Between June 2023 and January 2024, 14 female patients who had breast surgery at Clinica Villa Fiorita S.p.A. (Aversa, Italy) were included in the present study. On May 29, 2019, the study was authorized (approval no. 3/19) by the IRCCS Pascale Ethics Committee (Naples, Italy). Based on the Declaration of Helsinki (16), all human samples were taken only after each patient and healthy donor provided written informed consent (Table I). All the procedures listed below were carried out in compliance with the applicable rules and laws.

Ex-vivo 3D PDOs from samples of patients with BC were generated (17), by using a protocol approved by our local Ethics Committee. Moreover, all patients involved in the present study approved and signed a written informed consent form to authorize use of their tumor samples for research purpose. On the day of collection, all new tumor tissue samples need to be treated sterilely and stored on ice. Prior to being used in tissue transplantation, tumors are extracted from patients and physically dissected into tissue fragments. Tumor pieces were minced and digested enzymatically using collagenase and hyaluronidase while being shaken for 12–18 h at 37°C at a low to moderate speed (for example, 20 × g). The amount of tissue that is recovered during surgery and the amount that is set aside as fragments determines how much tissue is needed. Using two disposable scalpels, the tissue is cut into fine pieces in a crisscross pattern. Then, a cell lifter was used to move the tissue to a 50-ml conical container. The digestive process was observed using both visual inspection and light microscopy of the reactions.

If the tissue is resistant to enzymatic digestion, a 100-µm cell strainer was used to strain the digested tissue into a fresh tube, removing all debris and undigested pieces. The organoid fraction should be heavily enriched for refractile, solid clusters of tumor cells called ‘organoids’ and should contain few single cells (18,19). To preserve their three-dimensional structure, cells were implanted in Matrigel to create 3D cultures.

Glycyrrhizin (Glycyrrhizic Acid; cat. no. 167409; Selleck Chemicals) is a direct HMGB1 inhibitor that prevents oxidative stress and inflammatory molecule development that is dependent on HMGB1. Glycyrrhizin, a naturally occurring anti-inflammatory and antiviral triterpene, is employed in clinical settings to block HMGB1′s chemoattractant, mitogenic and intranuclear DNA-binding properties. Studies using NMR and fluorescence have shown that glycyrrhizin binds directly to HMGB1 [K(d) ~150 µM]. The two arms of both HMG boxes form two shallow concave surfaces with which it interacts (20). Glycyrrhizin 50 µM was administered for 72 h while examining samples from 14 PDOs in this pilot investigation.

Total RNA was extracted using TRIzol reagent (Thermo Fisher Scientific, Inc.). Reverse transcription was performed to convert 1 µg of isolated RNA into cDNA using SensiFAST Reverse Transcriptase (Bioline), following the manufacturer's instructions. Expression levels of genes encoding for HMGB-1, Cyclin A, Cyclin B, Cyclin D and Cyclin E were analyzed using RT-qPCR. Applied Biosystems' PRIMER EXPRESS software was used to generate gene-specific primers. The SYBR Green PCR Master Mix (Applied Biosystems) was used for the amplifications. A denaturation step at 95°C for 10 min (stage 2), a starting step at 50°C for 2 min (stage 1), and 40 cycles at 95°C for 15 sec and 60°C for 1 min (stage 3) comprised the thermal cycling conditions. Every sample was performed twice in 25 µl reactions with an Applied Biosystems QuantStudio 7 Flex. The 2^−ΔΔCq method was used to calculate relative expression values after normalizing to Actin, which was employed as an internal control gene (21). Non-specific signals generated by primer dimers were excluded from analysis using non-template controls. RT-PCR primer sequences were reported in Table SI.

To characterize the PDOs, confocal microscopy analysis was conducted. Images were acquired three days post-formation. Cell nuclei were stained with Hoechst 33342 (cat. no. H3570; Thermo Fisher Scientific, Inc.) at a final concentration of 1 µg/ml for 1 h at 37°C. To characterize HMGB-1, β-Catenin and NF-κB expression levels in tumor and non-cancerous ex vivo PDOs, the organoids were incubated with antibodies against HMGB-1 (cat. no 3935S; Cell Signalling Technology, Inc.), β-Catenin (cat. no sc-7963; Santa Cruz Biotechnology, Inc.) and NF-κB (cat. no 3865015; Sony Biotechnology, Inc.) at a 1:100 dilution for 1 h at room temperature (RT). Subsequently, they were incubated with secondary antibodies Alexa Fluor™ 488 Rabbit (cat. no A21206; Invitrogen; Thermo Fisher Scientific, Inc.), Alexa Fluor^® 594 Goat anti-mouse IgG (cat. no 405326; BioLegend, Inc.) and PE (cat. no 3D6C02; BioLegend, Inc.), respectively, diluted 1:200 for 1 h at RT in the dark. Maximal projection images were acquired using confocal microscopy (Mica; Leica Microsystems GmbH) at ×63 magnification and saved in TIFF format.

Annexin V-FITC Kit-AAD Kit (cat. no IM3614; Beckman Coulter, Inc.) was utilized, according to manufacturer's instructions, to analyze the cell cycle of PDOs three days post-formation, with a minimum of 50,000 single-cell events recorded. Briefly, PDOs cells suspension with 1X ice-cold PBS was centrifuged for 5 min at 500 × g at 4°C. The pellet was resuspended in 100 µl 1X Binding Buffer solution, containing 10 µl of Annexin A5-FITC and 20 µl of 7-AAD Viability dye, and incubated for 15 min on ice in dark conditions. Subsequently, 400 µl of 1X Binding Buffer solution were added and the cell preparation was analyzed by flow cytometry using CytoFLEX (Beckman Coulter, Inc.). The percentages of cells in apoptosis were calculated using the Michael Fox algorithm. Later, apoptotic cell analysis was conducted using Kaluza Analysis Software 2.1 (Beckman Coulter, Inc.).

The DNA-Prep Reagents kit (cat. no 6607055; Beckman Coulter, Inc.) was utilized, according to manufacturer's instructions, to analyze the cell cycle of PDOs 3 days post-formation, with a minimum of 10,000 single-cell events recorded. Briefly, PDOs cells suspension was fixed and permeabilized with DNA PREP RPL for 5 min at RT. Then, 1 ml of DNA PREP STAIN was added for 1 h at RT. All the samples were analyzed using CytoFLEX (Beckman Coulter, Inc.). The percentages of cells in the G1, S, and G2/M phases were calculated using the Michael Fox algorithm. Subsequently, cell cycle analysis was conducted using Kaluza Analysis Software 2.1 (Beckman Coulter, Inc.).

The Gene Expression Profiling Interactive Analysis (GEPIA; http://gepia2.cancer-pku.cn/#analysis) web server has been a valuable and highly cited resource for gene expression analysis based on tumor and normal samples from the TCGA and the GTEx databases. GEPIA2 is an updated and enhanced version to provide insights with higher resolution and more functionalities (22). A total of three separate experiments were conducted in triplicate, and the assay results were reported as the mean ± SD. GraphPad Prism software, version 9.0 (GraphPad Software Inc.; Dotmatics), was used to conduct statistical analysis. Tukey's test was used after the ordinary one-way ANOVA statistical test to compare the data. P<0.05 was considered to indicate a statistically significant difference.

In a preliminary study, HMGB1 mRNA expression was analyzed in normal and breast tumor tissues using GEPIA2; although no significant difference was observed, the HMGB1 mRNA expression was higher in breast tumor tissues compared with normal (t=1085; n=291) (Fig. 1). In our laboratories, in parallel ex vivo 3D cultures obtained from BC surgical specimens were generated to study molecular mechanisms involving HMGB1 protein. Briefly, human breast tumor and non-tumor tissues were obtained from a limited number of patients undergoing surgery to obtain PDOs. Despite the difficulties related to the collection of samples from patients and the maintenance of the 3D ex vivo cultures, which represent the main limitations of organoid studies, 14 tissues obtained from human breast biopsy samples were collected to derive 14 ex vivo PDOs primary cell cultures and they were divided into two groups; fibroadenoma (n=7) and BC (n=7). RT-qPCR results revealed higher HMGB1 mRNA levels in tumor PDOs subgroups compared with the non-tumor PDO subgroups in which low levels of HMGB1 mRNA expression were recorded (Fig. 2A). This result was confirmed by the localization of HMGB1 through immunofluorescence labeling of PDOs in which the signal was higher specifically in tumor-PDOs compared with fibroadenoma PDOs (Fig. 2B).

The processes of cell cycle progression and cell proliferation are subject to precise and orchestrated control in normal cells. However, uncontrolled cell proliferation resulting from aberrant cell cycle progression represents a pivotal characteristic of cancer. To understand the regulatory basis of these mechanisms and a possible involvement of HMGB1, treatment was performed with 50 µM of Glycyrrhizin, a licorice root extract known as selective HMGB1 inhibitor on PDOs samples for 72 h. Propidium iodide (PI) is a commonly used fluorescent dye in conjunction with annexin V to determine whether cells are viable, undergoing apoptosis, or necrotic. This is achieved through differences in plasma membrane integrity and permeability. The Annexin V/PI staining method is a common approach for detecting cells undergoing apoptosis.

Following 72 h of Glycyrrhizn treatment, a pro-apoptotic effect induced by HMGB1 inhibition was found. In the treated tumor PDOs, a significant increase in apoptotic cells was observed with a mean of apoptotic rate increased to 30% than in the untreated control (Fig. 3A). The primary characteristics of apoptosis include DNA fragmentation and damage. To ascertain whether Glycyrrhizin-induced apoptotic cells affected cell cycle distribution, a further analysis of the cell cycle was conducted after 72 h of Glycyrrhizin treatment by flow cytometry. The PI fluorescent dye can enter cells with compromised membranes and bind stoichiometrically to DNA, thereby enabling the assessment of the proportion of cells at each stage based on the levels of DNA present.

By comparing the proportions of cells in each phase, the percentage of cells in the G2-M phase gradually increased in the treated PDOs samples while a clear reduction in the G0-G1 phase was observed, suggesting that HMGB1 inhibition may arrest the transition from G2 to M phase in tumor PDOs so that their proliferation capacity is weakened, and the cell has a reduced vitality (Figs. 3B, S1 and S2).

Tumor-associated cell cycle defects are often mediated by alterations in cyclin activity, in these cases inducing uncontrolled proliferation as well as genomic and chromosomal instability. Therefore, selective inhibition of the cyclin pathway may provide therapeutic benefit against some human malignancies. Considering this, we chose to investigate the gene expression trend of cyclins to also support the results obtained by cell cycle analysis. By blocking HMGB1, we observed an opposite trend of cyclins A, B, D and E between tumor and non-cancerous PDOs (Fig. 3C and D). Following glycyrrhizin therapy, our data showed a significant decrease in all cyclins examined in tumor samples; this decrease in expression is not seen in non-tumor samples that continue to be active. Emerging data suggests that the use of an HMGB1 inhibitor could be a therapeutic strategy that selectively targets tumor cells.

It has been reported that interactions between the Wnt/β-Catenin and PI3K/AKT/mTOR pathways are often hyperactivated in numerous solid tumors, such as breast and ovarian cancer and act downstream on NF-κB (23). To study a possible role of HMBG1 in Wnt/β-Catenin modulation on NF-κB also in BC, preliminary experiments were performed observing that inhibiting HMGB1 causes a blockade of both β-Catenin and NF-κB.

The transcription levels were investigated in this initial stage of the study because the number of organoids produced is dependent on the ex vivo proliferation index as well as the size of the starting tissue. Thus, HMBG1 was blocked and its downstream effects were investigated by administering glycyrrhizin for 72 h to BC PDO. Lowering of NF-κB and β-catenin expression levels had the most significant effect (Fig. 4).

Inhibition of HMGB1-induced signalling by switching off HMGB1 causes downregulation of the Wnt/β-Catenin/NF-κB pathway, key regulator of anti-apoptotic genes and consequently the initiation of apoptosis. Glycyrrhizin (50 µM) was administered to organoids in order to examine this aspect, and then immunofluorescence labeling with some targets related to the pathway under investigation was used. The results of the localization analysis demonstrated that the organoids treated with glycyrrhizin did not change morphologically, but rather function. Indeed, following treatment, β-catenin (Fig. 5A) and NF-κB (Fig. 5B) signals were lost; this effect suggests a direct and guided involvement of HMGB1, which is totally inactivated by glycyrrhizin (Fig. 5C).