Interleukin-8 Dedifferentiates Primary Human Luminal Cells to Multipotent Stem Cells

Abstract

During aging, cellular plasticity and senescence play important roles in tissue regeneration and the pathogenesis of different diseases, including cancer. We have recently shown that senescent breast luminal cells can activate their adjacent stromal fibroblasts. In the present report, we present clear evidence that these senescence-related active fibroblasts can dedifferentiate proliferating primary human luminal cells to multipotent stem cells in an interleukin-8 (IL-8)-dependent manner. This was confirmed using recombinant IL-8, while the truncated protein was not active. This IL-8-related dedifferentiation of luminal cells was mediated through the STAT3-dependent downregulation of p16^INK4A and the microRNA miR-141. Importantly, these in vitro-generated mammary stem cells exhibited high molecular and cellular similarities to human mammary stem cells. They have also shown a long-term mammary gland-reconstituting ability and the capacity to produce milk postdelivery. Thereby, these IL-8-generated mammary stem cells could be of great value for autologous cell therapy procedures and also for biomedical research as well as drug development.

Keywords: IL-8; STAT3; cell differentiation; stem cells.

FIG 1

Active stromal fibroblasts induce EMT and stemness properties in normal breast luminal cells. Exponentially growing NBL-10 cells were exposed to SFM or SFCM from the indicated fibroblasts for 24 h. (A) Cells were collected and utilized to assess the proliferation rate and migration/invasion capacities using the RTCA-DP-xCELLigence system. Data are representative of results from different experiments performed in triplicate. (B) Whole-cell lysates were prepared from NBL-10 cells treated as indicated and then used for immunoblot analysis. (C) IL-8-activated fibroblasts induce EMT and stemness properties in normal breast luminal cells. Exponentially growing NBL-10 cells were exposed to SFM or SFCM from the indicated fibroblasts for 24 h. Whole-cell lysates were prepared and used to assess the level of the indicated proteins by immunoblotting. The numbers below the bands indicate the corresponding expression levels after loading correction against GAPDH.

FIG 2

Active breast stromal fibroblasts dedifferentiate luminal cells to multipotent stem cells. (A) Attached NBL-10 cells were treated with SLAF-SFCM for 24 h, subcultured in growing medium for a week, and then transferred to ultra-low-attachment plates. Images show hollow mammospheres formed 2 days after in-suspension culture. Bar, 50 μm. (B) SLAF- and ILAF-generated mammospheres were digested, 1,000 single cells were cultured in ultra-low-attachment 96-well plates in triplicate, and the newly formed mammospheres were counted and plotted as means ± SD. This was repeated 5 times. (C) SLAF-generated mammospheres were cytospun, fixed, and utilized for immunofluorescence analysis using the indicated antibodies. Representative images are shown. Bars, 50 μm. (D) Mammospheres were generated as indicated above, digested, and analyzed by flow cytometry. Representative dot plots show the phenotypes of the mammospheres obtained after treatment with the indicated media. (E) Mammospheres were digested to single cells, and 1,000 cells were seeded onto 96-well plates for 1 week and then counted. This process was repeated several times. Error bars represent means ± SD. (F) Cell extracts from the indicated cells and mammospheres were either sham treated or heated and amplified by PCR using the TeloTAGGG telomerase PCR ELISA (Roche). The PCR products were hybridized to antidigoxigenin-peroxidase, ELISA reactions were performed, and the OD was then measured at 450 nm using a standard ELISA plate reader (x-Mark; Bio-Rad). These experiments were performed in triplicates.

FIG 3

SLAF- and ILAF-generated mammospheres can be differentiated into both luminal and myoepithelial cells. (A) SLAF-generated mammospheres were digested into single cells, seeded for a week onto 0.1% gelatin-coated plates to allow differentiation, and photographed using a Floid cell imaging station. Red arrows point to luminal cells, while green arrows point to myoepithelial cells. Bar, 40 μm. (B) Immunofluorescence assay for differentiated SLAF-generated mammospheres using the indicated antibodies. Images were taken using the Floid cell imaging station. Bars, 40 μm. (C) The mammospheres generated by SLAF- and ILAF-SFCM were digested into single cells, seeded for a week onto 0.1% gelatin-coated plates to allow differentiation, and used to assess the percentages of luminal and myoepithelial cells by flow cytometry. Representative flow cytometry dot plots are shown, where blue dots represent cells with luminal cell features, red dots represent cells with myoepithelial cell features, and the numbers represent the proportions of each cell type. (D) A layer of prolactin-containing Matrigel was added on top of 5-day-differentiated SLAF- or ILAF-generated mammospheres. Complex acinar and branched acinar structures were formed after 3 days and imaged using the Floid cell imaging station. Bars, 100 μm. (E and F) SLAF- and ILAF-generated mammospheres were digested into single cells, and 0.5 × 10⁶ cells were injected into cleared mammary fat pads of 3-week-old female NOD/SCID mice, using SFM-treated cells as a control. At 8 weeks postinjection, the reconstituted fat pads were excised and subjected to carmine whole-mount staining (WMS). Hematoxylin and eosin (H&E) staining and immunohistochemistry (E) as well as immunofluorescence (F) assays were performed on FFPE tissues from reconstituted fat pads using human-specific antibodies against the indicated proteins. A normal mouse fat pad was used as a control. Bars, 25 μm.

FIG 4

IL-8 dedifferentiates luminal cells to multipotent stem cells. (A) Whole-cell lysates were prepared from NBL-10 cells treated as indicated and were utilized to assess the levels of the indicated proteins by immunoblotting using specific antibodies. The numbers below the bands indicate the corresponding expression levels after loading correction against GAPDH. (B) Luminal cells (NBL-10) were treated as indicated, and the proportions of cells with the stem cell-like phenotype (CD44^high/CD24^low) were assessed by flow cytometry. Representative flow cytometry dot plots showing the proportions of gated cells are presented. (C) IL-8-generated mammospheres (IL-8-MaSC) were digested into single cells and seeded for a week onto 0.1% gelatin-coated plates to allow for differentiation. Representative images are shown. Bars, 50 μm. NBL-10 cells were treated with IL-8 (2.5 ng/ml) for 24 h and then subcultured in ultra-low-attachment plates for 2 days. (D) IL-8-generated mammospheres were digested, 1,000 single cells were cultured in ultra-low-attachment 96-well plates in triplicate, and the newly formed mammospheres were counted (2nd generation) and plotted as means ± SD. This was repeated 5 times.

FIG 5

IL-8 dedifferentiates luminal cells to multipotent stem cells. (A) Immunofluorescence assay for cytospun IL-8-MaSC mammospheres using antibodies against the indicated proteins. Bars, 25 μm. (B) Immunofluorescence assay for differentiated IL-8-MaSC mammospheres using antibodies against the indicated proteins. Bars, 25 μm. (C) A layer of prolactin-containing Matrigel was added on top of 5-day-differentiated IL-8-MaSC. Complex acinar and branched acinar structures were formed after 3 days and imaged using the Floid cell imaging station. Bar, 100 μm. (D) IL-8-MaSC were digested into single cells, and 0.5 × 10⁶ cells were injected into cleared mammary fat pads of 3-week-old female NOD/SCID mice, using SFM-treated cells as controls. At 8 weeks postinjection, the reconstituted fat pads were excised and subjected to carmine whole-mount staining. (E) IL-8-MaSC-related reconstituted fat pads were digested into single cells, and 0.5 × 10⁶ cells were retransplanted into cleared mammary fat pads of 3-week-old female NOD/SCID mice, using SFM-treated cells as controls. Outgrowths were excised at 8 weeks posttransplantation and subjected to carmine whole-mount staining.

FIG 6

IL-8 dedifferentiates luminal cells to multipotent stem cells. (A) NBL-34 cells were cultured in SFM containing IL-8 (2.5 ng/ml) for 24 h and then subcultured in ultra-low-attachment plates for 2 days. A phase-contrast image of hollow mammospheres in suspension culture is shown. (B) IL-8-generated mammospheres were cytospun, fixed, and utilized for immunofluorescence analysis using the indicated antibodies. Representative images are shown. TRITC, tetramethyl rhodamine isothiocyanate. Bars, 40 μm. (C) IL-8-generated mammospheres were digested into single cells, seeded for a week onto 0.1% gelatin-coated plates to allow differentiation, and photographed using the Floid cell imaging station. (D) Immunofluorescence assay for differentiated IL-8-generated mammospheres using the indicated antibodies. Images were taken using the Floid cell imaging station. Bars, 40 μm. (E) IL-8 induces MaSC in 3D culture. NBL-10 cells were seeded at 1 × 10⁶ cells per well prior to exposure to either SFM, SFM containing IL-8 (25 ng/ml), or SFCM from SLAF or ILAF for 24 h. Cultures were maintained in growing medium for 2 weeks. The discs were then fixed in 10% formalin, paraffin embedded, sectioned, and utilized for either hematoxylin and eosin staining or immunofluorescence analysis utilizing the indicated antibodies. Bars, 200 μm. (F) Truncated IL-8 recombinant protein fails to generate MaSC. NBL-10 cells were exposed to SFM or SFM containing either full-length IL-8 or truncated IL-8 (IL-8t) (2.5 ng/ml) for 24 h. Cells were then cultured in MaSC medium for 3 days, and whole-cell lysates were then prepared and utilized to assess the level of the indicated proteins using specific antibodies. (G) Representative phase-contrast images of the indicated cells and mammospheres. Bars, 50 μm. (H) NBL-10 cells were exposed to SFM or SFM containing either full-length IL-8 (2.5 ng/ml) or 2 different concentrations of full-length IL-6 (1 ng/ml [1] and 3.5 ng/ml [3.5]) for 24 h. (Left) Cells were then cultured in MaSC medium for 3 days, and whole-cell lysates were prepared and utilized to assess the levels of the indicated proteins using specific antibodies. (Right) Representative phase-contrast images of the indicated cells. Bars, 50 μm. The numbers below the bands indicate the corresponding expression levels after loading correction against GAPDH.

FIG 7

IL-8-related transformation of luminal cells to MaSC is STAT3 dependent. NBL-10 cells were transfected with STAT3-specific siRNAs (STAT3si-1 and STAT3si-2) or a scrambled sequence used as a control, and cells were then either sham treated or challenged with IL-8 (2.5 ng/ml) for 24 h. (A and B) Whole-cell lysates were prepared from the indicated cells and utilized to assess the levels of the indicated proteins by immunoblotting. The numbers below the bands indicate the corresponding expression levels after loading correction against GAPDH. (C) The migration/invasion as well as proliferation capacities of the indicated cells were assessed using the RTCA-DP-xCELLigence system. (D) The indicated cells were plated in ultra-low-attachment 6-well plates in stem cell medium and monitored daily for mammosphere formation. Representative images are shown. Bars, 50 μm.

FIG 8

IL-8-dependent induction of MaSC is mediated through p16/miR-141 downregulation. (A) Total RNA was prepared from the indicated cells and utilized to assess the level of mature miR-141 by qRT-PCR. Error bars represent means ± SD, with values from 3 independent experiments. *, P = 0.000114. (B and C) NBL-10 cells transfected with STAT3 siRNA or control siRNA were exposed to SFM or SFM containing IL-8 (2.5 ng/ml) for 24 h. Cells were then collected and utilized to prepare total RNA, and the level of mature miR-141 was assessed by qRT-PCR (B), or cells were utilized to prepare whole-cell lysates, and the levels of the indicated proteins were assessed by immunoblotting (C). (D) NBL-10 cells were transduced with the lentiviral supernatant containing either a control plasmid or a plasmid bearing the CDKN2A open reading frame (ORF). Three days later, cells were exposed to either SFM or SFM containing IL-8 (2.5 ng/ml) for 24 h. After 3 days, cells were collected and utilized to prepare total RNA, and the levels of the indicated transcripts were assessed by qRT-PCR. Error bars represent means ± SD, with values from 3 different experiments. (E) NBL-10 cells were transduced with the lentiviral supernatant containing either a control plasmid or a plasmid bearing pre-miR-141. Three days later, cells were exposed to either SFM or SFM containing IL-8 (2.5 ng/ml) for 24 h. After 3 days, cells were then cytospun and used for immunofluorescence analysis utilizing the indicated antibodies. Bars, 50 μm.

FIG 9

IL-8-related generation of MaSC is mediated through STAT3 activation and miR-141 downregulation. (A) Total RNA was prepared from the indicated cells and used to assess the levels of the indicated transcripts by qRT-PCR. Error bars represent means ± SD, with values from 3 different experiments. (B and C) NBL-10 cells transfected with STAT3 siRNA or control siRNA were transduced with the lentiviral supernatant containing a control plasmid or a plasmid bearing miRZip-141. Cells were exposed to SFM or SFM containing IL-8 (2.5 ng/ml) for 24 h and were then collected to prepare whole-cell lysates and utilized for immunoblot analysis (B) (the numbers below the bands indicate the corresponding expression levels after loading correction against GAPDH) or cytospun and utilized for immunofluorescence analysis using the indicated antibodies (C). Bars, 50 μm.

FIG 10

IL-8-generated MaSC and breast natural MaSC exhibit similar gene expression patterns. (A) Total RNA was prepared from the indicated samples and utilized for microarray global gene expression analysis. Representative scatterplots show the log intensity values. (B) Total RNA was prepared from the indicated cells and utilized to assess the levels of the indicated transcripts by qRT-PCR. Error bars represent means ± SD, with values from 3 independent experiments. (C and E) Venn diagrams showing the percentages of differentially expressed genes between the indicated cells. (D and F) Hierarchical cluster analysis of the indicated samples using the expression data set.

FIG 11

Reconstituted humanized mammary fat pads produce milk after pregnancy and delivery. IL-8-generated mammospheres were injected into cleared mouse fat pads (nipples 4 and 5); at 6 weeks posttransplantation, these females were mated; and when the females were pregnant, all nipples but 4 and 5 were cauterized and removed. Delivery occurred after 3 weeks of pregnancy. (A) Image showing the presence of milk in a reconstituted fat pad. (B) Reconstituted fat pads used for lactation were also subjected to carmine whole-mount staining. (C) Pups with milk spots in their stomachs. (D) Mammary glands were collected postdelivery and also from virgin mouse and subjected to immunohistochemistry on FFPE tissues using human-specific anticasein antibody.