Literatura académica sobre el tema "Array hybridisation analysis"

AbstractObjectiveEfficient diagnosis of an underlying genetic aetiology in a patient with congenital heart disease is essential to optimising clinical care. Copy number variants are one aetiology of congenital heart disease; the majority are identifiable by targeted fluorescence in situ hybridisation or array comparative genomic hybridisation, not by classical cytogenetic analysis. This study assessed the utility of array comparative genomic hybridisation as a first-tier diagnostic test for neonates with congenital heart disease.Study designA prospective chart review of neonates with congenital heart disease in the Cardiac Intensive Care Unit at Children’s Hospital of Pittsburgh of UPMC was performed. Patients were tested by array comparative genomic hybridisation and classical cytogenetic analysis simultaneously. Data collected included all chromosome abnormalities detected, physical examination findings, and imaging results. McNemar’s test was used to compare detection of array comparative genomic hybridisation and classical cytogenetic analysis.ResultsOf 45 patients, three (6.7%) had an abnormality detected by classical cytogenetic analysis and an additional 10 (22.2%) had a copy number variant detected by array comparative genomic hybridisation, highlighting an increased detection rate (p=0.008). Several of these copy number variants had unclear clinical significance, requiring additional investigation. The prevalence of dysmorphology and/or comorbidity in this population was 72%. Identification of dysmorphic features was greater when assessed by a geneticist than by providers of different subspecialties.ConclusionsArray comparative genomic hybridisation has significant clinical utility as a first-tier test in this population, but it carries the potential for incidental findings and results of uncertain clinical significance. Collaboration between cardiologists and medical geneticists is essential to providing optimal clinical care.

AbstractBackground: CHD is the leading cause of mortality due to birth defects. Array comparative genomic hybridisation (aCGH) detects submicroscopic copy number changes and may improve identification of the genetic basis of CHD. Methods: This is a retrospective analysis of 1252 patients from a regional referral centre who had undergone aCGH. Of the patients, 173 had CHD. A whole-genome custom-designed oligonucleotide array with >44,000 probes was used to detect copy number changes. Results: Of the 1252 patients, 335 (26.76%) had abnormal aCGH results. Of the 173 patients with CHD, 50 (28.9%) had abnormal aCGH results versus 284 (26.3%) of 1079 non-cardiac patients. There were six patients with CHD who had well-described syndromes such as Wolf–Hirschhorn, trisomy 13, DiGeorge, and Williams. Of the patients with CHD, those with left-sided heart disease had the highest proportion (14/31; 45.13%) of abnormal aCGH results, followed by those with conotruncal heart disease (10/29; 34.48%), endocardial cushion defects (13/50; 26%), complex/other heart disease (12/52; 23.08%), and patent ductus arteriosus (1/11; 9.09%). Conclusions: Patients with CHD are at a substantial risk of having microdeletions and microduplications. The incidence of abnormalities on aCGH analysis is higher than identified with karyotype, and identification of copy number changes may help identify the genetic basis of the specific heart defects. However, aCGH may not have a significant diagnostic yield in those with isolated CHD. Further research using larger data sets may help identify candidate genes associated with CHD.

AbstractThe study aimed to show the chromosomal copy number variations responsible for the aetiology in patients with isolated conotruncal heart anomaly by array comparative genomic hybridisation and identify candidate genes causing conotruncal heart disease. A total of 37 patients, 17 male, and 20 female, with isolated conotruncal heart anomalies, were included in the study. No findings indicated any syndrome in terms of dysmorphology in the patients.Results:Copy number variations were detected in the array comparative genomic hybridisation analysis of five (13.5%) of 37 patients included in the study. Three candidate genes (PRDM16, HIST1H1E, GJA5) found in these deletion and duplication regions may be associated with the conotruncal cardiac anomaly.Conclusion:CHDs can be encountered as the first and sometimes the single finding of many genetic disorders in children. It is thought that genetic tests, especially array comparative genomic hybridisation, may be beneficial for children with CHD since the diagnosis of genetic diseases in these patients as early as possible will help to prevent or reduce complications that may develop in the future. Also, it would be possible to detect candidate genes responsible for conotruncal cardiac anomalies with array comparative genomic hybridisation.

The oviduct epithelium undergoes marked morphological and functional changes during the oestrous cycle. To study these changes at the level of the transcriptome we did a systematic gene expression analysis of bovine oviduct epithelial cells at oestrus and dioestrus using a combination of subtracted cDNA libraries and cDNA array hybridisation. A total of 3072 cDNA clones of two subtracted libraries were analysed by array hybridisation with cDNA probes derived from six cyclic heifers, three of them slaughtered at oestrus and three at dioestrus. Sequencing of cDNAs showing significant differences in their expression levels revealed 77 different cDNAs. Thirty-seven were expressed at a higher level at oestrus, for the other 40 genes expression levels were higher at dioestrus. The identified genes represented a variety of functional classes. During oestrus especially genes involved in the regulation of protein secretion and protein modification, and mRNAs of secreted proteins, were up-regulated, whereas during dioestrus particularly transcripts of genes involved in transcription regulation showed a slight up-regulation. The concentrations of seven selected transcripts were quantified by real-time RT-PCR to validate the cDNA array hybridisation data. For all seven transcripts, RT-PCR results were in excellent correlation (r>0.92) with the results obtained by array hybridisation. Our study is the first to analyse changes in gene expression profiles of bovine oviduct epithelial cells during different stages of the oestrous cycle, providing a starting point for the clarification of the key transcriptome changes in these cells.

AimsTo evaluate a reverse-hybridisation assay (strip assay) designed for the sensitive detection of 10 mutations in codons 12 and 13 of the KRAS gene. The strip assay relies on mutant-enriched PCR followed by reverse-hybridisation of biotinylated amplification products to oligonucleotide probes immobilised as an array of parallel lines on nitrocellulose test strips.MethodsThe strip assay was used to analyse genomic DNA isolated from 120 formalin-fixed paraffin-embedded (FFPE) ovarian tissue samples. The samples were analysed in parallel using a biochip-based protocol (biochip assay) covering the same mutation spectrum, and results were compared with respect to sensitivity, specificity and operational input.ResultsThe strip assay identified 19 (16%) of 120 FFPE samples to carry a KRAS mutation; results were in agreement with those obtained by biochip hybridisation. Both assays had an analytical sensitivity of 1% when performed on FFPE-extracted DNA with approximately the same operational input needed for post-PCR processing. In contrast to the biochip assay, strip assay hybridisation may be automated to a large extent.ConclusionsThe strip assay is an accurate and sensitive tool for the low to medium throughput detection of KRAS mutation in genomic DNA isolated from FFPE tissue.

Despite significant advances in diagnosis and treatment, breast cancer remains the leading cause of cancer-related deaths in Australian women. Colorectal cancer is the second most common cancer in both males and females; after prostate and breast cancer, respectively, and excluding non-melanocytic skin cancer. Both breast cancer and colorectal cancer follow a common progressive course of illness; presenting (at least initially) with benign symptoms that can be treated by ablation (or removal) of the affected area. Cancer progression is associated with breakdown of tissue barriers (such as basement membranes), leading to the spread of cancer cells (via the vasculature or lymphatic system), and the establishment of secondary metastatic disease at green-field sites. Secondary tumours presenting in the lungs, ovaries, liver, bone, or brain are associated with chronic-debilitating symptoms that are difficult to treat, and will result in death. In the case of breast and colon cancer, effective early therapeutic intervention does have a significant impact upon patient survival. Tumour progression in breast and colon carcinomas is characterised by invasion of the surrounding stroma, and the acquisition of stromal characteristics, by previously epithelial cells. This progression is associated with the expression of extracellular proteases (ECPs) and increased motility. The process of mesenchymal transformation that tumour cells undergo is also referred to as the epithelial to mesenchymal transition (EMT). In general terms the aim of the study, presented in this thesis, was to investigate gene expression in cancer biology; and to characterise changes in breast cancer and colon cancer, with a focus on those genes, and gene products that may play a role in metastasis, including a family of ECPs, the matrix metalloproteinases (MMPs). In our laboratory, we have applied methods in microdissection, differential display polymerase chain reaction amplification (DD-PCR), and array hybridisation analysis to identify gene expression patterns in late stage archival formalin fixed paraffin embedded (FFPE) breast tumour biopsies that may be indicative of the EMT; or the response to the surrounding stroma/interstitium to the presence of the tumour.' The quality of nucleic acid obtainable from FFPE material presents a considerable challenge for gene expression studies. In order to identify tissue specific gene expression patterns, DD-PCR products, amplified from message obtained following segregation of tumour tissue from surrounding stroma, was hybridised to arrayed cDNA libraries created from stromal tumours, or sarcomas. In this way, 21 known genes, or expressed sequence tags (ESTs), were identified. These included the cytoskeletal element and EMT marker, vimentin, the mammary developmental factor and, signal transducer and activator of transcription (STAT)-3, and the cargo selection protein (TIP47). Seventeen genes showed differential expression in either the tumour, or stromal fractions. When applied to transformed breast cancer cell lines (MDA-MB-435 & T47D) DD-array analysis revealed a further 17 genes that were differentially regulated in invasive cells, compared with those displaying a less invasive phenotype. Six of the ESTs identified by DD-PCR array analysis, had no known (or predicted) function. For example, bcaf-2 was identified as the 3'-end of a putative open reading frame (ORF) localised to chromosome 6, while bcaf-10 showed homology with a known ORF. In order to analyse the expression of these bcafs further, a stromal cell culture model, representative of the original osteosarcoma cDNA libraries from which they were obtained, was used. In this model, CD14' (or adherent) peripheral blood mononuclear cells (PBMCs) treated with macrophage colony stimulating factor (M-CSF), can be allowed to differentiate into macrophage-like (ML) cells; while cells treated with M-CSF, and the receptor activator of NF-KB ligand (RANKL) will differentiate into multinucleate osteoclast-like (OCL) multinucleate giant cells. Uniquely, the stromal EST, bcaf-2 was expressed only by RANKL-treated (or OCL) cells. bcaf-2 and other ESTs, identified by DD-PCR analysis (and recently published) are the subject of on going research in our laboratory. The role of RANKL in mammary gland development and bone metastasis suggested that the identification of a RANKL-regulated stromal factor in breast tissue (bcaf-2) was not an artefact. RANKL is a membrane-bound, member of the tumour necrosis factor (TNF)-a cytokine super family. In order to test the hypothesis that RANKL might act as an inflammatory cytokine to regulate clinically significant stromal gene expression in the breast, we employed quantitative real time PCR analysis to examine the relative levels of selected members of a group of metal dependent ECPs, the matrix metalloproteinases (MMPs). RNA was extracted from ML cells and OCL cells, as well as RANK positive breast cancer cell lines (T47D, MDA-MB-435 & MCF-7). When the relative levels of protease mRNA were compared we demonstrated a significant (>20- fold) specific increase in collagenase (collagenase 2lMMP-8 and collagenase 3lMMP-13), and the tissue inhibitor of MMP (TIMP)-2 expression in M-CSF and RANKL treated PBMCs cells. When the assay was applied to RANKL treated breast cancer cell lines (MCF-7, T47D & MDAMB- 231), minor (40-fold) but potentially significant alterations in stromal protease gene expression were observed. The changes observed did not however, support the hypothesis that RANKL might act as an inflammatory cytokine to induce significant alterations in ECP expression in breast cancer cells. To investigate the role of RANKL as a driver of EMT in aberrant breast epithelium, total message (mRNNcDNA) from T47D, MCF-7, MDA-MB-231 cells, and message from the same cell lines treated with RANKL were compared by comparative fluorescent cDNA microarray analysis. Of the 1,700 targets available on the arrays, this study identified 160 that were differentially expressed in RANKL treated cells. The results suggest that RANKL may promote rather than suppress a mammary epithelial phenotype in breast cancer. In fact a putative mesenchymal to epithelial transition (MET) was observed following microscopic analysis, and this finding is the subject of on going research in our laboratory. Sporadic structural alterations in certain mitogenic factors represent important early events in cancer progression, while inherited mutations govern familial susceptibility to disease. In colon cancer, a close link exists between Winglessllnt (WNT) signalling, disease pathology, and the expression of MMPs. To examine the relationship between protease expression and structural genetic alterations in this EMT-linked signalling pathway, and others, we applied combined QPCR analysis of MMP expression and PCR-Single Strand Conformation Analysis (SSCA) to 26 colonic tumours, and patient-matched normal colonic mucosa. In this study, significant correlations between the expression of ECPs, and a key mediator of WNT signalling (p-catenin) were identified. While tumours possessing specific functional mutations in K-Ras, were found to group with phenotypic clustering based on protease gene expression. This result may be due to an interruption of normal interactions between RasIRaf signalling and transforming growth factor (TGF) P signalling, via Sma- and Mad- related protein (SMAD) signalling. These results demonstrate that the already identified link between mutations in kinase signalling, and aspects of gross colon tumour morphology (such as dysplasia) may be due to aberrant MMP expression patterning. The final aim of this research was to utilise methods developed in microdissection and specific Q-PCR analysis, to identify whether tumour-stroma differences in MMP gene expression might be used as markers of disease pathology. Total RNA from tumour, and biopsy-matched adjacent stromal tissue were segregated from 35 FFPE archival breast tumour biopsies. Comparison with stroma identified specific associations between TIMP-2 expression in the stroma and lymph node involvement, as well as stromelysin-3 (MMP-I I ) and TIMP-I expression and calcification of the tumour. Furthermore, a significant correlation was identified in the pattern of gelatinase (gelatinase AIMMP-2 & gelatinaseB1MMP-9) expression; while no significant correlation was identified in tumour-stroma MMP gene expression differences, and tumour grade, or hormone receptor status. These results suggest that coordinated changes within the tumour, and proximal stromal tissues (rather than tissue specific changes per se), regulate pathologically significant changes in breast carcinogenesis. In conclusion, this thesis describes the use of novel techniques in specific and global gene expression analysis that permitted examination of stromal gene expression changes in epithelial tumour progression. Microdissection facilitated localisation of expression to particular tissues, while cell culture models provided material with which to optimise and demonstrate the efficacy of techniques used (where tumour material itself was not abundant). Furthermore, we have identified significant and specific correlations between general stromal protease gene expression changes, a putative mammary epithelial differentiation factor (RANKL), alterations in growth factor signalling, and epithelial tumour pathology in the breast and colon. The combination of techniques developed in this study may assist in improvement of categorisation of tumours in clinical pathology. Specifically, the development of novel grading systems that link underlying molecular genetic changes with changes in tumour pathology. These processes may assist to improve diagnosis and provide more effective patient/tumour-specific drug therapies.

Despite significant advances in diagnosis and treatment, breast cancer remains the leading cause of cancer-related deaths in Australian women. Colorectal cancer is the second most common cancer in both males and females; after prostate and breast cancer, respectively, and excluding non-melanocytic skin cancer. Both breast cancer and colorectal cancer follow a common progressive course of illness; presenting (at least initially) with benign symptoms that can be treated by ablation (or removal) of the affected area. Cancer progression is associated with breakdown of tissue barriers (such as basement membranes), leading to the spread of cancer cells (via the vasculature or lymphatic system), and the establishment of secondary metastatic disease at green-field sites. Secondary tumours presenting in the lungs, ovaries, liver, bone, or brain are associated with chronic-debilitating symptoms that are difficult to treat, and will result in death. In the case of breast and colon cancer, effective early therapeutic intervention does have a significant impact upon patient survival. Tumour progression in breast and colon carcinomas is characterised by invasion of the surrounding stroma, and the acquisition of stromal characteristics, by previously epithelial cells. This progression is associated with the expression of extracellular proteases (ECPs) and increased motility. The process of mesenchymal transformation that tumour cells undergo is also referred to as the epithelial to mesenchymal transition (EMT). In general terms the aim of the study, presented in this thesis, was to investigate gene expression in cancer biology; and to characterise changes in breast cancer and colon cancer, with a focus on those genes, and gene products that may play a role in metastasis, including a family of ECPs, the matrix metalloproteinases (MMPs). In our laboratory, we have applied methods in microdissection, differential display polymerase chain reaction amplification (DD-PCR), and array hybridisation analysis to identify gene expression patterns in late stage archival formalin fixed paraffin embedded (FFPE) breast tumour biopsies that may be indicative of the EMT; or the response to the surrounding stroma/interstitium to the presence of the tumour.' The quality of nucleic acid obtainable from FFPE material presents a considerable challenge for gene expression studies. In order to identify tissue specific gene expression patterns, DD-PCR products, amplified from message obtained following segregation of tumour tissue from surrounding stroma, was hybridised to arrayed cDNA libraries created from stromal tumours, or sarcomas. In this way, 21 known genes, or expressed sequence tags (ESTs), were identified. These included the cytoskeletal element and EMT marker, vimentin, the mammary developmental factor and, signal transducer and activator of transcription (STAT)-3, and the cargo selection protein (TIP47). Seventeen genes showed differential expression in either the tumour, or stromal fractions. When applied to transformed breast cancer cell lines (MDA-MB-435 & T47D) DD-array analysis revealed a further 17 genes that were differentially regulated in invasive cells, compared with those displaying a less invasive phenotype. Six of the ESTs identified by DD-PCR array analysis, had no known (or predicted) function. For example, bcaf-2 was identified as the 3'-end of a putative open reading frame (ORF) localised to chromosome 6, while bcaf-10 showed homology with a known ORF. In order to analyse the expression of these bcafs further, a stromal cell culture model, representative of the original osteosarcoma cDNA libraries from which they were obtained, was used. In this model, CD14' (or adherent) peripheral blood mononuclear cells (PBMCs) treated with macrophage colony stimulating factor (M-CSF), can be allowed to differentiate into macrophage-like (ML) cells; while cells treated with M-CSF, and the receptor activator of NF-KB ligand (RANKL) will differentiate into multinucleate osteoclast-like (OCL) multinucleate giant cells. Uniquely, the stromal EST, bcaf-2 was expressed only by RANKL-treated (or OCL) cells. bcaf-2 and other ESTs, identified by DD-PCR analysis (and recently published) are the subject of on going research in our laboratory. The role of RANKL in mammary gland development and bone metastasis suggested that the identification of a RANKL-regulated stromal factor in breast tissue (bcaf-2) was not an artefact. RANKL is a membrane-bound, member of the tumour necrosis factor (TNF)-a cytokine super family. In order to test the hypothesis that RANKL might act as an inflammatory cytokine to regulate clinically significant stromal gene expression in the breast, we employed quantitative real time PCR analysis to examine the relative levels of selected members of a group of metal dependent ECPs, the matrix metalloproteinases (MMPs). RNA was extracted from ML cells and OCL cells, as well as RANK positive breast cancer cell lines (T47D, MDA-MB-435 & MCF-7). When the relative levels of protease mRNA were compared we demonstrated a significant (>20- fold) specific increase in collagenase (collagenase 2lMMP-8 and collagenase 3lMMP-13), and the tissue inhibitor of MMP (TIMP)-2 expression in M-CSF and RANKL treated PBMCs cells. When the assay was applied to RANKL treated breast cancer cell lines (MCF-7, T47D & MDAMB- 231), minor (40-fold) but potentially significant alterations in stromal protease gene expression were observed. The changes observed did not however, support the hypothesis that RANKL might act as an inflammatory cytokine to induce significant alterations in ECP expression in breast cancer cells. To investigate the role of RANKL as a driver of EMT in aberrant breast epithelium, total message (mRNNcDNA) from T47D, MCF-7, MDA-MB-231 cells, and message from the same cell lines treated with RANKL were compared by comparative fluorescent cDNA microarray analysis. Of the 1,700 targets available on the arrays, this study identified 160 that were differentially expressed in RANKL treated cells. The results suggest that RANKL may promote rather than suppress a mammary epithelial phenotype in breast cancer. In fact a putative mesenchymal to epithelial transition (MET) was observed following microscopic analysis, and this finding is the subject of on going research in our laboratory. Sporadic structural alterations in certain mitogenic factors represent important early events in cancer progression, while inherited mutations govern familial susceptibility to disease. In colon cancer, a close link exists between Winglessllnt (WNT) signalling, disease pathology, and the expression of MMPs. To examine the relationship between protease expression and structural genetic alterations in this EMT-linked signalling pathway, and others, we applied combined QPCR analysis of MMP expression and PCR-Single Strand Conformation Analysis (SSCA) to 26 colonic tumours, and patient-matched normal colonic mucosa. In this study, significant correlations between the expression of ECPs, and a key mediator of WNT signalling (p-catenin) were identified. While tumours possessing specific functional mutations in K-Ras, were found to group with phenotypic clustering based on protease gene expression. This result may be due to an interruption of normal interactions between RasIRaf signalling and transforming growth factor (TGF) P signalling, via Sma- and Mad- related protein (SMAD) signalling. These results demonstrate that the already identified link between mutations in kinase signalling, and aspects of gross colon tumour morphology (such as dysplasia) may be due to aberrant MMP expression patterning. The final aim of this research was to utilise methods developed in microdissection and specific Q-PCR analysis, to identify whether tumour-stroma differences in MMP gene expression might be used as markers of disease pathology. Total RNA from tumour, and biopsy-matched adjacent stromal tissue were segregated from 35 FFPE archival breast tumour biopsies. Comparison with stroma identified specific associations between TIMP-2 expression in the stroma and lymph node involvement, as well as stromelysin-3 (MMP-I I ) and TIMP-I expression and calcification of the tumour. Furthermore, a significant correlation was identified in the pattern of gelatinase (gelatinase AIMMP-2 & gelatinaseB1MMP-9) expression; while no significant correlation was identified in tumour-stroma MMP gene expression differences, and tumour grade, or hormone receptor status. These results suggest that coordinated changes within the tumour, and proximal stromal tissues (rather than tissue specific changes per se), regulate pathologically significant changes in breast carcinogenesis. In conclusion, this thesis describes the use of novel techniques in specific and global gene expression analysis that permitted examination of stromal gene expression changes in epithelial tumour progression. Microdissection facilitated localisation of expression to particular tissues, while cell culture models provided material with which to optimise and demonstrate the efficacy of techniques used (where tumour material itself was not abundant). Furthermore, we have identified significant and specific correlations between general stromal protease gene expression changes, a putative mammary epithelial differentiation factor (RANKL), alterations in growth factor signalling, and epithelial tumour pathology in the breast and colon. The combination of techniques developed in this study may assist in improvement of categorisation of tumours in clinical pathology. Specifically, the development of novel grading systems that link underlying molecular genetic changes with changes in tumour pathology. These processes may assist to improve diagnosis and provide more effective patient/tumour-specific drug therapies.
Thesis (PhD Doctorate)
Doctor of Philosophy (PhD)
School of Health Sciences
Faculty of Health Sciences
Full Text

Im ersten Teil der Arbeit wurden Strategien zur Analyse von Transkripten erarbeitet. Die ersten Versuche zielten darauf ab, in mit Glaskapillaren genommenen Einzelzellproben verschiedener Gewebeschichten RT-PCR durchzuführen, um spezifische Transkripte nachweisen zu können. Dies gelang für eine Reihe von Genen aus verschiedenen Pflanzenspezies. Dabei konnten sowohl Transkripte stark wie auch schwach exprimierter Gene nachgewiesen werden.
Für die Erstellung von Gewebe-spezifischen Expressionsprofilen war es notwendig, die in vereinigten Zellproben enthaltene mRNA zunächst zu amplifizieren, um eine ausreichende Menge für Arrayhybridisierungen zu erhalten. Vor der Vermehrung wurde die mRNA revers transkribiert. Es wurden daran anschließend verschiedene Amplifikationsstrategien getestet: Die neben Tailing, Adapterligation und anderen PCR-basierenden Protokollen getestete Arbitrary-PCR hat sich in dieser Arbeit als einfache und einzige Methode herausgestellt, die mit so geringen cDNA-Mengen reproduzierbar arbeitet. Durch Gewebe-spezifische Array-hybridisierungen mit der so amplifizierten RNA konnten schon bekannte Expressionsmuster verschiedener Gene, vornehmlich solcher, die an der Photosynthese beteiligt sind, beobachtet werden. Es wurden aber auch eine ganze Reihe neuer offensichtlich Gewebe-spezifisch exprimierter Gene gefunden. Exemplarisch für die differentiell exprimierten Gene konnte das durch Arrayhybridisierungen gefundene Expressionsmuster der kleinen Untereinheit von Rubisco verifiziert werden. Hierzu wurden Methoden zum Gewebe-spezifischen Northernblot sowie semiquantitativer und Echtzeit-Einzelzell-RT-PCR entwickelt.
Im zweiten Teil der Arbeit wurden Methoden zur Analyse von Metaboliten einschließlich anorganischer Ionen verwendet. Es stellte sich heraus, daß die multiparallele Methode der Gaschromatographie-Massenspektrometrie keine geeignete Methode für die Analyse selbst vieler vereinigter Zellinhalte ist. Daher wurde auf Kapillarelektrophorese zurückgegriffen. Eine Methode, die mit sehr kleinen Probenvolumina auskommt, eine hohe Trennung erzielt und zudem extrem geringe Detektionslimits besitzt. Die Analyse von Kohlenhydraten und Anionen erfordert eine weitere Optimierung. Über UV-Detektion konnte die K+-Konzentration in verschiedenen Geweben von A. thaliana bestimmt werden. Sie lag in Epidermis und Mesophyll mit ca. 25 mM unterhalb der für andere Pflanzenspezies (Solanum tuberosum und Hordeum vulgare) publizierten Konzentration. Weiter konnte gezeigt werden, daß zwölf freie Aminosäuren mittels einer auf Kapillarelektrophorese basierenden Methode in vereinigten Zellproben von Cucurbita maxima identifiziert werden konnten. Die Übertragung der Methode auf A. thaliana-Proben muß jedoch weiter optimiert werden, da die Sensitivität selbst bei Laser induzierter Fluoreszenz-Detektion nicht ausreichte.
Im dritten und letzten Teil der Arbeit wurde eine Methode entwickelt, die die Analyse bekannter wie unbekannter Proteine in Gewebe-spezifischen Proben ermöglicht. Hierzu wurde zur Probennahme mittels mechanischer Mikrodissektion eine alternative Methode zur Laser Capture Microdissection verwendet, um aus eingebetteten Gewebeschnitten distinkte Bereiche herauszuschneiden und somit homogenes Gewebe anzureichern. Aus diesem konnten die Proteine extrahiert und über Polyacrylamidgelelektrophorese separariert werden. Banden konnten ausgeschnitten, tryptisch verdaut und massenspektrometrisch die Primärsequenz der Peptidfragmente bestimmt werden. So konnten als Hauptproteine im Mesophyll die große Untereinheit von Rubisco sowie ein Chlorophyll bindendes Protein gefunden werden.
Die in dieser Arbeit entwickelten und auf die Modellpflanze Arabidopsis thaliana angewandten Einzelzellanalysetechniken erlauben es in Zukunft, physiologische Prozesse besser sowohl räumlich als auch zeitlich aufzulösen. Dies wird zu einem detaillierteren Verständnis mannigfaltiger Vorgänge wie Zell-Zell-Kommunikation, Signalweiterleitung oder Pflanzen-Pathogen-Interaktionen führen.
The subject of this thesis was the analysis of single plant cells in respect to their contents of i) transcripts, ii) inorganic cations and anions, iii) metabolites like amino acids and carbohydrates as well as iv) proteins. One task was the transfer of existing methods to single cell analysis on leaf tissues of the model plant Arabisopsis thaliana L., the second one was the refinement and the development, respectively, of new protocols for the analysis of such picoliter samples. For cell type specific sampling two different complimentary methods were applied: Using micro glass capillaries specific single cell contents could be harvested from intact plants, whereas typical sample volumes were in the picoliter range. Even the sampling of inner cell types such as companion cells could be demonstrated. Using mechanical micro dissection of embedded tissue a larger amount of homogenous tissue could be collected.
Because single cell samples contain only femtogram amounts of mRNA, direct detection of transcripts is impossible. Therefore, two amplification protocols were applied to the cell samples: The first procedure makes use of specifically primed RT-PCR for amplification. Several genes derived from different plants and tissues could be detected after successful RT-PCR, including high as well as low expressed genes. The second method was developed to monitor the activity of many genes in parallel using array hybridisation with filters containing the cDNA of as many as 16.000 ESTs. For this purpose, unspecific RT-PCR as it is applied in the differential display was used to amplify different transcripts in just one reaction. However, in these tissue specific array hybridisations the expression patterns of several hundreds genes could be monitored. These included known tissue specific expression patterns (of mainly photosynthesis related genes) as well as a couple of unknown expression patterns. To verify the tissue specificity of gene activity some results were reconsidered using tissue specific northern blot hybridisations and real time RT-PCR, respectively.
Secondly, metabolites (including inorganic ions) were investigated: Because gas chromatography-mass spectrometry does not reveal the sensitivity which in necessary for the analysis of even multiple pooled single cell samples capillary electrophoresis was applied for these studies. This method has a high potential as it needs only small amounts of starting material, has uncomparable low detection limits and exhibits a high number of theoretical plates.
The analysis of inorganic anions and carbohydrates needs further optimisations. Using UV absorption-detection potassium could be detected in different cell types whereas the concentrations in mesophyll and epidermis were found around 25 mM each. These concentrations are lower than in other species as Solanum tuberosum or Hordeum vulgare. For investigations of amino acids the cell samples were derivatized to make the use of laser induced fluorescence-detection capable. In samples derived from pumpkin (Cucurbita maxima) mesophyll twelve amino acids could be detected and identified. The transfer of this method to A. thaliana derived samples exhibited no results which may be due to the low concentration of free amino acids in these plants.
Finally, a method was developed with which the existence of known and unknown proteins in tissue specific samples could be monitored. For this, mechanical micro dissection was used to: After embedding and sectioning the tissue of interest was cut out by an vibrating steel chisel to get homogenous samples. The proteins contained in these tissue pieces were extracted and separated by one dimensional SDS polyacrylamid gel electrophoresis. Several protein bands could be detected after staining with either silver or coomassie blue stain. These bands were cut out and sequenced by mass spectrometry. The large subunit of rubisco as well as one chlorophyll binding protein could be identified as the major proteins within the mesophyll.
The single cell analysis methods which were developed and applied to the model plant A. thaliana in this thesis allow a better spatial as well as temporal resolution of analysis. This will lead to a more detailed understanding of physiological processes like cell to cell communication, signalling or plant-pathogen interactions.

Introduction: The main aim was to investigate Dukes B cancer prognosis using array comparative genomic hybridisation (aCGH). The cancers were typed for microsatellite instability (MSI), APC loss of heterozygosity (LOH) and ploidy. A secondary aim was the selection, on the basis of aCGH results, of potential colorectal cancer genes using In Situ Hybridisation (ISH). Sporadic adenomas were also investigated with aCGH to investigate the timing of chromosomal instability in tumourigenesis. Methods: Dukes B cancers collected in Oxford, Leeds and Harrow were examined using aCGH based on a set of 3452 BAC clones at ~1Mb spacing. MSI was assessed using BAT26 and D5S346. D5S346 gave LOH at the APC gene locus. Ploidy was assessed using flow cytometry (FACS). From the aCGH data genes in regions of copy number gain were chosen for ISH analysis. Results: 79 Dukes B cancers (43 good outcome, 36 bad outcome on the basis of 5 year survival) were investigated with aCGH. The most commonly gained chromosomes across all cancers were 13, 20 and 7, and the most commonly lost were 22, 18 and 14. Comparing survival groups; chromosome 6 was more often lost in cancers associated with good outcome, chromosome 16 was more often gained in microsatellite stable cancers associated with good outcome, and chromosome 22 more often gained in microsatellite stable cancers associated with poor outcome. Chromosome 13 showed greater than single copy number change significantly more often in bad outcome cancers. Several new areas of small genomic gain and loss were detected. Four candidate genes were identified (CDX2, RHOA, FLT1 and ARHGEF1). None showed a relationship with outcome. Large scale chromosomal changes were found in 10 of14 sporadic adenomas. Conclusions: Array CGH did identify some difference between good and bad outcome Dukes B cancers. However on the basis of this data, the technique could not be used as a useful clinical tool to predict prognosis.

Micro-scale polymerase chain reaction (micro-PCR) systems offer substantial advantages over macro-scale systems. Smaller sample volumes are required, and faster process times are feasible. Thermal control of micro-PCR systems is a substantial technical challenge, however. The PCR process requires the fluid sample to be cycled through three temperature ranges — typically 90–95°C, 50–65°C and 72–77°C for denaturation, hybridisation and replication respectively. Durations of the three steps are required to be in the ratio of 4:4:9. In this paper, the thermal analysis of a continuous flow micro-PCR device is reported. The objective of the analysis is to optimize the thermal performance of the device for fast amplification cycles with high efficiency - an efficient PCR features rapid heating and cooling between steps, and good temperature uniformity within each step. The device comprises an array of parallel microchannels formed within a polypropylene substrate to carry fluid, with the base of the substrate mounted on an aluminium carrier. Substrate depth is 500 micron, and each channel is 60 micron wide by 40 micron deep. Thermoelectric cells (TECs) are bonded to the carrier, and powered by a thermoelectric controller with feedback from sensors embedded in the carrier. A Pyrex Glass slide is bonded to the substrate to form closed channels. Arrays of film heaters mounted on the slide adjacent to the channel are used to establish the required temperature regions along the channel. By pumping the fluid at a fixed flow rate, temperature cycling of specific period is achieved. Thermal analysis of the substrate is performed using an approximate closed-form solution, in conjunction with Finite Element (FE) and Computational Fluid Dynamics (CFD) simulations. The analysis is used to conduct a parametric study in order to determine the optimum configurations of substrate materials, cooling conditions, heaters and flow rates required to impose specific temperature cycles. The use of thermoelectric cells is shown to increase the rate of change of temperature between the various regions, improving the efficiency and decreasing the cycle time of the PCR process. Cycle times of 6s or less are shown to be feasible, yielding benefits in time saved for multiple amplifications. Finally, the analysis is also used to identify the dimensionless parameters which govern the thermal characteristics of the device, illustrating the importance of the Biot number.