Male CBA mice (ARC, Perth) were maintained in standard cages with irradiated food and water supplied ad libitum. The primary cancer cell line mouse CRC cell (MoCR) was derived from a dimethyl hydrazine-induced primary colon carcinoma in the CBA mouse and maintained in vivo by serial passage in the flanks of CBA mice.28 For passage and experimentation, subcutaneous tumors were teared apart, passed through a filter, treated with Ethylenediaminetetra-acetic acid (EDTA) and washed in phosphate-buffered saline (PBS) to make a single cell suspension. CRLMs were induced by intrasplenic injection of 5×10⁴ tumor cells followed by splenectomy. Metastases are fully established by 21 days following tumor induction (figure 1A,B).

Tissues were collected at days 15 and 21 post-tumor induction and were used to evaluate tumor growth and Tcell immune infiltration by immunohistochemistry (IHC) and fluorescent-activated cell sorting (FACS) (figure 1A).

Mice received tumor induction and were separated into two groups: control (saline) and treatment (captopril). Captopril (pH adjusted to 7.4) was administered daily via intraperitoneal injection adjusted to the body weight of each mouse (750 mg/kg), from the day 1 post-tumor induction to endpoint. Control mice received an equal volume of saline. Tumor growth was assessed by stereology on day 15 or 21 following tumor induction.

At each endpoint, animals were terminally anesthetized; laparotomy was performed; and the abdominal cavity was examined for indications of extrahepatic metastases, paying close attention to the splenic bed. The liver was carefully perfused with 20 mL saline solution, then excised and harvested. Immediately following excision, liver weights were recorded. Livers were then fixed in formalin (10%) (Sigma Aldrich, Castle Hill, New South Wales, Australia) for 24 hours and then transferred to 70% ethanol.

Images of the whole liver were taken to examine the tumor distribution and burden load. The liver was then transversely sliced into sections of 1.5 mm thickness using a tissue fractionator. For large livers (saline control), every second slice was sampled for analysis. For smaller livers (captopril-treated tumors), every slice was taken for analysis. Liver slices were placed on a clear plastic plate; a digital camera (Nikon Coolpix 5000, E500) was used to capture the images, and these were analyzed using image analysis software (Image-Pro Plus, Perth, Australia).

Tumors were visualized as distinctive white/cream-colored areas against the red/brown liver tissue. Each tumor outline was traced using image analysis software to determine the area occupied by the tumor. Stereology technique was used to determine tumor burden.

Formalin-fixed paraffin tissue sections (4 µm) were used with an indirect peroxidase labeling technique (Envision Plus, DAKO, Australia). Following deparaffinization and rehydration, endogenous peroxidase activity was blocked with 3% H₂O₂, and non-specific binding inhibited with 10% normal goat serum (01–6201 Thermo Fisher scientific). Heat-induced epitope retrieval was used. Antigens were visualized by immunohistochemical staining using their respective antibodies diluted as follows: CD3 1:100 (Clone A0452 DAKO), CD4 1:1000 (4SM95 e-Bioscience), PD-L1 1:500 (E1L3N, Cell Signaling Technology) and negative controls were stained without the corresponding primary antibodies. Following primary antibody treatment, sections were incubated with a horseradish peroxidase-labeled polymer secondary antibody. The antigen–antibody complex was visualized by diaminobenzene peroxidase substrate solution (DAKO, Australia). Viable tumor was determined by Haemotoxylin and Eosin (H&E) staining.

Slides were scanned at ×20 magnification (Aperio Scanscope AT Turbo, Leica Biosystems). Image analysis software (Aperio ImageScope) measured the number of positive cells within designated areas. Given the size of immune cells, the Positive Pixel Count v9 (V.9.1) algorithm was applied as a basis for detecting immune marker positivity, with the intensity thresholds adjusted manually to remove background artifacts and to account for variable differences in cell size (hue value (center): 0.1, hue width: 0.33 and color saturation threshold: 0.15). CD3⁺ cell distribution between tumor periphery and core was examined at day 15 (online supplementary figure S1A) and day 21 post-tumor induction. Same conditions were applied to measure CD4⁺ cells at day 15 (online supplementary figure S1B) . The tumor ‘periphery’ was defined to include approximately an equal length of the leading edge of tumor cells and the adjacent stromal interface defined as the inner invasive margin. The ‘core of the tumor’ was defined as the total tumor area excluding the periphery. For each of these regions (periphery/leading edge or core), a total area up to of 100 µm per pixel containing the highest density of IHC⁺ cells was drawn to encompass these areas. When the inner invasive tumor margin (TM) or center of the tumor was smaller than 5 µm per pixel, the positivity of the whole tumor and the invasive margin was considered the same. PD-L1 expression was counted in the membranous and cytoplasmic tumor region using the same staining intensity thresholds (intensity threshold weak: 220–175, medium: 175–100, strong: 100–0, and negative pixels: −1) .29

At endpoints of 15 days following tumor induction, livers were excised as described earlier, and caudate lobes were separated for processing for flow cytometry. Tumors were macroscopically isolated from liver parenchyma, and both tumor and macroscopically tumor-free liver were treated with tissue digest medium (0.25 mg/mL Liberase, Sigma and DNAse, Sigma) at 37 degrees for 40 min. Tissues were then pushed through a stainless-steel mesh, treated with red blood cell lysis buffer, and resuspended in FACS wash buffer (10% bovine serum albumin/5 mM EDTA/0.01% sodium Azide in PBS PH 7). 1×10⁶ cells were stained with antibody cocktails directly conjugated to fluorochromes (CD45-BV510, CD3-PE, CD8-PECy7, PD-1-APC, CD4-APC-Cy7 (BD Bioscience) and CD103-FITC (Miltenyi biotec) and the viability dye, DAPI (additional file 1: online supplementary table 1). Each sample was analyzed on a FACS CANTO II (BD Biosciences) and data were analyzed by FlowJo™ v10. Lymphocytes were identified as live (DAPI negative), single cells that were negative for autofluorescence, CD45⁺ and CD3⁺ (online supplementary figure S2A). Fluorescence minus one (FMO) controls were used to accurately gate CD4⁺ or CD8⁺ T cells and then the expression of PD-1 and CD103 on these T cell subsets (online supplementary figure S2B), including CD3⁺ cells that were expressing neither CD4 nor CD8, termed double-negative T cells (DNTs).

All data were presented as the mean value of each group ±SE of the mean. Correlation matrix was generated using correlation coefficient r (or rs) for each pair of variables: length (µm), area (µm²), TM positivity, tumor core positivity and total positivity. A heat map of R² calculating p values (two-tail) was generated. Statistical analysis was conducted using GraphPad Prism V.8.3.0 (GraphPad Software, San Diego, California, USA) using both parametric and non-parametric analytical tests as appropriate. All statistical tests were two-sided, and a p-value of 0.05 was considered statistically significant.

The stages of metastatic development of the CRLM model have been previously characterized (figure 1B).28 These experiments confirmed previous findings by this laboratory18 that captopril treatment has profound effects over metastatic growth after the metastases were fully established by 21 days following tumor induction (figure 1C,D). In this study, at day 15 endpoint, there was a trend in the reduction of tumor load by 3.94% difference. In contrast, liver collected at 21 days post-tumor induction from the group treated with captopril had significantly less tumor burden than the control (28.61% difference) (figure 1D). In this study, we also examined tumor viability as this is a better indication of treatment efficacy. Quantitation of live tumor area demonstrates that the percentage of viable tumor area in the captopril treatment group was significantly less, compared with the control group at 21 days post-tumor induction (figure 1E,F). Similar to the findings for tumor burden, while there is a trend, no significant differences in tumor viability could be determined at day 15 (figure 1F). Captopril treatment reduced the number of metastases, though this reduction did not reach any significance, the size of metastatic foci was in fact significantly reduced (online supplementary figure S3).

We used IHC to examine changes in CD3⁺ and CD4⁺ lymphocytes within the tumor tissues due to captopril treatment. Significant increases were seen in CD3⁺ lymphocytes in the treated group at day 15 (figure 2A,C). On the other hand, CD4⁺ lymphocytes were significantly decreased at this time point (day 15) as seen in figure 2B,D. Visual examination of the staining shows that lymphocyte infiltration tends to accumulate around the intratumoral vessels (figure 2A,B). Significant increases in the number of CD3⁺ lymphocytes were observed in both TM and intratumoral assessment in the captopril-treated group (figure 2A).

Interestingly, both density and distribution of CD3⁺ tumor-infiltrating lymphocytes inversely correlate with the area of liver metastases (online supplementary figure S4A). While the frequency of CD4⁺ cells in the control group significantly decreased in the same pattern as the CD3⁺ lymphocytes between the two time points; in the treated group, the CD4⁺ T cells did not significantly change compared with day 15, and in fact, at day 21, the frequency of CD4⁺ in the treated group is relatively higher than at day 15, but it is not significantly higher compared with that of the control at day 21 (figure 2C,D). These results suggest that the increased infiltration of CD3⁺ T cell population at day 15 was clearly not CD4⁺ T cells and that at day 21, these are no longer within the tumor in elevated numbers, giving impetus for a thorough investigation of day 15 T cell phenotype changes induced by captopril (figure 2).

Using IHC we also investigated whether the tumor in our model expresses PD-L1 (figure 3A), and whether captopril treatment alters tumor PD-L1 expression. While there is a strong expression of PD-L1 in the tumor, this expression was not significantly altered at any time point due to the treatment (figure 3B) and was not tumor size dependent (figure 3C). No significant difference was found between different populations of tumor cells expressing PD-L1 at different intensities between the control group and captopril at day 15 (figure 3D).

Flow cytometry analysis of dissected tumor tissues was used to identify the phenotype and the subpopulations of CD3⁺ T cells in tumor that were observed by IHC to be modulated by captopril at day 15 (figure 4A). We also analyzed the T cell subpopulation infiltration into liver tissues that had been macroscopically dissected from tumors. During liver harvesting, livers were perfused via portal vein cannulation with 20 mL of saline to remove peripherally circulating lymphocytes from the analysis; thus, data represent tissue-infiltrating cells.

FACS confirmed that captopril treatment significantly increased the infiltration of CD3⁺ T cells into the tumor. Interestingly, these changes were also seen within the liver parenchyma (figure 4B) indicating that CD3⁺ T cells, trafficking from circulation and/or dividing within the tissue in both liver and tumor, are affected by captopril.

The phenotype of CD3⁺ T cells was examined in terms of the proportion of these cells expressing CD4 and CD8 and those expressing neither CD4 nor CD8, termed DNT (figure 4C,D) and were found to be differentially modulated by captopril. Both in the tumor and liver tissues, there is a decrease in the proportion of CD4⁺ and an increase in the proportion of DNT. On the other hand, there is an increased proportion of CD8⁺ T cells in the tumor, while in the liver, this subpopulation decreases (figure 4C,D).

When considering these populations as a proportion of total leukocytes (CD45⁺) in these tissues, we find that in the liver, captopril treatment does not change either CD4⁺ or CD8⁺ subpopulations, whereas the DNT population is significantly increased (figure 4E). In the tumor, we also see a decrease in the CD4⁺ population while the CD8⁺ and DNT subpopulations significantly increased (figure 4F). This may indicate that in the liver tissue, the increase in CD3⁺ T cells induced by captopril is due to the infiltration of DNTs. Within the tumor, however, the changes induced by captopril are more complex and may reflect changes in the TME that affect the ability of each subpopulation to exist there.

To investigate changes in the activation state of the CD3⁺ T cell subtypes, the cells were also stained for expression of Programmed cell death-1 (PD-1), an inhibitory receptor upregulated following T-cell activation by cognate antigen and is specific for checkpoint ligand PD-L1. Captopril treatment increased the proportion of CD8⁺ and DNT subpopulations expressing PD-1 in both liver and tumor (figure 5A,B).

When considering these activated populations as a proportion of total leukocytes (CD45⁺) in these tissues, we find that in the liver and tumor tissues, captopril treatment again increases activated CD8⁺ and DNTs; however, in the tumor, the proportion of total leukocytes that is activated CD4⁺ significantly decreases (figure 5C,D). This may indicate a role for captopril in the activation of CD8⁺ and DNT but also in suppressing the activation of CD4⁺ T cells within the tumor.

We further investigated the proportion of CD8⁺ T cells expressing CD103 to delineate the effect of captopril on the tissue resident CD8⁺ T cell (T_RM) population. The percentage of CD8⁺ T cells expressing CD103⁺ was significantly reduced in both liver and tumor tissues in the treated group (figure 5E). However, within the total leukocyte population, the proportion of CD8⁺ CD103⁺ T cells was significantly increased in tumor and was not affected in the liver (figure 5F).

This suggests a potential mechanism of how captopril may control tumor growth by improving the cytotoxic T cell response.