Fig 2: PDX tumor sphere cells can be genetically engineered for functional studies of metastasis in vivo.a Schematic shows workflow for lentiviral engineering of PDX sphere cells in vitro followed by an analysis of their capacity to form primary tumors and spontaneous metastasis following transplantation in vivo. b Representative images show GFP expression in PDX sphere cultures following lentiviral transduction in vitro. HCI010 P0 sphere cells were transduced with GFP lentivirus at MOI = 25 and seeded in MAT-E culture at a density of 2 × 105 cells per well. Images show z-stack micrographs of individual wells containing P1 spheres 1 week later. Scale bar = 800 µm. c Images show GFP expression in representative P1 PDX spheres two weeks post transduction in MAT-E culture. Scale bar= 50 µm. d Representative flow cytometry plots show transduction efficiency of P1 PDX sphere cells two weeks post transduction and growth in MAT-E culture. The bar graph quantifies the percentage of PDX cells positive for GFP in each well (n = 3). Values represent the mean ± s.d. e Images show representative PDX tumors produced 60 days post orthotopic transplantation. 1 × 105 CD298+MHC-I-GFP+ P1 PDX cells were sorted by flow cytometry and transplanted into each mouse (n = 5). Scale bar= 5 mm. f Representative flow cytometry plots show the percent of human CD298+MHC-I- human tumor cells that retain GFP expression following tumor growth in vivo. The bar graph quantifies the percentage of human cells positive for GFP in each tumor (n = 5). Values are represented as the mean ± s.d. g Images show representative GFP+ metastatic lesions in the lungs of transplanted animals from e. A dashed red outline highlights the perimeter of the lesions. Scale bar = 50 µm. h Representative flow cytometry plots show the percent of human CD298+MHC-I- human metastatic cells in the lung that retain GFP expression. The bar graph quantifies the percentage of human cells positive for GFP in each lung (n = 5). Values are represented as the mean ± s.d.

Fig 3: Detection of CD298+ tumor cells in the bone using flow cytometry. (a) Representative flow cytometry plot of CD298 staining in non-tumor-inoculated (naïve) mouse bone marrow used as a negative control for the MCF7 model (representative of n = 3 mice). Dead cells (PI+) have been excluded. (b) Representative flow cytometry plots of bone marrow from MCF7-inoculated -E2 (n = 10 mice) and +E2 (n = 8 mice) mice. Dead cells (PI+) have been excluded. (c,d) Quantitation of total number and percent of CD298+ tumor cells from (b). (e) Representative flow cytometry plot of CD298 staining in non-tumor-inoculated (naïve) mouse bone marrow used as a negative control for the SUM159 model (representative of n = 3 mice). Dead cells (PI+) have been excluded. (f) Representative flow cytometry plot of bone marrow from SUM159-inoculated mice (n = 8 mice). Dead cells (PI+) have been excluded. (g,h) Quantitation of total number and percent of CD298+ tumor cells from (f). (c) Mann-Whitney U-Test, *P < 0.05 and **P < 0.01.

Fig 4: PDX tumor sphere cells maintain tumorigenic potential and form spontaneous metastasis in vivo.a Schematic overview of workflow and timeline for assessing the capacity of cultured PDX cells to generate primary tumors and spontaneous metastasis following orthotopic transplantation. b Serial dilution and transplantation analysis to determine the tumorigenic capacity of cultured cells. HCI010 cells were cultured in MAT-E conditions and injected orthotopically into NOD/SCID mice at increasing dilution (5.0 × 103 - 3.0 × 105). Table (top) shows the fraction of tumors generated at each dilution. Representative images (bottom) show primary tumors generated from orthotopic transplantation of 1 × 105 cells. Scale bar = 1 cm. c Histopathological and molecular marker analysis of tumors generated from uncultured (HCI010) and cultured (HCI010 MAT-E) cells. The top panels show the representative histopathological appearance of tumor sections stained with hematoxylin and eosin (H&E). Scale bar = 50 µm. The middle and bottom panels show representative images of IF staining for basal (KRT5), luminal (KRT18), and myoepithelial (SMA) cell markers. Scale bar = 50 µm. d Quantification of spontaneous metastasis in animals transplanted with 1 × 105 cultured HCI010 cells. Representative plots (top) show quantification of CD298+MHC-I- human metastatic PDX cells in distal tissues by flow cytometry. Bar graph (bottom) shows quantification of the percent of metastatic cells in a cohort of transplanted animals (n = 5–14). Fractions indicate the number of tissues with metastasis, defined by >0.005% CD298+MHC-I- cells. Data are represented as the mean ± s.d. LU lung, LN lymph node, BM bone marrow, BR brain, PB peripheral blood. e Comparison of lung metastatic burden in NOD/SCID and NSG animals transplanted orthotopically with 1 × 105 cultured HCI010 cells. Representative plots show CD298+MHC-I- human metastatic PDX cells in the lungs 8 weeks after transplant by flow cytometry. The bar graph shows the quantification of the percentage of metastatic cells in a cohort of transplanted animals. P value determined by unpaired t test. Data are represented as the mean ± s.d. Fractions indicate the number of tissues with metastasis, defined by >0.005% CD298+MHC-I- cells. f Representative images of metastatic lesions in the lungs of NSG animals transplanted with cultured HCI010 cells identified by H&E staining and histopathological analysis. Scale bar = 50 µm.

Fig 5: NME1 promotes lung metastasis from patient tumor cells.a Schematic of experimental workflow to investigate the effect of NME1 overexpression on spontaneous metastasis from cultured PDX tumor cells. b The box plot shows normalized gene expression of NME1 in metastatic (red, n = 435 cells) and primary tumor cells (blue, n = 684 cells). The lower and upper hinges correspond to the first and third quartiles, and the midline represents the median. The upper and lower whiskers extend from the hinge up to 1.5 * IQR (inter-quartile range). Outlier points are indicated if they extend beyond this range. c Primary tumor weight and growth kinetics. 2 × 104 HCI010 PDX cells transduced with +NME1-GFP or +GFP lentivirus were transplanted orthotopically and monitored for 60 days (n = 6 each condition). The line graph shows tumor volume calculated by caliper measurements (top panel). The bar graph shows tumor weight measured at the endpoint (bottom panel). Values are expressed as mean ± s.d. P values determined by unpaired t test. d Images show GFP signal identifying metastatic nodules in representative lungs from orthotopically transplanted mice from c. Scale bar = 5 mm. e Comparison of lung metastatic burden generated from control versus NME1-overexpressing PDX cells. Flow cytometry plots show the percentage of human metastatic cells in the lungs of three representative mice per group, identified by CD298+MHC-I- (top left panels) and GFP (bottom left panels). The bar graph quantifies the frequencies of CD298+MHC-I- metastatic cells in the lungs of each animal in the cohort. Values are expressed as mean ± s.d. P value determined by unpaired t test.