Thus, even in xenografts that appear to be pure carcinoma, the precursor stem cell exists covertly and grows when primary cultures are established in vitro. The fact that these variants established from xenografts have been consistently preneoplastic stem cells rather than carcinomas supports this concept. When clones are established directly from xenografts by modifying the tissue disaggregation method, other karyotypes have been isolated.

Thus, we believe that cells with other karyotypes do develop within the lesions in vivo but are overgrown by the dominant stem cell in vitro. We have recently isolated tumorigenic variants (i.e., carcinomas form directly with no precursor lesion) by serially transplanting lesions by trocar before establishing cells in culture.

Although the MCFIOAT xenografts progress to carcinoma with approximately a 25% incidence, rapidly growing xenografts are rare. Carcinoma has generally been diagnosed only after removal of the lesion and histologic examination. Indeed, the largest lesion to date has weighed 200 mg and most weigh less than 50 mg. When one considers the natural history of human breast cancer, which takes years to develop, the slow growth of these relatively early carcinomas is not surprising.

However, results with the MIB-1 antibody against the Ki-67 nuclear proliferation antigen clearly demonstrate that the proUferative rate increases with progression in the MCFIOAT xenograft model from 5% in simple hyperplastic lesions to 10% in CIS and 20% in invasive cancers. These rates resemble those reported for lesions from human patients (Barbareschi et al., 1992; PaveUc et al., 1992; Jensen, V. et al., 1995).

The MCFIOAT model should be valuable to analyze genetic alterations in the development of breast cancer. Specific genes can be experimentally altered in the cells to determine the effect on the incidence of cancer in xenografts, and new genes may be identified by comparing cells before and after spontaneous progression in xenografts.

A number of genes implicated in breast cancer are being analyzed in the MCFIOAT system. Immunohistochemical studies with DOT antibody to p53 (detects both wild type and mutant forms) found that staining was not a feature of hyperplastic lesions, CIS, or adenocarcinomas developing in MCFIOAT xenografts; significant increased expression of p53 was detected only in squamous carcinomas (Iravani et al., 1996). A number of investigators have found that/753 is not mutated in MCFIOA (Diella et al., 1993; Gudas et al., 1995; Merlo et al., 1995) nor MCFIOAT cells that form preneoplastic lesions in xenografts (Chen, Y.Q. et al.. 1995; Shekhar et al., 1995).

Taken together, these data suggest that mutated p53 does not play a significant role in producing the preneoplastic phenotype or in progression to adenocarcinoma but may be important in generation of squamous carcinoma. However, the expression of a conformationally altered, but not mutated, p53 does increase with successive transplant generations (MCF10AT1 / MCF10AT2B/ MCF10AT3B) and is accompanied by diminished normal/755 function (Shekhar et al., 1995).

Although erbB-2 is rarely amplified in hyperplastic human biopsy specimens, erbB-2 overexpression can be detected beginning at the stage of atypical hyperplasia in the MCFIOAT xenograft model (Iravani et al., 1996). As in human cancer, the proportion of erbB-2 - overexpressing MCFIOAT lesions increases at the stage of CIS and decreases somewhat in invasive adenocarcinomas. Although erbB-2-expression increases two- to threefold with in vivo passage (Wang and Miller, unpublished), MCFlOAneoT cells transfected with erbB-2 (MCFIOHE) are not tumorigenic (Ciardiello et al., 1992), nor do they form persistent lesions in immune-deficient mice (Miller, unpublished).

As discussed previously, because clones of MCFlOAneoT may not form persistent preneoplastic xenograft lesions, we hypothesize that a defect in addition to the T24 ras is necessary to produce the preneoplastic phenotype observed in MCFlOAneoT and MCFIOAT variants. We suggest that the MCFIOHE cells, which were cloned from MCFlOAneoT after transfection with the erbB-2 expression construct (Ciardiello et al., 1992), lack that hypothetical genetic defect. Growth in vivo selects for the second genetic defect and clones of MCFIOAT variants do form lesions.

Fred Raymond Miller
Advances in Oncobiology

References