Virchows Archiv

, Volume 446, Issue 1, pp 10–14

Immunophenotype of lymphocytic infiltration in medullary carcinoma of the breast

Authors

    • Department of PathologySaitama Medical Center, Saitama Medical School 1981
  • Jun-ichi Tamaru
    • Department of PathologySaitama Medical Center, Saitama Medical School 1981
  • Goi Sakamoto
    • Department of Breast PathologyCancer Institute and Hospital, Japanese Foundation for Cancer Research
  • Kiyoshi Ohnisi
    • Department of SurgerySaitama Medical Center
  • Shinji Itoyama
    • Department of PathologySaitama Medical Center, Saitama Medical School 1981
Original Article

DOI: 10.1007/s00428-004-1143-9

Cite this article as:
Kuroda, H., Tamaru, J., Sakamoto, G. et al. Virchows Arch (2005) 446: 10. doi:10.1007/s00428-004-1143-9

Abstract

Medullary carcinoma (MC) of the breast is characterized by large anaplastic cells and infiltration by benign lymphocytes. Patients with this pattern of breast carcinoma are considered to have a better prognosis than those with other histological subtypes. We reviewed cases of primary breast carcinoma that were surgically resected between 1990 and 2004. Of these, 13 cases of medullary carcinoma of the breast with lymphocyte infiltration were reported. Tests for CD3, CD4, CD8, CD20, CD56, TIA-1, and granzyme B were performed on paraffin sections. We found that the MC contained very few NK cells, as assessed by their reactivity with the CD56 antibodies. However, MC had a significantly greater percentage of CD3, CD8, TIA-1, and granzyme B lymphocytes infiltrating the stroma of the tumor. Furthermore, more CD8-positive than CD4-positive T-cell lymphocytes were present within the tumor cell nests in MC, as opposed to the proportion in usual ductal carcinoma. The infiltrating cytotoxic/suppressor T cells in MC represent host resistance against cancer, and the high grading of the T-cell infiltration could explain, in part, a key mechanism controlling the good prognosis for this type of tumor and solve the pathological paradox of MC.

Keywords

Medullary carcinomaCD8CD56TIA-1Granzyme B

Introduction

In 1949, Moore and Foote described their experience with a well-circumscribed form of breast carcinoma with lymphoid stroma [10]. It was called medullary carcinoma (MC) and was recognized to have a favorable prognosis compared with other histological subtypes. However, it was not singled out as a distinct entity. In 1977, Ridolfi et al. redefined the histological criteria for typical and atypical MC [15]. In 1988, Wargotz and Silverberg found that cases having only one of the “atypical” features listed by Ridolfi et al.—intraductal carcinoma, sparse mononuclear infiltrate, or focal marginal infiltration—should not be excluded from the typical MC category [20]. There is also controversy over the prognosis in MC. Most authors state that MC has a more favorable prognosis when compared with invasive ductal carcinoma. However, numerous authors make the error of grading MC according to prognostic factors applicable to other breast carcinomas. In MC, neither the nuclear grading of the modified Bloom–Richardson scoring system nor mitotic activity applicable for other types of breast carcinoma is effective [12]. Some authors consider that this better prognosis is due, at least in part, to the presence of the lymphocytic infiltrate that is closely associated with the tumor [2, 5, 17, 18, 21, 22]. Questions persist concerning the specific characteristics of this histological subtype of breast carcinoma. One possible hypothesis for the improved prognosis in MC is that infiltrating lymphocytes in MC possess cytotoxic activity. The aims of this study were to characterize the immunophenotype of the lymphocytic infiltration and to define the association of these findings with the established prognosis of MC.

Materials and methods

The subjects were cases of primary breast carcinoma that were surgically resected at Saitama Medical Center, Saitama Medical School between 1990 and 2004 and were reviewed for the histological presentation of MC. Of these, 13 cases were diagnosed as MC with abundant lymphocytic infiltration (Fig. 1, Fig. 2, Fig. 3). Paraffin-embedded slides stained with hematoxylin and eosin were used for the histological analysis. Histological typing was done using the criteria of Wargotz and Silverberg classification (Table 1) [20]. All of the histological slides were reviewed by two pathologists without knowledge of the clinical outcome. In addition, ten control specimens of infiltrating ductal carcinoma of the breast were selected from recent surgical samples that had been studied in our department and that contained abundant lymphocytic infiltration with nuclear grade 3.
Fig. 1

Low-power microphotograph of medullary carcinoma. Carcinoma was distributed in sharply defined bands separated by stroma permeated predominantly by lymphoid cells

Fig. 2

Low-power microphotograph of medullary carcinoma showing the broad, interanastomosing sheets of the tumor cells with indistinct cell borders, as described by Ridolfi et al. [15]

Fig. 3

High-power microphotograph of medullary carcinoma. This carcinoma had mitotic figures and pleomorphic nuclei

Table 1

Histological criteria for the diagnosis of medullary carcinoma. Modified from Wargotz and Silverberg [20]

Primary*

Predominantly circumscribed border

Syncytial growth>75%

Presence of admixed stromal mononuclear infiltrate

Grade 2 or 3 nuclei

Secondary**

Microscopically completely circumscribed

2+ to 3+ mononuclear infiltrate

Absence of in situ carcinoma

*Diagnosis of invasive ductal carcinoma is mandatory if any one of these criteria is not met

**Diagnosis of invasive ductal carcinoma is mandatory if two or more of these criteria are not met

Immunohistochemical staining was performed according to the following protocol. Consecutive sections from formalin-fixed, paraffin-embedded tissue blocks were cut at 5-μm intervals, deparaffinized, and dehydrated with xylene and graded alcohol. Pressure cooker pretreatment (1 cycle of 1 min) was performed at pH 6.0 for all antibodies. Endogenous peroxidase activity was quenched with 3% hydrogen peroxide at room temperature for 30 min. Sections were blocked with 10% goat serum for 60 min and then incubated with the primary antibody for 1 h at room temperature. Tests for CD3, CD4, CD8, CD20 (L26), CD56, TIA-1, and granzyme B were performed on paraffin sections. The primary antibodies and dilutions are summarized in Table 2.
Table 2

Monoclonal antibodies used in the study and the immunophenotypes of lymphocytes in breast cancer

Antibody

Clone

Source

Dilution

Main cellular expression

Anti-CD3

M7254

Dako-Japan

×100

T-cell lineage

Anti-CD4

1F6

Novocastra

×20

Helper T lymphocytes

Anti-CD8

C8/144B

Dako-Japan

×100

Cytotoxic/suppressor T lymphocytes

Anti-CD20

L26

Dako-Japan

×100

B-cell lineage

Anti-CD56

1B6

Novocastra

×50

NK cells, activated T cells

Anti-TIA-1

2G9

Immunotech

×50

Cytotoxic granule in cytotoxic

T cells and NK cells

Granzyme B

GrB-7

Monosan

×20

Cytotoxic granule in cytotoxic

T cells and NK cells

For every case, the evaluation of the immunoreactions was done on slides, with lymphocytes infiltrating the MC as well as the control. The staining result for each antibody was scored by two authors, and a consensus score was obtained (Fig. 4, Fig. 5, Fig. 6). Grading of lymphocyte infiltration in cancer tissues was made according to the following standards, which are a modified method of the study by Ben-Ezra and Sheibani: 0=no positive cells; 1+=0–50%; 2+=50–75%; 3+=75–100% [2]. Multiple group comparisons were analyzed by Mann–Whitney U test. All statistical tests were considered significant at the P<0.05 level.
Fig. 4

Significant numbers of CD8-positive cells were seen within surrounding stroma in medullary carcinoma

Fig. 5

Only a few CD56-positive cells were seen within the tumor cell nests and surrounding stroma in one of the medullary carcinoma cases

Fig. 6

Large numbers of granzyme-B-positive cells were seen infiltrating the stroma in medullary carcinoma

Results

The results of the immunohistochemical analysis for immunophenotype are summarized in Table 3. The grading of CD3-positive T-lymphocytes infiltrating the stroma was significantly greater than that of CD20-positive T lymphocytes in the MC and control cases (P<0.05). In contrast, no significant difference was found between the grading of CD3-positive lymphocytes between these two groups in the surrounding stroma. The grading of CD8-positive lymphocytes in the surrounding stroma was significantly greater than that of the CD4-positive lymphocytes in the MC, as compared with those in the control cases (P<0.05). There were only two cases showing staining with CD56 antibodies, which detect NK cell-associated antigens. In these cases, the percentage of cells stained was usually less than the total percentage of T cells present.
Table 3

The relationship between the histological classification and grading of lymphocyte infiltration with immunological markers of breast cancer

Cases

CD3

CD4

CD8

CD20

CD56

TIA-1

Granzyme B

MC cases

1

3

1

2

1

0

2

2

2

3

0

3

1

0

3

1

3

3

1

3

1

0

3

2

4

2

1

2

1

0

2

3

5

3

0

3

1

0

2

2

6

3

1

3

1

0

3

2

7

3

0

3

1

1

3

2

8

3

1

2

1

0

2

2

9

3

1

3

1

0

3

1

10

1

1

2

1

0

2

1

11

3

2

3

1

0

3

2

12

3

1

3

1

1

2

2

13

3

1

2

1

0

2

1

Control cases

1

3

3

1

1

0

2

1

2

3

3

2

1

0

1

1

3

3

2

1

1

0

2

1

4

3

3

2

1

0

2

1

5

3

3

3

0

0

2

1

6

3

3

2

1

0

3

1

7

3

3

2

1

0

3

0

8

3

3

2

1

0

3

1

9

3

3

2

1

0

2

1

10

3

3

2

2

0

3

1

*Significance was evaluated by Mann-Whitney’s U test, and the p value indicates whether any significant difference existed between MC and general types of breast carcinoma

The grading of TIA-1-positive cells was nearly equal to that of those stained with CD8. The grading of TIA-1-positive cells was not significantly greater than that of CD8-positive cells in the MC and control groups. The percentage of stromal TIA-1-positive cells was higher in the MC cases, although not significantly, than in the control cases.

The grading of granzyme-B-positive cells was consistently lower than that of TIA-1-and CD8-positive cells. However, the grading of granzyme-B-positive cells was markedly higher in cases of MC, in contrast to that of the control cases (P<0.05).

Discussion

MC of the breast is a rare tumor that has a favorable prognosis [1, 4, 12, 13, 15, 20]. The characteristic histological features of this tumor are large anaplastic cells and infiltration by benign lymphocytes. Numerous authors make the error of grading MC according to prognostic factors applicable to other breast carcinomas. In MC, neither the nuclear grading of the modified Bloom–Richardson scoring system nor mitotic activity applicable for other types of breast carcinoma is effective. These problems are what we call the “paradox of MC.” Some authors consider that this better prognosis is due, at least in part, to the presence of the lymphocytic infiltration, which is closely associated with the tumor. However, Ridolfi and Rapin classified MC using general diagnostic criteria and found that only about 30% of the cases fulfilled the strict criteria for that diagnosis [12, 15]. Even so, most authors, following Moore and Foote, retain this characteristic for the diagnosis [10].

As in previous studies of infiltrating ductal breast carcinomas, we found that the MCs contained very few NK cells, as assessed by their reactivity with CD56 antibodies. NK cells are a subpopulation of lymphocytes that have non-induced cell-mediated cytotoxicity against malignant tumor cells. They help protect the body against neoplasms and their metastatic dissemination by interacting with and destroying tumor cells. Therefore, it is not unreasonable to expect a higher percentage of CD56-positive cells with the NK phenotype to be present within MC. Bhan et al. reported that many NK cells, assessed by their reactivity with Leu-7 [3], were found in MC, which seems to differ from our results. However, similar to our findings, Ben-Ezra and Sheibani reported non-Leu-7-positive cases in 12 MC cases, in contrast to the study by Bhan et al. [3]. However, they did not describe the criteria required for a case to be considered MC. Furthermore, the result of the low percentage of CD56-positive cells might have been due to the small number of cases followed in our study. Thus, the NK-positive cell is scanty in the stroma of MC, and the mechanisms have still not been clarified.

Several authors have attempted to profile the lymphocyte subpopulations in breast carcinoma [2, 7, 8, 9, 16, 19]. These reports have shown a predominance of T cells and a paucity of B cells and NK cells. Attempts at further classification of these lymphocytes have led to disparate results, with some studies showing a predominance of T-helper cells and others showing a predominance of T-suppressor/cytotoxic cells. Furthermore, a survival advantage of marked lymphocytic infiltration has been demonstrated for several types of solid malignancies, including breast, renal, colon, malignant melanoma, and several others [6, 9, 11, 14].

Recently, antibodies that recognize antigens associated with the cytotoxic phenotype have become available by paraffin section. TIA-1 is a granule-associated RNA-binding protein that defines a subpopulation of CD8-positive T cells with cytotoxic activity. Granzyme B is a serine protease, expressed specifically by activated cytotoxic lymphocytes. Yakirevich et al. reported that MC is infiltrated by an increased percentage of lymphocytes with cytotoxic potential and activity, TIA-1- and granzyme-B-positive cells [21]. For these reasons, TIA-1 and granzyme B are considered to be capable of suppressing MC progression.

In this study, MC had significantly higher percentages of CD3-, CD8-, TIA-1-, and granzyme-B-positive lymphocytes infiltrating the stroma of the tumor than those in the usual breast cancer. Thus, the increase in infiltrating cytotoxic lymphocytes in MC suggests that cytotoxically active lymphocytes may be an important mechanism governing the favorable prognosis.

In conclusion, our results indicate that cancer nests of MC are infiltrated by lymphocytes that express immunohistochemical markers of cytotoxic potential and activity. These results could explain, in part, a key mechanism controlling the good prognosis for this type of tumor and solve the pathological paradox of MC.

Acknowledgements

We sincerely thank A. Adachi, M.D., I. Miura, M.D., Y. Toyosumi, M.D., S. Momose, M.D., T. Ougida, C.T., M. Abe, C.T., T. Saito, C.T., K. Kiuchi, C.T., K. Ohsawa, C.T., K. Matsuno, C.T., K. Hanami, C.T., Y. Ohno, C.T. and T. Masuda, C.T. for their technical assistance and advice.

Copyright information

© Springer-Verlag 2004