Premature cleavage and polyadenylation in the minor intron of PTEN modulate its expression and generate a functional long non-coding RNA Authors: Mariam Elesnawy, Rim Elghandour, Hana Hasna, Aisha Fakhroo, Boshra Al-Sulaiti, Ihab Younis Published: 2026 ABSTRACT SummaryThe tumor suppressor, PTEN, contains a minor intron (mi-INTs) at the start of the gene. mi-INTs regulate the expression of their host genes because their inefficient splicing can be rate-limiting. Whether the PTEN mi-INTs regulate PTEN expression is unclear. PTEN levels are tightly controlled in cancer cells, especially breast cancer, since small reductions promote tumorigenesis. Cancer cells, therefore, employ numerous mechanisms to reduce PTEN. Here, we describe a previously unexplored mechanism that modulates PTEN through its mi-INT. In breast cancer cells, this intron shows low splicing efficiency, causing about half of the PTEN pre-mRNA to retain it and fail to produce functional protein. Unlike many mi-INTs that act as reversible molecular switches, this intron is irreversibly processed by premature cleavage and polyadenylation to generate a long noncoding RNA, PINC. Exogenous PINC alters endogenous PTEN protein levels and cellular proliferation, even in PTEN-mutant cells, indicating PTEN- dependent and independent functions.Subject areas: Molecular biology, Cell biology Introduction Most genes in eukaryotes consist of exons that contain the information for protein expression, which are interrupted by relatively larger intervening sequences called introns. Since both exons and introns are transcribed into pre-mRNAs, it is necessary that the introns are spliced out and the exons joined together to form mature mRNAs.1 Alternative splicing is a physiological process that generates different isoforms from the same pre-mRNA, leading to the production of different versions of the encoded protein of any given gene.2,3,4,5,6,7,8,9 Conversely, the disruption of normal splicing has been shown to lead to a large array of diseases, including cancer.10,11,12 Aberrant splicing in cancer does not only affect tumor suppressor genes (TSGs) and oncogenes, but splicing deregulation is linked to all hallmarks of cancer, driving tumor progression and metastasis.10,13,14,15,16 While most introns are spliced out by the canonical spliceosome, around 0.4% of human introns (∼800 out of >200,000 introns) utilize a specialized spliceosome and are termed U12 introns or minor introns (mi-INTs) due to their low abundance.17 mi-INTs are unique, as they are highly conserved and are non-randomly embedded in genes that play essential cellular processes.18 They act as molecular switches, a post-transcriptional mechanism to promptly express hundreds of genes when needed.19 The deregulation of mi-INTs splicing could be detrimental to health, resulting in a wide range of ailments.17,18,19,20 Interestingly, the first of eight introns of the widely studied TSG, phosphatase and tensin homolog, PTEN, is a mi-INT that we have previously shown to be regulated in response to cellular stress.19 The PTEN protein functions in cell cycle regulation, metastasis, and apoptosis by downregulating the PI3K/AKT pathway.21,22,23,24 The PTEN gene is a highly mutated locus in sporadic human tumors.25 Reduced expression of PTEN has been linked to several types of cancers, with the mammary gland being the most sensitive.26,27 Given its importance in regulating cell growth and apoptosis, PTEN levels are highly regulated in cells, and even a subtle impairment can lead to tumorigenesis.28 In fact, PTEN exemplifies the continuum model of TSGs.28,29,30 This model suggests that complete loss of PTEN protein is not always required for tumor progression, and a reduction as low as 20% can lead to tumor formation in lymph nodes and mammary glands,28 whereas 50% or 80% of PTEN protein is not enough to suppress tumors in mice that are hemizygous or hypomorphic, respectively.31,32,33,34 Thus, cancer cells have developed several mechanisms to reduce PTEN levels either slightly or drastically. These include genetic alterations (e.g., point mutations, truncating mutations, and deletions), transcriptional inhibition (e.g., by modulating levels of transcription factors), epigenetic silencing, post-transcriptional repression (e.g., by inhibitory miRNAs), and post-translational modifications (e.g., phosphorylation, ubiquitylation, acetylation, and oxidation), which alter protein level, function, localization, and interactions.22,25,35,36,37,38,39,40,41,42,43 In certain cancers, mutations that cause partial or complete loss of PTEN account for 17–46% of cases,44,45,46 however, in other cancers, reduced PTEN levels are not readily explained by the above mechanisms, suggesting an unexplored process for inactivating this important tumor suppressor. Given that (1) subtle reduction in PTEN can lead to tumor formation, (2) aberrant splicing can lead to protein loss associated with tumorigenesis,11 and (3) miINTs are known to be bottlenecks for the expression of their host genes,19 we hypothesize that the regulation of PTEN pre-mRNA splicing, specifically its mi-INT, presents a novel mechanism by which tumor cells alter PTEN expression. In this study, we used breast cancer cell lines to show that PTEN expression is indeed dependent on its mi-INT splicing. Interestingly, we find that the low efficiency of PTEN mi-INT splicing and the consequential retention of the mi-INT led to premature cleavage and polyadenylation (PCPA) in this intron, producing a previously unidentified PTEN lncRNA that functions both dependently and independently of the PTEN protein in regulating cellular proliferation. Results PTEN minor intron splicing is altered in breast cancer cells As a key tumor suppressor, it is expected that PTEN levels and/or function would be reduced in breast cancer cells. However, the known mechanisms for down-regulating PTEN, such as genetic alterations, transcriptional inhibition, epigenetic silencing, and post-transcriptional and post-translational modifications25 do not account for all breast cancer cases, suggesting an unknown mechanism for reducing PTEN in breast cancer cells. Given that the PTEN gene harbors a mi-INT, which could be a rate-limiting factor for mRNA production, we tested whether PTEN’s mi-INT splicing efficiency is reduced in breast cancer and whether this would affect PTEN protein expression. We hypothesized that two isoforms could be formed from PTEN pre-mRNA, where isoform 1 is fully spliced and does not include intron 1, hence can produce functional PTEN protein, while isoform 2 has mi-INT retention, which leads to no PTEN protein being formed (Figure 1A). As the PTEN mi-INT is the first intron, it is expected that intron retention would severely affect the open reading frame, leading to the reduced expression of the encoded protein. To test this hypothesis, we used RT-PCR from 3 different breast cancer cell lines (MDA-MB-231, MDA-MB-468, and MCF-7) (Figure 1B). RT-PCR primers (Table 1) were used to detect the exon1-exon2 junction (spliced PTEN) and exon1-intron1 junction (unspliced PTEN). Because of the limited range of detection of PCR, we infer intron retention by an exon-intron signal at the 5′ or 3′ end of PTEN’s mi-INT. This confirmed that intron 1 was retained at a high rate, as indicated by the bands for unspliced PTEN. This was quantified using RT-qPCR, which shows that all three cell lines have more unspliced than spliced PTEN (Figure 1C). As a functional PTEN protein can only be translated from the spliced mRNA, the high levels of unspliced PTEN RNA indicate that a large fraction of the transcription output from the PTEN gene is not translated into a functional protein. Thus, the level of PTEN protein in these cells is below what would be predicted from a transcriptionally active gene. Finally, western blot analysis showed that PTEN protein is indeed expressed from two of the cell lines, supposedly from the fully spliced mRNA (Figure 1D). Of note, MDA-MB-468 cells show no expression of PTEN protein due to a known nonsense mutation.47,48 PTEN minor intron splicing is altered in breast cancer cells (A) Schematic of PTEN minor intron splicing showing the two isoforms PTEN pre-mRNA can form to result in functional PTEN protein or no PTEN protein with minor intron retention. (B) Representative (N = 3) RT-PCR for spliced and unspliced PTEN on MDA-MB-231, MDA-MB-468, and MCF7 cell lines. GAPDH was used as a loading control. (D) Representative qPCR for spliced and unspliced PTEN on MDA-MB-231, MDA-MB-468, and MCF7 cell lines. Data show the relative RNA expression after normalization to the loading control, Actin. Error bars represent standard deviations from three experiments. (D) Representative (N = 3) western blot showing PTEN and GAPDH (loading control) protein expression in MDA-MB-231, MDA-MB-468, and MCF7 cell lines. Table 1Primer and AMO sequencesPrimerSequenceGAPDH (PCR)TGGAAGGACTCATGACCACAGT (F)AGGTCCACCACTGACACGTT (R)Spliced PTEN (PCR)CCAGACATGACAGCCATCATC (F)CCTGTATACGCCTTCAAGTC (R)Unspliced PTEN (PCR)CCAGACATGACAGCCATCATC (F)ATGACCTAGCAACCTGACCA (R)PTEN Int1-Ex2 (PCR)CTGTCCATGTGGAAGTTACC (F)CCTGTATACGCCTTCAAGTC (R)Beta actin (qPCR)GGGAAATCGTGCGTGACATTA (F)GGCCATCTCTTGCTCGAA (R)Spliced PTEN (qPCR)CAAAAGGAGATATCAAGAGGATGGA (F)CAGGAAATCCCATAGCAATAATGTT (R)Unspliced PTEN (qPCR)CAGCCATCATCAAAGAGATCGTT (F)TGGATACAGGTCAAGTCTAAGTCGAA (R)PTEN PINC cloningTTTaagcttATGACAGCCATCATCAAAGAG (F)TTTggatcctagccctcaggaagagacca (R)d(T)VN-BamH1-EcoR1GGATCCGAATTCTTTTTTTTTTTTVN (R)BamHI-EcoR1GGATCCGAATTC (R)PTEN intron 1 binding site 1/1.5 kbGCTGGCAAGAGGGGTTTACTCCAGC (F, PCR1)GGTAGAGTGACACTACAGTGTC (F, PCR2)PTEN intron 1 binding site 1/2.1 kbGCTGGCAAGAGGGGTTTACTCCAGC (F, PCR1)CATGTGAGCTGTGACATGACTGGC (F, PCR2)PTEN intron 1 binding site 1/2.5 kbGCTGGCAAGAGGGGTTTACTCCAGC (F, PCR1)CTGTTTATTGGAGTGGGAAGGC (F, PCR2)PTEN intron 1 binding site 1.5/2.1 kbGGTAGAGTGACACTACAGTGTC (F, PCR1)CATGTGAGCTGTGACATGACTGGC (F, PCR2)Control AMO5′-CCTCCTACCTCAGTTACAATTTATA-3′PTEN AMO5′-CGCAGAAATGGATACAGGTCAAGTC-3′U6atac AMO5′-AACCTTCTCTCCTTTCATACAACAC-3′ Primer and AMO sequences Inhibiting PTEN minor intron splicing affects cell viability Our data suggest that PTEN’s mi-INT retention is a mechanism that reduces the expression of PTEN in cells, leading to inefficient tumor suppression. To assess these effects and confirm that PTEN’s mi-INT is a molecular switch, we designed antisense morpholino oligonucleotides (AMOs) to specifically bind the exon 1/intron 1 junction (Table 1) and thus selectively inhibit the splicing of PTEN’s mi-INT to only produce isoform 2 (Figure 2A). The efficiency of the AMOs was verified in all three cell lines, and an initial dosing experiment showed that 1 μM PTEN AMO was the lowest concentration to almost completely inhibit intron 1 splicing without causing side effects, as evidenced by the drastic decrease in spliced PTEN levels (Figure 2B). Western blot analysis confirmed that PTEN protein was reduced upon transfecting cells with PTEN AMO in MDA-MB-231 and MCF7 (Figure 2C). These data confirm that modulating the splicing of PTEN’s mi-INT directly affects the levels of encoded protein. Inhibiting PTEN minor intron splicing affects cell viability (A) Schematic of PTEN AMO that inhibits minor intron splicing, resulting in the generation of isoform 2 that does not encode PTEN protein. (B) Representative qPCR for spliced and unspliced PTEN in cells transfected with control and PTEN AMO. Data show the relative RNA expression after normalization to the loading control, Actin. Error bars represent standard deviations from three experiments. (C) Representative (N = 3) western blot shows PTEN and Actin (loading control) protein expression in cells transfected with control (C) and PTEN AMO (P), respectively, in each cell line. (D) RealTime-Glo MT cell viability assay results represented as relative cell number for cells transfected with control and PTEN AMO. Results are normalized to day 0 and are for up to 4 days post-transfection. For each condition, 10 wells were used as replicates. Error bars represent standard deviations from three experiments, unless otherwise indicated. To confirm that modulating PTEN protein levels via mi-INT splicing is consequential on cancer cell phenotype, cell proliferation was measured after transfecting cells with PTEN AMO. Selectively targeting PTEN mi-INT splicing increased the proliferation of MCF7 cells by 2-fold but had only a modest effect in the triple-negative (TN) cell line, MDA-MB-231. On the other hand, MDA-MB-468 cells showed no changes in proliferation, which is expected as this cell line does not express a functional PTEN protein (Figure 2D). The modest effect in MDA-MB-231 cells is also expected as PTEN has been shown in advanced tumors to constrain malignant phenotypes, such as invasion and migration, rather than reduce baseline proliferation.49,50,51 These results confirmed that as the PTEN AMO inhibits the splicing of intron 1 of PTEN, reducing the translation of a functional tumor suppressor protein, the expected increase in cell proliferation is observed. A retained PTEN minor intron is processed into a poly-adenylated lncRNA Given that alternative splicing produces more unspliced than spliced PTEN isoforms, we further characterized the unspliced PTEN mRNA. First, we assessed whether the unspliced isoform is structurally as we predicted in Figure 2A. The presence of this isoform would suggest that we could correct the low splicing efficiency of the mi-INT to enhance PTEN protein production as we have previously shown.19 To test this, we designed primers that specifically amplify the exon 1-intron 1 junction, which would be the beginning of the unspliced transcript, or the intron 1-exon 2 junction, which would be the end (Figure 3A). RT-PCR for RNA extracted from all three cell lines consistently showed the expected spliced PTEN, as shown by the primer pair that detects exon 1-exon 2 junction (Ex1-Ex2). However, when the exon-intron boundaries were amplified, the primer pair that detects exon 1-intron 1 boundary (Ex1-Int1) showed a clear signal, whereas those primers that detect intron 1-exon 2 boundary (Int1-Ex2) showed no signal. We confirmed the ability of the primers (Int1-Ex2) to amplify the 3′ end of the intron using genomic DNA as input (Figure 3B). This is also supported by RT-qPCR data that shows that there is a signal only from primers at the start of the unspliced transcript (Exon 1-Intron 1 boundary), but when primers at the Intron 1-Exon 2 boundary are used on the same samples, these are not detected. These data strongly suggest that PTEN’s intron 1 is not fully retained as an intact intron that can be spliced upon demand, but only the 5′ end of it accumulates in cells. A retained PTEN minor intron is processed into a poly-adenylated lncRNA (A) Schematic of isoform 2 with minor intron retention that does not encode PTEN protein, and arrows point to exon 1-intron 1 junction (unspliced at 5′ end of intron 1), and intron 1-exon 2 junctions (unspliced at 3′ end of intron 1). (B) Representative RT-PCR for PTEN exon 1-exon 2 junction (spliced), exon 1-intron 1 junction (unspliced at 5′ end of intron 1), and intron 1-exon 2 junctions (unspliced at 3′ end of intron 1) in MDA-MB-231, MDA-MB-468, and MCF7 cell lines. GAPDH was used as a loading control, and genomic DNA as a positive control. (C) RNA-Seq data follow U6atac AMO transfection in MDA-MB-231 cells, showing PTEN expression in control and U6atac AMO conditions. (D and E) Representative 3′ RACE on RNA extracted from cells transfected with U6atac AMO (D) and control and PTEN AMO (E) on different primers binding different regions in intron 1. (F) Expression of PTEN minor intron in various normal tissues. Data was extracted from the GTEx project. This could be explained if premature cleavage and polyadenylation (PCPA) were to occur in PTEN’s mi-INT. For this to happen, an intronic polyadenylation signal (iPAS) needs to be utilized. Consequently, the intron is cut, and a polyA tail is added, generating a stable polyadenylated transcript. To test this, we used a polyA detector tool (https://apa.cs.washington.edu/detect), which identified several iPASs, including a strong one at position 2469 in the intron. To confirm the enrichment of the polyadenylated RNA and rule out that our data are a side effect of PTEN AMO and its binding to the 5′ splice site of the PTEN pre-mRNA, we inhibited PTEN splicing through the U6atac AMO, which we have previously used to inhibit all mi-INT splicing by masking the catalytic subunit of the minor spliceosome.19 As such, U6atac AMO does not interact with PTEN pre-mRNA, and its effects are indirect. As shown in Figure 3C, transfecting MDA-MB-231 cells with U6atac AMO caused the exact pattern of PCPA at the same location (red arrow). Interestingly, we also observed a complete loss of Exon 2 signal (blue arrow) and the spliced Exon 1-Exon 2 junction reads, suggesting that the complete inhibition of the splicing of PTEN’s mi-INT generates a long noncoding RNA (lncRNA) at the expense of spliced mRNA. To verify the usage of these iPASs, 3′ RACE was employed on cells transfected with U6atac AMO. Utilizing several forward primers in the intron and oligodT reverse primers, we scanned for polyadenylated RNAs in the intron. This showed a clear band corresponding to a polyadenylated RNA at ∼2500 in PTEN’s intron 1 (Figure 3D), supporting our hypothesis that the retained intron 1 of PTEN is processed into a polyadenylated RNA that generates an lncRNA. This was also detected in samples transfected with PTEN AMO (Figure 3E). Finally, to further support this, we checked the genotype-tissue expression (GTEx) Project for the expression of PTEN and its mi-INT splicing efficiency in various normal tissues. Interestingly, many tissues showed some level of intron 1 retention (Figure 3F). However, upon closer inspection, the intron signal did not conform to the typical intron retention but was rather reminiscent of PCPA that has been previously observed upon the inhibition of U1 snRNP.52,53 Figure 3F shows examples of 12 tissues with varying levels of detectable RNA-seq reads in intron 1 of PTEN. Remarkably, all the tissues show an exact point of attrition at around 2500 bases into the intron, which is highly consistent with the data observed from U6atac AMO transfection and 3′RACE. The signal of spliced exons (exon 1 and exon 2 shown in Figure 3F) does not seem to be affected, suggesting that in normal physiological conditions most of the pre-mRNA is spliced properly and goes on to produce a PTEN protein, whereas a small fraction of the pre-mRNA undergoes intronic premature cleavage and polyadenylation (iPCPA), does not form a full-length transcript, and goes on to produce a lncRNA. Analysis of the sequence of Exon 1 plus 2500 bases from Intron 1 did not show any significant open reading frames except for a 46 amino acid polypeptide that would be initiated at the original PTEN start codon and extends into intron 1. Mutating this ORF showed no functional difference (see later in discussion and data not shown), suggesting that this polyadenylated RNA is a long non-coding RNA that we termed PTEN intronic non-coding (PINC) RNA. PINC RNA affects cell viability differently in various breast cancer cell lines To check whether PINC RNA is a functional lncRNA, we cloned Exon 1 plus 2523 bases of Intron 1 (based on the above analysis) into an expression vector and over-expressed them in breast cancer cell lines. Various cell lines were transfected with 250 ng of the expression plasmids (empty or containing PINC) for 48 h. RNA and proteins were then collected to check for the overexpression of the lncRNA and for any effects on the endogenous PTEN expression. We used 250 ng of plasmid to achieve an overexpression that is close to physiological conditions and to avoid non-specific toxicity effects. To measure PINC expression, the PTEN unspliced primers were used because data in Figure 3 confirmed that most, if not all, unspliced PTEN’s mi-INT is processed into an lncRNA. This showed that PINC overexpression was 15, 13, and 7-fold above control in MDA-MB-231, MDA-MB-468, and MCF7, respectively (Figure 4A). While RT-qPCR showed that PINC overexpression was successful, it also showed that this amount of overexpression did not have any significant effects on endogenous spliced PTEN, as there were no changes to spliced PTEN levels after PINC overexpression (Figure 4A). However, western blot analysis showed an increase in PTEN protein levels with PINC overexpression in MDA-MB-231 and MCF7 cell lines (Figure 4B). Interestingly, PINC overexpression caused varying effects on the proliferation of different breast cancer cell lines. As shown in Figure 4C, PINC overexpression caused a significant decrease in the proliferation of MCF7, a modest decrease in the proliferation of MDA-MB-468, and no effect on the proliferation of MDA-MB-231. PINC RNA affects cell viability differently in various breast cancer cell lines (A) Representative qPCR for spliced and unspliced PTEN on MDA-MB-231, MDA-MB-468, and MCF7 cell lines transfected with pcDNA3.1A (control) and PINC RNA-expressing plasmid. Data shows the relative RNA expression after normalization to the loading control, actin. Error bars represent standard deviations from three experiments. Comparison between groups was done using t-test, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. (B) Western blot shows PTEN and Actin (loading control) protein expression in MDA-MB-231, MDA-MB-468 and MCF7 cell lines transfected with control (C) and PINC (P) plasmids. (C) Cell viability using RealTime-Glo MT Cell Viability Assay results represented as relative cell number for MDA-MB-231, MDA-MB-468, and MCF7 cell lines transfected with control and PINC plasmids. Results are normalized to day 0 and are for up to 3 days post-transfection. For each condition, 10 wells were used as replicates. Error bars represent standard deviations from three experiments, unless otherwise indicated. Taken together, we conclude that in MCF7 PINC regulates PTEN levels post-transcriptionally, given the lack of effect on spliced mRNA but the increase in PTEN protein levels. Thus, PINC acts in a PTEN-protein-dependent manner in MCF7. However, we note that some of the effects are not protein-dependent as indicated by MDA-MB-468’s decrease, albeit modest, in viability when these cells lack PTEN protein. The lack of effect in MDA-MB-231 is similar to that of the AMO, given that overexpressing PTEN in these cells is not expected to cause a robust reduction in baseline proliferation. Discussion As a tumor suppressor, PTEN plays a pivotal role in regulating cellular proliferation, with even a small decrease in its levels leading to tumor formation in some tissues, including the breast. As such, many mechanisms are employed to ensure a tight regulation of PTEN expression and its activity in normal tissues. Deregulating any of these mechanisms can lead to cancer. While genomic alterations such as point mutations, truncating mutations, and deletions are well documented, other non-genetic aberrations can reduce PTEN’s expression and/or activity. These include transcriptional inhibition, epigenetic silencing by promoter methylation, post-transcriptional repression by inhibitory miRNAs, and post-translational modifications.22,25,35,36,37,38,39,40,41,42,43 In this study, we focused on a less explored mechanism: alternative splicing. An early study by Sarquis et al.54 showed an interesting correlation between PTEN’s differential expression, its splice variants (SVs), and the pathogenesis of breast cancer. Different SVs have an effect on downstream elements of the AKT pathway, such as Cyclin D1. For example, increased SV5b protein expression enhances cyclin D1 promoter activity, opposite to PTEN’s function.54 On the other hand, Agrawal and Eng55 demonstrated that SV5b is overexpressed in patients with sporadic breast cancer, while in patients with Cowden syndrome, SV3a is overexpressed, suggesting that the differential expression of PTEN SVs may be the cutoff between some benign and cancerous tumors.55,56 While these reports establish the importance of alternative splicing in regulating PTEN’s activity, they all focus on alternative exon splicing. As PTEN’s first intron is a mi-INT and we have demonstrated that it is inefficiently spliced in cells, we investigated here the regulation of intron retention and its correlation with PTEN expression and activity. Moreover, we have previously shown that PTEN expression can be regulated by increasing the efficiency of its mi-INT splicing in a p38 MAPK-dependent manner in cervical cancer cells.19 Thus, we hypothesized that the deregulation of PTEN’s mi-INT splicing is a novel mechanism to limit PTEN expression in cancer. Our data confirms that in several breast cancer cell lines, the first intron in a large fraction of PTEN’s pre-mRNA is not spliced efficiently, producing an isoform that cannot be translated into a functional protein. The use of an antisense morpholino that blocks the 5′ splice site of intron 1 resulted in an almost complete loss of PTEN spliced mRNA, and consequently the encoded protein, but resulted in unexpected effects on cellular proliferation. While MCF7 cells responded as expected when a tumor suppressor such as PTEN was reduced, with increased proliferation, the two TN cell lines used in this study showed little to no decrease in cellular proliferation. The TN cells are models for advanced tumors that are not expected to respond robustly to reducing PTEN levels. Moreover, MDA-MB-468 cells harbor a nonsense mutation in the PTEN gene that leads to complete loss of the protein.47,48 Thus, it is not expected that MDA-MB-468 cells would respond to a reduction in PTEN expression, which was indeed the case. This led us to conclude that in MCF7 cells, the expression and function of the PTEN protein are dependent on proper splicing of its mi-INT. Upon further inspection of PTEN’s mi-INT, we identified at least 12 tissues from GTEx in which the first 2500 bases of the intron are notably accumulated as opposed to the rest of the intron. We confirmed this experimentally in the breast cancer cell lines, showing a 100% biased accumulation of the 5′ end of the intron as opposed to the 3′ end when mi-INT splicing was inhibited using U6atac AMO, which inhibits the mi-INT splicing machinery without binding to the target pre-mRNAs. We opted to use this alternative approach to inhibit PTEN’s mi-INT splicing to avoid the side effects of direct AMO binding to PTEN’s mi-INT and causing artificial stabilization of the pre-mRNA. While the U6atac AMO, such as the GTEx data, resulted in the accumulation of the first 2500 bases of the intron, it was unique in causing the loss of signal in Exon 2 and Exon1/Exon2 junction reads. This strongly suggested that PCPA occurred, resulting in a previously unidentified transcript. Notably, the detection of this novel transcript in various normal tissues (GTEx) indicates that the mechanism is not cancer-specific. Our data show that this transcript is polyadenylated and is unlikely to have an open reading frame. We thus termed it PINC RNA. Remarkably, when PINC RNA was exogenously expressed from a plasmid, it caused different effects on proliferation in the various cell lines without affecting the endogenous expression of PTEN’s spliced mRNA, but with a noticeable effect on PTEN protein. Our data show that overexpressing PINC in MCF7 causes a post-transcriptional increase in PTEN protein, leading to a significant reduction in cellular proliferation. This observation directly links PINC’s function with the post-transcriptional regulation of the PTEN protein. While the proliferation of MDA-MB-231 cells does not respond to increased PTEN protein as previously published, we still observed an increase in PTEN protein levels when PINC was overexpressed, confirming that PINC regulates PTEN levels and/or acts in a PTEN protein-dependent manner. Intriguingly, MDA-MB-468 cells, which are unable to produce a PTEN protein due to a nonsense mutation, showed reduced, albeit modest, proliferation upon the exogenous expression of PINC RNA. This suggests a different mechanism of action for PINC that is independent of the PTEN protein. We thus concluded that while the main pathway for PINC is to regulate PTEN levels post-transcriptionally, some of its effects might be independent of the PTEN protein. We classified PINC RNA as a lncRNA. lncRNAs are typically similar to protein-coding mRNAs in terms of their transcription regulation and post-transcriptional processing (5′ end capping, 3′ end polyadenylation, and alternative splicing),57 but they are less conserved, are expressed at lower levels than mRNAs, and tend to have more tissue-specific expression.57,58,59,60 There are many reasons why a lncRNA would have different effects in different cell lines. Two of the most studied mechanisms by which lncRNAs function in cells, including cancer, are sequestering and thus inhibiting miRNAs and RNA binding proteins (RBPs).61,62,63,64 Given that the expression profiles of miRNAs and RBPs are different in the cell lines studied (unpublished data), it is reasonable to speculate that potential targets might be missing or over-represented in one cell line vs. another. Thus, the effect of sequestering them would be different. In this study, we used the full-length PINC RNA, but we also observed similar effects with a smaller functional fragment (∼340 bases) of PINC (unpublished data). We used this smaller fragment to narrow down potential miRNAs and RBPs that might be sequestered by PINC. Indeed, we identified hsa-miR-542-3P and hnRNP D as the top candidates. Both of these are differentially expressed in the breast cancer cell lines (unpublished data). Sequestering hsa-miR-542-3P or hnRNP D, or a combination of both (and potentially other differentially expressed targets) would be expected to affect cellular proliferation as both of them have been previously shown to play a role in regulating cancer cell behavior.65,66,67,68,69,70 Further detailed work is needed to delineate the exact mechanism by which PINC RNA regulates cellular proliferation, including its sequestration of hsa-miR-542-3P and hnRNP D. It was surprising that our study unmasked a previously unidentified mechanism for regulating a gene as studied as PTEN. On the other hand, PINC RNA expression could have been overlooked as inefficient splicing, which is common in cancer.71,72 In addition, it was likely overlooked as high PINC RNA production is dependent on both iPCPA and alternative splicing, specifically intron retention. While alternative cleavage and polyadenylation toward the 3′ end of pre-mRNAs is a prevalent process and well documented,73,74,75 iPCPA has recently emerged as a significant mechanism to regulate gene expression.52,53 Conversely, there are only a few documented cases of alternative splicing leading to the production of lncRNAs from what is typically a protein-coding gene. These include the PNUTS and PD-L1 protein coding genes that can be alternatively spliced to produce lncRNAs that promote tumor progression.74,76,77 This study provides the first evidence of a PTEN intronic ncRNA (PINC) that is produced when the first intron of PTEN is inefficiently spliced, followed by intronic cleavage and polyadenylation. Overexpressing PINC RNA in breast cancer cells has significant effects on cellular proliferation. While the mechanisms by which PINC RNA regulates proliferation are still poorly understood, they are likely to be a complex combination of post-transcriptional regulation of PTEN protein production and/or PTEN protein-independent mechanisms. It would be important to further understand this relationship, as regulating PTEN levels or activity can have profound therapeutic effects, especially in breast cancer. Limitations of the study While this study proposes that PINC affects phenotypes through miRNA and RBP sequestration, this mechanism is speculative and remains to be explored further. The authors also acknowledge that the heterogeneity in phenotype in different cell types is not fully explained. Additional experimentation is needed to clearly distinguish PTEN-dependent and independent pathways. Resource availability Lead contact Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Ihab Younis (iyounis@andrew.cmu.edu). Materials availability PINC-expressing plasmids and other reagents generated in this study will be made available on request, but we may require a completed materials transfer agreement if there is potential for commercial application. Data and code availability •Data: The data underlying this article will be shared on reasonable request to the lead contact. Data: The data underlying this article will be shared on reasonable request to the lead contact. •Code: This article does not report original code Code: This article does not report original code •Other items: Any additional information required to reanalyze the data reported in this article are available from the lead contact upon reasonable request. Other items: Any additional information required to reanalyze the data reported in this article are available from the lead contact upon reasonable request. Acknowledgments We are grateful to the members of the Biological Sciences program, especially Maya Kemaldean, for comments on the manuscript. This publication was made possible by the grant, NPRP 10-0117-170178, from the Qatar National Research Fund (a member of Qatar Foundation). Additional funding was provided by the generous support of the Qatar Foundation through the Carnegie Mellon University in Qatar’s seed research program; grant no. 38781.1.1390031. The findings and statements made herein reflect the work and are solely the responsibility of the authors. Author contributions M.E.: conceptualization, methodology, validation, formal analysis, investigation, supervision, and writing – review and editing. R.E.: conceptualization, methodology, validation, formal analysis, investigation, supervision, and writing – review and editing. H.H. conceptualization, validation, writing – review and editing, formal analysis, and investigation. A.F. conceptualization, validation, formal analysis, and investigation. B.A. conceptualization, validation, formal analysis, and investigation. I.Y.: conceptualization, methodology, validation, formal analysis, resources, data curation, visualization, project administration, writing – original draft, writing – review and editing, and funding acquisition. Declaration of interests The authors declare no competing interests. STAR★Methods Key resources table REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesanti-PTEN antibodyCell Signaling TechCat# 9552; RRID:AB_10694066anti-GAPDH antibodyCell Signaling TechCat# 97166; RRID:AB_2756824anti-Beta ActinCell Signaling TechCat# 4967; RRID:AB_330288Chemicals, peptides, and recombinant proteinsPTEN antisense morpholinoGeneTools47-10Apr19ACritical commercial assaysRealTime-Glo™ MT Cell Viability AssayPromegaG9711Neon™ NxT Electroporation System 100-μL KitThermo FisherN10096Experimental models: Cell linesMCF7ATCCHTB-22MDA-MB-231ATCCHTB26MDA-MB-468ATCCHTB-132OligonucleotidesSee Table 1 for Primer and AMO SequencesN/AN/ARecombinant DNAPINC in pcDNA3.1A plasmidThis paperN/ASoftware and algorithmsRSubreadopen-source R/Bioconductor packagesN/ADESeqopen-source R/Bioconductor packagesN/AedgRopen-source R/Bioconductor packagesN/AIllumina BCL2Fastq Conversion SoftwareIlluminaN/ApolyA detector toolhttps://apa.cs.washington.edu/detectN/AOtherRNeasy Plus Mini KitQiagen74136ProtoScript® II First Strand cDNA Synthesis KitNew England BioLabsE6560LOneTaq® 2X Master Mix kitNew England BioLabsM0482L Experimental model and study participants details Human breast cancer cell lines MDA-MB-231, MDA-MB-468 and MCF-7 were used as the experimental model for human breast cancer cells. These cells were purchased from ATCC and maintained as described below. This study did not include any genome editing of these cells and their genotype was assumed to be similar to the original ones obtained from ATCC. Routine Mycoplasma testing was done using PCR. For transformation experiments, DH5alpha strain was used as high efficiency, chemically competent E. coli. Method details Cell culture Human breast cancer cell lines MDA-MB-231, MDA-MB-468 and MCF-7 were maintained in Dulbecco’s modified Eagle’s Medium (DMEM) with high glucose, GlutaMAX, supplemented with 10% Fetal Bovine Serum (FBS), 1% (100U/ml) Penicillin-Streptomycin mixture, and 1% non-essential amino acids. Cells were grown at 37°C in a 5% CO2 humidified incubator. At 70–80% confluence, cells were passaged. Transfection with AMOs Control, PTEN and U6atac antisense morpholino oligonucleotides (AMOs) were synthesized by GeneTools (Sequences in Table 1), and up to 1 μM were transfected into 1 million cells using electroporation by Neon 100 μL Transfection system (Invitrogen). Cells were resuspended in resuspension buffer R, provided with the transfection kit. The AMOs were then added to the cells and electroporated, according to the manufacturer’s protocol. Transfected cells were then plated in complete medium lacking Penicillin-Streptomycin. RNA and proteins were harvested from transfected cells one day post transfection. Transfected cells were also plated for cell proliferation. PINC plasmid design and transfection For the cloning of PTEN’s intronic noncoding RNA, or PINC, the coding part of Exon 1 plus 2523 bases of Intron 1 were cloned into pcDNA3.1A. For exogenous expression, 250 ng of pcDNA3.1A (control) or the plasmid expressing PINC were transfected into 250,000 breast cancer cells using Lipofectamine 3000 reagent according to manufacturer’s recommendation. Total RNA and proteins were extracted 48 h post-transfection as described below. Cell proliferation assay Cell proliferation was evaluated using RealTime-Glo MT Cell Viability Assay (Promega). After transfection, cells were plated at a density of 2000–2500 cells/well in an opaque-walled 96-well plate. For each treatment, 10 wells were used as replicates. The MT Cell Viability Substrate and NanoLuc Enzyme were heated to 37°C, diluted in complete media and added at a concentration of 1X for a total volume of 100 μL per well. Cells were then incubated for 30 min at 37°C. Luminescence was measured on a GloMax Explorer Multimode Microplate Reader (Promega) at 37°C, with an integration time of 0.3 s per well. Measurements were taken at the time of plating (Day 0), and subsequently on a daily basis for 3–4 days. RNA and protein extraction RNA was extracted from cells following transfection or inhibitor treatments using Qiagen’s RNeasy Plus Mini Kit (Catalog #: 74136), according to manufacturer’s recommendations. Protein was harvested from live cells using RIPA buffer with protease and phosphatase inhibitor cocktail. Protein concentrations were measured using the Bio-Rad protein assay. RT-PCR and qPCR For reverse transcription, 1 μg of RNA was converted to cDNA using the ProtoScript II First Strand cDNA Synthesis Kit (New England BioLabs), according to the manufacturer’s protocol. cDNA was then diluted 10-fold in nuclease-free water and used as input for PCR. Primer sequences are in Table 1. PCR was carried out using OneTaq 2X Master Mix kit (New England BioLabs), according to manufacturer’s recommendations. Cycling conditions were: denaturing of DNA at 94°C for 30 s; 32 cycles of 30 s at 94°C for denaturation, 30 s at 50°C for annealing, and 30 s at 68°C for extension; and a final extension at 68°C for 5 min. Amplified products were run on 2% TAE agarose gels and imaged in a UV box. For qPCR, expression of spliced and unspliced PTEN was quantified using PowerTrack SYBR Green master mix (Applied Biosystems). Spliced PTEN primers were designed to bind exon 1 (forward) and exon 2 (reverse), whereas unspliced PTEN primers were designed to bind exon 1 (forward) and exon 1-intron 1 junction (reverse). qPCR was performed on the QuantStudio 6 system (Applied Biosystems), and relative expressions were normalized to beta actin levels. Primer sequences are in Table 1. Western blotting Extracted total protein lysates were diluted in Tris-Glycine SDS buffer and β-mercaptoethanol, then denatured at 95°C for 5 min and loaded on 4–12% NuPAGE Bis-Tris gels (Invitrogen). Samples were transferred to a nitrocellulose membrane using a semi-wet transfer. The membrane was blocked with 5% BSA in TBST for 2 h at room temperature, and then incubated with primary antibodies overnight at 4°C. The primary antibodies used were: rabbit monoclonal anti-PTEN antibody (Cell Signaling Technology) and mouse monoclonal anti-GAPDH antibody or rabbit anti-Beta Actin (Cell Signaling Technology). Following washes with TBST, the HRP-conjugated secondary antibodies were added at 1:5000 dilution and incubated for 1 h at room temperature, followed by washes with TBST. Membrane was then imaged using the Li-Cor Odyssey Fc Imager. RNA-sequencing Total RNA from MDA-MB-231 and MCF7 cells that were transfected with AMOs was extracted as described above. Total RNA with a RIN number above 8 was used as input for library preparation using TruSeq Stranded mRNA kit from Illumina following the manufacturer’s protocol. Briefly, from 500 ng of total RNA, mRNA molecules were purified using poly-T oligo attached magnetic beads and then mRNA was fragmented. cDNA was generated from the cleaved RNA fragments using random priming during first and second strand synthesis. Barcoded DNA adapters were ligated to both ends of the DNA and then amplified. The quality of the library generated was checked on an Agilent 2100 Bioanalyzer system and quantified using Qubit system. Libraries that pass quality control were pooled and sequenced on a NextSeq2000 system at a minimum of 20 million paired-end reads (2x75 bp) per sample. Basic trimming and quality control were performed during the conversion of raw data to fastq using Illumina BCL2Fastq Conversion Software. For alignment and expression analysis, an R pipeline was used, including the following packages: RSubread, DESeq, and edgeR. PAS detection and 3′ RACE The polyA detector tool (https://apa.cs.washington.edu/detect) was used to identify PAS signals, where the PTEN minor intron sequence was inputted. For 3′ Rapid Amplification of cDNA Ends (RACE), total RNA was extracted as described above. PolyA tail-specific primer oligo-d(T)VN-BamH1-EcoR1 (Table 1) was used to reverse transcribe poly-adenylated mRNA into cDNA using ProtoScript II First Strand cDNA Synthesis Kit (New England BioLabs) as described above. The reverse-transcribed cDNA was then diluted 10-fold in nuclease-free water. The nested PCR included two rounds of amplification. For PCR1, several forward primers that hybridize to PTEN intron 1 were used (Table 1), while the reverse primer was the same polyA tail-specific primer oligo-d(T)VN-BamH1-EcoR1. Amplification conditions were as follows: 30 s at 94°C; 25 cycles of 30 s at 94°C, 30 s at 50°C, and 2 min at 68°C; and a final extension of 5 min at 68°C. For PCR2, 5 μL of the amplification product of PCR1 was used as input, and amplified using one of the many forward primers that hybridize to intron 1 (Table 1), in addition to a reverse primer that corresponds to BamH1-EcoR1 sequence (Table 1), ensuring that only poly-adenylated mRNA that were amplified in PCR1 are a proper template for PCR2. For PCR2, the same conditions were used as PCR1 except for the number of cycles, which was 35 in this case. Splicing analysis For qPCR analysis, the Ct for each primer pair was calculated, normalized to actin (DeltaCt) and then used to calculate relative expression as 2ˆ(-DeltaCt). The Splicing Index was calculated as the ratio of signal from primers that detect spliced (exon/exon) to total (exon/exon + exon/intron). For RNA-Seq (U6atac AMO and GTEx), the Splicing Index could not be inferred from the data due to the large size of the intron and possible usage of cryptic splice sites. Instead, a more qualitative approach was used to detect read bias for the 5′ end of the intron, as this suggests premature cleavage and polyadenylation. Splicing inhibition was inferred from the number of junction reads spanning exon1/exon2 relative to reads that span exon-intron boundaries plus intronic reads. Quantification and statistical analysis Statistical significance in this manuscript was measured using t test except for proliferation data, in which a non-parametric (Mann-Whitney test) was used. ∗, ∗∗, and ∗∗∗ indicate p values <0.5, 0.05, or 0.005, respectively. N = 3 for most experiments, unless otherwise indicated in the figure legends, where N represents the number of biological replicates conducted for each experiment. Error bars on all figures represent standard deviation. Published: April 14, 2026 References Associated Data Data Availability Statement •Data: The data underlying this article will be shared on reasonable request to the lead contact. Data: The data underlying this article will be shared on reasonable request to the lead contact. •Code: This article does not report original code Code: This article does not report original code •Other items: Any additional information required to reanalyze the data reported in this article are available from the lead contact upon reasonable request. Other items: Any additional information required to reanalyze the data reported in this article are available from the lead contact upon reasonable request.