Frontiers in Medicinal Chemistry covers developments in rational drug design, bioorganic chemistry, high-throughput screening, combinatorial chemistry, compound diversity measurements, drug absorption, drug distribution, metabolism, new and emerging drug targets, natural products, pharmacogenomics, chemoinformatics, and structure-activity relationships. This E-book series is essential for any medicinal chemist who wishes to be updated on the latest and the most important advances in the field.
Angiogenesis regulation strictly depends on the balance between pro- and anti-angiogenic molecules such as growth factors. HER3 is a receptor for the paracrine growth factor neregulin-1, and a transmembrane protein that tethers the ligand to its dimerization partners, the receptor tyrosine kinases HER2 and HER4 [ 48 ], and known to play important roles in the development and progression of the malignant phenotype in breast cancer [ 49 ]. The abnormal expression and activity of HER2 has been studied extensively in the context of prostate cancer [ 50 ], being found overexpressed in advanced tumors, either metastatic or homone-independent, but infrequently in primary, organ-confined tumors.
More controversial is the information available on the role of HER3, with reports of its overexpression in prostate cancer together with HER2, HER4, or both [ 51 , 52 ], but also of its overexpression only in metastatic tumors, in particular of a truncated form corresponding to the extracellular domains of HER3 [ 53 ].
Furthermore, several transcriptional profiling analyses have found overexpression of this gene in prostate cancer. IQGAP2 is a calmodulin-binding protein that participates in cell signalling and modulation of cytoskeletal dynamics [ 37 ], and its activity has been reported to be positively [ 54 ] and negatively associated with neoplastic phenotype. Levels of desmin transcripts were determined as an index of the contribution of stromal cells, suggesting that the overexpression of the analyzed genes are detected in tumor samples even in the presence of substantial stromal contamination Figure 5.
Forget Password Confirmation
Of particular interest is the observed upregulation of HER3 in prostate tumor tissues relative to normal tissues. Recent experimental evidence further highlights the importance of HER3 in conferring a malignant phenotype and a hormone-refractory state to prostate epithelial cells [ 61 ]. Thus, whenever HER3 is expressed it is reasonable to expect co-expression of at least one other member of the HER family.
Therefore, we determined by real-time RT-PCR the relative expression in our prostate tissue samples of the genes for all four members of the HER family of receptor tyrosine kinases. Our results show that HER4 is expressed at increased levels in 10 of 14 prostate tumor samples Fig. The expression values for each gene, previously normalized with respect to the S14r expression level in each sample, are shown as ratios of the normalized values in prostate cancer vs.
Quantitation of desmin expression levels was used to assess the degree of contribution of stromal components in the samples analyzed.
Discovery of a Platelet Derived Growth Factor Peptide-based Mimetic
Values equal to or above fold are shown as B Heatmap representation of the same data color scale as shown below. C Real-time PCR analysis for HER3 transcript levels of laser microdissected tumor and normal samples, compared with relative transcript levels in enriched non-microdissected tissues from the same cases. D Immunohistochemical analysis of HER3 on paraffin-embedded prostate tissue sections arranged in tissue microarrays see Methods. As mentioned in the Methods section, both tumor and normal tissues were carefully chosen to have similar representation of epithelial compartment.
However, to further ensure that the observed expression of HER3 was not due to a dilution effect of normal epithelial cells by stroma, we performed real-time PCR analysis of laser microdissected samples. For this, we selected four samples that had shown overexpression of HER3 in the enriched tumor samples described above, and two that had levels that did not differ significantly from non-tumor containing normal matched tissues.
Of the four samples in which the enriched tumor tissue had shown increased levels of HER3 transcript, three microdissected samples overexpressed HER3 Fig. In two of the microdissected samples, HER3 transcript levels were equal in normal and tumor microdissected epithelia, and this also corresponded to samples in which HER3 levels did not differ significantly between enriched tumor and normal prostate tissues Fig. This analysis showed that overexpression of HER3 in prostate tumor tissues is not due to simple enrichment of epithelial cells in comparison with non-tumor tissues.
To further confirm the cell type expressing HER3 in prostate tissues, immunohistochemical analysis with a monoclonal antibody to HER3 was performed on 16 prostate samples, arranged in duplicate 1-mm diameter cores in tissue microarrays, in which both tumor and normal glands were present. HER3 protein was found clearly overexpressed in tumor epithelia in 13 of the 16 cases In all cases, normal epithelia showed weak reactivities for HER3 Fig. In summary, our transcriptome re-analysis, validated by real-time RT-PCR of non-microdissected and microdissected samples and by immunohistochemical analysis, significantly reinforces previous immunohistochemical studies that reported high levels of expression of HER3 and HER4 in primary prostate cancer [ 51 , 52 ].
We have shown that the method presented here for the analysis of expression microarray data permits the classification of samples into meaningful categories and, simultaneously, to identify a subset of genes and their assignment to pathways most significantly contributing to the corresponding phenotypes, while allowing for a given gene to participate as significant in more than one cluster of samples.
The analysis of the yeast dataset validates the approach. Our results are consistent with biochemical pathways known to be activated in the different stress conditions analyzed, and the clustering of samples reflects the underlying similarity of the biochemical responses. In the application to the prostate cancer dataset, we have found that two pathways, one modulated by androgen receptor and a second one by signals that originate from cell surface growth factor receptors, are prominently active in the organ-confined, non-metastatic prostate cancer samples analyzed.
The latter pathway has been reported to be spuriously active in at least a subset of prostate tumors that have progressed to invasive and hormone-independent states [ 62 ]. Our results suggest that such altered activation may already be present in primary tumors. Although a prevailing model for prostate tumor progression is that acquisition of the capacity for metastatic and hormone independent growth proceeds through selection of rare populations of cells concealed among primary tumor cells, there is also evidence that a transcriptional program for metastasis may already be present in the bulk of primary tumors at the time of diagnosis [ 63 , 64 ].
Our analysis would be more consistent with the latter model. Finally, we have unveiled and validated several markers highlighted by the analysis of the prostate cancer dataset. Overexpression of HER2 and consequent increased signalling have been associated with advanced prostate cancer, development of hormone independent state and poor prognosis [ 65 , 66 ], but is infrequently observed in primary tumors [ 67 , 68 ]. On the other hand, our results suggest that, in primary prostate cancer, HER3, together or not with HER4, rather than receptor complexes involving HER2, could play important roles in the biology of these tumors.
It originally consists of spotted array measurements of genes in experimental conditions that include temperature shocks, hyper and hypoosmotic shocks, exposure to various agents such as peroxide, menadione, diamide, dithiothreitol, amino acid starvation, nitrogen source depletion and progression into stationary phase. Log-ratios were preprocessed following several steps: first data from genes with missing values were filtered out, and their missing values estimated with LSimpute [ 69 ] using the 'Adaptive' method.
Next, ratios were computed from the log-ratios and quantile-normalized experiment-wise using the normalizeQuantile function from the R package [ 70 ], so that all experiments had the same average sample distribution. Finally, ratios were log transformed again. The prostate cancer dataset chosen is described in [ 14 ].
It was originally obtained by hybridizations on Affymetrix U95A oligonuleotide arrays with probes for a total of 55 samples. Intensity values were preprocessed following several steps: first intensity data were thresholded, with intensities below 10 fixed at 10 and values above fixed at The thresholded values were log-transformed and then centered by the median of all experiments. Finally, genes were subjected to z-transformation per gene basis.
Q-mode Factor Analysis FA [ 9 ] seeks to find an underlying orthogonal factor model of an original X -matrix nxm where n are the number of samples and m the number of mRNA levels measured of the form:. L is the loadings matrix of size nxk , where k is the number of factors, and F the scores matrix of size kxm , while E is the residual matrix , which contains both the specific variance of the individual genes and the errors in the model see Figure 1.
We used the so-called principal factor solution to solve this factor model. Specifically, in a first step, and based on the correlation matrix R derived from X , communalities i. These communalities replaced the diagonal entries of the correlation matrix, which was subjected to diagonalization.
- Peptide Growth Factors!
- The Key: A True Encounter.
- Comparative Health Policy in the Asia Pacific (State of Health).
- Breakthroughs in statistics. - Methodology and distribution.
- Peptides Targeting Angiogenesis Related Growth Factor Receptors.
- Growth factor - Wikipedia.
- Add to Cart!
New communalities were computed from the loadings at the chosen dimensionality, obtained by scaling the eigenvector matrix P , as follows:. The new communalities again replaced the diagonal entries, and the process was iterated until convergence. Finally, we proceeded to rotate the factor loadings by means of a varimax rotation [ 9 ]. The effect of this rotation is to maximally align each of the samples with one factor in order to simplify the factor model and make it more readily interpretable.
Phyletic trees were derived by clustering samples in loadings space at the optimal dimensionality using average linkage [ 4 , 11 ]. Once sample clusters are defined, these are used to identify groups of genes contributing heavily to the specific character of different groups. Each gene on the list is subjected to a Student's t -test that measures the differential expression of the gene in the cluster as compared with the rest of the samples. The q -value is similar to the well known P -value, except that it is a measure of significance in terms of the false discovery rate, rather than the false positive rate [ 73 ].
The association between selected genes and biological functions was established by determining the hypergeometric distribution of genes on the annotation databases GO [ 12 ] or GenMapp [ 32 ]. With this distribution we computed the probability that at least x genes annotated within a given biological function according to GO or GenMapp in a cluster of size n the total number of genes per cluster selected in the previous step can be obtained by chance, given a population of N genes under consideration and given A , the total number of genes within N with that particular annotation.
These P -values are obtained according to:. An aggregated score for each cluster from the significant P -values i. The significance of this score is established by simulation.
We randomly selected samples of size n genes each the number of genes per cluster selected according to the q -value and computed a new s -score s r for each one. The Z -score is finally computed as:. We should emphasize that in spite of the apparent intricacy of the computational procedure, the computational complexity is similar to other biclustering methods, and operates within a highly constrained parameter space: in the factor analysis part of the program only the percentage of variance employed should be set, yielding a reduced number of dimensions or latent variables, usually below 5; the number of clusters is automatically determined in this space from the c -index, and has no free parameters, and the selection of genes relevant for each cluster only depends on the cutoff employed in the q -value.
In each instance, the tumor sample and its matching normal counterpart were obtained from the same case, upon removal by radical prostatectomy. In addition, samples were chosen such that the tumor and normal counterparts in each case had approximately equal representations of the epithelial compartment, as assessed microscopically. For each sample, 0. TaqMan probes and their corresponding primer sets were obtained from Applied Biosystems. All determinations were performed in triplicate and in at least two independent experiments. Levels of ribosomal S14r amplification were used as an endogenous reference to normalize each sample value of Ct threshold cycle and normal tissues were used as calibrators for their tumoral counterparts in each case.
The final results, expressed as n -fold differences in target gene expression were calculated as follows:. Sixteen paraffin embedded prostate samples were evaluated for HER3 expression by immunohistochemistry on a tissue microarray. The cases were represented in duplicated 1-mm diameter cores and always included normal prostatic glands adjacent to neoplastic foci in at least one of the cores. Reactions were detected after development with diaminobencidine and H 2 O 2 for 3 min.
Slides were counterstained with Harri's hematoxilin, dehydrated and mounted. As a negative control, the primary antibody was substituted for isotype-matched mouse IgG.
Discovery of a Platelet Derived Growth Factor Peptide-based Mimetic | dinlungdalditi.ga
Remote access to the program has been enabled by setting up a web-server where the program can be executed . Genome Res. Nat Biotechnol. Hartigan JA: Clustering algorithms. Nucl Acids Res. Mol Biol Cell. Cancer Res. J Biol Chem. Promotes apoptosis in chondrocytes, but can also promote cancer cell proliferation.