Exploring the molecular pathogenesis associated with T-cell prolymphocytic leukemia based on a comprehensive bioinformatics analysis.

As a rare hematological malignancy, T-cell prolymphocytic leukemia (T-PLL) has a high mortality rate. However, the comprehensive mechanisms of the underlying pathogenesis of T-PLL are unknown. The purpose of the present study was to investigate the pathogenesis of T-PLL based on a comprehensive bioinformatics analysis. The differentially expressed genes (DEGs) between T-PLL blood cell samples and normal peripheral blood cell samples were investigated using the GSE5788 Affymetrix microarray data from the Gene Expression Omnibus database. To investigate the functional changes associated with tumor progression, Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analyses were used on the identified DEGs, followed by protein-protein interaction (PPI) and sub-PPI analysis. Transcription factors and tumor-associated genes (TAGs) were investigated further. The results identified 84 upregulated genes and 354 downregulated genes in T-PLL samples when compared with healthy samples. These DEGs featured in various functions including cell death and various pathways including apoptosis. The functional analysis of DEGs revealed 17 dysregulated transcription factors and 37 dysregulated TAGs. Furthermore, the PPI network analysis based on node degree (a network topology attribute) identified 61 genes, including the core downregulated gene of the sub-PPI network, signal transducer and activator of transcription 3 (STAT3; degree, 13) and the core upregulated gene, insulin receptor substrate-1 (IRS1; degree, 5), that may have important associations with the progression of T-PLL. Alterations to cell functions, including cell death, and pathways, including apoptosis, may contribute to the process of T-PLL. Candidate genes identified in the present study, including STAT3 and IRS1, should be targets for additional studies.

Introduction

T-cell prolymphocytic leukemia (T-PLL) has an aggressive course and poor response to conventional therapy, with median survival times ranging between 7 and 30 months (1–4). Although chemotherapeutic drugs, including alemtuzumab and analogues, have significantly improved survival and response rates (5), the survival rate remains unsatisfactory. Furthermore, the comprehensive molecular mechanisms underlying the pathogenesis of T-PLL remain unknown.

The majority of T-PLL cells carry the recurrent chromosome translocations t(14;14)(q11;q32)/inv(14)(q11q32) or t(X;14)(q28;q11), which cause the activation of the genes T-cell leukemia/lymphoma 1A (TCL1A) or mature T-cell proliferation 1 (6). These genes and their associated pathways are likely to be involved in the progression of T-PLL. Integrated genomic sequencing has proven the importance of mutated DNA or genes in T-PLL (7). Bergmann et al (8) indicated that Janus kinase 3 (JAK3) inhibitors may be an option to treat patients with T-PLL with recurrent activating JAK3 mutations. Genes including TCL1A (9) and SWI/SNF-related matrix-associated actin dependent regulator of chromatin B1 (10) have been demonstrated to be associated with the disease progress of T-PLL. Furthermore, apoptosis has been induced in T-PLL by certain drugs, including bortezomib (11), and by the induction of certain proteins, including p53 (12), which indicates further the potential association between pathways associated with apoptosis and T-PLL. Specific genes and chromosomal loci are likely to be linked with disease progression in T-PLL (13), and identifying the significance of altered genes and pathways is vital to increasing the understanding of T-PLL. However, these genes and pathways have yet to be identified.

To investigate the molecular basis of T-PLL in the present study, a bioinformatics analysis of gene expression profile data (GSE5788) was performed. The differentially expressed genes (DEGs) in T-PLL were identified by comparing the microarray data from 6 T-cell T-PLL blood cell samples with those of 8 cluster of differentiation 3 (CD3)+ T-cell samples from healthy donors. Gene ontology (GO; http://www.geneontology.org/) function and Kyoto Encyclopedia of Genes and Genomes (KEGG; http://www.genome.jp/kegg/) pathway enrichment analyses were performed, followed by protein-protein interaction (PPI) network and sub-PPI network analyses. The aim was to elucidate the molecular mechanisms of T-PLL, which may aid in the selection of appropriate treatment strategies and the development of novel treatments for T-PLL.

Materials and methods

Samples

The expression profile dataset GSE5788 (13), which was created using the microarray platform Affymetrix Human Genome U133 Plus 2.0 Array (Santa Clara, CA, USA), was downloaded from the Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/). In order to compare the difference between the T-PLL cells and the normal T-cells, a total of 6 T-PLL blood cell samples including GSM135264 (the experimental group), and 8 CD3+ normal donor T-cell samples purified by immunomagnetic separation, including GSM135270 (the control group), were included in GSE5788. The preprocessing of the microarray data, including calculation of the robust multi-array average (14), was performed by using the Affy options in Bioconductor software (15) and the Affy microarray probe annotation file of Brain Array lab (16).

Screening DEGs

An empirical Bayes method based on the Limma package (17) in R software (https://journal.r-project.org/) was used to identify DEGs among the groups. A false discovery rate (FDR) <0.05 and log. of fold-change >1 were selected as the criteria for the identification of DEGs.

GO and pathway enrichment analysis

GO (18) functional enrichment analysis, including associated cellular component, molecular function (6) and biological process categories, was performed to identify functional enrichment of DEGs. KEGG pathway enrichment (19) was performed to predict the pathways that the previously identified DEGs were associated with. The Database for Annotation, Visualization and Integrated Discovery (DAVID 6.7 Jan. 2010) (20) was used to identify GO categories and significant KEGG pathways with the FDR set as <0.01.

Annotation of gene function

The Transcription Factor (TRANSFAC; http://www.gene-regulation.com/pub/databases.html) database provides information on eukaryotic transcription factors, binding sites, consensus binding sequences and regulated genes. All DEGs were screened based on the TRANSFAC database to identify whether they had a function in transcriptional regulation. Cancer gene databases, including TSGene (a database of tumor suppressor genes) (21) and a tumor-associated gene (TAG) database as described by Chen et al (22), were used to screen for identified cancer-promoting or -inhibiting genes.

PPI network and sub-PPI network construction

The Search Tool for the Retrieval of Interacting Genes (STRING, https://string-db.org/cgi/input.pl) (23) database made associations based on predicted or experimental PPI information. From the STRING database, protein encoding genes that interacted with specific genes were assembled to construct a PPI network. As the non-DEGs of the network may be associated with multiple DEGs, the result of network analysis may be that the role of non-DEGs is greater compared with that of the DEGs. To study the role of DEGs in PPI and to avoid the interference of non DEGs, the non DEGs associated with 1–2 DEGs were enrolled in the network. Interactions were included when they had a combined score >0.95.

The sub-PPI network was additionally investigated using BioNet software (24). The modules were constructed based on the sub-PPI network. A false discovery rate P≤0.005 was selected as the threshold for sub-PPI network construction.

Results

Identification of DEGs

The GSE5788 expression profile data from GEO was investigated to screen for DEGs between the experimental and control groups. A total of 438 DEGs in T-PLL blood sample cells, including 84 upregulated genes and 354 downregulated genes, were revealed (Table I).

Table I.

Differentially expressed genes in the present study (FDR <0.05 and |logFC| >1).

Type	Transcript count	Gene count
Downregulated	1,249	354
Upregulated	305	84
Total	1,554	438

Transcript count refers to the number of differentially expressed transcripts; Gene count refers to the number of DEGs. FDR, false discovery rate; |logFC|, log fold-change; DEG, differentially expressed gene.

GO enrichment analysis

To investigate the function changes in tumor development, GO enrichment analysis of the previously identified DEGs was performed using DAVID. The top 10 up- and downregulated DEGs according to P-value are listed in Table II. The downregulated DEGs were frequently enriched for ontology labels associated with immune function, including immune response (GO:0006955; P=3.21×10−9), and tumor progression, including cell death (GO:0008219; P=2.37×10−7). Upregulated DEGs were frequently enriched in cell proliferation (GO:0008283; P=7.12×10−4) and skin development (GO:0043588; P=9.00×10−4).

Table II.

Top 10 upregulated and downregulated differentially expressed gene ontologies identified by GO functional enrichment analysis.

A, Downregulated genes
GO ID	Description	Gene count	P-value
GO:0006955	Immune response	58	3.21×10−9
GO:0002376	Immune system process	77	1.40×10-8
GO:0008219	Cell death	66	2.37×10−7
GO:0016265	Death	66	2.53×10-7
GO:0046649	Lymphocyte activation	28	5.42×10−7
GO:0042110	T cell activation	23	5.57×10-7
GO:0051249	Regulation of lymphocyte activation	21	9.52×10−7
GO:0030098	Lymphocyte differentiation	18	1.44×10-6
GO:0044267	Cellular protein metabolic process	103	1.49×10−6
B, Upregulated genes
GO ID	Description	Gene count	P-value
GO:0045321	Leukocyte activation	30	1.86×10-6
GO:0008283	Cell proliferation	18	7.12×10−4
GO:0043588	Skin development	7	8.97×10-4
GO:0009913	Epidermal cell differentiation	5	1.02×10−3
GO:0006228	UTP biosynthetic process	2	1.57×10-3
GO:0042455	Ribonucleoside biosynthetic process	4	1.61×10−3
GO:0046051	UTP metabolic process	2	1.85×10-3
GO:0006213	Pyrimidine nucleoside metabolic process	3	2.09×10−3
GO:1901070	Guanosine-containing compound biosynthetic process	2	2.15×10-3
GO:0009163	Nucleoside biosynthetic process	4	2.21×10−3
GO:1901659	Glycosyl compound biosynthetic process	4	2.28×10-3

GO, Gene Ontology; UTP, uridine 5′-triphosphate.

KEGG pathway enrichment analysis

KEGG pathway enrichment analysis using DAVID was performed on the DEGs (Table III). The results revealed that the upregulated DEGs were frequently associated with tumor metastasis pathways, including apoptosis (P=6.20×10−5), immune response pathways, including graft-versus-host disease (P=1.61×10−4), and immune response or antigen reaction pathways, including Chagas disease (American trypanosomiasis; P=5.18×10−3). The downregulated DEGs were enriched in the malaria pathway (P=3.33×10−3).

Table III.

Top 10 downregulated pathways and a unique upregulated pathway significantly enriched by DEGs in T-cell prolymphocytic leukemia.

Regulation	KEGG pathway	Gene count	P-value
Down	Apoptosis	9	6.20×10−5
	Graft-versus-host disease	6	1.61×10−4
	T cell receptor signaling pathway	9	3.30×10−4
	Allograft rejection	5	8.48×10−4
	Type I diabetes mellitus	5	1.70×10−3
	Natural killer cell mediated cytotoxicity	9	1.75×10−3
	Autoimmune thyroid disease	5	3.96×10−3
	Antigen processing and presentation	6	4.39×10−3
	Chagas disease (American trypanosomiasis)	7	5.18×10−3
	Prion diseases	4	5.33×10−3
Up	Malaria	3	3.33×10−3

DEG, differentially expressed gene; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Functional annotation of DEGs

From the T-PLL blood sample microarray data, a total of 13 downregulated [including signal transducer and activator of transcription 3 (STAT3)] and 4 upregulated [including transcription factor 7 like 2 (TCF7L2)] transcription factors, as well as 27 downregulated (including FYN) and 10 upregulated (including TCL1A) TAGs were revealed to be associated with T-PLL (Table IV).

Table IV.

Functional analysis of DEGs.

	Downregulated		Upregulated
Type	Count	DEGs	Count	DEGs
Transcription factors	13	TBX21, STAT4, STAT3, RORA, NR3C2, NR3C1, MAF, IRF1, ID2, GTF2B, GATA3, CDK7, ARNTL	4	TCF7L2, NME2, KLF4, ENO1
Tumor-associated genes	27		10
Oncogenes	4	SET, KRAS, GNA13, FYN	1	TCL1A
Tumor suppressors	16	TGFBR3, TGFBR2, SP100, RARRES3, PPP3CC, MFHAS1, IRF1, ING3, HOPX, HECA, HBP1, FHIT, DOK2, CYLD, CDKN1B, CASP8	4	TCF7L2, RAB25, ESRP1, CDK2AP1
Others	7	STAT3, PRKCB, MAP3K5, MAF, EVI2B, EPS15, ATM	5	KLF4, GSTM1, FES, ENO1, DDR1

DEGs, differentially expressed genes.

Using the Disease Ontology database (25), the downregulated genes, including ARL6IP5, ATM, CCL4, CCL5, CDKN1B, CFLAR, FAS, GNLY, IL2RB, KAT2B, MAP3K5, SET and TRIM22, were revealed to be associated with leukemia and chronic lymphocytic leukemia.

PPI module investigation and pathway regulation analysis

A PPI network of DEGs associated with T-PLL was constructed based on the STRING database (Fig. 1). A total of 10 nodes with the highest degree were selected, including FYN (degree, 90), STAT3 (degree, 76), ATM (degree, 51), KAT2B (degree, 48), IRS1 (insulin receptor substrate-1; degree, 45), PSMD12 (degree, 41), PSMB1 (degree, 40), CDKN1B (degree, 40), CASP8 (degree, 38) and CYC1 (degree, 34).

Figure 1.

A protein-protein interaction network. The red nodes represent upregulated genes; the green nodes represent the downregulated genes; the yellow nodes represent genes in which expression level was unaltered between T-cell prolymphocytic leukemia and normal T-cells.

A sub-PPI network was constructed based on the aforementioned PPI network. BioNet software was used to analyze the sub-PPI network. A total of 61 gene nodes, including STAT3 (the most significantly downregulated gene in the sub-PPI network, degree, 14) and IRS1 (the most significantly upregulated gene in the sub-PPI network, degree, 6), were included in the sub-PPI network (Fig. 2).

Figure 2.

A sub-protein-protein interaction network investigation. The red nodes represent upregulated genes and the green nodes represent downregulated genes in the experimental group (T-cell prolymphocytic leukemia); square nodes represent high importance DEGs and circular nodes represent low importance DEGs in the network. DEGs, differentially expressed genes.

The KEGG pathway analysis was performed based on the DEGs in the sub-PPI network. As presented in Table V, the sub-network of STAT3 is involved in growth signal pathways [including the JAK-STAT signaling pathway (P=3.0×10−8)], cell differentiation pathways [including osteoclast differentiation (P=7.5×10−6)], cancer-associated pathways [including prostate cancer (P=4.7×10−7)], and viral disease-associated pathways [including hepatitis C (P=1.1×10−5)].

Table V.

The top 10 most KEGG-enriched pathways for the network module.

KEGG pathway	Gene count	P-value
Pathways in cancer	17	2.65×10−10
JAK-STAT signaling pathway	11	2.97×10−8
Prostate cancer	8	4.67×10−7
TGF-β signaling pathway	7	4.37×10−6
Cell cycle	8	5.88×10−6
Osteoclast differentiation	8	7.45×10−6
Colorectal cancer	6	9.43×10−6
Hepatitis C	8	1.05×10−5
Pancreatic cancer	6	1.91×10−5
T cell receptor signaling pathway	7	2.32×10−5

KEGG, Kyoto Encyclopedia of Genes and Genomes; TGF, transcription growth factor.

Discussion

T-PLL is a rare, aggressive T-cell leukemia, which has not been well characterized, particularly in terms of its molecular mechanisms. In the present study, the molecular pathogenesis of T-PLL was investigated based on a comprehensive bioinformatics analysis. The results identified 84 upregulated and 354 downregulated genes in T-PLL sample microarrays. These DEGs were associated with various functions including cell death, and various pathways, including apoptosis. A total of 17 dysregulated transcription factors and 37 dysregulated TAG were revealed based on functional analysis of DEGs. A PPI network analysis identified a total of 61 genes. The most significantly downregulated gene, STAT3 (degree, 14), and upregulated gene, IRS1 (degree, 6), may have significant associations with the pathogenesis and progression of T-PLL.

The dysregulation of specific genes, including transcription factors, and associated pathways is commonly associated with increased tumor cell proliferation, based on previous bioinformatics analyses (26–28). These genes and pathways perform important roles and are likely to be significant in the development of cancer (29). STAT3 is activated in various types of cancer, including gliomas and breast cancer (30,31). The STAT3 signaling pathway, including the upstream JAK signal transducer, has been reported to participate in the development of various cancer types (32,33).

Previous studies indicate that JAK2-STAT3 signaling is involved in the production of hepatic thrombopoietin (34) and the growth of hormone refractory prostate cancer cells (35). In the present study, the downregulated STAT3 was the core node of the sub-PPI network, and the DEGs connected to STAT3 were involved in pathways including JAK-STAT signaling. This result confirms that STAT3 and the JAK2-STAT3 pathway are associated with the progression of T-PLL.

Another gene identified to be significant was IRS1, a critical component of insulin signaling, which is also involved in cell proliferation and cancer development (36). IRS1 is associated with the progression of tumors, including lung cancer (37) and colorectal cancer (38). The significant upregulation of IRS1 in the present study indicated the close association of IRS1 with T-PLL, which was in accordance with the function of IRS1 in cancer identified in previous studies. The expression levels of various genes, including STAT3 and IRS1, were significantly altered in the tumor compared with the controls, implying they may be used as novel biomarkers for establishing a prognosis in T-PLL.

Novel drugs targeting specific pathways can be developed based on an understanding of the pathogenesis of T-PLL (5). KEGG pathway analysis in the present study revealed that apoptosis and T-cell receptor signaling were included among the enriched pathways identified. The majority of these outstanding pathways were enriched among the downregulated genes, indicating that the downregulation of genes in these pathways may act to inhibit T-cell activation, promoting disease progression. However, additional investigations are required to improve the understanding of the complex interaction of these dysregulated genes and associated pathways.

In conclusion, the mechanism of T-PLL was observed to be complicated. Various cell functions, including cell death, and pathways, including apoptosis, may be involved in the process. Identified candidate genes, including STAT3 and IRS1, may be targets for the additional study of T-PLL.