Summary of ICMS Journal Club presentation on Friday, July 30, 2021.
Title: Molecular mechanism for reward-related gene expression and behaviors by dopamine signaling
Speaker: Dr. Yasuhiro Funahashi, Senior Assistant Professor, Research Project for Neural and Tumor Signaling, Institute for Comprehensive Medical Science, Fujita Health University
On Friday, July 30, by Zoom, Dr. Funahashi gave a presentation to the Journal Club of the Institute for Comprehensive Medical Science (ICMS) of Fujita Health University. He discussed the function of the neurotransmitter dopamine (sometimes called “the happy hormone”) in a region of the brain called Nucleus accumbens (NAc). He gave a general introduction to the topic plus a summary of two of his recent papers. Each of these two papers revealed how dopamine signaling induced the phosphorylation of a transcription factor (Npas4 and Mkl2, respectively) in dopamine 1 receptor (D1R-) positive medium spiny neurons (MSNs) in the NAc:
1. Funahashi Y, Ariza A, Emi R, Xu Y, Shan W, Suzuki K, Kozawa S, Ahammad RU, Wu M, Takano T, Yura Y, Kuroda K, Nagai T, Amano M, Yamada K, Kaibuchi K. Phosphorylation of Npas4 by MAPK Regulates Reward-Related Gene Expression and Behaviors. Cell Rep. 2019 Dec 3;29(10):3235-3252.
2.Ariza A, Funahashi Y (shared first author), Kozawa S, Omar Faruk M, Nagai T, Amano M, Kaibuchi K. Dynamic subcellular localization and transcription activity of the SRF cofactor MKL2 in the striatum are regulated by MAPK. J Neurochem. 2021 Jun;157(6):1774-1788.
There were 18 participants, which is slightly more than our usual audience. By request of Dr. Funahashi, after introducing him, I gave a preview type of introduction to his presentation. Funahashi then gave a beautiful introduction into dopamine-signaling, after which he described in detail the Funahashi et al. 2019 paper and briefly summarized the Ariza, Funahashi et al. 2021 paper. This was followed by an engaging discussion in which six different participants asked questions (see the end of this post) and/or had comments. Overall, I felt that the event was highly appreciated by those participants with at least some knowledge about neuroscience, whereas the other part of the audience enjoyed the general part but had difficulties understanding the details. As for any journal club event, preparing for the meeting by reading the recommended article beforehand can help prohibit those difficulties.
The contents of the presentation
The nucleus accumbens (NAc) is part of the striatum and a central region of the brain used for validation. Neural activities associated with experiencing, remembering, and planning, which involve multiple different brain regions, can all be directed through the NAc for this validation. The measure of validation is the concentration of dopamine released by axons form neurons whose cell bodies reside in the ventral tegmental area (VTA) (for locations see Fig. 1). Within the NAc, the delivery of dopamine is locally not very precise because the number of dopaminergic VTA neurons is low, and each of these extensively branched cells releases dopamine to many other neurons. Thus, dopamine levels in the NAc seem to represent a rather general evaluation of anything what someone is experiencing/thinking/remembering at a given moment, making the region perfectly suited for learning by association.
Dopamine levels peak at moments of excitement, when something unexpectedly positive happens or can be expected to happen based on a cue (e.g., Pavlov’s dog). The dopamine peaks promote feelings of excitement and energy, which is why recreational drugs that increase their concentration such as cocaine or alcohol are popular. On the other hand, if something more negative than expected happens, dopamine levels decrease below baseline. If the dopamine levels are continuously low, as often found in depressed patients, people feel as if not having enough energy to engage in any activity.
Most neurons in the NAc are medium spiny neurons (MSNs; for their morphology see Fig. 1) and the vast majority of those either express the D1 dopamine receptor (D1R) or D2R. D1R is a stimulatory receptor and D1R-MSNs are stimulated if dopamine levels increase above baseline. This is associated with feelings of positivity and activity. On the other hand, D2R is an inhibitory receptor, and D2R-MSNs are stimulated if dopamine levels decrease below baseline. This is associated with negative feelings and inactivity. Based on how the respective pathways relay through other brain regions, the D1R-MSNs are part of what is called the “direct pathway” while the D2R-MSNs are part of what is called the “indirect pathway.” The direct and indirect pathways ultimately target the same neural activities and promote and inhibit them, respectively—this dual system explains how the dopamine levels in the NAc influence the possibility of engaging in an activity.
Biologically, dopamine levels are not intended to make us feel good but to help us to adjust our behavior for optimizing rewards. Therefore, after the dopamine rewards have helped us to adjust our behavior so that rewarding events become habitual (e.g., marrying the person you fell in love with or going to the same great bakery shop every day), the dopamine rewards (feelings of excitement) dwindle. At the neurocircuit level, habitualization of a rewarding behavior requires the strengthening of neural pathways associated with the behavior. This requires the building/strengthening of neural synapses, which requires protein synthesis, which in turn requires the activation of transcription factors. The finding of activated transcription factors or transcriptional coactivators in dopamine-stimulated D1R-MSNs was the topic of Dr. Funahashi’s studies.
The conditional place preference (CPP) test system for activating dopamine pathways in the mouse NAc
The excitement and energy that people and mice experience through cocaine derives from this drug directly increasing the effective dopamine concentrations in the NAc. Therefore, cocaine administration to mice is a prime experimental tool for studying dopamine function.
Dr. Funahashi used a two-chamber system in which mice learned that in one chamber they would receive an injection with saline (the control) whereas in the other they would receive an injection with cocaine (Fig. 2). Then, on the day following a three-day learning period, in a 15 min period in which the mouse could move freely between the rooms, the difference in the times spent between the cocaine-chamber and saline-chamber was measured as “Conditioned Place Preference (CPP)” score. The CPP score increases with increasing cocaine doses.
The initial screening for interesting transcriptional factors and the decision to study Npas4 and Mkl2
Dr. Funahashi then applied an assay to find potential transcription factors activated by dopamine in the mouse striatum (Fig. 3). His strategy aimed at finding proteins in striatal cell nuclei that bound to CBP after cocaine stimulation. CBP [cAMP-response element-binding protein (CREB) Binding Protein] is a transcriptional coactivator that binds many different transcription factors. By mass spectrometry analysis of the CBP-binding proteins in the nuclear extracts from striatum tissue from mice injected with cocaine, Dr. Funahashi found >400 proteins (Figs. 3 and 4). The isolation of CREB (Fig. 3 bottom-right and Fig. 4), which is a transcription factor that is stimulated in D1R-MSN by dopamine and known to bind CBP, provides evidence that the assay worked. After screening the >400 isolated proteins for novel transcription factors or transcriptional coactivators likely to play a role in D1R-MSN neural plasticity, Dr. Funahashi and coworkers decided to focus on Neuronal Per Arnt Sim domain protein 4 (Npas4) and MKL/myocardin-like protein 2 (Mkl2).
What was already known about Npas4 at the start of Dr. Funahashi’s study was:
- Npas4 is a brain-specific basic helix-loop-helix transcriptional factor that is expressed throughout the whole brain at a low level in the resting state, though it is enriched in the limbic system, such as the hippocampus, amygdala, and NAc.
- Npas4 plays a role in the expression of activity-dependent genes, such as c-Fos and brain-derived neurotrophic factor (BDNF) to control synaptic plasticity.
- Npas4 is required for normal social interaction and contextual memory formation in mice.
What was already known about Mkl2 was:
- Mkl2 is enriched in the brain.
- Mkl2 is a coactivator of the transcription factor SRF. SRF is essential for cocaine-mediated spine morphogenesis in the NAc and reward behavior.
In cocaine-stimulated mice, Npas4 and CBP colocalize in the nuclei of NAc D1R-MSNs
The amount of Npas4 bound to CBP in the mouse striatum was higher after cocaine stimulation (compare CPP versus Home in the left part of Fig. 5), in agreement with the hypothesis that CBP-Npas4 interaction was promoted by dopamine. For analyzing the location of Npas4 and CBP at the cell-type level, Dr. Funahashi used two transgenic mouse strains which expressed the YFP marker protein under either the D1R or D2R promoter. After cocaine stimulation, in the NAc of these mice, immunohistochemistry revealed that Npas4 and CBP colocalized in the nucleus of some D1R-MSN and D2R-MSN (Fig. 5 micrographs; YFP is indirectly stained as green by using antibodies); among the anti-Npas4-stained MSNs in the NAC, >75% were D1R-MSNs and <25% were D2R-MSNs (Fig. 5 upper-right).
Npas4 is phosphorylated by MAPK, and this phosphorylation induces Npas4 accumulation in the nucleus
Dr. Funahashi then performed experiments to better understand the nature of Nsap4 activation and its binding to CBP. This blog will not describe or show all details of those analyses.
He identified the interacting regions of CBP and Npas4 by pull-down assays using various Npas4 and CBP fragments. This showed that CBP-Npas4 binding concerned the N-terminal transactivation domain of CBP and the transactivation domain and a middle region of Npas4 (for details see Funahashi et al. 2019).
Because CREB binds CBP better after CREB activation by phosphorylation, Dr. Funahashi hypothesized that also Npas4 might be activated by phosphorylation and the thereby improved binding to CBP. Therefore, he first analyzed—using transfected COS7 cells—how a phosphatase inhibitor and several specific inhibitors of different kinases affected the binding of Npas4 to CBP in a GST-pull-down assay. The combined data revealed that phosphatase inhibition increased Npas4 phosphorylation levels and the binding of Npas4 to CBP and that these processes could be reversed by U0126 which inhibits MAP2K1/2 kinases (for details see Funahashi et al. 2019). Thus, phosphorylation of Npas4 improves its binding to CBP.
For confirmation, by cotransfection of Npas4 and constitutively active MAP2K1 to COS7 cells, Dr. Funahashi then showed that MAP2K1 promoted Npas4 phosphorylation (for details see Funahashi et al. 2019). MAP2K kinases phosphorylate the kinase MAPK (see the drawing at the right in Fig. 6), and by in vitro incubation of Npas4 and MAPK1, Dr. Funahashi provided direct evidence for phosphorylation of Npas4 by MAPK (for details see Funahashi et al. 2019). From previous studies, MAPK was already known to be phosphorylated in D1R-MSN as part of the cocaine-induced cAMP signaling pathway downstream of D1R. Thus, finding MAPK as the responsible kinase for Npas4 phosphorylation made perfect sense.
Then Dr. Funahashi determined that Npas4 has six S/T phosphorylation sites by analyzing how partial or mutated Npas4 could be phosphorylated and by performing mass spectrometry analysis (for details see Funahashi et al. 2019). Dr. Funahashi and coworkers succeeded in establishing specific antibodies for the phosphorylated state of four of the six sites, with the antibodies recognizing the phosphorylated state of Npas4 residue T247 showing the highest specificity (for details see Funahashi et al. 2019). That is why in many assays anti-pT247 was chosen to represent Npas4 phosphorylation.
Fig. 6 is dedicated to an experiment using a mouse striatal neuron cell line that was stimulated by forskolin (FSK), which is an inducer of cAMP and in D1R-MSNs is known to induce the phosphorylation of MAPK (Nagai et al. 2016). Fig. 6 exemplifies some of Dr. Funahashi’s findings listed above, but in a more relevant setting (instead of in COS7 cells or in vitro), while additionally indicating that phosphorylation of Npas4 enhanced its accumulation in the nucleus.
Western blot results, with examples and histogram quantifications shown in the middle left of Fig. 6, reveal that FSK induces phosphorylation of Npas4 (represented by T427 phosphorylation), Rasgrp2, and MAPK. Furthermore, these results show that only for Npas4 and MAPK those effects are reversed by U0126 which inhibits MAP2K kinases, which agrees with Npas4 being phosphorylated by MAPK (downstream of MAP2K). For making the micrographs at the bottom left of Fig. 6, because the amounts of endogenous Npas4 were too low for immunocytochemistry analysis, the striatal neural cells were transfected by a GFP-Npas4 fusion. The micrographs reveal that FSK induces phosphorylation of GFP-Npas4 and induce its accumulation in the cell nucleus.
The drawing at the right of Fig. 6 summarizes Dr. Funahashi’s conclusions about the dopamine-induced cascade in D1R-MSNs that lead to the activation of Npas4 through its phosphorylation by MAPK.
Phosphorylation of Npas4 enhances its transcriptional activity
Transcriptional activity of the transcription factor Npas4 was analyzed by using recombinant COS7 cells in which a luciferase gene was placed behind the promoter of BDNF (brain-derived neurotrophic factor) gene. Transfection of these cells for expression of Npas4 gave a large increase in luciferase expression, especially if the cells were cotransfected for expression of constitutively active (CA) form of MAP2K1 or stimulated by FSK (Fig. 7 upper histograms). Additionally, the importance of Npas4 phosphorylation in this luciferase reporter assay was checked by transfecting the cells with GFP-Npas4 wildtype, or with variants thereof in which the serines and threonines of the six phosphorylation sites were replaced by either alanine (the “6A” mutant) or glutamic acid (the “6E” mutant). The lower histogram in Fig. 7 shows that the prohibition of Npas4 phosphorylation in the 6A variant reduced the Npas4 stimulatory effect on expression from the BDNF promoter, whereas use of the 6E variant, believed to functionally represent constitutively phosphorylated Npas4, increased the stimulatory effect. Thus, phosphorylation of Npas4 increases its ability to promote transcription from the BDNF promoter, and this may lead to neural plasticity involved in learning and memory (Fig. 7 drawing at the right).
Npas4 and its phosphorylation are important for reward-related learning and memory
Finally, Dr. Funahashi provided evidence that Npas4 and its phosphorylation are important for reward-related learning and memory. He did this by measuring the cocaine-induced CPP (conditioned place preference) scores in the above described two-chamber system (shown again at the top right of Fig. 8) after modification of Npas4 expression:
- In global Npas4 knockout mice compared to wildtype, the CPP scores were reduced >60% (see Funahashi et al. 2019).
- Specific expression of a dominant negative form of Npas4 in striatum D1R-MSNs by using a Cre-Lox recombination system resulted in CPP scores that were >50% lower than in control mice. Meanwhile, the expression of this Npas4 form in D2R-MSNs ha no significant impact on CPP scores (Fig. 8 middle/right bottom).
- Specific deletion of Npas4 in striatum D1R-MSNs of Npas4 conditional knockout mouse (Npas4 fl/fl) by using a Cre-Lox recombination system resulted in CPP scores that were >80% lower than in control mice (see Funahashi et al. 2019).
- In these mice (described in 3) cocaine-induced CPP was restored by the expression of Npas4-WT, but not Npas4-6A, which is evidence for the importance of Npas4 phosphorylation in learning and memory (see Funahashi et al. 2019).
Thus, using the CPP assay system, Dr. Funahashi showed in vivo that Npas4, and phosphorylation of Npas4, are important for reward-related learning and memory.
Concluding remarks about the Funahashi et al. 2019 study
The combined results of the Funahashi et al. 2019 study were summarized by Dr. Funahashi in Fig. 9. The important new findings of the study were that MAPK phosphorylates the transcription factor Npas4, and that this phosphorylation is important for learning and memory.
Figure 9. This figure was used as a slide in Dr. Funahashi’s presentation and is his summary of the Funahashi et al. 2019 study.
The Ariza, Funahashi et al. 2021 study
For this blog, it would be too much to also provide a detailed description of Dr. Funahashi’s summary of the Ariza, Funahashi et al. 2021 study. That study found, by largely using identical or similar approaches as in the Funahashi et al. 2019 study, that MAPK is also important for the phosphorylation of the transcriptional coactivator Mkl2. This phosphorylation induced the nuclear localization and activation of Mkl2. Phosphorylated Mkl2 was found to function in complexes with the transcriptional coactivator CBP and transcription factor SRF, and to enhance the transcription from Npas4 and c-Fos promoters. It is reasonable to assume that this cascade occurs in D1R-MSNs of the NAc, but direct evidence at the cell type level was not provided. The effect of Mkl2 phosphorylation on learning and memory was not investigated.
Questions by the participants
The questions by the participants, plus the answers by Dr. Funahashi, that I remember are (edited for clarity):
Question 1.
Fig. 1F of the Funahashi et al. 2019 paper shows the ratio of D1R-MSNs versus D2R-MSNs among the NAc MSNs that could be stained with anti-Npas4, 30 min after cocaine injection. What was the percentage of NAc D1R-MSNs that were positive for staining with anti-Npas4?
Answer: The percentage of NAc D1R-MSNs stained with anti-Npas4 was very small. It suggests that only a small population of D1R-MSN is activated by cocaine-induced CPP.
Question 2.
In the Funahashi et al. 2019 study, in wildtype mice, the conditioned place preference (CPP) training resulted in times spent in the cocaine-conditioned versus control chamber of approximately 8:45 min versus 6:15 min (approximately 150 seconds difference within a 15 min test period), which does not indicate a very strong preference. Would the effect of Nsap4 (conditional) knockout on CPP be bigger or smaller if the preference for the cocaine-conditioned chamber would be higher?
Answer: This has not been tested.
Question 3.
Did you observe, or have there been described, any other phenotypes of the Nsap4 knockout mice in either appearance or behavior?
Answer: I haven’t observed any significant differences in appearance. As shown at the beginning, it has been reported that Npas4KO mice show impaired social behavior and cognitive function. It has also been reported that Npas4KO mice are more likely to exhibit epilepsy-like seizures.
Question 4.
The Funahashi et al. 2019 study showed that a BDNF (brain-derived neurotrophic factor) promoter in transfected COS7 cells could be activated by Npas4 expression. You, therefore, speculate that Npas4 in D1R-MSN stimulates neural plasticity by upregulation of BDNF. However, some scientists believe that NAC MSNs do not express BDNF. Can you elaborate on that?
Answer: We and others have found upregulation of BDNF transcripts in the NAc upon conditioning of mice with cocaine. However, this expression remains to be addressed at the cell type level.
Question 5.
Which events do you believe happen downstream of BDNF?
Answer: BDNF can promote dendritic spine formation (neural plasticity) and the strengthening of neural connections (long term potentiation), and we speculate that an Npas4-BDNF-involving pathway in D1R-MSNs contributes to dopamine-mediated reward learning.
Question 6.
Why, among the many CBP-binding proteins, was Npas4 selected for further investigation?
Answer: The various CBP-binding proteins that we identified were analyzed by gene ontology (GO) and KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway analysis. We were especially interested in finding transcription factors involved in neural plasticity and learning that had not been well studied yet. Npas4 fulfilled these criteria, and an additional advantage was us having access to a knockout mouse.
Question 7.
Your approach for finding new transcription factors, that were activated in dopamine-stimulated D1R-MSN, was restricted to those with an affinity for CBP. Common characteristics of your identified transcription factors Npas4 and Mkl2 are that upon stimulation they are phosphorylated and translocated to the nucleus. Would it be possible, as to not depend on an affinity for CBP, to directly compare nuclear proteins of unstimulated and stimulated cells by mass spectrometry and so to identify translocated and phosphorylated transcription factors?
Answer: Yes, that would be possible. Since the concentration of transcription factors is very low in the cell, I think that the efficiency of finding them will be higher if they are first concentrated using a protein such as CBP.
Question 8.
How would you like to proceed with this research in the future?
Answer: I will study the possible targeting of the activation levels of these transcription factors from a medical (drugs) perspective.
