Annals of Surgical Oncology

, Volume 19, Issue 5, pp 1499–1507

Loss of Heterozygosity at BRCA1 Locus Is Significantly Associated with Aggressiveness and Poor Prognosis in Breast Cancer

Authors

  • Satoko Okada
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Hiroyuki Kitao
    • Department of Molecular Oncology, Graduate School of Medical SciencesKyushu University
  • Sayuri Akiyoshi
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Nami Yamashita
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Hiroshi Saeki
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Eiji Oki
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Masaru Morita
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Yoshihiro Kakeji
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Yoshihiko Maehara
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
Breast Oncology

DOI: 10.1245/s10434-011-2166-5

Cite this article as:
Okada, S., Tokunaga, E., Kitao, H. et al. Ann Surg Oncol (2012) 19: 1499. doi:10.1245/s10434-011-2166-5

Abstract

Background

BRCA1 and BRCA2 are two major tumor suppressor genes for hereditary breast and ovarian cancer. In sporadic breast cancer, although somatic mutations of these genes are rare, loss of heterozygosity (LOH) at BRCA1 and BRCA2 loci is common.

Methods

LOH at BRCA1 and BRCA2 loci were investigated in 202 Japanese invasive breast cancer patients. The relationships between LOH at these gene loci and clinicopathologic characteristics were analyzed.

Results

Among 166 informative cases for both BRCA1 and BRCA2 loci, 69 (41.6%) and 52 (31.3%) tumors revealed LOH at BRCA1 and BRCA2 loci, respectively. LOH at BRCA1 LOH or BRCA2 locus was associated with higher nuclear grade (P < 0.0001, P = 0.0187). LOH at BRCA1 locus was associated with estrogen receptor and progesterone receptor negativity (P = 0.001 and P = 0.015) and significantly shorter disease-free survival (P < 0.0001), distant metastasis-free survival (P < 0.0001), and overall survival (P < 0.0001). In contrast, LOH at BRCA2 locus had no associations with estrogen receptor or progesterone receptor status and prognosis. LOH at BRCA1 locus was independently associated with poor prognosis in terms of disease-free survival (hazard ratio 3.08, 95% confidence interval [CI] 1.58–6.18, P = 0.0009), distant metastasis-free survival (hazard ratio 5.18, 95% CI 2.35–12.19, P < 0.0001), and overall survival (hazard ratio 4.97, 95% CI 1.84–15.1, P = 0.0013).

Conclusions

LOH at BRCA1 locus could be an independent prognostic biomarker useful in identifying a subgroup of patients with poor prognosis.

BRCA1 and BRCA2 are two major tumor suppressor genes, and women with germ line mutation in these genes are at increased risk of developing breast, ovarian, and other cancers.1 BRCA1 and BRCA2 play important roles in the repair of DNA double-strand breaks by homologous recombination. Loss of function of either protein leads to a dramatic increase in genomic instability, which could be underlying mechanism of cancer predisposition.2 BRCA1 has relatively broad cellular roles, which include DNA repair, transcriptional regulation, and chromatin remodeling. In contrast, BRCA2 is also involved in various cellular responses, but the major function is as a master regulator of homologous recombinational repair through the regulation of RAD51 activity.1,3,4

Loss of heterozygosity (LOH), the loss of a functional allele at a heterozygous locus, is the most common type of somatic alteration found in primary human breast tumors.5 Tumor of patients from BRCA1 and BRCA2 families typically exhibit a LOH or other somatic alterations of BRCA1 or BRCA2, respectively.6 In sporadic breast cancers, although these genes are rarely inactivated, LOH at BRCA1 and BRCA2 loci is common.2,711 LOH at BRCA1 and BRCA2 loci has been reported in 20–63% and in 20–54% of sporadic cases.1124 These observations suggest that LOH at these loci may reduce the function of BRCA1 or BRCA2, which probably contributes to tumorigenesis. Although there are many reports regarding the LOH at BRCA1 and BRCA2 loci in the sporadic breast cancers mentioned above, the associations between LOH at BRCA1 and BRCA2 loci with clinicopathologic features of breast cancer are inconsistent.

Recent gene profiling research via microarray revealed that BRCA1-mutant tumors show similar gene expression profiles to basal-type tumors, while BRCA2 tumors are mainly luminal type and lack a clear pathologic phenotype that is distinct from sporadic tumors.2 Thus, there are probably different paths of tumorigenesis in BRCA1 and BRCA2 mutation carriers. If the LOH at BRCA1 or BRCA2 loci causes the reduced function of these proteins, the breast tumors with LOH at BRCA1 or BRCA2 loci might have some common features with BRCA1- or BRCA2-mutant tumors. Most studies regarding LOH at BRCA1 and/or BRCA2 loci were performed in the 1990 s or early 2000 s. Few reports have focused on the differences in clinicopathologic and biologic features or prognosis among patients with LOH at BRCA1 and/or BRCA1 loci.

In this study, we extensively analyzed the incidence of LOH at BRCA1 and BRCA2 loci using our system for microsatellite analysis, called high-resolution fluorescent microsatellite analysis, and investigated the relationship between the LOH at the BRCA1 and/or BRCA2 loci and the clinicopathologic characteristics in Japanese breast cancer.25 We found that tumors with LOH at BRCA1 locus, regardless of LOH status at BRCA2 locus, have quite different clinicopathologic features from those of tumors with retention of heterozygosity (ROH) at both loci or LOH exclusively at BRCA2 locus. Our data also indicated that LOH at BRCA1 locus could be an independent poor prognostic factor in sporadic breast cancer.

Materials and Methods

Patients and Specimens

Five hundred thirty-one women underwent surgery for primary breast cancer at the Department of Surgery and Science, Kyushu University, between 1994 and 2007. Among them were 487 invasive breast carcinoma cases without neoadjuvant chemotherapy or neoadjuvant endocrine therapy. In the 487 cases, fresh samples for genomic DNA analyses could be obtained from 202 cases, which we utilized in this study. The study protocol was approved by the institutional review boards. Informed consent was obtained from all patients before tissue collection. Twenty-nine women had a family history of breast cancer. There was no information about mutations in BRCA1 and BRCA2; however, none of the patients had strong family history for breast or ovarian cancer.

LOH Analysis

Immediately after surgery, the specimens were placed in liquid nitrogen and stored at −80°C. Peripheral blood samples were also obtained from each patient. Genomic DNA was isolated as previously described.26 LOH at the BRCA1 and BRCA2 loci was analyzed with microsatellite markers for BRCA1 (D17S855 and D17S579) and BRCA2 (D13S171 and D13S267) using our microsatellite analysis system, which has been described in detail elsewhere.25,27 The PCR primer sequences (forward and reverse) were as follows: D17S855, 5′-CATGCCAAGACTGAAACTCC-3′ and 5′-AAACCCCAATGCCATAATGG-3′; D17S579, 5′-AGTCCTGTAGACAAACCTG-3′ and 5′-CAGTTTCATACCAAGTTCCT-3′; D13S171, 5′-CAATGGAACCAAATGTGGTC-3′ and 5′-AGTCCGATAATGCCAGGATG-3′; and D13S267, 5′-GGCCTGAAAGGTATCCTC-3′ and 5′-TCCCACCATAAGCACAAG-3′. The cases showing heterozygosity were defined as being informative. The peak value of tumor DNA was compared with that of the normal control DNA, and a reduction of ≥30% in the peak value of tumor DNA compared with that of normal control DNA was considered to indicate LOH. Other cases were considered to be ROH.28,29

Evaluation of ER, Progesterone Receptor, and HER2

ER, progesterone receptor (PR), and HER2 status were examined as previously reported.30 ER and PR were considered positive if ≥1% of the nuclei of cancer cells were stained. Tumors were considered HER2 positive only if they were scored 3+ by immunohistochemistry or 2+ by immunohistochemistry in combination with HER2 amplification (ratio of >2.0) by fluorescence in situ hybridization.

Statistical Analyses

All statistical calculations were performed by JMP software (SAS Institute, Cary, NC). The relationship between LOH or ROH at the BRCA1 and BRCA2 loci and clinicopathologic factors were assessed by χ2 and Fisher’s exact probability tests. Survival curves were plotted by the Kaplan–Meier method, and the log rank test was used to assess associations between variables and survival. Survival time was evaluated by the Cox proportional hazard model in multivariate analyses. Values of P < 0.05 were considered to be statistically significant.

Results

LOH at BRCA1 and BRCA2 Gene Loci in Breast Cancer

A total of 202 breast cancer samples were analyzed. The number of informative cases and the incidence of LOH at each microsatellite marker are listed in Table 1. LOH was detected in 79 (42.5%) of 186 cases informative for either D17S855 or D17S579, and in 56 (31.0%) of 180 cases informative for either D13S171 or D13S267. There were 166 cases informative for both BRCA1 and BRCA2, with 69 (41.6%) and 52 (31.3%) considered to show LOH at BRCA1 and BRCA2, respectively (Table 1). These values are within the ranges described in previous reports, including those in Western countries. The following analyses were performed with the 166 cases informative for both BRCA1 and BRCA2.
Table 1

Incidence of LOH at BRCA1 and BRCA2 gene loci

Gene

Locus

Informative cases, n

Cases with LOH, n (%)

Total cases analyzed

 BRCA1

D17S855

140

53 (37.9%)

D17S579

154

60 (39.0%)

D17S855 and/or D17S579

186

79 (42.5%)

 BRCA2

D13S171

134

38 (29.1%)

D13S267

125

35 (28.0%)

D13S171 and/or D13S267

180

56 (31.0%)

Informative cases for both BRCA1 and BRCA2

 BRCA1

D17S855 and/or D17S579

166

69 (41.6%)

 BRCA2

D13S171 and/or D13S267

166

52 (31.3%)

Associations between LOH at BRCA1 and BRCA2 Loci and Clinicopathologic Characteristics

First, we evaluated the associations between LOH at BRCA1 and BRCA2 loci and clinicopathologic characteristics, regardless of the presence of LOH at another gene locus (Table 1). LOH at BRCA1 locus was significantly associated with older age (P = 0.0475), lymph vessel invasion (P = 0.0131), higher nuclear grade (P < 0.0001), ER negative (−) and PR− status (P = 0.001 and P = 0.015) and HER2 positive (+) status (P = 0.0011). Tumors with LOH at BRCA1 locus were less frequently of the hormone receptor (HR) +/HER2− subtype (P = 0.0001). LOH at BRCA1 locus was not associated with tumor size, lymph node metastasis or vascular invasion. On the other hand, LOH at BRCA2 locus was significantly associated with lymph vessel invasion (P = 0.046) and higher nuclear grade (P = 0.0187). Tumors with LOH at BRCA2 locus were less frequently of the HR+/HER2− subtype and more frequently of the HR−/HER2− subtype (Table 2).
Table 2

Relationships between LOH at BRCA1 or BRCA2 locus and clinicopathologic features in breast cancer, regardless of the presence of LOH at another gene locus

Factor

Subcategory

BRCA1

P

BRCA2

P

ROH

LOH

ROH

LOH

n = 97 (58.4%)

n = 69 (41.6%)

n = 114 (68.7%)

n = 52 (31.3%)

Age, mean ± SE

 

53.4 ± 1.2

57.6 ± 1.6

0.0475

55.0 ± 1.1

55.4 ± 1.9

0.9569

T

1

36 (37.1%)

18 (26.1%)

0.4442

35 (30.7%)

19 (36.5%)

0.5898

2

49 (50.5%)

40 (58.0%)

 

63 (55.3%)

26 (50.0%)

 

3

9 (9.3%)

7 (10.1%)

 

10 (8.8%)

6 (11.5%)

 

4

3 (3.1%)

4 (5.8%)

 

6 (5.3%)

1 (2.0%)

 

n

Negative

63 (66.3%)

40 (58.0%)

0.2758

76 (67.3%)

27 (52.9%)

0.0812

Positive

32 (33.7%)

29 (42.0%)

 

37 (32.7%)

24 (47.1%)

 

ly

Negative

64 (73.6%)

37 (54.4%)

0.0131

74 (70.5%)

27 (54.0%)

0.046

Positive

23 (26.4%)

31 (45.6%)

 

31 (29.5%)

23 (46.0%)

 

v

Negative

79 (90.8%)

58 (85.3%)

0.29

94 (89.5%)

43 (86.0%)

0.5277

Positive

8 (9.2%)

10 (14.7%)

 

11 (10.5%)

7 (14.0%)

 

Nuclear grade

1

58 (61.1%)

13 (19.4%)

<0.0001

57 (50.9%)

14 (28.0%)

0.0187

2

14 (14.7%)

13 (19.4%)

 

15 (13.4%)

12 (24.0%)

 

3

23 (24.2%)

41 (61.2%)

 

40 (35.7%)

24 (48.0%)

 

ER

Negative

18 (18.6%)

29 (42.0%)

0.001

27 (23.7%)

20 (38.5%)

0.0534

Positive

79 (81.4%)

40 (58.0%)

 

87 (76.3%)

32 (61.5%)

 

PR

Negative

31 (32.0%)

35 (50.7%)

0.015

41 (36.0%)

25 (48.1%)

0.1409

Positive

66 (68.0%)

34 (49.3%)

 

73 (64.0%)

27 (51.9%)

 

HER2

Negative

83 (85.6%)

44 (63.8%)

0.0011

88 (77.2%)

39 (75.0%)

0.7581

Positive

14 (14.4%)

25 (36.2%)

 

26 (22.8%)

13 (25.0%)

 

Subtype

HR+, HER2−

75 (77.3%)

30 (43.5%)

0.0001

79 (69.3%)

26 (50.0%)

0.0102

HR+, HER2+

8 (8.3%)

10 (14.5%)

 

10 (8.8%)

8 (15.4%)

 

HR−, HER2+

6 (6.2%)

15 (21.7%)

 

16 (14.0%)

5 (9.6%)

 

HR−, HER2−

8 (8.3%)

14 (20.3%)

 

9 (7.9%)

13 (25.0%)

 

Associations between LOH at BRCA1 and BRCA2 Loci and Prognosis

Next, we investigated the associations between LOH at BRCA1 and BRCA2 loci and the prognosis of breast cancer patients. LOH at BRCA1 locus was associated with significantly worse prognosis with shorter disease-free survival (DFS) (log rank P < 0.0001, hazard ratio 3.20, 95% confidence interval [CI] 1.77–5.98), distant metastasis-free survival (DMFS) (log rank P < 0.0001, hazard ratio 4.82, 95% CI 2.38–10.56) and overall survival (OS) (log rank P < 0.0001, hazard ratio 5.65, 95% CI 2.39–15.53) (Fig. 1a–c). In contrast, LOH at BRCA2 locus was not significantly associated with the prognosis of breast cancer patients in terms of DFS (log rank P = 0.1725, hazard ratio 1.53, 95% CI 0.83–2.75), DMFS (log rank P = 0.3243, hazard ratio 1.41, 95% CI 0.70–2.73) or OS (log rank P = 0.3272, hazard ratio 1.49, 95% CI 0.65–3.28) (Fig. 1d–f). The negative impact of LOH at BRCA1 locus on prognosis was observed irrespective of ER status and node involvement. LOH at BRCA1 locus was associated with significantly worse prognosis with shorter DMFS both in ER+ (log rank P < 0.0001, hazard ratio 5.71, 95% CI 2.46–14.32) and ER− groups (log rank P = 0.0421, hazard ratio 3.98, 95% CI 1.05–25.89), and also in node-negative (log rank P = 0.0421, hazard ratio 2.90, 95% CI 1.04–8.65) and node-positive (log rank P < 0.0001, hazard ratio 7.77, 95% CI 2.78–27.62) groups (Fig. 1g–j).
https://static-content.springer.com/image/art%3A10.1245%2Fs10434-011-2166-5/MediaObjects/10434_2011_2166_Fig1_HTML.gif
Fig. 1

Associations between LOH at BRCA1 and BRCA2 loci and prognosis. The association between LOH at BRCA1 and BRCA2 loci, regardless of LOH at another gene locus, and the prognosis of breast cancer patients was analyzed. LOH at BRCA1 locus was associated with significantly worse prognosis than ROH at BRCA1 locus in terms of shorter DFS (P < 0.0001) (a), DMFS (P < 0.0001) (b), and OS (P < 0.0001) (c). LOH at BRCA2 locus was not significantly associated with DFS (d), DMFS (e), or OS (f). The negative impact of LOH at BRCA1 locus on prognosis was observed regardless of ER status or node involvement. g ER+ patients (P < 0.0001). h ER− patients (P = 0.0421). i Node-negative patients (P = 0.0412). j Node-positive patients (P < 0.0001)

Associations between LOH at BRCA1 or BRCA2 Locus Alone and LOH at Both Loci with Clinicopathologic Features

Among 166 cases analyzed, 34 cases (20.5%) exhibited LOH at both BRCA1 and BRCA2 loci, 79 cases (47.6%) were ROH at both loci, 35 cases (21.1%) exhibited LOH exclusively in BRCA1 locus, and 18 cases (10.8%) exhibited LOH exclusively at BRCA2 locus, respectively. We next investigated the clinicopathologic features in these four groups based on the combination of LOH at BRCA1 and BRCA2 loci (Table 3). Compared with tumors with ROH at both BRCA1 and BRCA2 loci, tumors with LOH exclusively at BRCA1 locus were associated with lymph vessel invasion (P = 0.0367), higher nuclear grade (P = 0.0004), ER− status (P = 0.0281) and HER2+ status (P = 0.0047) (Table 3, group A vs. B). Tumors with LOH exclusively at BRCA2 locus were associated with younger age (mean 47.1 ± 2.6 years), but not with the other clinicopathologic features evaluated (Table 3, group A vs. C). Tumors with LOH at both BRCA1 and BRCA2 loci were significantly associated with lymph vessel invasion (P = 0.0098), higher nuclear grade (P < 0.0001), ER− and PR− status (P = 0.0016 and P = 0.0105) and HER2+ status (P = 0.0434). These features are similar to those of the tumors with LOH exclusively at BRCA1 locus. Tumors with LOH at both BRCA1 and BRCA2 loci were more frequently of the HR−/HER2− subtype and less frequently of the HR+/HER2− subtype (P = 0.0003) (Table 3, group A vs. D).
Table 3

Relationships between LOH exclusively at BRCA1 locus, LOH exclusively at BRCA2 locus, or LOH at both loci and clinicopathologic features in breast cancera

Factor

 

A

B

 

C

 

D

 

BRCA1

ROH

LOH

A vs. B

ROH

A vs. C

LOH

A vs. D

BRCA2

ROH

ROH

P

LOH

P

LOH

P

 

n = 79 (47.6%)

n = 35 (21.1%)

 

n = 18 (10.8%)

 

n = 34 (20.5%)

 

Age

<50 y

30 (38.5%)

14 (41.4%)

0.6971

12 (66.7%)

0.0297

9 (27.3%)

0.2532

≥50 y

48 (61.5%)

19 (57.6%)

 

6 (33.3%)

 

24 (72.3%)

 

Mean ± SE

54.9 ± 1.3

55.4 ± 2.2

0.9559

47.1 ± 2.6

0.0096

59.9 ± 2.3

0.059

Tumor size (pT)

1

29 (36.7%)

6 (17.1%)

0.1549

7 (38.9%)

0.4701

12 (35.3%)

0.9902

2

41 (51.9%)

22 (61.9%)

 

8 (44.4%)

 

18 (52.9%)

 

3

6 (7.6%)

4 (11.4%)

 

3 (16.7%)

 

3 (8.8%)

 

4

3 (3.8%)

3 (8.6%)

 

0 (0.0%)

 

1 (3.0%)

 

Nodal status

Negative

54 (69.2%)

22 (62.9%)

0.5068

9 (52.9%)

0.2064

18 (52.9%)

0.1011

Positive

24 (30.8%)

13 (27.1%)

 

8 (47.1%)

 

16 (47.1%)

 

ly

Negative

54 (77.1%)

20 (57.1%)

0.0367

10 (58.8%)

0.1374

17 (51.5%)

0.0098

Positive

16 (22.9%)

15 (42.9%)

 

7 (41.2%)

 

16 (48.5%)

 

v

Negative

63 (90.0%)

31 (88.6%)

0.8228

16 (94.1%)

0.58

27 (81.8%)

0.2551

Positive

7 (10.0%)

4 (11.4%)

 

1 (5.9%)

 

6 (18.2%)

 

Nuclear grade

1

49 (62.8%)

8 (23.5%)

0.0004

9 (52.9%)

0.2084

5 (15.2%)

<0.0001

2

9 (11.5%)

6 (17.7%)

 

5 (29.4%)

 

7 (21.2%)

 

3

20 (25.6%)

20 (58.8%)

 

3 (17.7%)

 

21 (63.6%)

 

ER

Negative

14 (17.7%)

13 (37.1%)

0.0281

4 (22.2%)

0.6633

16 (47.1%)

0.0016

Positive

65 (82.3%)

22 (62.9%)

 

14 (77.8%)

 

18 (52.9%)

 

PR

Negative

26 (32.9%)

15 (42.9%)

0.3104

5 (27.8%)

0.6704

20 (58.8%)

0.0105

Positive

53 (67.1%)

20 (57.1%)

 

13 (72.2%)

 

14 (41.2%)

 

HER2

Negative

67 (84.8%)

21 (60.0%)

0.0047

16 (88.9%)

0.6478

23 (67.7%)

0.0434

Positive

12 (15.2%)

14 (40.0%)

 

2 (11.1%)

 

11 (32.3%)

 

Subtype

HR+, HER2−

61 (77.2%)

18 (51.4%)

0.021

14 (77.8%)

0.4121

12 (35.3%)

0.0003

HR+, HER2+

6 (7.6%)

4 (11.4%)

 

2 (11.1%)

 

6 (17.6%)

 

HR−, HER2+

6 (7.6%)

10 (28.6%)

 

0 (0.0%)

 

5 (14.7%)

 

HR−, HER2−

6 (7.6%)

3 (8.6%)

 

2 (11.1%)

 

11 (32.4%)

 

ly lymph vessel invasion, v vascular invasion

aClinicopathologic features were compared between tumors with ROH at both BRCA1 and BRCA2 loci (A) and tumors with LOH exclusively at BRCA1 locus (B), between tumors with ROH both BRCA1 and BRCA2 loci (A) and tumors with LOH exclusively at BRCA2 locus, (C), and between tumors with ROH at both BRCA1 and BRCA2 locus (A) and tumors with LOH at both BRCA1 and BRCA2 (D)

Prognosis of Patients with Tumors with LOH at BRCA1 or BRCA2 Locus Alone and LOH at Both BRCA1 and BRCA2 Loci

Next we investigated the associations between LOH exclusively at BRCA1 locus, LOH exclusively at BRCA2 locus and LOH at both BRCA1 and BRCA2 loci with prognosis. LOH at both BRCA1 and BRCA2 loci and LOH exclusively at BRCA1 locus showed similar negative associations with prognosis in terms of DFS, DMFS, and OS. On the other hand, LOH exclusively at BRCA2 locus did not significantly affect prognosis as compared with that of patients with ROH at both BRCA1 and BRCA2 loci (Fig. 2).
https://static-content.springer.com/image/art%3A10.1245%2Fs10434-011-2166-5/MediaObjects/10434_2011_2166_Fig2_HTML.gif
Fig. 2

Associations between LOH at BRCA1 or BRCA2 locus alone and LOH at both BRCA1 and BRCA2 loci with prognosis. LOH at both BRCA1 and BRCA2 loci, and LOH exclusively at BRCA1 locus showed similar and significant negative associations with prognosis. There was no difference in the prognosis of patients with LOH exclusively at BRCA2 locus compared with that of patients with ROH at both BRCA1 and BRCA2 loci. a DFS. b DMFS. c OS

Multivariate Analyses for DFS and OS

Multivariate Cox hazard analyses with models including prognostic factors identified in univariate analyses (tumor size, nodal status, nuclear grade and LOH at BRCA1 locus) revealed that LOH at BRCA1 locus was independently associated with poor prognosis relative to ROH at BRCA1 locus in terms of DFS (hazard ratio 3.08, 95% CI 1.58–6.18, P = 0.0009) and DMFS (hazard ratio 5.18, 95% CI 2.35–12.19, P < 0.0001). In models including tumor size, lymph node metastasis, nuclear grade, ER status and LOH at BRCA1 locus, LOH at BRCA1 locus (vs. ROH at BRCA1 locus) was also an independent prognostic factor for OS (hazard ratio 4.97, 95% CI 1.84–15.1, P = 0.0013) (Table 4).
Table 4

Multivariate analysis of factors predicting DFS, DMFS, and OS

Variable

Evaluation

Hazard ratio (95% CI)

P

DFS

 Tumor size (pT)

1, 2

1

0.2039

3, 4

1.63 (0.75–3.24)

 

 Nodal status

Negative

1

0.0011

Positive

2.73 (1.49–5.06)

 

 Nuclear grade

1, 2

1

0.5369

3

1.23 (0.63–2.43)

 

 BRCA1

ROH

1

0.0009

LOH

3.08 (1.58–6.18)

 

DMFS

 Tumor size (pT)

1, 2

1

0.2486

3, 4

1.64 (0.69–3.53)

 

 Nodal status

Negative

1

0.0026

Positive

2.86 (1.45–5.80)

 

 Nuclear grade

1, 2

1

0.7911

3

0.90 (0.42–1.96)

 

 BRCA1

ROH

1

<0.0001

LOH

5.18 (2.35–12.19)

 

OS

 Tumor size (pT)

1, 2

1

0.0619

3, 4

2.40 (0.95–5.58)

 

 Nodal status

Negative

1

0.0077

Positive

3.07 (1.34–7.44)

 

 Nuclear grade

1, 2

1

0.5808

3

0.76 (0.29–2.08)

 

 ER

Positive

1

0.0545

Negative

2.35 (0.98–5.79)

 

 BRCA1

ROH

1

0.0013

LOH

4.97 (1.84–15.1)

 

Discussion

LOH at the BRCA1 and BRCA2 loci is common in sporadic breast cancers, which possibly attributes the tumorigenesis of sporadic breast cancers.2,10 There are many reports regarding the LOH at BRCA1 and BRCA2 loci in sporadic breast cancers, but the biologic consequences of such LOH are not clear. In this study, we determined the presence of LOH at BRCA1 and BRCA2 loci simultaneously by means of our system for microsatellite analysis, high-resolution fluorescent microsatellite analysis.25 We found that LOH at BRCA1 locus was associated with aggressive features of breast cancer and was independent poor prognostic marker.

LOH at BRCA1 locus, regardless of LOH at BRCA2 locus, was significantly associated with higher tumor grade (grade 3), ER− and PR− status, and HER2+ status (Tables 2 and 3). LOH at BRCA1 locus was also associated with poor prognosis and was an independent prognostic factor for DFS, DMFS and OS (Fig. 1; Table 4). Tumors with LOH exclusively at BRCA1 and tumors with LOH at both BRCA1 and BRCA2 loci had many features in common (Table 3). These findings are consistent with those of an earlier report.18 Moreover, the prognosis of patients with LOH at both BRCA1 and BRCA2 loci, and of those with LOH exclusively at BRCA1 locus was similar and significantly worse than that of patients with LOH exclusively at BRCA2 locus or ROH in both gene loci (Fig. 2). BRCA1 is a master regulator of genome integrity playing important roles in the DNA damage response, including cell cycle checkpoint control and DNA repair.31 Failure of BRCA1 function leads to genomic instability occurring via several pathways resulting in gross chromosomal rearrangements, chromosomal missegregation and aneuploidy.6 It is possible that LOH at BRCA1 locus induces severe chromosomal instability, resulting in gross genetic changes such as LOH of various regions that may include BRCA2. In our study, tumors with LOH at BRCA1 loci were of a higher nuclear grade, indicating a more rapidly dividing aggressive tumor, and more likely ER− and PR−, similar to familial BRCA1 cancers. These observations indicate that LOH at BRCA1 locus might induce the reduction of BRCA1 function.

On the other hand, tumors with LOH exclusively at BRCA2 locus were not associated with any clinicopathologic features and prognosis except for younger age (Table 3; Fig. 1). These data differed from previous studies, which showed a negative impact of LOH at BRCA2 locus on prognosis.22,24 These studies could not exclude a possible effect of LOH at BRCA1 locus because they did not evaluate it.24 As shown in Table 3 and Fig. 2, the clinicopathologic characteristics and the prognosis of tumors with LOH exclusively at BRCA2 locus are similar to those of tumors with ROH at both BRCA1 and BRCA2 loci, most of which were HR+/HER2−. This reminds us that BRCA2-associated cancers are related to the luminal subtype and not significantly different from sporadic tumors.32 Intriguingly, the age of the patients with tumors with LOH exclusively at BRCA2 locus was significantly younger (mean 47.1 ± 2.6 years) than that of patients with tumors with ROH at both BRCA1 and BRCA2 loci (mean 54.9 ± 1.3 years) (Table 3). Several basic studies have reported that disruption of BRCA2 increases the spontaneous mutation rates.33,34 We speculate that in the mammary epithelium with LOH at BRCA2 locus, the activity of BRCA2 might be reduced and somatic mutations may occur in many other genes, including tumor suppressor genes, which promote tumorigenesis at a relatively young age.

In conclusion, tumors with LOH at BRCA1 locus have quite different clinicopathologic features from those of tumors with LOH exclusively at BRCA2 locus or ROH at both loci. LOH at BRCA1 locus is an independent prognostic marker for breast cancer and it might be useful as a molecular marker of prognosis and predict unfavorable outcomes. These findings should be verified in a larger patient population in combination with basic research.

Acknowledgment

We are grateful to M. Makikusa and T. Shishino, M. Kiyota, and Y. Kubota for their valuable technical assistance. This study was supported by grants from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

Copyright information

© Society of Surgical Oncology 2011