2.3. Data and Statistical Analysis
Statistical analysis was done using the Statistical Package for Social Science (SPSS) version 18. A difference between mean values is considered significant at p < 0.05. The results were presented as mean ± SD. All the figures were made by GraphPad Prism version 8.
Results
CSCs were effectively enriched by 50 µg/mL CDDP
The percentage of CD44+ CD24- cells was increased using different concentrations of DOX, CDDP, and PTX, however, the most significant enrichment was observed with 50 µg/mL CDDP (50 µg/mL CDDP vs . EAC; p = 0.000000363) (Fig. 1 ). This concentration was therefore selected for subsequent in vivo experiments.
Tumorigenicity of CDDP-Resistant EAC Cells vs. Parent EAC Cells
Based on the proportion of CD44+CD24- cells, CDDP-resistant cells had a higher proportion of CSCs than the parent EAC cells (3.4% vs . 0.1%). CDDP-resistant cancer cells showed a heightened ability to form tumors compared with the sensitive cells where the CDDP-resistant cells-derived tumors were initiated faster and significantly larger than their parent counterparts (Fig. 2A ).
On day 17 when the tumors were harvested, all the resistant cells-derived tumors were significantly larger than their corresponding sensitive tumors except for indomethacin where the difference between indomethacin-receiving resistant and sensitive tumors was not statistically significant (p = 0.44) (Fig. 2B ).
Indomethacin effectively and selectively reduced the tumorigenicity of CDDP-resistant cells in in vivo mouse experiments
Indomethacin in combination with CDDP decreased the tumor volumes in mice bearing parent SEC more significantly than CDDP (SEC vs . SEC-combination; p = 0.00015212, SEC vs. SEC-CDDP;p = 0.00627464), whereas indomethacin alone was not able to significantly reduce the tumor volumes (SEC vs. SEC-indomethacin;p = 0.143). In contrast, for CDDP-resistant tumors, indomethacin dramatically reduced the tumor volumes compared to CDDP alone or combination of both (SEC/CDDP vs. SEC/CDDP-indomethacin; p= 0.00000037, SEC/CDDP vs. SEC/CDDP-CDDP; p = 0.01159254, SEC/CDDP vs. SEC/CDDP-combination; p = 0.00002975) (Fig. 2 ).
The growth rates of the sensitive SEC tumors exceeded those of the resistant tumors
The growth rates of the tumors were computed after the start of the treatment, unexpectedly, the growth rates of all the resistant tumors were lower than their analogous sensitive ones (Fig. 3A ). Noteworthily, the growth rates of the CDDP-resistant tumors were initially faster than the SEC tumors before the treatment has been commenced (growth rate of SEC/CDDP = 0.093952 vs. SEC = 0.058813) (Fig. 3B ).
In a similar pattern to the tumor volumes, the combination of indomethacin and CDDP more effectively impeded the growth rates of the parent SEC tumors than either drug alone (SEC vs.SEC-combination; p = 0.00005500, SEC vs. SEC-CDDP;p = 0.00233442, SEC vs. SEC-indomethacin; p = 0.12394883), while the growth rates of the resistant tumors were slowed down the most by indomethacin (SEC/CDDP vs.SEC/CDDP-indomethacin; p = 0.00013892, SEC/CDDP vs.SEC/CDDP-CDDP; p = 0.04591864, SEC/CDDP vs.SEC/CDDP-combination; p = 0.00503803) (Fig. 3A ).
CDDP-resistant tumors exhibited extensively deregulated miRNAs expressions
We determined the relative expressions of 4 miRNAs that are deregulated in murine breast cancer and involved in drug resistance. Compared with the untreated sensitive tumors, the untreated resistant tumors showed lower expressions of miR-7 and miR-145 besides overexpression of miR-21 and miR-22 (p = 0.037, 0.00024, 0.001, 0.0000136, respectively) (Fig. 4 ).
MiRNAs Expressions are Altered in Response to Indomethacin
miR-7 expression was significantly increased by indomethacin (SECvs . SEC-indomethacin; p = 2.17*10-21, SEC/CDDP vs . SEC/CDDP-indomethacin; p = 8.33*10-7). In contrast, the conventional chemotherapy further downregulated miR-7 in both sensitive and resistant tumors, however, it did not reach statistical significance for the latter (SEC vs . SEC-CDDP; p = 0.037309067230362626, SEC/CDDP vs . SEC/CDDP-CDDP; p = 0.9029500148694073). Strikingly, indomethacin showed an ability to overcome the downregulatory effect of CDDP and even upregulate miR-7 when both used together (SEC vs . SEC-combination; p = 0.00000823, SEC/CDDP vs . SEC/CDDP-combination; p = 0.00129) (Fig. 4A ).
Unlike miR-7 , indomethacin reduced the expression ofmiR-145 (SEC vs . SEC-indomethacin; p = 0.000000057341, SEC/CDDP vs . SEC/CDDP-indomethacin; p = 0.0098) whereas the effect of CDDP was as the same as its effect onmiR-7 (SEC vs . SEC-CDDP; p = 1.88*10-11, SEC/CDDP vs . SEC/CDDP-CDDP;p = 0.000065). Unexpectedly, the combination of both indomethacin and CDDP downregulated miR-145 to a less extent (SEC vs . SEC-combination; p = 0.0000000677, SEC/CDDP vs . SEC/CDDP-combination; p = 0.046) (Fig. 4B ).
For miR-21 , indomethacin significantly downregulated its expression (SEC vs . SEC-indomethacin; p =0.029, SEC/CDDPvs . SEC/CDDP-indomethacin; p =0.032) while it has been overexpressed by CDDP (SEC vs . SEC-CDDP; p = 0.00000083, SEC/CDDP vs . SEC/CDDP-CDDP; p =0.000091). The combination of both drugs has not significantly decreased the expression of miR-21 (SEC vs . SEC-combination; p = 0.058, SEC/CDDP vs . SEC/CDDP-combination; p = 0.059) (Fig. 4C ).
The levels of miR-22 in tumor tissues were significantly reduced by CDDP, indomethacin, and the combination (SEC vs . SEC-CDDP;p = 0.024, SEC/CDDP vs . SEC/CDDP-CDDP; p =0.000125) (SEC vs . SEC-indomethacin; p =0.021, SEC/CDDP vs . SEC/CDDP-indomethacin; p =0.000027) (SEC vs . SEC-combination; p = 0.008, SEC/CDDP vs . SEC/CDDP-combination; p = 1.28*10-10) (Fig. 4D ).
Indomethacin negatively affected the stem cell markers and suppressed the CDDP-induced increase of CD44+CD24- cells
Indomethacin decreased the percentage of CD44+ cells (p = 7.9*10-10) and increased that of CD24+ cells (p = 1.1*10-7), whereas CDDP showed an opposite effect (p = 3.7*10-11 and 0.003 for CD44+ and CD24+, respectively), so as a result, the fraction of CD44+ CD24- cells is considered to be gone up by CDDP unlike indomethacin. Strikingly, when indomethacin was used along with CDDP it has not only abolished the CDDP-associated elevation of CD44+ CD24- cells, but the proportions were further dropped below control levels (p = 1.7*10-10 and 0.019 for CD44+ and CD24+, respectively) (Fig. 5 ).
Both CDDP and indomethacin exerted a similar effect on the percentage of SCa-1+ cells where both drugs decreased the SCa-1+ population (p = 4.7*10-5, 1.099*10-12, and 1.088*10-12 for CDDP, indomethacin, and the combination, respectively) (Fig. 5 ).
Indomethacin showed favorable responses for CD62L+, CD4+, and CD117+ populations
Indomethacin raised the percentages of CD62L+, CD4+, and CD117+ cells (p = 1.08*10-12, 1.08*10-12, and 1.59*10-4, respectively) while a reverse effect was posed by CDDP, however, the decline was not statistically significant for CD62L+ cells (p = 3.7*10-6 and 0.009 for CD4+ and CD117+ cells, respectively). As expected, the combination treatment showed levels that are in between those of indomethacin and CDDP (p = 5.79*10-6, 0.022, and 0.57 for CD62L+, CD4+, and CD117+ cells, respectively) (Fig. 5 ).
Histological Examination
Upon examination of the H & E stained tumor sections, CDDP-resistant tumors were of a higher grade than the parent SEC tumors featuring extensive mitosis with pleomorphism and perineural invasion indicating the aggressiveness of such tumors. In the case of the parent SEC tumors, the three treatment modalities showed different degrees of necrosis with the highest extent of necrotic cells appeared with the co-treatment (Fig. 6 ). Remarkably, despite CDDP showed extensive areas of necrosis, there was a detectable tumor vascular embolus suggesting that CDDP did not target the metastasis-driving cells; CSCs (Fig. 6B2 ).
Discussion and Conclusion
Breast cancer patient’s mortality is on the rise due to the lack of medications that could curb the activity of CSCs which drive distant metastasis and promote disease recurrence after chemotherapy, both of which are the primary causes of cancer-related deaths [22, 23]. Current directions in cancer treatment aim to identify therapies that target CSCs thereby eliminating the roots of the disease.
NSAIDs have proven their chemoprophylactic effects in BC, and several studies have demonstrated that NSAIDs can even reduce the risk of BC recurrence [11]. NSAIDs reduced metastasis in cancer patients, of them, BC patients were included [24]. Since CSCs are the ones responsible for relapse and metastasis [6], we propose, based on these studies and our findings, that the therapeutic effects of NSAIDs might be ascribed to their anti-CSC effect.
Although there is extensive research on the CSC-inhibitory effect of NSAIDs, much of the research done investigated the modulation of stemness genes and epithelial mesenchymal transition (EMT) markers by NSAIDs as possible mechanisms for their CSC-suppressing ability [25-27] but little has been done on whether NSAIDs combat CSCs by affecting either miRNAs that are implicated in the drug resistance or the immune system whom CSCs can circumvent.
Herein, we aimed to examine the effect of indomethacin on two oncomirs and two tumor suppressor miRNAs in BC besides showcasing a little cue on the likely involvement of the antitumor immunity as a mechanism underlying the CSC-suppression mediated by indomethacin.
Much evidence indicate that CSCs subset markedly expands following chemotherapy [9] and we used this approach to enrich CSCs originally existing in EAC as these cells represent only a very small fraction of cancer population (0.1%) which could have hindered our ability to either effectively induce resistant tumors or discover a potentially pronounced anti-CSC effect of the candidate drugs. However, the percentage of the CDDP-enriched CSCs was 3.4% which is 34-fold higher than the portion in the parent EAC. As such, this CSCs-enriched population proved to be highly tumorigenic upon implantation in mice compared to their more differentiated progeny when an equivalent number of cells were injected and our result is in harmony with a previous report [28]. Intriguingly, although the tumors derived from the CDDP-resistant cells exhibited larger volumes, it grew slower than the parent tumors. These seemingly paradoxical behaviors could be potentially accredited to a relative shortage of oxygen and nutrients in quite big tumors which in turn retards the growth rate of such tumors [21].
Liu et al. studied the growth kinetics of BCSCs and proposed mathematical models thereafter. According to his study, non-stem cancer cells exert negative feedback on the proliferation rate and self-renewal ability of CSCs [29]. That could elaborate on why the antitumor effect of indomethacin on CDDP-resistant tumors was far more significant than the co-treatment. As CDDP preferentially targets the bulk tumor cells, it will be successful in letting the tumor sizes down, but meanwhile, the negative feedbacks on CSCs will be eliminated, therefore, CSCs will symmetrically proliferate at a high rate, and this is consistent with the dramatic increase in the CSCs fraction after CDDP treatment.
On the other hand, indomethacin selectively targets CSCs leaving the differentiated cancer cells and their negative feedbacks in action. As a result, the division of CSCs will be shifted towards an asymmetric one at a slower pace and this agrees well with the sharp drop in the CSCs population post indomethacin treatment. It is commonly believed that the optimum therapeutic strategy to achieve complete remission of cancer is to combine drugs that target both CSCs and non-CSCs. Based on our findings and the proposed models by Liu, Johnson et al. , the removal of the differentiated tumor cells will not make the outcome any better [29].
Concerning the more differentiated SEC tumors, although indomethacin has not significantly reduced their tumors sizes Liu, Johnson et al. , expected for such CSCs-targeted therapies to suppress tumor growth and inhibit relapse on the long term. On the contrary, relapse is probable with conventional chemotherapies despite the powerful fall in tumor size because of the accompanied CSCs enrichment [29]. This makes us recall the histological examination of SEC-CDDP tumors when we found a notable tumor vascular embolus indicating the advert of metastasis and subsequent relapse.
We then examined the relative expression of different miRNAs. BothmiR-7 and miR-145 were more downregulated in the CDDP-resistant tumors. A similar finding was observed in a study conducted by Pogribny et al. where miR-7 levels were 2.3 fold more downregulated in CDDP-resistant MCF-7 compared to MCF-7 cells [30]. Others reported that gall bladder cancer cells were sensitized to CDDP through miR-145 -mediated regulation of MRP1 [31].
Contrarily, both miR-21 and miR-22 were overexpressed in SEC/CDDP tumors compared to the non-resistant ones. miR-21 was found to be involved in CDDP resistance through targeting the tumor suppressor phosphatase and tensin homolog (PTEN) in gastric cancer [32] and neuroblastoma and was overexpressed in CDDP-resistant cell lines [33]. In another study, the sensitivity of MCF-7 cells to CDDP was enhanced by knockdown of dicer which is accompanied by lower expression of miR-21 . Collectively, these findings could explain the decreased sensitivity of the tumors derived from the CSCs-enriched population to CDDP treatment unlike the parent SEC tumors which responded better.
To examine whether indomethacin is capable of antagonizing the immunosuppressive impact of CDDP, we determined the levels of conventional blood and immune cells markers. CD4, CD117, and CD62L (otherwise known as L-selectin) are basic markers for T helper cells (Th) [34], hematopoietic stem cells (HSCs) [35], and white blood cells (WBCs) [36], respectively. CD4 is found on the surface of other immune cells such as dendritic cells [37] and natural killer T cells [38].
Although CDDP treatment has drastically shrunk the SEC tumors, its effect on the CDDP-resistant tumors was subtle. CDDP enriched CD44+ CD24- cells and further downregulated miR-7 and miR-145 while upregulatedmiR-21 . The mechanisms by which CDDP modulated these miRNAs are still obscure as most of the studies examined the consequence of miRNAs expression pattern on the sensitivity of cancer to CDDP but none, to the best of our knowledge, have tested the effect of CDDP on miRNAs expression.
Moreover, CDDP decreased circulating CD117+, CD62L+, and CD4+ cells as well. A possible explanation for such deficiency is the well-known CDDP-prompted myelosuppression effect. A study conducted by Zhao et al. showed that CDDP-provoked G2/M phase arrest in bone marrow cells resulting in bone marrow suppression [39].
On the other hand, albeit the SEC tumors sizes have not been significantly decreased by indomethacin, its effect on SEC/CPPD tumors was robust. Cells positive for CD44 or SCa-1 and those negative for CD24 have been reduced by indomethacin, additionally, it promoted the expression of miR-7 and declined that of miR-21 andmiR-22 . Majumder et al . have found that COX-2 induces BCSCs through EP4-mediated NOTCH/WNT axis and this effect was reversed with COX-2 inhibitor or EP4 antagonist [40]. This could clarify the indomethacin-associated decrease in CSCs markers. Furthermore, an earlier study demonstrated that pre-treatment of mammary cancer cells with COX-2 inhibitors caused suppression of radiotherapy-induced SCa-1+ cells [41] which supports the relevant finding of the indomethacin-accompanied reduction of SCa-1+ pool.
Moreover, as miR-7 blocks the growth of CSCs in BC [42], it might explain the decrease of CD44+CD24- cells in indomethacin-treated mice due to its upregulatory effect on miR-7 . Bourguignon et al . showed that miR-21 is induced in BC by binding of hyalouronan with CD44 [43], so miR-21 expression might be lowered in response to indomethacin due to the associated reduction in CD44 level. Consistently, it has been reported that the downregulation ofmiR-21 stimulated the differentiation of chemo-resistant colon cancer and consequently enhanced the sensitivity to therapy [44]. Our findings provide an illustration of the enhanced susceptibility of the established tumors to CDDP when used simultaneously with indomethacin.
Of note, despite miR-145 is thought to be a tumor suppressor, it displayed a contradictory effect where it suppressed the differentiation of BCSCs via the 3‘-untranslated region (3‘-UTR) of insulin receptor substrate 1 [45].
Furthermore, the percentages of cells positive for each of CD117, CD62L, and CD4 have been raised by indomethacin. Hoggatt et al.[46], have documented that NSAIDs can mobilize the HSCs from the bone marrow into the bloodstream. The mechanism of the NSAIDs mobilizing effect is based on inhibition of PGE2 synthesis which disrupts the production of osteopontin, a protein secreted by osteoblasts, that hooks the stem cells to the bone marrow. Since these HSCs give rise to all blood cell types, this might elucidate the indomethacin-mediated increase in blood cells expressing CD117, CD62L, CD4, and CD24. Considering this mobilizing effect, indomethacin therapy, unlike chemotherapy, can exclude the need for HSCs transplantation when used by cancer patients.
Aside from counteracting CDDP-resistance, indomethacin probably eradicates CSCs through immunostimulatory effects. It has been documented that PGE2 drives tumor progression by inducing myeloid-derived suppressor cells (MDSCs) and that EP antagonists block MDSCs differentiation [47]. MDSCs are immunosuppressive and they have been involved in BC promotion by conferring stem cell-like qualities as well as suppressing T-cell activation [48].
Interestingly, tumor-induced MDSCs downregulate the expression of L-selectin on CD4+ and CD8+ T cells impairing L-selectin dependent cell-mediated antitumor immunity [49] and this might explain the indomethacin-induced increase in CD62L+ cells. Furthermore, PGE2 reduces dendritic cells (DCs) differentiation and this type of immune cells is thought to eradicate CSCs [47, 50].
Intriguingly, the activity of the immune cells is influenced by miRNAs. For example, MDSCs highly express miR-21 which alters its differentiation and functional activity [51]. As indomethacin downregulates miR-21 , this could provide another support on how indomethacin might impede MDSCs activity. Moreover, miR-21downregulates the tumor suppressor programmed cell death-4 (PDCD4). PDCD4 inhibits the expression of the immunosuppressives IL-10 and IL-4 [52]. Both IL-4 and IL-10 promote CSCs survival [53, 54]. Furthermore, the expression of IL-12 is reciprocal to miR-21[52] and unlike IL-10 and IL-4, IL-12 inhibits CSCs [55]. Hence, indomethacin may kill CSCs through its downregulatory impact onmiR-21 and the subsequent reduction in IL-4 and IL-10 as well as elevation of IL-12.
Additionally, it has been shown that PGE2 suppresses T cell proliferation [56], thus, the apparent increase in peripheral CD4+ population could be accounted for the PGE2 synthesis inhibition by indomethacin. Moreover, we can expect that indomethacin might augment Th1 actions as long as PGE2 suppresses Th1 responses [56]. Notably, both IL-4 and IL-10 inhibit Th1 cell production and function, respectively [57, 58] whereas IL-12 is indeed involved in Th1 polarization [52]. Given these, we could deduce that the indomethacin-associated increase in peripheral CD4+ cells is potentially in favor of Th1 cell type.
In fact, recent studies have indicated that immunotherapeutic strategies in BC should be directed towards type 1 immune response that is rich in Th1 cells [59]. However, whether the indomethacin-associated CD4+ cells are tumor-antigen specific or even have a role in fighting CSCs needs to be investigated. Interestingly, elevated levels of antigen specific CD4+ T cells in the peripheral blood of BC patients who undergo anticancer therapy is not unfamiliar. This is exemplified by trastuzumab therapy which boosted the levels of circulating HER2 specific CD4+ Th1 cells in BC patients who otherwise normally have few circulating antigen-specific T cells able to eradicate their tumors [59].
Impressively, one study discovered the existence of anti-CSC specific CD4+ T cells in the peripheral blood of ovarian cancer patients [60]. This encourages us to complement the current work by conducting further studies on the phenotypic and functional characterization of the indomethacin-induced CD4+cells. Moreover, whether the high levels of CD4+ cells in the blood are correlated with the disease outcome remains to be explored. If this is the case, this could be harnessed as a prognostic tool instead of invasively monitoring the level of tumor-infiltrating lymphocytes.
There is also a link between miR-22 and the immune system, for instance, miR-22 is upregulated in DCs retarding their anti-tumor activity [61]. Thus, the indomethacin-induced downregulation ofmiR-22 could promote the CSC-suppressing ability of tumor-associated DCs.
In conclusion, this study shows that indomethacin has a CSC-suppressing effect in BC and inhibits CDDP-induced increase of CSCs presumably through modulation of miRNAs implicated in drug resistance as well as triggering an antitumor immune response. To this end, NSAIDs sound to have far-reaching mechanisms behind its CSC-suppressing activity that necessitate deeper investigations to exactly unravel the pathways of such mechanisms.