Bioinformatics analysis of the role of doxycycline in prostate cancer with microarray

Running title: Roles of doxycycline in prostate cancer

High lights:

  1. Significant pathways involved in the progression are analyzed.
  2. PPI network was constructed and total 392 common DEGs were identified.
  3. BUB1MYCIGF-1CCNE2, CYP2E1 and ALDH3B2 were identified as key genes.


Purpose: We aim to identify the key genes and pathways in prostate cancer (PC)with treatment of doxycycline for different time.

Methods: The microarray expression profile of GSE24261 was obtained from gene expression omnibus (GEO) database, including doxycycline treated samples and control samples at day 1 and day 2. Differentially expressed genes (DEGs) in doxycycline samples at different time point were identified using the limma package in R language. Gene ontology (GO) Term and kyoto encyclopedia of genes and genomes (KEGG) pathway analysis of DEGs were performed using the DAVID and protein-protein interaction (PPI) network was constructed using the Cytoscape. Moreover, the common DEGs in day1 and day2 doxycycline case samples were selected and analyzed.

Get Help With Your Essay

If you need assistance with writing your essay, our professional essay writing service is here to help!

Essay Writing Service

Results: Totally, 460 DEGs in day1 samples and 1336 DEGs in day2 samples were screened. The up-regulated DEGs were mainly enriched in metabolism of xenobiotics by cytochrome P450 and drug metabolism-cytochrome P450, while the down-regulated genes were mainly enriched in cell cycle and DNA replication. The common DEGs, aldehyde dehydrogenase 3 family, member B2 (ALDH3B2), cytochrome P450, family 2, subfamily E, polypeptide 1 (CYP2E1), v-mycavian myelocytomatosis viral oncogene homolog (MYC) insulin-like growth factor 1 (IGF1), cyclin E2 (CCNE2), were identified as key genes in response to doxycycline, which were involved in some significant pathways including cell cycle.

Conclusion: Doxycycline inhibits the expression of BUB1MYCIGF-1 and CCNE2 and increase the expression of CYP2E1 and ALDH3B2, leading to the inhibition of tumoregenesis.

Keywords: prostate cancer; doxycycline; pathway analysis; differentially expressed genes.


Prostate cancer (PC) is a cancer develops in prostate in the male reproductive system, which is predicted to account for approximately 240000 new cases in America and is the leading cause of cancer mortality [1]. The morbidity of invasive PC increases depending on patients` ages, with the risk that rises from 1 in 37 males between 40 to 59 of age to 1 in 8 males of 70 or older [2]. Radical prostatectomy combined with chemotherapeutic is the main therapy for PC, yet at least a third of the patients will have a recurrence and the cancer cells may spread from the prostate to other parts of human body, such as bones and lymph nodes [4]. However, it is little known on the efficient inhibition therapy of prostate, thus the research on the roles of some valid drugs, such as doxycycline, is of vital significance.

Doxycycline is a member of tetracycline family which has been used as antibiotics effectively for decades [5]. Doxycycline blocks shedding of MHC class I polypeptide-related sequences from a panel of human tumor cells and acted to increase their expression and cell surface translocation [6]. Besides, doxycycline also selectively enhanced the replication of oncolytic vaccinia in various tumor cell lines, which leads to increased sensitivity to these therapies [6]. In recent years, doxycycline has been used for treatment of various human cancers, such as breast cancer, in which doxycycline is highly effective in the inhibition of matrix metalloproteinases, the important enzymes in tumor cell invasion and metastatic ability [7]. Besides, the doxycycline has also been used for PC treatment [7]. However, the definite mechanism of doxycycline in PC cell has not been fully clarified.

Sanjeev K et al [8] take treatment of doxycycline on PC cells to analyze the mRNA expression in response to RunX2, an osteoblast master transcription factor and Gillian H et al [9] has identified RunX2-regulated genes in PC cells with treatment of doxycycline. Both two researched above make use of microarray of GSE24261, which was used in the present study. In the present study, we identified differentially expressed genes (DEGs) in PC cells using the microarray GSE24261, which was obtained based on PC cell samples treated with or without doxycycline at different time points.. Besides, pathways and protein-protein interaction (PPI) network were analyzed and common DEGs were identified to clarify the role of doxycycline in PC.

Materials and methods

Microarray data

The gene expression profile of GSE24261 was downloaded from Gene Expression Omnibus (GEO) database in National Center for Biotechnology Information (NCBI) ( based on the platform of GPL6883 (Illumina HumanRef-8 v3.0 expression beadchip). The study contains a total of 16 samples, including 4 day1 doxycycline case samples, 4 day1 control samples, 4 day2 doxycycline case samples and 4 day2 control samples.

Data preprocessing and DEGs analysis

By using Limma package [10] in Bioconductor and Illumina microarray probe annotation profile from Brain Array Lab, the probe-level data was converted into expression measures, during which Background Correction, Quantile Normalization and Probe Summarization were performed and the significantly DEGs in doxycycline case samples of PC cells were identified. A combination of FDR < 0.01 and |log2FC (fold change)| > 2 was used as the threshold.

Gene ontology (GO) and pathway enrichment analysis

GO analysis has become a widely used approach for the studies of large-scale genomic or transcriptomic data in function [12]. Kyoto encyclopedia of genes and genomes (KEGG) is a widely used collection of online database which deals with genomes, enzymatic pathways, and biological chemicals. [13] In this study, the functions and pathways of the screened DEGs in both day1 and day2 doxycycline case samples were analyzed using the DAVID from the GO and KEGG pathway database with the p-value < 0.05, respectively.

Functional annotation of DEGs

Functional annotation of DEGs was performed for the detection of oncogenes and tumor suppressor genes. Two databases, tumor suppressor genes (TSGs) database [14] and tumor associated genes (TAG) database [15] were used for screening tumor suppressor genes and oncogenes.

PPI network and sub-network construction


The Search Tool for the Retrieval of Interacting Genes (STRING) database [16] is a database of known and predicted protein interactions, including both experimental and predicted interaction information. Cytoscape [17] is an open source bioinformatics software platform, which is used for the visualization of molecular interaction networks and integrating with gene expression profiles and other state data. In this study, STRING was used to predict the interactions of selected DEGs [18] with Required Confidence (combined score) > 0.9 and Cytoscape was used to visualize the PPI network. [19]. Besides, sub-network was also constructed using ClusterOne [20] in Cytoscape and pathways were enriched on sub-network with p-value < 0.05.

Selection of common DEGs in samples at two time points

To analyze the effects of doxycycline works for different time, the common DEGs in both day1 and in day2 samples were identified. Afterwards, the functions and pathways of the screened common DEGs were analyzed using the DAVID from the KEGG pathway database with the p-value < 0.05.


DEGs selection

Totally, 460 DEGs (324 up-regulated and 136 down-regulated) in day 1 samples, and 1336 DEGs (707 up-regulated and 629 down-regulated) in day 2 samples were screened. The heat map of both day1 and day2 samples were showed in Figure 1A and Figure 1B, respectively.

GO and KEGG enrichment analysis of DEGs

GO and KEGG enrichment were performed with p-value < 0.01 for the functional and pathway analysis of DEGs, respectively. The main KEGG pathways of DEGs screened in day1 doxycycline case samples were listed in Table 1. The up-regulated genes were involved in metabolism of xenobiotics by cytochrome P450 and drug metabolism-cytochrome P450 (Table 1A), while the down-regulated genes participated in the cell cycle, TGF-beta signaling pathway and PC (Table 1B). In addition, the GO terms of the DEGs were also showed in Table 1. We can see from the results that the up-regulated genes were mainly enriched in the signal transduction and signaling (Table 1C), while the down-regulate genes were involved in negative regulation of binding and positive regulation of cellular component organization (Table 1D).

Find Out How Can Help You!

Our academic experts are ready and waiting to assist with any writing project you may have. From simple essay plans, through to full dissertations, you can guarantee we have a service perfectly matched to your needs.

View our services

The main KEGG pathways of DEGs identified in day2 doxycycline case samples were listed in Table 2. The up-regulated genes were involved in metabolism of xenobiotics by cytochrome P450 and drug metabolism-cytochrome P450 (Table 2A), while the down-regulated genes participated in the DNA replication and cell cycle (Table 2B). In addition, the GO terms of the up-regulated genes were mainly enriched in the metabolic process (Table 2C), while the down-regulate genes were involved in cell cycle (Table 2D).

Functional annotation of DEGs

Tumor suppressor genes (TSG) and oncogenes were screened. According to the results (Table 3), in the day1 doxycycline case samples, 4 up-regulated oncogenes and 2 down regulated oncogenes, and 27 up-regulated TSGs and 4 down-regulated TSGs were screened. In day2 doxycycline case samples, totally 15 oncogenes and 44 TSGs in up-regulated genes, and 17 oncogenes and 26 TSGs were screened, which is more than that in day1 doxycycline case samples.

PPI network construction

The network constructed of all DEGs in both day1 doxycycline case samples and day2 doxycycline case samples was shown in Figure 2. In the PPI network of day1 doxycycline case sample (Figure 2A), 102 nodes and 152 protein pairs were obtained. In these nodes, vascular endothelial growth factor A (VEGFA), cytochrome P450, family 2, subfamily E, polypeptide 1 (CYP2E1), glutathione S-transferase mu 1 (GSTM1), glutathione S-transferase alpha 5 (GSTA5), epoxide hydrolase 1, microsomal (EPHX1), glutathione S-transferase alpha 1 (GSTA1), cell division cycle 6 (CDC6), glutathione S-transferase mu 2 (GSTM2) and glutathione S-transferase alpha 4 (GSTA4) were identified as key genes.

In the PPI network of day2 doxycycline case samples, 548 nodes and 2501 protein pairs were obtained. Two sub-network modules, module A and module B, were constructed using ClusterOne and were shown in Figure 2B. BUB1 mitotic checkpoint serine/threonine kinase (BUB1), polo-like kinase 1 (PLK1), cyclin A2 (CCNA2), cell division cycle 20 (CDC20), baculoviral IAP repeat containing 5 (BIRC5) and cyclin B1 (CCNB1) were identified key nodes in the network. For further analysis of the modules, KEGG and GO term were performed on module A and module B, respectively (Table 4). According to the results, the main pathways those nodes enriched in were pathways related to cell cycle and mitotic, such as cell cycle, oocyte meiosis and p53 signaling pathway in module A, and DNA replication, mismatch repair and nucleotide excision repair in module B

Analysis on common DEGs

Totally, 392common DEGs were obtained, including 283 up-regulated genes and 109 down-regulated genes, and the heat map of these genes was shown in Figure 1C According to the pathways analysis (Table 5), up-regulated genes, such as aldehyde dehydrogenase 3 family, member B2 (ALDH3B2), cytochrome P450, family 2, subfamily E, polypeptide 1 (CYP2E1) and glutathione S-transferase alpha 1 (GSTA1), were mainly enriched in metabolism of xenobiotics by cytochrome P450 and Leukocyte transendothelial migration, while down-regulated genes, such as cyclin E2 (CCNE2), v-mycavian myelocytomatosis viral oncogene homolog (MYC) and insulin-like growth factor 1 (IGF-1), were mainly enriched in cell cycle, prostate cancer and p53 signaling pathway.


PC is the most common cancer found in men in America, which has a high rate of reccurence in spite of primary therapy [21]. As a widely used antibiotic in inflammation, doxycycline is also used in treatment of PC, but quite little is known about the mechanism of doxycycline in PC. In this study, PC cell samples with or without treatment of doxycycline at different time point have been used for the identification of key genes in PC in response to doxycycline. Our study showed that BUB1, IGF-1, MYC, ALDH3B2, CYP2E1, and CCNE2 were identified as key genes, which were involved in metabolism of xenobiotics by cytochrome P450, cell cycle, prostate cancer and p53 signaling pathway.

In the PPI network of day2 doxycycline case sample, we found that BUB1 was a down-regulated key gene enriched in cell cycle pathway. BUB1 is a member of BUB family, which encode proteins that are part of a large multi-protein kinetochore complex and believed to be key components of the checkpoint regulatory pathway [22]. Known as a serine/threonine-protein kinase, BUB1 plays a central role in mitosis and functions in part by phosphorylating members of the mitotic checkpoint complex and activating the spindle checkpoint [23]. This kinase accumulates at unattached kinetochores where it mediates the recruitment of mitotic arrest deficient (Mad) dimers [24] and prevents the premature separation of sister chromatids until all chromosomes are properly attached to kinetochores to make sure chromosome segregation is correct [25]. Its defects may cause the instability of chromosome or aneuploidy in human cancer cell lines and underexpression of BUB1 in colon cancer [26] and wilms tumor [27] has been reported. As aneuploidy is an important prognostic factor in PC, BUB1 plays a major role in PC tumorigenesis. Our data showed that with treatment of doxycycline for 2 days, BUB1 expressed down-regulated and was enriched in cell cycle, indicating that doxycycline inhibit the tumorigenesis and tumor growth by controlling the expression of genes associated with mitotic checkpoint, such as BUB1.

In this study, with treatment of doxycycline, the expression of MYC was down-regulated, which was enriched in cell cycle, TGF-beta signaling pathway, small cell lung cancer and bladder cancer. MYC is a proto-oncogene implicated in PC development, which is present on human chromosome 8q24 [28]. The MYC onco-protein is a transcription factor that regulates cellular processes including cell proliferation, metabolism, protein synthesis, mitochondrial function and stem cell renewal [29]. Associated with a partner protein MAX, MYC is able to alternatively associate with other proteins and bind DNA to influence transcription [30]. In PC cells, a region encompassing the MYC locus is somatically amplified at low levels in a subset of patients, which correlates with both worse prognosis and high historical grade. Based on our study and published reportd, overexpression of MYC associated with somatic genetic alterations including translocations and gene amplification lead to the tumorigenesis and growth of PC. In our findings, MYC was down-regulated, leading to the inhibition of main pathways related to cell cycle and TGF-beta signaling pathway, which is important for cytokines secretion and cellular proliferation and migration, and therefore, the growth of PC cells are decreased.

In the common genes, IGF-1 was expressed down-regulated with treatment of doxycycline and mainly enriched in PC, oocyte meiosis and pathway in cancer. IGF-1(encoded byIGF-1) is a member of insulin-like growth factor family and plays a role in prostate development and carcinogenesis [31]. IGF-1 stimulates mitogenic and antiapoptotic activities of prostate epithelial cells in both prostate development and tumorigenesis [31]. General view is that IGF-1 binds to its binding protein IGFBP-1 or IGFBP-3, which modulate the bioavailability of IGF-1 and regulated by insulin, promoting the proliferation of epithelial cells [32]. Doxycycline treatment causes the expression of IGF-1 down-regulated, which decreases the stimulation of cell proliferation and resistance to apoptosis and consequently inhibits the occurrence of PC.

In the present study, CCNE2, encoded by the down-regulated gene CCNE2, was enriched in significant pathways including cell cycle, prostate cancer, oocyte meiosis and p53 signaling pathway. CCNE family members are important regulators of S phage entry concurrent with DNA replication in cell cycle and they are frequently deregulated in some human cancers [33]. They drive the transition from G1 to S phase in mitotisis cycle through assembly of pre-replication complexes and activation of CDK kinases, which leads to the initiation of DNA synthesis [34]. Because of the fundamental role in promoting proliferation, CCNEs have been identified as oncogenes in various cancers [35]. As for CCNE2, it functions as a regulator of CDK2 and specifically interacts with CDK2 inhibitors [36], playing a key role in cell cycle G1/S transition. Our study may drawn a conclusion that doxycycline down-regulates the exhibition of CCNE2, leading to the restrain of G1/S transition in cell cycle, thereby the growth of tumor is inhibited.

Cytochrome P450 enzymes are a family of heme-containing enzymes involved in xenobiotics metabolism [37] and specific isoenzymes of the family have been identified in tumors [38]. CYP2E1 is a key enzyme in the metabolic activation of procarcinogens [39], which is up-regulated with doxycycline treatment in this study and enriched in drug metabolism-cytochrome P450 and metabolism of xenobiotics by cytochrome P450. CYP2E1 metabolizes some small molecules such as acetaminophen, ethanol and pro-carcinogenes like azo compounds and nitrosamines [40], during which toxic intermediates and excessive amounts of reactive oxygen species (ROS) are generated [41]. High ROS levels induces autophagy pathway and triggers accumulation of autophagy-regulated genes [42], thereby stimulating the formation of autophagosome in cancer [43]. In the present study, the express level of CYP2E1 was high, which increases ROS generation and inhibits the migration of the highly invasive cancer cells [38].

In the doxycycline case samples, another key gene enriched in drug metabolism-cytochrome P450 and metabolism of xenobiotics by cytochrome P450, ALDH3B2, was also expressed up-regulated. Aldehyde dehydrogenase (ALDH) is a family of cytosolic isoenzymes that are responsible for oxidizing intracellular aldehydes and contributing to the oxidation of retinol to retinoic acid in early stem cell differentiation [33]. Some member, such as ALDH1 enzymes, has been identified as being responsible for the resistance of progenitor cells to chemotherapeutic agents [35]. ALDH3 is induced by chlorinated compounds and polycyclic aromatic hydrocarbons and in hepatoma and lung cancer cell lines it is directly correlated with the degree of deciation [34]. ALDH3B2 encodes a member of ALDH family, which is a isozymes play a role in the detoxification of aldehydes generated by alcohol metbollism and peroxidation of lipid. The role of ALDH3B2 is similar with CYP2E1 in response to doxycycline and this gene is a target of the drug, whose up-regulated expression decreases the tumorigenesis.

In conclusion, our study identified key genes in prostate cancer cells with treatment of doxycycline at different time point, including BUB1, MYC, IGF-1, ALDH3B2, CYP2E1 and CCNE2. Doxycycline inhibits the expression of BUB1, MYC, IGF-1 and CCNE2 while increases the expression of CYP2E1 and ALDH3B2, leading to the inhibition of tumoregenesis. However, the sample size is relatively less in this study and no experiments have been performed to confirm our conclusion, which is a limitation. Therefore, more analysis and experiments should be performed for further research.