Spheroids and organoids to study long noncoding RNAs in cancer: how they could replace animal models Authors: Ana Maria Capela, João Pessoa, Hugo Arede, Bruno Bernardes de Jesus Published: 2026 ABSTRACT Spheroids, organoids, and further three-dimensional cellular models represent an intermediate stage of supra-cellular complexity between monolayered cell cultures and animal models. In the present review, we identified conditions in which spheroids and organoids could replace animal models in biomedical research, using long noncoding RNA (lncRNA) research in cancer as a study object. In tumor spheroids and patient-derived organoids, chemosensitivity can be enhanced through therapeutic overexpression or silencing of specific lncRNAs (depending on the lncRNA and context), which then decreases spheroid/organoid size or formation efficiency. These outcomes consistently mimic observations in xenograft mouse models, in which the same lncRNA intervention decreases tumor volume. Therefore, as a proxy of tumor growth assessment, animal models could be largely replaced with spheroids or organoids, taking advantage of their inexpensiveness and patient-specific features, respectively. The use of animal models could be restricted to bystander or pleiotropic effects, such as the assessment of parameters including metastasis and survival rates. Introduction Biomedical research is based on a set of research models with varying levels of complexity, including simpler ones, such as cell lines, or complex systems, such as animal models, before, and ultimately, translation to humans, if possible. Animal models properly represent the complexity of human patients, as opposed to the oversimplified nature of monolayered mammalian cell cultures. In this process, experimentation in animal models remains an essential step, despite the ethical issues, elevated costs, and facility requirements involved. As such, any strategies to decrease their usage should be considered. Three-dimensional (3D) in vitro cellular models, which include spheroids, organoids, “organs-on-chips”, and others, cannot fully replace animal models. Nevertheless, they provide an intermediate stage of supra-cellular complexity between monolayered two-dimensional (2D) cell cultures and animal models. Their use is key to decreasing the need for animal models in research, for which their exact potential needs to be assessed. A general replacement of the commonly used 2D cell culture method with 3D cellular models has been recommended (Jensen & Teng, 2020). Guidelines for the replacement of animal models with organoids and “organs-on-chips” have also been proposed (Hartung & Smirnova, 2025; Park et al., 2024). Several reviews have detailed the preparation, advantages, limitations, applications, and prospective developments of 3D cellular models (mostly spheroids and organoids) in cancer and other research fields (El Harane et al., 2023; Gunti et al., 2021; Živković & Opačak-Bernardi, 2025). Nevertheless, the exact conditions (and their limitations) under which 3D cellular models can replace animal models have not been uncovered yet. The present review identifies conditions under which spheroids and organoids could decrease the need for animal models in biomedical research. Here, we focused on the 3D cellular models that are most frequently used to study the roles of long noncoding RNAs (lncRNAs) in specific cancer types. lncRNAs are RNA molecules with a length longer than 200 nucleotides, which are not translated into proteins. Therefore, their function is directly determined by their sequence, structure, and subsequent associations with other RNAs, DNA, or proteins. The function of several lncRNAs has been detailed in several cellular processes, from cellular reprogramming to aging and cancer. Multiple lncRNAs, such as H19 imprinted maternally expressed transcript (H19), metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), HOX antisense RNA (HOTAIR), and noncoding RNA activated by DNA damage (NORAD), have been associated with poor prognosis in multiple cancers (Cai et al., 2014; Capela et al., 2024; Fang & Fullwood, 2016; S. Fu et al., 2020; Soghli et al., 2021; J. Yang et al., 2021). For instance, NORAD is upregulated in triple-negative breast cancer, and its silencing sensitizes cancer cells to chemotherapy (Alves-Vale et al., 2023; Capela et al., 2024). In the present review, we selected recently published original research articles in which at least one 3D cellular model was used to assess the impact of cellular levels of one lncRNA on the 3D model’s behavior and therapy response. Throughout this article, we use the terms “upregulation” and “downregulation” to indicate the association between 3D cellular model formation and its alterations in endogenous levels of a specific lncRNA, relative to a cognate 2D cellular model. Many of the cited studies utilized both an animal model and a 3D cellular model, in which we compared the outcomes obtained from the two models. With this approach, we uncovered conditions under which the 3D cellular model could replace the animal model. Spheroids: the simplest 3D cellular models The simplest 3D cellular models are spheroids, which in the literature can also be termed as spheres or cell spheroids. When prepared from specific cell types, they may also have specific designations, such as neurospheres (generated from neuronal cells) or mammospheres (generated from breast cells). When using cancer cells, the spheroids can be termed as tumorspheres, onco-spheroids, or multicellular tumor spheroids (MCTS) (Mitrakas et al., 2023). Spheroids are prepared through in vitro cell growth under conditions that disable their adherence to a surface (for example, by using ultra-low adherence [ULA] cell culture plasticware), or through other methods including hanging-drop, bioprinting, or spinner flasks. Under such conditions, the cells form intercellular interactions, generating a 3D aggregate resembling a sphere (Białkowska et al., 2020; Decarli et al., 2021). Spheroid preparation is relatively simple and inexpensive. When prepared from cancer cells, this 3D system can be regarded as a simplistic mimic of a tumor (Arora et al., 2024). Spheroids containing one or more cell types are termed homotypic or heterotypic, respectively, in which different cell types may not directly interact. In addition to the combination of different types of cells (i.e., cancer cells and cancer-associated fibroblasts [CAFs], cancer cells and immune cells, cancer cells, and endothelial cells), the complexity of spheroids can be increased by including extracellular matrix (ECM)-based biomaterials, which can promote the self-organization of the supracellular structures (Sousa et al., 2025). They can also be subjected to perfusion in microfluidic devices to study dynamic flow (to supply culture medium and remove waste products, to connect the medium—and therefore the secretome—of different wells growing different cells, among other goals) (Białkowska et al., 2020). The potential of this model is due to its versatility for customization, as exemplified above. Furthermore, spheroids can be used for testing the effects of drugs and alterations in lncRNA levels, and they can be personalized by using patient-derived cells. Table 1 summarizes lncRNA studies in spheroids. Collectively, these studies show that spheroids can have different levels of specific lncRNAs relative to the same cell line cultured in a 2D monolayer. Identical alterations could be observed in patient tissues relative to healthy individuals. They also demonstrate that lncRNAs are convenient molecules that modulate spheroid formation in multiple models of cancer. Table 1Studies that use spheroids to study lncRNAsCancer typelncRNASummaryRefAdrenocortical carcinomaH19 imprinted maternally expressed transcript (H19)H19 was ↓ expressed in adrenocortical carcinoma. Silencing H19 with siRNAs did not impact spheroid growth, but deacetylase inhibitors ↓ spheroid growth and ↑ H19 expression(Di Fazio et al., 2022)Bladder cancerFOXF1 adjacent non-coding developmental regulatory RNA (FENDRR)FENDRR was ↓ expressed in bladder cancer. FENDRR overexpression ↓ spheroid formation efficiency(Dai et al., 2025)Metastasis associated lung adenocarcinoma transcript 1 (MALAT1)MALAT1 was ↑ expressed in “leader” cells of patient-derived spheroids(Li et al., 2024c)Breast cancerAFAP1 antisense RNA 1 (AFAP1-AS1)AFAP1-AS1 was ↑ expressed in breast cancer tissues. AFAP1-AS1 silencing in spheroids grown in Matrigel ↓ hypoxia-induced 3D channel formation(García-Hernández et al., 2024)Deleted in lymphocytic leukemia 2 (DLEU2)DLEU2 was ↑ expressed in breast cancer cell lines and tissues, and was also ↑ expressed in spheroids. DLEU2 silencing ↓ spheroid formation efficiency(Islam et al., 2024)H19 imprinted maternally expressed transcript (H19)H19 expression in breast cancer was associated with poor prognosis. Silencing H19 ↓ spheroid formation efficiency(Shima et al., 2018)Long intergenic non-protein coding RNA 52 (LINC00052)LINC00052 was ↓ expressed in spheroids(Muñoz-Galindo et al., 2019)Long intergenic non-protein coding RNA 511 (LINC00511)Knockout of LINC00511 ↓ spheroid numbers, dimension and formation frequency number, as well as stemness levels(Azadbakht et al., 2022)LINC00511 was ↑ expressed in breast cancer tissues. LINC00511 overexpression ↑ the diameter of spheroids and stemness levels, and accelerated proliferation and invasion. Mutating LINC00511 (which ↓ the amount of wild-type LINC00511) ↓ spheroid diameter, stemness, proliferation and migration, and promoted apoptosis(Tan et al., 2023)Long intergenic non-protein coding RNA 520 (LINC00520)LINC00520 was ↑ expressed in human tumors. Silencing LINC00520 ↓ the number of cell protrusions and cell migration of spheroids in Matrigel(Henry et al., 2016)Long intergenic non-protein coding RNA 885 (LINC00885)LINC00885 was ↑ expressed in breast ductal carcinoma. Overexpressing LINC0085 in normal breast cells ↑ spheroid growth and branching(Abba et al., 2020)RP11-20F24.2 [ENSG00000234918]; SKAP1 antisense RNA 2 (SKAP1-AS2); RP11-206M11.7 [ENSG00000244468.1]; CTD-2566J3.1 [ENSG00000258837.1]; urothelial cancer associated 1 (UCA1); LASP1 neighbor (LASP1NB); long intergenic non-protein coding RNA 1535 (LINC01535); CTB-119C2.1 [ENST00000608362.1]; MIR924 host gene (MIR924HG); PRRT3 antisense RNA 1 (PRRT3-AS1); X inactive specific transcript (XIST); RP11-782C8.3 [ENSG00000232274.2]; cytoskeleton regulator RNA (CYTOR); MIR4435-2 host gene (MIR4435-2HG); CTD-2538C1.2 [ENSG00000267475.1]; long intergenic non-protein coding RNA 857 (LINC00857); RP11-425M5.7 [ENSG00000276603]; BLCAP apoptosis inducing factor antisense 1 (BLCAP-AS1); RP11-383J24.6 [ENSG00000237586]Extracellular matrix protein-enriched 3D cultures were used to identify differentially expressed lncRNAs. The identified ↑ expressed lncRNAs were as follows: RP11-20F24.2, SKAP1-AS2, also known as THRA1/BTR, RP11-206M11.7, CTD-2566J3.1, UCA1, LASP1NB also known as LINC00672, LINC01535, CTB-119C2.1 MIR924HG, and PRRT3-AS1. The identified ↓ expressed lncRNAs were as follows: XIST, RP11-782C8.3, CYTOR also known as LINC00152, MIR4435-2HG, CTD-2538C1.2, LINC00857, RP11-425M5.7, BLCAP-AS1, and RP11-383J24.6(Nuñez-Olvera et al., 2022)Cervical cancerAFAP1 antisense RNA 1 (AFAP1-AS1)CD44v6+ cells, which express more stemness-related genes, formed more and larger spheroids than CD44v6-. Silencing AFAP1-AS1 ↓ spheroid number and size(Xia et al., 2021)WWTR1 antisense RNA 1 (WWTR1-AS1)WWTR1-AS1 was ↑ expressed in cervical squamous cell carcinoma tissues. Overexpressing WWTR1-AS1 ↑ spheroid formation and the amount of CD133+ cells(Zhou et al., 2024)HOXA cluster antisense RNA 2 (HOXA-AS2), urothelial cancer associated 1 (UCA1), HOX transcript antisense RNA (HOTAIR), HIF1A antisense RNA 1 (HIF1A-AS1), imprinted in Prader-Willi syndrome (IPW)Spheroids were more resistant to cisplatin than 2D monolayer cells. Cisplatin treatment in secondary spheroids (which were fist cultured in spheroids, then in a 2D monolayer, and later in spheroids) ↓ secondary spheroid formation enriched in CSCs. HOXA-AS2, UCA1, HOTAIR, HIF1A-AS1 and IPW were ↑ expressed in cisplatin-treated spheroids. Downregulating HOXA-AS2, UCA1, HOTAIR, HIF1A-AS1 and IPW ↓ spheroid formation, thus enhancing the effect of cisplatin(M. Li et al., 2019)CholangiocarcinomaLon noncoding RNA Polycystic Kidney Disease 2-2-3 (Lnc-PKD2-2–3)Lnc-PKD2-2–3 was ↑ expressed in CCA tissues. In DAVID bioinformatics, expression of “lnc-PKD2-2–3 was correlated with several oncogenes and stemness-associated pathways, including DGAT2, GPAM, HEPACAM, ATP1A1, PAFAH1B3, LOC728342, STAB2, SH2D3A and GTF2I”. Overexpressing Lnc-PKD2-2-3 led to ↑ number of spheroids(Qiu et al., 2019)Colorectal cancerDLGAP1 antisense RNA 2 (DLGAP1-AS2)DLGAP1-AS2 was ↑ expressed in CD133+ rectal cancer stem cells. DLGAP1-AS2 silencing ↓ spheroid diameter(Xiao et al., 2022)H19 imprinted maternally expressed transcript (H19)H19 was ↑ expressed in colorectal cancer tissues. Silencing H19 ↓ the number of formed spheroids, while leukemia inhibitory factor ↑ the spheroid count(M. Zhu et al., 2024)Hypoxia-inducible factor-2α promoter upstream transcript (HIF2PUT)Silencing HIF2PUT ↓ spheroid formation and the self-renewal capability of cells(Yao et al., 2015)Long intergenic non-protein coding RNA 1315 (LINC01315)Silencing LINC01315 ↓ spheroid formation, and cultivation with CD133+/CD44+ cancer stem cell (CSC)-derived exosomes ↑ spheroid formation. However, simultaneous silencing of LINC01315 and CSC exosome cultivation reversed the effect of the exosomes on spheroid growth(Li, et al., 2022d) Metastasis associated lung adenocarcinoma transcript 1 (MALAT1)MALAT1 was ↑ expressed in colorectal cancer cells with high CD95 levels. Silencing of MALAT1 ↓ spheroid formation(Gao et al., 2024)TINCR ubiquitin domain containing (TINCR)Overexpression of TINCR ↑ radioresistance of spheroids, while silencing TINCR had the opposite effect(Kang et al., 2019)TMPO antisense RNA 1 (TMPO-AS1)TMPO-AS1 was ↑ expressed in colorectal cancer cells. Silencing TMPO-AS1 ↓ spheroid formation efficiency(Ye et al., 2022)Endometrial cancerLong intergenic non-protein coding RNA, regulator of reprogramming (LINC-ROR)↑ LINC-ROR expression was associated with the pluripotent state of endometrial cancer tissues. LINC-ROR silencing ↓ spheroid differentiation(Zhou et al., 2014)Esophageal squamous cell carcinomaProtein disulfide isomerase family A member 3 pseudogene 1 (PDIA3P1)Silencing PDIA3P1 ↓ spheroid formation and size, while PDIA3P1 overexpression had the opposite effects(T. Huang et al., 2024)(Sex determining region Y)-box 2 overlapping transcript (SOX2OT)SOX2OT was ↑ expressed in esophageal squamous cell carcinoma tissues and spheroids, as compared to 2D monolayer cells. SOX2OT silencing ↓ spheroid number and size, and ↑ sensibility to docetaxel treatment(Haghi et al., 2022)Gastric cancerADAMTS9 antisense RNA 2 (ADAMTS9-AS2)ADAMTS9-AS2 was ↓ expressed in gastric cancer tissues. ADAMTS9-AS2 overexpression ↑ spheroid number(F. Wang et al., 2020)GliomaLong intergenic non-protein coding RNA, regulator of reprogramming (LINC-ROR)LINC-ROR was ↓ expressed in glioma tissues. Overexpressing LINC-ROR ↓ spheroid number, while LINC-ROR silencing had the opposite effect(Feng et al., 2015)Nuclear paraspeckle assembly transcript 1 (NEAT1)NEAT1 was ↑ expressed in glioblastoma multiform spheroids(Yoon et al., 2020)Prostate cancer associated transcript 1 (PCAT1)PCAT1 was ↑ expressed in glioma CSCs. Silencing PCAT1 ↓ spheroid formation, while PCAT1 overexpression had the opposite effect(Zhang et al., 2019a)Small nucleolar RNA host gene 9 (SNHG9)SNHG9 was ↑ expressed in glioma stem cells, as compared to glioma cells. SNHG9 overexpression ↑ spheroid formation efficiency, while SNHG9 silencing had the opposite effect(Wang et al., 2022a)HOXD cluster antisense RNA 2 (HOXD-AS2), Long intergenic non-protein coding RNA 1116 (LINC01116)HOXD was ↑ expressed in glioma cells and glioblastoma tissues. Silencing HOXD-AS2 or LINC01116 ↓ spheroid formation(Deforzh et al., 2022)Hepatocellular carcinomaICAM4 antisense RNA 1 (ICAM4-AS1)ICAM4-AS1 (also known as ICR) was ↑ expressed in PVTT (portal vein tumor thrombus, common in liver cancer) tissues and hepatocellular carcinoma cell lines. Silencing ICAM4-AS1 ↓ spheroid formation(W. Guo et al., 2016)Lung cancerDifferentiation antagonizing non-protein coding RNA (DANCR)DANCR was ↑ expressed in non-small cell lung cancer (NSCLC). Silencing DANCR ↓ spheroid size(Nicolescu et al., 2022)HOX transcript antisense RNA (HOTAIR)HOTAIR was ↑ expressed in a wide variety of cancers, including lung cancer. Collagen I supplementation in lung cancer spheroids ↑ HOTAIR expression(Y. Zhuang et al., 2013)HOTAIR was ↑ expressed in cisplatin-resistant groups of patients and in cisplatin-resistant cell lines, and also correlated with the pathological grade. Overexpressing HOTAIR ↑ spheroid formation, while silencing HOTAIR had the opposite effect(M. Y. Liu et al., 2016)JPX (just proximal to XIST) transcript, XIST activator (JPX)JPX was ↑ expressed in lung adenocarcinoma tissues. Overexpressing JPX ↑ the number of cells in spheroids(Mosca et al., 2024)lncRNA-AC026356.1 [ENSG00000274964]lncRNA-AC026356.1 was ↑ expressed in chemo-resistant lung cancer cells, and in spheroids. lncRNA-AC026356.1 silencing ↓ spheroid number(Zhang et al., 2023c)Long intergenic non-protein coding RNA 1224 (LINC01224)LINC01224 was ↑ expressed in irradiation-resistant NSCLC cells. LINC01224 silencing ↓ spheroid formation(Fu et al., 2021b)Long intergenic non-protein coding RNA 1232 (LINC01232)LINC01232 was ↑ expressed in NSCLC cells. Silencing LINC01232 ↓ spheroid number and formation(L. Zhu et al., 2022)PITPNA antisense RNA 1 (PITPNA-AS1)PITPNA‐AS1 was ↑ expressed in lung squamous cell carcinoma . PITNAS-AS1 overexpression ↑ spheroid formation(B. hao Peng et al., 2022)MelanomaJUN inducer (JUNI) or JUN divergent transcript (JUN DT)JUNI was ↑ expressed in cancer cells after stress induction. Doxorubicin treatment ↓ cell survival of JUNI-silenced spheroids(Kumar et al., 2024)OsteosarcomaHypoxia-inducible factor-2α promoter upstream transcript (HIF2PUT)HIF2PUT expression was correlated with stem cell-like properties and CSCs of tumors. Silencing HIF2PUT ↑ spheroid formation rate(Y. Wang et al., 2015)Ovarian cancerHLA complex P5 (HCP5)HCP5 knockout ↓ spheroid formation and size(Moradi et al., 2024)Long intergenic non-protein coding RNA 958 (LINC00958)LINC00958 was ↑ expressed in ovarian cancer cells. Cells with higher basal levels of LINC00958 had ↑ sphere formation ability; however, LINC00958 silencing reversed this increase. Co-culture with exosomes from the cell line with the highest LINC00958 expression also ↑ the spheroid formation ability of the cells with lowest LINC00958 expression. However, silencing LINC00958 in the highest LINC00958 expressing cell line reverted the previous effect(Yan et al., 2024)Pancreatic cancerLong stress-induced non-coding transcript 5 (LSINCT5)Patients with low LSINCT5 had ↑ survival rates. LSINCT5 knockout reduced spheroid formation, number, and size(Y. Dai et al., 2020)Papillary thyroid carcinomaH19 imprinted maternally expressed transcript (H19)H19 was ↑ expressed in papillary thyroid carcinoma stem cells and papillary thyroid carcinoma tissues. H19 silencing ↓ spheroid formation, while H19 overexpression had the opposite effect(Li et al., 2018a)Prostate cancerH19 imprinted maternally expressed transcript (H19)H19 was ↑ expressed in prostate cancer cells enriched in stem cell properties, and even more ↑ expressed in prostate cancer cell lines and other cancer cell lines. Overexpression of H19 ↑ spheroid number and size, while H19 silencing had the opposite effect(Roy et al., 2015)NUTM2A antisense RNA 1 (NUTM2A-AS1)NUTM2A-AS1 was ↑ expressed in NSCLC and prostate cancer. Silencing NUTM2A-AS1 ↓ spheroid number(H. Han et al., 2024) Table 1Studies that use spheroids to study lncRNAsCancer typelncRNASummaryRefAdrenocortical carcinomaH19 imprinted maternally expressed transcript (H19)H19 was ↓ expressed in adrenocortical carcinoma. Silencing H19 with siRNAs did not impact spheroid growth, but deacetylase inhibitors ↓ spheroid growth and ↑ H19 expression(Di Fazio et al., 2022)Bladder cancerFOXF1 adjacent non-coding developmental regulatory RNA (FENDRR)FENDRR was ↓ expressed in bladder cancer. FENDRR overexpression ↓ spheroid formation efficiency(Dai et al., 2025)Metastasis associated lung adenocarcinoma transcript 1 (MALAT1)MALAT1 was ↑ expressed in “leader” cells of patient-derived spheroids(Li et al., 2024c)Breast cancerAFAP1 antisense RNA 1 (AFAP1-AS1)AFAP1-AS1 was ↑ expressed in breast cancer tissues. AFAP1-AS1 silencing in spheroids grown in Matrigel ↓ hypoxia-induced 3D channel formation(García-Hernández et al., 2024)Deleted in lymphocytic leukemia 2 (DLEU2)DLEU2 was ↑ expressed in breast cancer cell lines and tissues, and was also ↑ expressed in spheroids. DLEU2 silencing ↓ spheroid formation efficiency(Islam et al., 2024)H19 imprinted maternally expressed transcript (H19)H19 expression in breast cancer was associated with poor prognosis. Silencing H19 ↓ spheroid formation efficiency(Shima et al., 2018)Long intergenic non-protein coding RNA 52 (LINC00052)LINC00052 was ↓ expressed in spheroids(Muñoz-Galindo et al., 2019)Long intergenic non-protein coding RNA 511 (LINC00511)Knockout of LINC00511 ↓ spheroid numbers, dimension and formation frequency number, as well as stemness levels(Azadbakht et al., 2022)LINC00511 was ↑ expressed in breast cancer tissues. LINC00511 overexpression ↑ the diameter of spheroids and stemness levels, and accelerated proliferation and invasion. Mutating LINC00511 (which ↓ the amount of wild-type LINC00511) ↓ spheroid diameter, stemness, proliferation and migration, and promoted apoptosis(Tan et al., 2023)Long intergenic non-protein coding RNA 520 (LINC00520)LINC00520 was ↑ expressed in human tumors. Silencing LINC00520 ↓ the number of cell protrusions and cell migration of spheroids in Matrigel(Henry et al., 2016)Long intergenic non-protein coding RNA 885 (LINC00885)LINC00885 was ↑ expressed in breast ductal carcinoma. Overexpressing LINC0085 in normal breast cells ↑ spheroid growth and branching(Abba et al., 2020)RP11-20F24.2 [ENSG00000234918]; SKAP1 antisense RNA 2 (SKAP1-AS2); RP11-206M11.7 [ENSG00000244468.1]; CTD-2566J3.1 [ENSG00000258837.1]; urothelial cancer associated 1 (UCA1); LASP1 neighbor (LASP1NB); long intergenic non-protein coding RNA 1535 (LINC01535); CTB-119C2.1 [ENST00000608362.1]; MIR924 host gene (MIR924HG); PRRT3 antisense RNA 1 (PRRT3-AS1); X inactive specific transcript (XIST); RP11-782C8.3 [ENSG00000232274.2]; cytoskeleton regulator RNA (CYTOR); MIR4435-2 host gene (MIR4435-2HG); CTD-2538C1.2 [ENSG00000267475.1]; long intergenic non-protein coding RNA 857 (LINC00857); RP11-425M5.7 [ENSG00000276603]; BLCAP apoptosis inducing factor antisense 1 (BLCAP-AS1); RP11-383J24.6 [ENSG00000237586]Extracellular matrix protein-enriched 3D cultures were used to identify differentially expressed lncRNAs. The identified ↑ expressed lncRNAs were as follows: RP11-20F24.2, SKAP1-AS2, also known as THRA1/BTR, RP11-206M11.7, CTD-2566J3.1, UCA1, LASP1NB also known as LINC00672, LINC01535, CTB-119C2.1 MIR924HG, and PRRT3-AS1. The identified ↓ expressed lncRNAs were as follows: XIST, RP11-782C8.3, CYTOR also known as LINC00152, MIR4435-2HG, CTD-2538C1.2, LINC00857, RP11-425M5.7, BLCAP-AS1, and RP11-383J24.6(Nuñez-Olvera et al., 2022)Cervical cancerAFAP1 antisense RNA 1 (AFAP1-AS1)CD44v6+ cells, which express more stemness-related genes, formed more and larger spheroids than CD44v6-. Silencing AFAP1-AS1 ↓ spheroid number and size(Xia et al., 2021)WWTR1 antisense RNA 1 (WWTR1-AS1)WWTR1-AS1 was ↑ expressed in cervical squamous cell carcinoma tissues. Overexpressing WWTR1-AS1 ↑ spheroid formation and the amount of CD133+ cells(Zhou et al., 2024)HOXA cluster antisense RNA 2 (HOXA-AS2), urothelial cancer associated 1 (UCA1), HOX transcript antisense RNA (HOTAIR), HIF1A antisense RNA 1 (HIF1A-AS1), imprinted in Prader-Willi syndrome (IPW)Spheroids were more resistant to cisplatin than 2D monolayer cells. Cisplatin treatment in secondary spheroids (which were fist cultured in spheroids, then in a 2D monolayer, and later in spheroids) ↓ secondary spheroid formation enriched in CSCs. HOXA-AS2, UCA1, HOTAIR, HIF1A-AS1 and IPW were ↑ expressed in cisplatin-treated spheroids. Downregulating HOXA-AS2, UCA1, HOTAIR, HIF1A-AS1 and IPW ↓ spheroid formation, thus enhancing the effect of cisplatin(M. Li et al., 2019)CholangiocarcinomaLon noncoding RNA Polycystic Kidney Disease 2-2-3 (Lnc-PKD2-2–3)Lnc-PKD2-2–3 was ↑ expressed in CCA tissues. In DAVID bioinformatics, expression of “lnc-PKD2-2–3 was correlated with several oncogenes and stemness-associated pathways, including DGAT2, GPAM, HEPACAM, ATP1A1, PAFAH1B3, LOC728342, STAB2, SH2D3A and GTF2I”. Overexpressing Lnc-PKD2-2-3 led to ↑ number of spheroids(Qiu et al., 2019)Colorectal cancerDLGAP1 antisense RNA 2 (DLGAP1-AS2)DLGAP1-AS2 was ↑ expressed in CD133+ rectal cancer stem cells. DLGAP1-AS2 silencing ↓ spheroid diameter(Xiao et al., 2022)H19 imprinted maternally expressed transcript (H19)H19 was ↑ expressed in colorectal cancer tissues. Silencing H19 ↓ the number of formed spheroids, while leukemia inhibitory factor ↑ the spheroid count(M. Zhu et al., 2024)Hypoxia-inducible factor-2α promoter upstream transcript (HIF2PUT)Silencing HIF2PUT ↓ spheroid formation and the self-renewal capability of cells(Yao et al., 2015)Long intergenic non-protein coding RNA 1315 (LINC01315)Silencing LINC01315 ↓ spheroid formation, and cultivation with CD133+/CD44+ cancer stem cell (CSC)-derived exosomes ↑ spheroid formation. However, simultaneous silencing of LINC01315 and CSC exosome cultivation reversed the effect of the exosomes on spheroid growth(Li, et al., 2022d) Metastasis associated lung adenocarcinoma transcript 1 (MALAT1)MALAT1 was ↑ expressed in colorectal cancer cells with high CD95 levels. Silencing of MALAT1 ↓ spheroid formation(Gao et al., 2024)TINCR ubiquitin domain containing (TINCR)Overexpression of TINCR ↑ radioresistance of spheroids, while silencing TINCR had the opposite effect(Kang et al., 2019)TMPO antisense RNA 1 (TMPO-AS1)TMPO-AS1 was ↑ expressed in colorectal cancer cells. Silencing TMPO-AS1 ↓ spheroid formation efficiency(Ye et al., 2022)Endometrial cancerLong intergenic non-protein coding RNA, regulator of reprogramming (LINC-ROR)↑ LINC-ROR expression was associated with the pluripotent state of endometrial cancer tissues. LINC-ROR silencing ↓ spheroid differentiation(Zhou et al., 2014)Esophageal squamous cell carcinomaProtein disulfide isomerase family A member 3 pseudogene 1 (PDIA3P1)Silencing PDIA3P1 ↓ spheroid formation and size, while PDIA3P1 overexpression had the opposite effects(T. Huang et al., 2024)(Sex determining region Y)-box 2 overlapping transcript (SOX2OT)SOX2OT was ↑ expressed in esophageal squamous cell carcinoma tissues and spheroids, as compared to 2D monolayer cells. SOX2OT silencing ↓ spheroid number and size, and ↑ sensibility to docetaxel treatment(Haghi et al., 2022)Gastric cancerADAMTS9 antisense RNA 2 (ADAMTS9-AS2)ADAMTS9-AS2 was ↓ expressed in gastric cancer tissues. ADAMTS9-AS2 overexpression ↑ spheroid number(F. Wang et al., 2020)GliomaLong intergenic non-protein coding RNA, regulator of reprogramming (LINC-ROR)LINC-ROR was ↓ expressed in glioma tissues. Overexpressing LINC-ROR ↓ spheroid number, while LINC-ROR silencing had the opposite effect(Feng et al., 2015)Nuclear paraspeckle assembly transcript 1 (NEAT1)NEAT1 was ↑ expressed in glioblastoma multiform spheroids(Yoon et al., 2020)Prostate cancer associated transcript 1 (PCAT1)PCAT1 was ↑ expressed in glioma CSCs. Silencing PCAT1 ↓ spheroid formation, while PCAT1 overexpression had the opposite effect(Zhang et al., 2019a)Small nucleolar RNA host gene 9 (SNHG9)SNHG9 was ↑ expressed in glioma stem cells, as compared to glioma cells. SNHG9 overexpression ↑ spheroid formation efficiency, while SNHG9 silencing had the opposite effect(Wang et al., 2022a)HOXD cluster antisense RNA 2 (HOXD-AS2), Long intergenic non-protein coding RNA 1116 (LINC01116)HOXD was ↑ expressed in glioma cells and glioblastoma tissues. Silencing HOXD-AS2 or LINC01116 ↓ spheroid formation(Deforzh et al., 2022)Hepatocellular carcinomaICAM4 antisense RNA 1 (ICAM4-AS1)ICAM4-AS1 (also known as ICR) was ↑ expressed in PVTT (portal vein tumor thrombus, common in liver cancer) tissues and hepatocellular carcinoma cell lines. Silencing ICAM4-AS1 ↓ spheroid formation(W. Guo et al., 2016)Lung cancerDifferentiation antagonizing non-protein coding RNA (DANCR)DANCR was ↑ expressed in non-small cell lung cancer (NSCLC). Silencing DANCR ↓ spheroid size(Nicolescu et al., 2022)HOX transcript antisense RNA (HOTAIR)HOTAIR was ↑ expressed in a wide variety of cancers, including lung cancer. Collagen I supplementation in lung cancer spheroids ↑ HOTAIR expression(Y. Zhuang et al., 2013)HOTAIR was ↑ expressed in cisplatin-resistant groups of patients and in cisplatin-resistant cell lines, and also correlated with the pathological grade. Overexpressing HOTAIR ↑ spheroid formation, while silencing HOTAIR had the opposite effect(M. Y. Liu et al., 2016)JPX (just proximal to XIST) transcript, XIST activator (JPX)JPX was ↑ expressed in lung adenocarcinoma tissues. Overexpressing JPX ↑ the number of cells in spheroids(Mosca et al., 2024)lncRNA-AC026356.1 [ENSG00000274964]lncRNA-AC026356.1 was ↑ expressed in chemo-resistant lung cancer cells, and in spheroids. lncRNA-AC026356.1 silencing ↓ spheroid number(Zhang et al., 2023c)Long intergenic non-protein coding RNA 1224 (LINC01224)LINC01224 was ↑ expressed in irradiation-resistant NSCLC cells. LINC01224 silencing ↓ spheroid formation(Fu et al., 2021b)Long intergenic non-protein coding RNA 1232 (LINC01232)LINC01232 was ↑ expressed in NSCLC cells. Silencing LINC01232 ↓ spheroid number and formation(L. Zhu et al., 2022)PITPNA antisense RNA 1 (PITPNA-AS1)PITPNA‐AS1 was ↑ expressed in lung squamous cell carcinoma . PITNAS-AS1 overexpression ↑ spheroid formation(B. hao Peng et al., 2022)MelanomaJUN inducer (JUNI) or JUN divergent transcript (JUN DT)JUNI was ↑ expressed in cancer cells after stress induction. Doxorubicin treatment ↓ cell survival of JUNI-silenced spheroids(Kumar et al., 2024)OsteosarcomaHypoxia-inducible factor-2α promoter upstream transcript (HIF2PUT)HIF2PUT expression was correlated with stem cell-like properties and CSCs of tumors. Silencing HIF2PUT ↑ spheroid formation rate(Y. Wang et al., 2015)Ovarian cancerHLA complex P5 (HCP5)HCP5 knockout ↓ spheroid formation and size(Moradi et al., 2024)Long intergenic non-protein coding RNA 958 (LINC00958)LINC00958 was ↑ expressed in ovarian cancer cells. Cells with higher basal levels of LINC00958 had ↑ sphere formation ability; however, LINC00958 silencing reversed this increase. Co-culture with exosomes from the cell line with the highest LINC00958 expression also ↑ the spheroid formation ability of the cells with lowest LINC00958 expression. However, silencing LINC00958 in the highest LINC00958 expressing cell line reverted the previous effect(Yan et al., 2024)Pancreatic cancerLong stress-induced non-coding transcript 5 (LSINCT5)Patients with low LSINCT5 had ↑ survival rates. LSINCT5 knockout reduced spheroid formation, number, and size(Y. Dai et al., 2020)Papillary thyroid carcinomaH19 imprinted maternally expressed transcript (H19)H19 was ↑ expressed in papillary thyroid carcinoma stem cells and papillary thyroid carcinoma tissues. H19 silencing ↓ spheroid formation, while H19 overexpression had the opposite effect(Li et al., 2018a)Prostate cancerH19 imprinted maternally expressed transcript (H19)H19 was ↑ expressed in prostate cancer cells enriched in stem cell properties, and even more ↑ expressed in prostate cancer cell lines and other cancer cell lines. Overexpression of H19 ↑ spheroid number and size, while H19 silencing had the opposite effect(Roy et al., 2015)NUTM2A antisense RNA 1 (NUTM2A-AS1)NUTM2A-AS1 was ↑ expressed in NSCLC and prostate cancer. Silencing NUTM2A-AS1 ↓ spheroid number(H. Han et al., 2024) Studies that use spheroids to study lncRNAs The molecular mechanisms that mediate spheroid formation through alterations in lncRNA levels are starting to be uncovered. In general, increased spheroid formation is correlated with increased stemness, which could be a driving force. For example, cytoplasmic lncRNA H19 has oncogenic properties in several cancer types, associated with increased stemness, through mechanisms that involve microRNA inhibition (J. Yang et al., 2021). In papillary thyroid carcinoma, H19 expression enhances cancer stem-like properties (Li et al., 2018a). In breast cancer, H19 inhibits the let-7 microRNA, causing an increase in the levels of its target, the LIN28 RNA-binding protein (F. Peng et al., 2017). In prostate cancer, silencing of H19 decreased both spheroid formation and the expression of the stemness markers octamer-binding transcription factor 4 and SRY-box transcription factor 2 (Roy et al., 2015). Long intergenic non-protein coding RNA, regulator of reprogramming (LINC-ROR), also located primarily in the cytoplasm, was upregulated in esophageal squamous cell carcinoma and positively correlated with stemness marker SRY-box transcription factor 9; their interaction was mediated by microRNAs (L. Wang et al., 2017). In endometrial cancer stem cells, LINC-ROR was also positively correlated with stemness. The miR-145 microRNA could downregulate LINC-ROR and inhibit spheroid formation (Zhou et al., 2014). However, LINC-ROR was downregulated in glioma tissues, and its expression was negatively correlated with the stemness marker Krüppel-like factor 4 (Feng et al., 2015). For this lncRNA, increased stemness was also correlated with increased spheroid formation. HOTAIR, which functions in both the cytoplasm and nucleus, is upregulated in colorectal cancer, where its silencing inhibits stemness and tumorigenicity by inhibiting the miR-211-5p microRNA and upregulating the fms-like tyrosine kinase-1 (Y. Huang et al., 2021. In non-small cell lung cancer, both HOTAIR and stemness markers were upregulated (M. Y. Liu et al., 2016). In ovarian cancer, HOTAIR maintains stemness by targeting the miR-206 microRNA and upregulating the T-box transcription factor 3 (Y. Zhang et al., 2020). HOTAIR is another lncRNA whose upregulation increases both stemness and spheroid formation. Additional lncRNAs have been shown to promote stemness and spheroid formation, such as MALAT1, a predominantly nuclear lncRNA, that enhances triple-negative breast cancer tumorigenesis and stemness by downregulating DNA methyltransferase 1 through the miR-137 microRNA and the B-cell lymphoma/leukemia 11 A transcription factor (Hu et al., 2023). These findings indicate that the H19, LINC-ROR, MALAT1, and HOTAIR lncRNAs promote spheroid formation by enhancing stemness through gene regulation networks involving microRNA inhibition and protein upregulation. Spheroids can have reduced sensitivity to chemotherapeutic drugs, relative to identical 2D-cultured cells. For example, cervical cancer spheroids present a higher resistance to cisplatin treatment than the same 2D-cultivated cells. In comparison with 2D-cultured cells, cervical cancer spheroids had increased levels of lncRNAs, including homeobox A cluster antisense RNA 2 (HOXA-AS2), urothelial carcinoma associated 1 (UCA1), HOTAIR, hypoxia-inducible factor 1 alpha antisense RNA 1 (HIF1A-AS1), and imprinted in Prader-Willi syndrome (IPW). Silencing of these lncRNAs enhanced the cytotoxicity of cisplatin, resulting in decreased spheroid formation (M. Li et al., 2019). Furthermore, in melanoma, silencing of JUN inducer (JUNI) increased spheroid sensitivity to doxorubicin (Kumar et al., 2024). In esophageal squamous cell carcinoma, the SOX2 overlapping transcript (SOX2OT) lncRNA was upregulated in spheroids, relative to 2D cultures. SOX2OT silencing decreased the number and size of spheroids, where it enhanced docetaxel toxicity (Haghi et al., 2022). These examples demonstrate that the therapeutic silencing of lncRNAs, whose upregulation is associated with enhanced spheroid formation, may enhance their sensitivity to chemotherapeutic agents. Such enhanced sensitivity can be generally assessed by monitoring any decrease in spheroid size. Alterations of the lncRNA levels can also change the differentiation stage of spheroids. For example, spheroids, generated from patient’s endometrial cancer samples cultured in stem cell medium and differentiated upon basic fibroblast growth factor (bFGF) depletion, had lower levels of LINC-ROR. On the contrary, overexpression of LINC-ROR inhibited spheroid differentiation (Zhou et al., 2014). Spheroids can be dissociated and re-assembled in secondary spheroids. Spheroid formation is widely used to generate cellular models enriched in cancer stem cells (CSCs). Consequently, they can be used to infer self-renewal capabilities and stemness. For example, in esophageal squamous cell carcinoma, some stemness markers were upregulated in spheroids (Haghi et al., 2022). In addition, both the size and number of breast cancer spheroids were increased in the presence of the stemness marker leukemia inhibitory factor (M. Zhu et al., 2024). Therefore, studying spheroid formation and growth is also useful to assess the stemness properties of cancer cells. In addition to lncRNA silencing, spheroid size can be altered by treatment with specific small molecules, or lncRNA-interacting proteins, resulting in similar effects on spheroid size. Treatment of adrenocortical carcinoma spheroids with deacetylase inhibitors induced H19 overexpression and reduced their size (Di Fazio et al., 2022). Spheroid formation can also be manipulated by adding complex mixtures that affect lncRNA expression. In ovarian cancer, the secretome of mesenchymal stem cells (MSCs) derived from the amniotic fluid or the chorionic villi caused a shift in lncRNA expression and decreased spheroid formation (Vaiasicca et al., 2024). These approaches provide alternative ways to influence lncRNA expression and effects on spheroid formation and size. Cancer spheroids can mimic ex vivo tumor growth Multiple studies have combined cancer cell spheroids with xenografts in mouse models, showing consistent results in their responses to lncRNA manipulation. Specific examples of studies that have been published in several cancer types are summarized in Table 2. Overall, these studies indicate that tumor growth can be correlated, and even represented, by spheroid growth, in multiple cancer types. Their global analysis supports that in vitro spheroids could robustly mimic in vivo tumors regarding their response to the manipulation of lncRNA levels. Nevertheless, properties including metastasis and animal survival still require the use of animal models. Further research is needed to determine the parallelism between these models and how they can accurately represent time and dose-response in cancer. It is expected that study complementation with these 3D models may decrease the number of animals needed and restrict the use of animal models to properties that cannot be mimicked in vitro. Table 2Studies that use spheroids and xenografts to study lncRNAsCancer typelncRNASummaryRefBladder cancerBrain cytoplasmic RNA 1 (BCYRN1)BCYRN1 seemed ↑ expressed in bladder cancer. BCYRN1 silencing ↓ spheroid formation, as well as ↓ tumor weight and tumor volume in xenografted mice(Arima et al., 2024)Breast cancerColon cancer-associated transcript 2 (CCAT2)CCAT2 was ↓ expressed in tumor tissues from luminal breast cancer patients. CCAT2 overexpression ↓ the number and diameter of spheroids, and the diameter of tumors in xenografted mice(Xie et al., 2023)Differentiation antagonizing non-protein coding RNA (DANCR)DANCR was ↑ expressed in triple negative breast cancer cell lines. Silencing of DANCR ↓ the size of the networks formed by spheroids growing in Matrigel and the size of tumors in xenografted mice(Nicolescu et al., 2023)H19 imprinted maternally expressed transcript (H19)H19 was ↑ expressed in breast cancer tumors and cell lines. H19 overexpression ↑ the number and diameter of spheroids, while H19 silencing ↓ tumor volume in xenografted mice(F. Peng et al., 2017)H19 was ↑ expressed in breast cancer xenograft tumors and cell lines. Silencing of H19 ↓ the number and diameter of spheroids, while H19 overexpression ↑ tumor volume in xenografted mice(F. Peng et al., 2018)HOX transcript antisense RNA (HOTAIR)HOTAIR was ↑ expressed in breast cancer tissues. HOTAIR silencing ↓ spheroid formation and tumor size in xenografted mice, and ↑ survival rates of xenografted mice(R. Han et al., 2022)Long intergenic non-protein coding RNA 52 (LINC00052)LINC00052 was ↓ expressed in patients with worse prognosis. LINC00052 expression ↓ with spheroid formation. LINC00052 silencing did not affect spheroid formation, but ↑ cell migration in a zebrafish xenograft model(Sanchez-Lopez et al., 2021)Long intergenic non-protein coding RNA 3000 (LINC03000)LINC03000, also known as LncMat2B was ↑ expressed in cisplatin-resistant breast cancer cells. LINC03000 overexpression ↓ spheroid growth and survival rates in zebrafish(Garcia-Venzor et al., 2020)MBNL1 antisense RNA 1 (MBNL1-AS1)MBNL1-AS1 was ↓ expressed in breast cancer tissues and cell lines. ↓ expression of MBNL1-AS1 was associated with ↓ survival rates in patients. MBNL1-AS1 silencing ↑ spheroid number and diameter, and tumor size in xenografted mice, while its overexpression had the opposite effects(Ding et al., 2022)Metastasis associated lung adenocarcinoma transcript 1 (MALAT1)MALAT1 was ↑ expressed in triple-negative breast cancer tissues and cell lines. Silencing MALAT1 ↓ spheroid formation and tumor volume in xenografted mice, while MALAT1 overexpression ↑ tumor volume in this mouse model(Hu et al., 2023)TOMM22 divergent transcript (TOMM22-DT)TOMM22-DT, also known as Lnc408, was ↑ expressed in breast cancer stem cells. Silencing of TOMM22-DT ↓ the number of spheroids and tumor volume in xenografted mice, while TOMM22-DT overexpression had the opposite effects(Wen et al., 2021)X inactive specific transcript (XIST)XIST was ↑ expressed in breast cancer cell lines. XIST silencing ↓ the number of spheroids and their size, as well as tumor volume in xenografted mice(Ma et al., 2023)Colorectal cancerB4GALT1 antisense RNA 1 (B4GALT1-AS1)B4GALT1-AS1 was ↑ expressed in colon cancer cell lines. B4GALT1-AS1 silencing ↓ the number of spheroids and the size of tumors in xenografted mice(Y. Zhang et al., 2019a, b)BACE1 antisense RNA (BACE1-AS)BACE1-AS was ↑ expressed in colorectal cancer tissues and high expression of BACE1-AS was associated with ↓ patient survival rates. BACE1-AS overexpression ↑ the number and diameter of spheroids. BACE1-AS knockout ↓ the number and diameter of spheroids, as well as the number of metastatic foci in the livers of xenografted mice(Wang et al., 2023b)Breast cancer antiestrogen resistance 4 (BCAR4)BCAR4 was ↑ expressed in colorectal cancer tumors. Overexpression of BCAR4 ↑ sphere formation efficiency. In xenografted mice, BCAR4 silencing ↓ tumor weight and volume(Ouyang et al., 2019)CERS6 antisense RNA 1 (CERS6-AS1)CERS6-AS1 was ↑ expressed in colorectal cancer tissues. CERS6-AS1 silencing ↓ spheroid formation efficiency and tumor weight in xenografted mice(Zhao et al., 2022)GATA2 antisense RNA 1 (GATA2-AS1)GATA2-AS1 was ↑ expressed in colon adenocarcinoma tissues. Silencing of GATA2-AS1 ↓ the number of spheroids, as well as the tumor weight and volume in xenografted mice(Pan et al., 2022)HOX transcript antisense RNA (HOTAIR)HOTAIR was ↑ expressed in colorectal cancer tumors, and high expression of HOTAIR was associated with ↓ patient survival rates. HOTAIR silencing ↓ spheroid formation and tumor volume in xenografted mice(Y. Huang et al., 2021)Long intergenic non-protein coding RNA 1315 (LINC01315)LINC01315 was ↑ expressed in colon adenocarcinoma tissues. LINC01315 silencing ↓ the number of formed spheroids, and tumor size and weight in xenografted mice. LINC01315 overexpression ↑ spheroid formation(Li et al.., 2022c)Pvt1 oncogene (PVT1)PVT1 was ↑ expressed in colon cancer tissues and high expression of PVT1 was associated with ↓ patient survival rates. PVT1 overexpression ↑ the number of spheroids and tumor volume in xenografted mice(Lai et al., 2021)Testis development related 1 (TDRG1)TDRG1 was ↑ expressed in colorectal cancer cells. Silencing of TDRG1 ↓ spheroid size and number, as well as tumor formation in xenografted mice(Hong et al., 2022)Cutaneous squamous cell carcinomaPvt1 oncogene (PVT1)PVT1 was ↑ expressed in cutaneous squamous cell carcinoma tissues. PVT1 knockout ↓ spheroid area, as well as the volume and senescence levels in xenografted mouse tumors(Li et al., 2024b)Esophageal squamous cell carcinomaLong intergenic non-protein coding RNA, regulator of reprogramming (LINC-ROR)LINC-ROR was ↑ expressed in esophageal squamous cell carcinoma tissues. LINC-ROR silencing, as well as cisplatin, ↓ spheroid formation, as well as tumor mass and volume in xenografted mice(L. Wang et al., 2017)Gastric cancerASB16 antisense RNA 1 (ASB16-AS1)ASB16-AS1 was ↑ expressed in gastric cancer tissues and cells. ASB16-AS1 silencing ↓ spheroid formation efficiency, as well as tumor weight and volume in xenografted mice(Fu et al., 2021a)Long intergenic non-protein coding RNA 3145 (LINC03145)LINC03145, also known as CRART16, was ↑ expressed in gastric cancer tissues. LINC03145 overexpression ↑ number and sprouting of spheroids, tumor size in xenografted mice, and microvessel formation in chick chorioallantoic membrane assays(Zhang et al., 2022a)Long intergenic non-protein coding RNA 924 (LINC00924)LINC00924 was ↑ expressed in gastric cancer tissues, and high LINC00924 expression was associated with lower survival rates of patients. LINC00924 overexpression ↑ spheroid formation and peritoneal metastasis in mice. LINC00924 silencing ↓ spheroid formation and prevented tumor growth in xenografted mice(He et al., 2022b)WT1 antisense RNA (WT1-AS)WT1-AS was ↓ expressed in gastric cancer stem cells as compared to non-stem gastric cancer cells. WT1-AS overexpression ↓ spheroid size and tumor size in xenografted mice. Overexpression of WT1-AS ↓ tumor growth, stemness, and metastasis in mouse xenografts(Zhang et al., 2023b)GliomaFOXD2 adjacent opposite strand RNA 1 (FOXD2-AS1)FOXD2-AS1 was ↑ expressed in glioma tissues, and ↑ levels of FOXD2-AS1 were associated with ↓ patient survival rates. Silencing of FOXD2-AS1 ↓ spheroid formation and tumor weight in xenografted mice(Wang et al., 2022c)Hepatocellular carcinoma up-regulated long non-coding RNA (HULC)HULC was ↑ expressed in glioblastoma tissues. HULC silencing ↓ spheroid formation, as well as tumor weight and volume in xenografted mice. HULC overexpression ↑ tumor weight in xenografted mice(Li et al., 2022b)Nuclear paraspeckle assembly transcript 1 (NEAT1)NEAT1 was ↑ expressed in glioma cells upon infection with adenovirus. Silencing of NEAT1 ↓ the number of spheroids. Injection of spheroids in mice was associated with larger tumors, as compared to the injection of 2D-cultured cells. Inoculation of mice with higher numbers of cells in spheroids also correlated with ↓ survival rates(Zang et al., 2020)Small nucleolar RNA host gene 10 (SNHG10)SNHG10 was ↑ expressed in glioma cells. Silencing of SNHG10 ↓ spheroid formation efficiency, as well as tumor volume and weight in xenografted mice(Jin et al., 2020)Head and neck squamous cell carcinomaVentricular heart development associated lncRNA (VHRT)VHRT, also known as MASCC1, was differently expressed in metastatic head and neck squamous cell carcinoma, as compared to non-metastatic tissues of the same cancer type. Low VHRT expression was associated with ↑ patient survival. Silencing of VHRT ↓ spheroid weight and volume, as well as proliferation markers in xenografted mouse tumors. Overexpression of VHRT ↑ spheroid formation, as well as tumor formation and lymph node metastasis in xenografted mice(Wang et al., 2023c)Hypopharyngeal squamous cell carcinomaLong intergenic non-protein coding RNA 662 (LINC00662)LINC00662 was ↑ expressed in hypopharyngeal squamous cell carcinoma tissues, and patients with high expression of LINC00662 had ↓ survival rates. LINC00662 overexpression promoted spheroid formation, and ↑ tumor volume in xenografted mice(B. Zhang & Ye, 2024)Liver cancerFGD5 antisense RNA 1 (FGD5-AS1)FGD5-AS1 was ↑ expressed in hepatocellular carcinoma tissues and cell lines, being correlated with tumor stage and size. Patients with higher FGD5-AS1 expression showed ↓ survival. FGD5-AS1 silencing ↓ spheroid diameter and volume, as well as the weight, stemness, and proliferation of xenografted mouse tumors. FGD5-AS1 overexpression ↑ spheroid diameter(He et al., 2022a)FRMD6 antisense RNA 1 (FRMD6-AS1)FRMD6-AS1 was ↑ expressed in hepatocellular carcinoma tissues and cell lines. High expression of FRMD6-AS1 was correlated with ↓ survival of patients. Overexpression of FRMD6-AS1 ↑ spheroid formation and tumor size and weight in xenografted mice(Sun et al., 2023)KCNQ1 opposite strand/antisense transcript 1 (KCNQ1OT1)KCNQ1OT1 was ↑ expressed in hepatocellular carcinoma tissues, and high expression of KCNQ1OT1 was correlated with lower patient survival rates. KCNQ1OT1 knockout ↓ the number of spheroids and tumor volume in xenografted mice(Majumdar et al., 2024)Lipoprotein(a) like 2 (pseudogene) (LPAL2)LPAL2 was ↓ expressed in hepatocellular carcinoma tissues and CD133+ cells Low expression of LPAL2 was correlated with lower patient survival. LPAL2 silencing accelerated spheroid formation and ↑ spheroid size, as well as tumor growth and weight in xenografted mice(Lin et al., 2022)lncRNA regulator of Akt signaling associated with HCC and RCC (LNCARSR)LNCARSR, also known as LncRNA ARSR, was ↑ expressed in cisplatin-resistant hepatocellular carcinoma, in CSCs from primary liver cultures, and in spheroids derived from human primary hepatocellular carcinoma cells, relative to 2D cultures. When plated again in 2D-settings, the CSCs from trypsinized spheroids had ↓ expression of LNCARSR. Silencing of LNCARSR ↓ spheroid number and size, and resulted in a smaller population of CSC in hepatome cells when injected into mice. LNCARSR was also associated with chemoresistance in patients(Yang et al., 2019a, b)Long intergenic non-protein coding RNA 261 (LINC00261)LINC00261 was ↓ expressed in TGFB1-induced epithelial-mesenchymal transition hepatocellular carcinoma cell lines. Overexpression of LINC00261 ↓ spheroid number and size and ↓ tumor weight and growth rate in xenografted mice(Chen & Xiang, et al., 2022b)Long intergenic non-protein coding RNA 680 (LINC00680)LINC00680 was ↑ expressed in hepatocellular carcinoma tissues and cell lines, was correlated with the tumor stage and size, and patients with high expression of LINC00680 had lower survival rates. Overexpression of LINC00680 ↑ primary and secondary spheroid formation. Silencing LINC00680 ↓ primary and secondary spheroid formation, and, in xenografted mice, modified stemness markers and enhanced antitumor potential of the drug tested(Shu et al., 2021)OIP5 antisense RNA 1 (OIP5-AS1)OIP5-AS1 was ↑ expressed in hepatoblastoma cells. Silencing OIP5-AS1 ↓ the number of spheroids and tumor weight in xenografted mice(W. Jiang et al., 2022)Small nucleolar RNA host gene 3 (SNHG3)SNHG3 was ↑ expressed in hepatocellular carcinoma primary tumors. SNHG3 silencing ↓ spheroid formation efficiency and ↑ survival rates in xenografted mice(Guo et al., 2024a)Small nucleolar RNA host gene 5 (SNHG5)SNHG5 was ↑ expressed in spheroids, as compared to 2D monolayer cells. Silencing SNHG5 ↓ spheroid formation and ↓ tumor growth, volume, and weight in xenografted mice. Spheroids were enriched in CSCs and had ↑ expression of SNHG5(Li & Hu et al., 2022a)Small nucleolar RNA host gene 9 (SNHG9)SNHG9 was ↑ expressed in hepatocellular carcinoma tissues and cell lines, including CD133+ hepatocellular carcinoma CSC cells. Overexpressing SNHG9 ↑ spheroid formation of CD133- cells. Silencing SNHG9 ↓ spheroid formation of CD133+ cells, as well as tumor volume and weight in xenografted mice(Yang et al., 2023b)Lung cancerADAMTS9 antisense RNA 1 (ADAMTS9-AS1)ADAMTS9-AS1 was ↓ expressed in most cancer types (including lung adenocarcinoma), in both tissues and cell lines. Overexpressing ADAMTS9-AS1 ↓ spheroid formation as well as tumor growth and weight in xenografted mice(Wang et al., 2023a)ASAP1 intronic transcript 1 (ASAP1-IT1)ASAP1-IT1 was ↑ expressed in NSCLC tissues and cell lines. It was correlated with tumor differentiation, tumor node metastasis (TNM), and lymph node metastasis (LNM). ASAP1-IT1 was also ↑ expressed in spheroids, as compared to 2D monolayer cells. Overexpressing ASAP1-IT1 ↑ spheroid number. Silencing ASAP1-IT1 ↓ spheroid formation and tumor growth in xenografted mice(Liu et al., 2021a)GPAT4 and GINS4 antisense RNA 1 (GPAT4-AS1)GPAT4-AS1 ,also known as GIAT4RA, was ↓ expressed in lung tissues (both adenocarcinoma and squamous carcinoma) and associated with higher survival of patients. GPAT4-AS1 was also associated with tumor size, clinical stage, and lymphatic metastasis. Overexpressing GPAT4-AS1 ↓ spheroid size and number, as well as tumor growth and weight in xenografted mice. Silencing GPAT4-AS1 had the opposite effects(Yang et al., 2019b)Long intergenic non-protein coding RNA 1419 (LINC01419)LINC01419 was ↑ expressed in lung adenocarcinoma tissues and cell lines. Overexpressing LINC01419 ↑ spheroid formation. Silencing LINC01419 ↓ tumor growth, weight, volume, and proliferative levels in xenografted mice(Chen et al., 2022a)Long intergenic non-protein coding RNA 467 (LINC00467)LINC00467 was ↑ expressed in lung adenocarcinoma tissues and cell lines, and high expression of LINC00467 was associated with ↓ survival. Silencing LINC00467 ↓ spheroid number and tumor size and weight, as well as ↓ proliferation and ↑ apoptosis levels in xenografted mice(Chang & Yang, 2020)MNX1 antisense RNA 1 (head to head) (MNX1-AS1)MNX1-AS1 was ↑ expressed in lung tissues, including lung adenocarcinoma, and also in CSCs. Silencing MNX1-AS1 ↓ spheroid formation of CSCs and the tumor growth rate and weight in xenografted mice(Lv et al., 2023)OIP5 antisense RNA 1 (OIP5-AS1)OIP5-AS1 was ↑ expressed in lung cancer cells, and was also ↑ expressed in spheroids, as compared to 2D monolayer cultures. Silencing OIP5-AS1 ↓ spheroid number and size. Overexpressing OIP5-AS1 ↑ spheroid number and size, as well as tumor formation rates in in vivo experiments(Mao & Li, 2022)SOX2 overlapping transcript (SOX2OT)SOX2OT was ↑ expressed in lung tumors and resistant cell lines. Patients with high SOX2OT expression had ↓ survival. SOX2OT was also ↑ expressed in CSC-derived spheroids, as compared to 2D monolayer cells. CSCs isolated from spheroids had ↑ capacity to initiate solid tumor formation in vivo, as compared to 2D monolayer cells. Silencing SOX2OT ↓ spheroid formation and, when combined with GL1T repression, tumor growth in xenografted mice also ↓(Dong et al., 2023)Testis development related 1 (TDRG1)TDRG1 was ↑ expressed in non-small cell lung cancer cells. Silencing TDRG1 ↓ spheroid number and size. When inoculated into mice it ↓ tumor formation rate and stem cell frequency. Overexpressing TDRG1 had the opposite effects(Lu et al., 2023)MedulloblastomaLOXL1 antisense RNA 1 (LOXL1-AS1)LOXL1-AS1 was ↑ expressed in medulloblastoma tumors and cell lines. Patients with ↑ expressed LOX1-AS1 had ↓ survival rates. LOXL1-AS1 silencing ↓ spheroid formation, which was ↑ under LOXL1-AS1 overexpression. Xenografted mice with LOXL1-AS overexpression had ↓ survival rates(Do et al., 2024)Oral squamous cell carcinomaTINCR ubiquitin domain containing (TINCR)TINCR was ↓ expressed in oral squamous cell carcinoma tissues, specifically in poorly differentiated tissues, and patients with low expression of TINCR had ↓ survival. TINCR was also positively correlated with the pathological differentiation stage. TINCR was also ↓ expressed in spheroids, as compared to 2D monolayers, a process that was associated with dedifferentiation. Overexpression of TINCR ↓ tumor size, weight, and volume in xenografted mice(Z. Zhuang et al., 2021)OsteosarcomaB4GALT1 antisense RNA 1 (B4GALT1-AS1)B4GALT1-AS1 was ↑ expressed in osteosarcoma tissues, cell lines, and spheroids, as compared to 2D monolayer cells. Silencing B4GALT1-AS1 ↓ spheroid size and number, and, in xenografted mice, ↓ tumor size, weight, seeding ability, and formation rate(Li et al., 2018b)WAC antisense RNA1 (WAC-AS1)WAC-AS1 was ↑ expressed in most cancer types (including osteosarcoma), specifically, in metastatic tissues, as compared to non-metastatic and normal tissues. WAC-AS1 expression was also correlated with lung metastasis, and patients with high expression of WAC-AS1 had ↓ survival rates. Overexpressing WAC-AS1 ↑ spheroid number, as well as tumor growth and metastasis in xenografted mice(Yang et al., 2023c)Ovarian cancerHOX transcript antisense RNA (HOTAIR)HOTAIR was ↑ expressed in ovarian cancer (specifically in stages I/II) and was correlated with tumor stage. Plasma circulating levels of HOTAIR showed high specificity and sensitivity as a diagnostic biomarker in ovarian cancer. HOTAIR was also ↑ expressed in spheroids, as compared to 2D monolayer cells. Silencing HOTAIR ↓ spheroid formation and the rate of tumorigenesis, as well as proliferation levels, in xenografted mice(Y. Zhang et al., 2020)Pancreatic cancerDDIT4 antisense RNA 1 (DDIT4-AS1)DDIT4-AS1 was ↑ expressed in pancreatic ductal adenocarcinoma tissues and cell lines, and was also correlated with tumor size, tumor stage, and lower survival rates. Overexpressing DDIT4-AS1 ↑ spheroid formation. Silencing DDIT4-AS1 ↓ formation of both primary and secondary spheroids. Silencing DDIT4-AS1 in cells injected into mice, and injecting the silencing machinery intratumorally, ↓ both tumor incidence, growth, and weight in xenografted mice. Silencing DDIT4-AS1 and treating with gemcitabine further ↓ tumor incidence, growth, weight, and proliferation levels, and ↑ apoptosis levels, in xenografted mice(Zhang et al., 2022c)Pvt1 oncogene (PVT1)PVT1 was ↑ expressed in spheroids, as compared to 2D monolayer cells. Curcumin treatment, alone or in combination with gemcitabine, ↓ PVT1 expression in tumors, as well as tumor weight, in xenografted mice(Yoshida et al., 2017)Prostate cancerMBNL1 antisense RNA 1 (MBNL1-AS1)MBNL1-AS1 expression was correlated with the stemness index, prognosis, tumor progression, recurrence, and drug resistance, and inversely correlated with immune responsiveness and immunotherapy effects. Silencing MBNL1-AS1 ↑ spheroid formation from CSCs and ↑ proliferative and invasive abilities in xenografted mice(J. Liu et al., 2022)Wilms’ tumorLong intergenic non-protein coding RNA 667 (LINC00667)Silencing LINC00667 ↓ spheroid formation from primary cells. Knocking-out LINC00667 ↓ tumor growth, volume, and weight in xenografted mice(Liu et al., 2020a)EMX2 opposite strand/antisense RNA (EMX2OS)EMX2OS was ↓ expressed in Wilms’ tumor patients. Overexpressing EMX2OS ↓ the number of spheroids and tumor weight in xenografted mice(Zhang et al., 2023a) Table 2Studies that use spheroids and xenografts to study lncRNAsCancer typelncRNASummaryRefBladder cancerBrain cytoplasmic RNA 1 (BCYRN1)BCYRN1 seemed ↑ expressed in bladder cancer. BCYRN1 silencing ↓ spheroid formation, as well as ↓ tumor weight and tumor volume in xenografted mice(Arima et al., 2024)Breast cancerColon cancer-associated transcript 2 (CCAT2)CCAT2 was ↓ expressed in tumor tissues from luminal breast cancer patients. CCAT2 overexpression ↓ the number and diameter of spheroids, and the diameter of tumors in xenografted mice(Xie et al., 2023)Differentiation antagonizing non-protein coding RNA (DANCR)DANCR was ↑ expressed in triple negative breast cancer cell lines. Silencing of DANCR ↓ the size of the networks formed by spheroids growing in Matrigel and the size of tumors in xenografted mice(Nicolescu et al., 2023)H19 imprinted maternally expressed transcript (H19)H19 was ↑ expressed in breast cancer tumors and cell lines. H19 overexpression ↑ the number and diameter of spheroids, while H19 silencing ↓ tumor volume in xenografted mice(F. Peng et al., 2017)H19 was ↑ expressed in breast cancer xenograft tumors and cell lines. Silencing of H19 ↓ the number and diameter of spheroids, while H19 overexpression ↑ tumor volume in xenografted mice(F. Peng et al., 2018)HOX transcript antisense RNA (HOTAIR)HOTAIR was ↑ expressed in breast cancer tissues. HOTAIR silencing ↓ spheroid formation and tumor size in xenografted mice, and ↑ survival rates of xenografted mice(R. Han et al., 2022)Long intergenic non-protein coding RNA 52 (LINC00052)LINC00052 was ↓ expressed in patients with worse prognosis. LINC00052 expression ↓ with spheroid formation. LINC00052 silencing did not affect spheroid formation, but ↑ cell migration in a zebrafish xenograft model(Sanchez-Lopez et al., 2021)Long intergenic non-protein coding RNA 3000 (LINC03000)LINC03000, also known as LncMat2B was ↑ expressed in cisplatin-resistant breast cancer cells. LINC03000 overexpression ↓ spheroid growth and survival rates in zebrafish(Garcia-Venzor et al., 2020)MBNL1 antisense RNA 1 (MBNL1-AS1)MBNL1-AS1 was ↓ expressed in breast cancer tissues and cell lines. ↓ expression of MBNL1-AS1 was associated with ↓ survival rates in patients. MBNL1-AS1 silencing ↑ spheroid number and diameter, and tumor size in xenografted mice, while its overexpression had the opposite effects(Ding et al., 2022)Metastasis associated lung adenocarcinoma transcript 1 (MALAT1)MALAT1 was ↑ expressed in triple-negative breast cancer tissues and cell lines. Silencing MALAT1 ↓ spheroid formation and tumor volume in xenografted mice, while MALAT1 overexpression ↑ tumor volume in this mouse model(Hu et al., 2023)TOMM22 divergent transcript (TOMM22-DT)TOMM22-DT, also known as Lnc408, was ↑ expressed in breast cancer stem cells. Silencing of TOMM22-DT ↓ the number of spheroids and tumor volume in xenografted mice, while TOMM22-DT overexpression had the opposite effects(Wen et al., 2021)X inactive specific transcript (XIST)XIST was ↑ expressed in breast cancer cell lines. XIST silencing ↓ the number of spheroids and their size, as well as tumor volume in xenografted mice(Ma et al., 2023)Colorectal cancerB4GALT1 antisense RNA 1 (B4GALT1-AS1)B4GALT1-AS1 was ↑ expressed in colon cancer cell lines. B4GALT1-AS1 silencing ↓ the number of spheroids and the size of tumors in xenografted mice(Y. Zhang et al., 2019a, b)BACE1 antisense RNA (BACE1-AS)BACE1-AS was ↑ expressed in colorectal cancer tissues and high expression of BACE1-AS was associated with ↓ patient survival rates. BACE1-AS overexpression ↑ the number and diameter of spheroids. BACE1-AS knockout ↓ the number and diameter of spheroids, as well as the number of metastatic foci in the livers of xenografted mice(Wang et al., 2023b)Breast cancer antiestrogen resistance 4 (BCAR4)BCAR4 was ↑ expressed in colorectal cancer tumors. Overexpression of BCAR4 ↑ sphere formation efficiency. In xenografted mice, BCAR4 silencing ↓ tumor weight and volume(Ouyang et al., 2019)CERS6 antisense RNA 1 (CERS6-AS1)CERS6-AS1 was ↑ expressed in colorectal cancer tissues. CERS6-AS1 silencing ↓ spheroid formation efficiency and tumor weight in xenografted mice(Zhao et al., 2022)GATA2 antisense RNA 1 (GATA2-AS1)GATA2-AS1 was ↑ expressed in colon adenocarcinoma tissues. Silencing of GATA2-AS1 ↓ the number of spheroids, as well as the tumor weight and volume in xenografted mice(Pan et al., 2022)HOX transcript antisense RNA (HOTAIR)HOTAIR was ↑ expressed in colorectal cancer tumors, and high expression of HOTAIR was associated with ↓ patient survival rates. HOTAIR silencing ↓ spheroid formation and tumor volume in xenografted mice(Y. Huang et al., 2021)Long intergenic non-protein coding RNA 1315 (LINC01315)LINC01315 was ↑ expressed in colon adenocarcinoma tissues. LINC01315 silencing ↓ the number of formed spheroids, and tumor size and weight in xenografted mice. LINC01315 overexpression ↑ spheroid formation(Li et al.., 2022c)Pvt1 oncogene (PVT1)PVT1 was ↑ expressed in colon cancer tissues and high expression of PVT1 was associated with ↓ patient survival rates. PVT1 overexpression ↑ the number of spheroids and tumor volume in xenografted mice(Lai et al., 2021)Testis development related 1 (TDRG1)TDRG1 was ↑ expressed in colorectal cancer cells. Silencing of TDRG1 ↓ spheroid size and number, as well as tumor formation in xenografted mice(Hong et al., 2022)Cutaneous squamous cell carcinomaPvt1 oncogene (PVT1)PVT1 was ↑ expressed in cutaneous squamous cell carcinoma tissues. PVT1 knockout ↓ spheroid area, as well as the volume and senescence levels in xenografted mouse tumors(Li et al., 2024b)Esophageal squamous cell carcinomaLong intergenic non-protein coding RNA, regulator of reprogramming (LINC-ROR)LINC-ROR was ↑ expressed in esophageal squamous cell carcinoma tissues. LINC-ROR silencing, as well as cisplatin, ↓ spheroid formation, as well as tumor mass and volume in xenografted mice(L. Wang et al., 2017)Gastric cancerASB16 antisense RNA 1 (ASB16-AS1)ASB16-AS1 was ↑ expressed in gastric cancer tissues and cells. ASB16-AS1 silencing ↓ spheroid formation efficiency, as well as tumor weight and volume in xenografted mice(Fu et al., 2021a)Long intergenic non-protein coding RNA 3145 (LINC03145)LINC03145, also known as CRART16, was ↑ expressed in gastric cancer tissues. LINC03145 overexpression ↑ number and sprouting of spheroids, tumor size in xenografted mice, and microvessel formation in chick chorioallantoic membrane assays(Zhang et al., 2022a)Long intergenic non-protein coding RNA 924 (LINC00924)LINC00924 was ↑ expressed in gastric cancer tissues, and high LINC00924 expression was associated with lower survival rates of patients. LINC00924 overexpression ↑ spheroid formation and peritoneal metastasis in mice. LINC00924 silencing ↓ spheroid formation and prevented tumor growth in xenografted mice(He et al., 2022b)WT1 antisense RNA (WT1-AS)WT1-AS was ↓ expressed in gastric cancer stem cells as compared to non-stem gastric cancer cells. WT1-AS overexpression ↓ spheroid size and tumor size in xenografted mice. Overexpression of WT1-AS ↓ tumor growth, stemness, and metastasis in mouse xenografts(Zhang et al., 2023b)GliomaFOXD2 adjacent opposite strand RNA 1 (FOXD2-AS1)FOXD2-AS1 was ↑ expressed in glioma tissues, and ↑ levels of FOXD2-AS1 were associated with ↓ patient survival rates. Silencing of FOXD2-AS1 ↓ spheroid formation and tumor weight in xenografted mice(Wang et al., 2022c)Hepatocellular carcinoma up-regulated long non-coding RNA (HULC)HULC was ↑ expressed in glioblastoma tissues. HULC silencing ↓ spheroid formation, as well as tumor weight and volume in xenografted mice. HULC overexpression ↑ tumor weight in xenografted mice(Li et al., 2022b)Nuclear paraspeckle assembly transcript 1 (NEAT1)NEAT1 was ↑ expressed in glioma cells upon infection with adenovirus. Silencing of NEAT1 ↓ the number of spheroids. Injection of spheroids in mice was associated with larger tumors, as compared to the injection of 2D-cultured cells. Inoculation of mice with higher numbers of cells in spheroids also correlated with ↓ survival rates(Zang et al., 2020)Small nucleolar RNA host gene 10 (SNHG10)SNHG10 was ↑ expressed in glioma cells. Silencing of SNHG10 ↓ spheroid formation efficiency, as well as tumor volume and weight in xenografted mice(Jin et al., 2020)Head and neck squamous cell carcinomaVentricular heart development associated lncRNA (VHRT)VHRT, also known as MASCC1, was differently expressed in metastatic head and neck squamous cell carcinoma, as compared to non-metastatic tissues of the same cancer type. Low VHRT expression was associated with ↑ patient survival. Silencing of VHRT ↓ spheroid weight and volume, as well as proliferation markers in xenografted mouse tumors. Overexpression of VHRT ↑ spheroid formation, as well as tumor formation and lymph node metastasis in xenografted mice(Wang et al., 2023c)Hypopharyngeal squamous cell carcinomaLong intergenic non-protein coding RNA 662 (LINC00662)LINC00662 was ↑ expressed in hypopharyngeal squamous cell carcinoma tissues, and patients with high expression of LINC00662 had ↓ survival rates. LINC00662 overexpression promoted spheroid formation, and ↑ tumor volume in xenografted mice(B. Zhang & Ye, 2024)Liver cancerFGD5 antisense RNA 1 (FGD5-AS1)FGD5-AS1 was ↑ expressed in hepatocellular carcinoma tissues and cell lines, being correlated with tumor stage and size. Patients with higher FGD5-AS1 expression showed ↓ survival. FGD5-AS1 silencing ↓ spheroid diameter and volume, as well as the weight, stemness, and proliferation of xenografted mouse tumors. FGD5-AS1 overexpression ↑ spheroid diameter(He et al., 2022a)FRMD6 antisense RNA 1 (FRMD6-AS1)FRMD6-AS1 was ↑ expressed in hepatocellular carcinoma tissues and cell lines. High expression of FRMD6-AS1 was correlated with ↓ survival of patients. Overexpression of FRMD6-AS1 ↑ spheroid formation and tumor size and weight in xenografted mice(Sun et al., 2023)KCNQ1 opposite strand/antisense transcript 1 (KCNQ1OT1)KCNQ1OT1 was ↑ expressed in hepatocellular carcinoma tissues, and high expression of KCNQ1OT1 was correlated with lower patient survival rates. KCNQ1OT1 knockout ↓ the number of spheroids and tumor volume in xenografted mice(Majumdar et al., 2024)Lipoprotein(a) like 2 (pseudogene) (LPAL2)LPAL2 was ↓ expressed in hepatocellular carcinoma tissues and CD133+ cells Low expression of LPAL2 was correlated with lower patient survival. LPAL2 silencing accelerated spheroid formation and ↑ spheroid size, as well as tumor growth and weight in xenografted mice(Lin et al., 2022)lncRNA regulator of Akt signaling associated with HCC and RCC (LNCARSR)LNCARSR, also known as LncRNA ARSR, was ↑ expressed in cisplatin-resistant hepatocellular carcinoma, in CSCs from primary liver cultures, and in spheroids derived from human primary hepatocellular carcinoma cells, relative to 2D cultures. When plated again in 2D-settings, the CSCs from trypsinized spheroids had ↓ expression of LNCARSR. Silencing of LNCARSR ↓ spheroid number and size, and resulted in a smaller population of CSC in hepatome cells when injected into mice. LNCARSR was also associated with chemoresistance in patients(Yang et al., 2019a, b)Long intergenic non-protein coding RNA 261 (LINC00261)LINC00261 was ↓ expressed in TGFB1-induced epithelial-mesenchymal transition hepatocellular carcinoma cell lines. Overexpression of LINC00261 ↓ spheroid number and size and ↓ tumor weight and growth rate in xenografted mice(Chen & Xiang, et al., 2022b)Long intergenic non-protein coding RNA 680 (LINC00680)LINC00680 was ↑ expressed in hepatocellular carcinoma tissues and cell lines, was correlated with the tumor stage and size, and patients with high expression of LINC00680 had lower survival rates. Overexpression of LINC00680 ↑ primary and secondary spheroid formation. Silencing LINC00680 ↓ primary and secondary spheroid formation, and, in xenografted mice, modified stemness markers and enhanced antitumor potential of the drug tested(Shu et al., 2021)OIP5 antisense RNA 1 (OIP5-AS1)OIP5-AS1 was ↑ expressed in hepatoblastoma cells. Silencing OIP5-AS1 ↓ the number of spheroids and tumor weight in xenografted mice(W. Jiang et al., 2022)Small nucleolar RNA host gene 3 (SNHG3)SNHG3 was ↑ expressed in hepatocellular carcinoma primary tumors. SNHG3 silencing ↓ spheroid formation efficiency and ↑ survival rates in xenografted mice(Guo et al., 2024a)Small nucleolar RNA host gene 5 (SNHG5)SNHG5 was ↑ expressed in spheroids, as compared to 2D monolayer cells. Silencing SNHG5 ↓ spheroid formation and ↓ tumor growth, volume, and weight in xenografted mice. Spheroids were enriched in CSCs and had ↑ expression of SNHG5(Li & Hu et al., 2022a)Small nucleolar RNA host gene 9 (SNHG9)SNHG9 was ↑ expressed in hepatocellular carcinoma tissues and cell lines, including CD133+ hepatocellular carcinoma CSC cells. Overexpressing SNHG9 ↑ spheroid formation of CD133- cells. Silencing SNHG9 ↓ spheroid formation of CD133+ cells, as well as tumor volume and weight in xenografted mice(Yang et al., 2023b)Lung cancerADAMTS9 antisense RNA 1 (ADAMTS9-AS1)ADAMTS9-AS1 was ↓ expressed in most cancer types (including lung adenocarcinoma), in both tissues and cell lines. Overexpressing ADAMTS9-AS1 ↓ spheroid formation as well as tumor growth and weight in xenografted mice(Wang et al., 2023a)ASAP1 intronic transcript 1 (ASAP1-IT1)ASAP1-IT1 was ↑ expressed in NSCLC tissues and cell lines. It was correlated with tumor differentiation, tumor node metastasis (TNM), and lymph node metastasis (LNM). ASAP1-IT1 was also ↑ expressed in spheroids, as compared to 2D monolayer cells. Overexpressing ASAP1-IT1 ↑ spheroid number. Silencing ASAP1-IT1 ↓ spheroid formation and tumor growth in xenografted mice(Liu et al., 2021a)GPAT4 and GINS4 antisense RNA 1 (GPAT4-AS1)GPAT4-AS1 ,also known as GIAT4RA, was ↓ expressed in lung tissues (both adenocarcinoma and squamous carcinoma) and associated with higher survival of patients. GPAT4-AS1 was also associated with tumor size, clinical stage, and lymphatic metastasis. Overexpressing GPAT4-AS1 ↓ spheroid size and number, as well as tumor growth and weight in xenografted mice. Silencing GPAT4-AS1 had the opposite effects(Yang et al., 2019b)Long intergenic non-protein coding RNA 1419 (LINC01419)LINC01419 was ↑ expressed in lung adenocarcinoma tissues and cell lines. Overexpressing LINC01419 ↑ spheroid formation. Silencing LINC01419 ↓ tumor growth, weight, volume, and proliferative levels in xenografted mice(Chen et al., 2022a)Long intergenic non-protein coding RNA 467 (LINC00467)LINC00467 was ↑ expressed in lung adenocarcinoma tissues and cell lines, and high expression of LINC00467 was associated with ↓ survival. Silencing LINC00467 ↓ spheroid number and tumor size and weight, as well as ↓ proliferation and ↑ apoptosis levels in xenografted mice(Chang & Yang, 2020)MNX1 antisense RNA 1 (head to head) (MNX1-AS1)MNX1-AS1 was ↑ expressed in lung tissues, including lung adenocarcinoma, and also in CSCs. Silencing MNX1-AS1 ↓ spheroid formation of CSCs and the tumor growth rate and weight in xenografted mice(Lv et al., 2023)OIP5 antisense RNA 1 (OIP5-AS1)OIP5-AS1 was ↑ expressed in lung cancer cells, and was also ↑ expressed in spheroids, as compared to 2D monolayer cultures. Silencing OIP5-AS1 ↓ spheroid number and size. Overexpressing OIP5-AS1 ↑ spheroid number and size, as well as tumor formation rates in in vivo experiments(Mao & Li, 2022)SOX2 overlapping transcript (SOX2OT)SOX2OT was ↑ expressed in lung tumors and resistant cell lines. Patients with high SOX2OT expression had ↓ survival. SOX2OT was also ↑ expressed in CSC-derived spheroids, as compared to 2D monolayer cells. CSCs isolated from spheroids had ↑ capacity to initiate solid tumor formation in vivo, as compared to 2D monolayer cells. Silencing SOX2OT ↓ spheroid formation and, when combined with GL1T repression, tumor growth in xenografted mice also ↓(Dong et al., 2023)Testis development related 1 (TDRG1)TDRG1 was ↑ expressed in non-small cell lung cancer cells. Silencing TDRG1 ↓ spheroid number and size. When inoculated into mice it ↓ tumor formation rate and stem cell frequency. Overexpressing TDRG1 had the opposite effects(Lu et al., 2023)MedulloblastomaLOXL1 antisense RNA 1 (LOXL1-AS1)LOXL1-AS1 was ↑ expressed in medulloblastoma tumors and cell lines. Patients with ↑ expressed LOX1-AS1 had ↓ survival rates. LOXL1-AS1 silencing ↓ spheroid formation, which was ↑ under LOXL1-AS1 overexpression. Xenografted mice with LOXL1-AS overexpression had ↓ survival rates(Do et al., 2024)Oral squamous cell carcinomaTINCR ubiquitin domain containing (TINCR)TINCR was ↓ expressed in oral squamous cell carcinoma tissues, specifically in poorly differentiated tissues, and patients with low expression of TINCR had ↓ survival. TINCR was also positively correlated with the pathological differentiation stage. TINCR was also ↓ expressed in spheroids, as compared to 2D monolayers, a process that was associated with dedifferentiation. Overexpression of TINCR ↓ tumor size, weight, and volume in xenografted mice(Z. Zhuang et al., 2021)OsteosarcomaB4GALT1 antisense RNA 1 (B4GALT1-AS1)B4GALT1-AS1 was ↑ expressed in osteosarcoma tissues, cell lines, and spheroids, as compared to 2D monolayer cells. Silencing B4GALT1-AS1 ↓ spheroid size and number, and, in xenografted mice, ↓ tumor size, weight, seeding ability, and formation rate(Li et al., 2018b)WAC antisense RNA1 (WAC-AS1)WAC-AS1 was ↑ expressed in most cancer types (including osteosarcoma), specifically, in metastatic tissues, as compared to non-metastatic and normal tissues. WAC-AS1 expression was also correlated with lung metastasis, and patients with high expression of WAC-AS1 had ↓ survival rates. Overexpressing WAC-AS1 ↑ spheroid number, as well as tumor growth and metastasis in xenografted mice(Yang et al., 2023c)Ovarian cancerHOX transcript antisense RNA (HOTAIR)HOTAIR was ↑ expressed in ovarian cancer (specifically in stages I/II) and was correlated with tumor stage. Plasma circulating levels of HOTAIR showed high specificity and sensitivity as a diagnostic biomarker in ovarian cancer. HOTAIR was also ↑ expressed in spheroids, as compared to 2D monolayer cells. Silencing HOTAIR ↓ spheroid formation and the rate of tumorigenesis, as well as proliferation levels, in xenografted mice(Y. Zhang et al., 2020)Pancreatic cancerDDIT4 antisense RNA 1 (DDIT4-AS1)DDIT4-AS1 was ↑ expressed in pancreatic ductal adenocarcinoma tissues and cell lines, and was also correlated with tumor size, tumor stage, and lower survival rates. Overexpressing DDIT4-AS1 ↑ spheroid formation. Silencing DDIT4-AS1 ↓ formation of both primary and secondary spheroids. Silencing DDIT4-AS1 in cells injected into mice, and injecting the silencing machinery intratumorally, ↓ both tumor incidence, growth, and weight in xenografted mice. Silencing DDIT4-AS1 and treating with gemcitabine further ↓ tumor incidence, growth, weight, and proliferation levels, and ↑ apoptosis levels, in xenografted mice(Zhang et al., 2022c)Pvt1 oncogene (PVT1)PVT1 was ↑ expressed in spheroids, as compared to 2D monolayer cells. Curcumin treatment, alone or in combination with gemcitabine, ↓ PVT1 expression in tumors, as well as tumor weight, in xenografted mice(Yoshida et al., 2017)Prostate cancerMBNL1 antisense RNA 1 (MBNL1-AS1)MBNL1-AS1 expression was correlated with the stemness index, prognosis, tumor progression, recurrence, and drug resistance, and inversely correlated with immune responsiveness and immunotherapy effects. Silencing MBNL1-AS1 ↑ spheroid formation from CSCs and ↑ proliferative and invasive abilities in xenografted mice(J. Liu et al., 2022)Wilms’ tumorLong intergenic non-protein coding RNA 667 (LINC00667)Silencing LINC00667 ↓ spheroid formation from primary cells. Knocking-out LINC00667 ↓ tumor growth, volume, and weight in xenografted mice(Liu et al., 2020a)EMX2 opposite strand/antisense RNA (EMX2OS)EMX2OS was ↓ expressed in Wilms’ tumor patients. Overexpressing EMX2OS ↓ the number of spheroids and tumor weight in xenografted mice(Zhang et al., 2023a) Studies that use spheroids and xenografts to study lncRNAs Organoids: laboratory-made mini-organs Organoids are miniature versions of organs or tissues. They are generated from biological materials containing adult or pluripotent stem cells through in vitro growth in the presence of ECM proteins and biochemical signals. These signals drive stem cell aggregation, self-organization, and differentiation into a physiologically active microscopical structure (Dutta et al., 2017; X. Liu et al., 2025; Yang & Hu et al., 2023a). When generated from cancer tissues, organoids can be termed tumoroids and provide a disease model of a specific cancer type, mimicking features of the native tumors and their respective patients (Guo et al., 2024b). In the context of cancer, organoids are quite versatile to study the effects of lncRNA overexpression or silencing. For example, lncRNA glioma radiation sensitizers 1 (lncGRS-1), which is upregulated in both adult and pediatric forms of glioma, was silenced in glioma organoids by CRISPR interference (CRISPRi) or antisense oligonucleotides (ASOs), decreasing their proliferation. Such silencing enhanced radiation sensitivity, further decreasing cancer cell proliferation in the organoid. Furthermore, organoids composed of healthy human astrocytes and glioma cells were prepared. In these organoids, lncGRS-1 silencing was not harmful to the healthy astrocytes (Liu et al., 2020b), demonstrating the promising safety of this potential therapeutic approach. Of note, lncGRS-1 orthologs do not exist in mice, chicken, or zebrafish, for which studies in animal models would not be feasible for this lncRNA. As such, 3D cellular models provide a viable alternative, with the advantage of being representative of the human species, unlike animal models. This study not only exemplifies organoids as additional 3D cellular models that can be used to assess the effects of lncRNA overexpression or silencing in cancer, but also accentuates the usefulness of 3D models in cases of poor conservation of lncRNAs in commonly used animal models. Patient-derived organoids can mimic ex vivo tumors Several studies have compared the impact of lncRNA overexpression or silencing on organoids and xenograft mouse models, and some also included spheroids (Table 3). For example, in colorectal cancer, POU6F2 antisense RNA 2 (POU6F2-AS1) silencing suppressed patient-derived organoid formation, while its overexpression had the opposite effect. POU6F2-AS1 silencing also decreased cell growth, and tumor volume and weight in a xenograft model (T. Jiang et al., 2024). Additional studies also assessed metastasis in xenograft models. In gastric cancer, silencing of C8orf76, which directly binds and induces lncRNA Dual specificity phosphatase 5 pseudogene 1 (DUSP5P1), suppressed the growth of patient-derived organoids and tumors in a xenograft mouse model. In mice, C8orf76 silencing also suppressed lung and liver metastasis (X. Wang et al., 2019). These studies demonstrate that the effects of lncRNA overexpression or silencing are consistent between organoids and tumors representing the same cancer type. Furthermore, animals provide additional information on metastasis formation. Table 3Studies that focus on the combined use of organoids and xenograftsCancer typelncRNASummaryRefBreast cancerDDIT4 antisense RNA 1 (DDIT4-AS1)DDIT4-AS1 was ↑ expressed in triple negative breast cancer tissues and cell lines, and was positively associated with the TNM stage, Ki67 proliferation marker, and patient survival. DDIT4-AS1 was dysregulated by autophagy induction. Simultaneous silencing of DDIT4-AS1 and paclitaxel treatment ↓ patient-derived organoid quantity, size, and cell viability, as well as tumor growth and proliferation levels in xenografted mice(Jiang et al., 2023)Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), nuclear paraspeckle assembly transcript 1 (NEAT1)MALAT1 was ↑ expressed in lung metastasis cancer cells (as compared to matched breast cancer primary tumors) and in advanced carcinomas (as compared to adenomas or hyperplasia). Silencing MALAT1 ↓ organoid size, branching, cellular motility, and collective cell migration, ↑ organoid compaction, and organized organoids in a polarized, differentiated manner, similar to normal mammary glands. Delivery of MALAT1 to MALAT1 knockout organoids rescued the branching to the same level of wild-type organoids. MALAT1 knockout mice developed cystic-filled tumors and had ↓ metastatic burden, as compared to controls consisting of poorly differentiated solid carcinomas. Silencing NEAT1 resulted in similar branching alterations to controls, but there was ↑ apoptotic cell death and ↓ in the number of dividing cells(Arun et al., 2016)PDCD4 antisense RNA 1 (PDCD4-AS1)PDCD4-AS1 was ↓ expressed in triple-negative breast cancer tissues, as compared to other breast cancer subtypes, specifically in the most aggressive stages. High expression of PDCD4-AS1 was correlated with ↑ patient survival rates. Normal breast cells were modified and selected from xenografts to be increasingly aggressive and were then cultured in Matrigel as organoid-like structures, where PDCD4-AS1 was ↓ expressed in the most aggressive stages of breast cancer(Jadaliha et al., 2018)CholangiocarcinomaTTN antisense RNA 1 (TTN-AS1)TTN-AS1 was ↑ expressed in cholangiocarcinoma tissues, circulating tumor cell exosomes, and blood. High expression of TTN-AS1 was associated with LNM and TNM. Organoids were established using circulating tumor cells of patients, and, in exosomes from their extracellular vesicles (EVs), TTN-AS1 was ↑ expressed. Such exosomes were collected from TTN-AS1 overexpressing organoids and co-cultured with cholangiocarcinoma cells, which were then injected into mice. TTN-AS1 overexpression in organoids resulted in downstream ↑ tumor volume, weight, and proliferation levels in xenografted mice, while silencing TTN-AS1 in organoids had the opposite effect(Zhou et al., 2024a) Colorectal cancerMaternally expressed 3 (Meg3)Meg3 was ↓ expressed in colorectal cancer tumors. Meg3 accumulated in the colonic crypts of tumors in xenografted mice and organoids were derived from colon cells of these mice. Silencing Meg3 in mice resulted in downstream ↑ organoid growth, while Meg3 overexpression ↓ organoid growth(Zhang et al., 2022b)POU6F2 antisense RNA 1 (POU6F2-AS1)POU6F2-AS1 was ↑ expressed in colorectal cancer tumors and cell lines and was associated with tumor diameter, depth of invasion, the LNM and TNM stages, and ↓ survival. Overexpressing POU6F2-AS1 ↑ patient-derived organoid formation. Silencing POU6F2-AS1 ↓ patient-derived organoid formation and ↓ volume, weight, proliferative levels, and lipid content of xenografted mouse tumors(T. Jiang et al., 2024)Small Cajal body-specific RNA 2 (SCARNA2)SCARNA2 was ↑ expressed in radioresistant patients, and patients with high expression of SCARNA2 had ↓ survival. Silencing SCARNA2 combined with radiotherapy ↓ patient-derived organoid growth, as well as tumor weight and progression in xenografted mice. Delivery of SCARNA2 silencing machinery directly to xenografted mice also ↓ tumor growth, weights, and proliferative levels, and ↑ apoptotic levels(Y. Chen et al., 2023)Gastric cancerDSP antisense RNA 1 (DSP-AS1)DSP-AS1, also known as ABL, was ↑ expressed in gastric cancer tumors and cell lines and was a predictive marker for gastric cancer prognosis. DSP-AS1 was also associated with ↓ survival of patients. Overexpressing DSP-AS1 ↑ organoid diameter, cell growth, and ↓ cisplatin/paclitaxel-induced apoptosis. Overexpressing DSP-AS1 in xenografted mice ↑ size, weight, growth, proliferative levels, and ↓ the number of apoptotic cells and thus cisplatin effect. Co-delivering silencing machinery with paclitaxel in xenografted mice ↓ tumor size, weight, growth, proliferative levels, and the number of apoptotic cells(Wang et al., 2022b)Dual specificity phosphatase 5 pseudogene 1 (DUSP5P1)In xenografted mice, overexpressing DUSP5P1 ↑ tumor growth, numbers of proliferating cells and ↓ numbers of apoptotic cells. Silencing of C8orf76, which directly binds and induces DUSP5P1, ↓ the growth of patient-derived organoids and xenografted mice(X. Wang et al., 2019)GliomalncRNA glioma radiation sensitizers 1 (lncGRS-1)Glioma cells were seeded on the surface of normal brain organoids, which the cancer cells invaded and transformed into tumors. LncGRS-1 silencing reduced tumor growth and sensitized it to radiotherapy, without affecting the healthy organoids(Liu et al., 2020b) MIR22 host gene (MIR22HG)MIR22HG was ↑ expressed and transcriptionally active in glioblastoma, as well as in its subtype with worst outcomes, especially in the infiltrating regions of the tumor, as compared to the core. Patients with high MIR22HG expression had ↓ survival, and MIR22HG was also validated as an independent prognostic indicator of patient survival. Silencing MIR22HG ↓ formation and the invading distance of spheroids. When co-cultured with brain organoids, silencing MIR22HG in spheroids ↓ the number of invading tumor cells. Silencing MIR22HG in xenografted mice ↓ growth, proliferative, and invasive markers and ↑ survival of tumor-bearing mice(M. Han et al., 2020)Hepatocellular carcinomaBBOX1 antisense RNA 1 (BBOX1-AS1)BBOX1-AS1 was ↑ expressed in hepatocellular carcinoma cell lines and tumors, especially in patients with vascular invasion and advanced TNM stage. High expression of BBOX1-AS1 was associated with ↓ survival. BBOX1-AS1 was also validated as an independent risk factor for hepatocellular carcinoma. In intermediate resistance organoids, silencing BBOX1-AS1 ↓ organoid growth, which was enhanced when coupled with sorafenib treatment. Silencing BBOX1-AS1 and treating with sorafenib ↓ tumor growth and proliferative levels in xenografted mice. Patient-derived organoids were evaluated for sorafenib IC50 and the most resistant organoids had ↑ expressed BBOX1-AS1 as compared to the less resistant. Overexpressing BBOX1-AS1 ↑ tumor size, proliferative levels, and the number of metastatic nodules in xenografted mice(Tao et al., 2023) Table 3Studies that focus on the combined use of organoids and xenograftsCancer typelncRNASummaryRefBreast cancerDDIT4 antisense RNA 1 (DDIT4-AS1)DDIT4-AS1 was ↑ expressed in triple negative breast cancer tissues and cell lines, and was positively associated with the TNM stage, Ki67 proliferation marker, and patient survival. DDIT4-AS1 was dysregulated by autophagy induction. Simultaneous silencing of DDIT4-AS1 and paclitaxel treatment ↓ patient-derived organoid quantity, size, and cell viability, as well as tumor growth and proliferation levels in xenografted mice(Jiang et al., 2023)Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), nuclear paraspeckle assembly transcript 1 (NEAT1)MALAT1 was ↑ expressed in lung metastasis cancer cells (as compared to matched breast cancer primary tumors) and in advanced carcinomas (as compared to adenomas or hyperplasia). Silencing MALAT1 ↓ organoid size, branching, cellular motility, and collective cell migration, ↑ organoid compaction, and organized organoids in a polarized, differentiated manner, similar to normal mammary glands. Delivery of MALAT1 to MALAT1 knockout organoids rescued the branching to the same level of wild-type organoids. MALAT1 knockout mice developed cystic-filled tumors and had ↓ metastatic burden, as compared to controls consisting of poorly differentiated solid carcinomas. Silencing NEAT1 resulted in similar branching alterations to controls, but there was ↑ apoptotic cell death and ↓ in the number of dividing cells(Arun et al., 2016)PDCD4 antisense RNA 1 (PDCD4-AS1)PDCD4-AS1 was ↓ expressed in triple-negative breast cancer tissues, as compared to other breast cancer subtypes, specifically in the most aggressive stages. High expression of PDCD4-AS1 was correlated with ↑ patient survival rates. Normal breast cells were modified and selected from xenografts to be increasingly aggressive and were then cultured in Matrigel as organoid-like structures, where PDCD4-AS1 was ↓ expressed in the most aggressive stages of breast cancer(Jadaliha et al., 2018)CholangiocarcinomaTTN antisense RNA 1 (TTN-AS1)TTN-AS1 was ↑ expressed in cholangiocarcinoma tissues, circulating tumor cell exosomes, and blood. High expression of TTN-AS1 was associated with LNM and TNM. Organoids were established using circulating tumor cells of patients, and, in exosomes from their extracellular vesicles (EVs), TTN-AS1 was ↑ expressed. Such exosomes were collected from TTN-AS1 overexpressing organoids and co-cultured with cholangiocarcinoma cells, which were then injected into mice. TTN-AS1 overexpression in organoids resulted in downstream ↑ tumor volume, weight, and proliferation levels in xenografted mice, while silencing TTN-AS1 in organoids had the opposite effect(Zhou et al., 2024a) Colorectal cancerMaternally expressed 3 (Meg3)Meg3 was ↓ expressed in colorectal cancer tumors. Meg3 accumulated in the colonic crypts of tumors in xenografted mice and organoids were derived from colon cells of these mice. Silencing Meg3 in mice resulted in downstream ↑ organoid growth, while Meg3 overexpression ↓ organoid growth(Zhang et al., 2022b)POU6F2 antisense RNA 1 (POU6F2-AS1)POU6F2-AS1 was ↑ expressed in colorectal cancer tumors and cell lines and was associated with tumor diameter, depth of invasion, the LNM and TNM stages, and ↓ survival. Overexpressing POU6F2-AS1 ↑ patient-derived organoid formation. Silencing POU6F2-AS1 ↓ patient-derived organoid formation and ↓ volume, weight, proliferative levels, and lipid content of xenografted mouse tumors(T. Jiang et al., 2024)Small Cajal body-specific RNA 2 (SCARNA2)SCARNA2 was ↑ expressed in radioresistant patients, and patients with high expression of SCARNA2 had ↓ survival. Silencing SCARNA2 combined with radiotherapy ↓ patient-derived organoid growth, as well as tumor weight and progression in xenografted mice. Delivery of SCARNA2 silencing machinery directly to xenografted mice also ↓ tumor growth, weights, and proliferative levels, and ↑ apoptotic levels(Y. Chen et al., 2023)Gastric cancerDSP antisense RNA 1 (DSP-AS1)DSP-AS1, also known as ABL, was ↑ expressed in gastric cancer tumors and cell lines and was a predictive marker for gastric cancer prognosis. DSP-AS1 was also associated with ↓ survival of patients. Overexpressing DSP-AS1 ↑ organoid diameter, cell growth, and ↓ cisplatin/paclitaxel-induced apoptosis. Overexpressing DSP-AS1 in xenografted mice ↑ size, weight, growth, proliferative levels, and ↓ the number of apoptotic cells and thus cisplatin effect. Co-delivering silencing machinery with paclitaxel in xenografted mice ↓ tumor size, weight, growth, proliferative levels, and the number of apoptotic cells(Wang et al., 2022b)Dual specificity phosphatase 5 pseudogene 1 (DUSP5P1)In xenografted mice, overexpressing DUSP5P1 ↑ tumor growth, numbers of proliferating cells and ↓ numbers of apoptotic cells. Silencing of C8orf76, which directly binds and induces DUSP5P1, ↓ the growth of patient-derived organoids and xenografted mice(X. Wang et al., 2019)GliomalncRNA glioma radiation sensitizers 1 (lncGRS-1)Glioma cells were seeded on the surface of normal brain organoids, which the cancer cells invaded and transformed into tumors. LncGRS-1 silencing reduced tumor growth and sensitized it to radiotherapy, without affecting the healthy organoids(Liu et al., 2020b) MIR22 host gene (MIR22HG)MIR22HG was ↑ expressed and transcriptionally active in glioblastoma, as well as in its subtype with worst outcomes, especially in the infiltrating regions of the tumor, as compared to the core. Patients with high MIR22HG expression had ↓ survival, and MIR22HG was also validated as an independent prognostic indicator of patient survival. Silencing MIR22HG ↓ formation and the invading distance of spheroids. When co-cultured with brain organoids, silencing MIR22HG in spheroids ↓ the number of invading tumor cells. Silencing MIR22HG in xenografted mice ↓ growth, proliferative, and invasive markers and ↑ survival of tumor-bearing mice(M. Han et al., 2020)Hepatocellular carcinomaBBOX1 antisense RNA 1 (BBOX1-AS1)BBOX1-AS1 was ↑ expressed in hepatocellular carcinoma cell lines and tumors, especially in patients with vascular invasion and advanced TNM stage. High expression of BBOX1-AS1 was associated with ↓ survival. BBOX1-AS1 was also validated as an independent risk factor for hepatocellular carcinoma. In intermediate resistance organoids, silencing BBOX1-AS1 ↓ organoid growth, which was enhanced when coupled with sorafenib treatment. Silencing BBOX1-AS1 and treating with sorafenib ↓ tumor growth and proliferative levels in xenografted mice. Patient-derived organoids were evaluated for sorafenib IC50 and the most resistant organoids had ↑ expressed BBOX1-AS1 as compared to the less resistant. Overexpressing BBOX1-AS1 ↑ tumor size, proliferative levels, and the number of metastatic nodules in xenografted mice(Tao et al., 2023) Studies that focus on the combined use of organoids and xenografts Several studies have also assessed the effects of lncRNA overexpression or silencing on the organoid response to chemo- or radiotherapy. In gastric cancer, overexpression of DSP antisense RNA 1 (DSP-AS1) was associated with larger organoid dimensions and increased tumor size in mice. Furthermore, organoids with overexpressed DSP-AS1 were more resistant to cisplatin/paclitaxel-induced apoptosis (Wang et al., 2022b). In rectal cancer, silencing of small Cajal body-specific RNA 2 (ScaRNA2) increased sensitivity to radiotherapy on patient-derived organoids and xenograft models (Y. Chen et al., 2023). In triple-negative breast cancer, silencing of DNA damage inducible transcript 4 antisense 1 (DDIT4-AS1) enhanced the efficacy of paclitaxel on patient-derived organoids and on a xenograft mouse model (Jiang et al., 2023). These studies further demonstrate that lncRNA silencing may also increase organoid sensitivity to chemo- or radiotherapy. Organoids can also be obtained from xenografted animals instead of patients. Maternally expressed gene 3 (Meg3) was overexpressed or silenced in xenografted mice. The xenografts were further used to generate organoids. In both xenografts and xenograft-derived organoids, Meg3 silencing or overexpression increased or decreased their size, respectively (S. Zhang et al., 2021). Importantly, organoids are gold standards of “personalized” medicine, reflecting personal aspects of individual patients. In hepatocellular carcinoma, organoids derived from different patients had slightly different levels of sorafenib resistance. Sorafenib and BBOX1 antisense RNA 1 (BBOX1-AS1) silencing inhibited organoid growth, with synergistic effects. On the other hand, BBOX1-AS1 overexpression in a xenograft model increased tumor size and metastasis (Tao et al., 2023). While xenograft models are necessary to assess metastasis and survival rates, patient-derived organoids have the advantage of better reflecting human patients’ features. The above-mentioned studies suggest that xenograft mouse models might be replaced with cognate organoids if only tumor growth or an indicator of therapy response needs to be assessed. Advanced 3D cellular models: underexplored alternatives? While spheroids and organoids have been extensively used in lncRNA cancer research, other 3D cellular models, specifically microfluidic systems such as “organs-on-chips”, have been less exploited in this framework. Microfluidic systems are devices that connect several channels between wells, where each well is filled with cells from the organ they represent. Immune and tumor cells can circulate through these channels and thus, any associated immune response can be monitored (Li et al., 2024a). In the context of triple-negative breast cancer, H19 was studied in microfluidic-based 3D devices that allow co-cultures and hypoxia studies. Because MSCs play an important role in mediating cancer cell invasion, they were co-cultured in microfluidic devices with breast cancer cells. In hypoxic conditions, MSCs had enhanced migration and guided breast cancer cells in a “cluster-sprout-infiltrating” way relative to normoxic conditions. In this study, H19 was induced and upregulated in both hypoxic conditions and in a time-dependent manner. Silencing H19 did not change the morphology of MSCs, but delayed their sprouting, infiltration, and reduced their migration distance. These outcomes impaired and retarded cancer cell migration. Interestingly, under hypoxic conditions, aspirin treatment downregulated H19 in MSCs, in a time-dependent manner. These effects were similar to shRNA-mediated silencing of H19 (An et al., 2022). Microfluidics can thus recreate cell intravasion events on the endothelial barrier of tumors, with precise control of growth-factor gradients and tunable endothelial permeability (Zervantonakis et al., 2012). Another study embedded spheroids in a microfluidic device and studied the effect and rate of internalization of lncRNA RNA component of 7SK nuclear ribonucleoprotein (RN7SK)-loaded exosomes. The lncRNA was synthesized and loaded into exosomes that were placed into side channels, and the spheroids were injected within a hydrogel in the central channel. The uptake of the exosomes led to an overexpression of RN7SK in the non-small cell lung cancer cells, whose viability was reduced, and resulted in the suppression of spheroid dispersion, mimicking an inhibition of the initial metastatic epithelial-mesenchymal transition step (Aghamiri et al., 2026). This shows that microfluidic devices are exceptionally well suited to study migration dynamics of cancer cells responding to both exosomes and cellular cues. There are also 3D organotypic cultures, such as laminin-rich ECM-based three-dimensional organotypic (lrECM 3D) cultures, where cells were seeded in Matrigel-coated plates, in which HOTAIR and its isoforms were studied. HOTAIR-N was found to be a major isoform in invasive carcinoma. Its upregulation in triple-negative breast cancer increased with culture time, and both HOTAIR and HOTAIR-N were upregulated in 3D conditions, as compared to 2D cultures. Silencing HOTAIR decreased invasive growth patterns, cell numbers, and viability. Furthermore, silencing or inhibition of integrin α2, Src, or BRD4 in 3D conditions resulted in downregulated HOTAIR, thus showing that manipulation of ECM-related proteins can impact lncRNAs (M. Li et al., 2017). These examples demonstrate that increasing the complexity of 3D cellular models may bring further insights into our knowledge of the impact of lncRNAs in cancer. The increased complexity of these 3D cellular models could also contribute to increasing their capability of replacing animal models in research. Nevertheless, assessing such capability requires further and extensive research using these advanced 3D cellular models. Discussion In the present review, we analyzed a representative sample of original research studies involving at least one 3D cellular model to assess the impact of lncRNA overexpression or silencing in different types of cancer. The most used 3D model was the spheroid, followed by the organoid. Many of these studies also involved an animal model, which, in most cases, was a mouse xenograft. In most of these studies, a large component of experimental work was carried out in 2D cellular cultures. However, we focused only on the experimental data obtained in 3D and animal models, since those were the most relevant to our review. Since the number of relevant studies is very high (and permanently increasing), we could not perform an exhaustive analysis. As such, most articles chosen for analysis were published between 2022 and 2025. To ensure a representative literature sample, the number of analyzed studies involving each 3D model or 3D+ animal model combination was proportional to the respective abundance within the literature. Our understanding was also limited by the level of detail provided in the methods sections of the original research publications. lncRNAs are disease context-dependent In our literature search, multiple solid cancer types were addressed. The breast, colorectal, lung, liver, and brain cancers were the most found. Overall, more than half of the studied lncRNAs were sense, being H19 and HOTAIR the most studied lncRNAs, followed by MALAT1, PVT1, and LINC-ROR. In most studies, lncRNA levels were overexpressed or silenced (with or without drug testing) before assembly of the 3D in vitro cellular model. We also noticed that the vast majority of the studies focused on lncRNAs with oncogenic or tumor-promoting functions that are naturally upregulated in the tumors, which were then experimentally manipulated through gene silencing. For correlative studies of lncRNA expression, most studies used databases, such as GENCODE, ENCODE, and GEO, which contain gene annotations and gene expression repositories (Pinkney et al., 2020). Some of them also analyzed a new cohort as an additional validation of their results. Because lncRNA function and impact are context-dependent, each lncRNA may have different effects in different cancer types. For example, silencing H19 in either colorectal cancer (M. Zhu et al., 2024) or papillary thyroid carcinoma (Li et al., 2018a) decreased spheroid formation capacity; however, in breast cancer, silencing this lncRNA increased spheroid formation (Shima et al., 2018). Another example is HIF2PUT, whose silencing decreases colorectal cancer spheroids formation and self-renewal capabilities (Yao et al., 2015). However, in osteosarcoma, its silencing resulted in increased spheroid formation rates (Y. Wang et al., 2015). In glioma tissues, LINC-ROR was downregulated and its overexpression decreased spheroid number (Feng et al., 2015). Furthermore, in esophageal squamous cell carcinoma tissues, LINC-ROR was, on the other hand, upregulated and its silencing decreased both the formation of spheroids and tumors in xenografted mice (L. Wang et al., 2017). These examples demonstrate the complexity of lncRNA mechanisms in cancer and spheroid formation. Therefore, any generalization across cancer types should be carefully considered. Although the same lncRNA can have different impacts in different cancer types, its response in 3D and animal models of the same cancer type is quite consistent. We found only one study with potentially conflicting results between a 3D and an animal model. In breast cancer, LINC00052 expression decreased upon spheroid formation. However, its silencing failed to affect spheroid formation, but enhanced cell migration in a zebrafish xenograft model (Sanchez-Lopez et al., 2021). Of note, the phenotypic processes studied in spheroids and zebrafish were not exactly the same, which could explain the different results. Most studies used only one cell line; however, using more cell lines would strengthen the generalizability of the results for their respective cancer type. Some studies compared expression levels between 2D and 3D models (Li et al., 2017). However, in 3D cellular models, most studies did not test any additional hypotheses beyond lncRNA overexpression/silencing or drug resistance, revealing some gaps and opportunities to explore in future research. Manipulation of lncRNA expression Several systems were used to manipulate lncRNA expression, most of them employed in monolayer cultures preceding both 3D assembly and xenograft injection. For example, for silencing, CRISPR/Cas genome editing was widely used for permanent knockout. This process can be challenging, as the editing efficiency can be low, the selection and optimization processes can be time-consuming, and the outcomes may compromise cell survival (if homozygotic) and bring off-target and unexpected outcomes. On the other hand, CRISPRi recruits transcriptional repression factors to the gene promoter, decreasing RNA transcription levels, which provides a less drastic effect than the knockout (Ghavami & Pandi, 2021). For transient knockdown of lncRNAs in 3D cellular models, two possibilities are small interfering RNA (siRNA) and short hairpin RNA (shRNA). An siRNA consists of a double-stranded RNA molecule that is chemically synthesized and transiently transfected into target cells, where it activates a protein complex that degrades target RNA. The same RNA-degrading mechanism can be activated through shRNA, in which target cells are transfected with a vector that expresses a single RNA molecule that self-hybridizes (generating a short hairpin-like structure) to mimic the siRNA duplex. While both siRNA and shRNA activate the same RNA degradation mechanism, siRNA is adequate for the degradation of cytoplasmic RNAs over short time periods, since it is delivered as RNA, into the cytoplasm, where its lifetime is short. On the other hand, shRNA is delivered as its expression vector, being continuously transcribed in the nucleus. Therefore, shRNA usually provides longer-lasting RNA degradation than siRNA and can degrade nuclear RNA. Nevertheless, if a prolonged siRNA effect is needed, repeated transfection may be necessary, and penetration into the core of already formed 3D structures can be very limited. An additional approach consists of ASOs, which are single-stranded DNA molecules that hybridize with target RNA, inducing its degradation by RNAse H. ASOs can be chemically modified to increase their binding affinity and lifetime. Since RNAse H can be found in both the nucleus and cytoplasm, ASOs can degrade both nuclear and cytoplasmic lncRNAs. Due to the enhanced stability of DNA relative to RNA, ASOs can provide longer-lasting effects than siRNA. (Lennox & Behlke, 2016; S. Zibitt et al., 2021). Thus, the choice of an RNA silencing approach could be guided by the subcellular location of the target lncRNA and the duration of the intended effect. For lncRNA overexpression in 3D cellular models, their endogenous transcription can be increased by CRISPR activation, which uses a CRISPR-associated protein to recruit transcriptional activators to the promoter (Chavez et al., 2016). Alternatively, lncRNA can be transcribed from an exogenous expression vector. The vector can be delivered to the cells by lipofection, electroporation (electrical pulses), or packaged inside artificial viruses. If an exogenously expressed lncRNA is expected to have a nuclear location, an appropriate nuclear retention signal should be added to its sequence (Wei et al., 2023). The choice of a specific lncRNA overexpression approach should be guided by the efficiency and homogeneity of the effect needed. Electroporation and viral infection usually result in higher percentages of expressing cells, with more homogeneous expression levels, relative to lipofection. Both organoids and mouse models contain multiple cell types, requiring efficient delivery of molecules for lncRNA manipulation and drugs to the intended cell type. One strategy for targeted delivery is the encapsulation of drugs inside polymeric or vesicular nanoparticles coated with ligands recognized by receptors upregulated on cancer cells. Fine-tuning of nanoparticle dimensions prevents renal filtration and phagocytosis. SiRNAs can also be delivered alongside therapeutic drugs (Ciftci et al., 2025), being convenient for use in animal models. Targeted delivery has several advantages over conventional delivery: it increases the solubility, chemical stability, and half-life of the drug. Furthermore, specific targeting allows delivery to the target site at better controlled rates and reduces its systemic distribution, decreasing any side effects (Tewabe et al., 2021). Furthermore, the use of expression vectors with cell type-specific promoters (Zheng & Baum, 2008) could enable lncRNA overexpression only in specific tissues. Thus, targeted drug delivery should be preferred in both organoid and animal research. Spheroids: simple 3D models In most reviewed studies, spheroids were simple, homotypic, and assembled in ULA plates, without ECM, perfusion, or co-culture conditions. Nevertheless, they generally showed a decreased and more heterogeneous response to chemotherapeutic drugs relative to 2D-cultured cells. This critical difference increases the similarity of the spheroid to a real tumor, with the advantage of being straightforward and inexpensive to generate. Some studies injected spheroids (or cells obtained from spheroids) directly into mice (Wang et al., 2020), while others used cells from mouse xenografts to generate spheroid. Furthermore, some studies used primary cell lines from patients to generate spheroids. Some studies even correlated therapy resistance in spheroid and/or animal models with the patient’s respective response. It was observed that overexpressing or silencing specific lncRNAs could modulate the efficiency of formation and dimensions of spheroids. Of note, the effects on spheroid size were consistent with those observed in mouse xenografts. Despite the versatility of spheroids as 3D cellular models, most of the reviewed studies only tested whether lncRNA overexpression or silencing affected spheroid formation, assessed as spheroid number and size (through a spheroid formation assay [SFA]). In these studies, increased stemness was associated with increased spheroid formation efficiency. Only a relatively small fraction of these studies utilized spheroids for testing experimental hypotheses, such as variations in drug sensitivity. In general, no additional experiments were performed using these 3D cellular systems. As such, most studies involving spheroids did not fully exploit this model to further study lncRNA function, interactions, localization, or any other properties. For example, many of these studies did not perform drug testing in 3D in vitro models to assess the potential of lncRNA manipulation toward a neoadjuvant therapy. Nevertheless, in future studies, their complexity could be further increased by introducing more cell types, distinct matrix components, or even perfusion; therefore, increasing their similarity to native tumors. Future studies using spheroids should take advantage of compatible techniques, including lncRNA sequencing using ribosome-bound RNA (Cornelissen et al., 2024), RNA-seq and single-cell RNA-seq (Yoon et al., 2020), or time-course visualization of protrusions/detached clusters and associated gene expression patterns (Li et al.,2024c). Organoids: more complex 3D models Organoids are more complex 3D cellular models than spheroids, with the advantage of being able to better reflect patient features, such as their personal level of resistance to chemotherapy. Unlike spheroids, a significant percentage of the studies involving organoids tested variations in radiation or drug sensitivity associated with lncRNA overexpression or silencing. As in spheroids, organoids also revealed consistent responses to lncRNA overexpression or silencing relative to mouse xenografts, which provides robustness to in vivo conclusions. The value of organoids is strengthened by their compatibility with advanced technologies. Differences in lncRNA expression levels in different cells (or subcellular locations) of the same organoid can be assessed using spatial transcriptomics (Williams et al., 2022). This approach is broadly divided into imaging-based (visualization of single RNA molecules by fluorescence microscopy) and sequencing-based (identification of RNA sequences isolated from specific cells or subcellular locations) technologies. A common requirement is the isolation of intact and viable cells, which may be technically challenging. In addition to profiling lncRNA expression, identifying novel lncRNAs that affect organoid phenotype is another important goal. CRISPR screenings (Y. He et al., 2026) are a powerful approach that uses CRISPR-associated proteins to mutate or affect the expression of multiple genes (using multiple corresponding guide RNAs), to uncover those associated with cell proliferation or drug sensitivity/resistance. CRISPR screenings could be used to uncover lncRNAs with therapeutic potential in cancer, using 3D cellular models. They may also be further adapted for lncRNA silencing or overexpression. For example, lncGRS-1 was identified through a CRISPR screening, which was further adapted for its silencing in organoids (Liu et al., 2020b). Thus, spatial transcriptomics and CRISPR screenings are innovative resources that enhance the capabilities of organoids and their potential to decrease animal use in research. Organoids could be further used to test additional properties, such as branching morphogenesis, apoptosis (Arun et al., 2016), exosome transcriptome sequencing (Zhou et al., 2024a), and in situ RNA detection methods that allow the visualization of the spatial distribution of gene expression patterns (Liu et al., 2021b). These examples demonstrate that these 3D models are compatible with the study of multiple parameters and the use of various techniques, which is an advantage for their capacity to replace animal models. We found a larger number of studies combining organoids with mouse xenografts than studies using organoids solely. This difference reflects the necessity of animal models to assess parameters that cannot be assessed by organoids, such as metastasis and animal survival. Nevertheless, their combined use allows assessing how they complement each other and identifying some of the aspects in which the animal model might be (at least in part) replaced with a cognate organoid. In vitro 3D and animal models: advantages, limitations, and possibilities Unlike animal models, both spheroids generated from human cells and patient-derived organoids are representative of the human species, providing ideal models of human diseases. They are critically important in studies of lncRNAs that are not conserved in mouse or other animal species, such as lncGRS-1. Spheroids and organoids are also advantageous relative to animal models for the study of gene mutations, considering the laborious process of generating genetically modified animal models. In Table 4, we provide a summary list of advantages and limitations of spheroids and organoids. As listed in Table 4, spheroids and organoids have intrinsic limitations, including the lack of a full immune system, as well as the absence or very incomplete stromal and vascular structures. Thus, studying the complexity of animal tumors and associated properties, such as the tumor microenvironment and angiogenesis, is limited in 3D cellular models. 3D cellular models can also have significant batch variability and be affected by variations in culture conditions, with impacts on reproducibility and standardization. The time and cost requirements associated with 3D cellular models, as well as their technical complexity, can be additional limitations. However, animal models also have considerable costs and requirements associated. In the near future, the development and commercial availability of ready-to-use 3D cellular models should contribute to decreasing their overall costs and technical challenges. Of note, the technical challenges of generating genetically modified animal models have decreased since biotechnological companies have started selling these resources. A similar outcome could be possible regarding 3D cellular models. Table 4Comparative assessment of 3D and animal models for lncRNA study Criterium3D cancer cellular models (spheroids and organoids)Animal modelsEthics requirementsSimple, required mainly if using patient cellsMandatory, as well as specific trainingMonetary costSpheroids are usually cheaper to prepare than organoidsAnimal models have higher costs than 3D cellular modelsRepresentativeness of human patientsYes. Any desired cell line may be used—human or animal, immortalized or primary. Organoids are more representative of the supra-cellular organization of organs, than spheroidsLimitedManipulation of lncRNA levels and fine-tuning of its deliveryStraightforward, direct, several approaches availableMore challenging to achieve in the whole organism or in specific body partsVisualization of lncRNA spatial locationLimited by cellular densityUsually requires animal sacrifice, organ removal, and further processingFeasibility of achieving statistically significant dataAssays in multiwell plates can simultaneously provide sufficient biological replicates for statistically robust conclusionsThe use and sacrifice of multiple animals is required for statistically robust resultsImmune response testingLimited; immune cells can be added and their response assessedUnrestricted, but with limited representativeness of the human speciesAssessment of the extracellular matrix and tumor microenvironmentLimited; spheroids are more customizable in both stiffness, chemical components, and cells, as compared to organoidsUnrestrictedVariability among biological replicatesDependent on the experimental assay, but possibly lower than in animal modelsSignificant variability among genetically identical animalsAssessment of lncRNA manipulation effect on the size of the “tumor”Identical Table 4Comparative assessment of 3D and animal models for lncRNA study Criterium3D cancer cellular models (spheroids and organoids)Animal modelsEthics requirementsSimple, required mainly if using patient cellsMandatory, as well as specific trainingMonetary costSpheroids are usually cheaper to prepare than organoidsAnimal models have higher costs than 3D cellular modelsRepresentativeness of human patientsYes. Any desired cell line may be used—human or animal, immortalized or primary. Organoids are more representative of the supra-cellular organization of organs, than spheroidsLimitedManipulation of lncRNA levels and fine-tuning of its deliveryStraightforward, direct, several approaches availableMore challenging to achieve in the whole organism or in specific body partsVisualization of lncRNA spatial locationLimited by cellular densityUsually requires animal sacrifice, organ removal, and further processingFeasibility of achieving statistically significant dataAssays in multiwell plates can simultaneously provide sufficient biological replicates for statistically robust conclusionsThe use and sacrifice of multiple animals is required for statistically robust resultsImmune response testingLimited; immune cells can be added and their response assessedUnrestricted, but with limited representativeness of the human speciesAssessment of the extracellular matrix and tumor microenvironmentLimited; spheroids are more customizable in both stiffness, chemical components, and cells, as compared to organoidsUnrestrictedVariability among biological replicatesDependent on the experimental assay, but possibly lower than in animal modelsSignificant variability among genetically identical animalsAssessment of lncRNA manipulation effect on the size of the “tumor”Identical Comparative assessment of 3D and animal models for lncRNA study Concerning the xenograft studies, we noted that, in many of them, only tumor growth was assessed. Only a fraction of these studies assessed metastasis, animal survival rates, or any additional in vivo parameters. As such, not all xenograft research took full advantage of their animal models, providing results that could have been mimicked using spheroids, organoids, or other 3D cellular systems. Overall, our literature survey has uncovered that spheroids, organoids, and mouse xenografts have similar responses to lncRNA overexpression or silencing (Fig. 1).Fig. 1Cancer spheroids and patient-derived organoids can mimic the response of mouse tumors to long noncoding RNA (lncRNA) overexpression or silencing. Immortalized cancer cells can be used to generate spheroids or mouse tumors, through self-aggregation or their implantation in mice, respectively. Organoids can be generated from any patient cells through their differentiation, and, less frequently, patient cancer cells can self-aggregate into spheroids in vitro. Cancer spheroids, mouse tumors, and patient-derived organoids generally share a common feature: the overexpression or silencing of specific lncRNAs inhibits their growth and enhances their drug sensitivity. While spheroids and organoids suffice to mimic tumor regression, animal models remain necessary to monitor in vivo parameters, including metastasis and survival rates. Original illustration created by the authors Cancer spheroids and patient-derived organoids can mimic the response of mouse tumors to long noncoding RNA (lncRNA) overexpression or silencing. Immortalized cancer cells can be used to generate spheroids or mouse tumors, through self-aggregation or their implantation in mice, respectively. Organoids can be generated from any patient cells through their differentiation, and, less frequently, patient cancer cells can self-aggregate into spheroids in vitro. Cancer spheroids, mouse tumors, and patient-derived organoids generally share a common feature: the overexpression or silencing of specific lncRNAs inhibits their growth and enhances their drug sensitivity. While spheroids and organoids suffice to mimic tumor regression, animal models remain necessary to monitor in vivo parameters, including metastasis and survival rates. Original illustration created by the authors Potential to replace animal models We propose that spheroids and organoids could fully replace xenograft animal models in studies in which only tumor growth is assessed. The use of co-cultures would also partially allow the study of the immune response (a common reason for choosing animal models), which would be important in the study of immune-related lncRNAs. Furthermore, future developments and upgrades in 3D cellular models are likely to decrease even further the need for animal models in biomedical research. The combination of spheroids with organoids could be one such development, which could allow testing invasive properties associated with cancer. However, in a specific study, brain-derived organoids (not composed of cancer cells) were co-cultured with glioblastoma spheroids. The goal was to evaluate the extent to which the cancer cells from the spheroid would (or not) metastasize to the organ. Parameters including the number of invading cancer cells and invaded organoid area were measured (M. Han et al., 2020). Therefore, the combination of the two 3D models enables the assessment of properties that cannot be assessed separately in one 3D model. Furthermore, microfluidic 3D cellular systems are still poorly explored in lncRNA cancer research and could be further used. For example, in breast cancer, a microfluidic 3D co-culture system was used to assess cancer cell migration under hypoxia. In this study, H19 silencing decreased cell migration, which was mediated by H19 and hypoxia (An et al., 2022). Specifically, this microfluidic 3D cellular model allowed the assessment of hypoxia effects. Therefore, more advanced 3D cellular models are likely to further decrease the need for animal models. In addition to decreasing the use of mice, non-rodents such as zebrafish could be further used. Zebrafish xenograft models could provide similar results when compared with 3D cellular models and evaluate the same features and properties in a shorter time period and for a lower material investment (Y. Liu et al., 2023). As a representative study, in a zebrafish xenograft, the long intergenic non-protein coding RNA 52 (LINC00052) was shown to be downregulated when breast cancer formed spheroids; LINC00052 downregulation increased cell migration in xenografted zebrafish embryos (Sanchez-Lopez et al., 2021). Further research should uncover the extent to which mice could be replaced with zebrafish as an animal model. In the present review, we focused on lncRNAs in cancer research. Our findings are likely to apply to other types of RNA (such as messenger RNAs or micro-RNAs) and possibly to other biomedical research fields in which animal models are involved, such as metabolism and/or neuroscience.