Expression of metalloproteases and their inhibitors in primary tumors and in local recurrences after mastectomy for breast cancer

Abstract

Aims

To investigate the expression of matrix metalloproteases (MMPs) and their inhibitors (TIMPs) in patients who develop local recurrence (LR) after mastectomy.

Methods

We analyzed the expressions of MMP-1, -2, -7, -9, -11, -13, -14, TIMP-1, -2, and -3, using immunohistochemical techniques, in primary tumors from patients without tumoral recurrence (n = 50), patients who developed distant metastasis (n = 50), and from patients who develop LRs (n = 25). LRs of the latter group were also analyzed for MMPs expression. All the patients underwent mastectomy.

Results

Score values for all MMPs and TIMPs were significantly higher in primary tumors of patients with distant metastasis. Primary tumors from patients with LR have lower expressions of MMPs and TIMPs compared with those from patients who developed distant metastasis, and with patients without recurrence for some MMPs. Remarkably, however, primary tumors from patients with LR showed significantly higher percentage of TIMP-1 and 2 expression in stromal cells compared to primary tumors from patients with distant metastasis or primary tumors from patients without tumoral progression. Furthermore, LRs had significantly higher MMP-9 expression than their corresponding primary tumors.

Conclusions

Our data indicate differences in MMPs/TIMPs expression between primary tumors of patients with LRs and of those with distant metastasis, both after mastectomy for breast cancer.

Introduction

Despite the increasing use of breast-conserving therapy, modified radical mastectomy remains as an important surgical technique in primary breast cancer. However, 10–18% of patients undergoing mastectomy will eventually develop a local recurrence (LR) (Lacour et al. 1983, 1987; Kaae and Johansen 1962; Carreno et al. 2007). LR is defined as the development of adenocarcinoma in one or more of the following locations: skin, subcutaneous tissues, or muscles of the ipsilateral chest wall. This is due, presumably, to the fact that scar tissue is a preferential site to lodge and grow circulating tumor cells. In fact, experimental studies have shown a preference of tumor cells to grow in scar tissue, this being either surgical scars (Alexander and Altemeier 1964; Fisher and Fisher 1959; Symmans et al. 2003; Durkan et al. 2003; Skipper et al. 1989) or radiated tissues (Dao and Yogo 1967; Fidler and Zeidman 1972). In addition, survival after a postmastectomy LR remains very poor, with 5-year survival rates ranging from 40 to 70% and disease-free survival ranging from 13 to 50% (Donegan et al. 1966; Spratt 1967; Karabali-Dalamaga et al. 1978; Bedwinek et al. 1981; Borner et al. 1994; Schuck et al. 2002), and is very often associated with distant metastases (Dao and Nemoto 1963). Hence, it would be extremely helpful to identify breast cancer patients at a higher risk of locoregional recurrence, which could be candidates for a more aggressive treatment, such as postmastectomy radiotherapy. There are a number of known parameters of the primary tumors, such as tumor size, nodal status, estrogen receptor status, and tumor grade (Orel et al. 1995; Pisansky et al. 1993; Stefanik et al. 1985; Sykes et al. 1989; Kamby et al. 1991; Willner et al. 1997; Barnes et al. 1991; Berstock et al. 1985; van Tienhoven et al. 1999; Kamby and Sengelov 1999), that directly correlate with the development of LR. However, there is a need to unravel new biological factors that could be implicated in the development of postmastectomy LR and in the outcome thereof.

Some of the best candidates for such a biological marker are the matrix metalloproteases (MMPs). Their role is well known in the degradation of connective tissue stroma and of basement membranes, which are key elements/barriers in tumor invasion and metastasis. In addition, recent data clearly challenge the classic dogma stating that MMPs promote metastasis solely by modulating the remodeling of extracellular matrix. Indeed, MMPs have also been attributed an impact on tumor cell behavior in vivo as a consequence of their ability to cleave growth factors, cell surface receptors, cell adhesion molecules, or chemokines/cytokines (Sternlicht and Werb 2001; Egeblad and Werb 2002; Turk et al. 2004; Rifkin et al. 1999; Manes et al. 1999; Noe et al. 2001). Furthermore, by cleaving of proapoptotic factors, MMPs are able to produce an aggressive phenotype by generating apoptotic resistant cells (Fingleton et al. 2001). MMPs may also regulate angiogenesis in cancer, both positively through their ability to mobilize or activate proangiogenic factors (Yu and Stamenkovic 2000; Stetler-Stevenson 1999), or negatively, via generation of angiogenesis inhibitors, such as angiostatin and endostatin, which are cleaved from large protein precursors (Dong et al. 1997; Cornelius et al. 1998; Ferreras et al. 2000). Consequently, several MMPs, in particular the gelatinases MMP-2 (Jones et al. 1999; Duffy et al. 2000; Talvensaari-Mattila et al. 2001, 2003; Baker et al. 2002; Grieu et al. 2004; Li et al. 2004; Sivula et al. 2005) and -9 (Li et al. 2004; Chantrain et al. 2004; Pellikainen et al. 2004; Vizoso et al. 2007), have been recently studied as prognostic factors in breast cancer, and, as a result, being associated with a poor outcome in various subsets of breast cancer patients. Likewise, it has been reported that several other MMPs, such as MMP-7 (Vizoso et al. 2007), -11 (Duffy et al. 2000; Vizoso et al. 2007), -14 (Jones et al. 1999; Vizoso et al. 2007; Mimori et al. 2001), and 13 (Nielsen et al. 2001), may be over-expressed and/or related to the clinical outcome in breast cancer. On the other hand, it is known that the activity of MMPs is specifically inhibited by tissue inhibitors of metalloproteases (TIMPs). Currently, four different TIMPs are known to exist: TIMPs 1, 2, 3, and 4 (Gonzalez et al. 2007). Nevertheless, it is now assumed that TIMPs are multifactorial proteins also involved in the induction of proliferation and the inhibition of apoptosis (Jiang et al. 2002; Baker et al. 1999). Thus, it has also been reported that some TIMPs, such as TIMP-1 (Vizoso et al. 2007; Ree et al. 1997; McCarthy et al. 1999; Nakopoulou et al. 2002; Schrohl et al. 2003, 2004) or TIMP-2 (Vizoso et al. 2007; Ree et al. 1997; Visscher et al. 1994; Remacle et al. 2000), may be over-expressed and/or related to the clinical outcome of breast cancer.

The present study, performed on women who had undergone mastectomy for breast cancer, had two aims: (1) to analyze the differences in the expression of MMPs and TIMPs between primary tumors: those of patients who developed a local lesion as the first event of tumoral recurrence, those of patients whose first sign of disease recurrence is the appearance of distant metastases, and, finally, those of patients without tumoral recurrence; and (2) to compare the expressions of MMPs and TIMPs in primary tumors and in paired isolated LRs.

Materials and methods

Patient characteristics

We analyzed the records of 1,087 women who were treated for breast cancer at Hospital de Jove (Gijón, Spain), Hospital de Cabueñes (Gijón, Spain), and at Hospital Central de Asturias (Oviedo, Spain), between 1990 and 2002. All the patients underwent a modified radical mastectomy. We identified a total of 98 patients with a first isolated breast cancer LR following primary mastectomy. None of the patients had tumoral involvement at the marginal resection in the mastectomy specimens. Patients with concomitant distant metastases at the time of the initial diagnosis were excluded from the study. None of the patients showed evidence of any other malignant tumor at the time of diagnosis. Specimens at the time of surgery from both primary tumors as well as from their corresponding recurrences were obtained from 25 out of the 98 patients of the study. The status of these 25 patients with respect to age, menopausal status, clinical tumoral stage, histological grade, hormonal receptor status, adjuvant radiotherapy, and/or systemic therapy, is listed in Table 1. The histological grade was determined according to criteria reported by the Nottingham modification of Bloom and Richardson score (Dixon et al. 1991), whereas nodal status was assessed histopathologically. LR was defined as any reappearance of tumor on the ipsilateral chest wall or mastectomy scar.

Table 1.

Basal characteristics from 25 patients with local recurrence and from 50 patients with distant metastasis, both as first manifestation after mastectomy, and from 50 patients without tumoral recurrence

Characteristics	Patients without tumoral recurrence	Primary tumors of patients with local recurrence	Primary tumors of patients with distant metastasis
	N (%)	N (%)	N (%)
Total cases	50	25	50
Age (median)
≤56	21 (42)	13 (52)	27 (54)
>56	29 (58)	12 (48)	23 (46)
Menopausal status
Premenopausal	16 (32)	9 (36)	13 (26)
Postmenopausal	34 (68)	16 (64)	37 (74)
T
T1–T2	28 (56)	2 (8)	21 (42)
T3–T4	22 (44)	23 (92)	29 (58)
Nodal status
Negative	28 (56)	12 (48)	21 (42)
Positive	22 (44)	13 (52)	29 (58)
SBR
I	19 (38)	2 (8)	12 (24)
II	24 (48)	10 (40)	25 (50)
III	7 (14)	7 (28)	13 (26)
Unknown		6
RE
Negative	15 (30)	10 (40)	32 (64)
Positive	35 (70)	15 (60)	18 (36)
RP
Negative	16 (32)	13 (52)	34 (68)
Positive	34 (68)	12 (48)	16 (32)
Radiotherapy
No	41 (82)	8 (32)	25 (50)
Yes	9 (18)	17 (68)	25 (50)
Chemotherapy
No	29 (58)	9 (36)	18 (36)
Yes	21 (42)	16 (64)	32 (64)
Tamoxifen
No	20 (40)	6 (24)	36 (72)
Yes	30 (60)	19 (76)	14 (28)

All these 25 patients were followed for disease recurrence and survival status by clinical and biochemical studies every 3 months for the first 2 years and then once a year. Radiological studies were performed once a year, or when considered necessary. The medium follow-up period was of 39 months (range 6–217 months). The end-point of the study was death, secondary to tumor progression.

This study also comprised one control group of patients who underwent mastectomy for invasive ductal breast cancer, and who develop distant metastasis as first manifestation of tumoral progression. An additional control group was that of patients without tumoral recurrence, who had a minimum of 5 years of follow-up. Both groups of patients were randomly selected among those treated between 1990 and 2001, and stratified by node status.

All patients were treated according to the guidelines used in our institutions, and the study adhered to national regulations, being approved by our institution’s Ethics and Investigation Committee.

Tissue arrays and immunohistochemistry

Breast carcinoma tissue samples were obtained at the time of surgery. Those used in the study were routinely fixed (overnight in 10% buffered formalin), paraffin-embedded tumoral samples stored in the Pathology laboratories. Histopathologically representative tumor areas were defined on hematoxylin and eosin-stained sections and marked on the slide. Tumor tissue array (TA) blocks were obtained by punching a tissue cylinder (core) with a diameter of 1.5 mm through a histologically representative area of each ‘donor’ tumor block, which was then inserted into an empty ‘recipient’ tissue array paraffin block using a manual tissue arrayer (Beecker Instruments, Sun Prairie, Wisconsin, USA) as described elsewhere (Parker et al. 2002). Collection of tissue cores was carried out under highly controlled conditions. Areas of non-necrotic cancerous tissue were selected for arraying by two experienced pathologists (L.O. González and A. M. Merino). Two cores (double redundancy) were employed for each case, as this method has been shown to correlate well with conventional immunohistochemical staining (Rimm et al. 2001). From the 50 tumor samples available, two TA blocks were prepared each containing 25 primary and secondary tumors samples, as well as internal controls including four normal breast tissue samples from two healthy women who underwent reductive mammary surgery. These latter samples contained epithelial components, in which the immunohistochemistry was negative with the antibodies of the study.

Two composite high-density TA blocks were designed, and serial 5-μm sections were consecutively cut with a microtome (Leica Microsystems GmbH, Wetzlar, Germany) and transferred to adhesive-coated slides. One section from each tissue array block was stained with H&E, and these slides were then reviewed to confirm that the sample was representative of the original tumor. Immunohistochemistry was done on these sections of TA fixed in 10% buffered formalin and embedded in paraffin using a TechMate TM50 autostainer (Dako, Glostrup, Denmark). Antibodies for MMPs and TIMPs were obtained from Neomarker (Lab Vision Corporation, Fremont, CA, USA). The dilution for each antibody was established based on negative and positive controls (1/50 for MMP-2, -7, -14 and TIMP-2; 1/100 for MMP-9, -13, TIMP-1 and -3; and 1/200 for MMP-1, -11). Negative control is DakoCytomation mouse serum diluted to the same mouse IgG concentration as the primary antibody. All the dilutions were made in Antibody Diluent, (Dako) and incubated for 30 min at room temperature.

Tissue sections were deparaffinized in xylene and then rehydrated in graded concentrations of ethyl alcohol (100, 96, 80, 70%, then water). To enhance antigen retrieval only for some antibodies, TA sections were microwave treated in a H2800 Microwave Processor (EBSciences, East Granby, CT, USA) in citrate buffer (Target Retrieval Solution; Dako) at 99°C for 16 min. Endogenous peroxidase activity was blocked by incubating the slides in peroxidase-blocking solution (Dako) for 5 min. The EnVision Detection Kit (Dako) was used as the reactivity detection system. Sections were counterstained with hematoxylin, dehydrated with ethanol, and permanently coverslipped.

For each antibody preparation studied, the location of immunoreactivity, percentage of reactive area. and intensity were determined. All the cases were semiquantified for each protein-stained area. An image analysis system with the Olympus BX51 microscope and analysis software (analySIS^®, Soft imaging system, Münster, Germany) was employed as follows. Tumoral sections were stained with antibodies according to the method explained above and counterstained with hematoxylin. There are different optical thresholds for both stains. Each core was scanned with a ×400 power objective in two fields per core. Fields were selected searching for the protein-reactive areas. The computer program selected and traced a line around antibody-reactive areas (higher optical threshold: red spots), with the remaining, non-stained areas (hematoxylin-stained tissue with lower optical threshold) standing out as a blue background. Any field has an area ratio of stained (red) versus non-stained areas (blue). A final area ratio was obtained after averaging two fields. To evaluate immunostaining intensity we used a numeric score ranging from 0 to 3, reflecting the intensity as follows: 0, no reactivity; 1, weak reactivity; 2, moderate reactivity; and 3, intense reactivity. This score was applied both to tumoral and to stromal cells. Using an Excel spreadsheet, the mean score was obtained by multiplying the intensity score (I) by the percentage of reactivity area (PA) and the results were added together (total score: I × PA). This overall score was then averaged with the number of cores that were done for each patient. If there was no tumor in a particular core, then no score was given. In addition, for each tumor the mean score of two core biopsy samples was calculated.

Data analysis and statistical methods

Differences in percentages were calculated with the chi-square test. Immunoreactivity scores for each protein were expressed as median (range). Comparison of immunoreactivity values between groups was made with the Mann–Whitney or Kruskal–Wallis tests. Comparison of immunoreactivity values between primary tumors and paired LRs were assessed by using the Wilcoxon test. Statistical results were corrected applying Bonferroni’s correction. In addition, a multivariate multinomial logistic regression model was used to evaluate simultaneously the influence of significant co-variables on the relationship between MMPs/TIMPs expressions and local recurrences. Probabilities of survival were calculated with the Kaplan–Meier method. Differences between curves were evaluated with the log rank test. The SPSS 17.00 program was used for all calculations. Statistical significance was considered at 5% probability level (p < 0.05).

Results

More than 3,000 determinations in cancer specimens, from 50 patients without tumoral recurrences, from 25 patients with LRs after mastectomy, and from 50 patients with distant metastasis after mastectomy were evaluated on TAs. Minimal internal variance of score data between duplicate tissue cores from the same patients was detected in the TAs, showing a high agreement for each protein (r > 0.95 and p < 0.0001, for each protein), in primary tumors as in LR tissues. Likewise, we have previously described a validation study for MMPs and TIMPs in invasive breast carcinomas (Vizoso et al. 2007).

Figure 1 shows examples of TAs with immunoreactivity for each protein being evaluated. There was wide variability in the immunoreactivity score values for each protein (Table 2). Immunoreactivity for all the proteins studied was localized predominantly in tumor cells, but also, in a significant percentage of cases, in stromal cells.

Fig. 1.

Up Examples of tissue arrays with immunostaining for each protein evaluated. a MMP-1, b MMP-2, c MMP-7, d MMP-9, e MMP-11, f MMP-13, g MMP-14, h TIMP-1, i TIMP-2, j TIMP-3. Down (left) 400× Immunostaining for TIMP-2 in mononuclear inflammatory cells, (right) 400× Immunostaining for MMP-11 in fibroblastic cells

Table 2.

MMPs and TIMPs immunoreactivity values in primary tumors from patients without tumoral recurrences, from patients with local recurrences, from patients with distant metastasis and in local recurrences

Factor	Primary tumors of patients without tumoral recurrence		Primary tumors of patients with tumoral recurrence		Primary tumors of patients with distant metastasis		Univariate p value*	Multivariate p value*	Local recurrences
	N	Median (range)	N	Median (range)	N	Median (range)			Median (range)
MMP-1	48	120 (0–277.5)	24	48.9 (0–285)	50	147.5 (0–285)	0.002	0.006	0 (0–204.4)
MMP-2	48	0	25	0 (0–121.2)	50	0 (0–246)	n.s.	n.s.	0 (0–94.8)
MMP-7	47	140.9 (0–258.4)	25	0 (0–130.8)	50	120 (0–270)	0.0001	0.0001	0 (0–53.8)
MMP-9	50	64.5 (0–156)	25	0 (0–52.2)	49	108 (0–273)	0.0001	0.01	36.2 (0–120.2)†
MMP-11	49	134 (0–279)	24	121.3 (0–257.1)	49	173 (0–277.7)	0.002	0.0001	150.9 (0–275.1)
MMP-13	50	65.7 (0–234)	22	45.19 (0–133.8)	50	60.6 (0–192.3)	0.05	0.05	44.9 (0–142.1)
MMP-14	50	81 (0–261)	24	73.4 (0.279.8)	50	85.3 (0–258.5)	0.05	n.s.	59.7 (0–159.8)
TIMP-1	50	132 (0–258)	24	116.6 (21.8–235.7)	49	158 (0–285)	0.0001	0.01	127.8 (28.2–217.2)
TIMP-2	50	72 (0–220)	24	80.2 (0–211.8)	49	142 (34–243)	0.0001	0.01	101.6 (39.4–215.9)
TIMP-3	50	101.7 (0–284.8)	24	53.1 (0–213.6)	50	126.9 (0–253.3)	0.03	n.s.	61.3 (0–268.2)

n.s. not significant

* p values for comparing the three groups of primary tumors primary tumors

^†0.006: local recurrence versus primary tumors of local recurrences (Wilcoxon test)

First, we compared the expressions of MMPs and TIMPs between the respective primary tumors of patients with LR, patients with distant metastasis as first manifestation of tumoral recurrence, and patients without tumoral recurrence. As can be seen in Table 2, score values for the majority of MMPs and TIMPs were significantly higher in primary tumors of patients with distant metastasis compared to those of patients with LR or without tumoral recurrence. Nevertheless, it is worth mentioning that there were lower score values for MMP-1, -7, and -9, and TIMP-3, in patients with LR compared to patients without tumoral recurrence. Likewise, we found significant differences in the expressions of these factors between the primary tumors of these three groups of patients with regard to the cellular type (stromal cells (fibroblasts or MICs)) expressing each factor within the tumor. Indeed, as seen in Table 3, stromal cells of primary tumors from patients with distant metastasis showed more frequently positive immunoreactivity for MMP-1, -7, -9, -11, -13, and -14, than the other two groups of patients. In addition, primary tumors from patients with LR showed significantly lower expression of MMP-1, -7, and -14 by stromal cells than the other groups. However, primary tumors from patients with LR had significantly high percentages of expression of TIMP-1 and -2 in their stromal cells (Table 3). It is of note that the percentage of expression of these MMPs and TIMPs by stromal cells is similar in LRs and in their primary tumor counterparts (Table 3). On the other hand it is also remarkable to mention that we found no significant associations between MMPs or TIMPs expression and clinico-pathological parameters, such as tumor size, nodal status or histological grade in our study population of patients with breast cancer (data not shown).

Table 3.

Cell type expressing MMPs and TIMPs in primary tumors from patients without tumoral recurrence, from patients with local recurrences, from patients with distant metastasis, and in local recurrences

Factor	Primary tumors of patients without recurrence	Primary tumors of patients with local recurrence	Primary tumors of patients with distant metastasis	Univariate p value*	Multivariate p value*	Local recurrences
	N (%)	N (%)	N (%)			N (%)
MMP-1
Tumoral cell	41 (85.4)	16 (66.7)	47 (94)	0.008	0.02	11 (45.8)
Fibroblast	38 (79.2)	4 (16.7)	45 (90)	0.0001	0.001	3 (12.5)
MIC	31 (64.6)	3 (12.5)	36 (72)	0.0001	0.001	4 (16.7)
MMP-2
Tumoral cell	12 (25)	5 (20)	21 (42)	0.08	n.s.	7 (28)
Fibroblast	8 (16.7)	4 (16)	15 (30)	n.s.	n.s.	6 (24)
MIC	0	0	1 (2)	n.s.	n.s.	0
MMP-7
Tumoral cell	39 (83)	6 (24)	46 (92)	0.0001	0.0001	3 (12)
Fibroblast	29 (61.7)	1 (4)	41 (82)	0.0001	0.001	1 (4)
MIC	20 (42.6)	1 (4)	32 (64)	0.0001	0.0001	0
MMP-9
Tumoral cell	29 (58)	9 (36)	44 (89.8)	0.0001	0.012	16 (64)
Fibroblast	0	1 (4)	27 (34.7)	0.0001	0.01	1 (4)
MIC	0	5 (20)	8 (16.3)	0.007	0.0001	4 (16)
MMP-11
Tumoral cell	39 (79.6)	18 (75)	46 (93.9)	0.05	0.005	19 (79.2)
Fibroblast	19 (38.8)	15 (62.5)	43 (87.8)	0.0001	0.05	19 (79.2)
MIC	0	13 (54.2)	27 (55.1)	0.0001	0.0001	15 (62.5)
MMP-13
Tumoral cell	37 (74)	15 (68.2)	40 (80)	n.s.	n.s.	14 (63.6)
Fibroblast	14 (28)	3 (13.6)	36 (72)	0.0001	0.0001	7 (31.8)
MIC	8 (16)	0	28 (56)	0.0001	0.01	2 (9.1)
MMP-14
Tumoral cell	45 (90)	14 (60.9)	45 (90)	0.002	n.s.	13 (56.6)
Fibroblast	40 (80)	4 (17.4)	39 (78)	0.0001	0.001	2 (8.7)
MIC	11 (22)	3 (13)	35 (70)	0.0001	0.01	7 (30.4)
TIMP-1
Tumoral cell	45 (90)	24 (100)	48 (96)	n.s.	n.s.	24 (100)
Fibroblast	24 (48)	19 (79.2)	26 (52)	0.03	n.s.	21 (87.5)
MIC	3 (6)	20 (83.3)	24 (48)	0.0001	0.0001	22 (91.7)
TIMP-2
Tumoral cell	39 (78)	23 (95.8)	47 (94)	0.02	n.s.	24 (100)
Fibroblast	6 (12)	21 (87.5)	33 (66)	0.0001	0.0001	20 (83.3)
MIC	5 (10)	20 (83.3)	27 (54)	0.0001	0.0001	20 (83.3)
TIMP-3
Tumoral cell	44 (88)	18 (72)	41 (82)	n.s.	n.s.	19 (76)
Fibroblast	19 (38)	16 (64)	36 (72)	0.002	0.01	10 (40)
MIC	20 (40)	2 (8)	28 (56)	0.0001	0.008	3 (20)

n.s. not significant

* p value for comparing the three groups of primary tumors

We also investigated possible differences of MMPs and TIMPs expression between primary tumors and their corresponding LRs. As can be seen in Table 2, we only found significant differences for MMP-9 score values. LRs had significantly higher MMP-9 expression than their corresponding primary tumors (p = 0.006; Fig. 2). It is also remarkable that multivariate analysis adjusted for potential confounders confirmed the more above point out as significant for the majority associations between MMPs/TIMPs expression and LR (Tables 2, 3). However, for each cellular type, non-significant differences were found between expressions of MMPs or TIMPs in primary tumors and in LRs (Table 3).

Fig. 2.

Comparative expression of MMP-9 in primary tumors and local recurrence

Our results showed a significantly negative, although low, correlation between MMP-11 expression in primary tumors and paired LRs (r = −0.438; p = 0.032); whereas there was a significant and positive correlation for TIMP-2 between these two paired sets (r = 0.438; p = 0.032). However, no significant correlations for other MMPs or TIMPs, between primary tumors and LRs were found (data not shown).

We also investigated the possible prognostic value of the expressions of MMPs and TIMPs, both in primary tumors as in their corresponding LRs. However, no significant prognostic value of MMPs or TIMPs expression was found (data not shown).

Discussion

To our knowledge, this is the first study comparing the MMPs/TIMPs expression in the primary tumors of breast cancer patients after mastectomy. The patients were grouped in three cohorts: patients who developed LR, patients who developed distant metastasis (both LR and metastasis as first manifestation of tumoral recurrence), and patients without tumoral recurrence. Our results demonstrate significant differences in MMPs and TIMPs expressions between these three groups. Primary tumors from patients with LR have lower expression of MMPs and TIMPs compared to those with distant metastasis, and even, for some MMPs, to those without recurrence. However, it was remarkable that primary tumors in the LR group showed significantly higher percentage of stromal cell expression of TIMP-1 and 2, compared to the other two groups. In addition, we also found evolutionary changes in MMP-9 expression of primary tumors versus their corresponding LRs, which may be of biological and clinical importance.

There are two main hypothesis to explain the origin and significance of LRs: the first one states that LR is caused by an incomplete initial removal of the tumor (Donegan et al. 1966; Auchincloss 1958; Scanlon 1985; Toonkel et al. 1983); the second one proposes that LR is a sign (the first one) of the disease being already disseminated (Crile 1972; Baral et al. 1985; Fisher et al. 1980; Gilliland et al. 1983; Papaioannou 1985; Valagussa et al. 1978). Although, at the present time, it is not possible to determine the cause of isolated LRs, our data shows a significant lower MMPs profile in the primary tumors of patients who develop LR as first manifestation of tumoral progression, suggesting that an incomplete removal could be the cause behind local recurrence. Indeed, the putative relationship between MMPs expression and tumor invasion and distant metastasis is widely accepted. In accordance with our previous reports (Gonzalez et al. 2007; Vizoso et al. 2007), our work demonstrates a high global expression of MMPs (scores values), but additionally we also show a high expression rates of MMPs in stromal cells from primary tumors of patients who develop distant metastasis as first manifestation of tumoral progression. This supports the notion that stromal cells, such as fibroblasts and MICs, at least, could play an active role in tumoral progression. In addition, it is also of note our finding of high expressions of TIMP-1 and -2 in LR fibroblasts. It is well known that TIMPs are multipotential proteins: in addition to their role inhibiting the MMPs, it is now assumed that TIMPs are multifactorial proteins involved in the induction of proliferation as well as in the inhibition of apoptosis (Jiang et al. 2002; Baker et al. 1999). It is remarkable that LR stromal cells have high expression of TIMP-1 and -2. This make us hypothesize that the properties of host stromal cells may contribute to the survival and proliferation of those tumor cells that are incompletely removed in the initial treatment of tumors.

It seems reasonable to consider that the detection of a first isolated LR may be, in part, a consequence of the absence of prior distant tumor development, which might be a selection criteria of low tumoral aggressiveness. Nevertheless, although certain subgroups with more favorable prognosis are believed to exist, the outcome of patients with local or regional breast cancer recurrence after mastectomy is often described as fatal (Veronesi et al. 1995), because many patients develop distant metastases within a short period of time (Aberizk et al. 1986; Bedwinek 1994). To identify subsets of patients differing in the clinical course of the disease, different well-known prognostic factors for primary tumors and/or locoregional tumoral recurrences have been described, such as large size of primary tumors, node-positive status and poorly differentiated grade (Orel et al. 1995; Pisansky et al. 1993; Stefanik et al. 1985; Sykes et al. 1989; Kamby et al. 1991; Willner et al. 1997; Barnes et al. 1991; Berstock et al. 1985; van Tienhoven et al. 1999; Kamby and Sengelov 1999; Pineiro et al. 2004). Nevertheless, it is remarkable that these same parameters are not of value in predicting the risk of LR versus distant metastasis after mastectomy for invasive breast cancer. Because of that, it seems reasonable to consider the possibility that prognostic factors drawn from the recurrence itself might predict the final outcome in a more accurately way. Although the present study does not include enough to reach a prognostic value, it is worth commenting our comparative evaluation of MMPs/TIMPs expression in primary tumors and in their corresponding LRs after mastectomy. Of the 10 parameters analyzed, we only found significant and weak correlations between the primary tumors and the paired LRs for MMP-11 and TIMP-2. This could be the sign of evolutional changes happening in LRs versus their corresponding primary tumors. Similarly, in a prior report, we found no significant concordance for androgen receptors, c-erbB-2 or ki67 (Carreno et al. 2007). Thus, all these data suggest that LRs evolve some of the biological features of their corresponding primary tumors.

Our data also demonstrate that LRs express significantly higher MMP-9 score values than their corresponding primary tumors. MMP-9 was mainly expressed by tumoral cells, which seems to reflect evolutionary biological changes in the LR tumoral cells, which could be of prognostic importance. MMP-9 (Gelatinase B) is related to tumor invasion and metastasis through its special capacity to degrade the type IV collagen found in basement membranes (Jones and Walker 1997), as well as to induce angiogenesis (Egeblad and Werb 2002). There are several reports showing that a high MMP-9 expression correlates significantly with tumoral aggressiveness and poor prognosis in breast cancer (Li et al. 2004; Chantrain et al. 2004; Pellikainen et al. 2004). It has also been described that as breast cancer progresses, activation of MMP-9 occurs during the late cancerous stage (Liotta and Kohn 2001).

In summary, our data indicate differences in MMPs/TIMPs expression between primary tumors of patients with LRs and of those with distant metastasis, both after mastectomy for breast cancer. Further studies are necessary to evaluate if high MMP-9 in local recurrences may be a valuable prognostic marker. We postulate that this marker could be potentially used to select candidates for further therapeutic strategies when local recurrence is the first tumor manifestation after mastectomy for breast cancer.