Expression of growth factor and chemokine receptors: new insights in the biology of inflammatory breast cancer

Abstract

Purpose: Recent studies have indicated that expression of chemokine receptors CXCR4 and CCR7 could be an indicator of the metastatic potential of breast cancer. Expression of CXCR4 and CCR7 along with the biomarkers HER2-neu and epidermal growth factor receptor (EGFR) was investigated in inflammatory breast cancer (IBC) to evaluate their prognostic implications.

Experimental design: CXCR4, CCR7, and EGFR were evaluated by immunohistochemical staining (IHC) of paraffin-embedded tissue sections. HER2-neu amplification was assessed by FISH and/or IHC. All patients received chemotherapy, surgery, and radiation.

Results: Forty-four cases diagnosed with IBC from 1994 to 2002 were included in the study. In all, 18 (40.9%) patients had positive CXCR4, 10 (22.7%) had positive CCR7, 21 (47.7%) had positive HER2-neu, and EGFR was positive in 12 of 40 patients (30%). The 5-year overall survival (OS) was 24.8% for CXCR4-positive disease versus 42.3% for CXCR4-negative patients (P = 0.53) and 20.0% for CCR7-positive disease versus 41.9% for CCR7-negative patients (P = 0.24). EGFR-positive disease had significantly worse OS compared with EGFR-negative disease (P = 0.01).

Conclusions: These data demonstrate the expression of growth factor and chemokine receptors in IBC. The expression of these receptors is associated with increased risk of recurrence and death, and thus, they may represent potential therapeutic targets in IBC.

CCR7, CXCR4, EGFR, HER2-neu, inflammatory breast cancer

introduction

Inflammatory breast cancer (IBC) is the most aggressive form of primary breast carcinoma and is associated with peculiar clinical and biological features and with a dismal outcome despite multimodality treatment approaches [1]. The incidence of IBC has been reported to be 0.7 cases per 100 000 person-years by the National Cancer Institute's Surveillance, Epidemiology, and End Results Program [2]. Studies comparing outcomes in patients with locally advanced breast cancer associated with clinical features of IBC versus those in patients with locally advanced breast cancer without features of IBC demonstrated better outcomes for the latter and a peculiar pattern of recurrence for IBC [2–5]. Despite combined multimodality treatment comprising chemotherapy, surgery, and radiation, the prognosis of IBC is poor; the 10-year disease-free survival rate is 20%–25% [1, 5, 6]. These data clearly indicate that current treatment modalities are inadequate and that a better understanding of the biological features of the disease is necessary if more effective interventions are to be developed.

IBC is characterized by the presence of high histological grade; high proliferation rate (e.g. elevated MIB1 expression, high S-phase proportion, or high thymidine labelling index); aneuploidy; and high levels of expression of p53 and MUC1, RhoC, E-cadherin, and growth factor receptors, such as HER2-neu and epidermal growth factor receptor (EGFR) [7–16]. Furthermore, IBC has been found to be frequently associated with a ‘basal-like’ phenotype; increased intratumoral microvessel density; and tumor angiogenesis- and lymphangiogenesis-related factors such as vascular endothelial growth factor (VEGF)-A, VEGF-C, and VEGF-D, Flt-1, KDR, Tie-1, and Tie-2 [17–19].

Recent studies have indicated that interactions between chemokine receptors in breast cancer cells and their ligands in host organs play a role in malignant dissemination and progression [20–28]. The CXCR4 ligand CXCL12/SDF-1α is expressed in liver, bone marrow, lung, and lymph nodes, whereas the CCR7 ligand CCL21 is highly expressed in lymph nodes of breast cancer patients as shown by quantitative RT-PCR analysis [20]. The chemokine receptors CXCR4 and CCR7 are expressed in some human breast cancer cells representing both primary breast tumors and metastases [20, 21]. Furthermore, CXCR4 has been identified as a key mediator in neoangiogenesis; its effect is mediated by up-regulating VEGF [29]. Similarly, VEGF has been shown to increase CXCR4 expression, forming a positive feedback mechanism to promote angiogenesis [30]. Moreover, we recently showed cross-talk between CXCR4 and HER2-neu/EGFR via HER2-neu and EGFR transactivation by the SDF-1α–CXCR4 axis [31]. CCR7 has been reported to play a key role in T-cell differentiation, and its presence on the cell surface divides memory T cells into two subsets: effector memory (CCR7⁻) and central memory (CCR7⁺) T cells, which have different immune functions in secondary immune responses [32]. CCR7 also has been associated with increased propensity for metastatic spread to lymph nodes [26, 33–36].

We postulated that, for a subset of patients with IBC, the expression of growth factor receptors may be associated with adverse features. Furthermore, we hypothesized that the expression of chemokine receptors could contribute to the peculiar clinical features of IBC, such as dermal lymphatic invasion, early locoregional spread, and subsequent distant metastasis. In the present study, therefore, we investigated the expression of CXCR4 and CCR7, along with the biomarkers HER2-neu and EGFR, in primary IBC to evaluate their expression patterns and prognostic implications.

materials and methods

patients and therapy

Forty-four patients with IBC (stage III disease) who were treated at The University of Texas M. D. Anderson Cancer Center from August 1994 to April 2002 were included in this analysis. Sections of paraffin-embedded tissue samples, core needle or incisional biopsies at the time of diagnosis (n = 26) or of mastectomy specimen after primary chemotherapy (n = 18), were provided by the Breast Tumor Bank of the M. D. Anderson Cancer Center after study approval by the Institutional Review Board. All patients had undergone pre-treatment diagnostic biopsy. The diagnosis of IBC was on the basis of clinical signs such as diffuse erythema, peau d'orange, tenderness, induration, and warmth, with or without evidence of extensive dermal lymphatic invasion on pathologic evaluation. Clinical evolution of <3 months was also required. Modified Black's nuclear grading system was used to evaluate the histologic grade of the tumors. Immunohistochemical staining (IHC) for hormonal receptor status was carried out using the modified avidin–biotin complex method in a DAKO Autostainer (DAKO, Carpinteria, CA) using primary antibodies against estrogen receptor (ER) α (ER, clone: 6F11; Novocastra, Newcastle upon Tyne, UK, 1 : 50) and progesterone receptor (PgR) (PgR Ab-9, clone: 1A6; Neomarker/Labvision Corporation, Fremont, CA, 1 : 30).

All patients with primary IBC received multimodal treatment, including chemotherapy, surgery, and radiation therapy. Patients first received pre-operative chemotherapy according to the institutional protocols during the study period [1]. Induction chemotherapy consisted of four to six cycles of an anthracycline-based regimen that included doxorubicin (500 mg/m²), cyclophosphamide (500 mg/m²), and 5-fluorouracil (500 mg/m²) every 21 days. Patients whose disease was in objective remission after chemotherapy underwent modified radical mastectomy. The majority of the patients (89%) also received paclitaxel after surgery, followed by adjuvant radiation therapy. Tamoxifen was given to patients with ER- and/or PgR-positive tumors following radiation therapy.

immunohistochemical analysis of CXCR4, CCR7, and EGFR

IHC by the avidin–biotin complex method was used to assess CCR7, CXCR4, and EGFR. Briefly, 4-μm tissue sections were deparaffinized and treated with 0.3% hydrogen peroxide [in phosphate-buffered saline solution (PBS) for CXCR4 or methanol for CCR7 and EGFR] to block endogenous peroxidase activity. For EGFR staining, the slides were pre-treated with 0.02% Protease XXIV (Sigma, St Louis, MO) for 2 minutes. For CXCR4 and CCR7 staining, non-specific binding was blocked by incubation in protein-blocking solution containing 5% normal horse serum and 1% normal goat serum in PBS for 20 minutes at room temperature. Sections were incubated with primary antibodies for CXCR4 (44717.111, IgG_2b, R&D Systems, Minneapolis, MN) at 1 : 150 dilution and for CCR7 (2H4, IgM, BD Biosciences, San Diego, CA) at 1 : 100 dilution, both for 18 hours at 4°C, and with EGFR mouse mAb at 1 : 50 dilution (clone: 31G7; Zymed, South San Francisco, CA) for 1 hour at room temperature. Secondary antibodies were rat anti-mouse-IgG_2b-HRP (Serotec Inc., Raleigh, NC) for CXCR4 and goat-anti-mouse-IgM-HRP (Jackson ImmunoResearch Laboratories Inc., West Grove, PA) for CCR7. For EGFR staining, the Envison+ method was employed on a Dako Autostainer instrument for the rest of the procedure according to the manufacturer's instructions. Color was developed by using diaminobenzidine as a chromagen. Sections were then counterstained with hematoxylin.

The intensity, staining percentage, and pattern of staining (nuclear, cytoplasmic, or membranous) were noted for CXCR4 and CCR7 as described elsewhere [26]. Briefly, the intensity was scored as low, moderate, or strong compared with the background staining. Positivity (high expression) was defined as >50% of tumor cells showing staining according to the following patterns: (i) moderate or strong cytoplasmic staining for CXCR4 and (ii) strong cytoplasmic staining for CCR7. For EGFR, any complete membranous staining was considered positive regardless of the intensity of the staining, in concordance with previous studies [37].

assessment of HER2-neu

HER2-neu amplification was determined either by FISH (n = 25) by using tissue array samples, as described elsewhere [38], or by estimating the overexpression of HER2-neu by IHC using whole slide samples for those missing cases among tissue array samples (n = 16) that could not be studied by FISH. The PathVysion™ HER2-neu DNA probe kit (Vysis, Downers Grove, IL) was used for FISH. Briefly, a locus-specific indicator HER2-neu SpectrumOrange/CEP17 SpectrumGreen (Vysis) was applied to the slide region of interest, and slides were placed on a pre-programmed, humidified slide warmer (Hybrite; Vysis) with the following settings: denaturation at 73°C for 5 minutes and hybridization at 37°C for 16 hours. The slides were incubated and immersed in pre-warmed 2× standard saline citrate buffer with 0.3% Nonidet P-40. They were air dried in the dark and counterstained with 4,6-diamidino-2-phenylindole dihydrochloride (DAPI) (Vysis). The signals were enumerated using an Olympus AX70 epifluorescence microscope (Olympus, Melville, NY) fitted with a SpectrumOrange, SpectrumGreen, and DAPI triple-filter set. At least 60 cells were scored, and the copy numbers of HER2-neu and CEP17 were recorded for each cell as signal counts. HER2-neu gene amplification was defined as an HER2-neu : CEP17 signal ratio ≥ 2.0.

For IHC, antigen retrieval was carried out by steaming the tissue sections in 10 mM citrate buffer (pH 6.0) for 45 minutes before incubation with the primary antibody (clone: AB8; Neomarker/Labvision Corporation), at 1 : 300 dilution at room temperature. HER2-neu expression was considered positive if >10% of the tumor cells showed a complete and strong membranous staining (3⁺) by IHC or if FISH showed gene amplification.

statistical analysis

Analyses were carried out using SAS 9.0 (SAS Institute, Cary, NC) and S-Plus 7.0 (Insightful Corporation, Seattle, WA). Patient characteristics were tabulated by CXCR4, CCR7, HER2-neu, and EGFR expression and compared between groups with Fisher's exact test. Disease (breast cancer)-specific survival (DSS) time and overall survival (OS) time were provided. OS or DSS was defined as the interval between the diagnosis and last follow-up. Patients who were alive or had died of a cause other than breast cancer were censored for analysis of DSS, whereas patients who were alive or had died of any cause (other than breast cancer) were censored for analysis of OS. OS was estimated by the Kaplan–Meier method and compared between groups with the log-rank test. Variables that were found to be significant in univariate analyses (P < 0.05) along with those other variables that were investigated as biomarkers in this study were further evaluated in a Cox regression model by using forward logistic regression analysis to explore the independent factors associated with poor prognosis.

The cumulative incidences of breast cancer-specific death and disease recurrence at any site were estimated according to the method described by Gooley et al. [39] and compared between groups using the method described by Pepe and Mori [40]. In the analysis of cumulative incidence of breast cancer-specific death, death from other causes was considered a competing risk. In the analysis of the cumulative incidence of disease recurrence, death before disease recurrence was considered a competing risk. Hormone receptor status was defined as ER and/or PgR positivity versus both negative. The co-expression of CXCR4 with CCR7, HER2-neu, or EGFR was considered as both positive versus both/either negative. P values <0.05 were considered statistically significant; P values are not presented for groups of fewer than 10 patients.

results

Forty-four cases of IBC diagnosed from 1994 to 2002 were included in the study. The patients' median age was 49 years (range, 29–73 years). Patient and tumor characteristics and their associations with biomarkers are shown in Tables 1, 2 and 3.

Table 1.

Patient and tumor characteristics by CXCR4 expression

Patient and tumor characteristics	CXCR4 low	CXCR4 high	P value
	N (%)	N (%)
	26 (59)	18 (41)
Age >50 years
No	15 (57.7)	9 (50)	0.76
Yes	11 (42.3)	9 (50)
Nuclear grade
Intermediate	2 (8.0)	5 (31.3)	0.09
High	23 (92.0)	11 (68.7)
Lymphovascular invasion
No	1 (5.9)	1 (7.1)	0.999
Yes	16 (94.1)	13 (92.9)
Estrogen receptor (ER)
Negative	12 (50.0)	11 (64.7)	0.52
Positive	12 (50.0)	6 (35.3)
Progesterone (PgR)
Negative	17 (68.0)	15 (83.3)	0.31
Positive	8 (32.0)	3 (16.7)
Hormone receptor (either ER or PgR)
Negative	8 (33.3)	9 (52.9)	0.33
Positive	16 (66.7)	8 (47.1)
CCR7
Negative	24 (92.3)	10 (55.6)	0.01
Positive	2 (7.7)	8 (44.4)
HER2-neu
Negative	14 (53.8)	9 (50.0)	0.999
Positive	12 (46.2)	9 (50.0)
EGFR
Negative	19 (82.6)	9 (52.9)	0.08
Positive	4 (17.4)	8 (47.1)

Table 2.

Patient and tumor characteristics by CCR7 expression

	CCR7 low	CCR7 high	P value
	N (%)	N (%)
	34 (77)	10 (23)
Age >50 years
No	18 (52.9)	6 (60.0)	0.73
Yes	16 (47.1)	4 (40.0)
Nuclear grade
Intermediate	6 (18.2)	1 (12.5)	0.999
High	27 (81.8)	7 (87.5)
Lymphovascular invasion
No	2 (8.7)	0 (0.0)	0.999
Yes	21 (91.3)	8 (100.0)
Estrogen receptor (ER)
Negative	18 (58.1)	5 (50.0)	0.72
Positive	13 (41.9)	5 (50.0)
Progesterone receptor (PgR)
Negative	24 (72.7)	8 (80.0)	0.999
Positive	9 (27.3)	2 (20.0)
Hormon receptor (ER or PgR)
Negative	13 (41.9)	4 (40.0)	0.999
Positive	18 (58.1)	6 (60.0)
HER2-neu
Negative	16 (47.1)	7 (70.0)	0.29
Positive	18 (52.9)	3 (30.0)
EGFR
Negative	23 (76.7)	5 (50.0)	0.13
Positive	7 (23.3)	5 (50.0)
CXCR4
Negative	24 (70.6)	2 (20.0)	0.01
Positive	10 (29.4)	8 (80.0)

Table 3.

Patient and tumor characteristics by EGFR expression

	EGFR negative	EGFR positive	P value
	N (%)	N (%)
	28 (70)	12 (30)
Age >50
No	14 (50.0)	7 (58.3)	0.74
Yes	14 (50.0)	5 (41.7)
Nuclear grade
Intermediate	3 (11.5)	3 (27.3)	0.34
High	23 (88.5)	8 (72.7)
Lymphovascular invasion
No	2 (9.5)	0 (0.0)	0.999
Yes	19 (90.5)	8 (100.0)
Estrogen receptor (ER)
Negative	13 (50.0)	6 (54.5)	0.999
Positive	13 (50.0)	5 (45.5)
Progesterone receptor (PgR)
Negative	18 (66.7)	10 (83.3)	0.45
Positive	9 (33.3)	2 (16.7)
Hormone receptor (ER or PgR)
Negative	8 (30.8)	5 (45.5)	0.46
Positive	18 (69.2)	6 (54.5)
CCR7
Negative	23 (82.1)	7 (58.3)	0.13
Positive	5 (17.9)	5 (41.7)
HER2-neu
Negative	11 (39.3)	8 (66.7)	0.17
Positive	17 (60.7)	4 (33.3)
CXCR4
Negative	19 (67.9)	4 (33.3)	0.08
Positive	9 (32.1)	8 (66.7)

histopathology and expression of biomarkers

Hormone receptors (ER and/or PgR) were expressed in 24 patients of 41 patients tested (58.5%). High nuclear grade was reported in 34 patients (83%) (Table 1). Tumors were positive for CXCR4 in 18 (40.9%) patients and for CCR7 in 10 (22.7%) (Figure 1a and b). Expression of CXCR4 was cytoplasmic in 38 tumors (1+, 2+, 3+), combined cytoplasmic and nuclear in 10 tumors, and nuclear alone in four tumors, and no staining of CXCR4 was observed in four cases. Similarly, expression of CCR7 was found to be cytoplasmic in 37 tumors (1+, 2+, 3+), combined cytoplasmic and nuclear in one tumor, and nuclear in one tumor and no staining of CCR7 was observed in seven cases. HER2-neu gene amplification was demonstrated in 14 patients by FISH test, whereas seven tumors showed high HER2-neu expression by IHC (3+). Overall, HER2-neu positivity (overexpression/amplification) was 47.7% (Figure 1c). Furthermore, EGFR was positive in 12 of 40 patients tested (30%, Figure 1d).

Figure 1.

Breast tumor samples with (A) CXCR4 expression with high cytoplasmic staining; (B) CCR7 expression with high cytoplasmic staining; (C) HER2-neu overexpression (3+ by immunohistochemistry); (D) positive expression of EGFR.

Tumors were positive for both CXCR4 and CCR7 in eight (18.2%) patients, for both CXCR4 and HER2-neu in nine (20.5%), and for both CXCR4 and EGFR in eight (18.2%). Patients with CXCR4-positive disease also had CCR7-positive disease more frequently than patients with CXCR4-negative disease (P = 0.01). Moreover, patients with CXCR4-positive disease tended to have EGFR-positive disease more frequently than patients with CXCR4-negative disease (P = 0.08) (Table 1).

outcome and associations with survival analyses

The median follow-up interval was 46.5 months (range, 11–127 months). In all, 32 (73%) of the 44 patients had died and median OS was 47 months. The 5-year OS rate was 36.5 [95% confidence interval (CI) 24.5–54.6]. Patients with EGFR-positive disease had a significantly lower OS rate than patients with EGFR-negative disease (P = 0.01) (Table 4, Figure 2). Patients with EGFR-positive disease also had a higher cumulative incidence of breast cancer-related death than patients with EGFR-negative disease, but this trend did not achieve statistical significance (P = 0.09) (Table 5, Figure 3). Interestingly, eight of the 12 patients (67%) with EGFR-positive disease also had CXCR4-positive disease. Similarly, patients whose tumors co-expressed CXCR4 and EGFR had a lower OS rate (Table 4, Figure 4) and an increased incidence of breast cancer-related death compared with others (Table 5, Figure 5).

Figure 2.

Overall survival in patients with EGFR (+) and (−) inflammatory breast cancer.

Figure 3.

Patients with EGFR-positive disease had also a higher cumulative incidence of breast cancer deaths compared with patients with EGFR-negative disease, but this trend did not achieve statistical significance (P = 0.09).

Figure 4.

Overall survival in patients with CXCR4 with co-expression of EGFR.

Figure 5.

Patients with co-expression of CXCR4 and EGFR had a trend for increased incidence of breast cancer deaths.

Table 4.

Overall survival analyses in patients with inflammatory breast cancer

	N	Median (months)	3-year estimate (%)	95% confidence interval range	5-year estimate (%)	95% confidence interval	P value
All	44	47	63.2	50.3% to 79.3%	36.5	24.5% to 54.6%
CXCR4
Negative	26	47	69.2	53.6% to 89.5%	42.3	27% to 66.3%	0.53
Positive	18	45	54.2	35.1% to 83.6%	24.8	9.9% to 62.1%
CCR7
Negative	34	53	64.1	49.7% to 82.6%	41.9	27.8% to 63%	0.24
Positive	10	45.5	60.0	36.2% to 99.5%	20.0	5.8% to 69.1%
HER2-neu
Negative	23	52	65.2	48.4% to 87.9%	37.9	22.3% to 64.7%	0.79
Positive	21	47	60.3	42.3% to 86%	33.9	18.1% to 63.5%
EGFR
Negative	28	54	67.0	51.4% to 87.2%	48.1	32.4% to 71.3%	0.01
Positive	12	37	50.0	28.4% to 88%	NA	NA
CXCR4 and HER2-neu
Both/either negative	35	47	65.7	51.7% to 83.5%	39.5	26.1% to 59.7%	NA
Both positive	9	56	50.8	25.7% to 100%	16.9	3% to 96.4%
CXCR4 and EGFR
Both/either negative	35	52	65.1	50.9% to 83.2%	44.0	30% to 64.5%	NA
Both positive	8	37	50.0	25% to 100%	NA	NA

NA, not available. Furthermore, P values are not presented for groups of fewer than 10 patients.

Table 5.

Cumulative incidence estimates for breast cancer-specific deaths and disease recurrence in patients with inflammatory breast cancer

	N (patients)	Breast cancer-specific deaths				Recurrence at any site
		N (events)	3 years (%)	5 years (%)	P value	N (events)	3 years (%)	5 years (%)	P value
CXCR4					0.54
Negative	26	18	30.8	53.8		16	64.0	64.0	0.60
Positive	18	12	45.8	67.0		12	61.1	66.7
CCR7					0.21
Negative	34	21	35.9	51.6		20	60.6	60.6	0.15
Positive	10	9	40.0	80.0		8	70.0	80.0
HER2-neu					0.69
Negative	23	17	34.8	62.1		17	69.6	73.9	0.52
Positive	21	13	39.7	55.4		11	55.0	55.0
EGFR					0.09
Negative	28	18	33.0	48.2		17	59.3	63.0	0.16
Positive	12	11	50.0	91.7		10	83.3	83.3
CXCR4 and HER2-neu					NA
Both/either negative	35	25	34.3	57.7		23	64.7	67.6	NA
Both positive	9	5	49.2	66.1		5	55.6	55.6
CXCR4 and EGFR					NA
Both/either negative	35	23	34.9	53.0		22	61.8	64.7	NA
Both positive	8	7	50.0	87.5		6	75.0	75.0

NA, not available. Furthermore, P values are not presented for groups of fewer than 10 patients.

The expression of chemokine receptors was associated with a trend for worse OS rate. The 5-year estimated OS rate for patients with CXCR4-positive disease was 24.8% (95% CI 9.9% to 62.1%), while that for patients with CXCR4-negative disease was 42.3% (95% CI 27% to 66.3%) (Table 4, Figure 6). The 5-year estimated OS rate for patients with CCR7-positive disease was 20.0% (95% CI 5.8% to 69.1%), while that for patients with CCR7-negative disease was 41.9% (95% CI 27.8% to 63%) (Table 4, Figure 7). Furthermore, the 5-year cumulative incidence of recurrence was higher in patients with CCR7-positive disease compared with CCR7-negative disease and in those whose tumors co-expressed CXCR4 and EGFR versus those that did not (Table 5). No statistical significance could be found, however, when variables including EGFR positivity along with CXCR4 and CCR7 positivity were further evaluated in a multivariate Cox regression model to explore the independent variables associated with decreased OS or DSS rates.

Figure 6.

Overall survival in patients with high and low CXCR4 expression.

Figure 7.

Overall survival in patients with high and low CCR7 expression.

discussion

The management of IBC has evolved in the last three decades, resulting in improvements in outcome for this aggressive form of breast cancer [1]. Recent analysis indicates that much of this improvement is related solely to the appropriate use of a multidisciplinary treatment approach and the introduction of effective chemotherapy [41]. In fact, while several retrospective analyses have demonstrated that IBC has peculiar and more aggressive features than non-IBC locally advanced breast cancer, there still is no IBC-specific treatment that could significantly improve the prognosis for these patients [4, 5].

This is the first study to investigate the expression of chemokine receptors (CXCR4 and CCR7) along with the expression of growth factor receptors (EGFR and HER2-neu) and to address whether expression of these biomarkers predicts increases in risk of recurrence and death in IBC. Although some of the studies that compared the IBC phenotype with the non-IBC phenotype demonstrated increased expression of some genes, including chemokines, i.e. CXCL2, CCL3/MIP1A, and CCL5/RANTES in gene expression profiling or real-time RT-PCR analyses [42, 43], studies failed to show an increased expression of chemokine receptors such as CXCR4 and CCR7 in IBC by far. In the current study, we demonstrated high levels of expression of CXCR4, EGFR, and HER2-neu amplification in IBC. The analysis clearly showed that, of the factors tested, EGFR expression is the strongest prognostic factor in IBC [7, 12, 13]. Buchholz et al. [44] similarly found that EGFR-positive disease was associated with a poor survival in patients with locally advanced non-IBC. Other intriguing observations did not reach statistical significance, probably because of the small sample size, but have to be considered in the generation of further hypotheses. For example, the expression of CXCR4 or CCR7 was associated with lower survival rate.

There has been increasing evidence in the literature showing the association between CCR7 expression in different cancer types and lymph node metastases [20, 26, 32–35]. Moreover, Günther et al. [36] reported that only CCR7 was significantly correlated with decreased survival among other chemokine receptors assessed in their study including CXCR4 and CXCR5. In the current study, there was a trend towards a decreased OS rate. In regards to the staining patterns of CCR7, we recently found that high cytoplasmic CCR7 expression in T1 breast cancers (21.5% in lymph node-positive tumors versus 8.5% in lymph node-negative tumors) was associated with lymph node metastases, whereas nuclear CCR7 was exclusively expressed in lymph node-negative tumors [26]. In concordance with these observations and previous studies in other cancer types, we demonstrated here that a significant percentage of tumors (23%) highly expressed cytoplasmic CCR7, whereas a predominantly nuclear expression of CCR7 was exclusively absent in this series of IBC.

In our previous study, we also found an increased high cytoplasmic CXCR4 expression in lymph node-positive tumors, whereas a predominantly nuclear expression of CXCR4 was significantly associated with lymph node negativity [26]. Interestingly, only cytoplasmic staining was exclusively detected in patients with CXCR4 expression similar to the findings with CCR7 expression in our series of IBC, whereas none of the patients showed a predominantly nuclear staining pattern. Spano et al. [45] similarly demonstrated that strong CXCR4 nuclear staining was associated with a better outcome in patients with early-stage non-small-cell lung cancer, indicating different localization patterns of chemokine receptors, nuclear versus cytoplasmic, seem to have different biological significance in cancer cells. Furthermore, high cytoplasmic CXCR4 expression rate was found to be much higher in patients (41%) with IBC compared with our previous studies on earlier stages of breast cancer that reported only 5% of lymph node-negative T1 tumors and 11% of lymph node-positive T1 tumors expressed high cytoplasmic CXCR4 [26]. Consistent with these data, the increased expression of CXCR4 in almost half the IBC samples appears to be linked with the metastatic process since our preliminary results showed that half the primary tumors of patients with stage I–III disease who developed distant organ metastases also were found to express high levels of CXCR4 [46]. Similarly, high expression of cytoplasmic or membranous staining of CXCR4 was associated with more advanced stages, lymph node metastases, and decreased survival rates in patients with colorectal cancer [47], and poor prognosis in primary cutaneous melanoma [48]. Even though high cytoplasmic expression of CXCR4 was also associated with worse survival in IBC in this study, it did not reach the statistical significance, probably again due to the small sample size.

Besides high expression of chemokine receptors, CXCR4 and CCR7, we also showed that half the patients had HER2-neu-positive tumors in concordance with previous studies that reported similar findings [8, 9, 11, 14]. Furthermore, the current study demonstrated the co-expression of CXCR4 and growth factor receptors, particularly HER2-neu and EGFR as published in some recent studies [26, 27, 30, 49]. We previously reported that only 2% of the T1 breast cancers showed expression of both high cytoplasmic CXCR4 and HER2-neu, and this expression phenotype was associated with extensive lymph nodal involvement (≥4) [26]. Interestingly, IBC was found to have this expression phenotype in higher frequencies (20.5%) in this study. Increased expression of growth factors, particularly HER2-neu and EGFR in IBC compared with non-IBC [42] appears to be more specific to the IBC phenotype that might be a useful marker for the targeted therapy when planning the systemic therapy for the patients.

On the basis of these novel data, we can formulate an interesting hypothesis that can support a novel therapeutic approach for IBC. We have recently completed a retrospective analysis demonstrating that patients with locally advanced breast cancer who demonstrate clinical characteristics of IBC have a worse prognosis and different pattern of recurrence than patients who have locally advanced breast cancer without features of IBC [5]. This supports the hypothesis that IBC is relatively chemoresistant and associated with micrometastatic spread at the time of diagnosis, mainly to lymphatics, and also to the bone marrow. This spreading translates to subsequent evidence of recurrence, primarily in soft tissue and bone. It appears that IBC constitutes mostly cells with a basal-like phenotype, and this study confirms these findings by highlighting the prognostic role of the EGFR pathway [12, 13, 19, 50]. These tumors are resistant to chemotherapeutic agents and appear to spread in a non-random but peculiar pattern indicating preferential ‘homing’. This important process takes place early and is driven by the expression of chemokine receptors (frequently co-expressed with growth factor receptors).

In conclusion, this study generated additional support for the concept that novel treatments based on the biological characteristics of IBC are required if the prognosis for patients with IBC is to improve. Whether novel treatment modalities targeting both EGFR [51] and CXCR4 can affect processes of growth, invasion, and metastasis needs to be investigated in future clinical trials.

Part of this study was presented at the American Society of Clinical Oncology Annual Meeting, 1 June 2005, and at the 30th European Society of Medical Oncology Annual Meeting, 1 October, 2006, respectively.

Author notes

†

These authors contributed equally to this work.

‡

Present address: Department of Surgery, Haseki Research Hospital, Istanbul, Turkey.