Propranolol Attenuates Surgical Stress–Induced Elevation of the Regulatory T Cell Response in Patients Undergoing Radical Mastectomy

Breast cancer is by far the most common female malignancy and is the leading cause of cancer-related death among women worldwide (1). Although surgery is a mainstay of treatment for most breast cancer patients, studies have suggested that perioperative immunosuppression may promote the growth of pre-existing micrometastases, thereby playing a key role in postoperative metastatic recurrences (2–4).

It has been well documented that surgery profoundly suppresses many components of cell-mediated immunity (CMI), including decreasing the number and activity of NK cells, helper-inducer T cells, and CTLs, as well as increasing the number of suppressor T cells. Immunosuppression begins immediately after surgery and lasts for hours to several days postoperatively (5–8). Regulatory T cells (Tregs), characterized by the classical marker proteins CD25 and FOXP3 (9), can exert a broad suppressive effect on antitumor immunity by direct or indirect mechanisms, ultimately resulting in immunological tolerance and escape of tumor cells (10, 11). Elevation of CD4⁺CD25⁺FOXP3⁺ Tregs in the peripheral blood is associated with a higher risk of tumor recurrence and a poor prognosis in breast cancer patients (11–13). It is reasonable to speculate that the number and activity of Tregs perioperatively may be predictive of clinical outcomes.

Surgery induces the release of catecholamines (CAs) and PGs at levels that correspond to the extent of surgical stress and tissue trauma. Recent data have demonstrated that CAs and PGs may be key factors that mediate the suppressive effects on antimetastatic CMI perioperatively (14–18). Both hormones have been repeatedly demonstrated to suppress most aspects of CMI as well as promote the Treg response in vitro (19–22). However, their effects on Tregs in breast cancer patients undergoing surgery are not clear.

Surgical stress promotes increases in epinephrine (E) and norepinephrine (NE) levels. The suppressive effects of E and NE on CMI are primarily mediated through their specific binding to the β-adrenergic receptors (β-AR) of immune cells, which promote tumor survival, migration, and resistance to anoikis (19, 23). These effects are thought to be inhibited by the nonselective β-adrenergic antagonist propranolol (24, 25). Interestingly, several epidemiological studies have demonstrated that the chronic use of β-blockers improves the long-term prognosis of breast cancer patients (26, 27). The proinflammatory PGs induced by surgical inflammatory responses are also known to suppress CMI and to protect tumors from immune destruction (28). In addition, animal studies indicate that the combined use of propranolol and etodolac (a selective cyclo-oxygenase [COX]-2 inhibitor), rather than either treatment alone, could improve NK activity and potentially reduce the postoperative recurrence and metastasis of tumors (16, 17). To date, the effect of short-term perioperative regimen of β-blockers and COX-2 inhibitors on suppressive Tregs has not been clinically assessed in breast cancer patients.

In this study, we initially investigated the perioperative effects of propranolol and parecoxib (another selective COX-2 inhibitor that is used more frequently in the perioperative period) on the level of CD4⁺CD25⁺FOXP3⁺ Tregs in the peripheral blood of breast cancer patients. An ex vivo IFN-γ ELISPOT assay was used to measure the suppressive activity of Tregs on the responses of CD4⁺ T cells to tumor Ags (TAs) perioperatively. The present work indicated that propranolol administration alone was able to abrogate the increased Treg level or activity observed in response to mastectomy. Hence, to further investigate whether similar changes would occur in Tregs exposed to stress hormones in vitro, we extended our experiments on the effects of differing concentrations of E and/or propranolol on Tregs using an established PBMC culture model. The PBMC culture system was previously used to study the effects of stress hormones (cortisol or catecholamines) on immune balance, including the transcription of regulatory FOXP3, IL-10, and TGF-β mRNA, as well as on the Th1/Th2 cytokine balance (22). In this work, we chose to detect Treg proliferation to more directly study the effects of catecholamines on Tregs over PBMCs from breast cancer patients. E, acting as an in vitro representative for surgical stress, promoted Treg proliferation, and propranolol treatment prior to E exposure conferred a protective effect in the prevention of Treg elevation. These results were consistent with our clinical observations. Thus, our study demonstrates the beneficial effect of propranolol in the inhibition of Treg responses in vivo and in vitro, and suggests that clinical treatment with propranolol during the perioperative period could improve CMI in breast cancer patients.

In the clinical trial part, the current study enrolled 154 women between the ages of 25 and 65 y who underwent a modified radical mastectomy for the treatment of primary breast cancer at the Third Xiangya Hospital of Central South University. Patients with a previous radical mastectomy, inflammatory breast cancer, preoperative chemotherapy or radiotherapy, an American Society of Anesthesiologists Physical Status of III–IV or greater, immunologic or endocrinologic disease, or any contraindications to propranolol or parecoxib were excluded from participating in the study. Ultimately, 106 patients were enrolled in this section (Fig. 1). Clinical data regarding patient characteristics and surgical procedures were recorded (Table I).

For the in vitro study, we further recruited 17 women diagnosed with primary breast cancer who newly registered for hospitalization at the Third Xiangya Hospital of Central South University, and they met the same exclusion criteria as described above.

The trial was approved by the Medical Ethics Committee of the Third Xiangya Hospital of Central South University and was registered with the Chinese Clinical Trial Registry (ChiCTR-IPR-14005271). All patients were required to provide written informed consent.

In the clinical trial, patients were randomly assigned to control, propranolol, parecoxib, and propranolol plus parecoxib groups. In the propranolol group, patients were treated with oral propranolol (20 mg three times daily) from the day of surgery until the third postoperative day. In the parecoxib group, patients received i.v. parecoxib (40 mg once daily) from the day of surgery until the second postoperative day. In the propranolol plus parecoxib group, patients received propranolol and parecoxib by the oral and i.v. routes, respectively. Peripheral blood samples (10 ml) were collected from each patient early on the morning of surgery (preoperation [Pre-OP]) and at the end of surgery (end of operation [End-OP]), as well as in the early morning on the first (postoperative day [POD] 1), third (POD3), and seventh (POD7) postoperative days.

For each time point, ELISA was used to measure the plasma levels of E, NE, and PGE₂ in 2 ml blood samples that were collected from patients in the control group (n = 10), the propranolol group (n = 10), the parecoxib group (n = 10), and the propranolol plus parecoxib group (n = 10). The ELISAs were performed in 96-well microtitration plates according to the manufacturer’s instructions. Commercial ELISA kits were purchased from R&D Systems (Minneapolis, MN) for the E and NE assays and from Assay Designs (Ann Arbor, MI) for the PGE₂ assay. All samples were assayed in triplicate.

Blood samples (2 ml) were collected from all patients at the Pre-OP, POD1, POD3, and POD7 time points, and PBMCs were isolated thereafter. Real-time PCR (RT-PCR) was used to determine the mRNA levels of FOXP3 and CTLA-4. Total RNA was extracted using TRIzol (Life Technologies), according to the manufacturer’s instructions. The RNA (1 μg) was then reverse transcribed with RT random primers using a High Capacity cDNA Reverse Transcription Kit; the cDNA products were used for RT-PCR amplification with the following primers: 5′-CTGACCAAGGCTTCATCTGTG-3′ and 5′-ACTCTGGGAATGTGCTGTTTC-3′ for the FOXP3; 5′-TACCCACCGCCATACTACCT-3′ and 5′-AACAACCCCGAACTAACTGC-3′ for the CTLA4; and 5′-TGACGTGGACATCCGCAAAG-3′ and 5′-CTGGAAGGTGGACAGCGAGG-3′ for the β-actin.

Quantitative real-time PCR was performed on a PRISM 7500 Real-Time PCR System (Applied Biosystems) using SYBR Select Master Mix (Applied Biosystems), according to the manufacturer’s protocols. The transcript generated from the β-actin gene was used as an internal control. The relative level of gene expression was represented as ΔCt = Ct^gene − Ct^reference, and the fold change of gene expression was calculated by the 2^−ΔΔCt method. Experiments were repeated in triplicate.

PBMCs were isolated from 2-ml blood samples using standard Ficoll density gradient centrifugation. Then, the PBMCs were incubated with FITC-conjugated anti-CD4, PE-Cy7–conjugated anti-CD25, allophycocyanin-conjugated anti–CTLA-4, and the appropriate fluorochrome-conjugated mouse IgGs as isotype controls for 20 min at room temperature in the dark. Intracellular staining of FOXP3 was performed with a PE-conjugated FOXP3 Ab, according to the manufacturer’s instructions. All of the Abs were purchased from BD Biosciences (San Jose, CA). Data were acquired with a BD FACS Canto II Flow Cytometry system. Results were analyzed with FACSDiva 6.1.3 software.

The assays were conducted, as described previously (29, 30). Blood (10 ml) was obtained from 29 patients at the Pre-OP and POD7 time points and divided into duplicate aliquots (5 ml × 2). PBMCs were extracted from the blood by centrifugation. In one of the aliquots of the PBMC samples, CD25^high cells were depleted using magnetic separation with MACS CD25 microbeads (Miltenyi Biotec), following the manufacturer’s instructions, and the efficacy of depletion was determined by flow cytometry (Fig. 4G, 4H). Thereafter, the cell viability of the two aliquots of PBMCs was determined by a trypan blue exclusion cell-counting assay. The cell densities were adjusted to 3.5 × 10⁶ cells/ml with AIM-V serum-free medium (Life Technologies). CD25^high-depleted or undepleted cells (100 μl) were cultured in triplicate in the presence of both CD3 mAb (2.5 μg/ml) and IL-2 (1 μg/ml) for nonspecific stimulation or exposed to a pool of purified MUC1 (1 kU/ml), P53 (1 μg/ml), and HER2/neu (1 μg/ml) for specific Ag stimulation in 96-well ELISPOT plates (Multiscreen IP plate; Millipore) that had been precoated with IFN-γ mAb (U-Cytech; Dutch). Wells containing PHA (1 μg/ml) or serum-free medium were used as positive or blank controls, respectively. The culture plates were placed in a humidified 5% CO₂ incubator at 37°C for 48 h, and the Ag-specific IFN-γ secretion was determined by an ELISPOT assay using an ImmunoSpot analyzer (Cellular Technology). The activated CD4⁺ T cells secreting IFN-γ were measured as spots per well at the single-cell level. The number of spots in the negative control, subtracted from the number of spots in the experimental wells, was recorded as the number of final spot-forming cells (SFC) per 10⁶ PBMCs, and positive responses were defined as having >5 SFC per 10⁶ PBMCs after subtracting the control value and at least a 50% increase compared with the control. Patients who had positive responses in the total CD4⁺ T cells were classified as responders (Fig. 7B). The number of SFC in CD25^high-depleted experimental wells after subtracting the value of the undepleted wells was considered to represent the inhibitory effects of CD4⁺CD25^high Treg on total CD4⁺ T cells; an increase of at least 50% over the undepleted wells was defined as a positive effect of depletion. Similarly, patients who had positive depletion effects of Tregs were identified as suppressors (Fig. 7B).

In the in vitro study, PBMCs were separated from 20 ml blood samples of 17 patients by centrifugation and resuspended at 10⁶ cells/ml in AIM-V serum-free medium (Life Technologies) containing 1% supplemental l-glutamine and 100 U/ml human rIL-2 (R&D Systems). Twenty-four–well, flat-bottom microtiter plates (Costar 3596, Cambridge, MA) were coated overnight at 4°C with anti-CD3 mAb (Santa Cruz; clone PC3/188A) (10 μg/ml) and anti-CD28 mAb (Santa Cruz; clone M-20) (5 μg/ml) in 0.5 ml PBS per well. Thereafter, 1 ml PBMCs (10⁶ per well) were cultured in the 24-well plates precoated with anti-CD3/CD28 with various concentrations of E (Sigma-Aldrich, St. Louis, MO) (0, 1, 5, 10 μM) or propranolol (Sigma-Aldrich) (1, 10, 20 μM), respectively. Each concentration was repeated in triplicate. The effect of propranolol was assessed at an E concentration of 5 μM. Propranolol was administered 30 min before E treatment. These concentrations were determined after a trial of different doses (0.1–50 μM for E and 0.1–100 μM for propranolol). We determined that high concentrations of E (50 μM) or propranolol (100 μM) caused marked cell death and the cell viability was too poor, whereas low concentrations (0.1 μM) seemed to have little effect on cell proliferation. After 3-d incubation at 37°C and 5% CO₂ in a humidified incubator, cells were washed and continuously expanded with anti-CD3/CD28 plus 100 U/ml IL-2 for another 4 d. On day 7, PBMCs were harvested, and the percentage of CD4⁺CD25⁺FOXP3⁺ Tregs was determined by flow cytometry, as described above.

Statistical analyses were performed using the SPSS statistical package (version 18.0; SPSS, Chicago, IL). Continuous variables were defined as means ± SD if they were normally distributed; otherwise, median values and interquartile ranges (25th–75th percentile) were represented. We used ANOVA, two-way ANOVA, paired or unpaired t tests, Mann–Whitney tests, Wilcox signed ranks tests, Fisher’s exact tests, and McNemar tests, as appropriate. The p values <0.05 were considered statistically significant in all studies, and all p values were two tailed.

The perioperative levels of E, NE, and PGE₂ in breast cancer patients with or without propranolol and parecoxib treatment were measured by ELISA (Figs. 1, 2, Table II). As shown in Fig. 2A and 2B, there were no significant differences among the three groups at the corresponding time points for either E or NE levels. The End-OP and POD1 time points showed sharply increased levels of E and NE compared with the Pre-OP time point (p < 0.05). E and NE then returned to levels that were similar to the Pre-OP time point at POD3 and POD7. Even though PGE₂ level increased at End-OP and POD1 in the control group (p < 0.05), but not in the parecoxib or the propranolol plus parecoxib groups, no significant differences were observed among the groups (Fig. 2C). These results suggested that the levels of E, NE, and PGE₂ all increased after surgery; moreover, parecoxib or propranolol plus parecoxib treatment had modest inhibitory effects on postsurgical PGE₂ level, whereas propranolol or propranolol plus parecoxib treatment had no obvious effects on the production of E and NE.

We then investigated the effects of propranolol and parecoxib on two Treg-related functional markers (FOXP3 and CTLA-4) (Fig. 3, Table III). Both in the control and parecoxib groups, the peripheral FOXP3 mRNA level decreased temporarily at POD1 (p < 0.05); these levels then increased to a much higher level at POD7 compared with the Pre-OP time point (p < 0.05). However, no significant time-dependent changes in the propranolol or propranolol plus parecoxib group were observed (p > 0.05); in these groups, FOXP3 mRNA level was significantly lower than in the control group at POD7 (p < 0.05) (Fig. 3A). As shown in Fig. 3B, no significant changes in the CTLA-4 mRNA level were observed between the control group and the treatment groups at any time point. These results suggested that propranolol treatment (alone or in combination with parecoxib), instead of parecoxib alone, was able to decrease the FOXP3 mRNA level, but not the CTLA-4 mRNA level, in the peripheral blood of patients at POD7.

Flow cytometry was used to measure the perioperative levels of circulating CD4⁺CD25⁺FOXP3⁺ and CD4⁺CD25^highFOXP3⁺ Tregs (as percentages in total CD4⁺ T cells) (Figs. 4, 5, Table III). Both CD25⁺ and FOXP3⁺ were gated accordingly using blank controls and isotype controls (Fig. 4B, 4C). The CD4⁺CD25^high was defined by a CD25-staining intensity in the CD4⁺ T population that was greater than in CD4⁻ cells (29). As shown in Fig. 5, the percentages of both Treg subsets decreased at POD1 compared with the Pre-OP values (p < 0.05); moreover, they began to increase and finally reached a significantly higher level at POD7 than the Pre-OP level in the control group (p < 0.05). A similar result was observed in patients treated with parecoxib alone. However, the percentage of Tregs in the propranolol group and the propranolol plus parecoxib group did not increase at POD7 compared with the control group. There were some typical examples, presented in flow cytometry images, of CD4⁺CD25⁺FOXP3⁺ Treg percentages at both Pre-OP and POD7 time points for all four groups (Fig. 6). We also labeled the surface molecule CTLA-4 and detected the frequency of CD4⁺CD25⁺CTLA-4⁺FOXP3⁺ Tregs, but no significant changes were observed (data not shown). These observations indicated that propranolol treatment alone or in combination with parecoxib could block surgical stress–induced increases in circulating Tregs after the seventh postoperative day.

Because propranolol treatment alone exerted an obvious effect on Tregs, we further tested the effects of propranolol on the TA sp. act. of perioperative Tregs with ex vivo highly sensitive ELISPOT assays (Fig. 7, Table IV). IFN-γ secretion by CD4⁺ T cells was measured perioperatively, and the impact of CD25^high Tregs on the responses of CD4⁺ T cells was analyzed in the control and propranolol groups. We observed no differences in each group at each time point after nonspecific stimulation by CD3 plus IL-2 (Fig. 7A).

We next tested the CD4⁺ T cell responses to TA-specific Ags; the magnitude of TA-specific experimental wells ranged from 6 to 328 SFC/10⁶ PBMC. At POD7, decreased CD4⁺ T cell function was observed compared with the Pre-OP time point in the control group (p < 0.05) (Fig. 7C, Table IV), whereas in the propranolol group the total CD4⁺ T cell function at POD7 did not change significantly (Fig. 7D). In addition, the total CD4⁺ T cell function was significantly higher in the propranolol group than in the control group at POD7 (p < 0.05). We also found a decrease in the percentage of TA-specific responders (percentage of responses to tumor Ags) at POD7 in the control group (p < 0.05), but not in the propranolol group (Fig. 7E). However, CD25^high Treg-induced suppression increased at POD7 compared with the Pre-OP time point in the control group (p < 0.05), but no difference between the two time points was observed in the propranolol group (p > 0.05, Table IV). Accordingly, the proportion of suppressors (percentage of suppressed responses) significantly differed between the two groups at POD7 (p < 0.05, Fig. 7F). Interestingly, we observed six patients whose TA-specific responses of CD4⁺ T cells were completely inhibited by CD25^high Tregs at POD7 in the control group (6 of 15), whereas only two patients with such responses were found in the propranolol group (2 of 14). These data suggested that the TA-specific Treg activity was increased after surgery, thereby suppressing the CD4⁺ T cell response. However, the immunosuppressive effect exerted by postsurgical Treg responses could be alleviated by propranolol.

PBMCs were exposed to varying doses of E (0, 1, 5, 10 μM) for 3 d and sequentially cultured for an additional 4 d in vitro. On day 7, the percentage of CD4⁺CD25⁺FOXP3⁺ Tregs was analyzed by flow cytometry to determine the effect of E on Treg proliferation. At that time, as shown in Fig. 8A, the percentage of CD4⁺CD25⁺FOXP3⁺ Tregs significantly increased in the presence of E (1, 5, 10 μM) relative to the absence of E (0 μM) (p < 0.05). In addition, there is a trend for increased proliferation with high concentrations (5, 10 μM) in comparison with the low one (1 μM) (p < 0.05). Interestingly, no significant differences were observed between the two high doses (5, 10 μM) (p > 0.05).

Second, to assess the ability of propranolol to block the E-induced elevation of CD4⁺CD25⁺FOXP3⁺ Tregs, three increasing doses of propranolol (1, 10, 20 μM) were administered 30 min prior to 5 μM E treatment, respectively. As shown in Fig. 8B, although the low dose of propranolol (1 μM) did not affect the upregulating effect of E (5 μM) on CD4⁺CD25⁺FOXP3⁺ Treg proliferation (p > 0.05), both the 10 and 20 μM doses effectively attenuated the E-induced promotion (p < 0.05), also with a significant dose-dependent reduction of the percentage of CD4⁺CD25⁺FOXP3⁺ Tregs (p < 0.05). Besides, propranolol treatment (10, 20 μM) alone had no direct effect on Treg proliferation relative to the unexposed control (0 μM) (data not shown).

These results revealed that 3-d treatment with E generated an enhancement of Treg proliferation that could be inhibited by the addition of propranolol.

For patients with primary breast cancer, surgery is the crucial and necessary treatment for eradicating the primary tumor. However, an increased risk exists for promoting pre-existing micrometastases in the perioperative period (4), which is attributed to the immunosuppressive effects caused by physiological mechanisms, such as excess postoperative release of CAs and PGs (2, 3, 6, 7). Therefore, it is critical to provide perioperative drug treatment to block CAs or PGs with the goal of alleviating their inhibitory effects on the antitumor immune response. To the best of our knowledge, the current study is the first to demonstrate the effect of the perioperative use of propranolol or parecoxib on the number and activity of peripheral Tregs in breast cancer patients who underwent a radical mastectomy.

In this study, perioperative drug treatment with propranolol or parecoxib (alone or in combination) was administered on the morning of surgery and was terminated 2 or 3 d after surgery. Although it has been demonstrated that immunosuppression begins even before surgery due to psychological distress and other related mechanisms (5), we chose this short-term perioperative drug treatment to target immunosuppression that was induced by the surgery itself. It has been reported that the stress response and inflammatory response markedly increase postoperatively and last for 2–3 d (7). In addition, short-term perioperative propranolol or parecoxib administration has been clinically proven to be safer than chronic drug administration and has additional benefits, such as reduction of the dosage of anesthetic and analgesic agents and protection of patients from stroke and cardiac complications (31).

Although both CAs and PGs increased postoperatively, propranolol, rather than parecoxib, appeared to alleviate the increase in the surgery-induced Treg response at POD7. It has been suggested that the surgical stress response has a greater influence on peripheral Tregs than the inflammatory response. It was important to show that the percentages of peripheral Tregs and FOXP3 mRNA expression decreased temporarily at POD1 and then increased markedly at the very time when the stress response abated. Ogawa et al. (7) reported that CD8⁺ suppressor T cells increased on the second day after surgery (when elevated CA levels had already returned to baseline) in gastrointestinal cancer patients. Additionally, MacConmara et al. (32, 33) demonstrated a significant increase in circulating Tregs on day 7 in trauma patients, but no increase on the first day following injury. The following mechanism may be responsible for the aforementioned changes in Treg levels. Augmented CAs levels (induced by surgical stress) cause gentle and temporary CMI activation in the very early phase; however, sustained stimulation can result in immunosuppression by elevating Treg levels.

Indeed, convincing evidence suggests that it is possible for CAs (stress) to affect Tregs in in vivo environment and in ex vivo cultures (22, 34, 35). The effects of CAs on Tregs are achieved through both indirect and direct mechanisms. Indirect mechanisms are exemplified partly by the fact CAs regulate TGF-β generation; TGF-β is crucial to the induction and proliferation of FOXP3⁺ Tregs. One in vitro study has shown that NE can decrease or increase hepatocyte production of TGF-β in a dose-dependent manner (36). Bhowmick et al. (34) suggested that the number of peripheral Tregs is mediated by sympathetic nervous system signals, and TGF-β acts as a bridge between the sympathetic nervous and immune systems. And TGF-β was decreased in PBMCs by short-term exposure to E at 24 h, whereas it was increased at day 11 in long-term E cultures; a significant increase in the expression of FOXP3 mRNA in cultures subjected to 11 d of E exposure has also been observed (22). These stress hormones are also believed to exert their effects directly on Tregs through surface β-AR stimulation and intracellular cAMP/protein kinase A signal activation (19, 37, 38). It has been shown in a human trial that acute psychological stress (lasting for 12 min) directly decreased the percentage of peripheral CD4⁺FOXP3⁺ regulatory T cells, mainly by activating β1-ARs (35). The present work indicated β-AR involvement in CA-induced Treg alterations by demonstrating that the administration of propranolol, a β-adrenergic antagonist, prevented the elevation of Tregs observed following surgery or in vitro E exposure.

While acknowledging the β-AR–based direct mechanism, we further focused our efforts on studying the ability of propranolol to block the E-enhancing effect on Tregs in vitro and provided more direct evidence to strengthen the in vivo observations. PBMCs from breast cancer patients were exposed to varying doses of E (0, 1, 5, 10 μM) or three increasing doses of propranolol (1, 10, 20 μM) administered 30 min prior to E (5 μM) for 3 d and were sequentially cultured for an additional 4 d. Thereafter, the percentage of CD4⁺CD25⁺FOXP3⁺ Tregs was determined. E (1, 5, 10 μM) significantly increased the Treg percentage on day 7 with a dose-dependent effect to some extent (no more than 5 μM). Furthermore, 10 μM propranolol attenuated the enhancement effect of E (5 μM) on Treg proliferation, which was completely blocked by 20 μM propranolol, however. The low dose of propranolol (1 μM) had no obvious effect on the E-induced promotion. Thus, the decreased frequency of Treg observed after treatment with 20 μM propranolol and concomitant E (5 μM) may represent a fallback effect that occurs following sufficient blockade of the β-AR. Our in vitro study demonstrated that CAs could promote Treg proliferation and that propranolol treatment was capable of mediating preventive effects for E-induced elevation of Tregs.

CTLA-4, as a CD28-family receptor that binds to B7, is a central mediator of Treg function (39). We did not observe obvious surface CTLA-4 expression in the population of CD4⁺CD25⁺FOXP3⁺ Tregs. One possible explanation is that the presence of functional CTLA-4 on the cell surface is infrequent (40) and undetectable in flow cytometry (<0.1%). It has been suggested that CTLA-4 expression was detectable early after activation by allogeneic dendritic cells (16 h) on the surface of CD4⁺CD25⁺ T cells (41). Additionally, peripheral CTLA-4 mRNA expression was not significantly different in the pre- and postoperative periods and did not differ among the groups.

Furthermore, the combination of propranolol and parecoxib did not exhibit higher efficacy than propranolol alone, indicating that these two drugs may not be synergistic. This finding differs from earlier animal studies demonstrating that only the combined use of propranolol and etodolac had a significant effect on postoperative immunosuppression and the recurrence-free survival rate (14–18). We hypothesize that two main reasons may account for this discrepancy. First, previous animal research studied the nonspecific cytotoxicity of NK cells on tumor cells, as all animals were injected with tumor cells after the surgical procedure and therefore had no adaptive immunity to the malignancy. Nevertheless, Tregs possess specific immunoprotective competence to tumors and differ from NK cells in the response to CAs or PGE₂. As reported in one study, nadolol blocks surgery-induced decreases in B and T cell proliferation but has no effect on the suppression of NK cell activity in rats (6). Second, the perioperative secretory status of CAs and PGE₂ may differ between humans and animals. The inflammatory responses to surgical incision are most likely more severe in animals than in humans. The reason why the increase of PGE₂ was not as robust as that of E and NE in our study may be due to the fact that a large quantity of PGs is secreted by the tumor itself, which partially masks the effect of surgical trauma on serum PGE₂ level after tumor resection. In addition, a slightly elevated postsurgical PGE₂ level was not associated with increased Tregs, possibly because surgical stress induced a decrease in Tregs at POD1, which antagonized the positive effect of PGE₂ on Tregs. Most importantly, it has been reported that propranolol completely inhibits surgery-induced metastasis in an animal model of ovarian carcinoma (42), and evidences suggest that the use of β-adrenergic antagonists reduces cancer progression via several mechanisms, including effects on tumor biology, angiogenesis, and CMI (24, 25, 27, 43). These abovementioned results suggest that safe, inexpensive, and receivable β-antagonists are likely to be effective in preventing the micrometastasis, supporting our conclusion that a single propranolol administration in the perioperative period can be immunoprotective and beneficial for breast cancer prognosis.

We also examined the effects of propranolol on perioperative Treg activity, with respect to the inhibition of the specific CD4⁺ T cell response to breast cancer Ags such as MUCI, P53, and HER2/neu, using ex vivo ELISPOT assays. Generally, these three TAs, which abundantly express CD4⁺ T cell epitopes, are overexpressed in breast cancer cells, and they can efficiently restimulate a broad repertoire of pre-existing effective CD4⁺ T cells (44, 45). Our findings demonstrated that specific CD4⁺ T cell responses to TAs were suppressed by the augmented activity of CD25^high Tregs at POD7 and concurred with the findings of a previous clinical study showing that IFN-γ production by CD4⁺ T cells was decreased postinjury; CD25^high Treg depletion could reverse this effect, and the injury augmented the potency of Tregs to inhibit T cell proliferation at day 7 after trauma (33). Tashiro et al. (8) also demonstrated that Con A- and PHA-stimulated T cell proliferation were significantly decreased 7 d after esophagectomy. Importantly, we showed that propranolol treatment alleviated surgical stress–induced elevation of Treg activity and the accompanying suppression of CD4⁺ T cell responses to further confirm the utility of perioperative propranolol treatment in breast cancer patients.

In conclusion, this study demonstrated that propranolol can efficiently block the enhancement of circulating Treg responses and attenuate the immunosuppressive effects of surgical stress in patients who underwent a radical mastectomy. However, a 3- to 5-y follow-up study to evaluate long-term relapse-free survival rate is needed to address the clinical significance of perioperative treatment with propranolol in breast cancer patients.

We thank Prof. Shuang Liu and Dr. Jingxuan He for assistance with flow cytometry and IFN-γ ELISPOT assays. We are also grateful to Prof. Wei Tian for critical reading of this manuscript and helpful discussions.

The authors have no financial conflicts of interest.