FVB.Cg-Tg (Wnt1)1Hev/J female mice (with activated wnt1 signaling) (11) were obtained from the Jackson Laboratory and housed under a controlled temperature in the animal facility of the University of Kentucky. All experimental procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Kentucky. Female mice at the age of 5 weeks were group housed (4 mice/cage) and randomly assigned into seven groups (20 mice/group): control, ethanol exposure, PD-L1 or PD-1 antibody injection, ethanol exposure plus PD-L1 or PD-1 antibody injection, and IgG isotype control group. Mice were exposed to ethanol by feed with an ethanol liquid diet (Bio-Serv, Flemington, NJ), while mice without the ethanol treatment were fed with an isocaloric liquid diet in which maltose was used to substitute for ethanol. The ethanol concentration in the ethanol diet was increased with up-titration by the following: week 1, 2% ethanol; week 2, 4% ethanol; weeks 3 and on: 6.7% ethanol. Diet was provided ad libitum for the experimental period. For ethanol-treated animals, after feed with 6.7% ethanol for one week the blood samples were collected in the early morning from the tails and the alcohol concentration was analyzed by Alcohol Analyzer AM1 (Analox Instruments, MA). For inhibition of PD-L1 or PD-1, rat anti-mouse PD-L1 antibody (rat Immunoglobulin G2b (IgG2b), Leinco Technologies) or PD-1 antibody (rat IgG2a, Leinco Technologies) were administered by intraperitoneal injection at the dose of 200 μg/mouse on the first day of the 6th, 8th and 10th week. Rat IgG2a and IgG2b (Leinco Technologies) at the same dosage were used as isotype/injection control. The treatment regimen was derived from previous studies in which PD-1 or PD-L1 antibody injection at similar dosages inhibited tumor growth with no adverse effects (12–15). The time (22 weeks of age) when 50% of the mice in control group developed mammary tumor was used to set the end point of the study so that the results can be compared to those of previous studies (16). During the experiment period, body weights of mice and general health condition were evaluated and recorded regularly. No apparent toxicity or significant body weight difference were observed among these animals. Mammary tumor size was measured using a caliper in millimeters. The mice were monitored closely after tumor formation. At the end of experiment or when tumor size reached 20 mm in maximum diameter, the mice were euthanized by Ketamine/Xylazine overdose and mammary tumors were collected, measured, and weighted. The tumor samples were thereafter used for histological and biochemical examination or flow cytometry analysis. During the experiment, a few mice (1 in ethanol, 2 in PD-L1, 1 in PD-1 and 2 in isotype control treatment groups) died for unknown reasons.

The method has been published previously (15, 17, 18) with minor modification. The mammary tumor tissues for flow cytometry analysis were minced into small pieces and incubated in digestion solution containing 2 mg/ml of collagenase A (Sigma), 10 μg/ml of DNase I (Roche), and 10% FBS in Dulbecco’s PBS with MgCl₂ and CaCl₂ (Sigma) for 45 min with shaking at 250 rpm at 37 °C. The dissociated tumor tissues were then passed through a 100 μm cell strainer to remove the tissue debris. The cell pellets were resuspended and incubated with 5 ml 1× red blood cell lysis buffer for 4 min. The cells were then washed with 25 ml PBS by centrifuging at 350x g for 5 min, and the single cells were re-suspended in 1 ml of PBS containing 2% FCS for flow cytometry analysis. The antibodies against PD-L1, PD-1, CD45, and CD8 (all from BioLegend) were used for cell population identification. Relative isotype staining was used as negative controls. The single cells were pre-incubated with CD16/CD32 antibody (BioLegend) for 15 min and then stained with CD45 or CD8a antibody (BioLegend) to eliminate non-specific binding of Fc receptors. CD45 staining was used to gate tumor-tissue-infiltrated immune cells (CD45⁺) from non-immune cells (CD45^-). CD8a staining was used to identify CD8 T cell population. For the detection of intracellular cytokine IFN-γ, perforin and granzyme B in CD8 T cells, a previously published protocol was used (15). Briefly, the single cells went through stimulation, fixation and permeabilization steps followed with intracellular staining with the respective antibodies (BioLegend). Acquisition of data was performed on a Cytometers BD Symphony A3 equipped with Diva software and analyzed using FlowJo 10.8 version.

The method has been published previously (15). Briefly, the mammary tumor tissues were fixed with paraformaldehyde (PFA) and embedded in paraffin. Tissue blocks were cut into 5 μm thick sections in a HM 325 rotary microtome (Thermo Scientific). Sections were then deparaffinized and placed in the antigen retrieval solution (Histo-VT One, Nacalai USA). After rinsing thoroughly in 0.1 M PBS, the sections were pre-incubated in a blocking solution and then incubated with mouse PD-L1/B7- H1 antibody (R&D systems). After that, the sections were rinsed and incubated with anti-mouse biotinylated secondary antibody IgG (H+L) (Vector Laboratories). After washing with 0.1 M PBS, the sections were incubated with Avidin/Biotin complex (Vectastain ABC Elite kit; Vector Laboratories) and then chromagen DAB (3′,3′ -diaminobenzidine, Vector Laboratories). The sections were observed under an Olympus CX43 biological microscope. For quantitative assessment of PD-L1 staining, the average optical density (OD) of the regions of interest was measured by Fiji Image J software.

The data were analyzed by SPSS 28 (SPSS) and bar graphs were created using GraphPad Prism 8 (Graphpad Prism). Mammary tumor free curves were generated using the Kaplan–Meier method, and the statistical significance of the difference among the different groups was measured using the Log-Rank test. Tumor size over time was analyzed by general linear model in which repeated measures were used. Univariate factorial ANOVA was used for numerical data analysis and results were presented as (mean ± SD). Partial eta squared (η²p) was used for calculating the effect size. The differences among groups were determined by post-hoc multiple comparisons. The difference between two groups was tested by independent samples t-test, and Cohen’s D effect size was calculated for t-test. A p value < 0.05 was considered significant.

Wnt1 signaling, including both canonical and non-canonical, plays a key role not only in embryogenesis but also in the development of several cancers, such as breast cancer (19). In this study, female FVB.Cg-Tg (Wnt1)1Hev/J transgenic mice with activated wnt-1 signaling at the age of 5 weeks were randomly assigned into 7 treatment groups (20 mice/group): control, ethanol exposure, PD-L1 or PD-1 antibody injection, ethanol exposure plus PD-L1 or PD-1 antibody injection, and IgG isotype control group (see Methods for details). For ethanol-treated animals, the average blood alcohol concentration was 80.17 ± 9.06 mg/dl, a human equivalent dose of binge drinking. At the 22 weeks of age when 50% of the animals in control group developed mammary tumors, tumorigenesis was analyzed by Kaplan-Meier plot shown as the percentages of mammary-tumor-free animals over time. As shown in Figure 1A , ethanol exposure significantly accelerated the formation of early onset mammary tumors than that in control group (solid red line vs solid black line, p < 0.05), which was reversed by the injection of PD-L1 or PD-1 antibodies (purple or orange dashed line, respectively, p < 0.05), while PD-L1, PD-1 antibody alone or IgG treatment groups have no significant differences from control group (p > 0.05). Ethanol substantially shortened the time for the animals to exhibit a 50% incidence of tumors from 22 (in control group) to 9 weeks of age, whereas co-treatment of PD-L1 or PD-1 antibody postponed it from 9 to around 16 weeks. However, as shown in Figure 1B , ethanol exposure apparently did not alter the tumor growth after the tumors were detected.

The results of the animal study suggested that the PD-L1/PD-1 pathway may play an important role in ethanol-enhanced mammary tumorigenicity. Therefore, PD-L1 levels on CD45-negative cells of the tumors (most are tumor cells) between control and ethanol groups were examined by flow cytometry. As shown in Figures 2A, C , ethanol treatment increased the percentage of PD-L1⁺ cells in the tumors compared to that in control group. In addition, ethanol exposure also increased PD-L1 protein levels on these cells ( Figures 2B, D ). To further examine PD-L1 levels in the tumor cells, immunohistochemistry staining was performed to the mammary tumor tissue sections. As shown in Figures 2E, F , PD-L1 expression was higher in the tumor tissues of ethanol treated group than that of control group.

Since the immune cells in the tumor microenvironment may promote or inhibit tumor growth (20) and the host anti-tumor immunity is mainly carried out by CD8 T cells, we next examined the effects of long-term ethanol exposure on tumor-infiltrating CD8 T cells. The flow cytometry data indicated that ethanol exposure did not alter the proportion of these T cells significantly ( Figure 3A ). In addition, since the PD-1/PD-L1 pathway could be used by tumors to escape immune surveillance, we also examined the effects of ethanol exposure on the percentage of PD-1⁺CD8⁺ T cells as well as on the level of PD-1 on tumor-infiltrating CD8 T cells. As shown in Figures 3B–E , ethanol exposure significantly increased the percentage of PD-1⁺CD8⁺ T cells ( Figures 3B, D ) as well as PD-1 protein levels on these cells ( Figures 3C, E ).

CD8 cytotoxic T cells play a vital role in anti-tumor immunity. The tumor-infiltrating CD8 T cells kill the tumor cells by releasing apoptosis-inducing cytotoxic proteins (e.g. interferon gamma, IFN- γ) or granules (e.g. perforin and granzyme B) (21). Therefore, the effects of ethanol exposure on cytotoxic function of CD8 T cells were first examined by comparing the proportion of IFN-γ⁺CD8⁺ T cells in the mammary tumor tissues among different groups. Our data showed the percentage of IFN-γ-producing CD8 T cells were significantly decreased in the mammary tumor tissues of ethanol-exposed mice, which was restored by the co-treatment of PD-L1 or PD-1 antibody ( Figure 4A ). Next, we analyzed the effects of ethanol exposure on CD8 T cells producing perforin or granzyme B. As shown in Figures 4B, C , the percentages of perforin⁺CD8⁺ and granzyme B⁺CD8⁺ T cells were reduced by ethanol exposure, which were also ameliorated by the co-treatment of PD-L1 or PD-1 antibody.