Evaluation of calcium-sensitive adenylyl cyclase AC1 and AC8 mRNA expression in the anterior cingulate cortex of mice with spared nerve injury neuropathy

Highlights • AC1 and AC8 are widely expressed in many regions of the mouse brain including the hippocampus, ACC, medial prefrontal cortex and midbrain regions, but AC1 is more highly expressed.• Findings suggest a potential role for AC8 in anxiety-like behaviors caused by spared nerve injury in mice.• SNI causes an increase in AC8 mRNA expression in NMDAR-2B (Nr2b) positive neurons in the contralateral ACC but does not affect AC1 mRNA expression.


Introduction
Adenylyl cyclases (ACs) catalyze the formation of cyclic AMP (cAMP) which is a key signaling molecule regulating the activity of protein kinase A (PKA) (Kandel, 2012). In addition to other actions such as phosphorylation of transcription factors, PKA plays an important role in regulating neuronal excitability by phosphorylating voltage gated ion channels, and it also regulates synaptic plasticity by coordinating trafficking and channel kinetics of ionotropic glutamate receptors including AMPA and NMDA receptors (Kandel, 2012;Wang and Zhang, 2012). There are 10 ACs (Adcy gene family) in most mammalian genomes (Kandel, 2012). Nine of these are membrane-bound ACs with 12 transmembrane segments ACs (1-9) and one is a soluble AC (sAC), each of these enzymes exhibits different regulatory properties and expression patterns (Sadana and Dessauer, 2009). Canonical activation of these ACs involves stimulation of a G protein-coupled receptor that signals via a Gα s subunit to bind and activate an AC enzyme. AC1 and AC8, which are predominantly expressed in the CNS, are known to be stimulated by Ca 2+ , in a calmodulin-dependent manner, and are commonly referred to as Ca 2+ -sensitive ACs (Xia and Storm, 1997;Wang and Zhang, 2012). In line with this neuron-specific expression, these ACs play specialized roles in synaptic plasticity in the spinal cord and brain of rodents (Wang et al., 2011;Griggs et al., 2019) where they augment NMDAR currents upstream of PKA phosphorylation of these channels (Liauw et al., 2005;Wang and Zhang, 2012). Genetic knockout of AC1 in mice demonstrates an important role for the Adcy1 gene in chronic hypernociception that occurs after nerve injury or other injuries (Vadakkan et al., 2006;Xu et al., 2008;Wang et al., 2011;Corder et al., 2013;Griggs et al., 2019;Zhou et al., 2021). Genetic knockout of Adcy8 suggests a specialized role for this enzyme in anxiety-like behavior in mice (Bernabucci and Zhuo, 2016) but does not support a role in mechanical hypersensitivity after inflammation, although formalin nocifensive behavioral responses are attenuated in mice lacking the Adcy8 gene (Wei et al., 2002). AC1 promotes mechanical hypersensitivity via an action in the spinal cord and in the anterior cingulate cortex (ACC) of mice (Liu et al., 2020) and plays a critical role in inflammatory, muscle, visceral and neuropathic injury models, and in promotion of latent nociceptive sensitization (Vadakkan et al., 2006;Xu et al., 2008;Wang et al., 2011;Wang and Zhang, 2012;Corder et al., 2013;Qiu et al., 2014;Brust et al., 2017;Griggs et al., 2019;Liu et al., 2020;Zhou et al., 2021). Small molecule inhibitors of AC1 have been developed, such as NB-001 and ST034307, and these molecules reverse mechanical hypersensitivity behaviors in mice and rats, consistent with the notion that AC1 promotes pain through synaptic plasticity mechanisms in the CNS (Wang et al., 2011;Brust et al., 2017;Cheng et al., 2019;Griggs et al., 2019;Zhou et al., 2021). Direct injection of NB-001 into the ACC inhibits mechanical hypersensitivity in multiple mouse pain models in both sexes (Wang et al., 2011;Liu et al., 2020;Zhou et al., 2021), suggesting that AC1 activity in the ACC is a key site for promotion of persistent pain. The ACC has long been recognized as an important brain region for the affective component of pain (Talbot et al., 1991;Rainville et al., 1997;Hofbauer et al., 2001). This is further supported by the fact that anterior cingulotomy has been recognized as a surgical option for the management of intractable cancer and non-cancer pain in humans (Hassenbusch et al., 1990;Wilkinson et al., 1999). Collectively, these lines of evidence makes AC1 an attractive potential target for pain therapeutics.
While Ca 2+ -sensitive ACs, encoded by Adcy1 and Adcy8, are recognized as important mediators of chronic changes in nociception in animal models (Wang et al., 2011;Wang and Zhang, 2012), their mRNA expression distribution in the frontal cortex has not been examined in the context of a neuropathic injury model in male and female mice. In these studies, we used RNAscope in situ hybridization to define changes in Adcy1 and Adcy8 expression in cellular populations in frontal cortical areas in the spared nerve injury (SNI) neuropathic pain model in mice. We find that Adcy1 is widely expressed in important brain regions for pain, where it colocalizes with neurons that express NMDA receptors, but only Adcy8 shows increased expression in the cingulate cortex in the SNI model. Our work raises the testable hypothesis that AC8 may contribute to anxiety in neuropathic pain.

Animals
All animal procedures were approved by the University of Texas at Dallas Institutional Animal Care and Use Committee. Female and male C57BL/6 mice were used for these experiments. Mice were 4 weeks of age at the start of experiments. For assessment of expression of markers across the mouse brain (experiment 1), a single male mouse was used. For the sham versus SNI mouse expression comparison (experiment 2), 12 mice were used (6 sham and 6 SNI) with 3 males and 3 females per group.

Surgery
Spared nerve injury was performed by the ligation and cutting of the tibial and peroneal branches of the left sciatic nerve trifurcation, leaving the sural branch intact. Sham surgeries were performed the same way but without ligating/cutting the nerve. A single injection of the antibiotic gentamicin (3 mg/kg s.q.) was given immediately following surgery. Mice were allowed to recover for 2 weeks following surgery before being assessed for neuropathy.

Mechanical withdrawal threshold
Tactile sensitivity was measured by probing the left outer surface of the left hind paw with a series of calibrated von Frey filaments. Withdrawal thresholds were calculated using the up-down method (Chaplan et al., 1994). Mechanical withdrawal thresholds were assessed in all animals before and after (2 weeks) SNI to confirm neuropathic pain.

Tissue preparation
Mice were decapitated under isoflurane and the brain was removed and frozen in powdered dry ice. The frozen brain was slowly embedded in OCT by adding thin layers of OCT around the sample in order to avoid tissue thawing. The OCT blocks were then stored in a − 80 • C freezer. On the day of the experiment, the tissues were removed from the − 80 • C, placed on dry ice, and transferred to the − 20 • C cryostat chamber for ~ 30 min for temperature acclimation. The brain was then sectioned coronally at 20 µm onto charged slides. For assessment of markers across the mouse brain, single sections of main regions throughout the brain were kept including the medial prefrontal cortex-prelimbic cortex (mPFC -prelimbic), hippocampus (CA1, CA2, CA3), dentate gyrus (DG), lateral amygdala (LA), basolateral amygdala (BLA), mediodorsal thalamus, medial habenula, septal nuclei, caudate putamen, lateral septum, piriform cortex, periaqueductal gray (PAG), retrosplenial cortex, and anterior cingulate cortex (Cg1, Cg2). These regions were located by following a mouse brain atlas (Paxinos and Franklin, 2004) and looking for anatomical landmarks on the specimen.
For the sham versus SNI mouse experiments, at 3 weeks post-SNI, the animals were euthanized and the brains were prepared as described above. 20 µm sections targeting Cg1, Cg2 anterior cingulate cortex (Bregma + 1.1 mm to − 0.22 mm) were kept. While sectioning, a hole was made using a needle through the lateral portion of the hemisphere contralateral to injury. The slides were stored in a − 80 • C freezer.

Rnascope and Nissl staining
On the day of RNAscope experiments, the slide box was removed from the − 80 • C freezer, placed on dry ice, and transferred to the lab. The slides were immediately immersed in cold 10% formalin (pH 7.4) and processed for RNAscope using the fresh frozen protocol with a 1-2minute Protease IV digestion as described by Advanced Cell Diagnostics (ACD; acdbio.com). The probes used are shown in Table 1. The combinations used are shown in individual images. The DAPI step was not performed on all sections. Instead, following completion of RNAscope, the slides were submerged in a coplin jar containing cold blocking solution (10% Normal Goat Serum, 0.3% Triton-X 100 in 0.1 M Phosphate Buffer (PB)) for 1 h at room temperature. The coplin jar was covered in tin foil to protect the slides from light-exposure. Following blocking, the slides were rinsed with 0.1 M PB. Each slide was then placed in lighttight humidity control tray and Blue Fluorescent Nissl stain (Molecular Probes Neurotrace 435/455; Invitrogen Cat N21479) diluted in blocking solution (1:250) was pipetted onto the slide within the hydrophobic boundaries. This was performed one slide at a time to prevent tissue drying. The slides were incubated for 30 min, and then washed in 0.1 M PB. The slides were then air dried, cover slipped with Prolong Gold and allowed to cure over-night.

Imaging and analysis
Images were acquired on an Olympus FV3000 confocal microscope at 20X magnification. For experiment 1, a single image was acquired of all major brain regions listed above. For experiment 2, one 20X image was acquired of the contralateral Cg1 and Cg2 from each section, and 3-5 sections were imaged per animal (n = 6 per group; 3 males and 3 females). The acquisition parameters were set based on guidelines for the FV3000 provided by Olympus (gain = 1, HV ≤ 600, offset = 4) and only laser power was adjusted. Given that probe intensity varies depending on the number of bound amplifiers (thereby, fluorescence intensity is not representative of abundance in this technology), each image was acquired using optimal settings to best visualize the mRNA puncta as instructed by ACD. Images were brightened/contrasted and analyzed in Olympus CellSens (v1.18).
To estimate mRNA abundance in experiment 2, ~25 Nr2b, Adcy1, and Adcy8 co-positive neurons were randomly selected across all laminae in each image and their somas (Nissl signal) were traced using the polygon ROI tool. The area of Adcy1 and Adcy8 signal within the ROI was measured using the Count and Measure tool which highlights the mRNA signal using a thresholded detection. A manual threshold was applied to each image so that all mRNA signal was highlighted within the ROI. Any signal detected that was smaller than 1 µm was automatically not highlighted by the program. Since mRNA puncta is estimated to be ~ 1.5 µm 2 in size, we divided the area of detected Adyc1 and Adcy8 signal by 1.5 to estimate the number of mRNA puncta. The number of puncta was then divided by the area of the ROI (neuronal soma) to calculate Adcy1 puncta/µm 2 or Adcy8 puncta/µm 2 .

Statistics
Statistical tests are described in figure legends. An α value of p < 0.05 was considered significant. Adcy1, Adcy8, Nr2a, and Nr2b probe specificity in the brain. We conduced RNAscope in situ hybridization for Adcy1/ Nr2a/Nr2b and Adcy1/Adcy8 on a single mouse brain. To confirm probe specificity, we compared expression of these markers to the in situ hybridization database on the Allen Brain Atlas. A) For example, in the CA1, CA2 and dentate gyrus (DG) subregions of the hippocampus, Adcy1 (red), Nr2a (green), Nr2b (white) and B) Adcy1 (red), and Adcy8 (green) show differential expression in abundance and region-specificity C) which closely resembles the expression map found on the Allen Brain Atlas. We confirmed there was similar expression in all other assessed regions as well. All other images can be found in supplementary Figs. 1-4. Panel A-B: Images are 20X; scale bar = 50 µm. Panel C: scale bar = 210 µm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig. 2. Adcy1, Adcy8, Nr2a, and Nr2b probe expression in the frontal cortex. We conduced RNAscope in situ hybridization for Adcy1/Nr2a/Nr2b and Adcy1/Adcy8 on a single mouse brain. A) Adcy1 (red), Nr2a (green), Nr2b (white) and B) Adcy1 (red), and Adcy8 (green) in the dorsal anterior cingulate cortex (Cg1), ventral anterior cingulate cortex (Cg2) and medial prefrontal cortex (mPFC)-prelimbic area. Adcy1 had higher expression in the cortex than Adcy8 C) which closely resembles the expression map found on the Allen Brain Atlas (coronal sections include Cg1/Cg2). We assessed coronal images of the ABA expression map when available. We confirmed there was similar expression in all other assessed regions as well. All other images can be found in supplementary Figs

Results
We first assessed expression of Adcy1, Adcy8, Nr2a (also known as N-Methyl D-Aspartate Receptor Subtype 2A; GluN2A; Grin2a), Nr2b (also known as N-Methyl D-Aspartate Receptor Subtype 2B; GluN2B; Grin2b) in a single male mouse. Our probe combinations in this set of experiments were Adcy1/Nr2a/Nr2b and Adcy1/Adcy8. We chose Nr2a and Nr2b as these NMDAR subunits predominate in the forebrain and are important downstream targets of AC activity (Liauw et al., 2005). We compared marker expression to the in situ hybridization database on the Allen Brain Atlas (ABA) and saw similar probe-specificity in all assessed regions. For instance, in the hippocampus, Adcy1 showed low expression in the CA1 and CA3 subregions but was more pronounced in CA2 and the DG (Fig. 1A-B) while Adcy8 had very low expression in CA1 (Fig. 1B). This pattern closely resembled the ABA expression map (Fig. 2C) supporting that our probes were target-specific. We saw widespread expression of Adcy1 throughout the brain and limited Adcy8 expression (Figs. 1-2; Supplemental Figs. 1-4). Adcy1 was highly expressed in the cortical areas of the brain including the dorsal (Cg1) and ventral (Cg2) areas of the mouse ACC, the prelimbic cortex, retrosplenial cortex c and amygdala, and appeared to co-express Nr2a and Nr2b ( Fig. 2; Supplemental Fig. 1). Adcy1 showed little expression in the caudate putamen, medial habenula, medio-dorsal thalamus, and PAG (Supplemental Figs. 1-2). Adcy1 showed higher expression in all brain regions examined than Adcy8 (Supplemental Figs. 3-4). These results largely parallel Allen Brain Atlas (mouse.brain-map.org) and mousebrain.org datasets (Zeisel et al., 2018). Additionally, data from the proteinatlas.org shows that AC1 protein expression is higher in the human brain than AC8 but both are enriched in brain versus other tissues (Thul et al., 2017).
After establishing the specificity of these probes, and the overlap of Adcy1 with Adcy8, Nr2a and Nr2b expression in many brain regions, we sought to determine if Adcy1 or Adcy8 mRNA expression was altered in the SNI model of neuropathic pain. SNI surgery was conducted, and the presence of mechanical hypersensitivity was established by von Frey testing 3 weeks after surgery. SNI mice showed mechanical hypersensitivity compared to sham controls (Supplemental Fig. 5). Brains were taken from mice and then processed for mRNA assessment. We examined the ACC because AC1 has been associated with neuropathic painlike behaviors through a mechanism that involves the ACC in mice (Xu et al., 2008;Wang et al., 2011;Brust et al., 2017;Zhou et al., 2021). To determine changes in mRNA expression we estimated signal abundance in neurons as shown in Fig. 3. We used a Nissl staining technique to label whole neurons, and then calculated the area of the neuron covered with mRNA signal for each probe to estimate abundance ( Fig. 3A-C). We then used this method to assess changes in Adcy1 and Adcy8 mRNA abundance in neurons of the dorsal (Cg1) and ventral (Cg2) areas of the mouse ACC on the side contralateral from SNI or sham surgery. We found an increase in Adcy8 abundance in neurons of the Cg1 area, but no change in Adcy1 mRNA (Fig. 4). In the Cg2 region neither Adcy1 nor Adcy8 were significantly increased by SNI (Fig. 5). We used an equal split of male and female animals in these experiments. No sex differences were noted, and there was little variability between samples, but the sample size was not large enough to directly assess sex differences.

Discussion
Our findings confirm that AC1 and AC8 are expressed throughout the mouse brain including many cortical regions. AC1 is more widely expressed, corroborating previous brain-wide expression mapping studies (Visel et al., 2006;Sanabra and Mengod, 2011), including those done by the Allen Brain Institute (mouse.brain-map.org) and the Human Protein Atlas (proteinatlas.org). There is extensive literature on the ACC's involvement in the affective aspects of pain in people and its involvement in rodent chronic pain-like models (Hassenbusch et al., 1990;Talbot et al., 1991;Rainville et al., 1997;Wilkinson et al., 1999;Hofbauer et al., 2001;Xu et al., 2008;Blom et al., 2014;Pereira et al., 2014;Sharim and Pouratian, 2016) . Additionally, AC1/AC8 have been shown to play a role in modulating injury-associated and anxiety-like behaviors in this brain region (Xu et al., 2008;Blom et al., 2014;Pereira et al., 2014;Gu et al., 2015;Kang et al., 2015;Sharim and Pouratian, 2016;Sellmeijer et al., 2018;Deng et al., 2019). Given these findings, we investigated whether AC1 and AC8 mRNA expression in the ACC changed after nerve injury. In the SNI mouse model of neuropathic pain, 3 weeks after injury, we found that Adcy8 mRNA expression was increased in the dorsal, but not ventral ACC, likely reflecting an increase in Adcy8 expression in neurons that either did not express Adcy8 previously, or only expressed it on a low level. Based on previous experiments demonstrating increased AC1 expression in the zymosan visceral inflammation model in mice (Liu et al., 2020), we expected to observe increased Adcy1 expression in the dorsal ACC, but our findings did not support our original hypothesis. Our findings do not discount that AC1 plays an important role in neuropathic pain-like behaviors in mice as both knockout of Adcy1 gene, and inhibition of AC1 with pharmacological blockers, reduces neuropathic mechanical hypersensitivity in mice (Wei et al., 2002;Wang et al., 2011). This role may be better explained by alterations in protein expression or AC1 activation, as AC1 protein but not mRNA levels are significantly increased in the absence of important translation regulators (Koga et al., 2015;Sethna et al., 2017). Further characterization of proteomic changes in AC1 and AC8 induced by neuropathic pain states will offer a deeper understanding of this relationship.
A hypothesis emerging from our findings that could be tested in future experiments is that increased expression of Adcy8 in the ACC may be linked to anxiety produced by neuropathic pain. The presence of neuropathic pain in humans often causes comorbid anxiety (Nicholson and Verma, 2004). While anxiety caused by peripheral nerve injury has not been consistently observed in mouse models (LaGraize et al., 2004;Urban et al., 2011;Sheahan et al., 2017), some studies suggest that mice and rats can develop anxiety after nerve injury (Seminowicz et al., 2009;Sellmeijer et al., 2018;Li et al., 2021), or develop changes in gene expression consistent with anxiety phenotypes (Descalzi et al., 2017). Previous studies using Adcy8 knockout mice suggest that AC8 may be important for persistent anxiety caused by environmental cues and stress (Schaefer et al., 2000;Bernabucci and Zhuo, 2016). A possible interaction of the Adcy8 gene with SNI-induced anxiety could be tested in future studies.
There are several limitations to our study. We were not powered to look at sex differences, but we did use mice of both sexes. Given the small experimental variability, we think our findings of increased Adcy8 expression in the Cg1 region likely occurs in both sexes. We only examined a single time point after nerve injury, but we chose a time point that is consistent with most published studies in the field. Anxiety may develop at later time points after injury (Seminowicz et al., 2009), so we may expect greater changes in Adcy8 as time goes on. We did not examine potential changes in other brain regions because the existing literature has focused mostly on Ca 2+ -sensitive ACs in the ACC in pain models. Finally, we did not clarify the cell types in the ACC that show changes in Adcy8 expression, but these neurons clearly express the NMDAR subunit, NR2B.
We conclude that the extensive literature on the role for AC1 in enhanced nociception after injury in mice cannot be explained by increased expression of Adcy1 mRNA in neurons in the ACC. This does not discount the possibility that protein levels may increase or that enzyme activity might increase due to enhanced Ca 2+ signaling in these neurons after peripheral nerve injury. Nevertheless, our work highlights a potentially unappreciated role for AC8 that can be further explored using transgenic mice or through pharmacological manipulation. The latter may require further development of tool compounds as most pharmacological development on Ca 2+ -sensitive ACs has focused on AC1.

Funding
Contract funding from Grünenthal GmbH. NIH grant NS065926

Author contributions
Study design, SS, SH and TJP; Experimental work, SS and HE; Data analysis, SS and HE; Wrote the paper, SS and TJP; Edited the paper, all authors.

Editorial disclosure
Given his role as Editor-in-Chief, Theodore J. Price had no involvement in the peer-review of this article and has no access to information regarding its peer-review. Full responsibility for the editorial process for this article was delegated to Associate Editor, Anna M.W. Taylor.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper other than that SH is an employee of Grunenthal GmbH. The work was funded by Grunenthal GmbH and NIH grant NS065926.