Researchers have previously described multiple functions for miR-708-5p in chemoresistance. COX-2 and PGE2 also have well documented roles in promoting resistance in cancer. Additionally, we have shown that COX-2 and miR-708-5p expression is inversely correlated in lung cancer cells and tumors (55). Therefore, we examined various chemoresistant aspects of miR-708-5p and its regulation of COX-2/mPGES-1 derived PGE2 in lung cancer cells. We tested the ability of ERL, PAC, and DEX to modulate miR-708-5p and AA signaling in lung cancer cells. We treated A549 cells with ERL, PAC, or DEX for 48 hours and measured changes in mature COX-2 and mPGES-1 expression. We found that COX-2 mRNA was significantly decreased after ERL and DEX treatment compared to a vehicle control, whereas mPGES-1 mRNA was unchanged after all treatments (Fig. 1A). We also measured COX-2 and mPGES-1 protein expression and observed similar results to RT-qPCR data (Fig. 1B). Next, we investigated changes in miR-708-5p expression after treatment with ERL, PAC, and DEX. We observed significantly higher miR-708-5p expression in A549 cells 24 hours after treatment of each therapy (Fig. 1C, p <.05, n = 3). Given these results, we explored transcription factors that may be regulating chemotherapeutic-induced miR-708-5p expression.
We examined known regulators of miR-708-5p and their correlation to miR-708-5p expression in NSCLC, LUAD, and LUSC patients. In the broader NSCLC subtype, every known regulator was significantly correlated with miR-708-5p expression (Table 1). Within NSCLC, the mRNA expression of transcription factors in LUAD were not significantly correlated with miR-708-5p expression, but transcription factor mRNA expression in LUSC tumors were highly correlated with miR-708-5p expression (Table 1). More specifically, CHOP was significantly positively correlated with miR-708-5p expression in NSCLC (Table 1, p = .207, R2 = .0418, p = 2.18x10-11) and LUSC (Table 1, p = .188, R2 = .0333, p = 2.32x10-5) tumors. It was previously shown that CHOP also induces p53, the most commonly mutated tumor suppressor in cancer (69). Given the importance of p53 in tumorigenesis, we
Table 1. miR-708-5p expression correlates with known miR-708-p regulators in lung cancer tumors.
|
|
NSCLC
|
LUAD
|
LUSC
|
Gene
|
miRNA
|
Correlation
|
pvalue
|
Correlation
|
pvalue
|
Correlation
|
pvalue
|
CHOP ^
|
miR-708
|
0.207
|
2.181E-11
|
0.0445
|
0.3085
|
0.188
|
0.00002323
|
GRα ^
|
miR-708
|
-0.236
|
1.825E-14
|
-0.0579
|
0.1851
|
-0.203
|
0.000004526
|
MYC ^
|
miR-708
|
0.333
|
4.618E-28
|
-0.000522
|
0.9905
|
0.167
|
0.0001741
|
E2F1 ^
|
miR-708
|
0.119
|
0.000132
|
0.13
|
0.002744
|
0.106
|
0.01712
|
CTBP2 *
|
miR-708
|
-0.19
|
8.999E-10
|
0.0799
|
0.06715
|
-0.0134
|
0.7638
|
RAD21 ^
|
miR-708
|
0.206
|
2.502E-11
|
0.15
|
0.0005638
|
0.154
|
0.0005529
|
C/EBP-β ^
|
miR-708
|
-0.271
|
8.746E-19
|
0.0728
|
0.09539
|
-0.382
|
6.762E-19
|
CTCF *
|
miR-708
|
0.16
|
2.434E-07
|
-0.0445
|
0.3086
|
0.0894
|
0.04538
|
TCGA mRNA/miRNA data showing correlation, and significance (pvalue) of miR-708-5p and various validated regulators of miR-708-5p expression in NSCLC (n = 864), LUAD (n = 442), and LUSC (n = 424) tumors. Italicized font indicates a significant negative correlation; underlined indicates significant positive correlation; and black font indicates no significant correlation. (^) represents a positive regulator of miR-708-5p expression, while (*) represents a repressor of miR- 708-5p expression.
decided to investigate its relationship with miR-708-5p. We discovered that p53 mRNA expression was positively correlated with miR-708-5p expression in NSCLC, LUAD, and LUSC tumors (Table 2). Given CHOP’s previously defined regulation of miR-708-5p, as well as CHOP and p53’s profound roles in apoptosis, we examined if ERL, PAC, and DEX were regulating CHOP and p53 expression in lung cancer cells.
ERL treatment promoted high levels of CHOP, while PAC and DEX more modestly increased CHOP protein expression (Fig. 2A). Moreover, high CHOP mRNA expression positively associated with increased survival rates in LUSC patients (Supplemental Fig. 1C, p = .014, n = 424), following similar patterns seen with miR-708-5p in NSCLC. We also observed that while ERL did not affect A549 p53 expression, PAC, and to a lesser extent DEX, induced p53 protein expression (Fig. 2A). Next, we analyzed how miR-708-5p expression affected survival rates in p53 WT and mutant (MUT) tumors. miR-708-5p expression had no effect on survival in LUAD WT (Supplemental Fig. 2A, p = .79, n = 240) or MUT (Supplemental Fig. 2B, p = .45, n = 263) p53 tumors. While miR-708-5p levels in LUSC WT p53 tumors had no effect on survival (Supplemental Fig. 2C, p = .91, n = 70), high miR-708-5p expression significantly enhanced survival in LUSC MUT p53 tumors (Supplemental Fig. 2D, p = .041, n = 389). Collectively, these data suggest that high miR-708-5p may improve survival rates in LUSC patients containing p53 mutations.
Given the ability of ERL, PAC, and DEX to regulate miR-708-5p possibly through p53, and CHOP, we examined whether miR-708-5p regulates p53 and CHOP expression in lung cancer cells. To do this, we used A549 cells, as are p53 WT. We found that A549 cells transiently transfected with miR-708-5p induced p53 and CHOP protein expression while simultaneously reducing COX-2, mPGES-1, and Survivin protein levels (Fig. 2B).
Table 2. Mature miR-708-5p and p53 mRNA expression are positively correlated in NSCLC tumors.
Subtype
|
Gene
|
miRNA
|
Correlation
|
Adj.R^2
|
pvalue
|
NSCLC
|
p53
|
miR-708
|
0.178
|
0.0306
|
9.66E-09
|
LUAD
|
p53
|
miR-708
|
0.107
|
0.00964
|
0.01374
|
LUSC
|
p53
|
miR-708
|
0.154
|
0.0219
|
0.0005222
|
TCGA mRNA/miRNA data showing correlation, adjusted R2, and significance (pvalue) of miR-708-5p and p53 mRNA in NSCLC (n = 864), LUAD (n = 442), and LUSC (n = 424) tumors. Underlined lettering indicates a significant positive correlation.
Given miR-708-5p’s ability to enhance p53 and CHOP protein expression, we examined the molecular and phenotypic consequences of combinatory miR-708-5p and chemotherapy treatments in lung cancer cells. First, we tested ERL alone or in combination with a NC miR/miR-708-5p in A549 cells. Western blot analysis revealed that ERL treatment alone decreased COX-2 protein expression while increasing CHOP protein levels (Fig. 3A). ERL + miR-708-5p treatment further reduced COX-2 protein expression, while also suppressing mPGES-1 and Survivin protein levels (Fig. 3A). CHOP expression was not enhanced further in the ERL + miR-708-5p samples, but ERL + miR-708-5p treatment increased p53 protein expression compared to vehicle, ERL, and ERL + NC miR samples (Fig. 3A). Next, we investigated PAC treatment alone or in combination with a NC miR/miR-708-5p. We found that PAC had no effect on COX-2, mPGES-1, or CHOP protein expression, but induced Survivin and p53 protein levels (Fig. 3B). Intriguingly, PAC + miR-708-5p combination treatment suppressed PAC-induced Survivin expression, while also reducing COX-2 and mPGES-1 protein levels (Fig. 3B). While PAC + miR-708-5p treatment did not further increase p53 expression, the combination treatment strongly induced CHOP protein expression 10.2 fold in A549 cells (Fig. 3B). Together, combination treatments of ERL/PAC + miR-708-5p suppress pro-tumorigenic signaling (COX-2, mPGES-1, Survivin) while activating pro-apoptotic pathways (CHOP, p53) greater than either therapy alone. Therefore, we further explored the phenotypic impact of combination treatments on lung cancer cells.
To test the combinatory phenotypic effects of ERL/PAC and miR-708-5p in lung cancer cells, we examined changes in proliferation and apoptosis via Ki-67 and Annexin V staining, respectively. Ki-67 is a marker for proliferating cells, while Annexin V detects externalized phosphatidylserine (PS), a commonly used apoptosis marker (70, 71). First, we examined whether ERL and miR-708-5p treatment enhanced anti-proliferative activities greater than either treatment alone. We found that ERL treatment alone significantly decreased A549 proliferating cells from 95% in vehicle treated samples to 73% in ERL treated samples (Fig. 4, p < .0001, n ≥ 3). While ERL + NC miR treatment was similar to ERL treatment alone, ERL + miR-708-5p further suppressed lung cancer cell proliferation, with 51% of combinatory treated A549 cells were Ki-67+ (Fig. 4, p < .0001, n ≥ 3). Next, we investigated how the combinatory treatment was altering lung cancer cell cycle progression. We discovered that ERL alone significantly reduced the percentage of A549 cells in G1 and G2/M phase (80% to 59%) while also inducing cells to accumulate in G0 phase (Fig. 4, p < .01, n ≥ 3). Interestingly, ERL + miR-708-5p treatment further reduced the percent of A549 cells in G1 and G2/M phase to 43% and significantly reduced the number of cells in S phase by half (Fig. 4, p < .01, n ≥ 3). This reduction in actively proliferating phases was paired with a significant accumulation of non-proliferation G0 phase A549 cells compared to ERL and ERL + NC miR treatments (Fig. 4, p < .0001, n ≥ 3). Collectively, these data reveal that ERL and miR-708-5p cooperate to enhance anti-proliferative activities in lung cancer cells. Given these data, we investigated the effects this combinatory treatment had on apoptosis.
While ERL and miR-708-5p have both been shown to induce apoptosis in lung cancer cells, researchers have not studied their combinatory potential. To examine this, we utilized flow cytometry and Annexin V staining. We observed that ERL significantly increased PS+ cell number compared to vehicle treatment (Fig. 5, p < .05, n ≥ 3). ERL + miR-708-5p treatment enhanced the percentage of PS+ cells from 17% in ERL treatments to 39% in ERL + miR-708-5p treated samples (Fig. 5, p < .0001, n ≥ 3). While these data reveal an increase in apoptosis, they do not distinguish between early and late apoptosis. As Figure 5E reveals, ERL + NC miR increased the percent of early apoptotic A549 cells (4% to 9.6%), while ERL + miR-708-5p treatment further intensified the early apoptotic population to 28% (Fig. 5F/G, p < .05, n ≥ 3). Furthermore, ERL + miR-708-5p increased late apoptotic events while no other treatment was significantly different from our vehicle control (Fig. 5F/H, p < .05, n ≥ 3). We conclude that while ERL induces apoptosis, ERL + miR-708-5p intensifies lung cancer cell death. These data, as well as the Ki-67 data, reveal an additive anti-tumor ERL and miR-708-5p combination therapy that reduces proliferation and survival greater than either treatment alone. Now that we have studied the combinatory effects of ERL and miR-708-5p, we repeated our studies with PAC and miR-708-5p.
We investigated the effect PAC alone, or in combination with miR-708-5p, had on lung cancer proliferation. We found that PAC, PAC + NC miR, and PAC + miR-708-5p treatment significantly reduced the percent of Ki-67+ A549 cells (Fig. 6A/B, p < .0001, n ≥ 3). Interestingly, while the PAC + miR-708-5p treatment decreased Ki-67 positivity, this treatment had a significantly higher Ki-67+ population compared to PAC and PAC + NC miR treated samples (Fig. 6B, p < .05, n ≥ 3). Next, we examined how PAC and miR-708-5p were altering the cell cycle. We discovered that PAC + miR-708-5p treatment enhanced the percent of A549 cells in G0 to 20.6%, albeit significantly less than PAC (31%) and PAC + NC miR (35%) treatments (Fig. 6, p < .01, n ≥ 3). PAC and PAC + NC miR treatments reduced the number of A549 cells in G1 phase from 56% to 39% and 29%, respectively, while increasing the percent of cells in G0 phase (Fig. 6, p < .0001, n ≥ 3). Moreover, PAC + miR-708-5p further decreased the G1 population to 13%, while significantly increasing the percent of A549 cells in G2/M phase (Fig. 6, p < .0001, n ≥ 3). While this may suggest the combination treatment is promoting proliferation, it does not take into account PAC’s anti-tumorigenic mechanism of action. PAC is a microtubule stabilizer that locks dividing cells in the G2/M phase. This PAC-induced stalling increases cellular stress, leading to apoptosis. Therefore, miR-708-5p enhanced the anti-proliferative effects of PAC, as it further decreased the percent of G1 cells while also enhancing the PAC’s G2/M-arresting effects (Fig. 6, p < .0001, n ≥ 3). While PAC regulates proliferation, we also need to investigate the effects of PAC and miR-708-5p combination treatment on lung cancer cell apoptotic rates.
Given PAC’s anti-tumor characteristics, as well as miR-708-5p’s pro-apoptotic functions, we studied the combinatory effects of these two treatments in lung cancer cells. PAC treatment alone increased the number of late apoptotic cells compared to vehicle control, while there was no significant increase in PS positivity or early apoptotic events (Fig. 7, p < .05, n ≥ 3). PAC + NC miR did not significantly affect PS positivity, early, or late apoptotic events when compared to vehicle control (Fig. 7, p = n.s, n ≥ 3). Conversely, combination PAC + miR-708-5p dramatically increased the percent of PS+ A549 cells from 11.5% to 39% (Fig. 7, p < .0001, n ≥ 3). Moreover, PAC + miR-708-5p treatment increased the number of early and late apoptotic/dead cells compared to vehicle control (Fig. 7, p < .05, n ≥ 3). Collectively, these data suggest that PAC + miR-708-5p treatment significantly enhances the pro-apoptotic effects of PAC on lung cancer cells. These data paired with our proliferation data form the basis for exploring the therapeutic combinatory potential of PAC and miR-708-5p in lung cancer.
While ERL and PAC are commonly used to treat lung tumors, their efficacy in the clinic is limited because of developed resistance to these drugs. Given the annotated functions of COX-2/mPGES-1 derived PGE2 and miR-708-5p in resistance, we investigated the role of AA signaling and miR-708-5p in ERL and PAC resistance. First, we created A549 ERL resistant (A549-ER) and PAC resistant (A549-PR) cell lines as previously described (66). We confirmed our cells were resistant by comparing chemotherapeutic-induced changes in proliferation and apoptosis in naïve (A549-WT) and resistant (A549-ER, A549-PR) cell lines (Supplemental Figs. 3-6). We found that COX-2 protein expression was higher in our A549-ER and A549-PR cells compared to A549-WT cells (Fig. 8A). Next, we found that miR-708-5p expression was significantly lower at baseline in A549-ER (-69%) and A549-PR (-66%) cells compared to A549-WT cells (Fig. 8B, p < .01, n = 3). We also examined the ability of ERL and PAC to induce miR-708-5p in our resistant cell lines. Beyond being expressed less in the resistant cell lines, miR-708-5p expression was no longer responsive to ERL (Fig. 8C, p < .05, n = 3) or PAC (Fig. 8D, p < .05, n = 3) treatment in resistant cells, respectively. As miR-708-5p is underexpressed and no longer responsive to ERL or PAC treatments in our resistant cells, we explored the phenotypic value of miR-708-5p treatment to overcome resistance in our lung cells.
Given enhanced AA signaling paired with decreased miR-708-5p expression in our resistant cell lines, we explored the ability of miR-708-5p to resensitize A549-ER and A549-PR cells to ERL and PAC treatments, respectively. While ERL treatment alone or in combination with NC miR did not affect Ki-67 positivity in A549-ER cells, ERL + miR-708-5p decreased A549-ER Ki-67+ cell number by 24% (Fig. 9, p < .001, n ≥ 3). ERL and ERL + NC miR treatments insignificantly reduced S and G2/M populations in A549-ER cells, whereas ERL + miR-708-5p significantly suppressed the number of cells in S and G2/M phase by 21% (Fig. 9, p < .05, n ≥ 3). Moreover, ERL and ERL + NC miR had no effect on G0 and G1 populations, but ERL + miR-708-5p treatment significantly decreased the G1 population by 20% in A549-ER cells (Fig. 9, p < .0001, n ≥ 3). Next, we investigated survival rates in A549-ER cells. We found that ERL and ERL + NC miR treatments had no effect on PS+ (Fig. 10, p = n.s., n ≥ 3) or apoptosis rates (Fig. 10, p = n.s., n ≥ 3) compared to vehicle control. On the other hand, ERL + miR-708-5p significantly increased the number of PS+ cells from 12% in vehicle samples to 48% in ERL + miR-708-5p samples (Fig. 10A/B, p < .0001, n ≥ 3). ERL + miR-708-5p treatment significantly elevated the percent of early apoptotic and late apoptotic/dead cells compared to vehicle, ERL, and ERL + NC miR treatments (Fig. 10F-H, p < .0001, n ≥ 3). Collectively, these data suggest that ERL + miR-708-5p represses proliferation and stimulates apoptosis in ERL resistant lung cancer cells. After we examined ERL and miR-708-5p’s effects on A549-ER cells, we also replicated our studies in A549-PR cells.
We first measured PAC induced changes in A549-PR proliferation. PAC alone or in combination with a NC miR had no effect on the number of Ki-67+ cells compared to vehicle control (Fig. 11, p = n.s., n ≥ 3). Conversely, combination treatment of PAC + miR-708-5p significantly reduced the percent of proliferating A549-PR cells from 95 to 82% (Fig. 11A/B, p < .05, n ≥ 3). More specifically, it appears PAC + miR-708-5p is reducing proliferation by driving A549-PR cells into G0 phase, as well as reducing the percent of cells in S phase by 5.4% (Fig. 11, p < .001, n ≥ 3). PAC alone or in combination with NC miR had no affect the percentage of A549-PR cells in S or G2/M phases (Fig. 11, p = n.s., n ≥ 3). Lastly, PAC + miR-708-5p significantly increased the G2/M A549-PR population by 10-16% compared to other treatments (Fig. 11, p < .0001, n ≥ 3). As previously stated, PAC stalls proliferating cells in G2/M phase by stabilizing microtubules. This prevents cell division, which leads to increased cellular stress and apoptosis. Therefore, while it appears PAC may be promoting cell division, it is actually locking cells into G2/M phase, which ultimately leads to apoptosis. Therefore, we examined if PAC + miR-708-5p treatment modulated apoptotic rates in A549-PR cells. We discovered that PAC alone increased the PS+ population from 13% to 17%, which was further increased in the PAC + miR-708-5p co-treatment to 51% (Fig. 12, p < .05, n ≥ 3). Moreover, it appears that PAC + miR-708-5p treatment is amplifying both early apoptotic and late apoptotic/dead events compared to other treatments in A549-PR cells (Fig. 12, p < .01, n ≥ 3). Together, these results suggest miR-708-5p may be an important component of PAC resistance in lung cancer cells. Co-treatment of PAC + miR-708-5p helps to overcome resistance, highlighting the therapeutic potential of miR-708-5p in naïve and chemotherapeutic-resistant lung tumors.