Nanogenomics for medicine
2006; Future Medicine; Volume: 1; Issue: 2 Linguagem: Inglês
10.2217/17435889.1.2.147
ISSN1748-6963
Autores Tópico(s)Gene expression and cancer classification
ResumoNanomedicineVol. 1, No. 2 EditorialFree AccessNanogenomics for medicineClaudio NicoliniClaudio NicoliniNanoworld institute, University of Genova, Corso Europa 30–16132, Genoa, Italy. Published Online:7 Aug 2006https://doi.org/10.2217/17435889.1.2.147AboutSectionsPDF/EPUB ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinkedInReddit Figure 1. Alginate-poly-L-lysine-alginate gene microarray.Alginate-poly-L-lysine-alginate matrix after (A) interaction with Cy3-labeled cDNA, detail of nine spots (fluorescence microscope image; scanning electron microscope picture and before, (B) showing highly ordered anodic porous alumina on the surface of the microarray spot, resulting at the end of photolithographic microstructuring technique and two-step anodization process [5].Figure 2. Gene protein interactions.The lines that connect single genes represent a physical interaction between proteins, confirmed by various experimental methods, or the involvement in the same metabolic pathway [9].Figure 3. Interaction map of leader genes calculated with bioinformatics predictions and from experimental data.A new approach for medical diagnostics and therapy, called 'nanogenomics', is emerging from the interplay of bioinformatics and biomolecular microarrays on a previously unseen scale. This editorial summarizes its major features with a few key examples of molecular genomics applications to medicine.DNA microarrays have emerged as one of the most promising methods for the analysis of gene expression [1,2]. This technique allows the study of a huge number of genes (>10,000) in only one experiment and, therefore, can draw a picture of a whole genome. However, the huge amount of data arising from microarray experiments often raises experimental complications and difficulties in analysis. Moreover, most genes displayed on an array are often not involved directly in the cellular process being studied.New technological developments in DNA microarraysCommercial arrays with a lower number of genes (usually 150–200) are currently available, although the genes displayed are usually chosen without a precise consideration of the particular target of the study. Recently, a novel bioinstrumentation system, DNA analyzer (DNASER), was introduced for real-time acquisition and elaboration of images from fluorescent DNA microarrays [3]. In this configuration, a white light beam illuminates the target sample, allowing the images to be grabbed on a high sensibility and wide-band charge-coupled device camera (ORCA II, Hamamatsu). This high-performance device enables the acquisition of DNA microarray images and their processing in order to recognize the DNA chip spots, to analyze their superficial distribution on the glass slide and to evaluate their geometric and intensity properties. This is different from conventional techniques, since the spots analysis is fully automated and the DNASER does not require any additional information regarding the DNA microarray geometry.Two new alternative means of producing DNA microarrays and protein chips were also introduced: the first based on a novel biomolecular patterning on glass [4] and the second based on the nanostructuring of porous alumina matrix for a biomolecular microarray [5]. In the latter case, a photolithographic microstructuring technique for the ordered nanopore array fabrication uses a negative resist with hydrophobic properties, increasing the specificity to linking biomolecules. Nanoporous alumina is formed by the anodic process and yields straight holes with a high aspect ratio. Its use as a substrate for DNA microarray or protein chip applications offers several advantages over conventional supports, making it very attractive to use as a support for biological sample microarray applications. Alginate-poly-l-lysine-alginate (APA) matrix functionalization, after interaction with Cy3-labeled cDNA, demonstrates the possibility of obtaining an homogeneous functionalization of the spots, as required by DNA microarray technology. The minor fluorescence dishomogeneity in the spots does not appear to create a problem in quantitative fluorescence analysis, as confirmed by the preliminary analysis, which yielded a small coefficient of variation of the integral fluorescence per spot over the entire APA matrix (Figure 1). The subsequent test of functionalization and oligonucleotide immobilization demonstrated that the spots are suitable as a matrix for DNA microarray application [2]. Also, the hydrophobic properties provided by the resist to the spot boundary surface avoids possible liquid spread outside the spot and increases specificity to the linking biomolecules. The bidimensional photonic crystal-like behavior of the high aspect ratio ordered hollowed structure is such that the embedded dots or macromolecules emit light more selectively and with a controlled pattern distribution, thus reducing cross-interference between adjacent points, which is typical in traditional planar microarrays, with photoactive molecules emitting light in a Lambertian distribution. This characteristic of APA enables the preparation of very dense microarrays, reducing the distance between the spots. Presently, APA allows the preparation of up to 10,000 spots/cm2, similar to that obtainable by conventional techniques. However, the APA technique is limited only by the SU8 lithography constraints, which are less limited than the current constraints used for traditional glass substrate techniques. Our present efforts are focused on testing porous alumina matrices for protein-chip application printing protein microarrays, using the surface chemistries developed recently for functional proteomics [6] to exploit the advantages offered by this material over planar conventional supports. Few biological systems have been used in this context, integrating this technological effort to bioinformatics in an innovative approach to identify the leader genes in every cellular process and in the single pathology.Bioinformatics of leader genesThe key genes involved are typically identified by iterative searches of gene-related databases, as derived from DNA microarray experimentation. This has revealed and predicted interactions among those genes, assigned scores to each of the genes according to the numbers of interactions for each gene weighted by significance of each interaction and also applied several types of clustering algorithms to the genes based on the assigned scores [7]. With all clustering algorithms applied, both hierarchical and K-means, the same six 'leader' genes involved in controlling the cell cycle of human T lymphocytes were invariably selected. Six genes were identified in the cell cycle of human T lymphocytes that appear to be uniquely capable of switching between the stages of the human T-lymphocyte cell cycle [2,8].In recent years, mostly high-throughput approaches have been used to identify key genes for particular cellular processes and the usage of pre-existing knowledge bases of gene and protein interactions, combined from heterogeneous data sources, is rare. The 'leader gene' search/statistics algorithm consists of: iteratively searching GenBank® and PubMed® databases to identify the genes with proven involvement in the given cellular process; querying the STRING database to establish links among the genes [9]; assigning STRING association-based scores to each gene; and clustering of gene lists according to those scores to yield the final leader gene list.The leader gene algorithm is also being applied to predict the genes involved in cell cycle progression of human T lymphocytes [2,8], osteogenesis [Marconcini L et al. Unpublished Data], inflammatory processes and kidney graft rejection [Brouard S, Sivozhelevov V, Nicolini C et al. Unpublished Data]. The leader genes approach, with the text-mining scoring option off, appears to provide a list of the few most important genes relevant for the given cellular processes, according to the already available experimental data, and should be useful in interpreting the microarray expression data and in guiding clinical trials. Caution appears to be needed in the identification of leader genes obtained using either text-mining or no text-mining scoring and clustering, which frequently do not have a single gene in common [Sivozhelevov V, Giacomelli L, Nicolini C, Unpublished Data]. Indeed, the no text-mining approach produces more valid results [Sivozhelevov V, Giacomelli L, Nicolini C, Unpublished Data]. To perform the above analysis, two different softwares are required and must be installed on a dedicated computer: MATLAB and GenePix® software.Gene expression during the human T-lymphocyte cell cycleThis bioinformatics algorithm, based on the scoring of importance of genes and a subsequent cluster analysis, allowed us to determine the most important genes (leader genes) in the human T-lymphocyte cell cycle [7]. This particular cellular system is very well known and was quantitatively characterized previously [10,11]; therefore, it is a good starting point to verify our algorithm. In particular, we identified 238 genes involved in the control of the cell cycle. Most importantly, only six of them had been identified previously as leader genes [2]. Interestingly, they are involved in cell cycle control at important progression points; namely, the most important four at the transition from G0 to G1 phase (MYC) [12], at the progression in G1 phase (CDK4) [13], at the transitions from G1 to S (CDK2) [14] and from G2 to M phases (CDC2) [15,16]. The two remaining leader genes (CDKN1A and CDKN1B) are inhibitors of cyclin–CDK2 or –CDK4 complexes and thereby contribute to the control of the G1/S phase transition and G1 progression [17,18]. We also confirmed the results by analyzing changes in gene expression after 24, 48 and 72 h, using our DNASER [3,4]. The validity of the DNASER measurements was confirmed by standard fluorescence microscopy equipped with a cooled-charge device (CCD). This experimental analysis proved that DNASER is appropriate for monitoring gene expression during the human T-lymphocyte cell cycle [2].We calculated a final map of interactions among these eight high-ranking genes in the cell cycle of human T lymphocytes, which is shown in Figure 2, also representing their neighboring genes [7,18].Molecular genomics in graft toleranceA French group at Inserm is studying tolerance of kidney grafts. In particular, they have examined patients tolerating a kidney graft without any treatment and patients undergoing chronic rejection. They also performed a microarray analysis and identified a list of genes able to classify the two different classes. While these data concern 35,000 genes of a pangenomic array, the bioinformatic analysis being carried out by a joint cooperation between INSERM and Genova University (Italy) is presently in progress [Marconcini et al. Unpublished Data]. This involves two parallel approaches that will be run independently and, finally, compared with each other: ab-initio analysis and experimental analysis. The former concerns the identification of genes involved in kidney graft tolerance and of their leader genes; the latter concerns microarray data and clinical data reduction. In the former, our completely ab-initio set of genes involved in kidney transplant tolerance will be used [Marconcini L et al. Unpublished Data; Sivozhelevov V, Giacomelli L, Nicolini C, Unpublished Data]. This effort represents a big challenge to our search for the automatic identification of the aforementioned leader genes. In particular, it is important to note that there is no obvious necessary direct correlation between leader genes identified by ab-initio research and the quantitative changes in expression that are monitored by experimental analysis.OsteogenesisWe tentatively identified seven leader genes in the osteogenic process [Marconcini L et al.Unpublished Data]. They were classified according to their involvement in osteogenesis subprocesses (cell adhesion and proliferation, ossification, skeletal development and calcium-ion binding). On this basis, we are presently constructing a simple array (2 × 4 × 4 genes, for a total of 16 genes repeated twice) using APA technology and displaying the leader genes and an equivalent number of controls, in order to begin to study the osteogenetic process at the molecular level.ConclusionsPutative leader genes undergo some changes in gene expression, however, the level of change is not necessarily related to their leader gene status. In fact, leader genes are the genes that interact most and are also involved in a process, not necessarily the ones with most varying expression. This is also apparent in the human T-lymphocyte cell cycle, where the gene with the most varying expression is CCNH, which is not one of the leader genes. Ab-initio 'leader plus class B' genes were scanned against the list of experimental genes with changing the expression levels (more numerous than the leader genes identified ab-initio). An example of the typical result is summarized in Table 1. These preliminary results demonstrate that the text-mining approach is dangerous to use for interpreting microarray-expression data. Indeed, theoretical and experimental leader gene sets are clearly different and, for this reason, we are attempting to check links between the two sets of leaders: links between theoretical and experimental leader genes using text-mining and the same without text-mining. The final results are still in progress and very preliminarily leader genes identified with text-mining appear to have almost no interactions shown. The validity of this approach in relating bioinformatics predictions and experimental data is, therefore, much lower than the 'no text-mining' approach.Leader genes identified with a completely independent bioinformatics prediction appear to interact closely with genes changing expression in experimental analysis. An important change in expression of these leader genes among the two conditions is not necessary for their function. This was also proved in our previous microarray study on the T-lymphocyte cell cycle: overall, MYC was the most important gene in the whole process (it is necessary to enter the G1 phase), however, it changes its expression very slightly from quiescent to replicating cells. Many questions have still to be answered before we reach a conclusion to this open question although, in any cell system, the change in expression of a particular gene could be considered as the ultimate consequence of a complex network (Figure 3) of biochemical interactions, whose most important nodes could be the leader genes, as identified with ab-initio prediction.Several questions emerge, including: what should be considered as the true leader gene set – the theoretical only or theoretical plus experimental? To attempt to answer this question, we are presently scanning the experimental leader genes for presence in our total (not leader) theoretical gene set in cooperation with INSERM. Our previous findings suggest that the mere changing in expression of a particular gene is not meaningful by itself, only if it is put in a proper framework [2,7,8]. This change can often be considered a consequence of a more complex network of events, starting from leader genes (identified with bioinformatic predictions) which often do not vary their expression until identified as significant using pangenomic arrays. The leader gene approach, tentatively validated by experimental analysis using DNA microarrays on a model system, suggests a more rational approach to such experimental techniques and methods. The application of bioinformatics studies and the identification of leader genes can predict the most important genes in a particular cellular process. In this way, it could become possible to design smaller microarrays that display only the most interesting genes for a specific cellular process and, thus, are much easier to interpret. Similarly, protein microarrays are also used for the study of protein–protein and protein–gene interactions [6,19] and, for the DNA microarrays, the leader gene approach could simplify their analysis by reducing the protein displayed to the most important ones to be subsequently tested by mass spectrometry or by ad hoc experimentation with NMR and x-ray crystallography.AcknowledgementsThis work was supported by an Organic Nanosciences and Nanotechnology FIRB Grant (RBNE01X3CE) from MIUR (Italy) to Fondazione Elba (Rome) and CIRNNOB at the University of Genova, and by a FIRB International on Functional Proteomics and Cell Cycle to CIRNNOB at the University of Genova.Table 1. Presence (%) of potential 'leader gene' candidates in a typical total pangenomic list, depending on the expansion algorithm and scoring-clustering algorithm.ExpansionText-miningNo text-miningSingle text-mining expansionScoring and clustering Text-mining594443No text-mining303629Bibliography1 Butte A: The use and analysis of microarray data. Nat. Rev. Drug. Discov.1,951–960 (2002).Crossref, Medline, CAS, Google Scholar2 Nicolini C, Spera R, Stura E, Fiordoro S, Giacomelli L: Gene expression in the cell cycle of human T lymphocytes: II. Experimental determination by DNASER technology. J. Cell. Biochem.97,1151–1159 (2006).Crossref, Medline, CAS, Google Scholar3 Nicolini C, Malvezzi M, Tomaselli A, Sposito D, Tropiano G, Borgogno E: DNASER I: layout and data analysis. IEEE Trans. Nanobioscience1,67–72 (2002).Crossref, Medline, Google Scholar4 Troitsky, V, Ghisellini P, Pechkova E, Nicolini C: DNASER II. Novel surface patterning for biomolecular microarray. IEEE Trans. Nanobioscience1,73–77 (2002).Crossref, Medline, Google Scholar5 Grasso V, Lambertini V, Ghisellini P et al.: Nanostructuring of a porous alumina matrix for a biomolecular microarray. Nanotechnology17,795–798 (2006).Crossref, CAS, Google Scholar6 LaBaer J, Ramachandran N: Protein microarrays as tools for functional proteomics. Curr. Opin. Chem. Biol.9,14–19 (2005).Crossref, Medline, CAS, Google Scholar7 Sivozhelevov V, Giacomelli L, Tripathi S, Nicolini C: Gene expression in the cell cycle of human T lymphocytes: 1. Predicted gene and protein networks. J. Cell. Biochem.97,1137–1150 (2006).Crossref, Medline, Google Scholar8 Giacomelli L, Nicolini C: Gene expression of human T lymphocytes cell cycle: experimental and bioinformatic analysis. J. Cell. Biochem. DOI 10.1002/jcb.20991 (2006) (Epub ahead of print).Google Scholar9 vonMering C, Jensen LJ, Snel B et al.: STRING: known and predicted protein–protein associations, integrated and transferred across organism. Nucleic Acids Res.33,D433–D437 (2005).Medline, Google Scholar10 Cantrell D: Protein kinase B (Akt) regulation and function in T lymphocytes. Semin. Immunol.14,19–26 (2002).Crossref, Medline, CAS, Google Scholar11 Abraham S, Vonderheid E, Zietz S, Kendall FM, Nicolini C: Reversible (G0) and nonreadily reversible (Q) noncycling cells in human peripheral blood. Immunological, structural and biological characterization. Cell Biophys.2,353–371 (1980).Crossref, Medline, CAS, Google Scholar12 Oster SK, Ho CS, Soucie EL, Penn LZ: The myc oncogene: marvellously Y complex. Adv. Cancer Res.84,81–154 (2002).Crossref, Medline, CAS, Google Scholar13 Modiano JF, Mayor J, Ball C, Fuentes MK, Linthicum DS: CDK4 expression and activity are required for cytokine responsiveness in T cells. J. Immunol.165,6693–6702 (2000).Crossref, Medline, CAS, Google Scholar14 Kawabe T, Suganuma M, Ando T, Rimura M, Hori H, Okamoto T: Cdc25C interacts with PCNA at G2/M transition. Oncogene21,1717–1726 (2002).Crossref, Medline, CAS, Google Scholar15 Baluchamy S, Rajabi HN, Thimmapaya R, Navaraj A, Thimmapaya B: Repression of c-Myc and inhibition of G1 exit in cells conditionally overexpressing p300 that is not dependent on its histone acetyltransferase activity. Proc. Natl Acad. Sci. USA100,9524–9529 (2003).Crossref, Medline, CAS, Google Scholar16 Torgler R, Jakob S, Ontsouka E et al.: Regulation of activation-induced Fas (CD95/Apo-1) ligand expression in T cells by the cyclin B1/Cdk1 complex. J. Biol. Chem.279,37334–37342 (2004).Crossref, Medline, CAS, Google Scholar17 Jerry DJ, Dickinson ES, Roberts AL, Said TK: Regulation of apoptosis during mammary involution by the p53 tumour suppressor gene. J. Dairy Sci.85,1103–1110 (2002).Crossref, Medline, CAS, Google Scholar18 Chang BL, Zheng SL, Isaacs SD et al.: A polymorphism in the CDKN1B gene is associated with increased risk of hereditary prostate cancer. Cancer Res.64,1997–1999 (2004).Crossref, Medline, CAS, Google Scholar19 Ramachandran N, Hainsworth E, Bhullar B et al.: Self-assembling protein microarrays. Science305,86–90 (2004).Crossref, Medline, CAS, Google ScholarFiguresReferencesRelatedDetailsCited ByLeader gene of Corynebacterium pseudotuberculosis may be useful in vaccines against caseous lymphadenitis of goats: a bioinformatics approachJournal of Veterinary Medical Science, Vol. 80, No. 8Nanotechnology for Omics-Based Ocular Drug DeliveryNanotechnology for Omics-Based Ocular Drug DeliveryOmics and Its Application in Clinical Nanotechnology and Nanodiagnostics5 January 2017Nanotechnology for Omics-Based Ocular Drug DeliveryProteomics and Proteogenomics Approaches for Oral DiseasesNanoproteomics enabling personalized nanomedicineAdvanced Drug Delivery Reviews, Vol. 64, No. 13Real-Time Quantitative Polymerase Chain Reaction Analysis of Patients With Refractory Chronic PeriodontitisJournal of Periodontology, Vol. 82, No. 7Nanoproteomics for nanomedicineClaudio Nicolini & Eugenia Pechkova22 July 2010 | Nanomedicine, Vol. 5, No. 5Nanogenomics in medicine2 November 2009 | Wiley Interdisciplinary Reviews: Nanomedicine and Nanobiotechnology, Vol. 2, No. 1DNA bridging of yeast chromosomes VIII leads to near-reciprocal translocation and loss of heterozygosity with minor cellular defects18 November 2008 | Chromosoma, Vol. 118, No. 2Bioinformatic Prediction of Leader Genes in Human PeriodontitisJournal of Periodontology, Vol. 79, No. 10Immunosuppressive drug-free operational immune tolerance in human kidney transplants recipients. Part II. Non-statistical gene microarray analysis1 January 2008 | Journal of Cellular Biochemistry, Vol. 103, No. 6 Vol. 1, No. 2 Follow us on social media for the latest updates Metrics History Published online 7 August 2006 Published in print August 2006 Information© Future Medicine LtdPDF download
Referência(s)