Introduction

The regulation of gene expression in eukaryotes is complex and includes different levels of regulation including epigenetic, transcriptional, translational, and post-translational control (Willemin et al. 2024; Wang et al. 2023). Enhancers are non-coding regulatory elements that enhance the transcription of associated genes when bound by specific transcription factors and are not necessarily in proximity to the gene body. Rather, they can be found upstream, downstream, or within the coding region and are brought into the vicinity of the promoter through genome folding (Deng et al. 2012). Super-enhancers (SE) are a class of regulatory sequences with enrichment for the binding of activators of transcription factors within the sequence (Pott and Lieb 2015). SEs tend to spread over large DNA regions and are found to be enriched next to genes with known oncogenic functions (Loven et al. 2013; Hnisz et al. 2013; Jia et al. 2020).

SEs can drive the expression of oncogenic long non-coding RNAs (lncRNAs) with tumor-promoting functions. For instance, the lncRNA urothelial cancer-associated 1 (UCA1), which is driven by a SE, is overexpressed in ovarian cancer and leads to tumorigenesis through YAP activation (Lin et al. 2019). LINC01503 lncRNA is located at a SE, and the binding of the transcription factor TP63 to the SE locus activates the expression of LINC01503. This activation leads to squamous cell carcinoma development, and the lncRNA levels are correlated with the shorter survival time of patients (Xie et al. 2018). The lncRNA LIMD1-AS1, which is also activated by a SE, is upregulated in glioma through its activation of CDK7, which contributes to cell proliferation (Chen et al. 2023).

A well-known SE that can express various lncRNAs in different types of cancers is the SE at the MYC locus (Pott and Lieb 2015; Xiang et al. 2014; Iaccarino 2017). The c-MYC (MYC) oncogene is known to be upregulated in 50–60% of all tumors, while its overexpression can be achieved by a variety of mechanisms. The MYC gene is located at the 8q24 locus, an area containing different enhancers that are known to be involved in diverse types of cancers. The enhancers are organized in topologically associating domains (TADs), while the binding of CTCF and cohesion proteins to the TADs contributes to their stability (Pombo and Dillon 2015; Dixon et al. 2016; Dekker et al. 2023). Several studies have identified that the long-range chromatin looping at the MYC locus, which leads to MYC hyperactivation, plays a critical role in cancer progression (Huppi et al. 2012; Lancho and Herranz 2018). A major factor in MYC upregulation that leads to cancer development is the binding of MYC to active binding sites on nearby enhancers, which results in high transcriptional activity (Lin et al. 2012; See et al. 2022). MYC binding to enhancers activates protein-coding genes and lncRNAs, which can indirectly regulate gene expression (Wang et al. 2020b). Those ncRNAs, which are known as super-enhancer-derived ncRNAs, play a critical role in tumorigenesis, metastasis, drug resistance, and more (Peng et al. 2019; Ge et al. 2019; Lee et al. 2020).

Two of the many lncRNAs expressed from the MYC locus are PVT1 and CCAT1. Plasmacytoma variant translocation 1 (PVT1) lncRNA is located on chromosome 8q24.21 and 53 kb downstream of the MYC locus (Parolia et al. 2018). PVT1 is overexpressed in various types of cancer (Liu et al. 2015; Kong et al. 2015; Zhang et al. 2018; Li et al. 2024). Through its interaction with MYC, PVT1 promotes progression, invasion, metastasis, and chemoradiotherapy resistance in different tumors (Shigeyasu et al. 2020; Ansari et al. 2019). Colon cancer-associated transcript 1 (CCAT1), also known as CCAT1-S or as cancer-associated region long non-coding RNA-5 (CARLo-5), was first discovered to be overexpressed in patients with colorectal cancer (CRC), while the highest expression was observed in the small intestine and esophagus (Nissan et al. 2012). CCAT1 maps to chromosome 8q24.2 and contains 2795 nucleotides. The transcript has two exons and an intron that is spliced. CCAT1 has two isoforms: CCAT1-L with an extended second exon, and 5L-CCAT1 with an extended 5′UTR. The dysregulation of CCAT1 expression affects tumorigenesis and clinical manifestations such as tumor size, metastasis, invasion, and patient survival (Wang et al. 2019; Zhan and Xian 2023).

One of the reasons for CCAT1 upregulation is its location at the SE and its interaction with MYC. A DNA loop between the MYC promoter and the enhancer locus 335 kb upstream of MYC (MYC-335) has been demonstrated (Ahmadiyeh et al. 2010). More specifically, the CCAT1-L gene is located 515 kb upstream of MYC, and CCAT1-L transcripts have a functional role in chromatin looping at the MYC locus. DNA fluorescence in situ hybridization (DNA FISH) experiments confirmed that the MYC gene and the CCAT1 gene are co-localized in the tissues of CRC patients (Xiang et al. 2014). Further studies have identified high expression of CCAT1 in breast cancer, lung cancer, osteosarcoma, and other types of cancers (Chen et al. 2016; Lai et al. 2018; Liu et al. 2019, 2023; White et al. 2014; Alaiyan et al. 2013). At present, the exact role of CCAT1 in cancer development is unknown.

The expression of the RNAs from the MYC locus has not been studied on the single-molecule RNA level. We found that the CCAT1 lncRNA is highly detectable in HeLa cells, particularly in comparison to other cell lines that are known to overexpress CCAT1 but have less detectable transcript levels. We, therefore, examined CCAT1 expression patterns in HeLa cells and found that the large observed foci of CCAT1 transcription were formed due to transcript accumulation at the gene locus. In addition, we found that CCAT1 transcripts are post-transcriptionally spliced, and this occurs shortly after their release from the gene. Interestingly, under splicing inhibition conditions, unspliced transcripts were also found in the cytoplasm, suggesting that these transcripts are a variant that the cell can recognize as suitable for export.

Materials and methods

Cell culture

HeLa and RKO cells were maintained in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) (Biological Industries, Beit-Haemek, Israel) containing 10% fetal bovine serum (FBS; HyClone Laboratories, Logan, UT, USA), 100 IU/mL penicillin, and 100 μg/mL streptomycin (Biological Industries). Cells were grown at 37 °C and 5% CO2. HT-29 and HCT116 were maintained in McCoy's 5A Medium (Biological Industries) containing 10% FBS, 100 IU/mL penicillin, and 100 μg/mL streptomycin. For transcription inhibition, cells were grown on coverslips and incubated at 37 °C for 2, 3, and 4 h with either actinomycin D (ActD) (5 µg/mL, Sigma-Aldrich, Rehovot, Israel) or 5,6-dichloro-1-β-d-ribofuranosylbenzimidazole (DRB) (25 µg/mL, Sigma-Aldrich) before fixation for 20 min in 4% paraformaldehyde (PFA). For splicing inhibition, cells were grown on coverslips and incubated at 37 °C for 1 or 6 h with pladienolide B (PLB) (0.5 µM, Santa Cruz Biotechnology, Dallas, TX, USA) before fixation for 20 min in 4% PFA.

Total RNA purification

Total RNA was produced using TRI Reagent (Sigma-Aldrich), and DNA was removed using the TURBO DNA-free kit (Invitrogen) according to the manufacturer's instructions. Synthesis of complementary DNA (cDNA) was performed using the RevertAid™ First Strand cDNA Synthesis Kit (Fermentas/Thermo Fisher Scientific), by taking 1 μg RNA for each sample. Semi-quantitative reverse transcription polymerase chain reaction (RT-PCR) was performed using an Eppendorf thermocycler amplification for 20–35 cycles (depending on the saturation level of the genes amplified) using 1 min denaturation at 94 °C, 1 min annealing at 50 °C, 1 min extension at 72 °C; and 72 °C for 10 min for final extension. The following primers were used:

CCAT1 forward: TCCATCTGGAGCATTCACTG

CCAT1 reverse: AGCCATACAGAGCCAACCTG

c-MYC forward: AATGAAAAGGCCCCCAAGGTAGTTATCC

c-MYC reverse: GTCGTTTCCGCAACAAGTCCTCTTC

PVT1 forward: GCTGTCAAAGAGGCCTGAAG

PVT1 reverse: ACATTTCCTGCTGCCGTTTT

18S forward: TGTGCCGCTAGAGGTGAAATT

18S reverse: TGGCAAATGCTTTCGCTTT

RNA fluorescence in situ hybridization (RNA FISH)

Cells were seeded on 18 mm coverslips and fixed for 20 min in 4% PFA, then washed in 70% ethanol overnight. Coverslips were then washed twice with 10% formamide for probes purchased from Stellaris or fluorescent light-up aptamer (FLAP) probes (Tsanov et al. 2016) diluted in 4× saline-sodium citrate (SSC). For single-molecule FISH (smFISH) on endogenous transcripts, fluorescence-labeled DNA probes targeting the c-MYC exon sequence (570 nm, ~10 ng probe, Stellaris), CCAT1 intron sequence (670 nm, ~10 ng probe, Stellaris), CCAT1-L sequence (570 nm, ~10 ng probe, FLAP), CCAT1-5L (570 nm, ~10 ng probe, FLAP), PVT1 exon sequence (570 nm, ~10 ng probe, FLAP)were hybridized overnight at 37 °C in a dark chamber in 10% formamide. The next day, cells were washed twice with 10% formamide diluted in 4× SSC for 30 min at 37 °C and then washed with 1× phosphate buffered saline (PBS). To reduce photobleaching, the slides were mounted in GLOX (catalase, glucose oxidase) buffer (pH 8.1 mM, 2× SSC, 0.4% glucose) supplemented with 3.7 ng of glucose oxidase (Sigma-Aldrich G2133-10KU) and 1 μL catalase (Sigma-Aldrich 3515) prior to imaging.

RNA FISH and immunofluorescence

For smFISH on endogenous transcripts, fluorescence-labeled DNA probes targeting the CCAT1 exon sequence (570 nm, 10 ng, Stellaris) and the CCAT1 intron sequence (670 nm, 10 ng, Stellaris) were hybridized overnight at 37 °C in a dark chamber in 10% formamide together with the primary antibody, anti-SRRM2 (rabbit, Abcam, ab122719). The next day, cells were washed with 10% formamide, and a secondary antibody (Alexa Fluor 488 goat anti-rabbit IgG; Abcam) was added for 30 min at 37 °C, followed by washing with 1× PBS. To reduce photobleaching, the slides were mounted in GLOX buffer.

Quantitative RNA FISH

Following RNA FISH experiments, three-dimensional (3D) stacks of cells were acquired using a wide-field fluorescence microscope at 60× magnification. Specifically, 51 z-planes were acquired for each cell with 300 nm steps. After the acquisition, the images underwent deconvolution using Huygens software and were transferred to Imaris software (Oxford Instruments, Abingdon, UK) for image processing. In Imaris, the signal of each RNA spot was evaluated using “spot object” and the transcription site was designated using “surface object.” To calculate the number of free CCAT1 transcripts during mitosis vs. interphase, each spot was counted, following the average calculation for each group. To measure the distances of the RNAs from the transcription site, spots in the Cy5 channel were analyzed under “shortest distance from surface.”

To calculate the number of single CCAT1 transcripts at the transcription site, the sum of the fluorescence intensity of transcription sites was measured using the Imaris “surface tracker.” Next, the common value for the fluorescence signal of single transcripts was measured using the Imaris “spot tracker” and defined as a single RNA molecule. Then, the mean intensity of a single transcript was multiplied by the number of pixels covered by the transcription site. This value was extracted from the sum of intensity at the transcription site (TS).

$$\frac{{{\text{TS}}\;{\text{intensity}}}}{{{\text{Single}}\;{\text{transcript}}\;{\text{intensity}}}} = {\text{RNA}}\;{\text{number}}\;{\text{per}}\;{\text{TS}}$$

Fluorescence microscopy

Wide-field fluorescence images were obtained using the cellSens system based on an Olympus IX83 fully motorized inverted microscope (60× UPlanXApo objective, 1.42 NA) fitted with a Prime BSI scientific complementary metal–oxide–semiconductor (sCMOS; Teledyne) driven by CellSens software. Images were then created using the FIJI software package.

Statistical analysis

The experiments presented were carried out at least three times. Statistical analysis was performed using GraphPad Prism 10 software. For quantification of CCAT1 unspliced transcripts in interphase versus mitotic cells, data were analyzed with independent-samples t-tests. For quantification of CCAT1 transcripts at the transcription site, proximity to the transcription site, location of introns in the cytoplasm, percentage of transcription sites per cell, and semi-quantitative RT-PCR data were analyzed using one-way ANOVA, followed by Tukey's post hoc analysis. Treatment groups for which all values were constant (0% cells) were analyzed separately from other treatments using one-sample t-tests against a constant mean value of zero. Finally, a false discovery rate (FDR) correction was applied to adjust for multiple testing.

Results

CCAT1 lncRNA is highly abundant in the nucleus of HeLa cells

The MYC 8q24 locus has been demonstrated to express cancer-specific lncRNAs, including CCAT1 and PVT1 (Xiang et al. 2014; Jin et al. 2019). CCAT1 is located 515 kb upstream of the MYC locus, and PVT1 is 53 kb downstream of MYC (Jin et al. 2019). We examined whether the three genes could be transcriptionally active at the same time by detecting their active transcription sites using RNA FISH probes designed to hybridize with their cognate endogenous RNAs. The expression of these genes was tested in four human cancer cell lines. We tested various colorectal cancer (CRC) cell lines, as it is known that CCAT1 is expressed in CRC: HT-29 colorectal adenocarcinoma cells, RKO colon carcinoma cells, and HCT116 colorectal carcinoma cells (McCleland et al. 2016; Yang et al. 2019; Kam et al. 2014). We also decided to include a different type of cancer cell type, HeLa cervical cancer adenocarcinoma cells (Chen et al. 2021; Li et al. 2023). As expected, the active transcription sites of MYC and CCAT1 were found in all colorectal cell lines and were in close proximity (Fig. 1a). Surprisingly, HeLa cells had very large CCAT1 transcription foci relative to the other cell lines. MYC mRNAs were also abundant. Testing PVT1 lncRNA expression and CCAT1 yielded similar results, namely that PVT1 was expressed in all cells, and the highest signal was observed in HeLa cells (Fig. 1b). Altogether, we found that the MYC locus was transcriptionally active and that CCAT1 expression seemed to be the highest of the three genes expressed from this locus. Notably for HeLa cells, the cells typically exhibited 3–4 active transcription sites expressing these genes. Indeed, chromosome 8 usually appears in three copies in HeLa cells, while chromosome translocation is also common (Landry et al. 2013). When we tested the expression of all three genes using three different probes, we found that CCAT1, c-Myc, and PVT1 active genes were co-localized (Fig. 1c). Although CCAT1 lncRNA showed the strongest signal at the microscopy level, we wanted to determine whether the expression levels of CCAT1 were indeed relatively high. Semi-quantitative RT-PCR showed that MYC expression levels in HeLa cells were much higher than those of CCAT1 (Fig. 1d). This finding implies that the strong CCAT1 signals observed in the microscopy images at the gene locus do not mean that CCAT1 is expressed at remarkably high expression levels. Rather, the lncRNA might be accumulating at the active gene locus, as it is known that many lncRNAs associate with chromatin and with their own gene loci (Calandrelli et al. 2023).

Fig. 1
figure 1

CCAT1 lncRNA is highly abundant in nuclei of HeLa cells. Detection of a CCAT1 lncRNA (exon; green) with MYC mRNA (orange) or with b PVT1 lncRNA (orange) by RNA FISH in HT-29, RKO, HCT116, and HeLa cells. Large foci are the active genes and small dots are the single RNAs. Hoechst DNA stain is in blue. Boxed areas are enlarged. c RNA FISH in HeLa cells of MYC(orange), CCAT1 (exon; green), and PVT1 (purple) RNAs. Scale bars, 10 µm. d Expression levels of CCAT1, MYC and PVT1 RNAs in HeLa cells measured by semi-quantitative RT-PCR. The 18S gene was used as a housekeeping gene. Data were analyzed using one-way ANOVA, followed by Tukey's post hoc analysis. A significant difference was found in the relative expression levels between MYC to CCAT1 and PVT1. **P < 0.01, ***P < 0.001

Full size image

CCAT1 accumulates at the active gene locus

To test the hypothesis that the large size of the CCAT1 site of transcription is due to lncRNA accumulation at the locus, we examined how transcription inhibition affects the size of these transcription sites. Namely, if the enlarged size of the transcription site is due to transcript accumulation, then after treatment with a transcription inhibitor, the signal should not disappear, as would be expected from a normal active gene under transcription inhibition conditions (Brody et al. 2011; Darzacq et al. 2007). First, we used the transcription inhibitor 5,6-dichloro-1-beta-D-ribofuranosylbenzimidazole (DRB), which inhibits the CDK9 kinase (kinase subunit of positive transcription elongation factor) (Bensaude 2011) to inhibit RNA polymerase II (Pol II) transcription. Even after 4 h of treatment with DRB, CCAT1 transcription sites were still observed (Fig. 2a). However, the percentage of visible transcription sites per cell was dramatically decreased after DRB inhibition, and ~35% of the cells did not show any visible active genes (Fig. 2b). To confirm that the DRB treatment was effective, immunofluorescence against the SRRM2 protein was performed in parallel. As expected (Rino et al. 2007; Spector and Lamond 2011), nuclear speckles that were marked by SRRM2 and that usually have an irregular shape were transformed into rounded structures due to transcription inhibition. Many genes are known to associate with nuclear speckles when transcriptionally active (Faber et al. 2022; Kim et al. 2020; Belmont 2021). However, no association between the nuclear speckles and CCAT1 transcription sites was detected under normal or DRB-treated conditions. Similarly, when the transcription inhibitor actinomycin D (ActD) was applied, a similar but stronger effect was observed, as no visible CCAT1 transcription sites were observed in ~60% of the cells (Fig. 2c,d). The persistence of the CCAT1 transcription sites after transcription inhibition, rather than completely disappearing as would be expected, indicated that lncRNA accumulation was occurring at the site of transcription. Since introns are prominently detected at the site of transcription (Cote et al. 2024; Darzacq et al. 2007; Mor et al. 2010), we examined whether the intron and exon sequences of CCAT1 overlapped at these foci. Indeed, using RNA FISH probe sets against the CCAT1 exons or intron, we found that the intron sequences were most prominent at these foci. The latter were reduced in size when cells were treated with the transcription inhibitors (Supplementary Figure S1a, b), implying that these foci are most likely the sites at which CCAT1 is transcribed. This means that under regular conditions, the active CCAT1 genes produce CCAT1 transcripts, and a subpopulation of these transcripts remains associated with the gene locus.

Fig. 2
figure 2

CCAT1 lncRNAs accumulate on the gene locus. The transcription inhibitors a DRB or c ActD decreased CCAT1 detection at the site of transcription, but small foci were still observed (4 h of treatment). CCAT1 (exon; gray) was detected together with anti-SRRM2 (magenta) that marks nuclear speckles. Hoechst DNA stain is in blue. Boxed areas are enlarged. Scale bars, 10 µm. b The percentage of CCAT1 transcription foci per cell after treatment with DRB or d ActD. A minimum of n = 50 cells were selected for each analysis. Data were analyzed using one-way ANOVA, followed by Tukey's post hoc analysis. *P < 0.05, **P < 0.01, ****P < 0.0001

Full size image

Since we speculated that CCAT1 transcripts were accumulating at the gene locus, we assumed that during cell division, these transcripts would dissociate from the chromosome. Also, active transcription sites were not expected to be seen since the rates of transcription decrease dramatically during cell division (Hartl et al. 1993; Palozola et al. 2017). Indeed, no CCAT1 transcription sites were observed during cell division (Supplementary Figure S2), and no accumulation of CCAT1 transcripts was observed on metaphase chromosomes (Fig. 3a). Rather, many CCAT1 transcripts were now seen throughout the mitotic cell. The number of free CCAT1 transcripts in cells during metaphase was counted and compared with the number of free transcripts in cells during interphase. A large increase in free CCAT1 transcripts in the cell was observed (Fig. 3a, b). This indicates that during interphase, CCAT1 transcripts accumulate at the locus, and during mitosis, they are released from the chromosomes. We could now quantify the number of CCAT1 transcripts that were associated with the locus in cells during interphase. First, the numbers of CCAT1 loci were counted (n = 300 cells), which showed that ~50% of the cells contained at least three active sites of transcription (Fig. 3c). Next, the estimated number of CCAT1 transcripts associated with a single transcription site was calculated using the single CCAT1 transcript average intensity (Fig. 3d). Since the CCAT1 loci were not uniform in size, they were divided into several different subgroups. Large transcription sites contained ~40 transcripts, moderate-sized had ~10 transcripts, and the small ones contained less than 5 transcripts. Therefore, a cell that had three large loci would amount to a total of ~120 CCAT1 transcripts associated with the gene loci. This number is within the range of the average number of CCAT1 transcripts measured in the metaphase cells (Fig. 3b; ~150 transcripts). Altogether, these numbers correlate well, as there are also free CCAT1 transcripts in the nucleoplasm and cytoplasm of interphase cells.

Fig. 3
figure 3

CCAT1 transcripts are released from chromosomes during cell division. a Detection of CCAT1 RNAs by RNA FISH in cells at interphase (left) and metaphase (right). The top panels show original, with zoomed images presented at high signal intensity (boxes). The bottom panel shows analyzed images (Imaris). White spots show the CCAT1 RNA. Large pink dots are sites of transcription. Scale bar, 3 μm. b The total number of CCAT1 RNAs counted in interphase versus metaphase cells. Data were analyzed with independent-samples t-tests. c The distribution of CCAT1 transcription sites per cell under steady-state conditions. d Transcription foci were divided into subgroups according to their size. Each group contained different numbers of transcripts. A minimum of n = 50 cells were selected for each statistical analysis. ****P < 0.0001

Full size image

CCAT1 transcripts undergo post-transcriptional splicing

We next wanted to examine whether the transcripts associated with the CCAT1 gene loci were spliced transcripts in all cell lines expressing CCAT1. Using the RNA FISH probe sets against the CCAT1 exons or intron, the intron signal overlapped with the site of transcription in HT-29, RKO, HCT116, and HeLa cells (Fig. 4a). Typically, intron signals are seen predominantly on sites of transcription, since much of the splicing occurs co-transcriptionally. Here, all transcription site signals contained intron signals, suggesting that the associated CCAT1 lncRNAs were unspliced at this stage. However, unspliced transcripts were also detected in the nucleoplasm of the HeLa cells (Fig. 4a, Supplementary Figure S1a, b). In general, the vast majority of the CCAT1 transcripts in the cells were spliced RNAs, and the unspliced transcripts were mostly found near the site of transcription. Unspliced transcripts were nuclear only. This was suggestive of post-transcriptional splicing occurring after release from the transcription site. Since unspliced RNAs can accumulate in nuclear speckles (Gordon et al. 2021; Mor et al. 2016), we examined whether there was an association of unspliced CCAT1 transcripts with nuclear speckles. RNA FISH applied to CCAT1 unspliced transcripts and staining for SRRM2, a core nuclear speckle protein, showed no association between the two (Fig. 4b).

Fig. 4
figure 4

CCAT1 unspliced transcripts are present in the nucleoplasm of HeLa cells and do not localize with nuclear speckles. a RNA FISH detection of CCAT1 exon (pink) and CCAT1 intron regions (green). HeLa cells displayed high levels of unspliced CCAT1 transcripts at the sites of transcription and low levels of unspliced transcripts in the nucleoplasm. HT-29, HCT116, and RKO cells showed low levels of CCAT1 introns (unspliced transcripts) at the site of transcription. b CCAT1 active genes and nucleoplasmic transcripts were not associated with nuclear speckles marked by anti-SRRM2 (cyan). Hoechst DNA stain is in blue. Boxed areas are enlarged. Scale bars, 10 µm

Full size image

Next, we focused on the location of the CCAT1 unspliced transcripts in the nucleus. Spliced CCAT1 transcripts were dispersed over the entire nucleus, while unspliced transcripts were localized mostly next to the site of transcription (Fig. 5a). Next, we measured the distance of CCAT1 unspliced transcripts from the gene foci (Fig. 5b). When calculating the distance of unspliced CCAT1 transcripts from the gene locus, we found that most transcripts were located less than 3 μm from the transcription site, and less than 16% of the unspliced transcripts were located more than 6 μm from the locus (Fig. 5c). In addition, to test whether the other CCAT1 isoforms CCAT1-L and 5L-CCAT1 were also post-transcriptionally spliced, RNA FISH was performed using probes that detect either CCAT1-L or CCAT1-5L only, together with probes directed to the intron. We observed that CCAT1-L and CCAT-5L expression overlapped with the intron signal in the nucleoplasm, and as observed with the CCAT1 common isoform, the unspliced transcripts were localized close to the transcription site (Supplementary Figure S3). The presence of CCAT1-L and CCAT1-5L unspliced transcripts in proximity to the gene loci confirms that the post-transcriptional splicing is common to the three different isoforms. Taken together, these findings suggest that the splicing of CCAT1 is post-transcriptional and occurs soon after release from the gene and before the transcripts diffuse away from the gene. This finding is consistent with the recently published evidence demonstrating that highly transcribed genes are post-transcriptionally spliced (Cote et al. 2024).

Fig. 5
figure 5

Unspliced CCAT1 transcripts are found in close proximity to the active gene. a CCAT1 spliced and unspliced RNA was detected in untreated cells and analyzed by Imaris; CCAT1 exon spots (cyan) and CCAT1 intron spots (red). Large pink dots are sites of transcription. b Distances of CCAT1 unspliced RNAs from the active genes in untreated cells are color-coded. Pink spots show the unspliced transcripts located less than 3 μm from the active gene; green spots between 3 to 6 μm; and white spots more than 6 μm. Large pink dots are sites of transcription. c Measurements of the distances of CCAT1 unspliced transcripts from the active genes. Data were analyzed with a one-way ANOVA, followed by Tukey's post hoc analysis. **P < 0.01, ****P < 0.0001. A minimum of n = 50 cells were selected for statistical analysis. Scale bar, 3 μm. d Schematic illustration of the measured distances of unspliced CCAT1 transcripts from the active genes

Full size image

Unspliced CCAT1 transcripts are exported to the cytoplasm during splicing inhibition

It was previously demonstrated that splicing inhibition can lead to the accumulation of unspliced transcripts at the site of transcription and in nuclear speckles (Cote et al. 2024). These transcripts are retained in the nucleus and are not exported to the cytoplasm. We did not find any accumulation of CCAT1 unspliced transcripts in nuclear speckles under regular conditions (Fig. 4b), and so we examined their fate under conditions of splicing inhibition. We used pladienolide B (PLB), a splicing inhibitor that binds to the SF3B1 subunit of the U2 small nuclear ribonucleoproteins (snRNP) and blocks spliceosome activity (Effenberger et al. 2017). When PLB was added to HeLa cells, the presence of CCAT1 at the gene locus was reduced already after 1 h (Fig. 6a). After 6 h of treatment with PLB, CCAT1 transcription sites not only did not show any visible accumulation of unspliced transcripts, but in a considerable percentage of the cells the large loci disappeared (Fig. 6b). Surprisingly, in some cells, unspliced transcripts appeared in the cytoplasm (Fig. 6a, c). As a control, we examined another transcript, the MKI transcript that encodes Ki-67, under splicing inhibition conditions, and unspliced MKI mRNAs were not detected in the cytoplasm (Supplementary Figure S4). Additionally, unspliced CCAT1 transcripts did not appear to be associated with nuclear speckles (Fig. 6a). Quantification of the unspliced transcripts showed that ~40% of the introns were found in the cytoplasm during splicing inhibition, while the rest were in the nucleus. We postulated that following splicing inhibition, CCAT1 transcripts, which normally are highly transcribed, undergo nuclear export despite intron retention. Taken together, these findings might suggest that CCAT1 at steady-state conditions is post-transcriptionally spliced and that the unspliced version is not detected as an aberrant RNA with a splicing defect but might be a normal candidate for export under certain conditions.

Fig. 6
figure 6

Splicing inhibition reduces CCAT1 expression and leads to the appearance of unspliced transcripts in the cytoplasm. a Pladienolide B (PLB) treatment for 1 and 6 h decreased CCAT1 detection at the active genes. Unspliced transcripts were detected at the cytoplasm after 6 h. CCAT1 exon (pink), CCAT1 intron (green), anti-SRRM2 for marking nuclear speckles (cyan). b The percentage of CCAT1 gene foci per cell after treatment with PLB for 6 h. c RNA FISH of CCAT1 exon and intron regions in untreated and splicing-inhibited conditions. CCAT1 exon (red), CCAT1 intron (green). The top panel represents original and zoomed images at higher intensity (boxes). The middle panel shows analyzed images (Imaris). Red spots show CCAT1 exon and green spots show CCAT1 intron. The bottom row presents a schematic illustration. Scale bar, 4 μm. d The percentage of unspliced transcripts in the cytoplasm compared with nuclei in control and under splicing inhibition conditions. Data were analyzed with b independent-sample t-tests and d one-way ANOVA, followed by Tukey's post hoc analysis. *P < 0.05, **P < 0.01, ****P < 0.0001. A minimum of n = 50 cells were selected for statistical analysis

Full size image

Discussion

Super-enhancers are large clusters of enhancers spread over large DNA regions that regulate gene expression via a variety of mechanisms. Similar to typical enhancers, SEs can be located far away from their associated gene, and in such cases, genome folding brings the enhancer into close proximity to the gene promoter, where their interactions are mediated through transcription activators and the transcriptional initiation complex, to control gene expression (Deng et al. 2012; Ye et al. 2020). SEs display enrichment of binding sites for transcription factors and activators relative to a typical enhancer (Whyte et al. 2013; Pott and Lieb 2015). SEs in cancer cells have been shown to display different expression patterns and enhancer usage relative to healthy cells (Loven et al. 2013). Therefore, SEs are considered key regulators of oncogene expression in different tumors (Tang et al. 2020).

The MYC locus is a tumor type-specific super-enhancer that expresses different super-enhancer-derived ncRNAs (Hnisz et al. 2013; Amjadi-Moheb et al. 2021). Within the course of our research, we focused on the expression of the CCAT1 lncRNA at the MYC locus. CCAT1 is upregulated in CRC already in the early phase of tumorigenesis as well as at late stages of the disease, and hence, CCAT1 can be used as a potential biomarker for screening, diagnosis, and prognosis for patients with CRC (Mizrahi et al. 2015; Xiao et al. 2021). In the current study, we examined lncRNA localization in human cell lines on the single RNA level and characterized CCAT1 expression, with an emphasis on its detection at the site of transcription in HeLa cells, which showed very prominent sites of CCAT1 transcription. We were surprised to observe this significant upregulation of CCAT1 in HeLa cells, since CCAT1 is mostly correlated with colon cancer (Mizrahi et al. 2015; Shang et al. 2020). However, CCAT1 is also expressed in ovarian cancer (Wang et al. 2020a). We found that CCAT1 was highly detectable relative to other cell lines that express CCAT1; several large sites of transcription were observed and high numbers of transcripts were detected in the nucleus and cytoplasm (Fig. 1a, b). We assumed that the large foci were transcription sites since active transcription sites contain the highest number of transcripts when examined by RNA FISH, and the overlapping with the MYC and PVT1 transcription sites strengthened this assumption. The relative RNA expression levels of MYC in HeLa cells were twofold higher than CCAT1 levels, while the PVT1 levels were similar to CCAT1 (Fig. 1d).

To verify that the foci were indeed sites of transcription, we used transcription inhibitors. Following treatment with DRB or ActD, the CCAT1 foci were still observed to a certain extent (Fig. 2a, c), although the number of cells exhibiting these foci decreased significantly. Also, the intron and exon signals predominantly co-localized at these foci, as would be expected from transcription sites that are the areas that contain most intronic sequences. The fact that the foci did not completely disappear, as would be expected of transcription sites during transcription inhibition, suggested that in addition to transcription, there is an accumulation of the lncRNAs on their gene locus. The accumulation of lncRNAs on chromatin is known (Guo et al. 2020; Schlackow et al. 2017; Calandrelli et al. 2023). The treatment with ActD had a stronger effect on transcription site disappearance (Fig. 2b, d). This effect may be due to the ActD inhibition mechanism, which functions through intercalation into GC-rich sequences and prevents RNA polymerase progression. The intercalation of ActD into the double helix may interrupt the binding of CCAT1 to chromatin, which subsequently reduces foci size. Therefore, we speculate that the large transcription sites observed are formed due to CCAT1 transcript accumulation on the gene locus and not due to an unusually high transcription rate modulated by RNA Pol II activity. These results correlate with previous studies showing that lncRNAs can modulate gene expression locally by accumulation at or near their site of transcription (Kopp and Mendell 2018; Gil and Ulitsky 2020). lncRNA accumulation on chromatin can be associated with its regulation or function. One of the earliest examples of RNA accumulation is the lncRNA XIST (Brockdorff et al. 1992; Brown et al. 1991), a key regulator of X inactivation, which mediates the silencing of the inactive X chromosome through the subsequent recruitment of epigenetic regulators (Clemson et al. 1996; Lee 2012). Another example is polyadenylated nuclear RNA (PAN) lncRNA, which is a viral RNA that accumulates and binds to its own promoter and robustly activates PAN lncRNA expression (Campbell and Izumiya 2020). Multiple studies have indicated that RNAs with processing defects, particularly in their splicing, accumulate in nuclear foci located near their site of transcription. For example, the mutant β-globin mRNA that has a defect in either splicing or 3′ formation has been reported to accumulate close to its site of transcription, even in the presence of transcription inhibitors, which under normal conditions cause the rapid release of wild-type human β-globin RNA from the vicinity of the gene (Custodio et al. 1999).

Co-transcriptional splicing of nascent RNA is a central mechanism for gene regulation in normal cells, while in some cancer cells, defective splicing machinery can shift splicing to occur post-transcriptionally (Bentley 2014). For example, a single missense mutation (Ser34Phe) in the zinc finger domain of the conserved splicing factor U2AF1 is common in different types of cancers, leading to the post-transcriptional splicing of β-globin and FXR1 mRNAs (Coulon et al. 2014). A major question in the splicing field is how tightly associated are the processes of transcription and splicing. The balance between co-transcriptional and post-transcriptional splicing is regulated. Deep sequencing studies have shown that many RNAs undergo co-transcriptional splicing (Tilgner et al. 2012; Sanchez-Escabias et al. 2022). lncRNAs, however, have been reported to be less efficiently spliced than protein-coding genes, regardless of transcript activity (Mukherjee et al. 2017). One possible explanation for this phenomenon is the lack of splicing enhancer sequences at lncRNA splice sites (Krchnakova et al. 2019). Typically, splice sites are abundant along the transcribed region, while the recognition of those sites is mediated by different serine- and arginine-rich (SR) proteins and hnRNPs (De Conti et al. 2013). In HeLa cells, splicing was found to be less efficient and could occur post-transcriptionally, and in some cases the lncRNAs remain unspliced (Tilgner et al. 2012; Schlackow et al. 2017). Also, intron excision in lncRNAs is slower than protein-coding genes, and in contrast, their exon skipping is higher (Mukherjee et al. 2017). This phenomenon can be partially explained by the existence of longer lncRNA introns relative to protein-coding genes (Krchnakova et al. 2019), or also by the finding that there is less RNA Pol II pausing at transcription start sites (TES) of lncRNAs versus protein-coding genes (Schlackow et al. 2017).

A recently published study has demonstrated that highly expressed genes can undergo post-transcriptional splicing close to the transcription site zone in addition to co-transcriptional splicing which occurs during transcription (Cote et al. 2024). This study revealed that the mobility of the RNA is slower at the transcription site proximal zone than within the nucleoplasm, suggesting that the slow-moving zone is where transcripts are finally spliced. Using RNA FISH applied to the CCAT1 exon and intron sequences revealed CCAT1 unspliced transcripts that were located in the nucleoplasm of the cells. This was observed only in HeLa cells. The existence of unspliced transcripts was not specific to the CCAT1 short isoform but was found for all three isoforms (Supplementary Figure S3a, b). We measured the distance that CCAT1 unspliced transcripts traveled from the transcription site and found that most of the unspliced transcripts were localized near the site of transcription, suggesting that this transcript undergoes post-transcriptional splicing not long after release from the gene. As CCAT1 is highly transcribed, a high number of transcripts should undergo splicing, and perhaps the delay in the splicing events led to the abundance of transcripts at the gene locus and the release of unspliced transcripts from the gene, and therefore to splicing in the nucleoplasm. This assumption is consistent with the data showing that lncRNA exhibits lower splicing rates than coding genes (Mukherjee et al. 2017). It is also possible that the accumulation of many unspliced CCAT1 transcripts on the gene locus form a structure that sequesters the RNAs from the splicing machinery. Only after release into the nucleoplasm can the spliceosome function on these transcripts. This scenario agrees with the observation of unspliced CCAT1 transcripts only in the region close to the gene. As mentioned, lncRNAs are less efficiently spliced since their exons contain fewer putative binding sites for SR proteins, and hence cannot generate the cooperative network of positive signals that is needed for recruiting the spliceosome to splice sites. Therefore, their splicing would be more dependent on how optimal their splice sites were (Krchnakova et al. 2019). However, the removal of introns from enhancer lncRNAs did not change their enhancing activity, meaning that introns are not essential for the activating function of lncRNAs. We find that the majority of CCAT1 transcripts are co-transcriptionally spliced, while the minority are post-transcriptionally spliced. This can result from inefficient splicing at the site of transcription, caused by weak splicing factor interactions with the CCAT1 transcripts.

When the cells were treated with splicing inhibitors, there was a significant decrease in CCAT1 accumulation at the transcription site, and surprisingly, unspliced transcripts were observed in the cytoplasm of some cells. This is a rare occurrence, since introns are non-coding sequences that do not reach the cytoplasm. Moreover, when splicing is inhibited, the unspliced transcripts are usually retained at the site of transcription or accumulate in nuclear speckles (Hasenson et al. 2022; Barutcu et al. 2022; Hall et al. 2006; Johnson et al. 2000; Mor et al. 2016). These transcripts do not exit the nucleus since they are detected as aberrant transcripts. The fact that unspliced versions of CCAT1 could be found in the nucleoplasm under normal conditions, and even in the cytoplasm under conditions of splicing inhibition, suggests that the unspliced versions are not defective but are recognized as legitimate variants and might serve roles that await discovery.

Although we do not know to what extent CCAT1-regulated looping engages in crosstalk with other aspects of MYC regulation, this study represents yet another component of the complicated MYC region. Our results indicate that the unusual CCAT1 transcription site size at the gene locus is due to the accumulation of CCAT1 transcripts, and not to exceptionally high transcription levels. This may suggest that a function of CCAT1 transcripts on the gene locus is in the maintenance of the long-range chromatin interactions at the MYC locus. This shows that the accumulation of CCAT1 at its site of transcription is poised to act as a modulator of gene expression in a locus-specific manner. Finally, as this region also expresses distinct lncRNAs in other types of human cancers, it will be of interest to learn whether other 8q24 lncRNAs behave similarly to CCAT1.