Immunohistological Patterns of Immunoglobulins in Dysplasias, Benign Neoplasms and Carcinomas of the Breast

Novel method for differentiation between Trastuzumab and host adaptive response

Humoral immune response to human epidermal growth factor receptor 2 (HER-2/neu or ErbB-2) has been detected in sera of breast cancer patients and shown to be an appropriate prognostic marker (Taylor et al., 2007). However, since Trastuzumab (Herceptin) is a widely used monoclonal antibody as cancer therapy agent for tumors over-expressing HER-2, there is a need for an efficient way to detect host-generated antibodies against HER-2 without the confounding effect of Herceptin. Here we describe a screening method developed to decipher between host antibodies against HER-2 and that of Herceptin. By producing a series of truncation mutants within the epitope of Herceptin, we were able to inhibit this binding. We demonstrated also that by a three amino acid substitution (PPF→SSS) we were able to abrogate Herceptin binding while generating a highly conserved HER-2 extracellular domain (ECD). By producing a stable cell line that expresses this mutated form of the human HER-2 ECD, we have a source of this protein to probe patient sera. Our method represents a proof of principle that mutated HER-2 which we constructed could be used to distinguish between a host response against HER-2 and the monoclonal antibody Herceptin targeting the same protein.

Focused antibody response in plasma cell-infiltrated non-medullary (NOS) breast cancers

Breast tumors with prominent plasma cell (PC) infiltrates often have a more favorable natural course that may plausibly be mediated by anti-tumor activity of the elicited antibodies. These breast tumor-associated PCs are typically IgG dominant in contrast to normal breast PCs, which are mainly IgA. It is our hypothesis that this PC infiltration represents a host immune response that is driven by one or more tumor antigens. Previously, we and others showed that medullary carcinoma (MC) had a focused repertoire and features suggestive of a protein antigen driven response. Infrequently, non-MC, not otherwise specified (NOS) breast tumors may exhibit heavy PC infiltrations, also of IgG isotype. In this first characterization of this favorable prognosis NOS subgroup, IgG heavy chain (Hc) and light chain (Lc) variable (V) regions from three PC-infiltrated NOS tumors were randomly cloned and sequenced. We found biased (V) gene usage by the infiltrating PCs and somatic hypermutation in the rearranged Ig Hc and Lc V regions that were compatible with antigenic selection of the progenitor B cells. The antibody response of NOS infiltrated breast cancer is repertoire-focused, with 13-68% of isolates being clonally reiterated in the samples. Each NOS patient used distinct Hc V-D-J and Lc V-J rearrangements, with her own immune response "footprint," but the overall pattern of gene usage followed that typical of exogenous antigen-induced immune responses. The data are consistent with the hypothesis that PC infiltrates infrequently arising in NOS tumors, as previously inferred for MC, are in response to one or more breast cancer-associated protein tumor antigens.

Antigen-driven oligoclonal expansion of tumor-infiltrating B cells in infiltrating ductal carcinoma of the breast

The objective of this study was to determine whether tumor-infiltrating B cells (TIL-B) of infiltrating ductal carcinoma (IDC) of the breast represent a tumor-specific humoral immune response. Immunohistochemical analysis of three Her-2/neu-negative IDC tumors from geriatric patients showed that TIL-B cluster in structures similar to germinal centers containing CD20(+) B lymphocyte and CD3(+) T lymphocyte zones with interdigitating CD21(+) follicular dendritic cells, suggesting an in situ immune response. A total of 29, 31, and 58 IgG1 H chain clones was sequenced from the three IDC tumors, respectively. Intratumoral oligoclonal expansion of TIL-B was demonstrated by a preponderance (45-68%) of clonal B cells. In contrast, only 7% of tumor-draining lymph node and 0% of healthy donor PBL IgG H chains were clonal, consistent with the larger repertoires of node and peripheral populations. Patterns and levels of TIL-B IgG H chain somatic hypermutation suggested affinity maturation in intratumoral germinal centers. To examine the specificity of TIL-B Ig, a phage-displayed Fab library was generated from the TIL-B of one IDC tumor. Panning with an allogeneic breast cancer cell line enriched Fab binding to breast cancer cells, but not nonmalignant cell lines tested. However, panning with autologous tumor tissue lysate increased binding of Fab to both tumor tissue lysate and healthy breast tissue lysate. These data suggest an in situ Ag-driven oligoclonal B cell response to a variety of tumor- and breast-associated Ags.

Medullary breast carcinoma. A reevaluation of 95 cases of breast cancer with inflammatory stroma

The hallmarks of diagnosis of medullary breast cancer (MedBC) used by the authors since 1977 have been that the tumor is well circumscribed, has syncytial architecture in greater than 75% of its surface, contains diffuse inflammatory infiltrate, has atypical nuclei, and forms no glandular pattern. In order to assess the clinical utility of these criteria, we studied a series of 95 previously untreated, surgically operable patients with breast carcinoma at the Institut Gustave-Roussy (IGR) between 1960 and 1979. A diagnosis of MedBC was initially made for these patients or suspected based on abundant inflammatory stroma observed in a histologic evaluation. Using these criteria, 26 cases were identified as typical medullary carcinoma (TMC), 23 cases as atypical medullary carcinoma (AMC), and 46 cases as nonmedullary carcinoma (NMC). The 26 cases of TMC represent a very small fraction of the total infiltrating operable breast carcinomas diagnosed at IGR during the same time period. The prognosis for these 26 patients was much more favorable than for the other groups. They had a 10-year disease-free survival of 92% compared with 53% for the AMC group and 51% for the NMC group. Neither distant metastasis nor secondary primaries of the same histology were seen. Therefore, it is possible with the use of strict histologic criteria to distinguish a group of patients with a much more favorable prognosis. This histologic diagnosis alone renders a most favorable prognosis for the patient even if other factors such as large tumor size and lymph node involvement are present and, by inference, the only therapy needed is the removal of all tumor. In contrast, atypical forms have a prognosis no different from other atypical types of breast carcinomas without inflammatory stroma, and adjuvant therapy appears to be justified if other factors warrant it.

Evaluation of nine different fixatives. 1. Preservation of immunoglobulin isotypes, J chain, and secretory component in human tissues

Eight cross-linking fixatives were tested for their ability to preserve IgG, IgA, IgM and IgD isotypes, J chain, and secretory component (SC) in paraffin-embedded specimens of human tonsils and colonic mucosa. The results were compared with the antigenic preservation afforded by cold 96% ethanol (with or without inclusion of a prefixation 48-hr washing period). A semiquantitative immunofluorescence scoring system was applied. In relation to the cytoplasmic scores of 3.0 assigned to IgG and IgA immunocytes after ethanol fixation, the following median scores were obtained with the other fixatives: routine formalin, 1.3 (IgG and IgA); glutaraldehyde (1%)-formalin, 0.3 (IgG and IgA); Baker's formol calcium, 2.3 (IgG) and 1.5 (IgA); formol sublimate, 1.0 (IgG) and 1.2 (IgA); acetic acid (2%)-formol saline, 2.0 (IgG and IgA); Bouin's fluid and Susa fixative, 0 (IgG) and 1.3 (IgA); and carbodiimide, 2.3 (IgG) and 2.0 (IgA). Most aldehyde-based fixatives afforded poorer result for IgA immunocytes when an anti-alpha-chain reagent of restricted specificity was applied. The result was usually slightly better for IgM cells than for IgG cells, but IgM did not resist fixation in Bouin's fluid and Susa fixative to the same extent as IgA. IgD immunocytes were poorly revealed except after fixation with Baker's formol calcium, carbodiimide, and acetic acid-formol saline, which afforded median fluorescence scores of 1.0-1.8 compared with 3.0 after ethanol fixation. J chain in IgA immunocytes was most intensely stained in sections of ethanol-fixed tissue denatured in acid urea (median score, 3) but was also well revealed after fixation with formalin, acetic acid-formol saline, Bouin's fluid or Susa fixative (median scores, 1.8-2.0). The results with most of the cross-linking fixatives were much less favorable for IgA, IgM, J chain and SC in colonic crypt epithelium than in immunocytes. Only carbodiimide afforded fluorescence scores for IgA and SC approaching those obtained after ethanol fixation (2.0 and 1.5 vs. 3.0 and 2.8). The effect of proteolytic unmasking depended on the fixative - not only for different antigens and locations in the tissue, but also for the same antigen in apparently similar cells.