Introduction/Goals

    This atlas contains microscopic images of intestinal tissues stained by multichannel immunofluorescence or traditional H&E. The images are stored at the Open Microscope Environment (OMERO) at Harvard Medical School, which provides a secure server and the appropriate imaging tools to make these images immediately accessible via your web browser.

   The tissues in this atlas are analyzed on tissue microarray (TMA) slides. Each TMA is composed of about 50 tissue samples, including samples that serve as internal controls. After each slide is stained and mounted, it is scanned to create a very high resolution image of the result, sufficient to examine cellular architecture in biopsies several mm wide.  For orientation, E-cadherin and Hoescht stains are used to highlight epithelial cells and nuclei, respectively, in most ways.  Anti-bodies are selected with emphasis is on intestinal transport and barrier-related proteins.

The atlas contains human biopsies taken from the duodenum, jejunum, ilium, proximal colon and distal colon. Images are being added continuously and include increasing numbers of biopsies. Corresponding mouse tissues are in the process of being added.  

First look at the Atlas


Each image in the Atlas is described by many sortable data fields.  The table display shows many of the fields (Fig. 1).

Figure 1. The spreadsheet structure of the atlas.

 

Some of the fields in the spreadsheet might not be self-explanatory:

  • The eye icon brings up a pop-up box with additional data on each image (Fig. 2).
  • OmeroID refers to the specific identification of the image given by the server. 
  • Detail Diagnosis, when available, rates the severity or other details regarding the subject’s or animal’s condition.
  • Pat ID # is a unique identification value assigned to each subject.
  • The slide icon opens an additional web page with specific information about the stain (Fig. 3).

Fig 2. Data accessible by pressing the eye icon – only data in the database is shown in this window, i.e. empty fields are not listed.

 

By pressing the eye icon, a pop-up box (Fig. 2) opens with an overview of the image information includes the following fields:

  • Antigen is the antibody-stained protein of interest.  Other antigens are also stained for orientation.
  • Ch1, Ch2, Ch3, and Ch4 indicate which primary antibody is labeled by four non-overlapping visual spectrums, each of which is a channel.  Two stains common to every section identify the cell morphologies: Hoescht stain to label the nuclei (usually in channel 1) and E-Cadherin to label the epithelial lateral membranes (usually in channel 2).
  • Core# ID identifies the specific tissue core. Each core stained is assigned a unique ID.  One subject may have more than one core in the various Atlas TMAs.

Fig 3. Antibody stain data retrievable by pressing the slide icon.

The stain data page provides a brief description of the antigen being stained followed by details of the specific primary and secondary antibodies used on the tissue section. 

Navigating the Atlas (Filter Data)

In order to search through the specimens press the “Filter Data” button to bring up a pop-up window with each of the searchable field.  In the filter data screen, the dataset can be sorted to only show a wide variety of useful combinations.  To look at CFTR stains in normal versus diseased tissue in duodenum of older subjects, for example, four fields would be toggled (Fig 4).  Presently, this limits ~4000 images down to a more manageable number of ~50 images. 

Fig 4. Filter window with selected parameters.

Each subject whose biopsies were used was assigned a unique identifier, making it possible to identify multiple biopsies from one individual. Filtering with the Pt ID #  field can thus bring up all the data from a specific subject.  For example, selecting only Subject 9 for the search brings up distal colon and proximal colon cores from a healthy 6-month-old girl (Fig. 5).  It is then possible to identify the two different tissue cores by the gross morphology seen the thumbnail image.  A search showing the stained biopsies from Subject 55, a healthy 2-year-old girl, provides stained tissue from the distal colon, proximal colon and jejunum.  Often, a diseased subject will have only one tissue core, taken from the tissue of interest with pathological hallmarks of the disease in question.  Take note that the TMA Name is used for internal analysis and should not be used for searching biopsy sites or the subject’s diagnosis, both of which have their own fields listed above. 

Fig 5. Filtered data set showing only stained tissue from Subject 9.

The biopsies were stained in groups of TMAs, each named according to the primary site of tissues within the TMA.  In the filter window, they can be selected using the TMA Name field.  TMAs named with the prefix “Ms” are composed of mouse tissue.  TMAs that are composed of tissues, mostly from one region, contain some tissues from other sites to serve as internal staining controls.  By looking at any particular stain from the same TMA, it is possible to compare expression of antigens in tissues that were processed in parallel.

Observing an Atlas Image

Clicking on the preview (thumbnail) image associated with a particular image will launch that image in OMERO PathViewer in a new window (Fig. 6, top panel).  The image is easily navigated using the computer mouse.  A left click will zoom into the clicked on area.  Scrolling the mouse wheel will zoom in and out.  Pressing down the left button and moving the mouse will drag the image.  The small image in the lower right screen maps the current view within the larger image.

         

   

Fig 6. Navigating the PathViewer to find the imaging options.

The Viewing Options tab, accessible from the Browse window (Fig. 6, lower row), shows the stain used for each channel and allows modification of viewing properties, including color assignment and scaling (brightness and contrast). Scaling has been applied in the same way to each channel on all tissue slices in the same TMA.  It may, however, be helpful to optimize scaling for individual images. This will not change pixel values, but will only affect the displayed values. The color used to display each channel can also be modified.  One may also view the far red channel, which, in some cases, has an additional stain that can be useful in tissue orientation, e.g. ß-actin.  By reassigning stain colors and scaling each color, one can bring all four channels into view at once (Fig. 7).

Fig 7. Adjusting the viewing parameters can provide an informative and colorful result.

Contact Information


If you have suggestions or other comments, please feel free to contact Jerrold R. Turner.  

This work is supported in part by NIDDK R24099803.

Jerrold R. Turner, MD, PhD
Professor of Pathology and Medicine

Brigham and
Women's
Hospital

Harvard
Medical
School

NRB 730
77 Avenue Louis Pasteur
Boston, MA 02115