The differentiation and integration of the hippocampal dorsoventral axis are controlled by two nuclear receptor genes

Reviewer #1 (Public Review):

The hippocampus is a structure in the cerebral cortex known to be compartmentalised into regions with different functions. Dorsal hippocampus is involved in cognitive functions such as declarative memory and spatial navigation and interconnects chiefly with the neocortex. Ventral hippocampus interconnects with limbic structures such as amygdala and hypothalamus and is involved in affective states and anxiety. What specifies this functional regionalisation during development is not well understood. The present study focuses on the role of transcription factors COUPTFI and COUPTFII, confirming a previously observed dorsal to ventral gradient of expression of COUPTFI in both embryonic and adult mouse hippocampus, and reporting that expression of COUPTFII is strongest in ventral hippocampus. The aim of the authors was then to probe the role of these transcription factors with the use of conditional knockout of one or both factors using RxCre+ mice (sometimes Emx1Cre+ for comparison). As predicted, COUPTFI insufficiency resulted in failure of the CA1 subregion of the dorsal hippocampus to develop properly (with concomitant loss of performance in a spatial memory task) COUPTFII knockdown had even more marked effects upon the ventral hippocampus with ectopic CA1/CA3 domains forming, while a double knockout lead to a drastic reduction in size of the hippocampus with subsequent effects upon the appearance of hippocampal synaptic circuitry and the capacity for adult neurogenesis (a feature of rodent hippocampus). In order to help explain the role of COUPTFI/II a role in regulating expression of two transcription factors LHX2 and LHX5, known to be crucial to hippocampal development, was tested by examining gene and protein expression. Changes in LHX2 and LHX5 was observed and a role for COUPTFI/II in regulating expression of these genes was postulated.

I believe the authors have largely achieved their aims and the results mostly support the conclusions, but, as discussed further below, there are some weaknesses in the data and some areas that could be expanded upon and improved. The methods are mostly appropriate. The use of the transgenic mice and the application of histological methods, especially tyramide amplified immunohistochemistry, is exemplary. However, I'm not sure a wide enough range of tests to explore the phenotype of the transgenic mice was employed to back the conclusions drawn by the authors. The introduction and discussion are nicely written and explain the general concepts and conclusions well. The work makes an important contribution to our understanding of brain development in general and hippocampal development in particular.

Turning to more specific comments, I must first point out that specification of the ventral hippocampus by expression of COUPTFII is not an entirely original finding, as it was suggested for the developing human hippocampus following immunohistochemical experiments illustrating COUPTFII expression to be confined to the ventral hippocampal structures of the medial temporal cortex (doi: 10.1093/cercor/bhx185). Of course, this study, unlike the present study, was restricted to fetal cortex, not adult, and also reported expression of COUP-TFI throughout dorsal and ventral hippocampal structures but without observing any dorsal to ventral gradient, however I feel its contribution to the field has been overlooked by the present study, and should be incorporated into the introduction and/or discussion.

More information about Rx-cre mice would be informative and could help explain the different phenotype observed when EMX1-cre mice were used to conditionally knock down COUPTFI/II expression.

The demonstration of antagonistic gradients of COUP-TFI and -TFII across the hippocampus is more convincing in the immunohistochemical preparations than in the western blots. The qualitative data presented in Fig.1p does not convincingly represent the quantitative data presented in Fig.1q. There seem to be multiple bands for COUP-TFII and I wonder exactly how quantifying this was approached?

Behavioural testing is limited to one test of dorsal hippocampus function. other tests for non-spatial memory, e.g. novel object recognition, or ventral hippocampus function, e.g. step through passive avoidance, might have lead to some interesting discriminations between the various knock down animals (see doi: 10.3389/fnagi.2018.00091).

Abnormalities in the trisynaptic circuit. No studies of actual synapses, either physiological or morphological, were carried out. I wonder to what extent these immunohistochemical studies just further reflect the abnormalities in hippocampal morphology presented earlier in the manuscript without specifically telling us about synaptic circuits? Although the immunohistochemical preparations are beautiful, they are inadequate on their own in telling us much about what sort of synaptic circuitry exists in the transgenic animals.

LHX2/LHX5 interaction. The immunohistochemical study, which shows clear differences in LHX5 and LHX2 protein expression at E14.5 in double knockdown mice is more convincing than the qPCR study at E11.5, which show surprisingly small differences in mRNA expression. Could the authors expand upon whether this is due to stage of development, or differences between mRNA and protein expression? Why hasn't both mRNA and protein expression data at both time points been presented?

Reviewer #2 (Public Review):

The authors Yang et al., examine the role of NR2F1/COUPTFI and NR2F2/COUPTF2 genes in hippocampus (HP) development, using two Cre lines, RxCre, and Emx1Cre. They report that loss of COUPTFI leads to a defective specification of dorsal CA1; loss of COUPTF2 leads to defects in the morphogenesis of the ventral HP with some ectopic CA field domains; loss of both results in a greatly shrunken hippocampus.

While the phenotypes are indeed interesting and important to examine carefully, there are major lacunae in (A) the authors' interpretation of the literature that sets up the problem (B) the data itself and the experimental design (C) the interpretation of the data. These are detailed below.

[A] Interpretation of the literature
A1: The author's interpretation of the Lhx5 mutant phenotype (line 74-76) missed the fact that the hem appears to be missing or greatly reduced (Zhao et al., 1999; Figure 4D,I; Miquelajáuregui et al., 2010 Figure 5). If the hem is deficient, shrinkage/ agenesis of hippocampus is not surprising. It is incorrect to conclude that Lhx5 has a role in the hippocampal primordium, not only because of the above, but also because Lhx5 expression has been well characterized to be limited to the early hem and CR cells, but is not known to be expressed in the hippocampal primordium. The immunohistochemistry data in Figure 5B showing Lhx5 presence in the vz of the hippocampal and neocortical primordium is perplexing and not what other studies in the literature show for this gene. This is a major point because "regulation of the Lhx2-Lhx5 axis" is one of the main conclusions of the study.

A2: The Lhx2<->Lhx5 inhibition is pitched as a mechanism, but there's no evidence in the literature for this nor in this study. Lines 78-79 "Intriguingly, deficiency of either Lhx5 or Lhx2 results in agenesis of the hippocampus, and more particularly, these genes inhibit each other" are an incorrect interpretation of the literature. The "agenesis" of the hippocampus in the Lhx5 mutant (Zhao et al., 1999) is likely to be because the hem is deficient (point A1 above). The Lhx2 mutant lacks a hippocampus (and neocortex) because the entire dorsal telencephalon has transformed into hem and antihem (Mangale et al., 2008). To cite this as "agenesis of the hippocampus" as originally described by Porter et al (1997) misinterprets a complex stepwise process that was elucidated subsequently in the literature.

Finally, it has not been shown that Lhx2 and Lhx5 inhibit each other- the literature cited does not contain this information. The phenotype reported by the authors may actually have a basis in the effect of loss of COUPTFI/ II on the hem, and a rostro-caudal variation in this effect (or in the timing of action of the Cre lines used) may explain the phenotype.

Problems in the experimental design:
B1: What is the expression domain and timing of RxCre? If it has a dorso-ventral bias in the early embryo, it could explain the regional difference in the COUPTF phenotypes. The authors must show the domain of Cre activation using an Ai9 reporter at E10.5-E11.5 and also at later embryonic stages to be able to interpret whether the shrunken hippocampal phenotype in the single and double mutants is a due to a defect in induction (from the hem), specification (in the early hippocampal primordium), or growth and maintenance (at later embryonic/ postnatal stages). A related point is whether COUPTFI expressed in the hem at E10.5-E11.5, since the earliest age shown is E14.5 which does show expression in the hem; likewise COUPTFII is shown to be expressed in the hem at E12.5. Emx1Cre acts in the hem and therefore the phenotypes could be partially explained by a deficit in the hem itself. Where RxCre acts is not shown and nor is it cited and the logic of shifting between RxCre and Emx1Cre is not clear. A comparison of the expression domains of these lines at relevant early and late embryonic ages is important.

B2:
Line 187: "We would like to investigate the correlation of the CH and/or amygdala anlage with the duplicated ventral hippocampal domains in the COUP-TFII mutant in detail in our future study."
This is inadequate, the effect of the mutation on the cortical hem may be central to the hippocampal phenotype and therefore is central to this study. Ectopic CA fields arising in unexpected places is a finding that needs an explanation, this is not a mere morphogenesis issue as implied in line 190.

B3: Questionable immunofluoresence data: Figure 5B panel h shows that Lhx2 expression extends into the region of the hem at E14.5, suggesting that the hem may in fact not have been specified in the first place. However, the choroid plexus appears to be LHX2 positive in the same image, which it isn't supposed to be, and this calls into question the quality and specificity of the immunofluoresence data. LHX5 staining in Figure 5B panel has been mentioned in point A1- it does not reflect the known expression pattern of this gene (Allen Brain atlas, Zhao et al., 2009). SOX2 also shouldn't be seen in the choroid plexus.

[C] Interpretation of the data
C1: In the COUPTFII mutant, the ectopic presence of HuB+ve cells is intriguing, however it is a stretch to conclude that these cells are born at the expense of CTIP2+ve cells (line 179) without experiments that examine this point.

C2: Line 251: "Unexpectedly, an ectopic nucleus was observed in the region of the prospected temporal hippocampus, indicated by the arrowhead, in the double-mutant mice (Figure 3Ag, h)"
These data are unclear and difficult to appreciate.

C3: The hippocampus is shrunken in the double mutants but the underlying cause has not been examined from the perspective of early cell cycle exit or cell death. How does the reduction of Tbr2+ and NeuroD1+ cells speak to the hippocampal defect? (Figure 5)

Reviewer #3 (Public Review):

In this manuscript, Yang et al. showed that two nuclear receptor genes, COUP-TFI and -TFII, displayed distinct expression patterns and functions during the development of the dorsal and ventral hippocampus. The phenotypes in the presented single and double conditional knockout mice are striking and intriguing, which expands our knowledge of hippocampus development, especially the ventral part. Nevertheless, the manuscript is a bit descriptive without in-depth molecular mechanisms.

My major concerns as follows:

1. Quantification and statistical analysis to support their conclusions are almost absent throughout the whole manuscript, especially in relation to the numbers of DG, CA1, and CA3 neurons.
2. Only TFI conditional knockout mice, not TFII knockout mice, were used to test for behavioral abnormalities. It is important to determine whether the abnormal ventral hippocampus in TFII loss leads to any psychiatric illness.
3. Behavior defects were only tested on TFI conditional knockout mice but not on TFII knockout mice. TFII loss predominantly affects the ventral hippocampus which is involved in psychiatric disorders, and this should be tested.