The lncRNA, CyKILRb acts as a ceRNA for tumor suppressive microRNA to augment oncogenic phenotype via induction of PIK3R2/PI3K/AKT axis Authors: Xiujie Xie, H. Patrick Macknight, Barbara Dziegielewska, Amy L. Lu, Ngoc Thanh Vu, Charles E. Chalfant Published: 2026 ABSTRACT The lncRNA CyKILR produces two splice variants with antagonistic roles in lung tumorigenesis. While its nuclear counterpart functions as a tumor suppressor, we revealed that the cytoplasmic variant, CyKILRb, is a potent oncogenic driver that activates the PI3K/AKT signaling axis. Specifically, downregulation of CyKILRb induced the loss of the PI3K activator PIK3R2 and the tumor promoter RPS6KB2, while simultaneously increasing the tumor suppressors, p21 and p27. CyKILRb ectopic expression produced the contrasting effect, and suppression of either PIK3R2, PI3K, or AKT attenuated CyKILRb-induced cell proliferation and clonogenicity. Combinatorial expression studies oriented the pathway from CyKILRb→↑PIK3R2→PI3K→AKT→enhanced oncogenicity. An in silico analysis comparing predicted microRNA response elements (MREs) across the CyKILRb and PIK3R2 transcripts identified shared contiguous complementary miRNA binding sequences. Of these miRs, miR-3151-5p, a tumor suppressor and known PIK3R2 regulator, was identified. This miR associated with both RNAs in cells, and treatment of NSCLC cells with a miR-3151-5p mimic reduced PIK3R2 and clonogenicity. Ectopic expression of CyKILRb containing a mutated MRE for miR-3151-5p showed an inability to upregulate PIK3R2 and tumor-promoting phenotypes in contrast to WT CyKILRb. These findings define a novel role for CyKILRb as a competing endogenous RNA for tumor suppressive miRs, which induces the PI3K/AKT axis. Introduction Lung cancer remains the leading cause of cancer-related mortality worldwide with non-small cell lung cancer (NSCLC) accounting for approximately 85% of these cases.1,2 Despite advancements in precision-based therapies as well as new immunotherapies, NSCLC remains a highly aggressive malignancy with limited treatment options for patients. In recent years, non-coding RNAs (ncRNAs) have emerged as critical functional regulators of gene expression, which significantly influences tumor progression and therapy resistance in human malignancies including NSCLC.3,4 The complexity of ncRNAs, particularly their interactions within the tumor microenvironment, adds another layer of challenge to NSCLC treatment.5 Among ncRNAs, long non-coding RNAs (lncRNAs) have garnered increasing attention for their role in modulating oncogenic and tumor suppressor pathways through diverse cellular mechanisms including chromatin remodeling, transcriptional regulation, and post-transcriptional modulation.6,7 Additionally, lncRNAs exhibit complex secondary and tertiary structures, enabling them to interact with DNA, RNA, and proteins to modulate cellular pathways at multiple levels.8 Several lncRNAs have been implicated in NSCLC progression and drug resistance with some functioning as oncogenes and others as tumor suppressors. For instance, MALAT1 is an oncogenic lncRNA that promotes metastasis by regulating epithelial-mesenchymal transition (EMT) and enhancing tumor cell migration.9 Conversely, MEG3 functions as a tumor suppressor by interacting with p53 and inhibiting cell proliferation and survival.10,11 The lncRNA HOTAIR (HOX Transcript Antisense Intergenic RNA) is reported as a contributor to resistance against tyrosine kinase inhibitors (TKIs) in EGFR-mutant NSCLC by epigenetically regulating drug-response genes.12 Similarly, LINC00460 has been reported to promote cisplatin resistance by modulating the PI3K/AKT signaling pathway.13 Given their crucial role in oncogenesis and therapy resistance, lncRNAs present promising targets for therapeutic intervention in NSCLC. However, further research is needed to elucidate the precise molecular mechanisms governing their function, structural dynamics, and regulatory interactions. Despite their functional significance in NSCLC, many lncRNAs remain uncharacterized as well as key aspects of their regulatory roles left unexplored.14 In this regard, previous work from our laboratory identified CyKILR, a novel lncRNA expressed in NSCLC with wild-type (WT) CDKN2A and STK11 genes.15 This lncRNA presented with two splice variants, CyKILRa and CyKILRb, based on inclusion/exclusion of exon 3.15 While CyKILRa (exon 3 included) functions as a tumor suppressor, CyKILRb (exon 3 excluded) promotes oncogenesis and was required to maintain tumorigenic capacity in NSCLC cells with the aforementioned oncogenotype. Still, the precise molecular mechanisms underlying the roles of these lncRNA splice variants remained unclear. In this study, we sought to elucidate the cellular mechanisms by which CyKILRb augments oncogenic signaling. Through deep RNA sequencing (RNA-seq), CyKILRb was discovered to modulate a crucial signaling cascade that governs cell survival, proliferation, and apoptosis resistance, specifically the PIK3R2/PI3K/AKT pathway. Additionally, a role for CyKILRb as a competing endogenous RNA (ceRNA) for tumor suppressive microRNAs (miRNA) was identified, which mediated the regulation of PIK3R2 expression in NSCLC cells. This comprehensive regulatory ability by one lncRNA as identified in this study underscores the complex interplay between lncRNAs and tumorigenesis, which may offer powerful future strategies for the development of precision-based treatments for specific forms of NSCLC. Results The cytoplasmic lncRNA, CyKILRb, induces the PI3K/AKT signaling pathway Our laboratory recently reported that the lncRNA, CyKILRb, was a potent tumor promoting factor. To delve into the mechanism of action for this lncRNA, the transcriptomic signatures related to CyKILRb expression were interrogated using an unbiased approach. Specifically, CyKILRb was downregulated in two NSCLC cell lines with WT STK11 and CDKN2A genes (H1299 and H2009) followed by deep RNA-seq and pathway analysis (Figure 1; Figure S1; Tables S1 and S2). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis revealed that the PI3K/AKT signaling pathway exhibited the most significant gene expression changes upon CyKILRb silencing (Figures 1A–1C). Additionally, four significantly altered genes (differentially expressed genes [DEGs]) associated with the PI3K/AKT pathway were identified based on the highest percent change and lowest p value: downregulation of PIK3R2 and RPS6KB2 with upregulation of CDKN1A (p21) and CDKN1B (p27) (Figure 1C). These effects were validated at the RNA and protein levels (Figures 1D and 1E; Figure S2) along with loss of AKT activation/phosphorylation (Figure 1E; Figure S2). Ectopic expression of CyKILRb in these NSCLC cell lines (Figure 2A) modulated the expression of these proteins and AKT activation/phosphorylation in the opposite direction as those observed with CyKILRb downregulation (Figures 2B and 2C). Additionally and in stark contrast to our reported effects of CyKILRb downregulation in NSCLC cells,15 CyKILRb ectopic expression induced a significant increase in both cellular proliferation and clonogenicity (Figures 2D and 2E). These data demonstrate that CyKILRb is a potent activator of the PI3K/AKT axis and regulates the expression of PIK3R2, PRK6B2, p21, and p27.Figure 1The cytoplasmic lncRNA, CyKILRb, augments the PI3K/AKT axis and modulates downstream tumorigenesis regulatorsH2009 and H1299 cells were treated with either scrambled control siRNA (SCR) or CyKILRb siRNA (siCYB). Total RNA and protein were extracted. RNAs were subjected to both deep RNA sequencing and quantitative reverse-transcription PCR (RT-qPCR). The RNA sequencing was then subjected to data filtering, clean-read alignment, gene expression analysis, and differential expression gene (DEG) analysis, (A) KEGG enrichment analysis of annotated DEGs was performed using R Phyper. The panel depicts the KEGG pathway enrichment analysis of differentially expressed genes following CyKILRb knockdown in H2009 and H1299 (red arrow designates the PI3K/AKT signaling pathway, which demonstrated the largest enrichment). Right inset: RT-qPCR results for CyKILRb RNA levels normalized to GAPDH mRNA levels for H2009 and H1299 cells treated with either scrambled control siRNA (SCR) or CyKILRb siRNA (siCYB). (B) KEGG signaling pathway analysis of the PI3K/AKT pathway comparing CyKILRb knockdown (siCYB) versus SCR in H2009 and H1299: upregulated genes = red box, downregulated genes = blue box. (C) Heatmap depiction showing the top DEGs in the PI3K/AKT signaling pathway following CyKILRb knockdown (siCYB) versus SCR in H2009 and H1299 (red = upregulated genes; blue = downregulated genes). (D) RNA expression levels of PIK3R2, RPS6KB2, CDKN1A, and CDKN1B normalized to GAPDH RNA levels using RT-qPCR analyses on total RNA from H2009 and H1299 cells following CyKILRb knockdown (siCYB) compared to SCR. (E) Total protein from H2009 and H1299 treated with CyKILRb siRNA (siCYB) versus SCR siRNA were analyzed by SDS-PAGE and western immunoblotting for the depicted factors, PIK3R2, RPS6KB2, CDKN1A (p21), AKT, phospho-AKT, GAPDH, and CDKN1B (p27). Data in (E) are representative of n = 3 repeated on at least two separate occasions, and quantification analysis of each protein level normalized to GAPDH is shown in Figure S2. Graphical data in this figure are presented as dot plots (n ≥ 3 replicates repeated on at least two separate occasions), and statistical analyses were performed using an unpaired t test with Welch’s correction; ∗∗∗p < 0.001.Figure 2Ectopic expression of CyKILRb activates the PI3K-AKT pathway and enhances NSCLC cell proliferation and clonogenic survivalThe designated NSCLC cell lines were transfected with either an empty vector control (Vec) or CyKILRb (CYB) plasmids. After 48 h post-transfection, the cells were analyzed by: (A) RT-qPCR analysis for CyKILRb RNA levels normalized to GAPDH RNA levels. (B) SDS-PAGE and western immunoblot analysis of key factors in the PI3K-AKT and tumorigenesis pathways. (C) Quantification of the protein levels for the depicted factors in (B) normalized to GAPDH (for all replicates). (D) Cell proliferation assay comparing CyKILRb ectopic expression (CYB) vs. vector control (Vec) after 96 h ggVennDiagram packages (4 days). The percent increase in proliferation is depicted with red text. (E) Clonogenic survival assay comparing CyKILRb ectopic expression (CYB) vs. vector control (Vec) after 7–21 days. Data in this figure are presented as dot plots (n ≥ 3 replicates repeated on at least two separate occasions), and statistical analyses were performed using an unpaired t test with Welch’s correction; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. The cytoplasmic lncRNA, CyKILRb, augments the PI3K/AKT axis and modulates downstream tumorigenesis regulators H2009 and H1299 cells were treated with either scrambled control siRNA (SCR) or CyKILRb siRNA (siCYB). Total RNA and protein were extracted. RNAs were subjected to both deep RNA sequencing and quantitative reverse-transcription PCR (RT-qPCR). The RNA sequencing was then subjected to data filtering, clean-read alignment, gene expression analysis, and differential expression gene (DEG) analysis, (A) KEGG enrichment analysis of annotated DEGs was performed using R Phyper. The panel depicts the KEGG pathway enrichment analysis of differentially expressed genes following CyKILRb knockdown in H2009 and H1299 (red arrow designates the PI3K/AKT signaling pathway, which demonstrated the largest enrichment). Right inset: RT-qPCR results for CyKILRb RNA levels normalized to GAPDH mRNA levels for H2009 and H1299 cells treated with either scrambled control siRNA (SCR) or CyKILRb siRNA (siCYB). (B) KEGG signaling pathway analysis of the PI3K/AKT pathway comparing CyKILRb knockdown (siCYB) versus SCR in H2009 and H1299: upregulated genes = red box, downregulated genes = blue box. (C) Heatmap depiction showing the top DEGs in the PI3K/AKT signaling pathway following CyKILRb knockdown (siCYB) versus SCR in H2009 and H1299 (red = upregulated genes; blue = downregulated genes). (D) RNA expression levels of PIK3R2, RPS6KB2, CDKN1A, and CDKN1B normalized to GAPDH RNA levels using RT-qPCR analyses on total RNA from H2009 and H1299 cells following CyKILRb knockdown (siCYB) compared to SCR. (E) Total protein from H2009 and H1299 treated with CyKILRb siRNA (siCYB) versus SCR siRNA were analyzed by SDS-PAGE and western immunoblotting for the depicted factors, PIK3R2, RPS6KB2, CDKN1A (p21), AKT, phospho-AKT, GAPDH, and CDKN1B (p27). Data in (E) are representative of n = 3 repeated on at least two separate occasions, and quantification analysis of each protein level normalized to GAPDH is shown in Figure S2. Graphical data in this figure are presented as dot plots (n ≥ 3 replicates repeated on at least two separate occasions), and statistical analyses were performed using an unpaired t test with Welch’s correction; ∗∗∗p < 0.001. Ectopic expression of CyKILRb activates the PI3K-AKT pathway and enhances NSCLC cell proliferation and clonogenic survival The designated NSCLC cell lines were transfected with either an empty vector control (Vec) or CyKILRb (CYB) plasmids. After 48 h post-transfection, the cells were analyzed by: (A) RT-qPCR analysis for CyKILRb RNA levels normalized to GAPDH RNA levels. (B) SDS-PAGE and western immunoblot analysis of key factors in the PI3K-AKT and tumorigenesis pathways. (C) Quantification of the protein levels for the depicted factors in (B) normalized to GAPDH (for all replicates). (D) Cell proliferation assay comparing CyKILRb ectopic expression (CYB) vs. vector control (Vec) after 96 h ggVennDiagram packages (4 days). The percent increase in proliferation is depicted with red text. (E) Clonogenic survival assay comparing CyKILRb ectopic expression (CYB) vs. vector control (Vec) after 7–21 days. Data in this figure are presented as dot plots (n ≥ 3 replicates repeated on at least two separate occasions), and statistical analyses were performed using an unpaired t test with Welch’s correction; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. PIK3R2 downregulation mitigates CyKILRb augmentation of tumorigenic phenotypes As CyKILRb induced the PI3K/AKT pathway, we next examined the role of the PI3K activator, PIK3R2, in modulating CyKILRb effects on this signaling pathway as PIK3R2 expression was significantly modulated by CyKILRb (a top 10 DEG based on the highest percent change and lowest p value; Table S1), and this factor is a known activator of PI3K.16 Ectopic expression of PIK3R2 significantly attenuated the effects of CyKILRb downregulation on AKT activation/phosphorylation as well as cellular proliferation and clonogenic survival (Figures 3A–3C) with a complete “rescue” of the ablated clonogenicity imparted by CyKILRb downregulation (Figure 3C). Cellular proliferation was also “rescued” by PIK3R2 expression (∼75% on average), but the proliferative capacity was not completely restored to the level of cells ectopically expressing PIK3R2 transfected with control small interfering RNA (siRNA) (Figure 3B). To determine whether PIK3R2 mediates CyKILRb-driven tumor growth, CyKILRb-knockdown cells with or without PIK3R2 overexpression were examined in vivo. Silencing CyKILRb significantly suppressed tumor growth, whereas PIK3R2 overexpression enhanced tumor growth (Figure 3D; Figure S3A). Importantly, restoration of PIK3R2 expression rescued the growth inhibition caused by CyKILRb knockdown and restored tumor growth to levels comparable to those observed with PIK3R2 overexpression alone (Figure 3D; Figure S3A). Additionally, the ectopic expression of constitutively active AKT (caAKT) completely “rescued” the suppression of both clonogenicity and proliferation induced by CyKILRb downregulation, but no significant effect on PIK3R2 expression was observed (Figures 3E–3G). These data demonstrate that CyKILRb modulates specific oncogenic phenotypes (e.g., clonogenic survival) by regulating PI3KR2 expression and AKT activation.Figure 3Ectopic expression of PIK3R2 or constitutively active AKT rescues the suppression of tumorigenic phenotypes induced by CyKILRb loss/downregulationThe depicted NSCLC cells lines were transfected with either scrambled siRNA (SCR) or CyKILRb-specific siRNA (siCYB). After 24 h, cells were transfected with an empty vector (Vec), PIK3R2, or constitutive-active AKT (caAKT) expression constructs. The following downstream analyses were then performed 24 h post-plasmid transfection except cellular proliferation (96 h) and clonogenic survival (7–21 days). (A and D) SDS-PAGE and western immunoblotting results for the protein levels of P-AKT, T-AKT, PIK3R2, and GAPDH. Data in (A) and (D) are representative of ≥3 biological replicates repeated on at least two separate occasions. (B and E) Cell proliferation analysis for the depicted NSCLC cells transfected with SCR or siCYB ± Vec, PIK3R2 (B), or caAKT (E) after 96 h. (C and F) Clonogenic survival analysis for the depicted NSCLC cells transfected with SCR or siCYB ± Vec, PIK3R2 (B), or caAKT (E) after 7–14 days. (D) Endpoint tumor volumes and growth curves of xenografts stably expressing either scrambled control shRNA (SCR) and control vector (Vec) (SCR-Vec), CyKILRb shRNA (shCYB) and Vec (shCYB-Vec), SCR and PIK3R2 ectopic expression (PIK3R2) (SCR/PIK3R2), or shCYB and PIK3R2 (shCYB/PIK3R2). PIK3R2 restoration rescued the tumor growth suppression induced by CyKILRb knockdown. Data are presented as mean ± SEM (n = 6 mice per group). Quantitative data depicted in this figure are presented as dot plots (n ≥ 3 replicates repeated on at least two separate occasions), and statistical analysis was performed using one-way ANOVA with Tukey’s multiple comparisons test; ns, not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. Ectopic expression of PIK3R2 or constitutively active AKT rescues the suppression of tumorigenic phenotypes induced by CyKILRb loss/downregulation The depicted NSCLC cells lines were transfected with either scrambled siRNA (SCR) or CyKILRb-specific siRNA (siCYB). After 24 h, cells were transfected with an empty vector (Vec), PIK3R2, or constitutive-active AKT (caAKT) expression constructs. The following downstream analyses were then performed 24 h post-plasmid transfection except cellular proliferation (96 h) and clonogenic survival (7–21 days). (A and D) SDS-PAGE and western immunoblotting results for the protein levels of P-AKT, T-AKT, PIK3R2, and GAPDH. Data in (A) and (D) are representative of ≥3 biological replicates repeated on at least two separate occasions. (B and E) Cell proliferation analysis for the depicted NSCLC cells transfected with SCR or siCYB ± Vec, PIK3R2 (B), or caAKT (E) after 96 h. (C and F) Clonogenic survival analysis for the depicted NSCLC cells transfected with SCR or siCYB ± Vec, PIK3R2 (B), or caAKT (E) after 7–14 days. (D) Endpoint tumor volumes and growth curves of xenografts stably expressing either scrambled control shRNA (SCR) and control vector (Vec) (SCR-Vec), CyKILRb shRNA (shCYB) and Vec (shCYB-Vec), SCR and PIK3R2 ectopic expression (PIK3R2) (SCR/PIK3R2), or shCYB and PIK3R2 (shCYB/PIK3R2). PIK3R2 restoration rescued the tumor growth suppression induced by CyKILRb knockdown. Data are presented as mean ± SEM (n = 6 mice per group). Quantitative data depicted in this figure are presented as dot plots (n ≥ 3 replicates repeated on at least two separate occasions), and statistical analysis was performed using one-way ANOVA with Tukey’s multiple comparisons test; ns, not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. CyKILRb enhances the clonogenic survival of NSCLC cells via the PI3K/AKT pathway Our findings strongly suggested that CyKILRb enhanced NSCLC clonogenicity through activation of the PI3K/AKT pathway via induction of PIK3R2. To further investigate this signaling link, PIK3R2 was acutely downregulated by siRNA in vector control and CyKILRb expressing NSCLC cells. PI3K or AKT were also inhibited concurrently with CyKILRb ectopic expression (Figure 4). Downregulation of PIK3R2 attenuated the ability of CyKILRb to activate the AKT pathway, downregulate p21 and p27, and enhance NSCLC clonogenic survival (Figures 4A and 4B). In contrast, the enhanced RPS6KB2 expression induced by CyKILRb ectopic expression was not significantly affected by siPIK3R2 (Figure 4A). Inhibition of either the PI3K or AKT also significantly suppressed the induction of AKT activation/phosphorylation and the enhancement of NSCLC clonogenic survival enhanced by CyKILRb ectopic expression (Figures 4C–4F). Moreover, the induction of PIK3R2 expression was unaffected by these inhibitors of the PI3K/AKT axis (Figures 4C–4E). Ectopic expression of caAKT “rescued” the loss of clonogenic survival imparted by down regulation of PIK3R2 (Figure S4A). Lastly, inhibition of PI3K blocked the enhancement of clonogenicity imparted by PIK3R2 ectopic expression (Figure S4B). Therefore, these data demonstrate that CyKILRb drives NSCLC clonogenicity via the PI3K/AKT axis and further orient the signaling pathway from CyKILRb → PIK3R2 → PI3K/AKT.Figure 4PIK3R2 downregulation or inhibition of the PI3K/AKT pathway mitigates CyKILRb effects on tumorigenic phenotypesThe depicted NSCLC cell lines were transfected with either an empty vector control (Vec) or CyKILRb (CYB) plasmids. After 24 h, cells were either further transfected ± scrambled siRNA (SCR) or PIK3R2-specific siRNA (siPIK3R2), or further treated ± a PI3K inhibitor (PI3Ki), AKT inhibitor (AKTi), or sham control (DMSO). The depicted downstream analyses were conducted 24 h post-siRNA transfection or treatment except for clonogenic survival (post 7–14 days). (A) Protein levels of PIK3R2, phospho-AKT (P-AKT), total AKT (T-AKT), p21, p27, RPS6KB2, and GAPDH (for normalization) using SDS-PAGE and western immunoblotting. Data in (A) are representative of three biological replicates repeated on at least two separate occasions. (B) Clonogenic survival analyses. (C and E) SDS-PAGE and western immunoblotting for PIK3R2, phospho-AKT (P-AKT), total AKT (T-AKT), and GAPDH (for normalization) of Vec or CYB + AKTi, PI3Ki, or DMSO treatment. Data are representative of three biological replicates repeated on at least two separate occasions. (D and F) Clonogenic survival analyses. Quantitative data depicted in this figure are presented as dot plots (n ≥ 3 replicates repeated on at least two separate occasions), and statistical analysis was performed using one-way ANOVA with Tukey’s multiple comparisons test (n ≥ 3 replicates); ns, not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. PIK3R2 downregulation or inhibition of the PI3K/AKT pathway mitigates CyKILRb effects on tumorigenic phenotypes The depicted NSCLC cell lines were transfected with either an empty vector control (Vec) or CyKILRb (CYB) plasmids. After 24 h, cells were either further transfected ± scrambled siRNA (SCR) or PIK3R2-specific siRNA (siPIK3R2), or further treated ± a PI3K inhibitor (PI3Ki), AKT inhibitor (AKTi), or sham control (DMSO). The depicted downstream analyses were conducted 24 h post-siRNA transfection or treatment except for clonogenic survival (post 7–14 days). (A) Protein levels of PIK3R2, phospho-AKT (P-AKT), total AKT (T-AKT), p21, p27, RPS6KB2, and GAPDH (for normalization) using SDS-PAGE and western immunoblotting. Data in (A) are representative of three biological replicates repeated on at least two separate occasions. (B) Clonogenic survival analyses. (C and E) SDS-PAGE and western immunoblotting for PIK3R2, phospho-AKT (P-AKT), total AKT (T-AKT), and GAPDH (for normalization) of Vec or CYB + AKTi, PI3Ki, or DMSO treatment. Data are representative of three biological replicates repeated on at least two separate occasions. (D and F) Clonogenic survival analyses. Quantitative data depicted in this figure are presented as dot plots (n ≥ 3 replicates repeated on at least two separate occasions), and statistical analysis was performed using one-way ANOVA with Tukey’s multiple comparisons test (n ≥ 3 replicates); ns, not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. CyKILRb modulates the sensitivity of NSCLC to cisplatin To investigate the functional role of CyKILRb and PIK3R2 in modulating the sensitivity of NSCLC to standard of care therapeutics, cisplatin responsiveness was assessed. Specifically, clonogenic survival assays were performed in H1299 cells following CyKILRb knockdown or overexpression, as well as PIK3R2 knockdown, in the presence or absence of cisplatin. CyKILRb knockdown synergistically enhanced the inhibitory effect of cisplatin on colony formation (Figure S5A). In contrast, CyKILRb overexpression significantly increased clonogenic growth under both basal conditions and cisplatin treatment compared with vector-transfected controls, indicating that CyKILRb promotes cell survival and reduces cisplatin sensitivity (Figure S5B). Consistently, PIK3R2 knockdown markedly suppressed colony formation and increased cisplatin sensitivity (Figure S5C). To further determine whether PIK3R2 mediates the effects of CyKILRb on cisplatin sensitivity, rescue experiments were performed. Silencing of CyKILRb significantly reduced colony formation and enhanced cisplatin-induced growth inhibition (Figure S5D). Importantly, ectopic expression of PIK3R2 partially restored colony-forming capacity in CyKILRb-depleted cells under both control and cisplatin-treated conditions (Figure S5D). These findings demonstrate that PIK3R2 functions downstream of CyKILRb and partially mediates the ability of CyKILRb to promote clonogenic survival and cisplatin resistance in NSCLC cells. CyKILRb is a ceRNA for miR-3151-5p, which regulates the PIK3R2/PI3K/AKT axis Many lncRNAs function as ceRNAs for miRNAs to modulate gene expression, and several top DEGs regulated by CYKILRb (e.g., PIK3R2, FTH1, HTRA1, TMEM158, ZNF703, and FOXD1) are all reported to be controlled by miRNAs (Tables S1 and S2).17,18,19,20,21 To determine whether CyKILRb possessed contiguous reverse and complementary sequences to known miRs, an unbiased bioinformatics approach was utilized to perform an in silico prediction by matching the CyKILRb sequence with miRNAs downloaded from miRBase using either canonical seed sites (nt 2–8) or non-canonical binding sites (nt 3–9). In this regard, 98 possible miRs were predicted to bind specific miRNA response elements (MREs) in the CyKILRb mRNA sequence (Figure 5A; Table S3). Statistical comparisons to the miRs predicted to bind MREs in the PIK3R2 mRNA (Table S4) showed 10 shared miRs (Figure 5A). Among the top candidates, miR-3151-5p showed a predicted non-canonical 3′–9′ interaction with CyKILRb at nucleotides 284–290, comprising a total of seven base pairs (Figure 5B). Notably, miR-3151-5p has been reported to harbor three binding sites within the 3′ untranslated region (3′ UTR) of PIK3R2 acting as both a tumor suppressor and a known regulator of PIK3R2 expression (Figures 5B and 5C).22 These data supported the hypothesis that CyKILRb acted as a ceRNA for miR-3151-5p to block the degradation of PIK3R2 through the AGO2 system (Figure 5C). In this regard, treatment of NSCLC cells with a miR-3151-5p mimic reduced both PIK3R2 and CyKILRb RNA levels as well as clonogenic survival (Figures 5D–5F). Additionally, treatment of NSCLC cells with the miR-3151-5p mimic suppressed both basal and CyKILRb-induced PIK3R2 expression and AKT activation/phosphorylation (Figure 5G; Figure S6). In contrast, treatment of NSCLC cells with a miR-3151-5p inhibitor increased both PIK3R2 expression and clonogenic survival (Figures 5D–5F) as well as “rescued” the loss of PIK3R2 expression and AKT activation/phosphorylation, basally and in response to CyKILRb downregulation (Figure 5H; Figure S6). These data demonstrate that miR-3151-5p regulates the PIK3R2/PI3K/AKT axis and NSCLC clonogenicity as well as attenuates the ability of CyKILRb to affect PIK3R2 expression and AKT activation.Figure 5CyKILRb acts as a competing endogenous RNA for miR-3151-5p to relieve miRNA-dependent repression of PIK3R2(A) In silico prediction for miRNA binding for CyKILRb and PIK3R2 was performed by matching both the CyKILRb and PIK3R2 sequences with miRNAs downloaded from miRBase using either canonical seed sites (nt 2–8) or non-canonical binding sites (nt 3–9). miRNA “hits” for both RNA sequences were then compared and overlapping miRNAs are depicted as a Venn diagram, which was generated in R with the ggVennDiagram package. The top 10 miRs with the longest complementary and contiguous sequences for either PIK3R2 or CYKILRb are depicted. (B) Sequence alignment and positional mapping of the miR-3151-5p binding sites in CyKILRb and the PIK3R2 3′ UTR. (C) Schematic model illustrating the mechanistic hypothesis for CyKILRb competing with PIK3R2 for miR-3151-5p binding. (D–H). The depicted NSCLC cells were transfected with either control miRNA (Con), miR3151-5p mimic (mimic), or miR3151-5p inhibitor (inhibitor) ± vector control (Vec), scrambled control siRNA (SCR), CyKILRb expression plasmid (CYB), or CyKILRb siRNA (siCYB). (D) RNA levels of PIK3R2 and CyKILRb in the depicted NSCLC cells normalized to GAPDH mRNA levels as analyzed by RT-qPCR. (E) Clonogenic survival analysis. (F–H) SDS-PAGE and western immunoblotting for PIK3R2, phospho-AKT (P-AKT), total AKT (T-AKT), and GAPDH (for normalization). Data in (F–H) are representative of ≥3 biological replicates repeated on at least two separate occasions. Quantitative data in this figure are presented as dot plots (n ≥ 3 replicates repeated on at least two separate occasions), and statistical analyses were performed using an unpaired t test with Welch’s correction; ∗∗p < 0.01; ∗∗∗p < 0.001. For quantitation of western immunoblotting data depicted in (G) and (H), see Figure S6. CyKILRb acts as a competing endogenous RNA for miR-3151-5p to relieve miRNA-dependent repression of PIK3R2 (A) In silico prediction for miRNA binding for CyKILRb and PIK3R2 was performed by matching both the CyKILRb and PIK3R2 sequences with miRNAs downloaded from miRBase using either canonical seed sites (nt 2–8) or non-canonical binding sites (nt 3–9). miRNA “hits” for both RNA sequences were then compared and overlapping miRNAs are depicted as a Venn diagram, which was generated in R with the ggVennDiagram package. The top 10 miRs with the longest complementary and contiguous sequences for either PIK3R2 or CYKILRb are depicted. (B) Sequence alignment and positional mapping of the miR-3151-5p binding sites in CyKILRb and the PIK3R2 3′ UTR. (C) Schematic model illustrating the mechanistic hypothesis for CyKILRb competing with PIK3R2 for miR-3151-5p binding. (D–H). The depicted NSCLC cells were transfected with either control miRNA (Con), miR3151-5p mimic (mimic), or miR3151-5p inhibitor (inhibitor) ± vector control (Vec), scrambled control siRNA (SCR), CyKILRb expression plasmid (CYB), or CyKILRb siRNA (siCYB). (D) RNA levels of PIK3R2 and CyKILRb in the depicted NSCLC cells normalized to GAPDH mRNA levels as analyzed by RT-qPCR. (E) Clonogenic survival analysis. (F–H) SDS-PAGE and western immunoblotting for PIK3R2, phospho-AKT (P-AKT), total AKT (T-AKT), and GAPDH (for normalization). Data in (F–H) are representative of ≥3 biological replicates repeated on at least two separate occasions. Quantitative data in this figure are presented as dot plots (n ≥ 3 replicates repeated on at least two separate occasions), and statistical analyses were performed using an unpaired t test with Welch’s correction; ∗∗p < 0.01; ∗∗∗p < 0.001. For quantitation of western immunoblotting data depicted in (G) and (H), see Figure S6. CyKILRb augments the PIK3R2/PI3K/AKT axis and oncogenic phenotypes via competing for miR-3151-5p To verify that the regulation of PIK3R2, AKT activation, and NSCLC clonogenic survival by CyKILRb was through this lncRNA acting as a ceRNA for miR-3151-5p at nucleotides 284–290 (UGCCCCA) of CyKILRb (MRE-WT: MRE wildtype), a mutant construct was generated (CYB-MRE-Mut: UAGCGCU) using site-directed mutagenesis (Figure 6A). Expression of CYB-MRE-Mut, in contrast to CyKILRb-MRE-WT, was unable to upregulate PIK3R2, activate AKT, or induce tumor-promoting phenotypes (Figures 6A–6D; Figures S3B and S7). To verify that CyKILRb physically interacted as a ceRNA with miR-3151-5p, two assays were utilized. In the first assay, biotin-labeled miR-3151-5p mimics were utilized coupled to RNA “pull-down” experiments, which demonstrated that miR-3151-5p associated with both WT CyKILRb and PIK3R2 transcripts (Figure 6E). These data for CyKILRb were confirmed using a luciferase reporter assay by cloning 170 bp of CyKILRb sequence containing either the MRE-WT or the MRE-Mut into the psiCHECK2 luciferase reporter vector. Dual-luciferase activity was measured 48 h post-transfection in the NSCLC, which also demonstrated that miR-3151-5p associated with CyKILRb with an MRE-WT, but not CyKILRb-MRE-Mut (Figure 6F). To determine whether CyKILRb and PIK3R2 are regulated through a shared miRNA-mediated mechanism, Argonaute RNA immunoprecipitation (Ago-RIP) was performed following transfection of miR-3151-5p mimics. Overexpression of miR-3151-5p significantly increased the enrichment of both CyKILRb and PIK3R2 transcripts in Argonaute-containing RNA-induced silencing complex (RISC) complexes compared with control-transfected cells, indicating that both transcripts are targets of miR-3151-5p (Figure 6G; Figure S8). Notably, knockdown of CyKILRb further increased the enrichment of PIK3R2 in Argonaute immunoprecipitates, suggesting enhanced recruitment of PIK3R2 to miR-3151-5p-loaded RISC complexes in the absence of CyKILRb (Figure 6G; Figure S8). These findings support a competitive endogenous RNA (ceRNA) mechanism in which CyKILRb and PIK3R2 compete for binding to miR-3151-5p, thereby modulating the availability of miR-3151-5p for repression of PIK3R2.These data, when combined with the entirety of the data presented herein, demonstrate that CyKILRb is a ceRNA for miR-3151-5p, which regulates PIK3R2 expression and the PI3K/AKT pathway to drive oncogenic phenotypes such as NSCLC clonogenicity.Figure 6miR-3151-5p binds at 284–290 nt of CyKILRb and is required for CyKILRb to modulate PIK3R2 expression and oncogenic phenotypes(A–C) Vector control (Vec), CYB-MRE-WT (UGCCCCA) (MRE-WT), or CYB-MRE-Mut (UAGCGCU) (MRE-Mut) plasmids were transfected into H1299 or H1792 cells. At 48 h post-transfection, downstream assays including RT-qPCR, SDS-PAGE/western immunoblotting, and clonogenic survival assays (7–14 days) were employed. (A) RT-qPCR analysis of PIK3R2 and CyKILRb RNA levels normalized to GAPDH mRNA levels. (B) SDS-PAGE and western immunoblotting for PIK3R2, phospho-AKT (P-AKT), total AKT (T-AKT), and GAPDH (for normalization) (representative of three biological replicates). For quantitation of western immunoblotting data depicted in (B), see Figure S7. (C) Clonogenic survival assay. Data in (A)–(C) are presented as dot plots (n = 3 replicates repeated on at least two separate occasions), and statistical analysis was performed using one-way ANOVA followed by Tukey’s multiple comparisons test (n ≥ 3 biological replicates). ns, not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. (D) Endpoint tumor volumes and growth curves of H1299 xenografts stably expressing vector control (Vec), CyKILRb-MRE-WT (MRE-WT), or CYB-MRE-Mut (MRE-Mut). Disruption of the miR-3151-5p response element abolished the tumor-promoting effect of CyKILRb. Data are presented as mean ± SEM (n = 6 mice per group). Statistical significance was determined by one-way ANOVA; ∗p < 0.05; ∗∗p < 0.01. (E) To assess the physical interaction between CyKILRb or PIK3R2 and miR-3151-5p, a biotin-labeled miR-3151-5p mimic was synthesized and used for RNA “pull-down” experiments. PIK3R2 and CyKILRb RNA levels (% input) pulled down by biotinylated miR-3151-5p in H1299 and H1792 cells were analyzed RT-qPCR analysis for PIK3R2 and CyKILRb normalized to the input RNA. (F) The CyKILRb sequence (∼170 bp) containing the MRE (WT or Mut) was cloned into the psiCHECK2 luciferase reporter vector, and dual-luciferase activity was measured 48 h post-transfection in H1299 and H1792. Relative luciferase activity (Renilla/Firefly) in H1299 or H1792 cells co-transfected with miR-Con or miR-3151-5p mimic (miR-mimic) and luciferase reporter vectors (vector control [Vec], MRE-WT or MRE-Mut) is depicted. Renilla luciferase activity was normalized to Firefly luciferase activity to control for transfection efficiency. (G) H1299 cells were transfected with CyKILRb siRNA (siCYB), scrambled control siRNA (SCR), or miR-3151-5p mimic (miR-mimic). Ago2-associated RNAs were isolated by RNA immunoprecipitation (RIP) using an anti-Ago2 antibody or control IgG, followed by RT-qPCR analysis. RNA enrichment was calculated as fold change (AGO/IgG). Left graph: enrichment of CyKILRb in Ago2 immunoprecipitates relative to IgG controls. CyKILRb knockdown reduced the enrichment CyKILRb in the Ago2 complex, and miR-3151-5p expression increased the association of CyKILRb with Ago2. Right graph: enrichment of PIK3R2 mRNA in Ago2 immunoprecipitates relative to IgG controls is depicted. Both CyKILRb knockdown and miR-3151-5p expression significantly increased Ago2-associated PIK3R2, indicating enhanced miRNA-mediated targeting of PIK3R2 mRNA. Data depicted in (E)–(G) are presented as dot plots (n ≥ 3 replicates repeated on at least two separate occasions), and statistical analysis was performed using one-way ANOVA with Tukey’s multiple comparisons test. ns, not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. miR-3151-5p binds at 284–290 nt of CyKILRb and is required for CyKILRb to modulate PIK3R2 expression and oncogenic phenotypes (A–C) Vector control (Vec), CYB-MRE-WT (UGCCCCA) (MRE-WT), or CYB-MRE-Mut (UAGCGCU) (MRE-Mut) plasmids were transfected into H1299 or H1792 cells. At 48 h post-transfection, downstream assays including RT-qPCR, SDS-PAGE/western immunoblotting, and clonogenic survival assays (7–14 days) were employed. (A) RT-qPCR analysis of PIK3R2 and CyKILRb RNA levels normalized to GAPDH mRNA levels. (B) SDS-PAGE and western immunoblotting for PIK3R2, phospho-AKT (P-AKT), total AKT (T-AKT), and GAPDH (for normalization) (representative of three biological replicates). For quantitation of western immunoblotting data depicted in (B), see Figure S7. (C) Clonogenic survival assay. Data in (A)–(C) are presented as dot plots (n = 3 replicates repeated on at least two separate occasions), and statistical analysis was performed using one-way ANOVA followed by Tukey’s multiple comparisons test (n ≥ 3 biological replicates). ns, not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. (D) Endpoint tumor volumes and growth curves of H1299 xenografts stably expressing vector control (Vec), CyKILRb-MRE-WT (MRE-WT), or CYB-MRE-Mut (MRE-Mut). Disruption of the miR-3151-5p response element abolished the tumor-promoting effect of CyKILRb. Data are presented as mean ± SEM (n = 6 mice per group). Statistical significance was determined by one-way ANOVA; ∗p < 0.05; ∗∗p < 0.01. (E) To assess the physical interaction between CyKILRb or PIK3R2 and miR-3151-5p, a biotin-labeled miR-3151-5p mimic was synthesized and used for RNA “pull-down” experiments. PIK3R2 and CyKILRb RNA levels (% input) pulled down by biotinylated miR-3151-5p in H1299 and H1792 cells were analyzed RT-qPCR analysis for PIK3R2 and CyKILRb normalized to the input RNA. (F) The CyKILRb sequence (∼170 bp) containing the MRE (WT or Mut) was cloned into the psiCHECK2 luciferase reporter vector, and dual-luciferase activity was measured 48 h post-transfection in H1299 and H1792. Relative luciferase activity (Renilla/Firefly) in H1299 or H1792 cells co-transfected with miR-Con or miR-3151-5p mimic (miR-mimic) and luciferase reporter vectors (vector control [Vec], MRE-WT or MRE-Mut) is depicted. Renilla luciferase activity was normalized to Firefly luciferase activity to control for transfection efficiency. (G) H1299 cells were transfected with CyKILRb siRNA (siCYB), scrambled control siRNA (SCR), or miR-3151-5p mimic (miR-mimic). Ago2-associated RNAs were isolated by RNA immunoprecipitation (RIP) using an anti-Ago2 antibody or control IgG, followed by RT-qPCR analysis. RNA enrichment was calculated as fold change (AGO/IgG). Left graph: enrichment of CyKILRb in Ago2 immunoprecipitates relative to IgG controls. CyKILRb knockdown reduced the enrichment CyKILRb in the Ago2 complex, and miR-3151-5p expression increased the association of CyKILRb with Ago2. Right graph: enrichment of PIK3R2 mRNA in Ago2 immunoprecipitates relative to IgG controls is depicted. Both CyKILRb knockdown and miR-3151-5p expression significantly increased Ago2-associated PIK3R2, indicating enhanced miRNA-mediated targeting of PIK3R2 mRNA. Data depicted in (E)–(G) are presented as dot plots (n ≥ 3 replicates repeated on at least two separate occasions), and statistical analysis was performed using one-way ANOVA with Tukey’s multiple comparisons test. ns, not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001. A high ratio of CyKILRb/CyKILRa correlates with increased tumor mutational burden and higher recurrence risk As our data showed that the ratio of CyKILRb to CyKILRa modulated the clonogenicity and tumor growth of NSCLC cells, human tumor datasets were analyzed for clinical outcomes linked to the expression of this lncRNA. In this regard, a high CyKILRb/a RNA ratio significantly correlated with an increased risk of disease recurrence for NSCLC, specifically stage 1 lung adenocarcinomas (LUADs) (Figures S9A–S9C and S9E). Additionally, a significant correlation was observed when examining the ratio of CyKILRb/a RNA levels in NSCLC expressing the lncRNA as to the mutational burden of NSCLC cells (Figures S9D and S9E). These data infer that a high ratio of CyKILRb versus CyKILRa is associated with higher mutational burden and disease recurrence in NSCLC, and thus, the examination of specific RNA splicing events may serve as prognostic biomarkers for specific clinical parameters. Discussion Using an unbiased molecular approach, this study uncovered a novel oncogenic mechanism in NSCLC driven by the cytoplasmic lncRNA splice variant, CyKILRb. In particular, CyKILRb promoted clonogenicity and cellular proliferation by activating the PI3K/AKT signaling pathway, primarily through upregulation of phosphoinositide-3-kinase regulatory subunit 2 (PIK3R2). PIK3R2 encodes the regulatory subunit (p85β) of PI3K, a well-established mediator of signaling pathways crucial for cell growth and survival.23 Mutations in PIK3R2 have been linked to various developmental brain disorders, including megalencephaly-polymicrogyria-polydactyly-hydrocephalus (MPPH) syndrome and bilateral perisylvian polymicrogyria (BPP).24 In cancer, the role of PIK3R2 is understudied, but the factor is considered oncogenic, particularly in certain cancers such as glioblastoma multiforme.25 Our study supports a role for PIK3R2 as an oncogenic factor, and in the context of NSCLC and CyKILRb expression, a major driver of the PI3K/AKT pathway, which fosters key cancer cell phenotypes such a clonogenic survival. Thus, PIK3R2 expression is a plausible biomarker for constitutively activation of the PI3K/AKT pathway in the absence of a PI3K mutation. As this pathway is hyperactive in 50%–73% of NSCLC with major consequences on oncogenesis, the targeting of PIK3R2 and CYKLRb expression becomes a plausible therapeutic option.26,27,28 Additionally, the high expression of either factor may guide precision-based therapies for use of AKT inhibitors, which has been a large hurdle for determining the patient population that may benefit from these therapies.29 Mechanistically, this study identified a regulatory process for CyKILRb in modulating PIK3R2 expression. Specifically, CyKILRb acted as a ceRNA for a specific miR (i.e., miR-3151-5p) to enhance the expression of PIK3R2. The regulation of PIK3R2 by miRs has been previously reported. For example, miRNA-126 targets PIK3R2 and suppresses the proliferation of NSCLC cells.30 Enhanced PIK3R2 expression has also been linked to loss of miR-30a-5p induced by constitutively active EGFR signaling and the development of resistance to TKIs in NSCLC.31 More recently, co-targeting of miR-126-3p and miR-221-3p was shown to inhibit PIK3R2 reducing lung cancer growth and metastasis. Thus, one key regulatory mechanism for PIK3R2 is via miRs in NSCLC, which our study further corroborates. Indeed, our findings demonstrate that CyKILRb also acts as a ceRNA, in particular, for miR-3151-5p, a tumor suppressive miR, which is a direct regulator of PIK3R2. Only one other study has implicated a lncRNA in modulating miR-3151 and PIK3R2 levels, specifically, the lncRNA HOTAIR.32 Our study expands upon this work and shows that directly CyKILRb associates with miR-3151-5p and requires this association to induce PIK3R2 expression followed by subsequent activation of the PI3K/AKT pathway to enhance NSCLC clonogenicity. Therefore, our study substantiates a key and understudied role for miR-3151-5p in tumorigenicity. Additionally, the presented study showed that downregulation of CyKILRb concomitantly increased the expression of the tumor suppressors, p21 and p27 when the PIK3R2/PI3K/AKT axis was inactivated. These changes were associated with reduced cell proliferation and clonogenic survival. The link to p21 and p27 is highly relevant as these factors are well-known regulators of NSCLC cell proliferation.33,34,35,36,37 Our data support the activation of the PI3K/AKT pathway by CyKILRb as an upstream regulator of the expression of p21 and p27, which is in line with reports on the AKT pathway restraining the function of these two tumor suppressors.38 For example, the AKT pathway regulates p27 expression in NSCLC by phosphorylating and inactivating the FOXO transcription factor, which normally promotes p27 transcription. AKT can also directly phosphorylate p27, which promotes its cytoplasmic localization. This ability of hyperactivated AKT signaling in NSCLC to reduce nuclear p27 contributes to increased cell proliferation, growth, and malignant progression. In regard to p21, hyperactivated AKT signaling in NSCLC also leads to the downregulation of p21 contributing to increased cell proliferation in some instances.36,39 Similar to p27, AKT signaling can also influence p21 function through post-translational modifications. For example, phosphorylation of p21 by AKT induces cytoplasmic localization analogous to p27, which functionally inactivates it and promotes a proliferative effect.33 Thus, CyKILRb may be a key upstream and previously overlooked regulator of these pathways. Also of note in this study is that the ectopic expression of PIK3R2 did not fully restore the tumorigenic phenotypes, such as cellular proliferation or cisplatin resistance, when CyKILRb was downregulated. Furthermore, induction of RPS6KB2 by CyKILRb was essentially independent of PIK3R2. These data infer that while PIK3R2 is a key downstream effector of CyKILRb for inducing the PI3K/AKT pathway, it is not the sole mediator of CyKILRb-regulated mechanisms. Thus, CyKILRb may regulate additional targets or interact with RNA-binding proteins, mRNA stability factors, or other signaling molecules that contribute independently to oncogenesis via separate signaling pathways. For example, our laboratory recently reported that CyKILRb may act as an upstream rheostat for several important miRs (i.e., miR-424–3P and miR-503-3P) that control the growth and clonogenicity of NSCLC cells.15 Moreover, the unique RNA structure and cytoplasmic localization of CyKILRb could facilitate scaffolding or decoy functions that extend beyond PIK3R2 activation and are not recapitulated by PIK3R2 overexpression alone. In this regard, our pathway analysis of deep RNA-seq data also showed strong effects on other known oncogenic pathways such as the JAK-STAT (Janus Kinase–Signal Transducer and Activator of Transcription) pathway.40 CyKILRb also contains MREs for at least 88 miRs making it unlikely that CyKILRb only affects cellular physiologies via miR-3151-5p. Consequently, CyKILRb may act as a broad regulator of multiple survival and oncogenic pathways as a ceRNA for multiple miRs as well as indirectly inducing key survival pathways through as of yet, uncharacterized mechanisms. Thus, CyKILRb is an intriguing anti-cancer target in the context of specific oncogenotypes in NSCLC. Indeed, our study strongly infers that the exclusive expression of CyKILRb in NSCLC with WT and active STK11 and CDKN2A is required to activate specific survival pathways to overcome these tumor suppressive pathways and maintain a tumorigenic phenotype. In summary, our findings define CyKILRb as a critical modulator of the PI3K/AKT signaling cascade through increasing the expression of PIK3R2, an activator of PI3K. Additionally, CyKILRb was identified as a potent oncogenic driver in NSCLC through activation of this pathway via its ceRNA ability inhibiting tumor suppressive miRs (i.e., miR-3151-5p). This complex signaling paradigm emphasizes the multifaceted nature of lncRNA-mediated signaling, where cellular localization and specific RNA-seqs contribute to diverse functional outputs. Targeting CyKILRb or its broader network may offer novel therapeutic strategies for NSCLC patients, especially those tumors exhibiting hyperactivation of the PI3K/AKT pathway and possessing a WT CDKN2A and STK11 genotype. Materials and methods Cell lines and materials NCI-358 (RRID:CVCL_1559), NCI-H522 (RRID:CVCL_1513), NCI-H2009 (RRID:CVCL_1514), NCI-H1299 (RRID:CVCL_0060), and NCI-H1792 (RRID:CVCL_5895) were purchased from ATCC. SuperScript IV VILO Master Mix with ezDNase Enzyme (11766050), TaqMan Universal PCR Master Mix (4324018), Opti-MEM™ I Reduced Serum Medium (31985062), Lipofectamine RNAiMAX Transfection Reagent (13778150), Lipofectamine 3000 Transfection Reagent (L3000015), Dynabeads Streptavidin Magnetic Beads (11205D), GlycoBlue Coprecipitant (AM9515), TRIzol LS Reagent (10296010), and Pierce Fast Western Blot Kit (35050) were from Thermo Fisher Scientific. Cell Titer-Glo 2.0 Assay Kit and Dual-Luciferase Reporter Assay System were purchased from Promega (G9242 and E1910). Biotin labled miR-control and miR-3151-5p were synthesized from Horizon. The mycoplasma contamination detection kit was purchased from InvivoGen (rep-mys-50). Cell culture All cells were cultured following the procedures previously described and were utilized within 6 months or 20 passages, whichever came first, from the time of receipt and thawing.15 Throughout the entire study, cell lines were subjected to Mycoplasma testing every 2 months using a Mycoplasma contamination detection kit. Parental cell lines were authenticated by STR (short tandem repeat) profiling. RNA extraction and RT-qPCR Total RNA was isolated from cultured cells using RNeasy Mini Kit from QIAGEN (74104) following the manufacturer’s instructions. RNA quantification was performed using a NanoDrop NP-1000 spectrometer, and RNA (2 μg) was used for the first-strand cDNA (complementary DNA) synthesis using SuperScript IV VILO Master Mix with ezDNase Enzyme. Subsequently, quantitative reverse-transcription PCR (RT-qPCR) was conducted on a Quant Studio 5 instrument (Applied Biosystems, Carlsbad, CA, USA) as previously described.41,42,43,44,45 The housekeeping gene GAPDH was utilized for normalization using the ΔΔCt method. The detection of PIK3R2 or CyKILRb follows the methods a reported previously.15 siRNA, miR mimic/inhibitor, and plasmid transfection Cells were plated in a full medium without antibiotics 24 h before transfection. Cells were transfected at 60% confluency. siRNA, miR mimic, miR inhibitor, or plasmid DNA were diluted in Opti-MEM I Reduced Serum Medium. For siRNA, miR mimic, or miR inhibitor transfection, Lipofectamine RNAiMAX was employed as described previously,15 while Lipofectamine 3000 was used for plasmid transfection. The final concentration of siRNA was 20 nM in the cell culture medium. For CyKILRb transfection, 3 μg plasmid and 6 μL lipofectamine 3000 were used for 106 cells in 6 mL of cell culture medium. For PIK3R2 and ca-AKT transfection, 12 μg plasmid and 30 μL lipofectamine 3000 were used for 1–1.5 × 106 cells in 6 mL cell culture medium.46 PI3K and AKT inhibitor treatment Cells were cultured in full medium until reaching 60%–80% confluency before inhibitor treatment. Cells were treated with PI3K inhibitor (LY294002, Tocris) at 6 μM or AKT inhibitor (Akt Inhibitor VIII, Cayman) at 10 μM for 6 h for protein assay and 24 h for RNA analysis as described previously.44 Subsequently, the full medium was replaced with serum-free medium when treated with inhibitor. After treatment, full medium was replaced and culture for an additional 24–36 h before the proliferation assay. Proliferation analysis To investigate cell proliferation, the Cell Titer-Glo Luminescent Cell Viability kit was utilized following the manufacturer’s instructions. Cells were initially seeded in 96-well plates at a density of 10,000–20,000 cells per well in RPMI1640 supplemented with 10% fetal bovine serum (FBS). After treatment (single treatment for 96 h; double treatment for 48 h) with CyKILRb, PIK3R2, caAKT vectors or siRNA ± miR control, miR mimic or miR inhibitor (total of 96 h for the proliferation assay), the plates were carefully decanted and refilled with 50 μL of Cell Titer-Glo reagent and 50 μL of phosphate-buffered saline (PBS) per well. The plates were then incubated for 10 min at room temperature in the dark. Subsequently, the luminescence readings were taken using the Gene5 Multi-Mode Microplate Reader (Molecular Devices, Sunnyvale, CA, USA) with a 250 ms integration. Clonogenic survival assays After plasmid transfection or siRNA treatment, cells were harvested and re-suspended in complete growth media. Subsequently, the cells (200–400) were seeded onto 6-well plates and allowed to grow until visible colonies formed as done, which typically occurred within 7–21 days depending on cell type.46,47 Once the colonies became visible, they were fixed with cold 20% methanol, stained with 0.2% crystal violet in 20% methanol, washed, and then left to air-dry as previously described.15 For the cisplatin studies, H1299 cells were transfected with CyKILRb siRNA, PIK3R2 siRNA, CyKILRb overexpression plasmid, or the corresponding negative controls using the appropriate Lipofectamine transfection reagents (Thermo Fisher Scientific) according to the manufacturer’s instructions. For rescue experiments, cells were co-transfected with CyKILRb siRNA and a PIK3R2 expression plasmid. Twenty-four hours after transfection, cells were treated with either cisplatin (1 μM) or vehicle control (0.1% DMSO) for 12 h. The treatment medium was then removed and replaced with fresh complete growth medium. Cells were subsequently cultured for 7–10 days to allow colony formation, with the medium changed every 3 days. At the end of the incubation period, colonies were fixed with 20% methanol and stained with 0.2% crystal violet. Colonies containing more than 50 cells were counted manually. Colony numbers were normalized to the corresponding control group and expressed as mean ± SD from three independent experiments. Western immunoblotting Cells were lysed in RIPA (Radioimmunoprecipitation Assay buffer) buffer containing Halt Protease and Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific, 78440). For protein concentration determination, the Pierce BCA Protein Assay kit (Thermo Fisher Scientific, 23227) was used.48,49 Whole-cell lysates were mixed with Laemmli loading buffer, boiled, separated by 4%–20% SDS-PAGE, and transferred to a polyvinylidene fluoride (PVDF) membrane as previously described.50 Subsequently, immunoblot analyses were performed using the following primary antibodies: anti-PIK3R2/p85 (cat. no. 83606-5-RR; 1:2,000, RRID: AB_3671219), p21 (cat. no. 10355-1-AP; 1:1,000, RRID: AB_2077682), p27 (cat no. 25614-1-AP; 1:1,000, RRID: AB_2880161), RPS6KB2 (cat. no. 26194-1-AP; 1:500), and GAPDH (cat. no. 60004-1-Ig; 1:50,000, RRID: AB_2107436) from ProteinTech; and anti-phospho-AKT (Ser473; cat. no. 4060S; 1:1,000, RRID: AB_2315049) and AKT (pan; Cell cat. no. 8596S; 1:2,000, RRID: AB_10941384) from Signaling Technology. Briefly, immunoblotting for the aforementioned antibodies was performed using the Pierce Fast Western Blot Kit, ECL Substrate (Thermo Fisher Scientific, Cat. No. 35050) according to the manufacturer’s instructions. Membranes were rinsed with 1× Fast Western Wash Buffer (a pre-incubation with an additional blocking agent is not utilized with this system) and incubated with primary antibodies diluted in Fast Western Antibody Diluent overnight at 4°C with gentle agitation. Membranes were subsequently incubated with Fast Western Optimized HRP Reagent Working Dilution for 10 min at room temperature, followed by three washes (5 min each) with 1× Fast Western Wash Buffer. Immunoreactive bands were detected using Pierce ECL Detection Reagents (Pierce, 35050) and visualized with a ChemiDoc Imaging System (Bio-Rad Laboratories, Hercules, CA, USA). Biotin-labeled miR3151-5p pull-down assay At 1.2 × 106 H1299 or 2.0 × 106 H1792 cells were seeded in 10 cm plates and transfected at 70%–80% confluency with 40 nM biotin-labeled miR-3151-5p mimic (Biotin-miR-mimic) or a biotin-labeled negative control (Biotin-miR-Con) (Horizon) using Lipofectamine RNAiMAX (36 μL each plate) according to the manufacturer’s instructions. At 48 h post-transfection, cells were washed twice with ice-cold PBS and lysed in RIP lysis buffer (20 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% NP-40, 2 mM EDTA) supplemented with protease inhibitor cocktail and RNase inhibitor. Cell lysates were cleared by centrifugation at 12,000× g for 10 min at 4°C. The supernatants were incubated with pre-washed Streptavidin Magnetic Beads at 4°C for 2 h with gentle rotation to capture biotin-labeled miRNA-RNA complexes. The beads were then washed five times with wash buffer to minimize nonspecific interactions. Bound RNAs were eluted and purified using TRIzol reagent. The enrichment of CyKILRb and PIK3R2 transcripts in the pull-down fractions was quantified by RT-qPCR. For each sample, 10% of the cell lysate was reserved as input prior to streptavidin pull-down. RT-qPCR Ct values obtained from the pull-down fraction were normalized to the corresponding input RNA, and data are presented as percentage of input (% input). Dual-luciferase reporter assays Dual-luciferase reporter assays were performed using the psiCHECK-2 vector (Promega), which contains Renilla luciferase as the primary reporter and Firefly luciferase as an internal control. Wild-type or mutant fragments (170 bp of CyKILRb sequence) containing the predicted miR-3151-5p binding site (UGCCCCA) were cloned downstream of the Renilla luciferase gene in the psiCHECK-2 vector. H1299 (4 × 103) or H1792 (7 × 103) cells were seeded in each well of 96-well plates and co-transfected at ∼70%–80% confluency with 50 ng of psiCHECK-2 reporter plasmid (Vector Control, Wild type-UGCCCCA, or Mutant-UAGCGCU) ± 40 nM miR-3151-5p mimic or negative control miRNA using Lipofectamine 3000 according to the manufacturer’s instructions. At 48 h post-transfection, cells were lysed, and luciferase activities were measured using the Dual-Luciferase Reporter Assay System (Promega) following the manufacturer’s protocol. Renilla luciferase activity was normalized to Firefly luciferase activity to control for transfection efficiency. Data were expressed as relative luciferase activity of Renilla to Firefly. Transcriptomics and signaling pathway analyses RNA-seq was conducted using 150 bp paired-end reads, aiming for a depth of approximately 30 million reads.15 The sequencing data were filtered with SOAPnuke (v.1.5.6) (removing reads containing sequencing adapter and low-quality base ratio), then the clean reads were stored in FASTQ format.51 DEG analysis was performed using the DESeq2 with q value ≤0.05.52 KEGG enrichment analysis of annotated different expression gene was performed by R Phyper (v.120.0) based on hypergeometric test. The significant levels of terms and pathways were corrected by q value with a rigorous threshold (q value ≤0.05). For the comparative analyses of the CyKILRb/a ratio and NSCLC tumor parameters, raw RNA-seq (FASTQ) data and clinical annotations were obtained from the Cancer Genome Atlas (TCGA) Lung Adenocarcinoma (TCGA-LUAD; dbGaP accession phs000178) project through the National Cancer Institute Genomic Data Commons (GDC). Access was approved by the National Institutes of Health under project #38229 entitled, RNA splicing analysis in lung cancer. Somatic mutation data, including tumor mutational burden, were obtained from the corresponding TCGA-LUAD (PanCancer Atlas) study through cBioPortal for Cancer Genomics. Transcript abundance was quantified as transcripts per million (TPM) using Salmon v.1.10.2 against the GENCODE Release 46 (GRCh38.p14; Ensembl Release 112) reference transcriptome. Transcript expression, clinical variables, and mutation data were merged using matched TCGA sample identifiers. Tumors were stratified into quartiles according to the CyKILRb/CyKILRa TPM ratio. Associations with recurrence and tumor mutational burden were evaluated using Spearman’s rank correlation, Fisher’s exact test, and univariable logistic regression. Statistical analyses were performed in Python v.3.11.15 using pandas v.2.3.3, SciPy v.1.17.1, and statsmodels v.0.14.6. In silico analysis of miRNA response elements In silico prediction was performed by matching the CyKILRb and PIK3R2 RNA sequences with miRNAs downloaded from miRBase using either canonical seed sites (nt 2–8) or non-canonical binding sites (nt 3–9). Specifically, miRNA target sites were identified using a custom Python-based scanner. Target sequences were scanned for antiparallel Watson-Crick complementarity to human miRNA seed regions. Canonical sites were defined by perfect pairing to miRNA positions 2–8 and classified as 8 mer, 7 mer-m8, 7 mer-A1, or 6 mer based on seed length and the presence of an adenine immediately upstream of the target site. Offset seed sites initiating at miRNA position 3 or 4 were also evaluated and classified using the same definitions. Seed matches were required to meet specified pairing constraints with optional allowance for G:U wobble. All sites were evaluated on both strands of the target sequence, and sites were reported with positional and sequence context. Ago-RIP Ago-RIP was performed using the miRNA Target IP Kit (Active Motif, Cat. No. 25500) according to the manufacturer’s instructions. Briefly, scrambled, siCyKILRb or miR3151-5P mimic transfected H1299 cells were lysed in the supplied lysis buffer and incubated with protein G magnetic beads coupled to a pan-Argonaute antibody or control IgG. Following immunoprecipitation and washing, RNA associated with Argonaute-containing complexes was purified using the kit reagents. Purified RNA was reverse-transcribed, and the abundance of CyKILRb and PIK3R2 transcripts in the immunoprecipitates was quantified by RT-qPCR. Data were normalized to input RNA (%) and expressed as fold enrichment relative to the IgG control. Xenograft tumor growth assay All animal experiments were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Virginia. H1299 stable cell lines were generated by transfection of expression plasmids or small hairpin RNA (shRNA) constructs followed by puromycin selection for 7 days. The experimental groups included: (1) vector control, (2) CyKILRb WT overexpression, (3) CyKILRb-miR-3151-5p MRE mutant overexpression, (4) sh-scramble control plus vector control, (5) sh-CyKILRb plus vector control, (6) sh-scramble plus PIK3R2 overexpression, and (7) sh-CyKILRb plus PIK3R2 overexpression. Female NSG mice aged 6–8 weeks were randomly assigned to the seven experimental groups (n = 6 mice per group). Cells were harvested during logarithmic growth, washed twice with PBS, and resuspended in a 2:1 mixture of PBS and growth factor-reduced Matrigel. A total of 106 viable cells in 100 μL were injected subcutaneously into the right flank of each mouse. Mice were monitored regularly for tumor development. When tumors became palpable and reached approximately 100 mm3, tumor dimensions were measured three times per week using digital calipers. Tumor volume was calculated using the formula: V = (L × W2)/2, where V (volume, mm3) represents tumor volume, L (length, mm) represents the longest tumor dimension and W (width, mm) represents the dimension perpendicular to the length. Animals were euthanized when the largest tumor volumes reached approximately 1,000 mm3. Tumor growth curves were generated by plotting mean tumor volume ±SEM at the endpoint. Statistical analyses were performed using a repeated measures ANOVA and one-way ANOVA with Tukey’s multiple comparisons test. Statistical analysis Graphing and statistics were performed using Prism GraphPad (Prism Software, San Diego, CA, USA). The following models were used for statistical analysis as described, unpaired t test with Welch’s correction were used when only two experimental groups were being compared.53 Repeated measures ANOVA were used when analyzing comparison groups that contained a dependent variable that was measured several times over the course of the experiment.54 One-way ANOVA with Tukey’s multiple comparisons test were used when comparing three or more independent groups.55,56 All data reported as dotplot or mean (SD) where applicable; p < 0.05 was considered statistically significant. All bioinformatic analyses were conducted using the R program (ns, not significant; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001). To compare miRNA target sets, unique miRNA identifiers were extracted from each analysis and visualized using Venn diagrams generated in R with the ggVennDiagram package. Data and code availability All RNA-seq data are submitted to NCBI Sequence Read Archive (SRA) database for public access (submission no. SUB15688856, RRID: SCR_001370).