Patients and samples
Tumor specimens were obtained from patients with PAs, who underwent transsphenoidal surgery at the Department of Neurosurgery of the Guangdong Provincial People's Hospital (Guangzhou, China) between January 2020 and June 2021. The diagnosis of PAs was based on clinical manifestations, biochemical features of hormonal secretion, magnetic resonance imaging (MRI), and histopathological analyses confirmed by two pathologists after surgical resections. NIPA is defined as the limitation of tumor mass within the sellar region, without any compression on peripheral structures (Fig. 1A, B). The definition of IPA is according to Knosp classifications into grades III–IV  (Fig. 1C, D). IPAs (n = 3) and NIPAs (n = 3) were selected for lncRNA microarrays. The details of these six PAs are shown in Table 1. Also, another 41 specimens of PAs, including IPAs (n = 21) and NIPAs (n = 16), were used for the validation by qRT-PCR. The clinical characteristics of the 41 patients with PAs are summarized in Table 2. Tumor dimensions were manually obtained from MRI. A microadenoma was defined by a maximal tumor diameter of < 10 mm; a macroadenoma was ≥ 10 mm; a large macroadenoma was ≥ 20 mm; while a giant adenoma was ≥ 40 mm. The dimensional indices of the tumors were measured and recorded in three orthogonal planes, namely: transverse (TR), anteroposterior (AP), and craniocaudal (CC). The tumor volumes were estimated using the following formula:V = (π × [TR × AP × CC])/6 .After surgical excision, all tissue samples were immediately frozen in liquid nitrogen and stored at −80 °C for further analyses. All procedures of this study were approved by the Ethics Committee of Guangdong Provincial People's Hospital. Informed consent was obtained from all patients.
Total RNA extraction and purification
Total RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, USA) according to the manufacturer’s protocol. Furthermore, the quantification of total RNA was evaluated by Bioanalyzer 2200 (Agilent, California, USA) and kept at −80 °C. The RNA samples with RNA integrity number (RIN) > 6.0 are acceptable for rRNA depletion and subsequent lncRNA purification. The purification of total RNA was validated by gel electrophoresis.
cDNA library construction
cDNA libraries were constructed for each pooled RNA sample using the NEBNext® Ultra™ Directional RNA Library Prep Kit (New England BioLabs Inc., MA, US) for Illumina according to the manufacturer’s instructions. The protocol comprises the following steps: mRNA molecules are fragmented into 150–200 bp using divalent cations at 94 °C for 8 min. The cleaved RNA fragments as templates were reverse-transcribed into first-strand cDNA. Subsequently, the second-strand cDNA was synthesized using Polymerase I and RNase H with reaction buffer. Target bands were harvested through AMPure XP Beads (Beckman coulter). The products were then purified and enriched by PCR to create the final cDNA libraries and quantified by Agilent 2200. The tagged cDNA libraries were pooled in equal ratio and used for 150 bp paired-end sequencing in a single lane of the Illumina HiSeq XTen. The experiments of library construction and RNA sequencing were completed at the Center of NovelBio Lab center of Novelbio lab (Shanghai, China).
Mapping and identification of differentially expressed genes
Before read mapping, clean reads were obtained from the raw reads by removing the adaptor sequences, reads with > 5% ambiguous bases (noted as N), and low-quality reads containing more than 20% of bases with qualities of < 20. The clean reads were then aligned to the human genome (version: GRCh38 NCBI) using the HISAT2 . HTSeq was used to count gene and lncRNA, while reads per kilobase per million mapped reads method was used to determine gene expression . We applied DESeq algorithm  to filter the differentially expressed genes, after the analysis of the level of significance, i.e, determination of P value, and false discovery rate (FDR) analysis under the following criteria: 1) fold change > 2 or < 0.5; 2) FDR < 0.05 .
Functional enrichment analysis
Gene Ontology (GO) analysis was performed to facilitate elucidating the biological implications of unique genes . We downloaded the GO annotations from NCBI (http://www.ncbi.nlm.nih.gov/), UniProt (http://www.uniprot.org/), and the Gene Ontology (http://www.geneontology.org/). Fisher’s exact test was applied to identify the significant GO categories and FDR was used to correct the P values. The Gene Ontology is structured as a directed acyclic graph, and each term has defined relationships to one or more other terms. GO-Tree was built based on the Gene Ontology Directed Acyclic Graph to provide user-friendly data navigation and visualization. We selected the significant GO-Term (P value < 0.01) in GO analysis based on the differentially expressed genes to construct the GO-Tree for summarizing the function affected in the experiment .
Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed to clarify intuitively the role and significance of target genes in the overall biological pathways . We selected the genes in enriched biological pathways and using Cytoscape for graphical representations of pathways . The KEGG database was used to build the network of genes according to the relationship among the genes, proteins, and compounds in the database.
Construction of the lncRNA-miRNA-target gene interaction network
The role of lncRNAs in IPAs was investigated by a lncRNA-miRNA-target gene interaction network. According to the lncRNA microarray results, the 10 most dysregulated lncRNAs in IPAs were selected and Cytoscape software was performed to map out an interaction network. Putative interactions between lncRNAs and miRNAs were predicted using the online databases Jefferson Computational Medicine Center-RNA22 v2 microRNA target detection (https://cm.jefferson.edu/rna22/Interactive/) and LncBase Predicted v.2 (http://carolina.imis.athena-innovation.gr/diana_tools/web/index.php?r=lncbasev2%2Findex-predicted). Then, the miRNA with the highest target scores were selected, and their target genes were evaluated by TargetScan  and miRanda . Finally, miRNAs and their target genes with high targeting-relationship scores were selected to construct the lncRNA-miRNA-mRNA interaction network. The interaction network was delineated using the Cytoscape software.
qRT-PCR assay for the validation
Total RNA was extracted using Trizol reagent (Invitrogen) for 41 specimens of PAs. Reverse transcription and qRT-PCR were performed using a Reverse Transcription Kit (Takara, Dalian, China) and PrimeScript RT Reagent Kit (Takara, Dalian, China), respectively, as previously described . The expression of lncRNA was measured by qRT-PCR. The sequences of the primers are listed in Table 3. The gene expression levels were normalized to actin. Gene expression levels were determined by the 2−ΔΔCt method and analyzed for statistical significance.
Measurement data are presented as mean ± standard error of the mean (SEM) and enumeration data are presented as percentages. Comparisons were performed using independent sample t-test between pairs of groups or one-way analysis of variance for more than two groups followed by Dunnett’s multiple comparison test. The receiver operating characteristic (ROC) curves and the area under the curve (AUC) were used to estimate the diagnostic power and accuracy of lncRNAs on invasive/noninvasive PAs. All statistical analyses were performed on Statistical Product and Service Solutions (SPSS) 25.0 software (SPSS Inc., Chicago, IL, USA). P < 0.05 was considered statistically significant, which is indicated in the figures.