Carbonic anhydrase IX (CA IX) protein (previously named as MN or MN/CA IX) is a cell membrane protein (Pastoreková et al, 1992; Závada et al, 1993), which is ectopically expressed in various human tumours, mostly carcinomas, for example, cervical, renal, colorectal, lung, mammary and others (Liao et al, 1994, 1997; Saarnio et al, 1998). It has been suggested that CA IX may serve as a biomarker in early stages of tumorigenesis; now it is turning out to be induced in certain tumours by hypoxia, which is connected with poor prognosis (Chia et al, 2001; Giatromanolaki et al, 2001; Koukorakis et al, 2001). In renal clear cell carcinoma (RCC), synthesis of CA IX is switched on by the loss of suppressor gene VHL (Ivanov et al, 1998). Until recently, the identification of CA IX in formaldehyde-fixed, paraffin-embedded tumour sections has been based on immunohistochemical analysis using MAb M75.
On the other hand, there were no reports on detecting CA IX in body fluids of cancer patients. For the assays of soluble antigens, which are usually present only in extremely low concentrations, two antibodies are needed, preferably monoclonal, directed to different epitopes. One of them serves for concentrating the antigen and the other for its detection and quantitation. However, all efforts to raise antibodies specific for epitopes different from the M75 epitope have failed. This represented the crucial problem, which we succeeded to solve.
Our approach was based on the following consideration: sequencing of Car9, the mouse homologue of human CA9 gene, revealed that the N-terminal, proteoglycan-like (PG) domain of the protein shows almost no homology between man and mouse. The carbonic anhydrase (CA) domain is relatively conserved in evolution. Therefore, the mouse will recognise human PG as ‘non-self’ and be able to respond to it immunologically, but it will not produce antibodies reacting with human CA domain. However, mice with the disrupted Car9 gene (knockout) should be able to produce antibodies against different antigenic sites of the CA IX protein.
Indeed, Zat'ovičová et al (2003), taking advantage of Car9−/− mice constructed by Ortová-Gut et al (2002), was able to produce a series of potent monoclonal antibodies (mAbs) specific for different antigenic sites, many of which are located in the CA domain (e.g. the V-10 antibody used in this report). The epitope of our original M75 antibody is located in the PG domain of CA IX and its amino-acid sequence is PGEEDLP (Závada et al, 2000).
Using newly obtained mAbs, we were able to develop tests for the s-CA IX antigen. To start with, we decided to use tumour cell cultures as an easy model system. The first set of questions we started to ask were: (1) Do the tumour cells containing cell-associated CA IX also shed this antigen into the medium during cultivation? (2) Is the soluble form of CA IX different from cell-associated molecules? (3) What is the concentration of CA IX in the media? (4) Do short-term cultures of human tumours also shed s-CA IX, and if they do, how much of it? (5) Can soluble CA IX be detected in the serum and urine of the patients? (6) How fast is the clearance of CA IX from the blood and from urine after surgical removal of the tumour? (7) What is the dependence of CA IX concentration in body fluids on the tumour size? (8) Do the blood and urine of healthy individiuals also contain CA IX?
In the present experiments, we chose the patients with RCC because these tumours usually express CA IX to a high level and in a high proportion of the patients.