Identification of two subtypes and prognostic characteristics of lung adenocarcinoma based on pentose phosphate metabolic pathway-related long non-coding RNAs.

Abstract

Retraction The abstract by Wu et al entitled, “Identification of Two Subtypes and Prognostic Characteristics of Lung Adenocarcinoma based on Pentose Phosphate Metabolic Pathway-related Long Non-Coding RNAs,” ( Journal of Clinical Oncology 40, no. 16, suppl e20531) published on June 1, 2022, has been retracted due to statistical errors that significantly changed the results and conclusion of the abstract. This abstract was retracted on November 4, 2022. e20531 Background: The most frequent histological lung cancer subtype associated with a high mortality rate is lung adenocarcinoma (LUAD). Thus, its typing is crucial for the assessment of patient characteristics. This study aimed to analyze the differences in subtypes and characteristics of advanced (Stage II, III, and IV) LUAD patients based on the pentose phosphate metabolic pathway-related long non-coding RNAs (lncRNAs), along with their potential regulatory mechanisms. Methods: Using the expression profiling and corresponding clinical information of LUAD patients from Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA), single-sample gene set enrichment analysis (ssGSEA) was used to determine the scores for the metabolic pathways for all samples. In addition, differential pathway scores between normal and tumor samples from TCGA were identified by rank-sum tests. Prognosis-related metabolic pathways were identified by univariate Cox analyses for both datasets. Pearson correlation coefficients between PENTOSE PHOSPHATE scores of the pentose phosphate samples and the lncRNAs of the corresponding datasets were calculated. Next, a consistency matrix was constructed by consistency clustering and the samples were clustered and typed using the KM algorithm using euclidean as the metric distance. Next, the clusterProfiler software package was used for functional annotation. Finally, functional evaluation was performed by analysis of the transcription factor (TF) activity and first-order partial correlation computation. Results: By clustering pentose phosphate-related lncRNAs from LUAD samples, we obtained two molecular subtypes of LUAD, C1, and C2. The C1 subtype was associated with a lower PENTOSE PHOSPHATE score and a good prognosis; conversely, the C2 subtype showed a higher PENTOSEPHOSPHATE score and was related to poorer prognoses. The findings of the gene set enrichment analysis (GSEA) suggested that C2 was markedly associated with energy metabolic pathways. In addition, the levels of expression of most immune cells were markedly higher in the C1 subtype relative to the C2 subtype. Some crucial immune checkpoints, including CTLA4, CD274, and CD47, were also significantly upregulated in the C1 subtype, leading to a higher score for clinical effect in the C1 subtype. Finally, one TF, BACH1, was found to be significantly upregulated in C1 subtypes; the pathways activated by this TF may be associated with tumor progression and poor prognoses. Conclusions: Precise typing of LUAD based on pentose phosphate metabolic pathway-related lncRNAs was confirmed. The differences in the characteristics between C1 and C2 subtypes may contribute to a better understanding of LUAD, thus yielding new targets for detection and treatment.