This article is mentioned in:

Introduction

Triple-negative breast cancer (TNBC) is defined as a breast carcinoma without estrogen receptor (ER) and progesterone receptor (PR) immunoexpression, and with a lack of amplification of human epidermal growth factor receptor-2 (HER2). The disease accounts for ~15% of all breast carcinomas (1–3). TNBC shows biological aggressiveness and a higher recurrence rate, with no benefit from endocrine or HER2-targeted therapies (1–3). A number of studies have previously been performed to identify additional prognostic markers to better classify TNBC and stratify it into subgroups with different clinical courses (3–8).

Lactoferrin (LF), an iron binding 78- to 80-kDa glycoprotein usually present in mammalian milk, has been immunohistochemically revealed in numerous human neoplastic conditions of different sites (9–22). An immunopositive LF rate with a large variability, ranging from 7.5 to 42% of cases, has previously been identified in breast carcinoma (23–27); however, LF was more often observed in low-grade ER/PR-positive ductal carcinomas, confirming a decrease in LF immunostaining in less differentiated and more aggressive breast carcinomas (25–27). Moreover, it has been hypothesized that increased LF levels LF may be associated with reduced ERα and PR expression, and possibly reduced HER2 expression, and could therefore contribute to TNBC phenotype development (24). Consequently, taking into consideration the downregulation of ER, PR and HER2 at post-transcriptional levels in TNBC cell lines (24), the present study analyzed the immunohistochemical distribution of LF in a cohort of human surgical TNBC patients.

Patients and methods

Patient cohort

The present study retrospectively investigated LF immunoexpression in a cohort of 65 TNBC cases that were surgically treated by breast-conserving surgery (lumpectomy, quadrantectomy, partial mastectomy or segmental mastectomy), at the polyclinic ‘G. Martino’ of Messina, as well as at the Humanitas Oncology Center of Catania, between January 2001 and June 2015, and were not previously subjected to any neoadjuvant treatment. All female patients (mean age, 59.9 years; range, 38–82 years) were diagnosed with TNBC, as 0% ER and PR cell staining was recorded upon immunohistochemical analysis, as well as a HER2 staining score of 0% upon immunohistochemical analysis or a score of 1+ and 2+ with no gene amplification, as verified by fluorescence in situ hybridization (6). For all cases, clinicopathological parameters, including age, grading and tumor stage, were recorded according to international guidelines. Data on the follow-up, including the recurrence of the disease, were available for all patients, with the exception of 3 who were lost to follow-up. The study was conducted in accordance with Good Clinical Practice guidelines and the Declaration of Helsinki, and was approved by the Local Ethics Committees of Polyclinic ‘G. Martino’ (Messina, Italy) and Humanitas Oncology Center (Catania, Italy).

Immunohistochemical methods

All surgical samples were fixed in 10% neutral formalin for 24–36 h at room temperature, and then embedded in paraffin at 56°C. From each tissue block, 4-µm sections were stained with hematoxylin and eosin (H&E) for microscopic examination. Parallel sections were cut and mounted on silane-coated glasses, then dewaxed in xylene and rehydrated in graded ethanols. Antigen retrieval was performed prior to the addition of the primary antibody lactoferrin [clone 1A1; dilution, 1:75; catalog no., K99172B; Biodesign International, Inc., Saco, ME, USA] by heating slides placed in 0.01 M citrate buffer at pH 6.0 in a microwave oven (750 W) for 3 cycles of 5 min each.

Immunohistochemical procedures, and positive and negative controls of LF staining were performed as previously described (18,19).

Immunohistochemical quantification

The analysis of immunostained sections was estimated by light microscopy using 20X and 40X objective lenses, and a ×10 eyepiece. Two pathologists used a double-headed microscope to perform the assessment of LF-immunostained sections on a consensus basis. The percentage of stained neoplastic cells [area of staining positivity (ASP)] was graded as follows: 0) no staining; i) >0 to 5% staining; ii) >5 to 50% staining; and iii) >50% staining. In addition, the intensity of staining (IS) (1, weak; 2, moderate; and 3, strong) was also taken into consideration. Successively, a LF intensity distribution (ID) score was calculated for each case by multiplying the ASP and IS values, as previously described (28,29); using this method, only cases showing an ID score of >2 were considered as positive for LF.

Moreover, data concerning the Ki-67 labeling index (LI) status were also available and had been evaluated by counting the percentage of positive nuclei per 1,000 malignant cells in up to 10 fields representative of the whole neoplastic portions. The median MIB-1 staining score value (20%) was utilized as a cut-off point to determine low and high Ki-67 expression, as described previously (30); this mean value corresponded to the value indicated by the majority of the St. Gallen Breast Cancer Panel (31).

Statistical analysis

The statistical association between LF immunoexpression and the various clinicopathological parameters was investigated using the χ2 test or Fisher's exact test, as appropriate.

Disease-free survival (DFS) and cancer-specific survival (CSS) were assessed by the Kaplan-Meier method, with the date of primary surgery as the entry data. The end point for the DFS analysis was disease progression. CSS was characterized as the length of survival to mortality due to TNBC or to the last follow-up date. Patients who succumbed as a result of diseases independent from TNBC were censored. The Mantel-Cox log-rank test was applied to assess the strength of the association between DFS or CSS and each of the parameters [age, histological grade, pathological tumor (pT) stage, pathological node (pN) stage (6)], Ki-67 LI and LF immunoexpression) as a single variable. Successively, a multivariate analysis (Cox regression model) with stepwise method was utilized to determine the independent effect of each variable on survival.

P<0.05 was considered to indicate a statistically significant difference. Data were analyzed using the SPSS package version 6.1.3 (SPSS, Inc., Chicago, IL, USA).

Results

The routinely stained H&E sections exhibited good morphology, confirming the histopathological diagnosis of ductal invasive breast carcinoma (BC) in all cases. The immunoistochemical confirmed the diagnosis of TNBC.

Clinicopathological and immunohistochemical data for LF for the 65 available TNBC cases analyzed are shown in Table I. The follow-up time ranged from 3 to 112 months (mean, 54.5 months).

Table I. Associations between clinicopathological characteristics and LF expression.

Table I.

Associations between clinicopathological characteristics and LF expression.

	LF immunoexpression
Parameter	Negative	Positive	P-value
Age			1.000
≤70 years	26	9
>70 years	25	8
pT			0.640
1	27	10
2	19	4
3	1	1
4	4	2
pN
N0	30	8
N+	21	9
Histological grade			0.410
2	20	6
3	31	11	1.000
Ki-67 LI			0.040a
Low (≤20%)	20	12
High (>20%)	31	5
Recurrence			0.010a
Absent	19	13
Present	29	4
Status			0.020a
Alive/SID	23	14
SD	25	3

a P<0.05. LF, lactoferrin; SD, succumbed to disease; SID, succumbed independently of disease; LI, labeling index.

LF immunostaining was mainly localized in the cytoplasm of neoplastic elements and occasionally in the nucleus of the same cells; 17 cases (26.15%) exhibited an ID score of >2 and were therefore considered as positive for LF (Fig. 1). Statistical analyses showed that LF positivity was significantly associated with a low Ki-67 LI (<20%; P=0.040), the absence of recurrence (P=0.010) and alive status (P=0.020).

Figure 1. (A) Evident LF immunoreactivity in TNBC, with strongly and slightly positive tumor cells in direct contact (original magnification, ×400). (B) Uniform granular cytoplasmic LF immunostaining (original magnification, ×320). Immunoperoxidase, with Mayer's hematoxylin counterstain.

Univariate analysis showed that a high pT stage (P=0.040) and the absence of LF immunoexpression (P=0.030) were significantly associated with shorter DFS times (Table II). Multivariate analysis for DFS demonstrated that patient age, pT stage and Ki-67 LI were independent variables (Table I).

Table II. Associations between LF and recurrence.

Table II.

Associations between LF and recurrence.

				Univariate		Multivariate
Variables	Patients, n	Disease progression, n (%)	P-value	HR (95% CI)	P-value	HR (95% CI)	P-value
Age, years			0.460		0.120		0.02a
≤70	33	15 (45)		1.0		1.0
>70	32	18 (56)		1.6 (0.8–3.8)		2,3 (1.1–4.1)
pT			0.250		0.040a		0.001a
1	36	15 (42)		1.0		1.0
2	21	12 (57)		1.3 (0.6–2.6)		1.0
3	2	1 (50)		1.6 (0.1–15.1)		1.0
4	6	5 (83)		3.8 (0.7–18.7)		5.9 (1.9–17.9)
pN			0.800		0.250
N0	35	17 (49)		1.0
N+	30	16 (53)		1.4 (0.7–2.7)		NSS
Histological grade			0.310		0.150
2	24	10 (42)		1.0
3	41	23 (56)		1.6 (0.8–3.8)		NSS
Ki-67 LI			0.080		0.150		0.007a
Low (≤20%)	31	12 (39)		1.0		0.3 (0.1–0.1)
High (>20%)	34	21 (62)		1.6 (0.8–3.8)		1.0
LF expression			0.010a		0.030a
Absent	48	29 (60)		2.9 (1.2–6.2)
Present	17	4 (24)		1.0		NSS

a P<0.05. LI, labeling index; LF, lactoferrin; pT, pathological tumor; pN, pathological node; HR, hazard ratio; CI, confidence interval, NSS, not statistically significant.

Univariate analysis for CSS showed that an age >70 years (P=0.007) and a high pT stage (P=0.030) were significant negative prognostic factors (Table III). The absence of LF immunoexpression was not associated with shorter CSS times, since statistical significance was not reached (P=0.060; Table III). At multivariate analysis, patient age (P=0.004), pT stage (P=0.008) and Ki-67 LI (P=0.040) emerged as independent variables (Table III).

Table III. Associations between LF and survival.

Table III.

Associations between LF and survival.

				Univariate		Multivariate
Variables	Patients, n	Succumbed to disease, n (%)	P-value	HR (95% CI)	P-value	HR (95% CI)	P-value
Age, years			0.120		0.007a		0.004a
≤70	33	10 (30)		1.0		1.0
>70	32	18 (56)		2.6 (1.2–5.2)		3.3 (1.4–7.4)
pT			0.070		0.030a		0.008a
1	36	11 (31)		1.0		1.0
2	21	11 (52)		1.7 (0.7–3.7)		1.0
3	2	1 (50)		2.1 (0.1–24.1)		1.0
4	6	5 (83)		4.3 (0.8–20.8)		4.4 (1.4–13.4)
pN			0.620		0.340
N0	35	14 (40)		1.0
N+	30	14 (47)		1.4 (0.6–3.6)
Histological grade			0.300		0.120
2	24	8 (33)		1.0
3	41	20 (49)		1.8 (0.8–3.8)
Ki-67 LI			0.150		0.610		0.04a
Low (≤20%)	31	12 (39)		1.0		0.3 (0.1–0.1)
High (>20%)	34	16 (47)		1.2 (0.5–2.5)		1.0
LF expression			0.020a		0.060a
Absent	48	25 (52)		2.9 (1.2–6.2)
Present	17	3 (18)		1.0

a P<0.05. LI, labeling index; LF, lactoferrin; pT, pathological tumor; pN, pathological node; HR, hazard ratio; CI, confidence interval.

Discussion

It is well known that lactating breast tissue, as well as ductules and intralobular duct epithelial cells in normal and dysplastic tissue, strongly stain for LF (25); consequently, LF has been considered a potential marker for glandular or acinar differentiation, similar to its use in other malignancies (9,18,22). However, in breast cancer, the association between LF expression and clinical parameters is not well defined (24), even if it has been hypothesized that tumors with high LF expression retain a certain degree of physiological control over cell growth, which may explain a good prognosis (25).

In the present TNBC series, an LF immunostaining rate (ID score) of >2 was found in ~26.15% of cases, and was significantly associated with a low Ki-67 LI (<20%), the absence of recurrence and an alive status. Therefore, given the aggressive course of TNBC, the presence of LF-immunopositive cases may identify a more indolent subtype of the disease with peculiar clinical characteristics, including less aggressive biological behavior and a more favorable prognosis; by contrast, shorter DFS and CSS times were significantly associated with the absence of LF immunoexpression.

However, a similar capability to select a TNBC subgroup with low biological aggressiveness has been attributed to androgen (AR) expression (6,7); in particular, AR has been shown to be a favorable prognostic factor of DFS and overall survival, with significantly decreased recurrence and mortality risks (6,7,32). Furthermore, in a series of 105 TNBC patients with stage II or III disease treated with neoadjuvant chemotherapy based on docetaxel and doxorubicin, the prognostic and the predictive role of Ki-67 has been analyzed, identifying two distinct subgroups of TNBC with different Ki-67 expression, responses and prognoses (33). In the present study, Ki-67 emerged as an independent variable for DFS and CSS in TNBC, further highlighting its prognostic value. In addition, TNBC LF-positive cases always expressed lower levels of Ki-67, displaying a significant inverse association between Ki-67 LI and LF immunoreactivity, thus representing a further element to assess a favorable group of TNBC characterized by a low risk of recurrence and a better prognosis.

In conclusion, the results of the present study appear to be notable with regard to the field of TNBC, although further validation in large prospective studies is required to enable LF to be a promising biomarker. In fact, together with previous studies, the results of the present study show that the first immunohistochemical application of LF in TNBC appears to favor the selection of patients with a less aggressive behavior, particularly in combination with Ki-67 status and AR expression. However, additional investigations are also required with regard to the potential of LF in cancer treatment, due to its nutraceutical function and its ability to potentiate chemotherapy.

References