Fig 2: Schematic outline of the Proximity Ligation Assay (PLA) technical procedure. A) Schematic outline of the Proximity Ligation Assay (PLA) technical procedure. PLA probes directed against MMP-9 and TIMP-1 are incubated with the plasma sample allowing a binding of antibodies to target epitopes. Enzymatic ligations of the two oligonucleotide strands can be carried out only when the two PLA probes are in proximity, due to complex formation between MMP-9 and TIMP-1. Forming a PCR amplicon the antibody to antigen binding is now converted into a DNA strand, which can be amplified and later detected by real-time qPCR. B) Principle and sequential design of the MMP-9:TIMP-1 proximity probes. Each polyclonal antibody (MMP-9 and TIMP-1) has been divided into two pools, with one pool conjugated to the 5’ oligonucleotide (5’end) and the other pool conjugated to the 3’oligonucleotide (3’end). When mixing MMP-9 (3’end) probe with TIMP-1 (5’end) probe and plasma, the two probes will come into proximity, if their target antigens form a complex. The connector oligonucleotide is then used to connect the two oligonucleotide arms. Adding a ligase, the two arms will be ligated together, now forming the template of a PCR amplicon. The flanking 20 base pairs of the conjugated oligonucleotide represent the unique primer specific sequence, while the central part represents a universal sequence matching the connector oligonucleotide. The sequences of the specific primers are illustrated in the lower section of the figure. Rew: reverse primer, Frw: forward primer.

Fig 3: Logistic regression ROC model combining TIMP1, SNAIL1 and BCL11A expression levels for patients’ discrimination. Equation: Ln(odds) = 2.02 + (0.31 × TIMP1expression) + (0.05 × SNAILexpression) + (0.55 × BCL11Aexpression); AUROC, area under the ROC curve; CI, confidence interval. Healthy controls, n = 30; patients, n = 32.

Fig 4: Impact of TIMP1 knock-down in MDA-MB-231 proliferation and dissemination in a murine orthotopic model. (A) CKs (EA1/EA3) expression in the lung tissue of the control group, confirming the presence of lung metastasis. The image was taken at 20× and respective magnification (square) at 40×; (B) Mean tumor volume in PLKO and SH2 (nreplica = 3 mice/condition, ntotal = 9 mice/condition; two-sided Mann–Whitney U-test; p = 0.0053) after sacrifice (upper image). Representative images of tumor volume in both experimental groups (down image); (C) ECAD (CDH1), NCAD (CDH12), CD133 and VIM expression levels in primary tumors generated by PLKO and SH2 cell variants (n = 2); (D) Representative images of hematoxylin-eosin staining of the different mouse tissues after the orthotopic injection of the control cell line (PLKO) and the TIMP1 silenced one (SH2). T: Primary tumor; LN: Axillary lymph node; L: Liver; LU: Lung. 4× images and their respective magnifications (square) at 20×.

Fig 5: Tissue inhibitor of metalloproteinase 1 (TIMP1) protein levels. (A) Box plot of TIMP1 protein levels in controls, patients with monoclonal gammopathy of undetermined significance (MGUS), and patients with multiple myeloma (MM). Pairwise comparisons of TIMP1 protein levels in patients with different disease statuses. (B) MGUS and MM, (C) smouldering myeloma (SMM) and MM, and (D) MM at newly diagnosis (NDMM) and at complete response (CR). (E) TIMP1 protein levels in corresponding patients with different disease statuses; SMM, NDMM, CR, progressive disease (PD). Different colors of lines indicate corresponding patients. (F) Box plot of TIMP1 protein levels by international staging system (ISS) stage. (G) Box plot of TIMP1 protein levels by Revised ISS stage. (H) Correlation between TIMP1 concentrations in bone marrow plasma and plasma cell percentage in bone marrow in patients. Each dot represents an individual patient. The dotted line is the least square line.