The morphologic features of primary bilateral breast carcinoma have been well elucidated, but it is not known whether tumors at two sides share a common genetic profile and undergo the same clinical course. To address this issue, morphologically comparable epithelial and stromal cells in 18 paired primary bilateral breast tumors were microdissected and subjected to comparisons for the frequency and pattern of loss of heterozygosity (LOH) and microsatellite instability (MI), as well as the profiles of comparative genomic hybridization. Of 18 paired bilateral epithelial samples assessed with 10 DNA markers at five chromosomes, 78 altered loci were found; of these, 23 (29.5%) displayed concurrent and 55 (70.5%) showed independent LOH, MI, or both. Of 18 paired bilateral stromal samples assessed with the same markers, 70 altered loci were seen; of these, 9 (12.9%) displayed concurrent and 61 (87.1%) showed independent LOH, MI, or both. Collectively, all the markers and 30 (83.3%) of 36 paired bilateral epithelial and stromal cells displayed significantly more (P < 0.01) independent than concurrent LOH, MI, or both. In contrast, the epithelial cells of a pulmonary small cell carcinoma metastasized to both breasts displayed concurrent LOH at each of the four altered loci. Of seven selected cases for comparative genomic hybridization, six (86%) displayed chromosomal changes, but none showed an identical pattern and frequency of changes in both breasts. The significantly higher rate of independent genetic alterations in morphologically comparable cells of paired bilateral primary breast tumors supports the notion that the development and clinical course of tumors in two sides differ substantially; consequently, different interventions might be needed for the optimal management of bilateral breast tumors.

Microdissected epithelial and stromal cells from 15 cervical small-cell carcinoma patients and 9 healthy control subjects were assessed for loss of heterozygosity with polymorphic DNA markers at chromosomes 3p and 11p. Among malignant lesions assessed with 7 markers at 3p, 21 allelic losses were detected from 193 informative samples. Of losses, 20 were in epithelial and 1 was in normal-appearing stromal cells. Among losses in epithelial cells, 16 were from 44 samples informative for 3 markers within 3p21.2–p14.2 (0.36 loss/sample), whereas only 4 were from 54 samples informative for 4 markers outside the region (0.09 loss/sample), suggesting a “hot spot” of genetic alterations within 3p21.2–p14.2. Among malignant lesions assessed with 2 markers within 11p14–p12, 15 losses were seen in 52 informative samples. Of losses, 10 were in epithelial and 5 were in normal-appearing stromal cells. Of 10 epithelial samples showing losses within 11p14–p12, 8 also displayed losses within 3p 21.2–p14.2, suggesting a concurrent involvement of these loci in tumor development or progression. The five losses in stromal cells were in four cases that showed no loss in epithelial cells with same markers, suggesting that stromal cells might play initiative roles in tumor development.

The acquisition of comparable quality and quantity of DNA extracts is the prerequisite to the success of comparative genetic analyses. Although several DNA extracting protocols on paraffin sections have been introduced, the importance of deparaffinization, the procedure for obtaining an adequate hematoxylin staining, the significance of the ratio of the cell number to the enzyme volume, and a practical means for monitoring the digestion process have not been sufficiently addressed. These, however, are the most important factors accountable for a failure of DNA extraction. To minimize the impact of these factors, we have developed several unique strategies, including: (1) incubating sections at 80 degrees C for 30-60 minutes prior to xylene treatment, (2) checking each section to insure the complete removal of paraffin; (3) treating hematoxylin stained sections or cells with de-staining solutions; (4) using a micrometer inserted into the eyepiece of a microscope to estimate the number of cells collected and adjusting the enzyme volume according to the cell number; and (5) monitoring the digestion process with a magnifier. With these strategies, we have been able to consistently obtain comparable quality and quantity of DNA extracts which yielded uniform PCR products regardless of variations in tissue embedding and processing.