Intraoperative radiotherapy for breast cancer treatment efficiently targets the tumor bed preventing breast adipose stromal cell outgrowth

A total of 20 breast cancer patients undergoing breast-conserving surgery with (study collective) and 21 without (control collective) IORT were recruited after written informed consent was obtained. All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee (No. 2013-589N-MA, Mannheim Ethics committee II) and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. In women of the study collective, tumor bed biopsies were taken before (pre) and after (post) IORT.

IORT was performed according to the TARGIT‑A protocol [4]: The Intrabeam® system (Carl-Zeiss Meditec AG, Oberkochen, Germany) was used for intraoperative irradiation (50 KeV x‑ray). After excision of the tumor and pathoanatomical confirmation of free margins via frozen section, the spherical Intrabeam applicator was adjusted in the wound cavity. Irradiation was accomplished with a dosage of 20 Gy. In patients undergoing breast-conserving surgery without IORT, biopsies of the tumor bed were taken pre and post sentinel node biopsy (SNB) to ensure a comparable time interval of approximately 30 min between both biopsies. Biopsies were taken using conventional scissors. Of course, patients of the control collective also underwent irradiation but in contrast to the study collective solely 3 to 4 weeks after surgery via conventional whole breast radiotherapy.

Histological subtypes and molecular phenotype of were comparable for IORT and controls (not shown).

Twenty-four hours after biopsy, breast ASC (bASC) were isolated using an established method [7]. Briefly, tissue was weighed and then cut into small pieces. The tissue pieces were digested in 10 ml 0.075% collagenase I (Sigma Aldrich) in prewarmed DMEM (Pan Biotech) for 30–45 min at 37 °C. 20 ml of prewarmed DMEM were added to dilute collagenase and then the cell suspension filtered through a 100 µm cell strainer (Falcon). The cell suspension was then centrifuged (1200 ×g, 10 min, RT) and then optionally treated with erythrocyte lysis buffer (1 ×, 10 min, centrifugation 1200 ×g, 10 min, RT). The pellet was resuspended and counted. Cells were plated in a T25 flask (Nunc easy Flask) in DMEM low glucose, penicillin/streptomycin (Sigma Aldrich), L‑glutamine (Gibco) and 10% human AB-serum (German Red Cross Blood Donor Service). Medium was changed bi-weekly until bASC reached a subconfluent stage (approximately 75%). Cells were then split using trypsin/EDTA. Cells were counted and aliquots cryopreserved in fetal bovine serum (FBS)/10% DMSO (dimethyl sulfoxide, WAK-Chemie). Remaining cells were seeded as passage 1 (p1) cells at 200 cells/cm². ASC growth rate was monitored recording cell number at every passage by calculating cell doublings (CD) and doubling time (DT):where Fcn is final cell number and Icn the initial cell number.

To define multipotent MSCs, the International Society for Cell and Gene Therapy has suggested a characteristic combination of functional properties and the specific surface marker phenotype. The hallmarks include plastic adherence, trilineage differentiation capability into adipocytes, chondrocytes and osteoblasts as well as the characteristic surface molecular marker expression, defined by the absence of hematopoietic markers and the presence of CD73, CD90, and CD105 (14,15). After p1, cells were subjected to differentiation assays. Immunophenotyping was performed after p2. Growth curves were calculated in p1 and p2.

Osteogenic and adipogenic differentiation was performed as described previously [7] using osteogenic and adipogenic induction/maintenance medium (Lonza). Briefly, after 21 days of induction, cells were stained with Oil Red O and von Kossa stain as described.

Multicolor immunophenotyping was also performed as described before [7]. Briefly, cells were trypsinized using trypsin/EDTA (Pan Biotech) and adjusted to 1 × 10⁵ cells/tube. FcR blocking reagent (Miltenyi Biotech) was added and incubated for 10 min before adding the titrated volume of antibody (see table S1). Unstained cells served as negative control. Cells were incubated 20 min, followed by two washes with PBS. Finally, Sytox Blue was added to exclude dead cells and cells were analyzed using a FACS Canto II analyzer running FACS Diva software (Becton Dickinson).

MSCs were isolated using collagenase digestion from breast tissue biopsies pre and post IORT (controls: pre and post sentinel lymph node resection). The sample weight was comparable with 0.45 ± 0.37 for pre non-IORT, 0.37 ± 0.26 post non-IORT and 0.56 ± 0.44 and 0.54 ± 0.29 g for pre and post IORT, respectively. In all, 95% of the pre IORT samples and 57% and 66% of the pre and post non-IORT samples yielded MSC cultures, growing beyond passage 2 (Table 1). However, adherent cells were observed in only one of the 20 samples post IORT. These, however, stopped proliferation already after a few days, reminiscent of a highly senescent phenotype marked by numerous stress fibers (Fig. 1a; [7]). In all other cultures, bASC in early in culture grew out as colonies forming cell monolayers with a typical fibroblastoid phenotype (Fig. 1a). The proliferation potential of non-IORT bASC, both pre and post, was comparable to ASC isolated from subcutaneous fat [7]. bASC from pre IORT samples proliferated significantly slower (Table 2). bASC from two exemplary cultures could be cultured up to 6–7 passages until reaching replicative senescence, without an indication of excessive or prolonged proliferation indicative of malignant transformation or tumorigenic origin (not shown).

Differentiation assays and immunophenotyping were performed to verify the MSC-like nature of bASCs. Nearly 90% of the presamples underwent adipogenic differentiation, whereas from the post-non-IORT samples only 55% showed adipogenic differentiation properties (Table 3 and Fig. 1b). All bASC presamples differentiated into the osteogenic lineage, whereas this proportion again was reduced to 77% in the post non-IORT samples (Table 3 and Fig. 1b).

The immune phenotype of bASC at p2 was highly typical for ASC and corresponded to the ISCT proposed positive and negative markers to distinguish MSCs from other cells [12, 13]. As expected, no difference in the marker expression of bASC was observed comparing pre/post non-IORT and pre IORT samples. Thus, all data were merged. Furthermore, cells expressed/non-expressed the typical MSC markers (Fig. 1c, d) at comparable intensities in all but one sample. Negativity of HLA class II indicates a nonactivated state of isolated bASC. Markers to control for contamination (such as lineage for hematopoietic cells) were not detected in the majority (Fig. 1e, f). Surprisingly, few samples contained a high proportion of CD31-postive cells, indicating an endothelial trait despite their fibroblastoid morphology. Other endothelial markers such as CD144 (VE-Cadherin) and CD106 (VCAM), however, were not expressed. In p2, CD34 was not expressed, although ASC may express CD34 early after isolation [13]. The CD146+ subset has been suggested to represent a population transitional between adipose stromal cells and pericytes. In fact, percentage of expressing cells varied markedly. Mean fluorescence intensity on positive cells, however, was highly comparable.