It depends a little bit about which tumour your are interested in as some may be more diffuse and infiltrative into adjacent normal tissue than others. However you could do this by laser capture microdissection. You would need a pathologist to determine areas with tumour and normal tissue on your FFPE slides prior to laser capture.

A simple H&E staining would clearly show the difference between normal and tumor tissue, the homology of the tissue would be disturbed in the tumor tissue with infiltration of immune cells. I have attached the protocol for H&E staining (its a Dako protocol)

I don't have enough pratical experience, but theoretically I think that in order to distinguish between normal tissue and tumor tissue of a single FFPE block, a pathologist needs to carry RNA extraction, amplification, hybridization followed by gene expression profiling. Although there is extreme cross-linking and degradation of RNA in FFPE blocks, yet use of commercial RNA extraction and amplification kits are available to generate functional gene array data in clinical practice. We need to extract from several sites in the tissue block available (of tumor tissue) when searching for normal within it. RNA recovered from tumor biopsies will be of poor quality as compared to normal tissue. But such severely degraded RNA can be cultured and FFPE microarray-based DASL assay can be applied to analyze mRNAs in total RNA. The changes in gene expression resulting from degradation of RNA can be detected; the tissue not having those changes in gene expression will be normal... Fran et al. 2007; Ravo et al., 2008; Abdueva et al. 2010

I would suggest, in addition to comments above, to use independent control DNA samples.

For DNA study, you can use normal DNA as the control from healthy donors (e.g. isolated from blood). I would suggest to involve more independent samples. As there may exist only minor differences in DNA from different organs or tissues it may work well. There also exist some control DNA samples available commercialy.

If you want, you may certainly isolate DNA from normal tissue, either in FFPE block or isolated from fresh material, freshly frozen (liquid nitrogen) tissue or stabilized material (AllProtect Tissue Regent, RNALater). But, according to you, you have no such samples.

Although you may compare tumor tissue and adjacent normal tissue, you may expect that most of stromal cells may be „normal“ but some may come from cancer cells. See:

The Origin of Stroma Surrounding Epithelial Ovarian Cancer Cells

Author(s): Akahane, Tomoko; Hirasawa, Akira; Tsuda, Hiroshi; et al.

Source: INTERNATIONAL JOURNAL OF GYNECOLOGICAL PATHOLOGY Volume: 32 Issue: 1 Pages: 26-30 DOI: 10.1097/PGP.0b013e3182518533 Published: JAN 2013

Furthermore, depending on issues investigated, in case of hereditary mutations, the „normal tissue“ of cancer patient would bear such mutations, but as expected the tumor tissue would possess many other mutations. For general conclusions comparing normal and cancer tissue (healthy vs cancer patients) it is very risky to make such conclusions based only on the adjacent normal tissue from tumor samples in FFPE blocks.

Interested in RNA research, it should be strongly recommended to use normal tissue from healthy donors (if applicable) or tissue not affected by the investigated disease, as the control. Ideally, pooled control sample from the same number of control patients as the cancer patients should be prepared, but it depends on complexity of research you will conduct.

Depending on the tumor type you use, you may select surface markers and flow sort them. see the follwoing article for this:

Curr Protoc Cytom. 2011 Jan;Chapter 7:Unit 7.37. doi: 10.1002/0471142956.cy0737s55.

High-resolution multiparameter DNA flow cytometry for the detection and sorting of tumor and stromal subpopulations from paraffin-embedded tissues.

Corver WE, ter Haar NT.

Karoly

You need a pathologist or a person who can distinguish this and then a laser microdissector.

As you are in Graz I would recommmend to ask Helut Popper.

Without a pathologist your findings could be not reliable