Motion based ex vivo (MOTEX) culture of breast tumor slices sustains microenvironment composition

Abstract

Personalized medicine for breast cancer (BrC) requires predictive biomarkers to select the optimal therapeutic option for each individual patient. Personalization of chemotherapy or immunotherapy responses is particularly challenging, as molecular markers do not appear to be sufficiently predictive for therapy response. Functional assays for therapy selection may be the solution for this dilemma. An interesting approach is ex vivo cultures of precision cut tumor slices, such as the MOtion-based Tissue EX vivo (MOTEX) method that we described previously. This culture method has the advantage that it carries all cell types in the tumor, including various immune cell populations. We here show, that macrophages, B-cells and T-cell populations are maintained in the MOTEX culture for several days without apparent loss of viability. Even treatment with the microtubule poison paclitaxel did not reduce immune cell abundance or viability significantly. Anthracycline-based chemotherapy, however, did affect immune cell composition, as expected based on its cytotoxic properties. Therefore, we conclude that MOTEX culture of BrC tissue slices can be used to investigate effect of treatments that involve the immune system. This opens perspectives to develop predictive assays for immune checkpoint inhibitor treatment and other therapeutic interventions that require immune components in the assay system.

Keywords: Breast cancer; Treatment prediction; Tumor infiltrating lymphocytes (TILs); Tumor microenvironment.

Fig. 1

Maintenance of TME in MOTEX culture for 3 days. (A) Schematic overview of tissue collection and handling. Tumor specimens were embedded in agarose and sliced using a Leica Vibratome. Next, generated slices were cultured in the MOTEX system for three days, with or without treatment. Formalin-fixed paraffin-embedded (FFPE) samples were then sectioned, stained and imaged. (B) Representative images of the H&E and CD45 staining of two primary BrC resection and two biopsy samples. (C) CD45 levels of the resection and (D) biopsy samples at day 0 and day 3 of culture (n=20 and 14, respectively). (E) Representative images of the Panel 1 MIF of resection BC material (sample M395). (F) Quantification results of markers representing the TME of the resection and (G) biopsy material at day 0 and day 3 of culture. The Y-axis indicates the percentage of cells positive for the indicated surface marker in the tissue slice. (H) Quantification results from Fig. F, with differences between all markers plotted separately. Significant differences are indicated in the graph based on the 2-way ANOVA test. (I) Quantification results from Fig. G, with differences between all markers plotted separately. Significant differences are indicated in the graph based on the 2-way ANOVA test.

Fig. 2

Similarities between tissue slices in primary breast cancer specimens. (A) H&E images of two primary tumor resection samples, where slices were assigned randomly (generated slices were mixed during slicing and randomly assigned to each condition; M587) and in a sequential order (the order of slicing was kept when assigning each condition; M591). (B) CD45 levels for these two samples. (C) H&E images of three primary tumor biopsy samples (B5, B9 and B10), showing differences in morphology between two biopsies – one fixed immediately after biopsy collection and the other one cultured for three days. (D) Levels of CD45⁺ cells and (E) Panel 1 MIF of the three selected samples. The Y-axis indicates the percentage of cells positive for the indicated surface marker in the tissue slice.

Fig. 3

Changes in the TME after treatment ex vivo. (A) Representative images of the CD45 DAB staining of M459 (resection) and B5 (biopsy) BC samples before and after treatment. (B) Quantification results of the CD45 staining before and after taxane and FAC treatment. For taxanes, a concentration of 10 nM docetaxel or paclitaxel was used, while for the FAC the average was calculated after treatment with all three different treatment concentrations (FAC 1x, 5x and 10x). Significant differences are indicated in the graph based on the 2-way ANOVA test. (C) Quantified results of the panel 2 MIF staining for the resection material samples without and with paclitaxel (P) or docetaxel (D) treatment. The Y-axis indicates the percentage of cells positive for the indicated surface marker in the tissue slice. In most samples, day 3 was used as a reference point and 10 nM taxane as treated sample. In some samples, due to lack of material, Day 0 or 5 was used as an untreated sample and 25 nM taxane treatment as treated samples (indicated with a “*” for days in culture or “**” for higher treatment concentration). AVG = average of all samples combined. (D) Quantification results from Fig. C, with differences between all markers plotted separately. Significant differences are indicated in the graph based on the 2-way ANOVA test. (E) Quantified results of the panel 2 MIF staining for the biopsy samples without and with FAC (indicated as T for treatment, calculated as average response to all tested FAC concentrations). The Y-axis indicates the percentage of cells positive for the indicated surface marker in the tissue slice. (F) Quantification of results from Fig. E, with differences between all markers plotted separately. Significant differences are indicated in the graph based on the 2-way ANOVA test.

Fig. 4

Assessment of the initial levels and composition of the TME and taxane response ex vivo. (A) Schematic overview representing resection samples with known taxane sensitivity, scored as either sensitive or resistant to the treatment. The sensitivity to taxanes is based on the EdU/pH3 ratio after 10nM of taxane treatment relative to the untreated sample. Graphs B-E show division of the samples on the X-axis, based on their REMIT score. Samples M457 and M520 were assessed within this study, scoring as resistant (67 %) and sensitive (24 %), respectively. (B-H) Scatter plots representing the CD45⁺ (B), total T-cell (C), total macrophages (D) and the M2 macrophage (E) levels in relation to the taxane sensitivity. Results from graphs C-E are based on the panel 2 MIF results represented in Fig. 1, Fig. 2. All graphs include the average levels calculated from day 0 and day 3 of untreated samples; exceptions are samples M412 and M587 were only day 3 was used and M457, M538, M581 and M591 were only day 0 was used (due to insufficient availability of material for the other time points). Panel 2 MIF results from sample M424 were not available, and this sample was only included in Fig. 3B.

Fig. 5

Assessment of the pretreatment TME composition and sensitivity to FAC ex vivo and anthracycline-based chemotherapy in vivo. (A) Representation of the TUNEL levels measured in the BREAST study [11] after treatment with 1xFAC, used to categorize the biopsy samples in a gradient from the least to most sensitive tumors ex vivo. (B) Scatter plot representing the CD45⁺ levels in relation to FAC sensitivity ex vivo. (C) Schematic overview of the ex vivo and in vivo sensitivity to anthracycline-based chemotherapy of biopsy samples, scored as sensitive or intermediate to the treatment. Patients with rCR were marked accordingly. (D, E) Average CD45⁺ levels divided between sensitive and intermediate and rCR and non-rCR patients.