PLC-gamma-1 phosphorylation status is prognostic of metastatic risk in patients with early-stage Luminal-A and -B breast cancer subtypes

Background Phospholipase Cγ1 (PLCγ1) is highly expressed in human tumours. Our previous studies reported that both stable and inducible PLCγ1 down-regulation can inhibit formation of breast-cancer-derived experimental lung metastasis. Further, high expression of PLCγ1 and its constitutively activated forms (i.e., PLCγ1-pY1253, PLCγ1-pY783) is associated with worse clinical outcome in terms of incidence of distant metastases, but not of local relapse in T1-T2, N0 breast cancer patients. Methods In the present retrospective study, we analysed the prognostic role of PLCγ1 in early breast cancer patients stratified according to the St. Gallen criteria and to their menopausal status. PLCγ1-pY1253 and PLCγ1-pY783 protein expression levels were determined by immunohistochemistry on tissue microarrays, and were correlated with patients’ clinical data, using univariate and multivariate statistical analyses. Results In our series, the prognostic value of PLCγ1 overexpression was restricted to Luminal type tumours. From multivariate analyses, pY1253-PLCγ1High was an independent prognostic factor only in postmenopausal patients with Luminal-B tumours (hazard ratio [HR], 2.4; 95% confidence interval [CI], 1.1–5.3; P = 0.034). Conversely, PLCγ1-pY783High was a remarkably strong risk factor (HR, 20.1; 95% CI, 2.2–178.4; P = 0.003) for pre/perimenopausal patients with Luminal-A tumours. Conclusions PLCγ1 overexpression is a strong predictive surrogate marker of development of metastases in early Luminal-A and -B breast cancer patients, being able to discriminate patients with high and low risk of metastases. Therefore, targeting the PLCγ1 pathway can be considered of potential benefit for prevention of metastatic disease. Electronic supplementary material The online version of this article (10.1186/s12885-019-5949-x) contains supplementary material, which is available to authorized users.

relapse [5][6][7]. Currently, adjuvant chemotherapy regimens are standard of care for treatment of early-stage disease that is oestrogen receptor (ER)-negative or human epidermal growth factor receptor (HER)2-positive. However, selecting chemotherapy for patients with ERpositive, HER2-negative disease is a more challenging task, due to the different risk profiles for disease relapse associated with this tumour subtype. Indeed, among patients with luminal tumours, there will be women at low risk of recurrence who will derive little benefit from chemotherapy combined with hormone therapy, and women at high-risk of recurrence where chemotherapy would be helpful. As far as the economic impacts on health care systems, is should be noted that adjuvant treatments are not devoid of toxicity. On this basis, above the search for new discriminatory biomarkers capable of selecting patients at risk of early relapse remains mandatory.
Among the lipid signalling metabolites in cells, the phosphoinositides are the most widely studied lipids due to their involvement in several cell signalling pathways. In particular, phosphoinositide-specific phospholipase C gamma 1 (PLCγ1) signalling is necessary for many physiological cellular processes (e.g., cell proliferation and differentiation) [8]. PLCγ1 is highly expressed in various tumours, including breast cancers [9][10][11]. We previously observed that PLCγ1 down-regulation strongly reduced formation of MDA-MB-231-derived lung metastases in nude mice [12]. Furthemore, lung metastasis formation from prostate cancer cells was significantly reduced by a dominant-negative fragment of PLCγ1 [13].
In human, using tumour cases as training and validation sets, we have shown that overexpression of activated PLCγ1 is a risk factor for distant relapse in T1-T2, N0 breast cancer patients undergoing adjuvant chemotherapy [14]. Therefore, in the present study, we investigated the prognostic role of PLCγ1 in these early breast cancer patients stratified according to the St. Gallen criteria and to their menopausal status.

Patients
We retrospectively reviewed the medical records of 979 consecutive patients (year range, 1995-2003) diagnosed with primary unilateral breast carcinoma at the "Regina Elena" National Cancer Institute, Rome, Italy. From the original series, only N0 patients with T1/T2 tumours were included in the present study (n = 414). The patients' and tumour characteristics are given in Table 1.
This study was reviewed and approved by the Ethics Committee of the Regina Elena National Cancer Institute. All of the patients were treated with quadrantectomy and received radiation therapy (n = 414), while 172 received chemotherapy without or with hormonal therapy, and 160 underwent only hormonal therapy. Patients with HER2-positive tumours did not receive trastuzumab, because this immune treatment was not available during the study period. The median follow-up was 79 months (range, 2-298 months). Follow-up data were collected from institutional records or from the referring physicians. During follow-up, 50 patients (12.1%) experienced local relapse. Distant relapse was seen in 65 (15.7%) of the patients.

Immunohistochemistry
The 414 breast cancer cases were distributed in 21 tissue microarrays (TMA) blocks assembled in duplicate. Briefly, TMAs were constructed by punching 2-mmdiameter cores of histologically proven invasive breast carcinoma areas, as previously described [14]. The tissue microarray sections were incubated with the mouse anti-PLCγ1 monoclonal antibody (sc-7290), and with the rabbit anti-PLCγ1-pY1253 (sc-22141-R) and anti-PLCγ1-pY783 (sc-12943-R) polyclonal antibodies, with all from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Although these antibodies were validated by Santa Cruz Biotechnology, their specificities were further validated using appropriate silenced breast cancer cell lines (Additional file 2: Figure S1). The anti-mouse and anti-rabbit EnVision kits (Agilent, Santa Clara, CA, USA) were used for signal amplification, as appropriate. For the control tissues, the primary antibody was excluded or substituted with non-immune serum or isotype-matched immunoglobulins. The immunohistochemical analysis was carried out by two pathologists (R.L., M.P.) by agreement, with both blinded to the clinicopathological information. The immunohistochemical results for the ER, progesterone receptor (PgR), Ki67 and HER2 status were obtained from the patients' hospital records.
Pearson's χ2 or Fisher's exact tests were used to asssess the relations between the tumour PLCγ1, PLCγ1-pY1253 and PLCγ1-pY783 protein expression and the patient clinicopathological parameters. Disease-free survival (DFS) was defined as the interval from surgery to the first of the following events: tumour relapse at local or distant sites. Local relapse-free survival (LRFS) and distant relapse-free survival (DRFS) were defined as the time from surgery to the occurrence of local and distant relapses, respectively. Kaplan-Meier plots were used for the survival analyses, and log-rank tests were applied to compare the survival curves between the patient groups.
Cox's proportional hazards models were used to evaluate the association of PLCγ1, PLCγ1-pY1253 and PLCγ1-pY783 expression with survival time, using covariates. The following covariates were computed in the multivariate models: tumour size, tumour grade, and ER, PgR, Ki-67, HER2, PLCγ1, PLCγ1-pY1253 and PLCγ1-pY783 status. The statistical software SPSS version 15.0 (SPSS, Chicago, IL, USA) was used throughout, and P < 0.05 was considered statistically significant.

Menopausal status
Using Kaplan-Meier plots, we also examined the distant recurrence rates associated with PLCγ1-pY1253 and PLCγ1-pY783 expression in luminal tumours with patients clustered according to menopausal status. We observed that pY1253-PLCγ1 High expression was significantly associated with lower DRFS rate in postmenopausal patients with Luminal-B tumours (P = 0.028), while pY783-PLCγ1 High expression was significantly correlated with increased risk of distant relapse in those patients with Luminal-A cancers and pre/perimenopausal status (P < 0.001) (Fig. 4).
Multivariate analyses confirmed that pY1253-PLCγ1 High was a significant independent prognostic factor for postmenopausal Luminal-B cancers (HR, 2.4: 95% CI, 1.1-5.3; P = 0.034), while over-expression of PLCγ1-pY783   Table S3). Significant negative prognostic value was seen for pY783-PLCγ1 High for women with Luminal-A tumours who were pre/perimenopausal and were treated with hormonal therapy, as well as those treated with chemotherapy plus hormonal therapy (P = 0.003, P = 0.001, respectively; Fig. 5). The two patients treated with radiotherapy alone were pY783-PLCγ1 Low , and they did not show distant metastasis events.

Discussion
The identification of criteria for accurate prognostication of disease relapse is crucial for the selection of patient candidates for adjuvant therapy. Recognizing patients with high recurrence risk can potentially enhance their treatment outcomes, with the adoption of more aggressive treatments from an earlier stage of the disease, which might ultimately offer better overall survival. Conversely, low-risk patients can undergo less aggressive therapy, and therefore they can enjoy a better quality of life.
Although the phosphoinositide cycle operates classically at the plasma membrane level, a phosphoinositide cycle operates also within the nucleus [21]. We previously observed a selective nuclear positivity for PLCγ1-pY1253 and PLCγ1-pY783 in patients with early breast cancer, indicating that the nuclear signalling of these activated forms of PLCγ1 may have a specific tumorigenic role [14]. The nuclear PLCγ1 can contribute to mammary carcinogenesis through the modulation of key pathways, including the phosphoinositide 3-kinase (PI3K) nuclear activation [22], and by regulating the expression of cell cycle regulators such as cyclin D1 and cyclin-dependent kinase 4 expression, and the nuclear export of the Cdk inhibitor p27-kip1 [23]. Conversely, the down-regulation of PLCγ1 expression in breast cancer cells results in decreased lung metastasis formation in mice [12]. However, the mechanism/s by which PLCγ1 favours migration and metastatisation remain unclear. PLCγ1 has an essential role in cytoskeletal changes needed for the acquisition of the metastatic phenotype [11,24], and dephosphorylation of PLCγ1 on residue Y783 inhibits PLCγ1 activation, thus blocking PLCγ1-activated rearrangement of the cytoskeleton, and cell migration [25]. PLCγ1 can contribute to metastatisation by direct [26] or indirect [12,26] activation of RAC1, thus inducing migration-supporting cellular structures, such as lamellipodia and filopodia.
In the present study, we have shown that the prognostic role of activated PLCγ1 expression is limited to ER-positive, Luminal breast tumours. Indeed, by using validated antibodies for immunohistochemistry on paraffin sections, different PLCγ1 phosphorylation sites were associated with different prognosis for the Luminal-A and -B molecular subtypes. Of note, pY1253-PLCγ1 High , but not pY783-PLCγ1 High , was a significant independent prognostic factor for postmenopausal patients with Luminal B cancers (HR, 2.4). On the other hand, this was reversed in the hormonal pre/perimenopausal setting, where pY783-PLCγ1 High , but not pY1253-PLCγ1 High , was a particularly strong and significant risk factor for metastatic relapse (HR, 20.1).
Several multigene assays can now be included in clinical practice, such as the Oncotype DX, Prosigna and MammaPrint assays. Compared to the use of standard prognostic criteria, these multigene assays can provide some improvements in the recognition of patients with early stage ER-positive, HER2-negative breast cancer that will be at risk of recurrence. Indeed, at present, randomised controlled trials [27] are ongoing to prospectively validate their clinical usefulness. The future of diagnostic/ prognostic testing in ER-positive breast cancer is likely to rely on devising and reliably deploying assays that can predict the benefits of additional therapies, including newer targeted therapies. As no particular technology holds the key, immunohistochemistry remains a well settled, widely diffuse, and low-cost technique, and so it can have a role in the choice of adequate treatment [28].
Although the relationship between oestrogen stimulation and PLCγ expression has been explored in depth, recent data [29] have indicate a role for PLCγ1 in the proliferation of ER-positive tumour cells. Cells must increase chaperone levels to fold and sort proteins required for ERα-dependent cell proliferation. The unfolded protein response (UPR), which is an endoplasmic reticulum stress sensor, controls protein folding homeostasis. The UPR is overexpressed in several tumours where an early, pathological, activation of UPR occurs before the accumulation of unfolded proteins. In ERαpositive breast and ovary cancer cells, 17β-oestradiol induces rapid anticipatory activation of the UPR that is strictly PLCγ1 dependent. ER-positive breast cancers (See figure on previous page.) Fig. 4 Kaplan-Meier plots in Luminal-A and Luminal-B subtypes. Kaplan-Meier estimates of DRFS in patients with Luminal-A and Luminal-B tumours according to menopausal status. In Luminal-A tumours, the patients with postmenopausal status showed distant relapse in 5% (3/62) PLCγ1-pY1253 Low and 11% (5/44) PLCγ1-pY1253 High , and in 5% (4/82) PLCγ1-pY783 Low and 17% (4/24) PLCγ1-pY783 High , and the patients with premenopausal status showed distant relapse in 4% (1/28) PLCγ1-pY1253 Low and 23% (5/22) PLCγ1-pY1253 High , and in 0% (0/37) PLCγ1-pY783 Low and 46% (6/13) PLCγ1-pY783 High . In Luminal-B tumours, the patients with postmenopausal status showed distant relapse in 12% (9/76) PLCγ1-pY1253 Low and 29% (18/63) PLCγ1-pY1253 High , and in 20% (22/109) PLCγ1-pY783 Low and 17% (5/30) PLCγ1-pY783 High , and the patients with premenopausal status showed distant relapse in 17% (4/24) PLCγ1-pY1253 Low and 31% (4/13) PLCγ1-pY1253 High , and in 16% (5/31) PLCγ1-pY783 Low and 50% (3/6) PLCγ1-pY783 High demonstrate elevated expression of a UPR gene signature that is also a prognostic marker associated to high risk of relapse and poor survival, and also resistance to tamoxifen therapy. Therefore, this PLCγ1-dependent anticipatory activation of the UPR defines a new role for oestrogens that can create a supportive environment for cancer cell proliferation and resistance to therapy, and might represent a new target in breast cancer [30]. Considering this aspect, it has also been reported that PLCγ1 activation downstream of FGFR-3 signalling is a critical event in the control of MAPK and PI3K activation, which can induce resistance to tamoxifen treatment [31].

Conclusions
Although the mechanisms involved remain to be defined, the activation of PLCγ1 as assessed by immunohistochemistry is a strong prognostic factor that can discriminate between high-risk and low-risk patients with hormone-receptor-positive early breast cancers. PLCγ1 might thus serve as a new target especially for treatment of Luminal-A pre/perimenopausal patients with T1-T2, N0 disease.