Breast Cancer (BC) patients and Healthy Controls were recruited at the Joliot Curie Institute, Cancer institute of Medicine of Cheikh Anta Diop University, of Hospital Aristide Le Dantec in Dakar (Senegal). Recruited HC are women free of known tumors or chronic inflammatory diseases. They live in Dakar.

Selected BC patients were newly diagnosed for breast cancer and had not undertaken any chemotherapy (CT) or radiotherapy. The major exclusion criteria included prior treatment with traditional plant-based remedies, blood transfusion during our screening and bacterial or viral infections such as tuberculosis, HIV and hepatitis B (immunodeficiency). These infectious diseases were systematically investigated by the medical team responsible for the monitoring. The tumor size was assessed using tape measure and ultrasound prior to any chemotherapy. Therefore, we sorted patients in two groups regarding the median value of the tumor size: group 1 with tumor size smaller than the median value and group 2 with tumor size greater than the median value. All selected patients were managed by the same medical staff and the CT protocol was in accordance with European Sarcoma Network Working Group (ESMO) recommendations specifying one cycle every 21 days. Usually, three cycles are recommended before undertaking a thorough clinical examination to evaluate therapy response. Each patient followed a treatment protocol depending on individual clinical disease characteristics and general health status [11]. Besides the assessments of the tumor size, drugs as doxorubicin (A), epirubicin (E), cyclophosphamide (C), methothrexate (M), 5-fluorouracil (F) were administered in 5 different neoadjuvant chemotherapy protocols (drug combinations: AC, CMF, FAC, FEC, EC). In terms of doses, drugs were administrated every three weeks: 10 - 20 mg/m² of body surface for doxorubicin, 60 - 120 mg/m² of body surface for epirubicin, 500 - 4000 mg/m² of body surface for cyclophosphamide and 40 - 60 mg/m² of body surface for methotrexate. 5-FU was used with to 1000 mg/m² every 3 weeks for 3 cycles.

After chemotherapy, clinical responses to chemotherapy were assessed according to RECIST1.1 criteria and patients were classified into 3 groups: complete response (CR), partial response (PR) and null response (NR = stable and progressive disease) [11].

Peripheral blood from all study participants (BC and HC) was sampled using EDTA vacutainers (Marque, ville, pays). For BC, sample collection was based on the treatment protocol. Three samples were collected per BC, corresponding to one before each CT cycle and are designated: Sample 1 (S1), Sample 2 (S2) and Sample 3 (S3). However, only one sample was obtained per HC. All samples were immediately centrifuged, and sera were stored at −80˚C and PBMCs isolated.

Informed consent was obtained from each participant and/or relatives prior to inclusion, after providing written or verbal information in their native language. The protocol 0196/2016/CER/UCAD was approved by Institutional Ethics Committee of Cheikh Anta Diop University (Dakar, Senegal).

Plasma samples were stored at −80˚C until performing the assay. Studied biological parameters were galectin-3, TNF-α, IL-1 and IL-6 levels in peripheral blood. They were measured by ELISA sandwich kits (Invitrogen^TM, Carlsbad, California, USA). According to the manufacturer’s instructions, the lyophilized standards were reconstituted to obtain for each parameter a stock of known concentration and serially diluted to generate a standard curve. Diluted capture antibodies were added to each well of the microtiter before incubation overnight at 4˚C. For each parameter, after washing with the appropriate solution with an adequate volume, diluted standards or samples were added and plates were incubated for 2 hours at room temperature. Detection antibodies were added to each well after a second wash and incubated for 1 hour at room temperature. Avidin-HRP was added to each well and incubated for 30 min at room temperature, before adding the substrate (15 - 20 min) and then stop solution. Optical densities were immediately determined using a microplate (Thermo Fisher^TM, Waltham, Massachusetts, USA) reader set to 450 nm. The results are expressed in pg/ml for cytokines and in ng/ml for galectin-3.

From BC patients and HC, Peripheral blood mononuclear cells (PBMCs) were isolated using standard Ficoll density gradient (Ficoll-Paque^TM PREMIUM Marlborough, UK) centrifuged within four hours after blood collection. Cells were washed once in 10% FBS/RPMI and once in phosphate-buffered saline (PBS) (Thermo Fisher^TM, Waltham, Massachusetts, USA) before a staining step for 30 minutes at 4˚C in FACS buffer (0.5% BSA/2mM EDTA/PBS) containing anti-FcγR (eBioscience^TM, San Diego, California, USA) to inhibit non-specific bindings. To identify T and B lymphocytes populations, 50 µl of PBMC suspension (1.106 cells/µl) were stained with three different panels of monoclonal antibodies labeled with fluorochromes. Panel 1: (CD3-PE, CD4-APC, CD8-PerCP, CD69-FITC), Panel 2: (CD3-PE, CD19-APC, CD69-FITC, HLADR-PercP) and Panel 3: (CD3-PeCy5, CD69-FITC, CD19-APC, Apo2.7-PE) (Becton Dickinson®, Franklin Lakes, New Jersey, USA). Stained cells were washed and suspended in 0.5%BSA/2mM EDTA/PBS and sorted in a FACS Calibur (Becton Dickinson®, Franklin Lakes, New Jersey, USA) using Cell Quest Pro® software. Results were analyzed by Flowjo® software version 10.0.6 (Becton Dickinson®, Franklin Lakes, New Jersey, USA). Apoptotic and activated T and B lymphocytes were Apo2.7+ and CD69+ (early activation for both T and B cells) or HLADR+ (late activation only for T lymphocytes) respectively.

The statistical analyses were performed using Statview^® software version 5.1 (SAS institute inc., Cary, North Carolina, USA). Differences between groups were evaluated with nonparametric Kruskal-Wallis H and Mann-Whitney U tests for continuous variables. Correlation analyses were performed using the nonparametric Spearman Rho test. Relationships between continuous dependent variables and age or tumor size as dichotomy nominal variables (high and low) have been performed using the Multiple Linear Regression (MLR) where dummy variable coding has been used for the tumor size status variable which has two categories. Graphs were performed in Statview^® and GraphPad Prism^® V5.03 for Windows (GraphPad, Playa La Jolla, California, USA). A p value < 0.05 was considered significant.

This study included 112 individuals aged from 20 to 78 years and distributed in two groups: 70 Senegalese Breast cancer (BC) patients (46 years range 24 - 78) and 42 healthy controls (HC) with no history of malignancy (26 years range 20 - 49). Peripheral blood from patients and controls was collected as described previously. The clinical and therapeutic characteristics of the patients are illustrated in Table 1.

a. nd*: data not available; A: doxorubicin; E: epirubicin; C: cyclophosphamide; M: methotrexate; F: 5-fluorouracil; CR: complete response; PR: partial response; NR: no response.

In terms of clinical aspects, most of the patients had a high Scarff Bloom Richardson (SBR) grade tumor. A rate of 50.8% of patients had grade II tumor and 9.2% had grade III versus 6.2% with grade I. The median value of tumor size was 60 mm (range 3 - 200).

After 3 neoadjuvant chemotherapy cycles, rate of patients with complete (CR), partial (PR) and no or low response (NR) to the cure were respectively 11% (n = 8), 59% (n = 39) and 29% (n = 20). For 3 patients, the response to the chemotherapy wasn’t assessed.

Median value of the gal-3 plasma concentration was 6.31 ng/ml (range 1.07 - 50.74) in patients group compared to versus 0.84 ng/ml in the controls. The gal-3 concentration was significantly higher in the patients group (Figure 1). After chemotherapy, this concentration seemed to be decreasing in patients with total and half therapy responses albeit without any statistical significance (Figure 1). We also found a positive correlation (Figure 2) between gal-3 concentration and tumor size (p = 0.010; rho = 0.48), before the first chemotherapy cure. For BC patients depending on therapy responses, a positive correlation was observed only in patients with no responses to chemotherapy (p = 0.041; rho = 0.68). Therefore, we sorted patients in two groups regarding their tumor size: group 1 (n = 11) with tumor size smaller than 60 mm (median value of tumor size) and group 2 (n = 19) with tumor size greater than 60 mm. We found a positive correlation only in group 2 (p = 0.019; rho = 0.55).

Before chemotherapy, median values in the BC group of TNF-α, IL-1 and IL-6 plasma concentrations were respectively 8.66 pg/ml (range 2.67 - 79.40), 4.530 pg/ml (range 3.01 - 16.77), 2.39 pg/ml (range 0.60 - 18.20) in BC group compare

to 5.89 pg/ml (range 1.84 - 16.95), 4.57 pg/ml (range 3.30 - 11.63), 1.65 (range 0.86 - 7.05) in the HC group.

However, a comparison of plasma inflammatory cytokines concentrations between patients and healthy controls has shown higher TNF-α rates in the BC patients compare to HC group (p = 0,024) (Figure 3). There weren’t any differences in the levels of IL-1 and IL-6 between the two groups.

Further, no correlation was found between gal-3 plasma concentration and the inflammatory cytokines in both HC and BC groups (p > 0.05).

Prior to the first chemotherapy cure, FACS results have shown statistically higher T and B lymphocytes apoptosis and activation rates in the patients’ group (Figure 4 and Figure 5).

Median values of T and B lymphocytes apoptosis, were respectively, 0.73% (range 0.00 - 2.63) and 0.99% (range 0.00 - 3.470) in the patients group in comparison to 0.13% (range 0.00 - 1.53) and 0.285 (0.00 - 1.00) found in the HC. After the first chemotherapy cure, we found an elevation in the lymphocytes apoptosis rates (Figure 3). Both the T and B lymphocytes apoptosis rates were significantly increased (p < 0.05).

The median values of CD69+ T and B lymphocytes, were respectively 2.39% (range 0.00 - 10.94) and 1.360% (range 0.00 - 61.50) in the patients’ group compared to 1.28% (range 0.52 - 5.78) and 0.83 (range 0.00 - 5.630) detected in the HC. Similar to the lymphocytes apoptosis rates, there was an increase in CD69+ T and B lymphocytes after the first chemotherapy cure. We have observed no effect of the first chemotherapy cure on HLA - DR+ T lymphocytes in patients (2.93%) and healthy HC (0.73%) as shown in (Figure 5). However, before the first chemotherapy cycle, CD69+ CD8+ T lymphocytes rate was higher in the PR group than in the other groups (p = 0.008) (Figure 6).

There wasn’t any correlation between gal-3 concentration and B lymphocytes apoptosis. But our results had shown a dynamic relationship between gal-3 concentration and T lymphocytes apoptosis (Table 2). Before therapy began, there was a negative correlation (p = 0.048; rho = −0.39) which became positive after the 2nd chemotherapy cure (p = 0.017; rho = 0.75). Also, there wasn’t any statistically significant correlation between the gal-3 plasma concentration and lymphocytes activation.