points:
    The PFC areas studied here may include a heterogeneous group that differs in stereotaxic location, laminar organization, and projection pattern. In "Anatomical considerations" in the discussion, Line 547: "the exact definition of the PrL subregions greatly varies between publications, just like the distinction between dorsal and ventral mPFC. Such inaccuracies can contribute to the still abundant contradictions in the literature and complicate the proper interpretation of the results." I agree. But what I think is lacking here is a characterization in a region-by-region manner of the laminar organization of the cell types you either identify by retrograde label (CAV-Cre anatomy, for example) or by molecular approaches (how the lamination of Ntsr1+ neurons vary between the areas you lump together here as PFC).

    I think this subdivision might help by defining these areas in stereotaxic coordinates and giving some idea of how defined cell types (defined by Cre driver or retrograde label or other marker) might vary in their laminar distribution across these areas. Maybe I am wrong, but my perception of Fig 1 and 2 is mainly that the laminar pattern of cortical labeling from VTA and NAc varies somewhat depending on where you assess it in cortex.

    The degree to which the result is novel depends somewhat on the credence given to prior efforts to unravel this connectivity (Line 494-502). In addition to the single axon reconstructions mentioned above, retrograde tracing with CAV-Cre (and HSV-flp) suggested that the PFC populations projecting to VTA and NAc were anatomically and molecularly distinct (Kim et al., 2017 Cell), with the VTA projections originating from neurons in deeper layers (further from the midline, Fig. 1) - as shown here. They do show that mPFC output has unique laminar origin (PFC-to-NAc is L5A, PFC-to-VTA is L5B, the retrograde tracing of Fig 1) and some molecular differences (VTA outputs express CTIP2, TCERG1L, and CHST8; NaC outputs express NPTX2, NRN1, and SCCPDH). This work devotes far less effort to the anatomical characterization that is presented quite beautifully here, instead addressing behavioral roles for these populations. Further, there is some prior work to suggest overlap in a subset of layer 5 cells (For example, NAc and VTA projecting neurons shown in rats (not mice as here); Gao et. Al, 2020 Neurobio Dis.).

  6. Version published to 10.1101/2022.02.17.480862v1 on bioRxiv