A Study of the Molecular Heterogeneity of LCIS

Institution: University of California, Los Angeles
Investigator(s): Sanford Barsky, M.D. -
Award Cycle: 2000 (Cycle VI) Grant #: 6JB-0044 Award: $150,000
Award Type: IDEAS II
Research Priorities
Biology of the Breast Cell>Pathogenesis: understanding the disease



Initial Award Abstract (2000)
Lobular carcinoma in situ (LCIS) is a poorly understood disease of the breast that involves the ductal and glandular systems. The human female breast has a series of ducts connected like the branches of a tree that end in glands analogous to the leaves of this tree. When the cells that comprise these end glands divide and form a disease called lobular carcinoma in situ, this is thought to increase the relative risk for the development of invasive breast cancer approximately 10 fold. Although all types of LCIS are regarded by pathologists, surgeons and oncologists in the same way as increasing breast cancer risk, the disease might not be the same disease in every patient. LCIS, when examined under the microscope, seems to take on different appearances: 1) cellular criteria: the appearance of the LCIS cells may differ ranging from small similar-appearing cells to larger more variable-appearing cells to cells that look like rings; 2) geographic criteria: LCIS can be present in only a single area of the breast, multiple areas present in both breasts, or in large areas involving both the glands as well as the ducts; 3) association criteria: LCIS can occur in isolation, or in association with infiltrating ductal or other type of breast carcinoma, or in association with infiltrating lobular carcinoma. These differing manifestations of LCIS suggest that it may be incorrect clinically to regard the disease as representing only one disease entity. In fact the disease may not be only morphologically different but also molecularly and biologically different with differing clinical implications.

Using a combination of laser capture microdissection (a new technique to retrieve cells from microscopic slides), loss of heterozygosity (LOH) and comparative genomic hybridization (CGH) studies (studies designed to investigate the molecular biology of cancer) and appropriate statistical analyses, we propose to analyze over 200 cases of LCIS available to us in previously collected pathological material 1) to determine in fact whether LCIS is a collection of different diseases; 2) to determine specifically whether some types of LCIS actually progress to invasive breast cancer more often than other types; and 3) to determine whether such molecular stratification might ultimately identify subsets of patients with differing breast carcinoma risk, treatment and prophylaxis options.


Final Report (2004)
Lobular carcinoma in situ (LCIS) is a poorly understood epithelial proliferative lesion of the breast that involves terminal ducts and acini and is thought to increase the relative risk for the development of invasive breast cancer approximately 10 fold. Although all types of LCIS are regarded by pathologists, surgeons and oncologists in the same way as increasing breast cancer risk, the disease might not be the same disease in every patient. LCIS may be, in fact, quite a heterogeneous disease. Some types of LCIS may be innocuous, some types associated with increased risk of invasive breast cancer and some types actually progress into invasive lobular carcinoma.

Using a combination of laser capture microdissection (a new technique to retrieve cells from microscopic slides), loss of heterozygosity (LOH), comparative genomic hybridization (CGH) studies (studies designed to investigate the molecular biology of cancer) and appropriate statistical analyses, in this completed study, we have analyzed 200 cases of LCIS available to us in archival pathological material and determined that LCIS is, in fact, a molecularly heterogeneous disease: Some types of LCIS share a common chromosomal loss (LOH) and gain (microsatellite instability) pattern with concomitant invasive lobular carcinoma suggesting that these LCIS types have, in fact, clonally progressed to invasive lobular carcinoma; other types of LCIS share a different chromosomal loss (LOH) and gain (microsatellite instability) pattern with the concomitant invasive carcinoma that is present suggesting that these LCIS types perhaps are only risk factors; still other types of LCIS show no obvious genomic alterations (losses or gains) suggesting that these LCIS types perhaps are innocuous. In the latter half of the study we analyzed all of the remaining cases and applied additional molecular probes (based on the sequencing of the entire human genome) to our group of 200 LCIS cases and confirmed that our three tier stratification scheme for LCIS held. Future studies supported by other grants will be prospective rather than retrospective in nature.


Symposium Abstract (2003)
Introduction: The lymphovascular embolus is a clump of breast tumor cells found within a lymphatic or blood vessel channel. The presence of such tumor cell clumps in breast cancer is a poor prognostic sign indicating a likelihood of local recurrence and distant metastases. The mechanisms behind the formation of these deadly tumor emboli and even basic questions concerning how these clumps find themselves within vessels remain unknown.

Methods: We used our previously developed human model of breast cancer in the “scid” and nude mouse (MARY-X), which exhibits florid lymphovascular tumor emboli to address these questions. We likened these questions to the story of the “ship in the bottle.”

Results: Injection of single cells of MARY-X failed to produce any lymphovascular emboli whereas injections of clumps of tumor cells, termed spheroids, produced florid lymphovascular emboli. Since the tumor cell clumps were held together by the overexpression of a strong adhesion molecule (E-cadherin), it seemed to us unlikely that the clumps could actually invade through the walls of vessels. It seemed to us more likely that the tumor cell clumps actually stimulated vessel endothelial lining cells to form and grow around them. To test this hypothesis we carried out and observed the following: MARY-X spheroids were co-injected with green fluorescence protein (GFP)-labeled murine embryonal fibroblasts. The emerging tumors exhibited florid lymphovascular tumor emboli in which GFP was present within both the tumor stroma as well as the endothelial lining cells which surrounded the tumor emboli. Bone marrow from GFP-transgenic male nude mice were tail-vein injected into lethally irradiated female mice, a technique which destroyed most of the recipient marrow. Mice exhibiting successful bone marrow engraftment were injected with MARY-X spheroids. The emerging tumors exhibited florid lymphovascular emboli in which GFP was again present within both the tu-moral stroma and the endothelial lining cells which surrounded the tumor emboli. Injected GFP-labeled spheroids alone produced no GFP signal within either the stroma nor the surrounding endothelial cells. The GFP present within the endothelial lining cells therefore was not the result of phagocytosis of liberated GFP but truly reflective of a mesenchymal / bone marrow origin.

Conclusion: Our studies indicate that the story of the lymphovascular embolus is the story of the “ship in the bottle.” It is a time-consuming and arduous task to have an artisan build a ship in a bottle, but a much simpler job to purchase an imitation at any flea market where the bottle has been built around the ship. So it is, by analogy, that it is much easier for the tumor cell clumps to stimulate vasculogenesis around themselves.

Should LCIS be regarded as a heterogeneous disease?
Periodical:Breast Journal
Index Medicus: Breast J
Authors: Barsky SH, Bose S
Yr: 1999 Vol: 5 Nbr: 6 Abs: Pg:407-412