Metastasis-suppressor NME1 controls the invasive switch of breast cancer by regulating MT1-MMP surface clearance

Membrane Type 1 Matrix Metalloprotease (MT1-MMP) contributes to the invasive progression of breast cancers by degrading extracellular matrix tissues. Nucleoside diphosphate kinase, NME1/NM23-H1, has been identified as a metastasis suppressor; however, its contribution to local invasion in breast cancer is not known. Here, we report that NME1 is up-regulated in ductal carcinoma in situ (DCIS) as compared to normal breast epithelial tissues. NME1 levels drop in microinvasive and invasive components of breast tumor cells relative to synchronous DCIS foci. We find a strong anti-correlation between NME1 and plasma membrane MT1-MMP levels in the invasive components of breast tumors, particularly in aggressive histological grade III and triple-negative breast cancers. Knockout of NME1 accelerates the invasive transition of breast tumors in the intraductal xenograft model. At the mechanistic level, we find that MT1-MMP, NME1 and dynamin-2, a GTPase known to require GTP production by NME1 for its membrane fission activity in the endocytic pathway, interact in clathrin-coated vesicles at the plasma membrane. Loss of NME1 function increases MT1-MMP surface levels by inhibiting endocytic clearance. As a consequence, the ECM degradation and invasive potentials of breast cancer cells are enhanced. This study identifies the down-modulation of NME1 as a potent driver of the in situ-to invasive transition during breast cancer progression.


CRISPR/Cas9 technology
CRISPR guides. Lentiviral plasmid guides targeting human NME1 and NME2 were generated in pLenti U6gRNA Cas9-GFP-Puro vector and were purchased from Merck-Sigma-Aldrich as well as the non-target guide (pLenti CRISPR-NT CONTROL). Two different guides were designed (https://www.milliporesigmabioinfo.com/bioinfo_tools/)   The clinical and pathological features of patients are summarized in Table S2. The TMA consisted of arrayed 1-mm diameter cores from synchronous in situ and microinvasive/invasive carcinomas and a matched core from adjacent non-tumoral breast tissue constructed as previously described [4].

Immunohistochemical (IHC) staining of breast tumor tissue microarray
Sections (3 µm) from paraffin-embedded tissue microarrays were dewaxed in xylene and rehydrated in a graded alcohol series before heat-induced antigen retrieval (60 minutes in 10 mM sodium citrate buffer, pH 6.0 at 90°C). Then, sections were incubated in 3% hydrogen peroxide solution for 5 min to inhibit endogenous peroxidase activity. The slides were further incubated with blocking serum for 15 min and then with the primary antibodies. Selective NME1 and NME2 pAbs were used: affinity-purified rabbit NME1 pAb [2] was used at a dilution of 1:3000; mouse NME2 mAb purchased from Kamiya Biomedical Company (Seattle, WA) was used at a dilution of 1:1000.
Immunolabeling was performed using the Dako Autostainer Plus and EnVision™ FLEX, Low pH kit with diaminobenzidine as chromogen according to the manufacturer's procedure (Dako, Santa Clara, CA). Slides were counterstained with hematoxylin before mounting. Images were acquired with the Philips Ultra-Fast Scanner. NME1 and NME2 levels in the different tumor biopsies were scored under the supervision of a pathologist in a blinded manner using the H-score method based on semi-quantitative assessment of the intensity of plasma membrane and cytoplasmic staining (graded as: 0, non-staining; 1, weak; 2, median; or 3, strong) and the percentage of positive cells. H-score range was from 0 to 300 for membranous or cytoplasmic staining and 0 to 600 for total staining.

Unsupervised hierarchical clustering
The membranous H-score of MT1-MMP and the total H-score of NME1 from in situ

Intraductal transplantation method
The intraductal xenograft model was carried out as previously described [4,7]. Briefly, 5x10 4 MCF10DCIS.com cells in 2 µl PBS were injected into the primary duct through the nipples of both mammary inguinal glands #4 of 8-10 weeks-old virgin female SCID mice. Mice were sacrificed at 4 and 7 weeks after injection by cervical dislocation.
Immediately after being euthanized, mammary glands were excised and processed for further study (including whole-mount and histological and IHC staining on sections).
The animal facility was granted approval (C-75-12-01) given by the French

Histological and immunofluorescence analysis of mouse tissue sections
Whole-mount carmine and hematoxylin and eosin (H&E) staining of tissue sections were performed as described in [4]. After whole mount staining, image acquisition was performed with an A1R Nikon confocal binocular microscope. Quantification of the tumor area was performed using Image J software. To retrieve antigens on paraffinembedded tissue samples, sections were incubated for 20 min in 10 mM sodium citrate buffer, pH 6.0 at 90°C. Then, after 60 min incubation in 5% fetal calf serum, sections were incubated overnight with diluted primary antibodies, washed and further incubated for 2 hrs at room temperature with appropriate secondary antibodies.

3D collagen I invasion assay
Details of the procedure have been described [8] . Briefly, Petri dishes were filled with 1.35 ml of neutralized native type I collagen and incubated overnight at 37°C to allow gelling. Cells were harvested and isolated using Moscona buffer and trypsin/EDTA, then seeded on the top of the collagen gel at the density of 0.33 x10 6 cells per dish.
Cells were cultured for 24 hrs in the absence or presence of GM6001 (10 µM). Noninvasive round-shape cells that remained at the surface of the gel, and invasive cells that inserted an invasive extension within the collagen gel were scored in twelve fields of 0.157 mm 2 . Invasion index was calculated as the number of cells with invasive extensions to the total cell number multiplied by 100 [8].

Quantification of pericellular collagenolysis
Cells treated with siRNAs against MT1-MMP, NME1, NME2, and non-targeting siRNA for 48 hrs, or stably overexpressing NME1 (or empty vector) were trypsinized, resuspended (2.5 x 10 5 cells/ml) in 0.2 ml of 2.2 mg/ml acidic-extracted type I collagen solution (Corning) with pH buffered to 7.5 and loaded on a glass coverslip. After gelling for 30 min at 37°C, complete medium was added and collagen-embedded cells were incubated for 24 hrs at 37°C in 5% CO2. After fixation in 4% paraformaldehyde in PBS at 37°C for 30 min, samples were incubated with ColI-3/4 C pAb (2.5 µg/ml) for 2 hrs at 4°C, washed extensively with PBS, and counterstained with Cy3-conjugated anti-rabbit IgG antibodies, 4',6-diamidino-2-phenylindole (DAPI), and Alexa Fluor-phalloidin to visualize cell shape. Image acquisition was performed with an A1R confocal microscope (Nikon) with a 40x oil objective. Quantification of the degradation spots was performed with a home-made plugin in ImageJ [9].

Multicellular spheroid outgrowth in 3D Matrigel
For the analysis of invasive outgrowth of MCF10DCIS.om cells, an overlay basement membrane assay was performed. Briefly, 6-well plates were coated with 12 mg/ml native Matrigel and allowed to solidify for 20 min at 37°C. MCF10DCIS.com cells (3 x 10 5 cells) were seeded as single cells onto the solidified basement membrane. After 7 days, cells were imaged in triplicate for development of invasive outgrowths by differential interference contrast (DIC) imaging using a 20x objective. Invasive growths were defined as consisting of two or more cells migrating away from their structure of origin. A minimum of 20 images were analyzed for each condition.

In situ proximity ligation assay (PLA)
To monitor the subcellular localization of protein-protein interactions at single molecule resolution, an in situ proximity ligation assay (PLA) was performed as previously Coverslips were analyzed on an inverted wide-field microscope.
Boxplots were generated with the R ggplot2 package.  threshold total NME1: 300.
Supplemental Figure S9. Validation of NME1 and NME2 knockout in

MCF10DCIS.com clones
Fraction of sequencing reads aligned to the reference amplicon sequence and for which an insertion or deletion was observed. For both NME1 and NME2 genes, two sites (referred to as A and B) were targeted for edition by CRISPR/Cas9 in MCF10DCIS.com cells. Corresponding targeted sites were sequenced after amplification with appropriate primers (see Table S1)