Cocaine regulation of Nr4a1 chromatin bivalency and mRNA in male and female mice

Cocaine epigenetically regulates gene expression via changes in histone post-translational modifications (HPTMs). We previously found that the immediate early gene Nr4a1 is epigenetically activated by cocaine in mouse brain reward regions. However, few studies have examined multiple HPTMs at a single gene. Bivalent gene promoters are simultaneously enriched in both activating (H3K4me3 (K4)) and repressive (H3K27me3 (K27)) HPTMs. As such, bivalent genes are lowly expressed but poised for activity-dependent gene regulation. In this study, we identified K4&K27 bivalency at Nr4a1 following investigator-administered cocaine in male and female mice. We applied sequential chromatin immunoprecipitation and qPCR to define Nr4a1 bivalency and expression in striatum (STR), prefrontal cortex (PFC), and hippocampus (HPC). We used Pearson’s correlation to quantify relationships within each brain region across treatment conditions for each sex. In female STR, cocaine increased Nr4a1 mRNA while maintaining Nr4a1 K4&K27 bivalency. In male STR, cocaine enriched repressive H3K27me3 and K4&K27 bivalency at Nr4a1 and maintained Nr4a1 mRNA. Furthermore, cocaine epigenetically regulated a putative NR4A1 target, Cartpt, in male PFC. This study defined the epigenetic regulation of Nr4a1 in reward brain regions in male and female mice following cocaine, and, thus, shed light on the biological relevance of sex to cocaine use disorder.

Despite decades of research, cocaine use disorder remains a global health problem for which there is currently no FDA approved treatment. Despite clear epidemiological evidence that cocaine abuse afflicts men and women differently 1 , the underlying neurobiology of how cocaine impacts each sex is not fully understood. Human sex differences in behavior include higher initiation of cocaine use by men than women, a greater percent of men addicted to cocaine than women 2 , and a faster shift from casual drug use to addiction in women than men 3 . Additionally, women enter treatment at an earlier time point than men, have a higher relapse rate 3 , both of which may reflect women being more vulnerable to cue-induced relapse than men 4 . Importantly, rodent models of cocaine abuse recapitulate sexually dimorphic behavior, such that female rodents are more responsive to drug-conditioned stimuli 5,6 and acquire cocaine-self administration at a faster rate than male rodents 7,8 . Recent research has indicated there are sexually dimorphic molecular effects of cocaine that might mediate behavioral phenotypes. For example, it was recently shown that cocaine broadly regulates distinct sets of proteins in males relative to females following cocaine self-administration in rodents 9 . Given the prominent sexually dimorphic cocaine-associated behavioral phenotypes and the distinct neuroadaptations for males and females 10 , sex must be considered a biological variable within drug research.
The histone code has historically proposed that engagement of individual HPTMs at a chosen locus can induce changes in protein recruitment and impact gene expression. However, the code also posits that HPTMs can be Results K4&K27 bivalency at Nr4a1 promoter in brain reward regions was present in both male and female mice. For epigenetic profiling of three reward-associated brain regions following cocaine treatment, we applied sequential ChIP to examine K4&K27 bivalency at Nr4a1 (Fig. 1). ChIP-qPCR and ChIP-Seq protocols that quantify HPTM deposition (the addition of the modification) at monovalent promoters are unable to quantify combinatorial HPTM deposition at single promoter. This weakness stems from the fact that individual HPTM profiling experiments cannot distinguish between bivalent promoter deposition of two HPTMs in a single population of cells or monovalent promoter deposition of two HPTMs from a mixed population of cells 33 . www.nature.com/scientificreports/ The sequential ChIP protocol is required to define true bivalency of H3K4me3 and H3K27me3 deposition at the same locus in a single population of cells (and nucleosomes) 33 . To measure bivalency, we split each single sample and performed H3K4me3 ChIP followed by H3K27me3 ChIP. From the same sample, we also performed H3K27me3 ChIP followed by H3K4me3 ChIP. In this way, sequential ChIP was assayed with either H3K4me3 or H3K27me3 as the first IP, and the other HPTM as the second IP. This approach allowed us to interpret differences in cocaine-regulated K4&K27 bivalency due to differences in rates of deposition or removal between the two HPTMs 47 . In each case, enrichment based on antibody order was quantified by qCHIP. As we found no differences in results of qCHIP by order of sequential ChIP, we concluded that both approaches demonstrated Nr4a1 K4&K27 promoter bivalency.
Cocaine enriched H3K27me3 and bivalency at the Nr4a1 promoter and did not alter Nr4a1 mRNA levels in male striatum. Following cocaine treatment, there was no change in Nr4a1 mRNA ( Fig. 2A) or H3K4me3 deposition at Nr4a1 in the male STR (Fig. 2B). In the same samples, there was increased H3K27me3 at Nr4a1 (Fig. 2C) (t 10 = 2.3, p = 0.0443) and an increase in K4&K27 bivalency at the Nr4a1 promoter (Fig. 2D,E) (D: t 5.973 = 2.663, p = 0.0376; E: t 9 = 2.298, p = 0.0476). In the male HPC following cocaine treatment, there was no change in Nr4a1 mRNA (Fig. 2F), but we observed depletion of H3K4me3 at Nr4a1 (Fig. 2G) (t 12 = 2.528, p = 0.0265). In the same samples, there was no change in H3K27me3 deposition or K4&K27 bivalency ( Fig. 2H,I,J). In the male PFC following cocaine treatment, there was no change in Nr4a1 mRNA (Fig. 2K) 3G). This was accompanied by no change in H3K4me3 or H3K27me3 deposition at Cartpt (Fig. 3H,I). These data show that cocaine induces Cartpt in the male PFC.
In the female HPC, there was no change in Nr4a1 mRNA following cocaine treatment when the Nr4a1 Exon 4/5 primer set was used, but we observed a trending increase in Nr4a1 mRNA when mRNA levels were evaluated with Exon 6/7 primers (trending, t 14 = 1.1841, p = 0.0869) (Fig. 4F). In the same samples, there was no change in Nr4a1 H3K4me3, H3K27me3 or K4&K27 bivalency ( Fig. 4G-J). In the female PFC following cocaine treatment, Nr4a1 mRNA was increased when measured at Nr4a1 E6/7 but not Nr4a1 E4/5 (Nr4a1 Exon 6/7: t 13 = 2.327, p = 0.0368) (Fig. 4K). In the same samples, there was no change in H3K4me3 (Fig. 4L), there was H3K27me3 recruitment (t 6.392 = 3.363, p = 0.0138) (Fig. 4M), and no change in K4&K27 bivalency (Fig. 4N,O). In sum, these data showed that cocaine treatment in female mice increased Nr4a1 mRNA expression in the STR along with maintenance of H3K4me3, H3K27me3, and K4&K27 bivalency at the Nr4a1 promoter. These data also showed that cocaine treatment in female mice increased Nr4a1 mRNA expression in the PFC, despite recruitment of H3K27me3.
Cocaine did not alter mRNA of drug-related markers, Cartpt mRNA, or Cartpt HPTMs in female mouse brain regions. As we saw independent increases in mRNA levels of Nr4a1 following cocaine in female mice in the PFC and the STR, we expected to find transcriptomic or epigenetic changes at Cartpt, a downstream target of Nr4a1 38 . Surprisingly, we found no transcriptomic or epigenetic changes (H3K4me3, H3K27me3) at Cartpt in female mice following cocaine in the STR ( Cocaine impacted correlational relationships within and between the transcriptomic and epigenetic profiles of Nr4a1 and Cartpt in male and female mice. As we hypothesized that we would see a linear relationship between Nr4a1 and Cartpt following cocaine, we utilized Pearson's correlation matrices to explore the relationships within our sample population between the transcriptomic and epigenetic profiles of Nr4a1 and Cartpt. Every significant correlation we observed can be found in Supplementary Tables 3-5 which also provides the corresponding figure(s) that show the data within each correlation in Fig. 6  www.nature.com/scientificreports/ either Exon 4/5 or Exon 6/7 across treatments (saline, cocaine) and sex (male, female) (Fig. 6A,B). In the STR of saline-treated females, we observed a positive correlation between (3) Nr4a1 mRNA and Cartpt H3K27me3 (Fig. 6B). In the STR of males following cocaine, we observed a positive correlation between (4) Cartpt mRNA and Cartpt H3K27me3, (5) Cartpt mRNA and Nr4a1 K4/K27 bivalency, and (6) Cartpt H3K27me3 and Nr4a1 K4/K27 bivalency (Fig. 6A). In the STR of females following cocaine, we observed a positive correlation between (7) Nr4a1 H3K4me3 and H3K4me3 at Cartpt and (8) Nr4a1 H3K27me3 and Cartpt H3K4me3 (Fig. 6B).

Discussion
There is emerging literature on the transcriptomic and epigenetic profiles of Nr4a1 and its regulation by cocaine.
Despite this progress, no study to date has explored basal and cocaine-activated regulation of Nr4a1 in both males and females in any organism. Additionally, few studies have assessed bivalency in the brain [35][36][37] , and none, to           www.nature.com/scientificreports/ our knowledge, have investigated cocaine regulation of bivalency. Here, we demonstrate that, following cocaine, Nr4a1 mRNA increases in female mouse STR and PFC, and does not change in male mouse STR, HPC, or PFC. We also find an increase in Nr4a1 K4&K27 promoter bivalency in the male STR following cocaine treatment. Finally, we show that the mRNA level of a putative Nr4a1 target, Cartpt, is activated in the male PFC.
Our results are the first to establish that there is a relationship between bivalency and cocaine in the brain, specifically we show induction of bivalency in the male STR. As this mechanism has not yet been linked to drugs of abuse, it provides a new mechanism of exploration for drug abuse research. Bivalent chromatin regulates DNA accessibility rather elegantly by transitioning euchromatin or heterochromatin to a poised, bivalent state that can quickly react to cellular cues. This flexibility is beneficial for cell differentiation 32 and stress responses 35 , but may underlie detrimental instability of gene expression in the adult brain. Consequently, bivalency is implicated in transcriptional dysregulation observed in Huntington's Disease 48 as well as multiple types of cancer [49][50][51][52] . We hypothesize that bivalency may similarly underlie transcriptional dysregulation in drug addiction, in both males and females. Our data highlight that the interplay between bivalency and cocaine warrants further investigation.
The current study profiles both ventral and dorsal STR, as well as HPC and PFC, other brain regions associated with reward neuropathology. Both the ventral STR (NAc) and dorsal STR are involved in cocaine reinforcement, and the interconnectivity between these regions mediates cocaine-seeking behavior 53 . Additionally, pharmacological studies have shown that dopamine transporters in both the NAc and dorsal STR have the same affinity for cocaine binding 54 , and that the connectivity between dorsal STR and NAc is necessary for the regulation of dopamine to mediate cocaine's reinforcing properties 55 . Our lab and others find that cocaine treatment increases Nr4a1 mRNA in the NAc in mixed sex populations 38 , as well as in males and females when assessed separately 39 . However, in order to analyze K4&K27 bivalency at the Nr4a1 promoter of a single mouse, it was necessary to isolate chromatin from combined STR and perform sequential ChIP. Both H3K4me3 and H3K27me3 are basally present at Nr4a1, with high enrichment for H3K4me3 in bulk tissue of the nucleus accumbens 31 as well as in cell-type specific tissue of the combined striatum 56 , showing that combined STR and NAc have the same enrichment pattern. Additionally, both H3K4me3 and H3K27me3 are present at baseline at Cartpt, with high co-enrichment for both marks in bulk tissue of the nucleus accumbens 31 and in cell-type specific tissue of the striatum 56 . Using sequential ChIP, we were able to identify the same patterns of enrichment at both modifications at baseline for Nr4a1 and Cartpt.
The strengths of the sequential ChIP approach in deciphering the histone code are limited by the technical requirement of bulk STR starting material, which obscures cell-type specific Nr4a1 regulation by cocaine and sex. The current literature contains complex and conflicting data on the expression pattern of Nr4a1 in direct pathway, Drd1 + , and indirect pathway, Drd2 + and/or A2a + , medium spiny neurons (MSNs) 39,40,57,58 . Studies suggest that Nr4a1 expression in Drd1 + MSNs, but not A2a + MSNs, functions in cocaine craving and relapse-like behavior. Drd1-specific signaling, such as the phosphoErk (pErk) cascade, increases in the NAc across incubation of cocaine craving 59 , and may regulate Nr4a1 in the Drd1 + MSN subtype 40,57,59,60 .
Importantly, the relationships with bivalency (H3K4me3/H3K27me3 vs H3K27me3/H3K4me3) do not always align at the level of a correlation. This is likely inherent to the biological mechanism of bivalency. Induction of H3K4me3/H3K27me3 bivalency is indicative of a transition from an activating to poised state (bivalency results in lower activation), while induction of H3K27me3/H3K4me3 bivalency is indicative of a repressive to poised state (bivalency results in higher activation). Additionally, we found that Nr4a1 mRNA levels did not always correspond with the canonical roles of H3K4me3, H3K27me3, and K4&K27 bivalency, in activating, repressing, and poising gene expression, respectively. For example, in the female PFC following cocaine treatment we observed concomitant increases in Nr4a1 mRNA and H3K27me3 at Nr4a1. We thus applied correlational analyses to examine these relationships more comprehensively. In the male STR, we found a positive correlation between Nr4a1 K4/K27 bivalency and Cartpt mRNA levels. Given that Nr4a1 is a putative activator of Cartpt expression 38 , this finding went against our expectation that Nr4a1 gene poising would be negatively correlated with Cartpt mRNA levels. Additionally, given the lack of transcriptomic or the epigenomic regulation of Nr4a1 in the male PFC and the vast increase in Cartpt mRNA levels, we hypothesize that Cartpt may additionally be regulated by other distinct transcription factors and/or HPTMs following cocaine. This is supported by our correlational data that shows minimal overlap in the transcriptomic and epigenetic profiles of Nr4a1 and Cartpt in either drug conditions in males or females or in any of the three regions examined. Another reason that some of the epigenetic and mRNA correlations may not have aligned could be due to the fact that our work focused solely on investigating Nr4a1 bivalency at the promoter. Epigenetic investigation of other regulatory regions, such as enhancers, which influence transcription via signaling to promoters 61 , could shed further light onto epigenetic control of gene regulation. For example, H3K4me3 is predominantly located at promoters, although it has been also been identified as a modifier of enhancers 62 . Future studies will expand investigation of bivalency beyond promoters to other genomic regulatory regions, provided such regions have been identified and functionally validated.
Based on our findings and the literature, we also hypothesize that additional HPTMs, such as acetylation, regulate expression of Nr4a1 and Cartpt. Intracranial HPC injection of the histone deacetylase (HDAC) inhibitor, Trichostatin A, increases Nr4a1 mRNA and protein levels 63 , as does HDAC3 inhibition in the dorsal HPC 64 . In cell lines, Nr4a1 expression is regulated by the recruitment of the histone acetyltransferase p300 or HDAC1 65 . It is interesting to consider that changes in STR histone acetylation, and methylation to a certain extent, are sensitive to the duration of cocaine treatment (acute versus chronic) and time since the last injection 21 , which may reflect the rapid turnover of acetylation, especially compared to methylation 47 . Although acetylation turnover is more rapid than methylation, we hypothesize that repeating the endpoints outlined in this study at a different time point, such as 30 min after the final injection versus 24 h, could prompt different levels of H3K4me3, H3K27me3, and K4&K27 bivalency. Specifically, we expect an initial decrease in H3K27me3 and K4&K27 bivalency (< 30 min following cocaine) followed by an increase in H3K27me3 and K4&K27 bivalency at later timepoints (> 24 h following cocaine). www.nature.com/scientificreports/ Nonetheless, our findings encourage future studies to investigate combinatorial HPTMs beyond K4&K27 bivalency. For example, H3K27me3 and H3K27ac bivalency are well described in the literature and represent a tractable starting point for further exploration 66,67 . We hypothesize that cocaine-induced depletion of H3K27ac acts synergistically with increased K4&K27 bivalency to inhibit the activation of Nr4a1 expression in males. We speculate that the persistence of H3K27me3 is sex-dependent, and thus H3K27me3 is likely bivalent with other HPTMs as well. This hypothesis is based in part on the sexually dimorphic expression of H3K27me3demethylase, Kdm6a, in rodent brain [68][69][70] . As Kdm6a is X-linked 71 , it may mediate sex-specific changes in HPTM expression upon cocaine exposure to drive epigenetic changes at Nr4a1 and other cocaine-responsive loci. It is possible that sexually dimorphic molecular changes underlie sexually dimorphic behaviors. For example, male rodents are less responsive to drug-conditioned stimuli 5,6 and acquire cocaine-self administration at a slower rate than female rodents 7,8 . Future studies will explore additional combinatorial HPTMs with H3K27me3 to probe these hypotheses.
Additionally, new methods are being developed to allow for future studies to conduct sequential ChIP profiling of small amounts of input material which will allow for profiling of more specific brain regions and cell-types. For example, the recently developed multi-CUT&Tag 72 allows for mapping HPTM co-localization in the same cell, and single-cell multi-CUT&Tag 72 can probe for bivalency in specific cell types. While single-cell CUT&Tag has been carried out in mouse brain to profile individual HPTMs 73 , these multi-CUT&Tag methods that uncover bivalency have yet to be applied in brain tissue as far as we are aware. Finally, in the current study we test and support the hypothesis that K4&K27 bivalency is increased at Nr4a1 in male mice following cocaine treatment, concomitant with increased H3K27me3 deposition. To test the sufficiency of loss of H3K27me3 in increased Nr4a1 mRNA, we can apply in vivo epigenetic editing to exogenously enrich Nr4a1 H3K27me3 in the presence of cocaine. This approach may also shed light onto the role of H3K27ac, given that dCas9-FOG1 interacts with the NuRD complex to cause HDAC1/2-mediated removal of H3K27ac 74 .
In closing, despite epidemiological evidence that cocaine abuse afflicts both men and women 1 , the underlying neurobiology of how cocaine impacts each sex is not fully understood. This is problematic as a growing body of knowledge demonstrates that cocaine causes distinct neuroadaptations for males and females 10 . The inclusion of both sexes within research investigating reward neurobiology is thus vital and necessary to the goal of treating addiction disorders. Future studies should include both males and females to understand how cocaine causes sex-specific neuroadaptations and for uncovering sexually dimorphic mechanisms that could lead to innovations for therapeutic intervention.

Methods and materials
Animals. Male and female mice on the C57BL/6 J background were used in this study. Mice were housed under a 12-h light-dark cycle at 23 °C with access to food and water ad libitum. Note that in compliance with ethical standards to minimize the use of mice, the mice used in this study were cre-negative offspring of R26-CAG-LSL-Sun1-sfGFP;A2a-cre and LSL-Sun1-sfGFP;Drd1-cre, generated for a separate study. All animal procedures were conducted in accordance with the National Institutes of Health Guidelines as well as the Association for Assessment and Accreditation of Laboratory Animal Care. Ethical and experimental considerations were approved by the Institutional Animal Care and Use Committee of The University of Pennsylvania. All experiments complied with ARRIVE guidelines.
Investigator-administered cocaine. Male and female mice were handled daily for three days prior to cocaine administration. Following handling, mice were given a daily cocaine hydrochloride intraperitoneal injection (20 mg/kg dissolved in 0.9% saline) or 0.9% saline injection for 10 days 38 . Saline injected mice were used as the control group in all experiments. Male and female mice were injected with cocaine at the same time. Mice were sacrificed one day after the final injection. STR, HPC, and PFC tissue was immediately collected from each animal (Fig. 1A,B) and stored at − 80 °C until processing. S3EQ nuclear and cytosolic fraction separation. S3EQ 45 was conducted as previously published with modified buffer volume. Tissue samples were homogenized in 350 μL cell lysis buffer (10 mM Tris-HCl (pH 8.0), 10 mM NaCl, 3 mM MgCl2 and 0.5% NP-40 in H 2 O) and spun for five minutes (1500 × g, 4 °C). The nucleicontaining pellet and RNA-containing cytosolic supernatant were separated and subjected to sequential ChIP or RNA extraction, respectively (Fig. 1C).

Single and sequential chromatin immunoprecipitation (ChIP) and ChIP-PCR (qCHIP). Prior to
initiating the protocol, the following solutions were prepared as previously described 45 www.nature.com/scientificreports/ stop cross-linking (500 rpm, 22 °C). Samples were then centrifuged for five minutes (5500 × g, 4 °C) and the supernatant was discarded. The pellet was resuspended in 200 μL nuclear lysis buffer, transferred to TPX tubes, (Diagenode, C30010010) and incubated on ice for ten minutes. Samples were then sonicated in a Bioruptor ® (Diagenode) for two runs (high setting, 30 s on, 30 s off, 15 cycles). Dilution buffer was added to reach 1000 μL. 10% input (relative to each IP) was collected from each sample for normalization later used in qChIP analysis. The remaining diluted chromatin was combined with antibody-bound beads (1.2 μL bead-antibody slurry/1 μL chromatin), dilution buffer, and placed on a rotator (O/N, 4 °C). Samples were then washed with 1 mL of icecold Wash Buffer 1, Wash Buffer 2, Wash Buffer 3, and TE Buffer. Samples were rotated for five minutes during each wash (RT). Following washes, samples were rocked with 200 μL of elution buffer for 20 min (500 rpm, RT), centrifuged for three minutes (14,000 × g, RT), and placed on a magnet. The supernatant from each sample was divided in half and transferred to fresh tubes, each containing 100 μL of elution buffer. One tube was collected as the single ChIP sample. The remaining tube was subjected to a second immunoprecipitation (sequential ChIP samples), in which each sample was subjected to antibody-bound beads that were bound to the reciprocal antibody (i.e., H3K4me3 ChIP samples were bound to H3K27me3 antibody-bound beads) and placed on a rotator (O/N, 4 °C). The sequential ChIP samples were processed according to the single ChIP protocol above. Single and sequential ChIP samples were then incubated for 4 h (65 °C) following the addition of 8 μL 5 M NaCl. Proteinase K (0.002 mg) was added, and incubation continued for two hours (65 °C). Samples were heat inactivated (15 min, 78 °C) and DNA was purified using the QiaAmp DNA Micro kit (Qiagen, 56304). qChIP was conducted as previously published 38 . Briefly, qChIP was run using Power SYBR™ Green PCR Master Mix (Life Technologies). The primer sequences that span − 100 to + 100bps of target promoter regions used to assess the presence of HPTMs 38 are noted in Supplemental Table 1. RNA Extraction and qPCR. The cytosolic fraction following S3EQ was mixed with 600 μL RLT buffer (Qiagen) and RNA was extracted using the RNAeasy Micro Kit (Qiagen, 74004). RNA concentration was obtained using a Qubit 4 Fluorometer and RNA HS Assay Kit (Invitrogen). cDNA was synthesized using the Bio-Rad iScript™ cDNA Synthesis Kit (Bio-Rad, 1708890). qPCR was conducted using Power SYBR™ Green PCR Master Mix (Life Technologies). The primer sequences used to assess mRNA levels are noted in Supplemental Table 2.
Pearson correlation analysis of qCHIP and qPCR. Pearson correlation analysis was used to assess the relationship between K4&K27 bivalent Nr4a1 promoter, Nr4a1 or Cartpt mRNA levels, and Nr4a1 or Cartpt monovalent H3K4me3 or H3K27me3. Scripts were written and executed in R v4.1.1. First, all values corresponding to relevant endpoints were compiled for each sex (male, female) and brain region (STR, HPC, PFC) combination. For a given combination, the dataset was further separated based on treatment (saline, cocaine). Analyses were conducted with input values (fold change) from qChIP and qPCR data. An endpoint (qChIP or qPCR) was excluded if it lacked at least three observations in both the saline and cocaine datasets. An all-againstall Pearson's correlation matrix was generated for the saline and cocaine datasets separately using the base R cor (method = "pearson") function. A correlation heatmap was generated from the Pearson correlation matrix using the corrplot() function from the corrplot package.
Statistical analyses. Statistical tests used reflect the experimental design 75 . Specifically, while male and female mice were treated with cocaine in a single cohort, the tissue for each sex and each brain region was processed and analyzed separately due to the technical limitations associated with processing large numbers of sample. We therefore did not apply ANOVA to analyze sex or region differences. G*Power: Statistical Power Analyses Software 76 (v3.1) was used to determine sample size. Final sample size was reduced during experimental execution (from n = 10 to n = 8 per group) due to previous empirical experimental endpoints and technical limitations. The Grubbs test was used to identify all outliers after an entire data set was complete and before any other statistical tests were applied (alpha = 0.05). Student's t-tests were used for qChIP and qPCR analyses as these experiments directly compared one factor (drug exposure) from two groups. Sex was not considered a variable within any analyses due to experimental execution. F-tests of variance were conducted in every analysis and Welch's correction was utilized when outliers were removed and when variances differed between groups. These data are shown as mean ± SEM. Graphpad Prism V9.3 was used for qChIP and qPCR analyses. qChIP data was analyzed by comparing CT values of each experimental group versus the control group following normalization of each sample's input using the ∆∆Ct method 77,78 . Bivalency at Cartpt was not assessed, as not all sequential-immunoprecipitated samples reached minimum signal intensity for qChIP (CT values were ≥ 38). qPCR data was analyzed by comparing CT values of each experimental group versus the control group following normalization of a housekeeping gene (Gapdh) using the ∆∆Ct method 77,78 . All experiments were carried out one to two times, and data replication was observed in all instances of repeated experiments. Additionally, every qChIP and qPCR dataset had a technical replicate conducted by an independent investigator to ensure the validity of the data. All data shown in final analysis come from the first experiment of replicates, and the data specifically comes from the technical replicate conducted by the first author. Regarding visual data representation, each data point in all graphs represent one animal (Figs. 2, 3, 4, 5). For Pearson correlation analysis, the p-value matrix of all correlations corresponding to each comparison was extracted using the two-sided cor.mtest(method = "pearson") function from the corrplot package with a 95% confidence interval. This statistical test for Pearson's product-moment coefficient was calculated following a t-distribution ( t = r √ n−2 √ 1−r 2 , where n = number of observations and r = calculated Pearson correlation coefficient) with degrees of freedom equal to length (number of complete observations)-2. Heatmap tiles were shaded red or blue only when satisfying a p value ≤ 0.05, where the null