Project Description with Specific Aims
Differential epithelial cell-type specific CD expression provides a powerful means of identifying cell types related by lineage and, thus, of distinguishing progenitor cells from intermediate cells and terminally differentiated epithelial cells. We have identified a number of such CD markers that are specific for the parenchymal cells of the prostate, including the fibromuscular stromal cells and the basal and luminal epithelial cell types of the prostate. These markers greatly augment cytokeratins for use in cell-type analysis and increase our ability to identify a greater number of cell types as exemplified by the differentiation of luminal and cancer cells (both of which express luminal-type cytokeratins). The overall goal of this project is to characterize and isolate epithelial stem cell populations from two urologic organs, the prostate and bladder. To realize this goal we propose the following aims.
Aim 1. To CD phenotype human prostate and bladder at high resolution with the use of confocal microscopy.
We will finish CD phenotyping of the bladder by immunohistochemistry to generate a CD chart for the constituent cell populations similar to the chart generated for the prostate (http://scgap.systemsbiology.net/figures/CD_specificity.php). This will allow us to compare the CD patterns of bladder and prostate basal cells. It will also allow us to select appropriate CD antibodies for sorting specific bladder cell populations for gene profiling, as was done for the prostate cell populations. We will use confocal microscopy and multiple CD antibody-dye conjugates to define at higher resolution possible epithelial cell types within the epithelium and stroma of prostate and bladder, respectively. We postulate that there are intermediate cell types that express both basal and luminal CD markers. It is also possible that subpopulations within the CD44+ basal cells could be identified by their unique CD phenotypes, and one of these could represent the stem cell population.
Aim 2. To CD phenotype the mouse prostate and bladder.
We will use the same protocol to characterize the corresponding mouse organs by CD antibodies. Many but not all of the human prostate cell-type specific CD antibodies have a mouse equivalent for which antibodies are available from the same commercial source. For this work, we have established collaboration with Dr. Robert Vessella, who will provide excess mouse prostate and bladder tissue for immunohistochemistry, and Dr. Peter Nelson at the Fred Hutchinson Cancer Research Center, who has done gene profiling of the mouse prostate (http://www.mpedb.org/). Although there are anatomical differences between the mouse and human prostate, there are no significant anatomic differences between the mouse and human bladders. The data will be presented as CD charts for the constituent cell populations of the mouse prostate and bladder. These results will allow us to determine how similar are the cell-type compositions of the two organs between human and mouse, especially if an epithelial stem cell CD pattern could be identified, and whether the respective stromal patterns are similar.
Our characterization of CD expression in bladder and prostate will look for differences based on the location within the respective organs. Specifically, in the prostate we will look for differences in CD profile based on (a) the zone of the human prostate (peripheral zone, transition zone, central zone) and the lobe of the mouse prostate (anterior, dorsolateral, and ventral), and (b) differences due to the proximal vs. distal location. CD phenotyping of the bladder will assess heterogeneity in six topographic regions of the bladder - three layers of urothelium (umbrella cells, intermediate cells, basal cells) and three layers of stromal cells [superficial lamina propria (the region of CD13 expression), deep lamina propria, and submucosa].
Aim 3. To profile more samples of basal and of stromal cells using uncultured cells.
Most CD-defined cell samples will come from MACS. These will be supplemented by several laser-capture microdissected (LCM) samples of each type to provide a cross check on the cell-specific transcriptome. The rationale of analyzing LCM samples is to ensure that “contamination” of MACS-sorted cell samples (typically <10% of cells in a MACS sort represent contaminants) does not affect the expression profile, specifically, the rank order of expressed genes to any degree. Since LCM samples would have to be amplified to obtain enough RNA for arrays, we will compare the expression profiles of several LCM samples with the expression profiles of RNA from MACS-sorted cells that is amplified. Since one consideration in using LCM cells is that intra-epithelial lymphocytes may unwittingly be collected as a minor contaminant, we might also analyze the profiles of LCM samples from immunofluorescent immunostained sections. By using sections immunostained with fluorochrome-conjugated CD45, which is lymphocyte-specific, we will be able to avoid capturing lymphocytes.
Aim 4. To confirm cell-type specific expression of genes that we identified by array analysis.
We will use a high throughput format to examine gene expression in tissues by in situ hybridization (with in vitro generated anti-sense and sense RNA probes). The test genes include genes that we observed to be differentially expressed between the stromal cells of the prostate and bladder, and those expressed in relatively high abundance by CD44+ basal/stem cells (excluding those whose basal cell expression is known). The use of an automated machine (Ventana) will allow us to study at least 10 different probe sets with multiple tissue section slides in one experiment. In this approach, probes are produced directly from PCR products without the need of cloning of gene sequences in bacteria.
As with the immunostained sections, we will assess both the magnitude and uniformity of expression of specific transcripts. We will assess the uniformity of expression of mRNA (by ISH) within a given region (zones of man and lobes of mouse) and in different zones/lobes of the prostate. A semiquantitative assessment of gene expression level will be made for each hybridization. Visual assessment of levels of signal intensity for specific RNA’s will be validated by performing quantitative PCR on laser captured cells.
Aim 5. Database.
The database consists of multiple components. At present, the components include (1) data from CD immunohistochemistry of human bladder and prostate, and (2) array data from the same sources. A third component will consist of in-situ hybridization data. The components will be expanded as data accumulates to include mouse bladder and prostate. All components will contain our protocols, in addition to data from the respective technologies, so that other labs can replicate our studies.
The immunohistochemistry component will contain images of CD immunohistochemically stained tissue grouped into the following categories: (a) human prostate, (b) human bladder, (c) mouse prostate, and (d) mouse bladder. We have developed a standardized way of naming the images [antibody, case number, block number, slide number, image number, parent image number (the lower magnification image that encompasses the image), optical magnification, type of tissue].
A second component will contain array analysis results (obtained after the appropriate number of replicates). The raw data can be obtained through links to our in-house database. Where available, all ancillary data - RT-PCR validation - will be included. The CD immunostain data and the array data for each sample and each category of sample will be integrated.
The third component will include in-situ hybridization data, with images and relevant methodology protocols.
All images will be annotated. The annotation will describe the tissue, distribution of reaction product in the tissue, localization patterns within histologic cell types, and an assessment of the level of expression of the protein (for the immunostaining data) or of the mRNA (for in-situ hybridization data). The images will be accessed by the CD number or gene name, respectively.