|[Frontiers in Bioscience 1, c4-15, November 1, 1996]|
THE FOUNDATION OF SUCCESSFUL RT IN SITU PCR|
Director, MGN Medical Research Laboratories, Setauket, New York 11733, USA
Received: 07/25/96; Accepted: 10/02/96; On-line 11/01/96
Assuming adequate protease and DNase digestion, target specific direct incorporation of the tagged nucleotide is routinely achieved with RT in situ PCR. To illustrate some of the key points for successful RT in situ PCR already discussed, the expression of matrix metalloprotease (MMP) and their inhibitors (tissue inhibitor of matrix metalloprotease -TIMP) in cancers will be used as a model system.
Studies of tumor invasion and metastases have focused on the degradation of the extracellular matrix (ECM) and the endothelial cell basement membrane. Enzymes that have been implicated in the degradation of these compartments include the MMP family. Two of these enzymes, designated MMP-2 and MMP-9, are potent gelatinases and their activity, in conjunction with their inhibitors, TIMP-1 and -2, have been correlated with the processes of tumor cell invasion and metastasis (24,25). The importance of the balance in production of MMPs and TIMPs in tumor cell invasion and metastasis has been suggested by several studies. The inhibition of tumor cell invasion and metastasis in animal models has been demonstrated using in vivo injections of TIMP (24,25). The inactivation of TIMP by transfection of mouse 3T3 cells with antisense DNA converted the phenotype of these cells from noninvasive to tumorigenic and metastatic in nude mice (24,25).
MMP and TIMP expression in cervical cancers serves as a good model to illustrate the major points discussed in the first section of this manuscript. The foundation for determining if the signal seen with RT in situ PCR is specific is based in the positive and negative controls, which should be done on the same slide, if possible, with the test. The intense, nuclear signal in all cell types with the positive control (no DNase) and the loss of this signal with DNase digestion is the primary way to determine if the key variables, especially protease time, have been optimized (Figure 5). Unless the variables are optimized, one will not get a target specific signal with RT in situ PCR. However, there are two other important indicators of specificity with RT in situ PCR that require some expertise in histologic analysis to appreciate: the restriction of the signal to certain cell types and the subcellular localization of the signal.
We used the RT in situ PCR technique to correlate the presence MMP-9 and MMP-2 and of TIMP-1 and TIMP-2 mRNAs with prognosis in 23 cases of cervical carcinoma (26). It is important to stress that most of these tissues showed, in addition to foci of invasive cancer, areas of noninvasive carcinoma in situ and normal epithelium. PCR-amplified MMP and TIMP cDNAs were restricted to the invasive cancer cells and the surrounding stromal cells. Thus, the adjacent carcinoma in situ and normal epithelial areas served as important "in-built" negative controls (26). One should be skeptical when obtaining a signal in many different cell types when doing RT in situ PCR. Also, as would be expected for human mRNAs (see Figure 5), the signal localizes to the cytoplasm, whereas the genomic based signal with the positive control is seen in the nucleus. A nuclear based signal after RT in situ PCR most likely is nonspecific and reflects inadequate protease digestion (1,26).
The ratios of the percentage of cancer and stromal cells expressing MMP-9 and MMP-2 to those expressing TIMP-1 and TIMP-2 were approximently one in those cancers with a good prognosis. This MMP to TIMP ratio in the cancer and stromal cells with a poor prognosis was significantly increased to 5.4 and 3.4 (p<0.0001), respectively, reflecting a marked reduction in the percentage of cells expressing TIMP in cancers with a poor prognosis. Expression of human papillomavirus open reading frames E6 and E7, which are important in cell transformation and immortalization (1) were equivalent in cervical cancers of good and poor prognosis (26).
We then studied the in vitro behavior of the Caski and HeLa cells in more detail. The matrigel system was used to study the invasiveness of these cancer cells. We used the RT in situ PCR technique to directly compare the expression of cells invading the matrigel with the non-invasive cells. There was a significant increase in the percentage of HeLa cells invading the matrigel expressing MMP-9 (33%) and MMP-2 (48%) when compared to the non-invasive cells (11% and 12%, respectively); there was no change in the percentage of cells expressing either TIMP, HPV E6 or E7 RNA. These simple experiments illustrate an important advantage of RT in situ PCR as compared to solution phase PCR; the simple and rapid quantification of the percentage of a given cell type expressing a transcript of interest.
These data derived from RT in situ PCR suggest that the balance of MMP-9 and -2 to TIMP-1 and -2 expression, but not HPV expression, is an essential factor in determining the invasiveness of cervical cancer. More specifically, the data suggest a multi-step process in the evolution of cervical cancer. Early invasion in cervical cancer probably requires activation of MMP-9 and MMP-2 expression. This is counterbalanced by an equivalent expression of TIMP-1 and TIMP-2. The controlled production of these two MMPs and two TIMPs may explain, in part, why such superficial cervical cancers, called microinvasive carcinomas, have a metastatic rate of less than 1% and the women have a survival rate of near 100%. More extensive invasion in clinical samples is strongly associated with a decreased in the percentage of cells expressing TIMP. The next step towards increased invasiveness requires the ability to enter microvessels. The data from the clinical samples, including direct detection of MMP and TIMP expression in tumors cells that had invaded the microvasculature, strongly suggests that, for this to occur, there must be an increased percentage of cells expressing MMP-9 and MMP-2 with a marked reduction in the expression of TIMP-1 and TIMP-2.