In contrast to molecular biological techniques based on the analysis of nucleic acid extracts from tissues or cells, e.g. PCR, Northern or Southern blot hybridisation, ISH allows the detection of nucleic acids in their morphological context. In the following paragraphs; selected examples of the application of ISH to the detection of DNA and RNA will be used to illustrate the power of this technique. No attempt is made to provide a comprehensive overview of the available literature, which would be beyond the scope of this chapter.
RNA ISH permits the analysis of numerous cells present in one section for expression of the same gene while single cell RT-PCR is suited for the analysis of multiple genes in one (or a limited number) of cells (Todd and Margolin, 2002). For phenotypic analysis in a diagnostic setting, immunohistochemistry clearly is the method of choice. Antibodies suitable for staining of paraffin sections are available against a vast array of relevant antigens, and the techniques required to obtain reproducible immunohistochemical staining results are widely available. Nevertheless, RNA ISH has maintained its role as a powerful tool in various research applications.
First, the role of RNA ISH in the validation of immunohistochemical reagents must not be underestimated. We have recently used ISH to localise transcripts of the w, which are involved in antigen receptor gene rearrangements during lymphocyte development. We found expression of these genes in a small subset of tonsillar lymphocytes localised at the interface between lymphoreticular tissue and stroma and adjacent to, but not within, germinal centres (Meru et al., 2002). These results were unexpected in view of several other papers reporting conflicting immunohistochemical staining patterns using RAG-'specific' antibodies but agreed well with the expression pattern of terminal transferase, another protein expressed in developing lymphocytes (Meru et al., 2002). Thus, if unexpected or novel results are obtained with immunohistochemistry, it appears a good idea to use RNA ISH for validation (and vice versa!).
In spite of the development of a large and still growing number of highly specific reagents, immunohistology may fail to provide information about the cellular source of a protein. This is a particular problem with secreted proteins. Thus, RNA ISH has been successfully employed to study the cellular origin of different extracellular matrix components and spatial and temporal expression patterns of collagen RNA transcripts, e.g. in liver fibrosis (Milani et al., 1995). Similarly, the identification of cytokine- or chemokine-producing cells may be difficult because these polypeptides may be secreted by one cell and bound by receptors on another cell, leading to false negative or false positive results, respectively. RNA ISH has proved useful in this context. Thus, Hodgkin and Reed-Sternberg (HRS) cells were demonstrated to produce a number of cytokines, e.g. IL-6 and IL-10, using this technique (Beck et al., 2001; Herbst et al., 1996, 1997). Similarly, expression of chemokines has been analysed by RNA ISH in tumours as well as in autoimmune disease (Romagnani et al., 2002; Teruya-Feldstein et al., 1999).
The intracellular demonstration of secreted proteins may not necessarily reflect protein synthesis at this location, but may be the result of active uptake due to membrane receptors or phagocytosis, or of passive influx as a supravital or fixation artefact. Such issues can be resolved by RNA ISH because RNA is unlikely to be subject to diffusion processes from one cell to another. For example, immunoglobulins of polyclonal origin may occasionally be detected by immunohistology in normal squamous epithelial cells or in giant cells of various histogenetic origins, such as anaplastic carcinomas, glioblastomas, sarcomas, or large cell lymphomas. Such reactivities can be observed in paraffin sections stained for immunoglobulins, but not in corresponding frozen sections, strongly suggesting an influx of serum proteins, probably as the result of poor fixation. The problem is more complex in the case of cells expressing certain types of receptors for constant domains of Ig (Fc-receptors), such as macrophages or B-lymphocytes. ISH for the detection of immunoglobulin gene transcripts, however, may not only be valuable in such situations, but within limits also allows analysis of the clonal composition, isotype usage and class switching at the single cell level.
Recent progress in the cloning of the human genome has opened a new field of application of RNA ISH (Frantz et al., 2001). Available databases contain sequences of roughly 30000 mRNAs. Most of these are not characterised, their functions are unknown, and there are no antibodies available for immuno-histological detection of the corresponding proteins. Using RNA ISH, signals detected, for example in Northern blots, can be attributed to specific cell types. Moreover, the identification of spatial and temporal expression patterns of newly identified genes may provide clues as to their function. In conjunction with tissue microarrays, RNA ISH allows the rapid generation of expression data for new and potentially interesting genes (Bubendorf et al., 2001; Frantz et al., 2001). RNA ISH has also been used for rapid confirmation of data on differential gene expression, e.g. in tumour-derived vs. normal endothelial cells (St Croix et al., 2000). In these situations, preparation of gene-specific probes suitable for RNA ISH is quickly achieved while generation of specific antibodies suitable for immunohistology would require months of work with uncertain outcome.
Finally, RNA ISH can be useful for the detection of RNA transcripts not translated into proteins, e.g. U6 RNAs. This has proved a useful approach to the detection of certain infectious agents. Thus, ISH for the detection of the small EBERs has become the method of choice for the detection of latent EBV infection (see below). Moreover, detection of ribosomal RNA species by ISH can be used for the identification of other microorganisms, such as mycobacteria and fungi (Boye et al., 2001; Hayden et al., 2002).
ISH has been used for the detection of specific DNA sequences in many different settings (for a review, see Niedobitek and Herbst, 1991). In recent years, two main areas of application of DNA ISH have emerged.
One major field is the detection of infectious agents. DNA ISH has been successfully employed for the demonstration of several bacteria, e.g. Helicobacter pylori, Chlamydia trachomatis, Haemophilus influenzae and Mycoplasma pneumoniae (Horn et al., 1988; Saglie et al., 1988; Terpstra et al., 1987; van den Berg et al., 1989). More recent developments in this field include the use of DNA probes for the detection of ribosomal RNA sequences specific for bacterial or fungal organisms (Boye et al., 2001; Hayden et al., 2002). Furthermore, DNA ISH has been used to study acute viral infections, e.g. in immunocom-promised individuals, and to investigate possible associations of DNA tumour viruses with human malignancies. A major application remains the detection of HPV in anogenital neoplasia. In addition to identifying virus types associated with low or high risk of progression, it has been demonstrated that the pattern produced by HPV DNA ISH using non-radioactive probes may be of relevance. Thus, a diffuse labelling of nuclei has been reported to indicate the presence of episomal viral DNA, while a punctate pattern is suggestive of virus integration into the host genome (Cooper et al., 1991). This may be of relevance, since it has been reported that a punctate pattern may indicate a poor prognosis (Gomez Aguado et al., 1996).
DNA ISH has also been used for the detection of EBV DNA in latent and lytic infection (Herrmann et al., in press; Niedobitek et al., 1989c). For the demonstration of latent EBV infection, this method has largely been replaced by ISH for the detection of the EBERs. These are expressed at very high copy numbers in all established forms of latent EBV infection (Niedobitek et al., 2001).
Therefore, they represent an ideal target for ISH using radiolabeled or nonradioactive probes (Niedobitek et al., 2001; Wu et al., 1990). Recently, the possibility of the existence of an EBER-negative form of EBV latency in hepatocellular and breast carcinomas has been raised (Bonnet et al., 1999; Sugawara et al., 1999, 2000). This, however, has not been confirmed by others (Chu et al., 2001a, 2001b; Herrmann and Niedobitek, 2003; Junying et al., in press). Thus, at present it would appear that EBER-specific ISH is a reliable method for the detection of latent EBV infection.
The second major field of application of DNA ISH with implications for diagnostic histopathology is interphase cytogenetics (Wolfe and Herrington, 1997). This term refers to the detection of chromosomal abnormalities in interphase nuclei and, in contrast to classical cytogenetics, does not require viable cells for the generation of metaphase spreads. Thus, interphase cytogenetics is applicable to conventional cytological or histological preparations, including paraffin sections (Wolfe and Herrington, 1997). Several groups have used interphase cytogenetics for the detection of chromosomal abnormalities in haematological tumours. Mantle cell lymphomas are characterised by t(11;14) juxtaposing the CCND1 gene to the immunoglobulin heavy chain locus. This results in overexpression of the cyclin D1 protein. Immunohistochemical detection of this protein, however, is temperamental. By contrast, interphase cytogenetics by ISH using fluorochrome-labelled probes (FISH) has demonstrated the t(11;14) in over 95% of cases (Belaud-Rotureau et al., 2002; Remstein et al., 2000). Amplification of the HER2 gene occurs in a significant proportion of breast carcinomas and is associated with poor prognosis. Moreover, information on HER2 amplification and overexpression has become relevant in view of immunotherapeutic approaches to breast cancer treatment using a humanised monoclonal antibody directed against the HER2 protein. Conventional assays for the immunohistochemical detection of HER2 protein expression are convenient but subject to technical and interpretational variability. Reliable assessment of HER2 gene amplification by ISH using fluorochrome (FISH) or colorimetric detection (CISH) is currently employed to identify breast cancer patients that may benefit from a therapy with HER2-specific antibodies (Figure 2.2) (Dandachi et al., 2002; Zhao et al., 2002). Similarly, amplification or deletion of the TOP2A gene coding for topoisomerase IIa may be visualised by FISH (Jarvinen et al., 2000).
Both DNA and RNA ISH can be combined with the immunohistological detection of proteins. Moreover, in RNA ISH, radioactive and non-radioactive probes can be combined in the same assay. Such techniques have been used in various settings. In virus research, double-labelling experiments have been used for the simultaneous demonstration of viral nucleic acids and virus-encoded proteins (Niedobitek et al., 1997b). Concurrent detection of virus DNA or RNA
Figure 2.2 Fluorescence ISH demonstrates amplification of the HER2 gene in a breast carcinoma (red signal). A colour version of this figure appears in the colour plate section transcripts and proteins associated with viral latency or replication may allow an assessment of the state of the virus in an infected cell. The simultaneous demonstration of viral DNA and transforming viral proteins may allow an assessment of the potential significance of a virus for the pathogenesis of malignant tumour. Combining ISH with immunohistochemical detection of lineage specific antigens, e.g. intermediate filaments or leukocyte differentiation antigens, may be used for determining the phenotype of virus-infected cells. This approach has proved useful for identifying B-cells as a major site of EBV persistence (Anag-nostopoulos et al., 1995; Karajannis et al., 1997; Niedobitek et al., 1992,1997a).
Was this article helpful?