Long noncoding RNA-dependent control of Myc transcriptional bursting

SUMMARY

Genes are transcribed in short periods of activity, called bursts. Bursts are initiated by enhancer-promoter contacts and dynamically controlled by the levels of enhancer- and promoter-produced RNAs through a negative feedback mechanism. Here, by direct visualization of nascent transcripts, we show that chromatin-associated long noncoding RNAs (lncRNAs) contribute to the regulation of transcriptional bursting. We find that production of Pvt1 raises the baseline of RNA concentration in the locus of the Myc proto-oncogene and acts locally and in a dose-dependent manner to decrease the duration of Myc bursting. Premature termination of Pvt1 led to higher Myc expression and transcriptional activities, resulting in increased cellular proliferation and advanced tumor development in autochthonous models of lung cancer. These findings point to a critical lncRNA-mediated mechanism for Myc regulation and suggest a potentially widespread role for lncRNAs in fine-tuning gene expression through local control of transcriptional bursting.

In brief

Li et al. show that the chromatin-associated long noncoding RNA Pvt1 is a negative regulator of Myc transcription by increasing local RNA concentration and limiting bursting from the Myc promoter. This mechanism serves to restrain Myc levels and restrict its proliferative and oncogenic activities.

Graphical Abstract

INTRODUCTION

Genes from bacteria to mammals are transcribed discontinuously in so-called “bursts” with promoters cycling between ON and OFF periods of active transcription.^1–3 Previous work has established that promoter-enhancer contacts initiate promoter bursts and are maintained for the duration of bursting,^4,5 while an RNA-mediated transcription feedback model has been proposed to explain the oscillating nature of bursts.⁶ In this model, low levels of local RNA transcription from the promoter and its enhancers initially promote the formation of transcriptional condensates to stimulate transcription, while high local RNA concentration above a threshold repels transcriptional condensates, terminating the transcriptional burst and reverting the locus to a transcriptionally inactive state.⁷ It has been proposed that other neighboring transcripts, such as long noncoding RNAs (lncRNAs), may also contribute to the regulation of transcriptional bursting.^7,8

LncRNAs are a diverse class of transcripts longer than 500 nucleotides that are transcribed and processed similarly to mRNAs but lack protein-coding potential.⁹ A subset of lncRNAs appear to associate stably with the chromatin at their sites of transcription,¹⁰ where they act locally to modulate the expression of neighboring protein-coding genes.¹¹ To date, multiple mechanisms of cis-regulation have been proposed, including lncRNA molecules scaffolding the recruitment of transcription factors, co-factors, and chromatin-modifying complexes, as well as the act of lncRNA transcription reconfiguring the local three-dimensional chromatin architecture.¹² However, for many cis-regulatory lncRNAs, the mechanism of local regulation remains unknown.

The myelocytomatosis (Myc) proto-oncogene is a master regulator of cellular proliferation during normal development and in pathological conditions, such as cancer.¹³ To maintain homeostasis, Myc levels are tightly controlled by post-transcriptional and post-translational mechanisms that promote rapid mRNA and protein turnover.^14,15 At the transcriptional level, Myc expression is dependent on numerous upstream and downstream enhancers, which are frequently co-amplified with Myc in cancer.^16,17 Myc transcription is also modulated by a neighboring lncRNA Pvt1 (plasmacytoma variant translocation 1). Pvt1 is initiated 50 kb downstream of Myc and gives rise to a set of abundant, chromatin-associated, alternative transcription initiation and splice isoforms that span more than 10 exons over a 300 kb-region that encompasses Myc downstream enhancers. Pvt1 was initially identified as a site of frequent viral integrations, translocations, and amplifications associated with increased Myc expression in cancer and was proposed to play an oncogenic role.^18–21 Subsequent studies revealed that the Pvt1 locus is also a negative regulator of Myc transcription by limiting the access of the Myc promoter to downstream enhancers in cancer¹⁶ and by giving rise to a p53-induced, Myc-repressive Pvt1 isoform, Pvt1b, in the presence of oncogenic and genotoxic stress.^22,23 From these conflicting studies, it was not clear whether constitutively transcribed isoforms of Pvt1 play an activating, repressive, or neutral role in Myc regulation.

In this study, we used complementary genetic approaches to investigate the contribution of constitutively expressed Pvt1 to Myc regulation. By direct visualization of endogenous nascent Myc transcripts in situ, we found that Pvt1 limits the duration of transcriptional bursting from the Myc promoter in an RNA- and dose-dependent manner. We further investigated whether this added layer of regulation was independent from known mechanisms of Myc control and how it affected Myc oncogenic functions in cellular and organismal models of cancer. Our findings reveal a lncRNA-dependent mechanism that limits Myc transcriptional output and demonstrate its physiological importance for restraining Myc oncogenic activity.

RESULTS

Dynamics of Pvt1 and Myc transcription

To investigate the dynamics of endogenous Pvt1 and Myc transcription in situ, we visualized Pvt1 and Myc nascent transcripts by single-molecule RNA fluorescence in situ hybridization (smRNA-FISH) using Pvt1 (Pvt1i) and Myc (Myci) intron-specific probes in wild-type (WT) primary mouse embryonic fibroblasts (MEFs; Figure 1A). We observed that Pvt1i foci co-localized with Myci signals, and Pvt1i and Myci puncta overlapped with exonic Pvt1 (Pvt1e) and Myc (Myce) foci, respectively, confirming the detection of nascent transcripts rather than excised introns (Figures 1B and S1). Quantification revealed that cells contained 1–4 Pvt1i foci per cell and averaged at 2.2 Pvt1i foci per cell, consistent with constitutive expression of Pvt1 in diploid cells (Figures 1A and 1B). Myci foci similarly averaged at 1.8 foci per cell but exhibited a wider distribution (Figure 1B). On the one hand, 56% of cells harbored one or no Myci signals, indicating that one or both of Myc loci were transcriptionally inactive (Figures 1A and 1B). On the other hand, 18% of cells harbored 5 or more Myci signals (Figures 1A and 1B). Altogether, the observed Pvt1 pattern confirmed constitutive expression and accumulation of Pvt1 in the chromatin at its sites of transcription, as previously reported.²³ In contrast, the Myc pattern suggested dynamic regulation of Myc transcription.

Figure 1. Pvt1 RNA production limits Myc transcription.

(A) smRNA-FISH visualization of nascent Pvt1 (Pvt1i, green) and Myc (Myci, red) (left) and mature (Myce, green) and nascent (Myci, red) Myc (right) using indicated probes in wild-type (WT) MEFs. DNA stained with DAPI, blue. Scale bar is 5 μM. Enlarged images highlight signal pattern and co-localization at sites of transcription.

(B) Quantification of the fraction of cells (n ≥ 50) with indicated number of Pvt1i and Myci foci per cell from experiment in (A). Foci were counted independent of size in two biological replicates.

(C) Schematic of the Pvt1/Myc locus, highlighting the insertion of the polyadenylation signal (PAS) in exon 1a of Pvt1.

(D) qRT-PCR detection of Pvt1a, Pvt1b, and total Pvt1 levels in RNA isolated from littermate WT (+/+) and homozygous PAS mutant (P/P) MEFs.

(E and F) smRNA-FISH visualization of nascent Pvt1 (Pvt1i, green) and Myc (Myci, red) (E) and mature (Myce, green) and nascent (Myci, red) Myc (F) using indicated probes in +/+ and P/P MEFs. DNA stained with DAPI, blue. Scale bar is 5 μM. Enlarged images highlight signal pattern and co-localization at sites of transcription.

(G) Violin plot showing the distribution of the number of Myci foci per cell in +/+ and P/P MEFs in three biological replicates (Rep). n indicates the number of cells scored in each experiment. Red lines indicate median.

(H) Average Myci foci per cell from experiments in (G).

(I) Distribution of Myci foci per cell from experiments in (G), showing fraction of cells with indicated number of Myci foci.

(J) RT-qPCR detection of nascent and spliced Myc in RNA isolated from +/+ and P/P MEFs.

(K) Representative immunoblot and quantification of Myc protein levels in whole-cell extracts (WCEs) from +/+ and P/P MEFs. Hsp90, loading control. Ladder markers (kilodalton [kDa]) indicated.

(L) Cumulative frequency distribution plot showing the upregulation of genes from GSEA Hallmark geneset MYC_TARGETS_V1 relative to an expression-matched control set in P/P relative to +/+ MEFs in RNA-seq analysis of 4 biological replicates. KS test used to determine p value.

In this figure, bar graphs show individual values and mean ± SEM of indicated number of biological replicates. Paired t test (D and H–K) and unpaired t test (G), *p < 0.05, **p < 0.01, and ***p < 0.01; ns, not significant.

See also Figures S1, S2, S3, and Table S1.

Development and characterization of Pvt1 premature termination mutant

To determine whether the production and accumulation of Pvt1 in the chromatin near Myc influences its transcriptional dynamics, we inserted the 49-nucleotide synthetic premature polyadenylation signal (PAS, P)²⁴ in exon 1a of endogenous Pvt1 in murine blastocysts (Figure 1C). Germline transmission of the P allele was confirmed by genotyping (Figure S2A). Heterozygous crosses established that homozygous mutant mice (Pvt1^P/P, P/P) were born at Mendelian ratio, were fertile, and did not display any apparent developmental or aging abnormalities (Figures S2B–S2D). We next isolated littermate WT (Pvt1^+/+, +/+) and homozygous P/P (PAS) MEFs and analyzed the effects of PAS insertion on isoform-specific and total Pvt1 levels by qRT-PCR. We observed that, compared to littermate +/+ controls, P/P MEFs exhibited a 47 ± 4% downregulation of the constitutively expressed Pvt1a isoform, accompanied by a compensatory 35 ± 7% increase of the downstream, p53-regulated Pvt1b isoform, which has previously been shown to be expressed at very low levels in absence of stress (Figure 1D).²³ Overall, there was a 27 ± 6% decrease in total Pvt1 levels in P/P compared to +/+ MEFs (p < 0.05, Figure 1D). A similar pattern was observed in RNA isolated from tissues from adult Pvt1^P/P mice compared to WT controls (Figure S2E). We concluded that insertion of the PAS element in exon 1a of Pvt1 led to a partial downregulation of Pvt1 expression.

Pvt1 controls the dynamics of Myc transcription

To investigate whether the Pvt1 mutation affected the dynamics of Myc transcription, we performed smRNA-FISH in +/+ and P/P MEFs. As above, Myci foci co-localized with Pvt1i and Myce signals, indicating nascent transcripts (Figures 1E and 1F). Interestingly, we observed increased number of Myci foci in P/P compared to +/+ MEFs (Figures 1E and 1F). We quantified the number of Myci foci per cell and confirmed a reproducible increase in P/P compared to +/+ MEFs across biological replicates (+/+: 2.5 ± 0.5, P/P: 4.2 ± 0.3 average Myci foci per cell, p < 0.05; Figures 1G and 1H). The difference was attributable to an approximately 50% increase in the fraction of cells with ≥5 Myci foci (+/+: 21 ± 4%, P/P: 32 ± 3% of cells, p < 0.05; Figure 1I). Conversely, the fraction of cells with 0 Myci foci was reduced in P/P compared to +/+ cells (+/+: 31 ± 8%, P/P: 16 ± 6% of cells, p < 0.05; Figure 1I). These findings indicated that high levels of Myc transcription occurred in a larger fraction of P/P MEFs, while inactive Myc loci were more frequent in +/+ MEFs. We concluded that, despite the partial effects of premature termination, Pvt1-mutant MEFs exhibited altered dynamics of Myc transcription.

Pvt1 negatively regulates Myc RNA and protein steady-state levels and function

Consistent with the increased frequency of cells with higher levels of Myc transcription, we observed significant increases in nascent and spliced Myc RNA levels by 14 ± 2% and 30 ± 26%, respectively, and in Myc protein levels by 33 ± 23% in P/P MEFs compared to +/+ controls (Figures 1J and 1K). Of note, we did not observe any significant changes in Myc protein translation or degradation rates in P/P compared to +/+ MEFs (Figures S3A–S3D). These findings led us to conclude that Pvt1 negatively regulates the steady-state Myc RNA and Myc protein levels, independent of other known mechanisms of Myc control.

We also examined the genome-wide consequences of Pvt1 mutation by performing RNA-seq analysis of +/+ and P/P MEFs in four biological replicates (Table S1). We validated an overall reduction, but not loss, of Pvt1a levels as well as a compensatory increase in Pvt1b levels, consistent with qRT-PCR analysis (Figures S3E and S3F). We also confirmed a consistent increase in Myc levels across all replicates (Figure S3F). Importantly, genes responsive to Pvt1 loss, determined by principal component analysis (PCA), were strongly enriched for Myc targets (Figures S3G–S3I). Similarly, a significant increase in the expression of Myc targets in P/P compared to +/+ MEFs was validated in comparison to a random expression-matched control set (Figure 1L). We concluded that Pvt1 restricts Myc RNA and Myc protein levels and limits the transcriptional output downstream of Myc.

Pvt1 abundance controls Myc transcription

We next asked whether Myc is responsive to the expression levels of locally produced Pvt1 transcripts. To overcome the partial knockdown of the PAS mutation, we reasoned that deletion of p53 in P/P MEFs will eliminate the compensatory expression of the p53-dependent Pvt1b isoform (Figure 2A). We crossed Pvt1^P/P; p53^Δ/+ mice and isolated MEFs from control Pvt1^P/P; p53^+/+ and three independent Pvt1^P/P; p53 ^Δ/Δ littermate embryos. Analyses were performed at passage 2–4 of primary MEF cultures, prior to any indirect, proliferation-mediated effects of p53 deficiency on Myc levels or the development of aneuploidy. While the constitutive isoform, Pvt1a, was not affected by p53 deficiency, the p53-dependent Pvt1b isoform was reduced by over 90% (Figure 2B). Consequently, the total transcriptional output from the Pvt1 locus in Pvt1^P/P; p53^Δ/Δ was reduced by 45 ± 6% compared to Pvt1^P/P; p53^+/+ littermate MEFs, corresponding to a 60 ± 6% reduction compared to WT MEFs (Figures 2A and 2B). The additional reduction of Pvt1 levels led to a notable increase in the number of Myci foci per cell (Figure 2C). Quantification confirmed an increase in the average number of Myci foci per cell from 3.2 in Pvt1^P/P; p53^+/+ MEFs to 5.5 ± 0.2 in Pvt1^P/P; p53^Δ/Δ MEFs (p < 0.05; Figures 2D and 2E). We also observed a significant increase in the fraction of cells with ≥5 Myci foci from 29% in Pvt1^P/P; p53^+/+ MEFs, consistent with the fraction reported in P/P MEFs (32 ± 3%; Figure 1I), to 47 ± 3% in Pvt1^P/P; p53^Δ/Δ MEFs (p < 0.05; Figure 2F). Finally, we observed a significant rise in the levels of nascent and spliced Myc RNA, by 54 ± 40% and 20 ± 7%, respectively (Figure 2G). We concluded that Pvt1 modulates Myc transcriptional dynamics and output in a dose-dependent manner.

Figure 2. Pvt1 is a dose-dependent repressor of Myc transcription.

(A) Schematic of the two alternative transcription start sites of Pvt1, highlighting the p53 response element (p53RE, pink star) and visualizing Pvt1 isoform transcriptional states (black [Pvt1a] or orange [Pvt1b] solid [high expression] or dashed [low/no expression] lines) in wild-type (Pvt1^+/+; p53^+/+), PAS mutant only (Pvt1^P/P; p53^+/+), and combined Pvt1 mutant and p53 knockout (Pvt1^P/P; p53^Δ/Δ) MEFs. Percentages indicate the observed transcriptional output of total Pvt1 in each cell line normalized to wild-type cells.

(B) qRT-PCR detection of Pvt1a, Pvt1b, and total Pvt1 levels in RNA isolated from littermate Pvt1^P/P; p53^+/+ and Pvt1^P/P; p53^Δ/Δ MEFs.

(C) Representative examples of smRNA-FISH visualization of nascent Pvt1 (Pvt1i, green) and Myc (Myci, red) (left) and mature (Myce, green) and nascent (Myci, red) Myc (right) using indicated probes in Pvt1^P/P; p53^Δ/Δ MEFs. DNA stained with DAPI, blue. Scale bar is 5 μM. Enlarged images highlight signal pattern.

(D) Violin plot showing the distribution of the number of Myci foci per cell in littermate Pvt1^+/+; p53^+/+ and three independent Pvt1^P/P; p53^Δ/Δ MEFs lines. n indicates the number of cells scored in each experiment. Red lines indicate median.

(E) Average Myci foci per cell from experiments in (D). Paired t test compares the three independent Pvt1^P/P; p53^Δ/Δ MEF lines to the littermate Pvt1^+/+; p53^+/+ control.

(F) Distribution of Myci foci per cell from experiments in (D), showing fraction of cells with indicated number of Myci foci. Statistical analysis performed as in (E).

(G) qRT-PCR detection of nascent and spliced Myc in RNA isolated from indicated MEFs.

In this figure, bar graphs show individual values and mean ± SEM of indicated number of biological replicates. Paired t test (B and G) and unpaired t test (D), *p < 0.05, **p < 0.01, and ***p < 0.01; ns, not significant.

Pvt1 restricts Myc oncogenic activities

Given the role of Pvt1 in limiting Myc expression, we next asked whether Pvt1 inhibition affects Myc functions, such as promoting cellular proliferation and oncogenic transformation.^13,25 Growth curve analyses revealed that multiple, independently isolated P/P MEF lines proliferated faster compared to their littermate +/+ controls (Figures 3A and 3B). Moreover, while all three +/+ MEF lines senesced at passage 10, two out of three P/P MEFs continued to proliferate beyond passage 14, indicating increased predisposition to spontaneous immortalization (Figure 3C). Additionally, the Pvt1 mutation cooperated with oncogenes, such as E1A, to promote cellular transformation, assessed by low-density colony formation assay (Figure 3D). Together, these findings indicated that Pvt1 restrains Myc oncogenic activities in vitro.

Figure 3. Pvt1 restrains Myc proliferative and oncogenic functions.

(A) Growth curve analyses showing cumulative cell number over passaging of three independent +/+ and P/P MEF littermate pairs.

(B) Fraction of BrdU-positive cells in (A).

(C) Representative brightfield images of cells from (A) visualized at passage 14. +/+ MEFs show features of cellular senescence, while P/P MEFs maintain high mitotic index (arrowheads).

(D) Representative examples and quantification of colony formation assay in +/+ and P/P cells expressing empty vector (EV) or E1A oncogene.

(E) Representative H&E images of lungs from 5-month-old Kras-LA1; Pvt1^+/+ (n = 18) and Kras-LA1; Pvt1^P/P (n = 18) mice. Enlarged images highlight grade 1 and mixed grade 1–2 tumors in Kras-LA1; Pvt1^+/+ and Kras-LA1; Pvt1^P/P lungs, respectively.

(F) Quantification of tumor burden in mice from (E).

(G) Quantification of the fraction of tumors assigned to AAH-1, 2–3, or 4+ grade categories in mice from (E). Numbers above bars indicate the number of tumors scored in each condition.

(H) Quantification of the number of pHH3-positive cells per tumor area in mice from (E).

(I) Schematic of experiment and representative H&E images of lungs from KP; Pvt1^+/+ (n = 8) and KP; Pvt1^P/P (n = 6) mice at 12 weeks post-tumor initiation (pti) by intratracheal infection with adenoviral Cre (AdCre).

(J) Quantification of tumor burden in mice from (E).

(K) Quantification of the fraction of tumors assigned to AAH-1, 2–3, or 4+ grade categories in mice from (I). Numbers above bars indicate the number of tumors scored in each condition.

In this figure, bar graphs show individual values and mean ± SEM of indicated number of biological replicates. Unpaired t test, *p < 0.05, **p < 0.01, and ***p < 0.01.

We also asked whether Pvt1 plays a tumor suppressive role at the organismal level. Spontaneous activation of a latent K-ras-G12D allele in lung epithelial cells in the K-ras^LA1 (LA1) mouse model has been shown to initiate the development of 20–40 lung lesions in 100% of animals within 4 months.²⁶ Importantly, in this model, tumors begin as atypical adenomatous hyperplasia (AAH) and progress to adenomas and adenocarcinomas in a synchronous fashion, allowing the investigation of modulators of tumor progression. Previous work has shown that Myc is required for K-ras-G12D-driven lung tumorigenesis,²⁷ allowing us to investigate the impact of Pvt1-mediated Myc regulation on tumor development. To this end, we crossed K-ras^LA1; Pvt1^P/+ mice to generate a cohort of littermate K-ras^LA1; Pvt1^+/+ (n = 18) and K-ras^LA1; Pvt1^P/P (n = 18) mice. Histopathological analysis of hematoxylin and eosin (H&E)-stained sections at 5 months of age revealed a 1.4-fold increase in tumor burden in K-ras^LA1; Pvt1^P/P compared to K-ras^LA1; Pvt1^+/+ mice (p < 0.05; Figures 3E and 3F). Additionally, we observed that approximately half of the tumors in K-ras^LA1; Pvt1^P/P animals were grade 2 adenomas or grade 3 adenocarcinomas, whereas 75% of tumors in K-ras^LA1; Pvt1^+/+ mice displayed characteristics of AAH and grade 1 tumors (Figures 3E and 3G). The increased tumor burden and advanced grade resulted from enhanced cellular proliferation, determined by immunohistochemistry for the mitotic marker phosphorylated histone H3 (pHH3; Figure 3H).

We further investigated whether p53-deficient tumors, which are expected to lack Pvt1b and thus harbor a further reduction in overall Pvt1 levels (Figure 2A), exhibit a more pronounced increase in tumor growth. We used a previously developed Cre-activatable K-ras^LSL-G12D/+; p53^FL/FL (KP) mouse model, in which lung tumorigenesis is initiated by intratracheal infection with adenoviral Cre (AdCre) to simultaneously activate a mutant K-ras G12D allele and inactivate p53, resulting in synchronous tumor development from adenoma to high-grade adenocarcinoma over 8–16 weeks.^28,29 We generated a cohort of littermate KP; Pvt1^+/+ (n = 8) and KP; Pvt1^P/P (n = 6) mice and harvested tumor-bearing lungs at 12 weeks post tumor initiation (pti) with AdCre. We observed a dramatic 2.6-fold increase in tumor burden in Pvt1 mutant compared to control lungs (KP; Pvt1^+/+: 10.3 ± 2.0%, KP; Pvt1^P/P: 27.0 ± 4.1%, p < 0.001; Figures 3I and 3J). In contrast, we did not observe any differences in tumor grade, consistent with p53 deficiency being a powerful driver of tumor evolution to advanced disease (Figure 3K). Altogether, these findings demonstrated that Pvt1 plays a dose-dependent, tumor suppressive role that restricts tumor growth in vivo.

Pvt1 does not significantly contribute to genome organization and chromatin accessibility in the Myc locus

To investigate the mechanism by which Pvt1 affects Myc transcriptional dynamics, we compared the chromatin accessibility of the Myc/Pvt1 locus in control and Pvt1 mutant MEFs. Analysis of genome-wide chromatin accessibility by ATAC-seq revealed limited changes in P/P compared to +/+ MEFs (14 peaks upregulated, 38 peaks downregulated; n = 4 biological replicates; FDR < 0.01), which included only one significantly depleted peak in the Myc/Pvt1 locus near the PAS insertion site (Figures 4A and 4B; Table S2). We concluded that changes in chromatin accessibility are not responsible for the differential Myc transcription in +/+ and P/P MEFs.

Figure 4. Pvt1 does not significantly contribute to chromatin accessibility and three-dimensional organization of the Myc locus.

(A) ATAC-seq read coverage of left, Myc, and right Pvt1 gene bodies, showing comparable patterns in +/+ and P/P MEFs in 4 biological replicates, with the exception of one significantly reduced peak in the promoter of Pvt1a (FDR < 0.01), indicated by a purple star.

(B) Volcano plot showing log2FC (fold change) of peaks in Pvt1 mutant (PAS) compared to wild-type (WT) MEFs from ATAC-seq experiment in (A). Dot colors represent peaks with significantly (FDR < 0.01) reduced (blue), increased (red), or unaltered (black) chromatin accessibility. Green arrow highlights the peak in the Pvt1a promoter.

(C) Top: ATAC-seq and RNA-seq profiles of the Myc/Pvt1 locus in +/+ and P/P samples included for reference. Middle: HiC arc plot visualization of loops in the Myc/Pvt1 locus detected from merged reads of two biological replicates of +/+ and P/P MEFs. Differential enrichment of loops between +/+ and P/P MEFs is not statistically significant (black) and can be attributed to increased sequencing depth in P/P compared to +/+ MEFs (Figure S4B). Bottom: arc plot visualization of significantly (FDR < 0.001) depleted (blue) or enriched (red) intrachromosomal contacts (ICs) in the Myc/Pvt1 locus in P/P compared to +/+ MEFs.

(D) Volcano plot showing log2FC (fold change) of chromosome 15 ICs in Pvt1 mutant (PAS) compared to WT MEFs from HiC experiment in (C). Dot colors represent ICs that are significantly (FDR < 0.001) depleted (blue), enriched (red), or unaltered (black). Green halos highlight differential contacts in the Myc/Pvt1 locus, illustrated in (C).

See also Figures S4 and Tables S2, S3, and S4.

We next examined whether Pvt1 affected the three-dimensional architecture of the Myc/Pvt1 locus by performing Hi-C in +/+ and P/P MEFs in two biological replicates. As expected, premature termination of Pvt1 did not alter genome-wide open and closed chromatin compartments, also known as A and B compartments, including on chromosome 15 (Figure S4A). Next, we focused on loops, which represent stable, long-range contacts between distant genomic locations on the same chromosome that define TADs (topologically associated domains). While we detected more loops in the Myc/Pvt1 locus in P/P compared to +/+ samples (Figure 4C, black lines), the strength of loops was not significantly altered between the samples (Table S3). Instead, the difference was attributed to increased sequencing depth in P/P samples (genome-wide loop count in +/+: 4,766 and 4,667 and P/P: 6,066 and 5,449, Figure S4B). Out of approximately 150 significant differential loops in P/P compared to +/+ MEFs, none were located near the Myc/Pvt1 locus (Figures S4C and S4D; Table S3). Lastly, we considered intrachromosomal contacts (ICs), which are low-count, transient associations that may be indicative of enhancer-promoter interactions. Out of over 15,000 ICs identified on chromosome 15, only 68 were differentially enriched between +/+ and P/P MEFs (FDR < 0.001; Figure 4C). Of those, 6 mapped to the Myc/Pvt1 locus, but none overlapped with the Myc promoter or gene body (Figures 4D and S4E; Table S4). Considering these observations, we concluded that Pvt1 does not appear to play a major role in the three-dimensional organization of the Myc/Pvt1 locus.

Pvt1 deficiency further enhances Myc transcription

We wondered whether the partial inhibition of Pvt1 in PAS mutants resulted in limited phenotypes. To elucidate the full extent of Myc regulation by Pvt1, we sought to generate cells that lack Pvt1 expression. We used dual gRNA CRISPR mutagenesis to generate a clonal cell line with a deletion of the promoter of Pvt1, including the alternative transcription start sites of both Pvt1 isoforms, Pvt1a and Pvt1b, in p53-deficient MEFs (Figures 5A, S5A, and S5B). We confirmed 79 ± 17% loss of expression of all Pvt1 isoforms by qRT-PCR and absence of Pvt1-specific signals by smRNA-FISH in promoter deletion cells (ΔPro) compared to a control gRNA-expressing clonal MEF line (Con; Figures 5B and S5C). As above, loss of Pvt1 led to a significant increase in the levels of nascent Myc by 65 ± 22% and spliced Myc by 36 ± 12% as well as a significant increase in Myc protein levels by 46 ± 20% (Figures 5C and 5D). Of note, Con and ΔPro cells grew at similar rates, indicating that the changes in Myc expression were due to Pvt1 loss and not resulting from differences in proliferation (Figure 5E).

Figure 5. Pvt1 controls Myc transcriptional bursting.

(A) Schematic of the Pvt1/Myc locus and Pvt1 promoter eletion strategy.

(B) qRT-PCR detection of Pvt1a, Pvt1b, and total Pvt1 levels in RNA isolated from p53-deficient control gRNA (Con) and Pvt1 promoter deletion (ΔPro) MEF lines.

(C) qRT-PCR detection of nascent and spliced Myc in RNA isolated from Con and ΔPro MEFs.

(D) Representative immunoblot and quantification of Myc protein levels in whole-cell extracts (WCEs) from Con and ΔPro MEFs. Hsp90, loading control. Ladder markers (kilodalton [kDa]) indicated.

(E) Growth curve analysis of Con and ΔPro MEFs.

(F) smRNA-FISH visualization of nascent Pvt1 (Pvt1i, green) and Myc (Myci, red) using indicated probes in Con and ΔPro MEFs. DNA stained with DAPI, blue. Scale bar is 5 μM. Enlarged images highlight signal pattern and co-localization at sites of transcription. Note p53-deficient cells are aneuploid (3N).

(G) Violin plot showing the distribution of the number of Myci foci per cell in Con and ΔPro MEFs in three biological replicates (Rep). n indicates the number of cells scored in each experiment. Red lines indicate median.

(H) Average Myci foci per cell from experiments in (G).

(I) Distribution of Myci foci per cell from experiments in (G).

(J) Left: representative loci illustrating inactive, active, and bursting Myc transcriptional states in smRNA-FISH co-localization of indicated probes; Pvt1 is constitutively transcribed. Bursting Myc loci are defined as nuclear regions containing ≥4 Myci independent foci within a 20 μm² area. Right: quantification of the fraction of cells with bursting Myc loci in Con and ΔPro MEFs.

(K) Combined IF-smRNA-FISH visualization of nascent Pvt1 (Pvt1i, green) and Myc (Myci, red) (left) and Brd4 (green) and Myc (Myci, red) (right) in Con and ΔPro MEFs. Scale bar is 5 μM. Enlarged images illustrate signal and co-localization patterns.

(L) Quantification of the fraction of Myci foci that were found to co-localize (Brd4+) or not co-localize (Brd4−) with Brd4 in experiment in (K) in 4 biological replicates.

In this figure, bar graphs show individual values and mean ± SEM of indicated number of biological replicates. Paired t test (B–D, H–J, and L) and unpaired t test (G), *p < 0.05, **p < 0.01, and ***p < 0.01; ns, not significant.

See also Figure S5.

As above, by smRNA-FISH, we observed a significant increase in the average number of Myci foci per cell (Con: 5.5 ± 0.4 and ΔPro: 9.3 ± 0.8 Myci foci per cell, p < 0.05; Figures 5F–5H and S5D) and a larger fraction of cells with high (≥13) Myci foci per cell in ΔPro compared to Con cells (Con: 8.7 ± 2.4%, ΔPro: 30.3 ± 3.8% of cells, p < 0.01; Figure 5I). The overall increase of Myci foci in these experiments compared to primary MEFs was attributed to a combined effect of p53 deficiency and triploid genomes in Con and ΔPro clonal lines. Consistent with the results above (Figures 4A and 4B), loss of Pvt1 in ΔPro MEFs did not significantly affect chromatin accessibility in the Myc/Pvt1 locus compared to Con MEFs (Figure S5E).

Altogether, the effects of Pvt1 loss on Myc transcription in ΔPro MEFs were consistent with and more robust than the observations in PAS MEFs (Figures 1E–1I), providing independent support for our conclusion that Pvt1 abundance negatively regulates Myc transcriptional dynamics.

Pvt1 limits Myc bursting

Closer examination of smRNA-FISH images revealed that whenever Con and ΔPro cells harbored multiple Myci foci, these frequently clustered near one of the loci, consistent with transcriptional bursts. We hypothesized that Pvt1 perturbations were upregulating Myc transcription by increasing the frequency of Myc bursting. To quantify Myc bursts, we defined a locus as “bursting” if it harbored ≥4 Myci foci in a fixed area (20 μm² circle), independent of focus size (Figure 5J, left). Consistent with our hypothesis, we observed that the fraction of cells with Myc bursting loci increased by 2.8-fold in ΔPro MEFs compared to Con MEFs (Con: 26.0 ± 0.5%, ΔPro: 72.8 ± 9.4% of cells, p < 0.01; Figure 5J, right). We concluded that Pvt1 controls Myc expression by limiting the frequency of Myc bursting.

Previous work has linked the onset of transcriptional bursting with co-localization with transcriptional condensates,³⁰ which contain Pol2, transcription factors, and co-factors, such as Brd4. To determine whether increased Myc bursting in ΔPro MEFs correlates with increased condensate localization at the Myc locus, we performed combined immunofluorescence (IF)-smRNA- FISH detecting Brd4, nascent Pvt1 (Pvt1i), and nascent Myc (Myci) (Figure 5K). As previously reported, we observed that Brd4 displayed homogeneous nuclear distribution of uniformly sized foci throughout euchromatin (Figure 5K). We scored the number of Myci foci that did not co-localize (Brd4−) or co-localized (Brd4+) with Brd4 and observed that, while over 60% of Myci foci did not co-localize with Brd4 in Con cells, over 60% of Myci foci co-localized with Brd4 in ΔPro (Figure 5L). This finding indicated that Pvt1 limits the frequency of interaction of transcriptional condensates with the Myc promoter.

Regulation of Myc bursting frequency and size by Pvt1 in cis

Lastly, we addressed whether Pvt1 affects the size of Myc bursts. We quantified Myci signal intensity in a triploid heterozygous Pvt1 promoter deletion clonal cell line that harbors two deletion (ΔPro) and one WT (+) alleles (Pvt1^{ΔPro/ΔPro/+}, Het; Figure 6A). Importantly, the ΔPro alleles, which lack Pvt1i signal, were compared to WT alleles, marked by Pvt1i signal, in the same cell (Figure 6A). We first quantified the fraction of alleles with bursting Myc, defined as above by foci count (Figure 5J), and confirmed the increased frequency of bursting at ΔPro alleles (60% and 40%) compared to WT alleles (14% and 12%; Figure 6B). Next, we quantified the Myci mean intensity (MI) signal in a fixed area (20 μm² circle) in two biological replicates and observed a significant increase in the average MI at ΔPro alleles (300.5 and 271.6 arbitrary units [a.u.]) compared to WT alleles (113.9 and 119.8 a.u.; Figures 6C and 6D). We also found a significant increase in the fraction of ΔPro alleles with MI ≥ 201 a.u. (58.1% and 48.3%) compared to WT alleles (20.7% and 27.3%; Figure 6E). We concluded that Pvt1 confers allele-specific control of Myc bursting frequency and size.

Figure 6. Allele-specific control of Myc bursting frequency and size by Pvt1.

(A) Representative smRNA-FISH visualization of nascent Pvt1 (Pvt1i, green) and Myc (Myci, red) using indicated probes in triploid p53-deficient MEFs with heterozygous Pvt1 promoter deletion (ΔPro/ΔPro/+, Het). DNA stained with DAPI, blue. Scale bar is 5 μM. Enlarged images highlight signal pattern at Pvt1-proficient wild-type (WT) and Pvt1-deficient mutant (ΔPro) loci.

(B) Quantification of the fraction of WT and ΔPro alleles in Het MEFs with bursting Myc loci, as defined in Figure 5J.

(C) Violin plot showing the distribution of the mean intensity (MI) of Myci foci at WT and ΔPro alleles in Het MEFs in two biological replicates (Rep). MI is represented as arbitrary units (a.u.). n indicates the number of alleles scored in each experiment. Red lines indicate median. Unpaired t test (G), *p < 0.05 and ***p < 0.01.

(D) Average Myci MI per allele from experiments in (C).

(E) Distribution of Myci MI per allele from experiments in (C).

In this figure, bar graphs show individual values and mean ± SEM of indicated number of biological replicates. Color of individual data points indicates paired samples.

DISCUSSION

In this study, we present evidence that constitutively expressed Pvt1 is a negative regulator of Myc expression and function. We observed that the inhibition of Pvt1 by premature termination or by promoter deletion led to increased Myc transcription, resulting in elevated Myc RNA and Myc protein levels and enhanced Myc transcriptional activity. Our conclusion contrasts previous observations that the Pvt1 RNA is a positive regulator of Myc protein stability¹⁸ or that the Pvt1 locus does not regulate Myc transcription through RNA-dependent mechanisms.¹⁶ The differences can be explained by our development of the Pvt1 premature termination PAS genetic model, which enabled unique insights through dissociation of RNA and DNA elements in the Pvt1 locus.

The finding that increased nascent transcription of Myc in Pvt1 mutant cells led to proportional increases in mature Myc RNA and protein levels was unexpected and indicated a surprising absence of negative feedback in Myc regulation. Our interpretation is that Pvt1 modulates Myc expression locally, autonomously, and independent of previously established post-transcriptional and post-translational Myc regulatory mechanisms.

As a consequence of restricting Myc transcription, we found that Pvt1 limited Myc transcriptional activity and cellular and organismal functions. The extent of the observed proliferation and tumor growth phenotypes was surprising given the significant but relatively minor increases in Myc levels. However, prior work has shown that Myc activity is tightly linked to its abundance,³¹ providing an explanation for how small changes in Myc levels can have significant effects on cellular proliferation and oncogenic transformation in vitro and in vivo. It is important to note that functional experiments were performed in the context of the hypomorphic Pvt1 PAS mutation. Complete loss of Pvt1 would be expected to cause larger changes in Myc levels and activity.

The data revealed that Pvt1 inhibition promoted cellular proliferation in primary cells and tumor growth in mouse models of lung adenocarcinoma. These findings point to constitutively expressed Pvt1 as a tumor suppressor factor. This conclusion is consistent with the presence of recurrent cancer-associated genetic aberrations in the Pvt1 locus, including deletions in the Pvt1 promoter^16,32 and Myc translocations with breakpoints within the Pvt1 promoter,²¹ both of which are expected to prevent Pvt1 transcription near Myc. Prior studies have also reported preferential viral integrations in the Pvt1 promoter in a murine model of T cell lymphomas induced by retroviral insertions.³³ While these insertions were reported to increase Pvt1 expression, they might have disengaged the Myc promoter from Pvt1-mediated negative feedback.

Our initial hypothesis was that Pvt1 regulates Myc by influencing chromatin accessibility and/or architecture at the Myc promoter, which interacts with the Pvt1 promoter, or at Myc enhancers, many of which are located within the gene body of Pvt1.^16,23 Analogous mechanisms have been proposed for other cis-regulatory lncRNAs.¹² However, we did not find significant differences in chromatin accessibility, and we did not observe significant changes in chromatin architecture, including in the formation of A/B compartments, loops, and ICs at the Myc/Pvt1 locus. Thus, although we looked exhaustively, we did not find evidence to support a role for Pvt1 in the regulation of the chromatin accessibility or the three-dimensional organization of the Myc locus.

The visualization of nascent Myc transcripts as clusters, whose frequency and size were elevated in cells lacking Pvt1, suggested a role for Pvt1 in modulating Myc transcriptional bursting. Moreover, we noticed that in cells lacking Pvt1, a greater proportion of nascent Myc foci co-localized with Brd4, a component of transcriptional condensates, which associate with promoters to initiate transcriptional bursts. Together, these observations indicate that Pvt1 is a local inhibitor of Myc bursting. This conclusion is consistent with a recent study, which used allele-specific single-cell sequencing to infer bursting parameters of lncRNAs and their protein-coding neighbors, and similarly observed that lncRNAs can modulate the burst kinetics of nearby genes.⁸ It remains to be determined whether specific lncRNA elements and contexts, such as chromatin-retention features, RNA stability determinants, or locus-specific organization, play a role.

We also considered what features of Myc bursting are controlled by Pvt1. In live-cell imaging studies, the output of transcriptional bursting is defined by frequency, duration, and amplitude.^4,5 Although these parameters are challenging to infer from static images, our observation that Pvt1 inhibition increased both the frequency and the size of bursts can be reconciled with an increase in burst duration. Indeed, longer duration will result in more frequent detection of bursts in static images, which we observed as increased fraction of loci with bursts, and will give rise to higher output per locus, which we observed as increased signal intensity per locus.

Taking all into consideration, we propose that Pvt1’s role in Myc regulation best fits in the context of the RNA-mediated negative feedback model. This model postulates that local RNA concentration plays a dynamic role in seeding of transcriptional condensates at low concentration and dissociating at high concentration, which in turn drive the cyclical nature of transcriptional bursting.⁷ We propose that Pvt1 contributes to the overall local RNA concentration in the vicinity of the Myc promoter and may play a role in seeding transcriptional condensates, although this was not addressed in this study (Figure 7). Myc burst duration lasts until the sum of chromatin-associated Pvt1 transcripts plus newly transcribed Myc RNA molecules reaches a threshold that triggers the RNA-mediated negative feedback mechanism that disperses transcriptional condensates and reverts the Myc locus to an inactive state. One prediction from this model is that changes in Pvt1 expression that alter the baseline RNA concentration to a lower setpoint will increase the duration of Myc bursts, resulting in more Myc molecules to reach the same threshold that triggers the RNA-mediated feedback mechanism (Figure 7, middle). This is consistent with the higher frequency, larger bursts, and increased localization of transcriptional condensates at the Myc locus observed in Pvt1 mutants compared to control cells in a dose-dependent manner. A second prediction from this model is that changes that increase baseline Pvt1 expression will lead to reduced Myc transcriptional duration and output (Figure 7, bottom). This prediction is consistent with published evidence that increased transcription from the Pvt1 locus, as a result of CRISPR activation (CRISPRa) or p53-dependent induction of Pvt1b isoform during stress, increases local RNA abundance to suppress Myc transcription.^16,22,23 Importantly, this model explains how changes in Pvt1 abundance can dynamically control Myc expression without altering chromatin accessibility or architecture.

Figure 7. Proposed model for role of Pvt1 abundance in the control of Myc burst duration.

Top: constitutively transcribed Pvt1 (green) maintains elevated levels of local RNA abundance in the vicinity of Myc promoter. This limits the number of Myc molecules (red) that are transcribed during each burst prior to reaching the local RNA concentration threshold that triggers dissociation from transcriptional condensates. Middle: reduced baseline Pvt1 abundance in loss-of-function studies permits the production of a greater number of Myc molecules in a longer burst before the same threshold is reached. Bottom: increased baseline Pvt1 abundance during the cellular response to stress limits the production of Myc molecules in a shorter burst before the same threshold is reached.

Limitations of the study

While in situ detection of Myc bursting by smRNA-FISH allowed detailed quantification of Myc bursting frequency and size across multiple genetic contexts, it could not dissect the temporal parameters of Myc bursting, which can be addressed by live-cell imaging. Additionally, further studies are required to elucidate what molecular features of Pvt1, such as overall length or specific sequence elements, are required for its repressive activity.

RESOURCE AVAILABILITY

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Nadya Dimitrova (nadya.dimitrova@yale.edu).

Materials availability

Mouse strains and cell lines generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

Data

RNA-seq, ATAC-seq, and HiC data have been deposited in NCBI GEO under accession numbers GSE306107, GSE306101, and GSE306102.

Code

This study does not report original code.

Other items

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

STAR★METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Mouse strains

Pvt1 PAS (P) mice were generated using CRISPR/Cas9-mediated engineering in C57BL/6J blastocysts at Jackson Laboratory. Briefly, embryos were electroporated with Cas9, a guide RNA targeting exon 1a of Pvt1 and PAS homology directed (HDR) templates, described in Table S5. Founders were crossed to wild-type C57BL/6J mice. Germline transmission was identified by PCR genotyping with GoTaq Green PCR Master Mix (Promega) with primers listed in Table S5. Correct alleles were confirmed by Sanger sequencing. Previously described Kras-LA1 mice²⁶ were generously provided by Dr. Jonathan Kurie (MD Anderson Cancer Center). Previously described p53 null (p53Δ/Δ)⁵⁰ and KP (Kras-LSL-G12D/+; p53F/F)²⁸ mice were purchased from Jackson Laboratory. Mice genotypes were confirmed by genotyping. To determine viability and normal development and for MEF isolation, mice with equal sex distribution were sacrificed between 1 and 3 months. For spontaneous tumor studies, mice with equal sex distribution were sacrificed at 5 months. For induced tumor studies, tumorigenesis was initiated in mice with equal sex distribution, aged between 3 and 6 months. Sex was not observed to have a significant effect on experimental outcomes. All mouse studies and procedures were conducted with the approval of the Yale University Institutional Animal Care and Use Committee.

Cell lines

All cell lines were maintained at 37°C in a humidified incubator with 5% CO₂. Primary MEFs were isolated from E13.5 embryos, resulting from timed matings between wild-type (WT) C57BL/6J mice or mice heterozygous for the allele of interest (Pvt1^P/+ or Pvt1^P/P; p53^Δ/+). Cell lines were tested for mycoplasma contamination and authenticated by PCR genotyping with primers, listed in Table S5. All experiments were performed with wild-type (WT) and littermate mutant MEFs between passages 2 and 8. When relevant, such as in RNA-seq, ATAC-seq, and Hi-C studies, sex of primary MEFs was determined and considered in analysis pipeline, as described in Methods Details. Primary MEFs were maintained in DMEM (Gibco) supplemented with 15% fetal bovine serum, 50 U mL⁻¹ penicillin-streptomycin, 2 mM L-glutamine, 0.1 mM non-essential amino acids, and 0.055 mM 2-mercaptoethanol.

Pvt1 promoter deletion was performed in previously described puromycin-sensitive, p53-deficient (PR, p53^LSL/LSL;Rosa26-CreER^T2) MEFs,^23,51 authenticated by genotyping and RNA analysis, using paired gRNA-mediated CRISPR mutagenesis. Paired gRNAs, listed in Table S5, were designed using CRISPETa tool⁵² and cloned downstream of U6 promoter in BRD1 lentiviral constructs (a gift from the Broad Institute) that co-expresses the puromycin resistance gene and spCas9. Control gRNA targeting dTomato (Con) was used as a negative control. Lentivirus was produced in 293T cells by co-transfecting BRD1 constructs with pCMV-dR8.2 dvpr (Addgene) and pCMV-VSV-G (Addgene) viral packaging constructs. Virus-containing supernatants supplemented with 4 μg/mL polybrene (Millipore Sigma) were used to infect PR MEFs by 3 consecutive lentiviral infections, delivered at 24 hr-intervals, followed by selection with 2 μg/mL puromycin (Millipore Sigma). Homozygous and heterozygous promoter deletion (ΔPro) mutants were identified in PR clones by genotyping with primers listed in Table S5 and validated by Sanger sequencing. PR MEFs and 293T cells were maintained in DMEM (Gibco) supplemented with 10% fetal bovine serum, 50 U mL⁻¹ penicillin-streptomycin, 2 mM L-glutamine, and 0.1 mM non-essential amino acids.

METHOD DETAILS

Cell treatments

To assess effects on Myc post-transcriptional regulation, cells were treated with 10 μM MG132 (Millipore Sigma) and 50 μg/mL Cycloheximide (Millipore Sigma). Empty vector or E1A pLPC puro retroviral constructs³⁴ were introduced into +/+ and P/P MEFs by 4 consecutive retroviral infections delivered at 12 hr-intervals with virus-containing supernatants from Phoenix ecotropic cells, supplemented with 4 μg/mL polybrene (Millipore Sigma), followed by selection with 2 μg/mL puromycin (Millipore Sigma).

Single-molecule FISH

Previously described (Pvt1i, Myci, Pvt1e, and Pvt1-ex1a)²³ and custom-designed (Myce) Quasar570 (Q570)- and Quasar670 (Q670)-conjugated FISH probe sets (Stellaris, LGC Biosciences) detecting Pvt1 and Myc nascent and mature transcripts are listed in Table S5. Single-molecule FISH (smRNA-FISH) was performed according to the manufacturer recommendations. Briefly, cells were grown on coverslips and fixed for 10 min in 4% methanol-free formaldehyde (ThermoFisher Scientific) diluted in RNAse-free 1xPBS (ThermoFisher Scientific) at RT, followed by 1xPBS washes. Cells were dehydrated overnight at 4°C in 70% EtOH diluted in DEPC-H₂O and stored in 70% EtOH for up to a week at 4°C. Coverslips were transferred to a hybridization chamber and equilibrated for 5 min in Wash Buffer A (Stellaris, LGC Biosciences) prepared with formamide (Millipore Sigma) according to manufacturer’s instructions. Cells were incubated overnight at 30°C with the indicated probes diluted 1:50 in Hybridization solution (Stellaris, LGC Biosciences). The next day, cells were washed 2 times for 30 min at 30°C in Wash Buffer A, incubated in Wash Buffer B (Stellaris, LGC Biosciences) for 5 min at RT, mounted in antifade reagent (Vectashield Mounting medium with DAPI, Vector Laboratories), and sealed with nail polish.

Combined IF-smRNA-FISH

Cells, grown on coverslips, were fixed in 4% methanol-free formaldehyde (ThermoFisher Scientific) diluted in RNAse-free 1xPBS (ThermoFisher Scientific) for 10 min at RT, washed twice with 1xPBS, permeabilized with 0.5% Triton X-100 diluted in 1xPBS for 10 min at RT, and washed three times with 1xPBS. Next, cells were blocked for 10 min in 4% BSA (Jackson Immunoresearch) in 1xPBS prior to incubation with anti-Brd4 primary antibody (1:250, Abcam) diluted in 1xPBS for 2 h at RT. Following three 1xPBS washes, cells were incubated with Alexa Fluor 488 AffiniPure Goat Anti-Rabbit IgG (H + L) secondary antibody (1:500, Jackson Immunoresearch) diluted in 1xPBS for 1 h at RT, followed by two 1xPBS washes. Cells were fixed in 4% formaldehyde for 10 min at RT, followed by two 1xPBS washes, prior to performing smRNA-FISH, as described above.

Image acquisition and analysis

smRNA-FISH and IF-smRNA-FISH images were captured using an Axio Imager 2 microscope system (Zeiss) with a PlanApo 63 × 1.4 oil DIC objective lens (Zeiss) using Zen 2.6 Pro software (Zeiss) and consistent settings (Binning - 2 × 2; Filter (target): exposure - Cy5 (Pvt1i): 2000 ms, Rhodamine (Myci): 2000 ms, EGFP (Brd4): 200 ms, DAPI (DNA); 20 ms). 8–10 z stack images at 0.25 μm distance were acquired and deconvolved using the Fast iterative method, followed by Maximum orthogonal projection, and Best Fit levels adjustment. Scoring of Myci foci per cell and per locus (within a 20 μm² circle) were performed manually. Quantification of Myci Mean Intensity (MI) per locus within a 20 μm² circle was performed using Measurement tool in arbitrary units (a.u.). MI of an equivalent area that lacked Myci foci was subtracted as a background in each cell. Images were exported as TIFF files and edited using Adobe Photoshop to highlight signals.

RNA isolation and qRT-PCR

For RNA-seq and qRT-PCR analysis, total RNA was isolated with the RNeasy Mini Kit (Qiagen). On-column DNAse digestion was performed with RNAse-free DNase Set (Qiagen) for analysis of nascent transcripts. Tissue RNA was isolated using TissueRuptor II (Qiagen), followed by RNeasy Mini Kit (Qiagen). 0.5–1 μg of total RNA was reverse transcribed using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). SYBR Green PCR master mix (Applied Biosystems) was used for quantitative PCR in triplicate reactions with primers, listed in Table S5. Relative RNA expression levels were calculated using the ddCt method compared to Gapdh and normalized to control samples.

Immunoblotting

Cells were collected, counted, and lysed in 2×Laemmli buffer (100 mM Tris-HCl pH6.8, 200 mM DTT, 3% SDS, 20% glycerol) at 0.5–1×10⁴ cells/μL. Samples were heated at 95°C for 7 min and passed through an insulin syringe. Protein from 1 × 10⁵ cells was separated on 10% SDS-polyacrylamide gels and transferred to nitrocellulose membranes (Bio-Rad). After blocking (5% milk, PBST), membranes were incubated overnight at 4°C in primary antibodies (anti-c-Myc (1:1,000, clone Y69, Abcam) and anti-Hsp90 (1:5,000, Cell Signaling Technology)), then 1 h at RT in secondary antibody (Peroxidase AffiniPure Donkey Anti-Rabbit IgG (H + L), 1:10,000, Jackson Immunoresearch). Protein bands were visualized using Pierce ECL Prime Western Blotting Substrate (ThermoFisher Scientific). Quantification of Myc and Hsp90 protein levels was performed using the rectangle selection and measure tools in ImageJ and Myc levels plotted relative to Hsp90 levels and normalized to negative control in relevant graphs. For cycloheximide experiments, Myc levels were normalized to negative control and half-life of Myc protein was determined using Prism8 software.

Cellular assays

To generate growth curves of control and Pvt1-mutant cells, cumulative cell numbers were plotted over time or passaging, as indicated, in three biological replicates. For BrdU incorporation assay, +/+ and P/P MEFs, grown on coverslips, were incubated with 10 μM BrdU (Millipore Sigma) for 4 h, prior to fixation in ice-cold 25% 0.05M glycine (pH 2.2)/75% ethanol in −20°C. Next day, cells were rehydrated in PBS and denatured in 4N HCl for 10 min at RT, followed by 3 × 5 min washes with PBS. Cells were blocked for 20 min in PBG (0.2% (w/v) cold water fish gelatin and 0.5% BSA in PBS) prior to 45 min incubation with FITC-conjugated anti-BrdU antibody (1:100, clone B44, BD Biosciences) diluted in PBG. Cells were washed 3 × 10 min in PBG and embedded in an antifade reagent (Vectashield Mounting medium with DAPI, Vector Laboratories). Images were captured using an Axio Imager 2 microscope system (Zeiss) with a PlanApo 63 × 1.4 oil DIC objective lens (Zeiss). BrdU-positive cells as a fraction of DAPI cells were quantified in ImageJ. For colony formation assays, 1 × 10³ EV (empty vector)- or E1A-expressing +/+ and P/P MEFs were plated in 6 cm dishes and monitored for colony formation. After 2 weeks, plates were washed with PBS, fixed in 25% MeOH/0.5% Crystal Violet for 10 min, and washed in ddH₂O.

Tissue and tumor analysis

Lung tumorigenesis was initiated in a cohort of littermate KP; Pvt1^+/+ and KP; Pvt1^P/P mice by intratracheal infection with 1 × 10⁷ pfu/mouse of high titer Adenoviral Cre (University of Iowa, Viral Vector Core Facility) as previously described.²⁹ Tumor-bearing lungs were inflated and fixed in 4% formaldehyde at 12 weeks post tumor initiation. Tissues dissected from adult ≥1 month old +/+ and P/P mice and 4% formaldehyde-inflated tumor-bearing lungs from 5-month old Kras-LA1; Pvt1^+/+ and Kras-LA1; Pvt1^P/P mice were fixed in 4% formaldehyde for 24 h prior to dehydration in 70% ethanol. Fixed tissues were embedded in paraffin, sectioned, and stained with hematoxylin and eosin (H&E) by Yale Pathology Tissue Services (YPTS). Tumor burden, quantified as the fraction of total lung area occupied by tumors, was determined in ImageJ. Tumor grade was scored using previously described criteria.^29,53 Briefly, Advanced Adenomatous Hyperplasia (AAH): peripheral focal proliferation of atypical cuboidal epithelial cells, adjacent to normal lung; grade 1: tumors with uniform nuclei showing no nuclear atypia; grade 2: tumors containing cells with uniform but slightly enlarged nuclei that exhibit prominent nucleoli; grade 3: tumor cells with enlarged, pleomorphic nuclei showing prominent nucleoli and nuclear molding; grade 4: tumor cells showing large, pleomorphic nuclei exhibiting a high degree of nuclear atypia, including abnormal mitoses and hyperchromatism, and containing multinucleate giant cells; grade 5: tumors with all the features of grade 4 tumors and also showing stromal desmoplasia surrounding nests of tumor cells. In overall survival studies, mice were euthanized at ethical endpoints.

Immunohistochemistry

Immunostaining was carried out on paraffin-embedded tissue sections using Vectastain Elite ABC Peroxidase kit (PK-6101, Vector Laboratories). Antigen retrieval was carried out by heating in a steamer with 10 mM Citrate buffer (pH 6.0) for 30 min at 95°C. Endogenous peroxidase activity was blocked with 3% hydrogen peroxidase or Dual Endogenous Enzyme Block (S-2003, Dako), followed by Avidin/Biotin block (SP-2001, Vector Laboratories), and CAS-Block (008120, Invitrogen). Tissues were incubated with pHH3 antibody (1:500, Cell Signaling Technology) at 4°C overnight. The signal was visualized with DAB (SK-4100, Vector Labs) and slides were counterstained with hematoxylin. pHH3 staining was evaluated as number of positive cells per tumor area using ImageJ.

RNA-seq

Reads were mapped with the STAR (2.7.11a) alignment tool³⁵ to a custom tandem repeat assembly (satellite DNA, rDNA, telomere) first allowing no mismatches with the arguments “–outFilterMismatchNmax 0 –alignEndsType EndToEnd –alignMatesGapMax 500 –outFilterMultimapNmax 1 –outSAMmultNmax 1 –outMultimapperOrder Random –alignIntronMax 1 –quantMode GeneCounts –clip3pAdapterSeq AAAAAAAAAA AAAAAAAAAA –clip3pAdapterMMp 0.1 0.1”. Reads that did not map perfectly to tandem repeat sequences were then mapped to the mm10 genome with command “–outReadsUnmapped Fastx –outFilterType BySJout –alignSJoverhangMin 8 –alignSJDBoverhangMin 1 –alignIntronMin 20 –alignIntronMax 1000000 –alignMatesGapMax 1000000 –outFilterMismatchNmax 999 –outFilterMismatchNoverReadLmax 0.04 –outFilterMultimapNmax 20 –outSAMmultNmax 1 –outMultimapperOrder Random –winAnchorMultimapNmax 1000 –outSAMattrIHstart 0 –chimSegmentMin 20 –chimOutType SeparateSAMold –quantMode GeneCounts –clip3pAdapterSeq AAAAAAAAAA AAAAAAAAAA –clip3pAdapterMMp 0.1 0.1”. The number of mapped reads per gene in each sample was quantified using HTseq³⁶ and the mouse gencode (vM25)⁵⁴ gene annotation. Differential Gene Expression (DEG) analysis was done with the DESeq2³⁷ package in R. PCA was done using the prcomp command in R with the DESeq2 library normalized gene counts. Genes having an FDR<0.01 were considered as differentially expressed. Gene set enrichment plots were generated using the clusterProfiler³⁸ package and genesets were retrieved from the Molecular Signatures Database³⁹ with the msigdbr package (V2025.1.Hs) in R. Annotated genes from chrX and chrY were excluded from all analysis due to differences in sex between MEF lines. To compare changes in Myc target expression, genes were randomly selected to match the distribution of expression of the Hallmark MYC V1 target gene set and the Kolmogorov-Smirnov test was used to compare distrubutions. Scatter and bar plots were generated with ggplot2.⁴⁰ BigWig files were generated with deeptools (V3.5.2)⁴¹ with command “–normalizeUsing CPM –ignoreDuplicates –binSize 20” and visualized with IGV (Integrative Genomics Viewer).⁴²

ATAC-seq

Cells were preserved by cryopreservation prior to ATAC-seq library preparation using Active. Motif Library preparation kit and sequencing by YCGA. Reads were mapped to tandem repeats using the STAR (2.7.11a) alignment tool³⁵ allowing no mismatches first with the following arguments: “–outFilterMismatchNmax 0 –alignEndsType EndToEnd –alignMatesGapMax 500 –outFilterMultimapNmax 1 –outSAMmultNmax 1 –outMultimapperOrder Random –alignIntronMax 1 –quantMode GeneCounts” and to the mm10 mouse genome using the following commands: “–outFilterMultimapNmax 5000 –outSAMmultNmax 1 –outMultimapperOrder Random –winAnchorMultimapNmax 5000 –alignEndsType EndToEnd –alignMatesGapMax 500 –outFilterMismatchNmax 999 –outFilterMismatchNoverReadLmax 0.04 –alignTranscriptsPerReadNmax 30000 –alignWindowsPerReadNmax 30000 –alignTranscriptsPerWindowNmax 300 –seedSearchStartLmax 30 –seedPerReadNmax 3000 –seedPerWindowNmax 300 –seedNoneLociPerWindow 1000 –alignIntronMax 1”. For each individual sample, MACS2⁴³ was used to call peaks in mouse genome with default parameters. All peaks (signalValue >2 and −log10(FDR) > 2) were merged into a single bed file that was used to retrieve counts from individual libraries with commands from Bedtools.⁴⁴ Significantly different peaks (|log2Fold change| > 0.6 and FDR<0.01) were determined using the DESeq2³⁷ package in R. Read coverage plots were generated with deeptools (V3.5.2)⁴¹ with command “–normalizeUsing CPM –ignoreDuplicates –binSize 20” and visualized in IGV. Butterfly plots were generated with ggplot2.⁴⁰

Hi-C

Cryopreserved cells were submitted to YCGA for in situ chromatin conformation capture. Hi-C library was prepared using an Arima-HiC kit according to the manufacturer’s instructions and paired-end sequencing was performed on Illumina Novaseq 6000. Paired-end reads were processed with the Juicer command tool (V1.6)⁴⁵ to transform raw mapped sequence data into a list of Hi-C contact matrices (.hic). The Juicer command processed reads that included mapping to the mm10 genome using the bwa aligner (V0.7.17)⁴⁶ and command “mem -SP5M^′′ and site annotation using the restriction enzyme file for the Arima cocktail. Potential differential chromosomal compartments were determined using the dcHiC command line tool⁴⁷ at 100kb resolution. The juicer_tools.3.0.0.jar tool was used to run the HICCUPS algorithm for annotating loops with merged biological replicate files (for MAPQ ≥ 30) and arguments “-k KR -r 25000,10000,5000 -f 0.2,0.2,0.2” and the HiCCUPSdiff algorithm with arguments “-r 5000,10000,25000 -f 0.2,0.2,0.2”. The merged replicate files and 2D annotation were visualized with Juicebox software.⁴⁸ For analysis of differential chromosome 15 contacts, the multiHiCcompare⁴⁹ package was used with individual biological replicate files (for MAPQ ≥ 30) with “zero.p = 0.8, A.min = 4” and cyclic loess normalization. Intrachromosomal contacts (logfc_cutoff = 0.5, logcpm_cutoff = 1) and normalized contact counts were plotted with ggplot2 (Wickham 2016). Significantly different contacts (FDR<0.001) were colored for increased (red) or decreased (blue) interactions and visualized on IGV.

QUANTIFICATION AND STATISTICAL ANALYSIS

Data in graphs are represented as mean ± SEM of biological replicates (n ≥ 3). Paired or unpaired t test, as indicated, was performed to determine statistical significance and indicated as follows; ns, not significant, *p < 0.05, **p < 0.01, and ***p < 0.001. Statistical analyses of RNA-seq, ATAC-seq, and Hi-C datasets were performed using established pipelines, detailed in Methods Details. Additional experimental details can be found in Figure legends, Main text, and STAR Methods sections.

Supplementary Material

Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2025.116439.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Brd4	Abcam	Cat#ab128874
pHH3	Cell Signaling	Cat#9701S
Myc, clone Y69	Abcam	Cat#ab32072
Hsp90	Cell Signaling	Cat#4877S
Alexa Fluor 488 AffiniPure Goat Anti-Rabbit IgG (H + L)	Jackson Immunoresearch	Cat#111-545-003
Peroxidase AffiniPure Donkey Anti-Rabbit IgG (H + L)	Jackson Immunoresearch	Cat#711-035-152
FITC-conjugated anti-BrdU antibody, clone B44	BD Biosciences	Cat#347583
Bacterial and virus strains
Adenoviral Cre (High Titer)	University of Iowa, Viral Vector Core Facility	N/A
Stable Competent E.coli (High Efficiency)	New England Biolabs	Cat#C3040H
Chemicals, peptides, and recombinant proteins
2-mercaptoethanol	Gibco	Cat#21985-023
Puromycin dihydrochloride	Millipore Sigma	Cat#P9620
MG132	Millipore Sigma	Cat#M7449
Cycloheximide	Millipore Sigma	Cat#C4859
Hexadimethrine bromide (polybrene)	Millipore Sigma	Cat#107689
Formaldehyde, 16%, methanol-free	ThermoFisher Scientific	Cat#28908
1xPBS, RNAse-free	ThermoFisher Scientific	Cat#J62851AK
Formamide	ThermoFisher Scientific	Cat#17899
Vectashield mounting media with DAPI	Vector Laboratories	Cat#H-1200
BSA, RNAse-free	Jackson Immunoresearch	Cat#001-000-162
Fast SYBR Green PCR Master Mix	ThermoFisher Scientific	Cat#4385614
GoTaq Green PCR Master Mix	Promega	Cat#M7123
BrdU	Millipore Sigma	Cat#B5002
Dual Endogenous Enzyme Block	DAKO	Cat#S-2003
Avidin/Biotin block	Vector Laboratories	Cat#SP-2001
CAS-Block	Invitrogen	Cat#008120
Stellaris® RNA FISH Hybridization Buffer	LGC Biosciences	Cat#SMF-HB1-10
Stellaris® RNA FISH Wash Buffer A	LGC Biosciences	Cat#SMF-WA1-60
Stellaris® RNA FISH Wash Buffer B	LGC Biosciences	Cat#SMF-WB1-20
Formaldehyde solution	Millipore Sigma	Cat#252549
Hematoxylin Solution, Gill No.2	Millipore Sigma	Cat#GHS232
Critical commercial assays
RNeasy Mini Kit	Qiagen	Cat#74106
RNAse-free DNase Set	Qiagen	Cat#79254
TissueRuptor II	Qiagen	Cat#9002755
High Capacity cDNA Reverse Transcription Kit	ThermoFisher Scientific	Cat#4368814
Pierce ECL Prime Western Blotting Substrate	ThermoFisher Scientific	Cat#32209
Vectastain Elite ABC Peroxidase Kit	Vector Laboratories	Cat#PK-6101
DAB Peroxidase (HRP) Substrate Kit	Vector Laboratories	Cat#SK-4100
Deposited data
RNA-seq	This paper	GEO: GSE306107
ATAC-seq	This paper	GEO: GSE306101
Hi-C	This paper	GEO: GSE306102
Experimental models: Cell lines
E13.5 MEFs Pvt1P/P and Pvt1+/+, littermate, primary	This paper	N/A
E13.5 MEFs Pvt1P/P; p53+/+ and Pvt1P/P; p53−/−, littermate, primary	This paper	N/A
PR MEFs, puromycin-sensitive: p53LSL/LSL; Rosa26-CreERT2	Olivero et al.23	N/A
293	ATCC	Cat#CRL-1573
Phoenix ecotrophic	ATCC	Cat#CRL-3214
Experimental models: Organisms/strains
Pvt1 PAS: B6-Pvt1P/P	This paper	N/A
p53 null: B6.129S2-Trp53tm1Tyj/J	Jackson Laboratories	RRID: IMSR_JAX:002101
Kras-LA1: 129-Krastm1Tyj/J	Generously provided by Dr. Jonathan Kurie, MD Anderson Cancer Center	RRID: IMSR_JAX:002674
KP: B6.129-Krastm4Tyj Trp53tm1Brn/J	Jackson Laboratories	RRID:IMSR_JAX:032435
Oligonucleotides
gRNAs sequences, see Table S5	This paper	N/A
Genotyping primers, see Table S5	This paper	N/A
qRT-PCR primers, see Table S5	This paper	N/A
Stellaris FISH Probes, Custom Assay with Quasar® 670 Dye, see Table S5	LGC Biosciences	Cat#SMF-1065-5
Stellaris FISH Probes, Custom Assay with Quasar® 570 Dye, see Table S5	LGC Biosciences	Cat#SMF-1063-5
Recombinant DNA
pCMV-dR8.2 dvpr	Addgene	Cat#8455
pCMV-VSV-G	Addgene	Cat#8454
pLPC-E1A	Dimitrova et al.34	N/A
BRD1-gRNA	Generously provided by Broad Institute	N/A
Software and algorithms
Zen2.6 Pro software	Zeiss	N/A
ImageJ	FIJI	N/A
Prism8	GraphPad	N/A
STAR (2.7.11a)	Dobin et al.35	N/A
HTseq	Anders et al.36	N/A
DESeq2	Love et al.37	N/A
clusterProfiler	Yu et al.38	N/A
Msigdbr (V2025.1.Hs)	Liberzon et al.39	N/A
ggplot2	Wickham et al.40	N/A
deeptools (V3.5.2)	Ramirez et al.41	N/A
IGV (Integrative Genomics Viewer)	Robinson et al.42	N/A
MACS2	Zhang et al.43	N/A
Bedtools	Quinlan et al.44	N/A
Juicer command tool (V1.6)	Durand et al.45	N/A
bwa aligner (V0.7.17)	Li and Durbin46	N/A
dcHiC command line tool	Chakraborty et al.47	N/A
Juicebox software	Robinson et al.48	N/A
multiHiCcompare	Stansfield et al.49	N/A

Highlights.

• Pvt1 controls the dynamics of Myc transcription in a dose-dependent manner

• Inhibition of Pvt1 leads to increased Myc expression and oncogenic activity

• Pvt1 does not regulate the chromatin architecture or accessibility of the Myc locus

• Pvt1 limits transcriptional condensate formation and bursting at the Myc promoter