In previously published studies, we have shown that apoptosis can be induced by overexpression of POX by stable transfection ( Liu Y, et al., Carcinogenesis, 26:1335, 2005;Liu Y, et al., Oncogene, 25:5640, 2006) or by increased expression with pharmacologic agents (Pandhare J, et al., J. Biol. Chem., 281:2044, 2006 ). Furthermore, we showed that POX activates both the intrinsic (mitochondrial) and extrinsic (death receptor) pathways (Liu Y, et al., Oncogene, 25:5640, 2006) for apoptosis. These effects are based on the generation of proline-dependent superoxide;co-expression of MnSOD in mitochondria abolished the POX-dependent apoptotic effects. Not only is the apoptotic cascade induced by overexpression of POX, but also progression through the cell cycle is disrupted. There were marked changes in regulatory proteins governing the G2/M checkpoint. Using RT-PCR we found that isoforms of GADD45 were increased. On Western blots, GADD45alpha which was undetectable with POX suppressed, showed a robust signal with POX expression. These findings clearly showed that POX overexpression not only induces apoptosis, but also blocks the G2/M transition, perhaps by inducing GADD (Liu, Y, et al., Cancer Res. , 69:6414, 2009). We also found that POX signaling downregulates COX2/PGE2 (Liu, Y, et al., Oncogene , 27:6729, 2008), and most interestingly, decreases the level of HIF-1alpha by increasing alpha ketoglutarate which increases prolyl hydroxylation of HIF-1alpha and its proteasomal degradation ((Liu, Y, et al., Cancer Res. , 69:6414, 2009). Thus, by a variety of mechanisms, the expression of POX can inhibit growth and initiate apoptosis by a variety of mechanisms. To test whether this POX-dependent apoptotic and growth-inhibition mechanism can be translated to animal models, we performed studies using DLD-tet-off-POX to form xenograft tumors in athymic mice differentially administered doxycycline (100 micrograms/ml) in their drinking water. We found that when doxycycline was removed from the drinking water and POX expressed, the engraftment of tumors was markedly inhibited (Liu, Y, et al., Cancer Res. , 69:6414, 2009). Evidence that POX functions as a tumor suppressor protein in human tumors was provided by immunohisto-chemical examination (IHC) of POX expression in a variety of human tumors with paired normal tissue from the same patient. Out of 36 pairs of tumors from the digestive tract, 28 showed marked decrease or undetectable levels in tumors as compared to normal tissue. From 6 pairs of tumors/ normal tissue from the kidney, all 6 tumors had decreased immunohistochemical levels of POX (Liu, Y, et al., Cancer Res. , 69:6414, 2009). Thus, POX appears to function as a tumor suppressor protein. We sought the mechanism for the loss of this tumor suppressor protein, but did not find any of the more common genetic or epigenetic mechanisms. Instead, we found that a microRNA, miR-23b*, suppressed POX translation by binding to its 3'-UTR. We first confirmed that POX is markedly decreased in tumor tissue. Choosing clear cell renal carcinoma, we obtained samples from an archival collection, in which both tumor and adjacent nonmalignant tissue are on the same slide. We examined every slice fitting these criteria (6) by IHC;every tumor had marked decrease or undetectable POX as compared to normal tissue. In addition, we obtained 16 pairs of frozen clear cell renal carcinomas with paired nonmalignant tissue and showed by Western immunoblots that 13 of the 16 tumors had markedly decreased POX expression as compared to paired, nonmalignant tissues. To investigate the mechanistic role of miRNAs, we first identified a group of miRNAs by computer and then tested them in vitro by comparing 3 renal cancer cell lines with normal kidney epithelial cells. Levels of miR-23b* were high in tumor cells and showed the highest differences between tumor and normal cells. Mimic miR-23b* markedly decreased the expression of POX and anti-miR-23b* increased the expression of POX. Importantly, the functional responses to POX expression (ROS generation, increased apoptosis, or inhibition of prolfieration) were affected by treating with mimic miR-23b* or anti-miR-23b*. These in vitro studies were validated in human tumors by examining the expression of POX compared to the expression of miR-23b* in the 16 paired tumor tissues examined for POX expression. We found there was negative correlation between POX levels and miR-23b*. Furthermore, using in situ hybridization, the expression of miR-23b* was increased in tumors. These studies are the first to show that the function of proline as a microenvironmental substrate for proline oxidase, a tumor suppressor protein, could be regulated by a specific miRNA, miR-23b*. Thus, the increased expression of miR-23b* in tumors and its inhibition of POX makes it a functional oncogene. We are excited that these mechanisms can restore a p53-independent, ROS-dependent programmed cell death. The blockade of miR-23b* to activate POX function in tumors may be used as an adjunct in chemotherapy. Studies are underway to test this possibility. We have also shown that the oncogene c-Myc regulates POX through miR-23b*. In cells stably transfected with a tet-off c-Myc construct, POX decreases as c-Myc is expressed. Importantly, c-Myc increases the level of miR-23b* which inhibits the translation of POX mRNA. These findings on POX complement the effects of c-Myc on glutaminase, an enzyme which plays an important role in cell proliferation. Thus, the effect of c-Myc on glutaminase and POX are oppositely directed. Furthermore, glutaminase is controlled by miR-23a/b, the complementary strand to miR-23b*. The increase in glutaminase is mediated by a decrease in the level of miR-23a/b while its decrease in POX is due to an increase in the level of miR-23b*.
Showing the most recent 10 out of 16 publications
