CD44 and CD24 coordinate the reprogramming of nasopharyngeal carcinoma cells towards a cancer stem cell phenotype through STAT3 activation

Cell surface proteins such as CD44 and CD24 are used to distinguish cancer stem cells (CSCs) from the bulk-tumor population. However, the molecular functionalities of CD24 and CD44, and how these two molecules coordinate in CSCs remain poorly understood. We found that nasopharyngeal carcinoma (NPC) cells with high expression of CD44 and CD24 proteins presented with pronounced CSC properties. Accordingly, a subpopulation of NPC cells with co-expression of CD44 and CD24 were specially enriched in high-stage clinical samples. Furthermore, ectopically expressing the epithelial-mesenchymal transition (EMT) regulator Twist was able to upregulate the stemness factors, and vice versa. This indicates a reciprocal regulation of stemness and EMT. Intriguingly, we found that this reciprocal regulation was differentially orchestrated by CD44 and CD24, and only simultaneous silencing the expression of CD44 and CD24 led to a broad-spectrum suppression of CSC properties. Oppositely, overexpression of CD44 and CD24 induced the reprogramming of parental NPC cells into CSCs through STAT3 activation, which could be blunted by STAT3 inhibition, indicating that CD44 and CD24 collaboratively drive the reprogramming of NPC cells through STAT3-mediated stemness and EMT activation. Consequently, targeting of the CD44/CD24/STAT3 axis may provide a potential therapeutic paradigm for the treatment of NPC through repressing CSC activities.


Irradiation selection
A total of 5 × 10 3 cells were seeded into 6-well cell culture plates in complete DMEM. After two hours of cell seeding, various doses from 0 to 12 Gy were given to determine the sublethal dose for each cell line. Irradiation was delivered at room temperature at 37.9 mGy/s using Rad Source RS 2000 X-ray biological irradiator (Rad Source Technologies, Inc., Suwanee, GA). The range of sublethal dose was 8-11 Gy in NPC cell lines. To establish radioresistant clones, three different irradiation doses were used to execute irradiation selection, including four rounds of 8 Gy, four rounds of 11Gy, and 8 Gy, 9 Gy, 10 Gy, 11 Gy, in each round of irradiation. After four rounds of irradiation selection, a single colony selection was performed to obtain homogeneous radioresistant clones. Combining the four rounds of irradiation and single colony selection, NPC radioresistant clones were obtained that had a radioresistant phenotype verified by survival fraction assay.
Tumor sphere formation assay 1×10 4 cells were trypsinized into individual cells and seeded into 1.2% soft agar coated petri-dishes with serumfree DMEM medium supplemented with 1% sodium pyruvate, 1% NEAA and 1% antibiotics-antimycotics solution. The soft surface rendered the cells unable to attach and they subsequently formed tumor spheres after a few days in suspension. The number of tumor spheres was counted after 10 days.

Side population selection
The cell concentration was adjusted to 10 6 cells/ ml and incubated with 1 μl verapamil (50 μM, Sigma-Aldrich, St. Louis, MO) or fumitremorgin C (FTC, 10 μM, Sigma-Aldrich) to block the ABCG2 transporter. Hoechst 33342 (5 μM, Sigma-Aldrich), the DNA binding dye, was then added at a final concentration of 5 μg/mL and the cells were incubated in a 37°C incubator for 90 min. The tube was shaken every 15 min to avoid cell precipitation. After 90 min, the cells were washed twice with PBS, 2 μg/ml propidium iodide (PI) was then added, and the cells were kept at 4°C in the dark before sorting using a BD FACSAria Flow Cytometer (BD Biosciences). After UV488 laser excitation, major cell populations containing Hoechst 33342 showed blue and red fluorescence light. A subset of side population cells pumped out the Hoechst dye via ABC-transporters and demonstrated low fluorescence expression. These side population cells exhibited high fluorescence when treated with verapamil or FTC to block the efflux of fluorescence. The patterns were compared with or without treatment with verapamil or FTC to isolate side population cells.
A total of 10 6 293T cells were seeded on 100 mm dishes and cultured overnight. Utilizing PolyJet™ DNA In Vitro Transfection Reagent (SignaGen Laboratories, Rockville, MD), cells were transfected with the viral vectors. The medium was replaced with complete DMEM 24 hours post-transfection. The virus-containing supernatants of the transfectants were collected and filtered through a 0.45 μm pore size filter 48 hours posttransfection. Viral infection was achieved by adding virus-containing supernatant supplemented by 8 ng/ml polybrene to target cells. Antibiotic selection 24 hours post infection was subsequently employed.

Real-time RT-PCR analysis
Total RNA was extracted with TRIsure™ reagent (Bioline Reagents Ltd, London, UK). The concentration and purity of total RNA was determined using a NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies, Inc., Wilmington, DE). Real-time PCR was performed using the SensiFAST™ SYBR Hi-ROX Kit (Bioline) on the ABI StepOnePlus™ Real-Time PCR machine (Applied Biosystems, Foster City, CA). The PCR primer sequences are listed in Supplementary Table S1.

Western blot
Total cellular protein was extracted in complete RIPA complete RIPA containing protease inhibitor, and phenylmethylsulfonyl fluoride (PMSF) (all from Sigma-Aldrich). The protein assay kit (Bio-Rad, Hercules, CA) was used to determine protein concentrations using a UVvisible spectrophotometer U-3300 (Hitachi, Tokyo, Japan). Aliquots containing 50 μg of total protein from each sample were separated by electrophoresis on 8%, 10%, or 12% SDS-polyacrylamide gels and electro-blotted to BioTrace , or anti-Ac-STAT3 (K685) (Cell Signaling), were used for hybridization in 4% non-fat dried milk; this was followed by incubation with appropriate secondary antibodies (Sigma-Aldrich). The protein bands were captured using a Luminescence Imaging System (Fujifilm, Tokyo, Japan).

Soft agar assay
The plates were first de-coated with 1.2% soft agar as a base. After the agar solidified, 1×10 4 cells were mixed with 0.4% soft agar and seeded on the base. The colonies grown in soft agar were counted after two weeks.

Wound healing migration assay
We utilized ibidi culture inserts (ibidi GmbH, Munich, Germany) for wound healing migration assay and performed according to the manufacture's protocol. The insert contains two reservoirs separated by a 500 μm thick wall. We placed the culture insert into a 24-well culture plate and added 1×10 4 cells into each reservoir for overnight culture. A 500 μm gap was created after removing the culture insert. The images of cell migration were captured by an EVOS FL Cell Imaging System (Thermo Fisher Scientific Inc.).

Transwell invasion assay
For invasion assay, we placed Millicell invasion chamber (8 μM pore size, Millipore, Darmstadt, German) with Matrigel (BD Biosciences) into a 24-well plate. In the upper compartment of the invasion chamber, 1×10 4 cells were seeded and filled with 200 μL serum-free DMEM. In the lower compartment of the invasion chamber, 600 μL complete DMEM with 10% FBS. After 24 hours incubation, the invasive cells located on the underside of the filter were fixed with 3:1 ratio of methanol and glacial acetic acid, stained with 2% crystal violet, and counted under a phase-contrast microscope.

Immunofluorescence staining
Cells were seeded on the cover slides for 24 hours and then fixed in 4% paraformaldehyde and permeabilized in 0.3% Triton X-100 in PBS. The fixed cells were then washed three times with PBS and blocked in FBS. The cells were incubated with primary antibodies such as anti-Oct3/4 (Cell Signalling), anti-Nanog (

Immunoreactivity evaluation for immunohistochemistry staining
Immunoreactivity was semi-quantified by a method that determines the score based on combined intensity and the percentage of cells with positive stain according to a previous report [4]. Briefly, compared with the positive control, the same degree of staining intensity was graded as 2. When the intensity was weaker than the control, the cells were graded as 1, and when stronger than the control, they were graded as 3. Negative staining was 0. We then counted about 1000 NPC cells and the percentage of cells with each intensity grade was estimated. The staining score for CD44 and CD24 antibodies (Abcam) was defined as the sum of the percentage of positive cells with each intensity level multiplied by the intensity grade (e.g., a case with 40% grade 3, 40% scored 2, and 20% scored 1 staining would be scored as 40×3+40×2+20×1=220).