Using the nationwide National Health Insurance Research Database (NHIRD) of Taiwan, we conducted a population-based nested case–control study. The study population consisted of 2.2 million subjects with heart or lung disease who held Taiwan National Health Insurance. Complete outpatient and inpatient electronic claim records, individual diagnoses, surgical procedures, and prescribed medications are available in the NHIRD database. The study cohort consisted of hypertension patients taking at least one anti-hypertensive drug for more than 300 days within 1 year. All patients who were longitudinally followed from the NHIRD between January 2000 and December 2010 and who were aged 18 years or older on 1 January 2000 were eligible for inclusion in this study. The study was approved on 17 December 2012 by the Institutional Review Board of the National Health Research Institutes (Taiwan) with a waiver of informed consent (protocol no. EC1011008-E). All procedures were followed in accordance with the ethical standards of the Institutional Review Board of the National Health Research Institutes and with the Helsinki Declaration of 1975.

We identified newly diagnosed sepsis by ICD-9-CM codes as evidence of bacterial or fungal infection and acute organ dysfunction (30–33). This definition of sepsis cases has been used in previous studies and was validated in a linked survey database. The index date referred to the first date of sepsis diagnosis. The one-year period preceding the index date was used for the assessment of beta-blocker exposure status. For each case, one control was randomly selected using the incidence density sampling method and matched by index date, 5-year age group, and sex. Nonselective beta-blockers (ATC code C07AA) and selective beta-blockers (ATC code C07AB) dispensed before the index date were identified from the NHIRD, and drug exposure was defined by having a drug prescription for at least 7 days. Exposure was defined using four different time frames. Current user status referred to patients with a selective or nonselective beta-blocker prescription that was filled within 90 days of the index date. Recent user status referred to patients with a selective or nonselective beta-blocker prescription filled between 90 and 365 days prior to the index date.

Seven-week-old female C57BL/6 J mice (20–25 g) were randomized into two groups: mice were pretreated with normal water or 0.375 mg/mL atenolol in drinking water (equivalent to ~75 mg/kg/day based on 5 mL daily water intake) for 14 days (n = 20 in each group). Mice were injected intraperitoneally with a single dose of lipopolysaccharide (LPS, 15 mg/kg; E. coli 055: B5 purchased from Sigma-Aldrich) on day 15 to induce sepsis as previously described (34). The survival rate of the LPS-induced septic mice was compared between the control group and the study group administrated with atenolol. At 12 h post-LPS injection, their inflammatory cytokines and immunomodulatory checkpoint molecules in serum were measured by the ProcartaPlex multiplex assay (Level Biotechnology, Taiwan) and the PD-L1 expression in F4/80 positive cells was detected using flow cytometry. All operations were executed in compliance with the applicable regulations. This study was approved by the Institutional Animal Care and Use Committee of Cardinal Tien Hospital, New Taipei, Taiwan (protocol no. 107A-006).

Human monocyte THP-1 cells (RRID: CVCL_0006) were obtained from ATCC and cultured in RPMI 1640 medium (Gibco, Invitrogen), containing 10% heat-inactivated fetal bovine serum (FBS), 2 mM L-glutamine, 20 IU/mL penicillin, 20 μg/mL streptomycin, and 0.05 mM 2-mercaptoethanol. Human monocyte-derived macrophages were differentiated from THP-1 cells by culturing for 7 days in completed RPMI 1640 medium containing 20 ng/mL 12-O-tetradecanoylphorbol-l3-acetate (PMA). Cells were pretreated with or without atenolol for 7 days and then stimulated with or without 100 ng/mL LPS, as indicated.

For ROS detection, cells (5 × 10⁵) were incubated with 5 μM H2DCFDA dye (Invitrogen, NY, USA) in 6-well plates at 37°C in a 5% CO2 incubator for 30 min. Following removal of the H2DCFDA dye, cells were collected by centrifugation at 300 g, resuspended in PBS, and then detected by flow cytometry. The positive rate of reactive oxygen species (ROS) was analyzed and computed using the Cytobank software (Cytobank, Beckman Coulter). For phagocytosis assay, cells were co-cultured with 100 μg pHrodo-conjugated Escherichia coli (Invitrogen, NY, USA) at 37°C for 2 h. Cells were washed with PBS, harvested by centrifugation, and detected by flow cytometry. The percentages of pHrodo-positive cells were analyzed using the Kaluza software (Beckman Coulter).

Cells were lysed with a RIPA buffer containing 150 mM NaCl, 50 mM Tris–HCL, 0.1% SDS, 1% Triton X-100, 0.5% Na deoxycholate, 2 mM EDTA, and 1× protease/phosphatase inhibitor cocktail (Roche Diagnostic). Whole cell lysates were collected by centrifugation at 13,000 rpm, heated in sample buffer at 95°C for 5 min, separated by SDS electrophoresis, and then transferred to nitrocellulose membranes by semi-dry electroblotting. The membranes were blocked with 5% non-fat milk in Tris-buffered saline containing 0.1% Tween 20 (TBST), and incubated overnight at 4°C with the corresponding primary antibodies. The membranes were then washed with TBST, incubated with horseradish peroxidase-conjugated secondary antibodies, washed again, and developed with enhanced chemiluminescence. The images were acquired by an Imaging System. All antibodies used in this study (PD-L1 Cat# 13684, RRID: AB_2687655; phospho-NF-κB p65 Cat# 3033, RRID: AB_331284; NF-κB p65 Cat# 8242, RRID: AB_10859369; phospho-Stat3 Cat# 9145, RRID: AB_2491009; STAT3 Cat# 9139, RRID: AB_331757; β-actin Cat# 3700, RRID: AB_2242334) were obtained from Cell Signaling Technology (MA, United States).

Total RNA was extracted from cells with the Genezol reagent (Genaid, Taiwan) and was submitted to reverse transcription using the TAKARA cDNA Reverse Transcription kit (Takara, Japan). Quantitative PCR was performed with the fluorescent dye SYBR Green methodology and a Rotor-Gene Q real-time PCR system. Primers for gene expression analysis were purchased from IDT (Singapore). These mean Cq values were used to normalize the steady-state target mRNA concentrations to those of the beta-actin by the 2(−ΔΔCq) method.

Data were expressed as mean and standard deviation or numbers with percentages. We compared the baseline characteristics between case and control groups by Student’s t test or chi-squared test according to the type of variables. Conditional logistic regression analysis was used to estimate type, dosage and duration of beta-blockers with regard to the risk of sepsis. The crude and adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were calculated with conditional logistic regression. ORs were adjusted for Charlson Comorbidity Index (CCI), age, and the use of other antihypertensive drugs. On the basis of the duration of beta-blocker treatment, we divided the patients into current users (≤ 90 days) and past users (91–365 days) and took nonusers as the reference. The mean daily dosage of beta-blocker treatment was calculated based on defined daily dose (DDD) divided by the cumulative period, and divided into two categories as less than 0.5 DDD/d and more than 0.5 DDD/d. A two-sided p value of less than 0.05 was considered statistically significant. SAS version 9.4 (SAS Institute, Cary, NC, United States) and GraphPad Prism Software (San Diego, CA, USA) were used for the analyses.

Between 2000 and 2010, a total of 64,070 cases diagnosed with sepsis were identified from the NHIRD and matched to another 64,070 control subjects. The case group had more underlying comorbidities, including acute myocardial infarction, congestive heart failure, peripheral vascular disease, stroke, dementia, rheumatologic disease, peptic ulcer disease, renal disease, liver disease, malignancy and diabetes mellitus; and higher CCI scores than the control group (Table 1). The associations between the use of selective and nonselective beta-blockers and the risk of sepsis are summarized in Table 2. For the selective beta-blocker users, current users were associated with a lower risk of sepsis than the non-users (adjusted OR (aOR), 0.842; 95% CI, 0.755–0.939), and this trend was also observed with the recent users (aOR, 0.773; 95% CI, 0.737–0.810). In addition, the mean daily dosage ≥0.5 DDD was associated with a lower risk of sepsis (aOR, 0.7; 95% CI, 0.676–0.725), but the risk of sepsis was not changed by beta-blockers at the mean daily dosage <0.5 DDD (aOR, 1.009; 95% CI, 0.97–1.051). In contrast, for the nonselective beta-blocker users, current users were associated with a higher risk of sepsis than non-users (aOR, 1.246; 95% CI, 1.141–1.36), but recent users were associated with a lower risk of sepsis than non-users (aOR, 0.955; 95% CI, 0.914–0.997). A higher risk of sepsis among nonselective beta-blocker users than non-users was observed at different dosages. The risk of sepsis among patients receiving different selective and nonselective beta-blockers is shown in Table 3. Among the six selective beta-blockers, metoprolol, atenolol and bisoprolol were found to be associated with lower risks of sepsis than non-users (metoprolol, aOR, 0.767; 95% CI, 0.607–0.967; atenolol, aOR, 0.915; 95% CI, 0.840–0.997; bisoprolol, aOR, 0.922; 95% CI, 0.852–0.998). For the other three selective beta-blockers, betaxolol, acebutolol and esmolol, the risks of sepsis were similar to control groups. However, the risk of sepsis did not vary among individual nonselective beta-blockers.

As atenolol was recorded as a commonly used beta-blocker in the NHIRD database and can reduce the incidence of sepsis in patients, the effects of atenolol on septic mortality were further examined in an animal model (Figure 1A). We monitored mouse survival after LPS injection with or without atenolol pretreatment for 14 days. Based on Kaplan–Meier survival curves, mice with LPS-induced sepsis had a 7-day survival rate of 42%. The reduction in sudden death on the second day after LPS injection was particularly noticeable. Atenolol administration increased the 7-day survival rate to 80% relative to that in mice with un-pretreated sepsis (Figure 1B). The levels of cytokines, pro-inflammatory IL-1β, IL-6, TNF-α, and anti-inflammatory IL-10, were also strongly increased after 12 h from LPS stimulation relative to those measured in saline-injected control mice (Figures 1C–F). Even though statistical significance was not reached, atenolol administration partially attenuated the LPS-induced increase in IL-1β, IL-6 and TNF-α, while there was no effect on IL-10 release. Several studies have indicated that serum soluble programmed cell death 1 ligand (sPD-L1) and soluble T cell Ig and mucin domain protein 3 (sTIM-3) levels were positively correlated with the severity of sepsis and can be used as a biomarker to predict the immunosuppressive-induced mortality in sepsis (35–37). Interestingly, atenolol administration significantly reduced the sPD-L1 level and partially diminished the sTIM-3 level in LPS-induced septic mice (Figures 1G,H). Furthermore, in the septic mice that were pretreated with atenolol, there was a reduction in the percentage of F4/80 positive macrophages that expressed surface PD-L1, with a decrease of approximately 20%, although this change did not reach statistical significance (Figure 1I). Since sPD-L1 and sTIM-3 are involved in the regulation of immune tolerance, we compared the correlation of sPD-L1 and sTIM-3 levels with inflammatory cytokines. Regardless of atenolol treatment, sTIM-3 was positively correlated with inflammatory cytokines in septic mice and was statistically significant with IL-10 (Figure 2). Unexpectedly, sPD-L1 was negatively correlated with inflammatory cytokines and showed statistical significance with IL-6 in septic mice, whereas atenolol treatment reversed this phenomenon to a positive correlation (Figure 2). Atenolol thus alleviated the imbalance between pro-inflammation and anti-inflammation of the immune system. This suggests that atenolol-fed mice may have lower susceptibility to LPS-driven sepsis and reduced mortality during disease.

Since the reduction of sudden death in the atenolol-treated group was particularly pronounced on the second day after LPS injection, we speculated that atenolol may have a regulatory role in innate immunity. Monocytes and monocyte-derived macrophages are key effectors and regulators of inflammation and innate immune responses. Considering that PD-L1 on monocytes/macrophages was upregulated in septic animals and PD-L1 blockade enhanced bacterial clearance and significantly improved their survival (38), we examined the effect of atenolol on PD-L1 expression in LPS-stimulated THP-1 cells. To simulate long-term use of drugs, we pretreated THP-1 cells with the indicated doses of atenolol for 7 days before LPS stimulation. Interestingly, atenolol treatment downregulated LPS-induced PD-L1 expression in monocytes/macrophages through the NF-κB and STAT3 pathways (Figure 3A). To understand whether atenolol affected LPS-induced dysfunction of monocyte/macrophages, we examined the pathogen clearance by THP-1 cells. However, atenolol had no effect on the phagocytic activity of monocytes/macrophages (Figure 3B). In contrast, atenolol treatment reduced the reactive oxygen species (ROS) level of THP-1 cells in a dose-dependent manner (Figure 3C). Although ROS play essential roles in monocyte/macrophage killing of pathogens (39), ROS production has also been recognized as one of the earliest innate immune responses of host cells in response to infection (40). Indeed, the expression of NF-κB and STAT3-dependent pro-inflammatory and anti-inflammatory genes were also attenuated by atenolol treatment (Figures 3D–G). These results suggest that atenolol was able to attenuate ROS-induced NF-κB and STAT3 activation, leading to a modest inflammatory response. Considering sPD-L1 may reflect PD-L1 levels expressed on the cell surface, our in vitro results also suggested that atenolol may regulate the expression of PD-L1 on monocytes/macrophages, which are the first-line innate immune cells to defend against infection.