Colorectal cancer remains one of the most common and deadliest malignancies worldwide, responsible for well over 900,000 deaths each year according to global cancer statistics. While much attention has focused on protein-coding genes that drive tumor growth, an enormous portion of the human genome is transcribed into RNA molecules that never become proteins. Among these, long non-coding RNAs, or lncRNAs, have emerged as master regulators of cancer biology, fine-tuning how genes behave inside tumor cells. Now, a new study published in Molecular Biology Reports has turned the spotlight on a particularly mysterious subgroup: antisense lncRNAs, which are transcribed from the opposite DNA strand of their neighboring protein-coding genes. Researchers led by Aida Houshmand, Reza Safaralizadeh, and Mohammad Khalaj-Kondori at the University of Tabriz in Iran have examined three of these antisense transcripts, DIAPH3-AS1, HMMR-AS1, and SPATA3-AS1, in colorectal cancer patients, asking whether their expression patterns could reveal anything about diagnosis, tumor characteristics, or the response to chemotherapy.
The rationale for this investigation lies in the way antisense lncRNAs are thought to operate. Natural antisense transcripts can overlap their corresponding sense genes and regulate them through a variety of mechanisms, including interference with transcription, alteration of messenger RNA stability, and recruitment of epigenetic modifiers to chromatin. In cancer, this fine-tuning can tip the balance toward uncontrolled proliferation, invasion, or resistance to therapy. Well-known examples such as HOTAIR and H19 have already established that lncRNAs can act as oncogenic drivers in colorectal cancer, while others like HAND2-AS1 have been linked to resistance against 5-fluorouracil, the backbone drug of colorectal cancer chemotherapy. Yet the three transcripts selected by the Tabriz team had never been characterized in this disease. In fact, SPATA3-AS1 had never been studied in any type of cancer at all, making this the first look anywhere at its behavior in a malignancy.
To conduct the study, the researchers enrolled ninety patients with colorectal cancer and collected paired tumor biopsies, one taken before the start of capecitabine-based chemotherapy and another taken afterward, allowing each patient to serve as their own internal control. A cohort of seventy-one healthy individuals provided normal colorectal mucosa for comparison. Capecitabine is an oral prodrug that is converted in the body to 5-fluorouracil, a compound that disrupts DNA synthesis in rapidly dividing cells, so tracking molecular changes across the treatment window offered a rare opportunity to observe how the lncRNA landscape responds to therapy in living tumors. Expression levels of all three antisense lncRNAs were quantified using quantitative real-time polymerase chain reaction, a technique that amplifies and measures specific RNA sequences with high sensitivity, and the results were normalized to the expression of the β-actin housekeeping gene using the 2−ΔCt method.
The statistical framework of the study was rigorous and appropriately matched to the data. Because expression measurements in biomedical samples often deviate from a normal distribution, the team employed non-parametric tests: the Mann–Whitney U test for comparisons between independent groups and the Wilcoxon signed-rank test for paired comparisons within the same patients before and after treatment. To evaluate whether any of the three lncRNAs could serve as diagnostic markers, the researchers performed receiver operating characteristic, or ROC, analysis, which plots true-positive rates against false-positive rates across different expression thresholds and summarizes the discriminative power of a test through the area under the curve, known as the AUC. An AUC of 1.0 represents perfect discrimination, while an AUC of 0.5 indicates performance no better than random chance.
The headline finding was striking in its consistency: all three antisense lncRNAs were significantly overexpressed in pre-treatment colorectal cancer tissue compared with healthy mucosa. The fold-changes ranged from 2.22 to 2.87, meaning that tumor tissue carried roughly twice to nearly three times the amount of these transcripts as normal tissue, with p-values below 0.001. This uniform upregulation across three separate transcripts suggests that the antisense RNA dimension of colorectal cancer biology is more than a random byproduct of cellular chaos. It hints that transcription from the opposite strand of key cancer-related genes may be systematically activated during tumor development, potentially reflecting dysregulation of the very sense genes they overlap, including DIAPH3, which has been implicated in colorectal cancer progression through its role in maintaining degradation of the epidermal growth factor receptor, and HMMR, whose protein product, the hyaluronan-mediated motility receptor RHAMM, is known to promote invasion and dissemination of colorectal cancer cells.
When the researchers compared biopsies taken before and after chemotherapy, a fascinating divergence appeared. HMMR-AS1 expression dropped significantly after treatment, falling by a factor of 1.60 with a p-value of 0.0012, and, crucially, its level in post-treatment tumors became statistically indistinguishable from that of healthy controls, with a p-value of 0.133. This chemo-responsiveness means that the transcript is not a static feature of the tumor genome but a dynamic molecular readout that reflects the biological state of the tumor as therapy reshapes it. In contrast, DIAPH3-AS1 and SPATA3-AS1 remained elevated even after capecitabine-based treatment, suggesting that their upregulation is more deeply entrenched in the tumor’s identity and less perturbed by the cytotoxic assault of chemotherapy. The transcript-specific nature of this response is itself an important lesson: antisense lncRNAs in the same disease do not behave as a uniform class, and each must be evaluated on its own terms.
The study also examined whether expression of these transcripts correlated with clinicopathological features, including TNM stage, lymph node metastasis, tumor site, and treatment response classified by the RECIST criteria into complete response, partial response, and stable disease. For the most part, the answer was no. None of the three lncRNAs was associated with stage, nodal spread, anatomical location, or how well patients responded to chemotherapy. However, one association stood out: HMMR-AS1 was significantly higher in well and moderately differentiated tumors than in poorly differentiated ones, with a p-value of 0.0067. Tumor differentiation grade describes how closely cancer cells resemble their normal counterparts, with poorly differentiated tumors generally behaving more aggressively. The link between HMMR-AS1 and better differentiation therefore adds a nuanced layer, implying that this transcript tracks with a particular biological state of the tumor rather than simply marking advanced disease.
On the diagnostic front, the results were encouraging but tempered by realism. ROC analysis of pre-treatment samples yielded AUC values between 0.652 and 0.699 for the three lncRNAs, indicating modest but genuine ability to distinguish cancer tissue from healthy mucosa. These values fall short of the 0.9-plus territory that would mark a standalone clinical diagnostic test, but they position the transcripts as plausible components of a multi-marker panel, a strategy increasingly favored in oncology where combinations of biomarkers often outperform any single molecule. Notably, HMMR-AS1’s diagnostic performance declined after treatment, with its AUC dropping to 0.569, a change fully consistent with its demonstrated chemo-responsiveness. In other words, once chemotherapy had pushed the transcript’s expression back toward baseline, its power to flag tumor tissue evaporated, a pattern that simultaneously undermines its use as a monitoring marker during therapy and confirms the biological reality of the treatment effect.
What makes this study resonate beyond its immediate findings is the glimpse it offers into an underexplored layer of cancer genetics. Most genomic screens prioritize protein-coding sequences, yet antisense transcripts like the ones studied here may coordinate the behavior of oncogenes and tumor suppressors from the shadows. Previous work has shown that antisense lncRNAs participate in essentially every hallmark of cancer, from evading cell death to activating invasion and metastasis. The Tabriz team’s data add colorectal cancer to the list of malignancies where DIAPH3-AS1 and HMMR-AS1 are dysregulated, echoing earlier reports of HMMR-AS1 overexpression in lung adenocarcinoma and epithelial ovarian cancer, where it has been linked to proliferation, metastasis, and prognosis through its interaction with microRNA axes such as the miR-138/SIRT6 pathway.
The road from these findings to clinical application will require further validation. The ninety-patient cohort, while substantial for a molecular study, will need to be expanded across diverse populations, and the modest AUC values suggest these transcripts are best deployed in combination with other markers rather than alone. Still, the demonstration that a first-time-characterized transcript like SPATA3-AS1 is reliably elevated in colorectal tumors, and that HMMR-AS1 responds dynamically to chemotherapy, opens concrete avenues for future research: longitudinal studies tracking these RNAs in blood or exosomes, functional experiments to determine whether they drive tumor behavior or merely mirror it, and trials testing whether their expression at diagnosis predicts long-term outcomes. As the dark matter of the genome continues to yield secrets, antisense lncRNAs are proving that the most consequential conversations in a cancer cell may be the ones conducted between two strands of DNA, read out in RNA that never makes a protein but shapes everything the proteins do.
Subject of Research: Expression of antisense long non-coding RNAs DIAPH3-AS1, HMMR-AS1, and SPATA3-AS1 in colorectal cancer and their association with chemotherapy response and clinicopathological features
Article Title: Expression of long non‑coding RNAs DIAPH3‑AS1, HMMR‑AS1, and SPATA3‑AS1 in patients with colorectal cancer: association with chemotherapy response and clinicopathological features
Article References: Houshmand, A., Safaralizadeh, R., & Khalaj-Kondori, M. (2026). Expression of long non‑coding RNAs DIAPH3‑AS1, HMMR‑AS1, and SPATA3‑AS1 in patients with colorectal cancer: association with chemotherapy response and clinicopathological features. Molecular Biology Reports, 53(1), Article 1623. https://doi.org/10.1007/s11033-026-12789-5
Image Credits: AI Generated
DOI: 10.1007/s11033-026-12789-5
Keywords: colorectal cancer, long non-coding RNA, antisense transcript, DIAPH3-AS1, HMMR-AS1, SPATA3-AS1, capecitabine, chemotherapy response, biomarker, ROC analysis, tumor differentiation, qRT-PCR
Cite Scienmag News
APA MLA Chicago
Nathaniel Bowman. (September 26, 2026). Three Hidden RNA Molecules Revealed as New Clues in Colorectal Cancer Detection and Treatment. Scienmag. https://scienmag.com/three-hidden-rna-molecules-revealed-as-new-clues-in-colorectal-cancer-detection-and-treatment/
Nathaniel Bowman. “Three Hidden RNA Molecules Revealed as New Clues in Colorectal Cancer Detection and Treatment.” Scienmag, 26 September 2026, https://scienmag.com/three-hidden-rna-molecules-revealed-as-new-clues-in-colorectal-cancer-detection-and-treatment/. Accessed 26 September 2026.
Nathaniel Bowman. “Three Hidden RNA Molecules Revealed as New Clues in Colorectal Cancer Detection and Treatment.” Scienmag. September 26, 2026. https://scienmag.com/three-hidden-rna-molecules-revealed-as-new-clues-in-colorectal-cancer-detection-and-treatment/
Copy citation Download RIS
Tags: antisense lncRNAs in cancer diagnosisantisense transcriptbiomarkercapecitabinechemotherapy responseColorectal cancerDIAPH3-AS1gene expression regulation by antisense lncRNAsHMMR-AS1Long non-coding RNAlong non-coding RNAs as tumor regulatorsnon-coding RNA-based cancer therapiesnon-coding RNAs and chemotherapy responsenon-coding RNAs in colorectal cancerqRT-PCRRNA interference mechanisms in tumor progressionRNA molecules as biomarkers for colorectal cancerROC analysisrole of antisense transcripts in gene regulationSPATA3-AS1SPATA3-AS1 in colorectal cancertumor differentiation
