Computational approaches in cancer multidrug resistance research: Identification of potential biomarkers, drug targets and drug-target interactions.

Like physics in the 19th century, biology and molecular biology in particular, has been fertilized and enhanced like few other scientific fields, by the incorporation of mathematical methods. In the last decades, a whole new scientific field, bioinformatics, has developed with an output of over 30,000 papers a year (Pubmed search using the keyword "bioinformatics"). Huge databases of mass throughput data have been established, with ArrayExpress alone containing more than 2.7 million assays (October 2019). Computational methods have become indispensable tools in molecular biology, particularly in one of the most challenging areas of cancer research, multidrug resistance (MDR). However, confronted with a plethora of different algorithms, approaches, and methods, the average researcher faces key questions: Which methods do exist? Which methods can be used to tackle the aims of a given study? Or, more generally, how do I use computational biology/bioinformatics to bolster my research? The current review is aimed at providing guidance to existing methods with relevance to MDR research. In particular, we provide an overview on: a) the identification of potential biomarkers using expression data; b) the prediction of treatment response by machine learning methods; c) the employment of network approaches to identify gene/protein regulatory networks and potential key players; d) the identification of drug-target interactions; e) the use of bipartite networks to identify multidrug targets; f) the identification of cellular subpopulations with the MDR phenotype; and, finally, g) the use of molecular modeling methods to guide and enhance drug discovery. This review shall serve as a guide through some of the basic concepts useful in MDR research. It shall give the reader some ideas about the possibilities in MDR research by using computational tools, and, finally, it shall provide a short overview of relevant literature.

High number of kinome-mutations in non-small cell lung cancer is associated with reduced immune response and poor relapse-free survival.

Lung cancer is the leading cause of cancer related death, and the past years' improved insight into underlying molecular events has significantly improved outcome for specific subsets of patients. In particular, several new therapies that target protein kinases have been implemented, and many more are becoming available. We have investigated lung cancer specimens for somatic mutations in a targeted panel of 612 human genes, the majority being protein kinases. The somatic mutation profiles were correlated to profiles of immune cell infiltration as well as relapse-free survival. Targeted deep sequencing was performed on 117 tumour/normal pairs using the SureSelect Human Kinome kit (Agilent Technologies), with capture probes targeting 3.2 Mb of the human genome, including exons and untranslated regions of all known kinases, kinase receptors and selected cancer-related genes (612 genes in total). CD8 staining was determined using Ventana Benchmark. Survival analyses were performed using SPSS. The number of mutations per sample ranged from 0 to 50 (within the 612 genes tested), with a median of nine. The prognosis was worse for patients with more than the median number of mutations. A significant correlation was found between mutations in one of selected DNA-repair genes and the total number of mutations in that tumour (p < 0.001). There was a significant inverse correlation between the number of infiltrating stromal CD8+ lymphocytes and the presence of EGFR mutations.

A literature network of human genes for high-throughput analysis of gene expression.

We have carried out automated extraction of explicit and implicit biomedical knowledge from publicly available gene and text databases to create a gene-to-gene co-citation network for 13,712 named human genes by automated analysis of titles and abstracts in over 10 million MEDLINE records. The associations between genes have been annotated by linking genes to terms from the medical subject heading (MeSH) index and terms from the gene ontology (GO) database. The extracted database and accompanying web tools for gene-expression analysis have collectively been named 'PubGene'. We validated the extracted networks by three large-scale experiments showing that co-occurrence reflects biologically meaningful relationships, thus providing an approach to extract and structure known biology. We validated the applicability of the tools by analyzing two publicly available microarray data sets.

Increased expression of IRF4 and ETS1 in CD4+ cells from patients with intermittent allergic rhinitis.

BACKGROUND: The transcription factor (TF) IRF4 is involved in the regulation of Th1, Th2, Th9, and Th17 cells, and animal studies have indicated an important role in allergy. However, IRF4 and its target genes have not been examined in human allergy. METHODS: IRF4 and its target genes were examined in allergen-challenged CD4(+) cells from patients with IAR, using combined gene expression microarrays and chromatin immunoprecipitation chips (ChIP-chips), computational target prediction, and RNAi knockdowns. RESULTS: IRF4 increased in allergen-challenged CD4(+) cells from patients with IAR, and functional studies supported its role in Th2 cell activation. IRF4 ChIP-chip showed that IRF4 regulated a large number of genes relevant to Th cell differentiation. However, neither Th1 nor Th2 cytokines were the direct targets of IRF4. To examine whether IRF4 induced Th2 cytokines via one or more downstream TFs, we combined gene expression microarrays, ChIP-chips, and computational target prediction and found a putative intermediary TF, namely ETS1 in allergen-challenged CD4(+) cells from allergic patients. ETS1 increased significantly in allergen-challenged CD4(+) cells from patients compared to controls. Gene expression microarrays before and after ETS1 RNAi knockdown showed that ETS1 induced Th2 cytokines as well as disease-related pathways. CONCLUSIONS: Increased expression of IRF4 in allergen-challenged CD4(+) cells from patients with intermittent allergic rhinitis leads to activation of a complex transcriptional program, including Th2 cytokines.

CLC and IFNAR1 are differentially expressed and a global immunity score is distinct between early- and late-onset colorectal cancer.

Colorectal cancer (CRC) incidence increases with age, and early onset of the disease is an indication of genetic predisposition, estimated to cause up to 30% of all cases. To identify genes associated with early-onset CRC, we investigated gene expression levels within a series of young patients with CRCs who are not known to carry any hereditary syndromes (n=24; mean 43 years at diagnosis), and compared this with a series of CRCs from patients diagnosed at an older age (n=17; mean 79 years). Two individual genes were found to be differentially expressed between the two groups, with statistical significance; CLC was higher and IFNAR1 was less expressed in early-onset CRCs. Furthermore, genes located at chromosome band 19q13 were found to be enriched significantly among the genes with higher expression in the early-onset samples, including CLC. An elevated immune content within the early-onset group was observed from the differentially expressed genes. By application of outlier statistics, H3F3A was identified as a top candidate gene for a subset of the early-onset CRCs. In conclusion, CLC and IFNAR1 were identified to be overall differentially expressed between early- and late-onset CRC, and are important in the development of early-onset CRC.

Patterns of genomic evolution in advanced melanoma.

Genomic alterations occurring during melanoma progression and the resulting genomic heterogeneity between metastatic deposits remain incompletely understood. Analyzing 86 metastatic melanoma deposits from 53 patients with whole-exome sequencing (WES), we show a low branch to trunk mutation ratio and little intermetastatic heterogeneity, with driver mutations almost completely shared between lesions. Branch mutations consistent with UV damage indicate that metastases may arise from different subclones in the primary tumor. Selective gain of mutated BRAF alleles occurs as an early event, contrasting whole-genome duplication (WGD) occurring as a late truncal event in about 40% of cases. One patient revealed elevated mutational diversity, probably related to previous chemotherapy and DNA repair defects. In another patient having received radiotherapy toward a lymph node metastasis, we detected a radiotherapy-related mutational signature in two subsequent distant relapses, consistent with secondary metastatic seeding. Our findings add to the understanding of genomic evolution in metastatic melanomas.

Photochemically induced gene silencing using small interfering RNA molecules in combination with lipid carriers.

Novel strategies for efficient delivery of small interfering RNA (siRNA) molecules with a potential for targeting are required for development of RNA interference (RNAi) therapeutics. Here, we present a strategy that is based on delivery of siRNA molecules through the endocytic pathway, in order to develop a method for site-specific gene silencing. To achieve this, we combined the use of cationic lipids and photochemical internalization (PCI). Using the human S100A4 gene as a model system, we obtained potent gene silencing in four tested human cancer cell lines following PCI induction when using the cationic lipid jetSI-ENDO. Gene silencing was shown at both the RNA and protein levels, with no observed PCI toxicity when using the jetSI reagent and an optimized PCI protocol. This novel induction method opens for in vivo site-specific delivery of siRNA molecules toward a sequence of interest.

Connectivity can be used to identify key genes in DNA microarray data: a study based on gene expression in nasal polyps before and after treatment with glucocorticoids.

CONCLUSIONS: The presented analysis of nasal polyposis using connectivity based on the PubGene literature co-citation network demonstrates that this tool can be used to identify key genes in DNA microarray studies of human polygenic diseases. OBJECTIVES: DNA microarray studies of complex diseases may reveal differential expression of hundreds of genes. According to network theory and studies of yeast cells, genes that are connected with several other genes appear to have key regulatory roles. This study aimed to examine if this principle can be translated to DNA microarray studies of human disease, using nasal polyposis as a base for the analysis. MATERIALS AND METHODS: The connectivity of differentially expressed genes from a previously described microarray study of nasal polyposis before and after treatment with glucocorticoids was determined. This was done using the literature co-citation network PubGene. RESULTS: In all, 166 genes were differentially expressed; 39 of these were previously defined as inflammatory and considered important for nasal polyposis. The connectivity of all differentially expressed genes was analysed using the PubGene literature co-citation network. Seventy-four of the 166 genes were connected to other genes. By contrast, the average number of connected genes among 100 sets of 166 randomly chosen genes was 31.5. A small number of the differentially expressed genes were highly connected, while most genes had few or no connections. This indicated a scale-free network. The most connected gene was interleukin-8, an inflammatory gene of known importance for nasal polyposis. Twenty-eight of the 74 connected genes were inflammatory (38%), compared with 11 of the 92 unconnected genes (12%), p < 0.0001. Since most evidence suggests that nasal polyps are inflammatory in their nature, this supports the hypothesis that connected genes have more disease relevance than unconnected genes.

Photochemically induced gene silencing using PNA-peptide conjugates.

The potential for exploration of peptide nucleic acid (PNA) as an experimental and therapeutic regulator of gene expression has been hampered by a poor delivery and a lack of site-specific targeting. In the present study, we have developed an efficient strategy for nuclear delivery of PNA by combining cationically charged PNA-peptide conjugates and photochemical internalization (PCI) technology. When using the S100A4 gene as a model system, a consistent downregulation to around 10% remaining protein signal was obtained in three selected cell lines. Furthermore, a dose-dependent and time-dependent inhibition of the S100A4 protein was demonstrated. A main benefit of the strategy proposed is the possibility of site-specific targeting.

Screening for germ line TP53 mutations in breast cancer patients.

The constant denaturant gel electrophoresis technique was used to screen for TP53 germ line mutations in 237 women with breast carcinoma (167 unselected patients, 30 patients with at least one first-degree relative with breast cancer, and 40 women diagnosed with breast cancer before age 35). A germ line mutation at codon 181 was noted in one of the unselected patients and a codon 245 mutation in one of the early-onset patients. Both had a family history of breast cancer and other malignancies suggestive of Li-Fraumeni syndrome. The codon 245 mutation was also present in this patient's affected mother.

Expression of a novel factor in human breast cancer cells with metastatic potential.

Clinical and experimental evidence suggests that tumor cells shed into the circulation from solid cancers are ineffective in forming distant metastasis unless the cells are able to respond to growth conditions offered by the secondary organs. To identify the phenotypic properties that are specific for such growth response, we injected carcinoma cells, which had been recovered from bone marrow micrometastases in a breast cancer patient who was clinically devoid of overt metastatic disease and established in culture, into the systemic circulation of immunodeficient rats. The animals developed metastases in the central nervous system, and metastatic tumor cells were isolated with immunomagnetic beads coated with an antibody that was reactive with human cells. The segregated cell population was compared with the injected cells by means of differential display analysis, and two candidate fragments were identified as up-regulated in the fully metastatic cells. The first was an intracellular effector molecule involved in tyrosine kinase signaling, known to mediate nerve growth factor-dependent promotion of cell survival. The second was a novel gene product (termed candidate of metastasis-1), presumably encoding a DNA-binding protein of helix-turn-helix type. Constitutive expression of candidate of metastasis-1 seemed to distinguish breast cancer cells with metastatic potential from cells without metastatic potential. Hence, our experimental approach identified factors that may mediate the growth response of tumor cells upon establishment in a secondary organ and, thereby, contribute to the metastatic phenotype.

S100A4 involvement in metastasis: deregulation of matrix metalloproteinases and tissue inhibitors of matrix metalloproteinases in osteosarcoma cells transfected with an anti-S100A4 ribozyme.

The biological function of the metastasis-associated gene S100A4 is not fully understood, although there is evidence indicating interactions between the gene product and the cytoskeleton. We have examined whether an association could exist between S100A4 and the regulation of matrix metalloproteinases (MMPs) and their endogenous inhibitors (TIMPs). For these studies, three clones of a highly metastatic human osteosarcoma cell line (OHS) transfected with a hammerhead ribozyme directed against the S100A4 gene transcript were used. The clones demonstrated different expression levels of S100A4 and also different metastatic capacity. In the clone with the most prominent down-regulation of S100A4, the mRNA levels of MMP2, membrane type (MT) 1-MMP, and TIMP-1 were significantly reduced in exponentially growing cultures. Western blots, gelatin zymography, and ELISA showed similar expression patterns of MMPs and TIMPs at the protein level. In the clones with an intermediate expression of S100A4, reduced expression of MT1-MMP and TIMP-1 was detected, whereas the expression of MMP-2 was at the same level as in the control cells. In contrast to the other factors, TIMP-2 was up-regulated in all of the clones independent of the extent of ribozyme-induced down-regulation of S100A4. The transwell chamber assay demonstrated that the capacity of the ribozyme-transfected cells to cross uncoated filters was reduced, relative to control cells, according to the reduction in the S100A4 expression level. The clone with the lowest reduction in S100A4 did not demonstrate different motility compared with control cells, whereas transfectants with only 5% S100A4 mRNA showed a 50% reduction in motility. Interestingly, this trend was even more striking when the capacity to cross Matrigel-coated filters was analyzed, as all the clones demonstrated between 40 and 75% reduced invasion. It is concluded that S100A4 may exert its effect on metastasis formation not only by stimulating the motility of tumor cells but also by affecting their invasive properties through influencing the expression of MMPs and their endogenous inhibitors.

p53 abnormalities in different subtypes of human sarcomas.

In this report we examined p53 alterations at the DNA, mRNA, and protein levels on tissue from 39 patients with different subtypes of sarcoma. Loss of heterozygosity for the chromosome 17p region was found in 60, 63, and 33% of 10 informative osteosarcomas, 11 malignant fibrous histiocytomas, and 6 leiomyosarcomas, respectively. In addition, 2 of 10 tumors belonging to a heterogeneous group of soft tissue sarcomas showed loss of heterozygosity. Elevated levels of p53 mRNA were found in six tumors, four had a truncated transcript, and in six patients no mRNA was detected. In most cases, elevated transcript levels were accompanied by overexpression of protein as studied by immunohistochemistry, whereas the presence of truncated transcripts was associated with negative immunostaining. Point mutations in exons 5, 7, or 8 of the TP53 gene were detected in seven tumors. Six of these expressed high levels of mRNA and protein, probably reflecting a point mutation in one of the alleles and loss of the other. Three of the mutations have not previously been described. Taken together, p53 abnormalities were found in approximately 65% of the osteosarcomas, malignant fibrous histiocytomas, and leiomyosarcomas examined and in 30% of the other soft tissue tumors. The results indicate that the TP53 gene is involved in the tumorigenesis of several sarcoma subtypes in a higher fraction of cases than was previously recognized.

Reversal of the in vivo metastatic phenotype of human tumor cells by an anti-CAPL (mts1) ribozyme.

The putative role of the CAPL gene in enhancing the development of human cancer metastasis was examined by transfecting human high-expressing osteosarcoma cells with a hammerhead ribozyme directed against the gene transcript. The ability of the ribozyme to cleave target mRNA in intact cells was demonstrated in a 5'-rapid amplification of cDNA ends assay. In transfected cells, a suppression of the capacity to give skeletal metastases upon intracardial injection into nude rats was observed in cell clones with reduced expression of CAPL mRNA and protein, whereas in vitro and in vivo cell proliferation and tumorigenicity were unchanged. The results provide direct evidence that the expression level of the CAPL-encoded protein can determine the metastatic potential of osteosarcoma cells, and they demonstrate an association between reduced gene expression and proliferation-independent inhibition of the metastatic capacity of human tumor cells. The effects of the specific cleavage of CAPL mRNA indicate that the gene product is involved in key cellular functions associated with the metastatic process and suggest that therapeutic modulation of the protein function may represent a novel approach for inhibiting the metastatic spread of cancer cells.

BRAF V600E mutation in early-stage multiple myeloma: good response to broad acting drugs and no relation to prognosis.

In this study, we analyzed the prevalence and clone size of BRAF V600E mutation in 209 patients with multiple myeloma and related the results to clinical phenotype, response and survival. Biopsies were screened for BRAF V600E by allele-specific real-time PCR (AS-PCR). Positive results were confirmed by immunohistochemistry, Sanger sequencing and, in three patients from whom we had stored purified myeloma cells, whole-exome sequencing. Eleven patients (5.3%) were BRAF V600E mutation positive by AS-PCR and at least one other method. The fraction of mutated cells varied from 4 to 100%. BRAF V600E-positive patients had no characteristic clinical phenotype except for significantly higher levels of serum creatinine (125 versus 86 µmol/l) Seven of eleven patients responded with at least very good partial response to alkylators, immunomodulatory agents or proteasome inhibitors. Progression-free and overall survival were similar in patients with and without the mutation. By this integrated approach, we found that patients with BRAF V600E mutation responded very well to broad acting drugs and there was no relation to prognosis in early-stage myeloma. In particular, a large mutated cell fraction did not correlate with aggressive disease.

Detection of DNA variation in cancer.

Detection of DNA variation in cancer is central to the identification of relevant genes and mutations involved in the tumourigenic process. Diverse methods exist for such detection. One category of methods is for the detection of frequent sites for larger DNA alterations in cancer. Such areas may provide clues to the positioning of relevant genes, such as loss of heterozygosity (LOH) as in the case of tumour suppressor genes. Another category of methods is for the detection of single base mutations within specific genes. Frequently, such mutations may obliterate normal protein function. Among the most well-known are DGGE, SSCP, the HOT-method and direct sequencing. The methods for detection of DNA variation of these different levels are discussed. Two methods are presented in more detail. At the large-scale level, two-dimensional DNA fingerprinting has the potential of revealing the extent and location of altered DNA regions. This method is demonstrated using a panel of breast cancer patients. As an example of methods for the small-scale level, a recent development from DGGE, constant denaturant gel electrophoresis (CDGE) is demonstrated. This method has successfully been applied for the detection of mutations in a number of genes. Results with this method in studies of the RB1 gene are given, and its applicability as a screening tool for base mutations is discussed.

Three per cent of Norwegian ovarian cancers are caused by BRCA1 1675delA or 1135insA.

Our aim was to determine the prevalence of two Norwegian BRCA1 founder mutations in ovarian cancer patients, to identify carriers and their families for medical follow-up, and to study histopathological factors. Of a cohort of 727 ovarian cancer patients, 615 gave informed consent to testing. 2.9% (18/615) of the tested patients were found to be carriers of BRCA1 1675delA (n = 13) or 1135insA (n = 5). The total frequency of the mutations was 4.7% (8/171) in patients below 50 years of age, and zero (0/144) in patients above 70 years of age. In patients below 70 years of age, the frequency of 1675delA and 1135insA mutations was 2.8% and 1.0%, respectively. Out of 13 patients with 1675delA mutation, 4 had breast cancer. 14/16 (87.5%) families fulfilled clinical criteria for familial breast-ovarian cancer. Median age of onset of ovarian and breast cancer was 51 years and 37 years, respectively. Mutation carriers tended to have tumours with unfavourable prognostic factors. This is, to our knowledge, the highest reported frequency of founder mutations in a national ovarian cancer cohort (less than in the Ashkenazis). It seems justified to offer such testing to ovarian cancer patients below 70 years of age in Norway, identify their risk of breast cancer and offer medical follow-up to the families.

The BRCA1 syndrome and other inherited breast or breast-ovarian cancers in a Norwegian prospective series.

Inherited breast cancer is a heterogenous group of diseases. We examined this heterogeneity in a prospective series of inherited breast and ovarian cancers, previously demonstrated to include 84% of inherited cancers. Ninety-two tumours (65 breast and 27 ovarian) in 82 patients from 70 kindreds were prospectively diagnosed. Fifteen of the breast cancers were in situ, 50 were infiltrating. 40 (49%) of the 82 women carried a BRCA1 mutation, whereas no mutation in BRCA2 was found. Approximately, two-thirds of the BRCA1 mutation carriers had one of the four most frequent Norwegian founder mutations. Ninety-five per cent of the epithelial ovarian cancers occurred in BRCA1 mutation carrying women versus 38% of infiltrating breast cancers and 7% of carcinoma in situ of the breast. The BRCA1 syndrome was phenotypically distinct with invasive, high grade, oestrogen receptor-negative breast cancers and epithelial ovarian cancers. Non-BRCA1/2 inherited breast cancers included carcinoma in situ and lobular carcinoma and were frequently bilateral. Non-BRCA1/2 inherited breast cancer is not associated with epithelial ovarian cancer and in breast cancers has distinct biological characteristics, indicating that the different subgroups of inherited breast cancer may need different healthcare services.

Genetic epidemiology of BRCA1 mutations in Norway.

Familial breast-ovarian cancer has been demonstrated to be frequent but unevenly distributed in Norway. This was assumed to be caused by the reduced population size created by the medieval Bubonic plagues 25 generations ago, and by the following rapid expansion. We have previously reported that four mutations account for 68% of the BRCA1 mutation carriers. Subsequent analysis has resulted in a total of 100 separate families carrying one of these founder mutations. The four mutations occurred on one specific BRCA1 haplotype each. The 1675delA, 816delGT and 3347delAG families originated from the South-West coast of Norway with a few families in the north, while the traceable ancestors of the 1135insA families clustered along the historical inland road from the South-East to mid-Norway. The carriers of each of the four mutations today are descendants of one or a few individuals surviving the plagues. We may identify the majority of BRCA1 mutation carriers in Norway by screening for local founder mutations.

A BRCA1 founder mutation, identified with haplotype analysis, allowing genotype/phenotype determination and predictive testing.

We searched for a founder mutation in a population from one geographic region of Norway with prevalent breast/ovarian cancer families. We sampled 33 breast/ovarian cancer families and determined haplotypes of four markers linked to the BRCA1 region. Of the affected 33 index women, 13 (39.4%) shared one haplotype. In five (15% of total), an identical mutation was indicated by an abnormal truncated protein test (PTT) of exon 11 and shown to represent a 1675delA mutation. In the other index women, PTT of exon 11 showed no abnormality. No other BRCA1 founder mutation of this prevalence is likely because no other haplotype was more frequent in affecteds than in controls. All families with the 1675delA mutation in this geographic region may be considered as part of one large kindred. This allows a genotype-phenotype correlation to be precisely determined and used in genetic counselling for predictive testing within this kindred. Identification of identical haplotypes between unrelated affected individuals may be used to estimate the extent of founder effects for any mapped disease, without knowledge of the specific founder mutation.