The present study aimed to investigate PA-SG therapy on metastatic HNSCC cell lines. Information on seven HNSCC cell lines was searched in the Cellosaurus (20) for a suitable metastatic HNSCC, such as HSC-1, HSC-2, HSC-3, HSC-4, Ca9-22, KB (15,21-23) and SAS (24) HSC-3 is the most suitable cell line for the present study being one of the first metastatic HNSCC cells established (Table I) (15). Moreover, the metastasis phenotype of HSC-3 in nude mice has been supported by an independent study (17). The other cells were not suitable for our current study, because they were not HNSCC (HSC-1), lacked the metastatic mice phenotype (HSC-2, HSC-4 and SAS) or were contaminated by another cell type (Ca9-22 and KB).

The HSC-3 cell line was purchased from Merck KGaA (cat. no. SCC193). Cells were cultured in DMEM (cat. no. 10569010; Thermo Fisher Scientific, Inc.) supplemented with 10% FBS (cat. no. 26140079; Thermo Fisher Scientific, Inc.) and 1% penicillin-streptomycin antibiotics (cat. no. 15140122l; Thermo Fisher Scientific, Inc.) at 37˚C with 5% CO₂. Cells were subcultured every 2-3 days by digestion with TrypLE™ (Thermo Fisher Scientific, Inc.). The subculture ratio is 1:6.

Stable cell lines were generated as previously described (14). Briefly, yeast CD (pCMVtight-UPRT-T2A-RFP-IRES-CD; cat. no. 126677; Addgene, Inc.) and HSV type 1 TK (pAL119-TK; cat. no. 21911; Addgene, Inc.) suicide genes were amplified by PCR and ligated to mVenus in the pROSA26 vector with human ROSA26 1 kb sequence (AC018506.5; bp 114,245 to 115,244) and promoter and polyA signal for mVenus. The resulting plasmids are mVenus-CD and mVenus-TK, which serve as templates. The single guide (sg)RNA targeting sequences were sgRNA1: 5'-TGTCGAGGTTATTGTAATAA-3' and sgRNA4: 5'-CCGTGGGAAGATAAACTAAT-3'. The sgRNA sequences were synthesized by Integrated DNA Technologies (Integrated DNA Technologies Pte. Ltd.) and cloned into pX330 vector according to instruction in the reference (25). The templates, sgRNA1, and sgRNA4 were co-transfected into HSC-3 cells to integrate the suicide genes into the ROSA26 locus. One day before transfection, 1x10⁵ HSC-3 cells were seeded into one well of a 12-well plate with 1 ml culture medium and cultured at 37˚C with 5% CO₂. On the day of transfection, 1 µg template encoding mVenus-TK, mVenus-CD or mVenus only with 1 µg sgRNA plasmid targeting the ROSA26 locus were transfected at 37˚C overnight into HSC-3 cells using FuGENE HD Transfection Reagent kit (cat. no. E2311; Promega Corporation). One day after transfection, the transfected HSC-3 cells were trypsinized and subcultured into a 96-well plate at a density of 1 cell/well. After 8 days of culture at 37˚C fluorescence-positive cell colonies were collected. The fluorescent images of HSC-3 mVenus-CD or mVenus-TK cell colonies were taken under Nikon inverted microscope (Eclipse Ti2-U, Nikon Instruments Inc.) with longpass GFP filter cube (excitation filter 480/30 nm, dichroic mirror 505 nm, barrier filter 515 nm) with the same exposure time and objective. The fluorescent intensity of ~20-40 individual cells was measure by ImageJ Software (Version 1.53a, National Institutes of Health). The cell colony with the highest fluorescent intensity was chosen for the experiments.

To extract genomic DNA, 1x10⁶ HSC-3 cells were incubated in 300 µl 0.5 M NaOH for 30 min at 37˚C. After centrifuging at 10,000 x g at room temperature for 10 min, HSC-3 supernatant was diluted to 1:1,000 in 0.1 M Tris-HCl (pH 8.0) as a template. The suicide gene inserts at the ROSA36 locus were amplified by Taq DNA Polymerase (cat. No. 11304011; Thermo Fisher Scientific, Inc.) using the following primer pair: Forward, 5'-CGGCCGAGACTTCTGGATGG-3' and reverse, 5'-CCCAGCTAAGGAAAAAGGATAAAATGAAAATCAAG-3', which target to ROSA26 locus, The thermocycling conditions used for PCR were as follows: Initial denaturation at 95˚C for 30 sec followed by 30 cycles at 95˚C for 30 sec, annealing at 60˚C for 15 sec and elongation at 72˚C for 90 sec. The PCR products were resolved on 1% agarose gel with ethidium bromide.

GCV and 5-FC were purchased from Sigma-Aldrich (cat. nos. PHR1593 and F7129, respectively). GCV was dissolved in DMSO at a concentration of 30 mM as a stock solution. 5-FC was dissolved in water at a concentration of 100 mM. HSC-3 mVenus-TK cells were treated with GCV and HSC-3 mVenus-CD cells were treated with 5-FC at 37˚C at indicated concentration for indicated duration. For the dose-response experiment, HSC-3 cells were treated with two-fold dilution of GCV or 5-FC ranging from ~0.1 µM to 100 µM for 3 days at 37˚C. For the time-course response experiment, HSC-3 mVenus-TK cells were treated with 25 µM GCV and HSC-3 mVenus-CD cells were treated with 100 µM 5-FC for 0, 24, 48, 72, and 96 h respectively at 37˚C. For the bystander effect experiment, the cells were treated with 25 µM GCV or 100 µM 5-FC for 72 h at 37˚C. Before treating cells, drugs were diluted to the desired concentration using the culture medium. For the prodrug treatments, the old culture medium was replaced with medium containing the prodrug.

MTT was dissolved in PBS at a concentration of 5 mg/ml (cat. no. M2003-1G; Signa-Aldrich; Merck KGaA). A total of 1x10⁴ cells was seeded into each well of a 96-well plate. Following drug treatment, 10 µl MTT reagent was added to each well. Following 1 h incubation at 37˚C, 100 µl solubilization solution (10% SDS in 0.01 M HCl) was added to each well. The plate was kept at 37˚C overnight. The soluble formazan was measured by using FlexStation 3 Multi-Mode Microplate Reader (Molecular Devices, LLC) at 570 nm wavelength. For the dose-response experiment, MTT assay was performed 3 days after drug treatment. For the time-course response experiment, MTT assays were performed every 24 h from 0 to 96 h. For the bystander effect experiment, MTT assay was carried out after prodrug treatment for 72 h.

Bystander effect is a phenomenon of suicide gene therapy, in which adjacent cancer cells that do not express suicide gene are killed by prodrug treatment. Bystander effects were quantitively measured using a method described previously (14). HSC-3 mVenus-TK or mVenus-CD cells were mixed with HSC-3 mVenus cells at the ratios of 100, 75, 50, 25 and 0% of suicide gene-positive cells. Then, the cell mixtures were treated with GCV at 25 µM or with 5-FC at 100 µM for three days. Next, cell viabilities were measured by MTT assay. The cell viability was plotted against the percentage of suicide gene positive cells ratio. The data were fitted to exponential equation: y=a^*e^-bx + c, where y is cell viability and x is percentage of suicide gene positive cells. In this exponential equation, b is the decay constant, which represents how fast the cell viability decrease with increase of percentage of suicide gene positive cell. The bigger b value means prodrug will kill more cells with the same ratio of suicide gene positive cells. Thus, a bigger b value represents a better bystander effect.

A total of 6x10⁵ cells was seeded into a 6-well plate with 2 ml culture medium with 10% FBS. Serum starvation was not performed during wound healing assay because serum starvation may complex HSC-3 cell migration (26) and would healing assay protocol without serum starvation is feasible (27). The wound was generated by using 1 ml pipette tip. Detached cells and cell debris were removed by washing with medium. Wound images were captured immediately with a Nikon inverted light microscope equipped with a 10x objective (Nikon Corporation). The healing images were captured after 24 h culture at 37˚C. Inhibition of cell migration was calculated as the ratio of the wound area of the healing image and the mean area of the wound image at three time intervals: 0-24, 24-48, and 48-72 h. Wound areas were measured using ImageJ Software (Version 1.53a, National Institutes of Health) with the Wound Healing Tool (28).

Wound healing was measured at three time periods: 0-24, 24-48 and 48-72 h. For the 0-24 h period experiment, a wound was generated immediately before prodrug treatment and the healing images were captured after 24 h prodrug treatment. For the 24-48 h period experiment, cells were treated with prodrug for 24 h before wound generation while the healing image was captured after a further 24 h for a total of 48 h prodrug treatment. For the 48-72 h experiment, cells were treated with prodrug for 48 h before the wound strip generation while the healing images were captured after a further 24 h for a total of 72 h prodrug treatment.

GraphPad Prism 9 software (GraphPad Software, Inc.) was used to perform statistical analysis. One-way analysis of variance with Tukey's post hoc test was performed for multiple comparisons. Data are expressed as the mean ± standard deviation or standard error of the mean from three independent experiments. P<0.05 was considered to indicate a statistically significant difference.

HSC-3 cell lines stably expressing mVenus tagged TK (mVenus-TK) or CD (mVenus-CD) were generated. The mVenus tag serves as a fluorescent marker to facilitate the screening of stable cell lines. A control cell line, which expresses the mVenus tag only, was generated as a negative control. Most HSC-3 cells were fluorescence-positive, indicating the three stable cell lines were homogenous (Fig. 1). To demonstrate the incorporation of genes into the ROSA26 locus, PCR was performed to amplify the DNA sequence around the insert site of the ROSA26 locus from HSC-3 mVenus, mVenus-CD or mVenus-TK cells. mVenus and suicide genes were successfully inserted into the ROSA26 locus of HSC-3 (Fig. S1A). The proliferation curve of HSC-3 mVenus and HSC3 mVenus-CD cells were the same as the wild-type HSC-3 cells. However, HSC-3 mVenus-TK cells grew more slowly than wild-type HSC-3 cells (Fig. S1). The proliferation rate of other independent HSC-3 mVenus-TK cells was also measured, which showed a slower proliferation rate than wild-type HSC-3 cells (data not shown). This indicated that the expression of mVenus-TK reduces the HSC-3 cell proliferation rate.

To evaluate the response of HSC-3 cells to GCV and 5-FC, HSC-3 cells expressing TK or CD were treated with a series of 2-fold dilutions of GCV or 5-FC for three days. The highest concentration of prodrug was 100 µM.

For HSC-3 mVenus-TK cells, GCV concentrations from 0.2 µM started to inhibit HSC-3 cell viability increase (Fig. 2A). However, high concentrations of GCV, including 50 and 100 µM GCV, inhibited HSC-3 mVenus cells viability increase, which did not express TK (Fig. 2A). The highest concentration of GCV that did not inhibit HSC-3 mVenus viability increase was 25 µM. At this concentration, GCV inhibited 83% of HSC-3 mVenus-TK cell viability increase. Cell viability reduction rate increased at GCV concentration >3.1 µM (Fig. 2A). Consequently, the dose-response curve was divided into two phases based on the viability reduction rate: i) A slow phase at GCV <3.1 µM and ii) a quick phase at GCV ≥3.1 µM. The shape of each phase was sigmoidal, whereas the dose-response curve of HSC-3 mVenus-TK to GCV was double sigmoidal (Fig. 2A).

5-FC started to inhibit the viability increase of HSC-3 mVenus-CD cells at a concentration of 3.1 µM (Fig. 2B). The inhibitory effect plateaued at 50 µM 5-FC. At 100 µM, 5-FC caused 78% inhibition of HSC-3 mVenus-CD cells viability increase, whereas 5-FC at this concentration did not inhibit the viability increase of HSC-3 mVenus cells (Fig. 2B).

To evaluate time-course response, HSC-3 mVenus-TK and HSC-3 mVenus-CD cells were treated with GCV at 25 µM and 5-FC at 100 µM. The cell viability was measured via MTT assay every 24 h.

GCV and 5-FC treatments started to inhibit the viability increase of HSC-3 mVenus-TK and HSC-3 mVenus-CD cells after 24 h treatment (Fig. 3A and B). By contrast, the two prodrugs did not inhibit HSC-3 mVenus cells viability increase. Compared with HSC-3 mVenus cells, cell viability of HSC-3 mVenus-TK and HSC-3 mVenus-CD cells was 78 and 107% after 96 h treatment with GCV and 5-FC, respectively. By contrast, viability of HSC-3 cells reached to 795 and 866% respectively. The time-course response profiles of HSC-3 cells to GCV or 5-FC treatment exhibited a similar profile (Fig. S2).

Bystander effects of TK/GCV and CD/5-FC on HSC-3 cells were investigated. The cell viability and the percentage of suicide gene positive cells were fitted to the exponential equation: y=a*e^-bx + c. The fitting results is satisfactory with R² equal to 0.9995 and 0.9934 for CD/5-FC and TK/GCV respectively (Table II). It was found that CD/5-FC pair had a bigger b value than the TK/GCV pair (Fig. 4 and Table II), indicating that CD/5-FC had a better bystander effect than TK/GCV on HSC-3 cells.

To measure the overall bystander effect, the % of area was calculated based on curves in Fig. 4. First, the area between fitted exponential curve and theoretical no bystander effect curve was calculated. This area is directly related to by-stander effect of suicide gene. Then the percentage of this area to the area of triangle surrounded by x-axis, y-axis, and the theoretical no bystander effect curve was calculated. As shown in Table II, CD/5-FC still has a larger % of area than TK/GCV (31.4 vs 13.2%), confirming that CD/5-FC has a better bystander effect than TK/GCV on HSC-3 cells.

To investigate the effect of prodrugs on HSC-3 cell migration, an improved wound-healing assay was performed. In the improved wound-healing assay, instead of evaluating prodrug effects within the first 24 h treatment, the effect of prodrugs on cell migration was evaluated during the last 24 h of prodrug treatment.

For HSC-3 mVenus-TK cells, 25 µM GCV inhibited cell migration at all three different prodrug treatment durations. By contrast, migration of HSC-3 mVenus cells was not affected by GCV treatment (Fig. 5A-C). GCV inhibited 57-95% of HSC-3 mVenus-TK cell migration in all prodrug treatment durations and the inhibitory effect increased by increasing the duration of treatments (Fig. 5D).

For HSC-3 mVenus-CD cells, 100 µM 5-FC inhibited cell migration in all treatment durations but had no effects on the HSC-3 mVenus cells (Fig. 6A-C). However, the inhibitory effect decreased from 98 to 89% during the last 24 h of the longest prodrug treatment (Fig. 6C and D).