A Glycolysis Outsider Steps into the Cancer Spotlight

Transcriptional regulation and post-translational modifications in the glycolytic pathway for targeted cancer therapy

Cancer cells largely rely on aerobic glycolysis or the Warburg effect to generate essential biomolecules and energy for their rapid growth. The key modulators in glycolysis including glucose transporters and enzymes, e.g. hexokinase 2, enolase 1, pyruvate kinase M2, lactate dehydrogenase A, play indispensable roles in glucose uptake, glucose consumption, ATP generation, lactate production, etc. Transcriptional regulation and post-translational modifications (PTMs) of these critical modulators are important for signal transduction and metabolic reprogramming in the glycolytic pathway, which can provide energy advantages to cancer cell growth. In this review we recapitulate the recent advances in research on glycolytic modulators of cancer cells and analyze the strategies targeting these vital modulators including small-molecule inhibitors and microRNAs (miRNAs) for targeted cancer therapy. We focus on the regulation of the glycolytic pathway at the transcription level (e.g., hypoxia-inducible factor 1, c-MYC, p53, sine oculis homeobox homolog 1, N6-methyladenosine modification) and PTMs (including phosphorylation, methylation, acetylation, ubiquitination, etc.) of the key regulators in these processes. This review will provide a comprehensive understanding of the regulation of the key modulators in the glycolytic pathway and might shed light on the targeted cancer therapy at different molecular levels.

Glucose-lightened upconversion nanoprobes for accurate cellular-discrimination based on Warburg effect

Accurate cellular-recognition based disease therapy is of significance for precision medicine. However, except of specific antibody-coupling strategy, very few probes have been reported to efficiently discriminate normal cells and lesion cells through cellular microenvironment. Herein, we proposed a glucose selectively-lightened upconversion nanoprobe to recognize cancer cells from a pile of normal cells based on Warburg effect, that indicated a heightened demand for glucose intake for cancer cells. The nanoprobes were constructed by mesoporous silica-coated upconversion nanoparticles (UCNP@mSiO2) with the crucial incorporation of a glucose-responsive modality, benzoboric acid (BA)-modified fluorescein molecules (FITC-BA). In cancer cells, the presence of elevated glucose concentrations triggered the transformation of FITC-BA to FITC-Glucose to recover nanoprobes' luminescence, however, the nanoprobes exhibited a shielded luminescent effect in healthy cells. To validate the hypothesis of accurate cellular-discrimination, a photodynamic therapy modality, riboflavin, with a specific ratio were also loaded into above UCNP@mSiO2 nanoprobes for effective production of reactive oxygen species to kill cells. It was found that 97.8% of cancer cells were cleaned up, but normal cells retained a nearly 100% viability after 10 min laser illumination. By leveraging the metabolic disparity from Warburg effect, the nanoprobes offer a highly accurate cellular discrimination, and significantly mitigate "off-target" damage commonly associated with conventional therapies.

Genome-wide neonatal epigenetic changes associated with maternal exposure to the COVID-19 pandemic

BACKGROUND

During gestation, stressors to the fetus, including viral exposure or maternal psychological distress, can fundamentally alter the neonatal epigenome, and may be associated with long-term impaired developmental outcomes. The impact of in utero exposure to the COVID-19 pandemic on the newborn epigenome has yet to be described.

METHODS

This study aimed to determine whether there are unique epigenetic signatures in newborns who experienced otherwise healthy pregnancies that occurred during the COVID-19 pandemic (Project RESCUE). The pre-pandemic control and pandemic cohorts (Project RESCUE) included in this study are part of a prospective observational and longitudinal cohort study that evaluates the impact of elevated prenatal maternal stress during the COVID-19 pandemic on early childhood neurodevelopment. Using buccal swabs collected at birth, differential DNA methylation analysis was performed using the Infinium MethylationEPIC arrays and linear regression analysis. Pathway analysis and gene ontology enrichment were performed on resultant gene lists.

RESULTS

Widespread differential methylation was found between neonates exposed in utero to the pandemic and pre-pandemic neonates. In contrast, there were no apparent epigenetic differences associated with maternal COVID-19 infection during pregnancy. Differential methylation was observed among genomic sites that underpin important neurological pathways that have been previously reported in the literature to be differentially methylated because of prenatal stress, such as NR3C1.

CONCLUSIONS

The present study reveals potential associations between exposure to the COVID-19 pandemic during pregnancy and subsequent changes in the newborn epigenome. While this finding warrants further investigation, it is a point that should be considered in any study assessing newborn DNA methylation studies obtained during this period, even in otherwise healthy pregnancies.

Circular RNA Ubiquitin-associated Protein 2 Silencing Suppresses Bladder Cancer Progression by Downregulating DNA Topoisomerase 2-alpha Through Sponging miR-496

Background

Circular RNAs (circRNAs) have been uncovered to be implicated in the malignant development of bladder cancer (BC).

Objective

Herein, this work aimed to investigate the role and mechanism of circRNA ubiquitin-associated protein 2 (circUBAP2) in BC progression.

Design setting and participants

Quantitative real-time polymerase chain reaction and Western blotting were used for the detection of genes and proteins.

Outcome measurements and statistical analysis

In vitro functional experiments were conducted using colony formation, 5-ethynyl-2'-deoxyuridine (EdU), Transwell, wound healing, and flow cytometry assays, respectively. A glycolysis analysis was conducted by assessing glucose uptake and lactate production. A murine xenograft model was established to perform in vivo experiments. The binding interaction between miR-496 and circUBAP2 or DNA topoisomerase 2-alpha (TOP2A) was verified using a dual-luciferase reporter assay.

Results and limitations

CircUBAP2 was highly expressed in BC patients, and high circUBAP2 expression showed a shorter survival rate. Functionally, knockdown of circUBAP2 could suppress BC cell growth, migration, invasion, and aerobic glycolysis in vitro, as well as impede BC growth in nude mice. Mechanistically, circUBAP2 acted as a sponge for miR-496, which targeted TOP2A. Moreover, circUBAP2 could indirectly regulate TOP2A expression through sequestering miR-496. Furthermore, a series of rescue experiments showed that miR-496 inhibition reversed the anticancer action of circUBAP2 knockdown on BC cells. Moreover, miR-496 could attenuate BC cell malignant phenotypes and aerobic glycolysis, which were abolished by TOP2A overexpression.

Conclusions

Silencing of circUBAP2 could suppress BC growth, invasion, migration, and aerobic glycolysis by the miR-496/TOP2A axis, suggesting a promising target for the molecular targeted therapies of BC.

Patient summary

Circular RNA ubiquitin-associated protein 2 (circUBAP2) was found to be associated with poor prognosis in bladder cancer (BC). Knockdown of circUBAP2 might suppress BC growth, invasion, migration, and aerobic glycolysis, indicating that it may be a new target for the development of molecular targeted therapy for BC.

A Machine Learning Analysis of Prognostic Genes Associated With Allograft Tolerance After Renal Transplantation

In this study, we aimed to identify transplantation tolerance (TOL)-related gene signature and use it to predict the different types of renal allograft rejection performances in kidney transplantation. Gene expression data were obtained from the Gene Expression Omnibus (GEO) database, differently expressed genes (DEGs) were performed, and the gene ontology (GO) function enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were also conducted. The machine learning methods were combined to analyze the feature TOL-related genes and verify their predictive performance. Afterward, the gene expression levels and predictive performances of TOL-related genes were conducted in the context of acute rejection (AR), chronic rejection (CR), and graft loss through heatmap plots and the receiver operating characteristic (ROC) curves, and their respective immune infiltration results were also performed. Furthermore, the TOL-related gene signature for graft survival was conducted to discover gene immune cell enrichment. A total of 25 TOL-related DEGs were founded, and the GO and KEGG results indicated that DEGs mainly enriched in B cell-related functions and pathways. 7 TOL-related gene signature was constructed and performed delightedly in TOL groups and different types of allograft rejection. The immune infiltration analysis suggested that gene signature was correlated with different types of immune cells. The Kaplan-Meier (KM) survival analysis demonstrated that BLNK and MZB1 were the prognostic TOL-related genes. Our study proposed a novel gene signature that may influence TOL in kidney transplantation, providing possible guidance for immunosuppressive therapy in kidney transplant patients.

Interaction between glycolysis‒cholesterol synthesis axis and tumor microenvironment reveal that gamma-glutamyl hydrolase suppresses glycolysis in colon cancer

Background

Metabolic reprogramming is a feature of cancer. However, colon cancer subtypes based on the glycolysis‒cholesterol synthesis axis have not been identified, and little is known about connections between metabolic features and the tumor microenvironment.

Methods

Data for 430 colon cancer cases were extracted from The Cancer Genome Atlas, including transcriptome data, clinical information, and survival outcomes. Glycolysis and cholesterol synthesis-related gene sets were obtained from the Molecular Signatures Database for a gene set variation analysis. The relationship between the genomic landscape and immune landscape were investigated among four metabolic subtypes. Hub genes were determined. The clinical significance of candidate hub gene was evaluated in 264 clinical samples and potential functions were validated in vitro and in vivo.

Results

Colon cancer cases were clustered into four metabolic subtypes: quiescent, glycolytic, cholesterogenic, and mixed. The metabolic subtypes differed with respect to the immune score, stromal score, and estimate score using the ESTIMATE algorithm, cancer-immunity cycle, immunomodulator signatures, and signatures of immunotherapy responses. Patients in the cholesterogenic group had better survival outcomes than those for other subtypes, especially glycolytic. The glycolytic subtype was related to unfavorable clinical characteristics, including high mutation rates in TTN, APC, and TP53, high mutation burden, vascular invasion, right colon cancer, and low-frequency microsatellite instability. GGH, CACNG4, MME, SLC30A2, CKMT2, SYN3, and SLC22A31 were identified as differentially expressed both in glycolytic-cholesterogenic subgroups as well as between colon cancers and healthy samples, and were involved in glycolysis‒cholesterol synthesis. GGH was upregulated in colon cancer; its high expression was correlated with CD4⁺ T cell infiltration and longer overall survival and it was identified as a favorable independent prognostic factor. The overexpression of GGH in colon cancer-derived cell lines (SW48 and SW480) inhibited PKM, GLUT1, and LDHA expression and decreased the extracellular lactate content and intracellular ATP level. The opposite effects were obtained by GGH silencing. The phenotype associated with GGH was also validated in a xenograft nude mouse model.

Conclusions

Our results provide insight into the connection between metabolism and the tumor microenvironment in colon cancer and provides preliminary evidence for the role of GGH, providing a basis for subsequent studies.

CircLPAR3 knockdown suppresses esophageal squamous cell carcinoma cell oncogenic phenotypes and Warburg effect through miR-873-5p/LDHA axis

BACKGROUND

Circular RNAs (circRNAs) have been identified to participate in regulating multiple malignancies. Herein, this study aimed to explore the clinical significance, biological function, and regulatory mechanisms of circRNA lysophosphatidic acid receptor 3 (circLPAR3) in esophageal squamous cell carcinoma (ESCC) cell malignant phenotypes and Warburg effect.

METHODS

The qRT-PCR and Western blot were used to detect the levels of genes and proteins. Glucose uptake and lactate production were detected to determine the Warburg effect. The effects of circLPAR3 on ESCC cell proliferation, apoptosis, and metastasis were evaluated by MTT, 5-ethynyl-2'-deoxyuridine (EdU), flow cytometry, wound healing, and transwell assays. The binding interaction between miR-873-5p and circLPAR3 or lactate dehydrogenase A (LDHA) was verified using dual-luciferase reporter and RIP assays. Xenograft mice models were established to conduct in vivo analysis.

RESULTS

CircLPAR3 is a stable circRNA and was increased in ESCC tissues and cells. Functionally, circLPAR3 knockdown suppressed ESCC cell Warburg effect, proliferation, metastasis, and induced apoptosis in vitro, and impeded xenograft tumor growth and Warburg effect in ESCC mice models. Mechanistically, circLPAR3 served as a sponge for miR-873-5p, which targeted LDHA. Moreover, circLPAR3 could regulate LDHA expression by sponging miR-873-5p. Thereafter, rescue experiments suggested that miR-873-5p inhibition reversed the anticancer effects of circLPAR3 silencing on ESCC cells. Furthermore, miR-873-5p overexpression restrained ESCC cell Warburg effect and oncogenic phenotypes, which were abolished by LDHA up-regulation.

CONCLUSION

CircLPAR3 knockdown suppressed ESCC cell growth, metastasis, and Warburg effect by miR-873-5p/LDHA axis, implying a promising molecular target for ESCC therapy.

Mitochondrial-targeting nanoprodrugs to mutually reinforce metabolic inhibition and autophagy for combating resistant cancer

Abnormal energy metabolism is one of the hallmarks of cancer and closely linked to therapy resistance. However, existing metabolic inhibitors suffer from inefficient cell enrichment and therapeutic effects. In this work, we developed an effective strategy to mutually reinforce the metabolic inhibition and autophagy for enhanced tumor killing efficacy and combating resistant cancer. First, mitochondrial homing moiety triphenylphosphonium and metabolic inhibitor lonidamine were grafted onto polylysine. After self-assembly of this functionalized polylysine, ferrocene and glucose oxidase were immobilized to afford additional chemotherapy functions, and the final product was named as FG/T-Nanoprodrug. Effective mitochondrial targeting and metabolic inhibition were observed in resistant cancer cells. In addition, owing to the inhibited metabolism, less glucose is consumed to allow FG/T-Nanoprodrug to produce excess reactive oxygen species (ROS) by glucose oxidase and ferrocene. The enhanced chemodynamic therapy increases the mitochondrial permeability to promote the release of cytochrome c from mitochondria, ultimately induces high levels of autophagy. The FG/T-Nanoprodrug demonstrated superior mutually reinforcing of metabolic inhibition (up to 3.7-fold compared to free lonidamine) and autophagy (up to 125.3-fold compared to free lonidamine) to effectively kill resistant cancer cell both in vitro and in vivo. Overall, this strategy could pave a new way to efficient treatment of resistant cancer and other metabolically abnormal diseases.

PFKFB4 is overexpressed in clear-cell renal cell carcinoma promoting pentose phosphate pathway that mediates Sunitinib resistance

BACKGROUND

Kinases play critical role in clear-cell renal cell carcinoma (ccRCC). We aim to exploit novel kinase that is both protumorigenic and drugable in ccRCC.

METHODS

Reproduction of public datasets with validation using microarray was performed to identify candidate gene. Functionality was studied using multi-omics with validation in vitro and in vivo.

RESULTS

6-Phosphofructo-2-Kinase/Fructose-2,6-Biphosphatase 4 (PFKFB4) was differentially expressed showing significantly higher expression in tumor than in normal kidney. PFKFB4 overexpression was associated with advanced tumor grade, stage and worsened prognosis. PFKFB4-knockdown significantly impaired fitness in cell proliferation, migration and wound healing. Despite being recurrently deleted on 3p, PFKFN4 mRNA remained actively transcribed by HIF1α. Metabolomics showed overexpressed PFKFB4 showed enriched metabolites in pentose phosphate pathway (PPP). Phosphoproteomics and immunoprecipitation showed PFKFB4 also phosphorylated NCOA3 which interacted with FBP1 to counteract overactive PPP flux, forming a regulatory loop. PFKFB4-knockdown overcame resistance to Sunitinib in vitro and in vivo both in xenograft and tail-vein injection murine models.

CONCLUSION

We concluded PFKFB4 was associated with PPP activity and the fine-tuning of which was mediated by its phosphorylation of NCOA3. Targeting PFKFB4 held promise to combat resistance to Sunitinib.

SRC-3, a Steroid Receptor Coactivator: Implication in Cancer

Steroid receptor coactivator-3 (SRC-3), also known as amplified in breast cancer 1 (AIB1), is a member of the SRC family. SRC-3 regulates not only the transcriptional activity of nuclear receptors but also many other transcription factors. Besides the essential role of SRC-3 in physiological functions, it also acts as an oncogene to promote multiple aspects of cancer. This review updates the important progress of SRC-3 in carcinogenesis and summarizes its mode of action, which provides clues for cancer therapy.

Identification and validation of a six-gene signature associated with glycolysis to predict the prognosis of patients with cervical cancer

BACKGROUND

Cervical cancer (CC) is one of the most common gynaecological cancers. The gene signature is believed to be reliable for predicting cancer patient survival. However, there is no relevant study on the relationship between the glycolysis-related gene (GRG) signature and overall survival (OS) of patients with CC.

METHODS

We extracted the mRNA expression profiles of 306 tumour and 13 normal tissues from the University of California Santa Cruz (UCSC) Database. Then, we screened out differentially expressed glycolysis-related genes (DEGRGs) among these mRNAs. All patients were randomly divided into training cohort and validation cohort according to the ratio of 7: 3. Next, univariate and multivariate Cox regression analyses were carried out to select the GRG with predictive ability for the prognosis of the training cohort. Additionally, risk score model was constructed and validated it in the validation cohort.

RESULTS

Six mRNAs were obtained that were associated with patient survival. The filtered mRNAs were classified into the protective type (GOT1) and the risk type (HSPA5, ANGPTL4, PFKM, IER3 and PFKFB4). Additionally, by constructing the prognostic risk score model, we found that the OS of the high-risk group was notably poorer, which showed good predictive ability both in training cohort and validation cohort. And the six-gene signature is a prognostic indicator independent of clinicopathological features. Through the verification of PCR, the results showed that compared with the normal cervial tissuses, the expression level of six mRNAs were significantly higher in the CC tissue, which was consistent with our findings.

CONCLUSIONS

We constructed a glycolysis-related six-gene signature to predict the prognosis of patients with CC using bioinformatics methods. We provide a thorough comprehension of the effect of glycolysis in patients with CC and provide new targets and ideas for individualized treatment.

Long non‑coding RNA BCYRN1 promotes glycolysis and tumor progression by regulating the miR‑149/PKM2 axis in non‑small‑cell lung cancer

Cancer cells use aerobic glycolysis to sustain their proliferation. Long non‑coding RNA brain cytoplasmic RNA 1 (BCYRN1) has been reported to act as an oncogene in non‑small‑cell lung cancer (NSCLC). The present study investigated the role of BCYRN1 in NSCLC glycolysis. BCYRN1 expression was detected in NSCLC cells and tissues using reverse transcription‑quantitative PCR. The effect of BCYRN1 on aerobic glycolysis was examined by measuring NSCLC cell glucose catabolism and lactate synthesis. The relationships between BCYRN1 and microRNA (miR)‑149, and between miR‑149 and pyruvate kinase M1/2 (PKM2) were measured using a dual‑luciferase reporter assay. Cell proliferation and invasion were analyzed by the Cell Counting kit‑8 assay and the Matrigel invasion assay, respectively. High BCYRN1 expression was observed in NSCLC tissues and cells compared with the corresponding controls. BCYRN1 induced glycolysis and upregulated the expression levels of PKM2 in NSCLC cells. In addition, BCYRN1 regulated miR‑149 expression levels, and miR‑149 inhibitor rescued the effects of si‑BCYRN1 on glucose consumption and lactate production. miR‑149 knockdown significantly enhanced the expression of PKM2. Furthermore, PKM2 inhibition significantly reversed the effects of miR‑149 inhibitor on glucose catabolism and lactate synthesis. Furthermore, PKM2 was involved in NSCLC cell proliferation and invasion, and BCYRN1 knockdown and miR‑149 overexpression inhibited both processes. The present study suggested that BCYRN1 was involved in cell glycolysis, proliferation and invasion during NSCLC via regulating miR‑149 and PKM2.

LncRNA LINRIS stabilizes IGF2BP2 and promotes the aerobic glycolysis in colorectal cancer

Background

Long noncoding RNAs (lncRNAs) play nonnegligible roles in the epigenetic regulation of cancer cells. This study aimed to identify a specific lncRNA that promotes the colorectal cancer (CRC) progression and could be a potential therapeutic target.

Methods

We screened highly expressed lncRNAs in human CRC samples compared with their matched adjacent normal tissues. The proteins that interact with LINRIS (Long Intergenic Noncoding RNA for IGF2BP2 Stability) were confirmed by RNA pull-down and RNA immunoprecipitation (RIP) assays. The proliferation and metabolic alteration of CRC cells with LINRIS inhibited were tested in vitro and in vivo.

Results

LINRIS was upregulated in CRC tissues from patients with poor overall survival (OS), and LINRIS inhibition led to the impaired CRC cell line growth. Moreover, knockdown of LINRIS resulted in a decreased level of insulin-like growth factor 2 mRNA-binding protein 2 (IGF2BP2), a newly found N⁶-methyladenosine (m⁶A) ‘reader’. LINRIS blocked K139 ubiquitination of IGF2BP2, maintaining its stability. This process prevented the degradation of IGF2BP2 through the autophagy-lysosome pathway (ALP). Therefore, knockdown of LINRIS attenuated the downstream effects of IGF2BP2, especially MYC-mediated glycolysis in CRC cells. In addition, the transcription of LINRIS could be inhibited by GATA3 in CRC cells. In vivo experiments showed that the inhibition of LINRIS suppressed the proliferation of tumors in orthotopic models and in patient-derived xenograft (PDX) models.

Conclusion

LINRIS is an independent prognostic biomarker for CRC. The LINRIS-IGF2BP2-MYC axis promotes the progression of CRC and is a promising therapeutic target.

Emerging roles of aerobic glycolysis in breast cancer

Altered aerobic glycolysis is a well-recognized characteristic of cancer cell energy metabolism, known as the Warburg effect. Even in the presence of abundant oxygen, a majority of tumor cells produce substantial amounts of energy through a high glycolytic metabolism, and breast cancer (BC) is no exception. Breast cancer continues to be the second leading cause of cancer-associated mortality in women worldwide. However, the precise role of aerobic glycolysis in the development of BC remains elusive. Therefore, the present review attempts to address the implication of key enzymes of the aerobic glycolytic pathway including hexokinase (HK), phosphofructokinase (PFK) and pyruvate kinase (PK), glucose transporters (GLUTs), together with related signaling pathways including protein kinase B(PI3K/AKT), mammalian target of rapamycin (mTOR) and adenosine monophosphate-activated protein kinase (AMPK) and transcription factors (c-myc, p53 and HIF-1) in the research of BC. Thus, the review of aerobic glycolysis in BC may evoke novel ideas for the BC treatment.

A Metabolism Toolbox for CAR T Therapy

The adoptive transfer of T cells expressing chimeric antigen receptors (CARs) through genetic engineering is one of the most promising new therapies for treating cancer patients. A robust CAR T cell-mediated anti-tumor response requires the coordination of nutrient and energy supplies with CAR T cell expansion and function. However, the high metabolic demands of tumor cells compromise the function of CAR T cells by competing for nutrients within the tumor microenvironment (TME). To substantially improve clinical outcomes of CAR T immunotherapy while treating solid tumors, it is essential to metabolically prepare CAR T cells to overcome the metabolic barriers imposed by the TME. In this review, we discuss a potential metabolism toolbox to improve the metabolic fitness of CAR T cells and maximize the efficacy of CAR T therapy.

6-Phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 and 4: A pair of valves for fine-tuning of glucose metabolism in human cancer

Background

Cancer cells favor the use of less efficient glycolysis rather than mitochondrial oxidative phosphorylation to metabolize glucose, even in oxygen-rich conditions, a distinct metabolic alteration named the Warburg effect or aerobic glycolysis. In adult cells, bifunctional 6-phosphofructo-2-kinase/fructose-2, 6-bisphosphatase (PFKFB) family members are responsible for controlling the steady-state cytoplasmic levels of fructose-2,6-bisphosphate, which allosterically activates 6-phosphofructo-1-kinase, the key enzyme catalyzing the rate-limiting reaction of glycolysis. PFKFB3 and PFKFB4 are the two main isoenzymes overexpressed in various human cancers.

Scope of review

In this review, we summarize recent findings on the glycolytic and extraglycolytic roles of PFKFB3 and PFKFB4 in cancer progression and discuss potential therapies for targeting of PFKFB3 and PFKFB4.

Major conclusions

PFKFB3 has the highest kinase activity to shunt glucose toward glycolysis, whereas PFKFB4 has more FBPase-2 activity, redirecting glucose toward the pentose phosphate pathway, providing reducing power for lipid biosynthesis and scavenging reactive oxygen species. Co-expression of PFKFB3 and PFKFB4 provides sufficient glucose metabolism to satisfy the bioenergetics demand and redox homeostasis requirements of cancer cells. Various reversible post-translational modifications of PFKFB3 enable cancer cells to flexibly adapt glucose metabolism in response to diverse stress conditions. In addition to playing important roles in tumor cell glucose metabolism, PFKFB3 and PFKFB4 are widely involved in multiple biological processes, such as cell cycle regulation, autophagy, and transcriptional regulation in a non-glycolysis-dependent manner.