The Moelcular Mechanism Underlying Differential Response Of Medulloblastoma Cells To Retinoic Acid

Backgrounds and Objectives:Medulloblastoma (MB) is the most common malignant tumour of central nervous system (CNS) during childhood with high invasiveness and mortality. It is originated from precursors of cerebellar neurons and occurs in most cases in vermiform of cerebellum which is located in posterior cranial fossa. The conventional therapy was surgical removal combined with chemo-and/or radiotherapy. For the past few years along with improvement of tumor pinpoint and optimization of surgical types and chemoradiation scheme, curative effect and survival rate were somewhat ameliorated. But the high recurrence rate, short-and long-term adverse effects includes neurological and/or mental disorders of conventional postsurgical adjuvant therapies have become therapeutic problem of neurooncology, which also lower the patient's living quality.MB belongs to primitive neuroectodermal tumors (PNET) and results from abnormal cell proliferation and deficient terminal differentiation. Normal cerebellum development is regulated by many signaling pathways. The development of MB is related with abnormalities of some signaling pathways which promote continuous proliferation of precursors and mature obstruction. So that, MB seems to keep the potential to return to differentiation route if being treated by suitable inducer(s), which may be provide new chemotherapy method. It has been proved that some drugs induce differentiation of MB cells in terms of neuron-like phenotype and expression of neuron biomarkers. Our previous studies indicated that resveratrol, a polyphenolic compound existed in grapes and many other natural foods, exhibited ideal anti-medulloblastoma effects by promoting either differentiation or apoptosis in time-and dose-dependent fashions, followed by the appearance of neuronal biomarkers such as neurogeninl (NGN1) and synaptophysin (SYP) and the activation of multiple signaling pathways related with neural development and tumorgenesis. The metabolic style and bioavailability of resveratrol in brain tumors of CNS is in investigation in order to identify internal molecular mechanism.In difference with resveratrol, all-trans retinoic acid (RA) is conventionally used differentiation-promoting agent in clinic. RA, the derivative of vitamin A, could inhibit growth of MB cells and facilitate differentiation. It has widespread biological effect in embroyo development, spermgenesis and visual development which obtained from animal food such as meat and vegetable food such as carrot. Vitamin A deficiency animals were susceptible to cancers in earlier studies. Meanwhile natural or artificial retinoids could inhibit tumorigenesis efficiently. Tumorigenesis accompanys with transformation of normal cells into abnormal strong proliferating and dedifferentiating ones. Because of the growth inhibitory and differentiation promoting effects, RA was used as first line differentiation-promoting agent in cancer chemotherapy in the clinical management of acute promyelocytic leukemia (APL). Earlier studies in our lab also proved and supported the above mentioned conclusion. Moreover, the RA differentiation-promoting function was correlated with apoptotic related gene-Fas, leukemia inhibitory factor (LIF) and telomerase activity. Nevertheless, the studies of our lab and others also indicated that the thera-peutic effects of RA were distinct in different MB cells which respond with growth-inhibition and differentiation or weakly or no response. So far the internal molecular mechanism underlying that phenomenon has not been clarified.It has been already proved that vitamin A (retinol) should turns into retinaldehyde after body admission, then convert into RA by oxidation. RA's pharmacologic function was implemented by its signaling pathway. Liposolubility makes extrinsic RA permeating from extracellular matrix into cytoplasm directly and exdogenous RA is synthesized by retinaldehyde dehydrogenase type 2(RALDH-2) in cytoplasm. Intracellular RA has two fates, one is decomposed by (cytochrome P450 26A1/CYP26A1) into polar products which will be eliminated out of cells, the other is to bind with cellular retinoic acid binding protein 2 (CRABP-Ⅱ) and form complex with retinoid acid receptor and retinoid x receptor (RAR and RXR) after transmitting from cytoplasm into nucleus. RAR and RXR, the intranuclear transcription factor, could induce transcription by binding with retinoic acid response element (RARE) within the promoter of the downstream genes. It is thus clear that CRABP, includes 2 isomers-Ⅰand-Ⅱ, CYP26A1 and RAR/ RXR subtypes play principal roles in transduction of RA signals. The main function of CRABP-Ⅰis to promote metabolic efficiency of CYP26A1. Both RAR and RXR have 3 subtypes, RARα/β/γand RXRα/β/γwhich function as dipolymer.A body of evidence shows that tumorigenesis is process of dysfunction and abnormality of one or more signaling pathways. Similar with the pathways mediated by Notch and Wnt, RA signaling pathway not only plays important roles in neural development but also links with the development of MB and other cancers such as leukemia. This signaling participates in cell fate decision through impact on proliferation and differentiation. Activation of RA signaling pathway induced by RAR/RXR binding with ligand promotes differentiation and apoptosis of cancers in cell type related fashion. Our previous work indicated that Med-3, a MB cell line, showed differentiation tendency after RA treatment. But UW228-2, another MB cell line, showed little response to RA treatment. The differential responses of MB cells to RA thus provide an ideal experimental model for exploring the underlying molecular mechanism leading to RA sensitivity or resistance. It has important significance for personalized RA therapy and prognosis-promoting in clinic.It has been known that the state of RA signaling pathway is correlated with its chemosensitivity. Classical abnormality of RA signaling compo-nents is RARa form fusion protein with promyelocytic leukemia (PML) which result in mature blockade of premyelocyte, inhibiting expre-ssion of tumor suppressor gene and PML's apoptosis-promoting function. RA could restart differentiation-regulating network by binding to RAR domain of oncoprotein PML-RARa which induce leukemic cells different-tiation or even apoptosis. The studies performed on oral squamous cell carcinoma revealed that decreased RARβ2 in cancer cells associated with loss of RA sensitivity. The binding of RA with RARy selectively inhibits growth and induces differentiation and apoptosis of neuroblastoma cells. RXRy was deficient in normal tissue of thyroid gland and overexpression in some thyroid cancers. If treated with 9-cis RA, the growth of thyroid cancer cells with RXRy over-expression was inhibited and subjected to apoptosis. In contrast no change could be observed in the thyroid cancer cells that RXRy silenced. In human mammary epithelial cells, CRABP-Ⅱ, which could regulate RA reaction ability, was regulated by AP2 factor. Other compo-nents of RA signaling pathway includes CYP26A1 also influence the effects of this drug. In view of the above, different types of abnormality of RA signaling in tumors, single or combination, makes it necessary to study on different RA sensitivity in MB from entirety of this pathway. So far no report was seen about entire RA signaling pathway on MB cells and tissues. In this respect by studying on MB cells and tissue, we analyze the state of RA signaling pathway components in different sensitivity MB cells and the correlation with RA effects, sensitive or resistant. On this basis, epigenetic mechanism which cause gene silence of RA signaling pathway was investigated for the purpose of providing reasonable base for improvement of MB therapeutic strategy.Materials and Methods:MB and tumor-surrounding noncancerous cerebella were provided by Department of pathology, the 1st affiliated hospital of Dalian Medical University and Shen-Jing hospital of China Medical University. Med-3 MB cell line was kindly provided by the doctors in the Department of neurosurgery, Kobe University School of Medicine and UW228-2 cell line was established and provided by the Department of neurological surgery, University of Washington at Seattle. By the methods of paraffin embedded tissue array immunohistochemistry(IHC), cell culture, immunocytochemistry(ICC), reverse transcription-polymerase chain reac-tion(RT-PCR), Western blotting(WB), mythylation-polymerase chain reaction(MP), DNA sequencing and RNA interference(RNAi), the following experiments were performed:1) The expression of RARaα/β/γ,RXRaα/β/γ, CRABP-Ⅰ/-Ⅱand CYP26A1 was detected in RA sensitive Med-3 cell line and RA resistant UW228-2 cell line. Transcription level change of them after treatment of RA were also studied.2) The correlation between promoter methylation and abnormal expression of CRABP-Ⅰand-Ⅱwas explored.3) Demethylation agent(5-Aza) was used to recover expression of silenced CRABP-Ⅱand then RA effects on growth, differentiation and apoptosis was observed.4) Directed siRNA of CRABP-Ⅱwere applied to transfect RA sensitive Med-3 cells and change of RA sensitivity were determined.5) By combination of paraffin embedded tissue array and IHC, CRABP-Ⅰ/-Ⅱand synaptophysin expression were analyzed. The data were statistically analyzed by Spearman Rank and Bivariate Correlation with SPSS 11.0 software.Results:1. The status of RARα/β/γ,RXRα/β/γ,CRABP-Ⅰ/-Ⅱand CYP26A1 in RA sensitive and resistant MB cell lines and transcription level change of them after treatment of RA.1.1 By HE staining and Flow cytometry analysis, different RA chemosen-sitivity in variant MB cell lines were studied. Neuron-shaped change and appearance of NGN 1, primary differentiating marker, were observed in RA sensitive Med-3 cell. RA treated Med-3 cell exhibited s-Phase arrest.1.2 Statistic analysis of OD value between normal and RA treated Med-3 cell for 48h and 72h is significant(p=0.006 and p=0.000).Oppositely Statistic analysis of OD value between normal and RA treated UW228-2 cell for 24h,48h and 72h,, there was no statistical difference. The difference of viable/nonviable ratio between normal and RA treatd Med-3 cell at time point 48h and 72h was significantly(p=0.000), but that in UW228-2 was insignificantly(p>0.05).1.3 By RT-PCR expression of CRABP-Ⅰ/-Ⅱwas lost in UW228-2, no matter RA treatment or no. Low level transcription of these two genes was observed in RA sensitive Med-3 cell, markedly elevated after RA addition.1.4 Further WB and ICC confirmed that lost of CRABP-Ⅰ/-Ⅱin UW228-2 and expression of them in Med-3 is regulable.2. Promoter hypermethylation induce silence of CRABP-Ⅱand RA resist-ance in UW228-2.2.1 Methylation PCR (MP) results showed that CRABP-Ⅱbut no CRABP-Ⅰ band could be observed at corresponding molecular weight location. Bisulfite sequencing indicated that there were 14 sites methylation of CpG islands.2.2 CRABP-Ⅱbut not CRABP-Ⅰexpression was recovered by 5-Aza, a demythylation agent, in dose-related fashion. Meanwhile no elevation was observed in RARa and RARβtranscription level after demethylation.2.3 RA sensitivity was overturned in resistant UW228-2 after 5-Aza treatment. Cell growth was inhibited. Apparent neuron-like morphology and cell death was induced. Furthermore, NGN1 andβ-Ⅲtubulin, the early neuronal differentiation biomarker, was expressed or in higher. Statistical analysis of MTT assay indicated that the difference between demythelation following with RA treatment groups(one group is single RA treatment after demethylation, the other group is that combined RA and 5-Aza treatment after demethylation.) and control groups was significantly (p<0.05).3. CRABP-ⅡsiRNA reduced RA-sensitivity of Med-3 cells1. Three designed siRNA aimed directly at different sites silenced CRABP-Ⅱat.different level, the highest inhibition ratio was 83%. The transfection efficiency of transfection agent was above 90%.2. CRABP-Ⅱ-RNAi makes the RA sensitive Med-3 cell resistant.2.1 No statistical significance of OD values from MTT assay among the four experimental groups (p>0.05) includes normal culture condition (Normal), transfected with Mock siRNA (Mock) or CRABP-Ⅱspecific siRNA (siRNA) and treated by CRABP-Ⅱspecific siRNA for 48 hours followed by 10μM RA treatment starting at 48 hour time point (siRNA+RA).2.2 CRABP-Ⅱoriented immunocytochemical staining results indicatd that it was markedly decrease after RNAi. H & E staining performed on normally cultured Med-3 cells and the cells treated by siRNA and RA combination showed that no differentiation and apoptosis were observed.2.3 Flow cytometry analysis of Med-3 cells harboring CRABP-ⅡsiRNA without and with 3 day 10μM RA treatment proved that no obvious cell cycle arrest and apoptosis index change.4. Expression of CRABP-Ⅰ/-Ⅱand Synaptophysinin in MB and tumor-surrounding noncancerous cerebella.In 104 MB tumor tissues,58 samples were diagnosed as the classic subtype,40 as large cell subtype(includes anaplastic ones) and 6 as nodular subtype(includes desmoplastic ones). CRABP-Ⅰ/-Ⅱand Synaptophysinin were constitutively expressed in 12 tumor-surrounding noncancerous cere-bella. Expression of them in MB tissues is as follows:4.1 Positive staining of CRABP-Ⅱwas brown in nucleus.45 (45/104,43.3%) of the 104 tumor samples were positive,44 (44/104,42.3%) negative and 15 (15/104,14.4%) partly positive in CRABP-Ⅱexpression.58 samples were diagnosed as the classic subtype,40 as anaplastic/large cell subtype and 6 as nodular subtype, and the above CRABP-Ⅱstaining patterns could be observed among the classic, anaplastic/large cell and nodular subtypes. In 45 CRABP-Ⅱpositive samples,24(24/58,41.4%) of them were classic, 19(19/40,47.5%) were large cell subtype and 2(2/6,33.3%) were nodular ones.4.2 Positive expression of CRABP-Ⅰwas brown particles in cytoplasm. 39(39/104,37.5%) of the 104 MB tissues were positive,5 (5/104,4.8%) partly positive and others (60/104,57.7%) negative in CRABP-Ⅰexpression. In 39 CRABP-Ⅰpositive samples,20(20/58,34.5%) of them were classic, 19(19/40,47.5%) were large cell subtype and no CRABP-Ⅰpositive staining was observed in nodular subtype.4.3 To synaptophysin, brown particles in cytoplasm represents positive staining.74(74/104,71.1%) of the 104 MB samples were positive, 24(24/104,23.1%) negative and 6 (6/104,5.8%) partly positive in synaptophysin expression. In 74 synaptophysin positive samples,38(38/58, 65.5%) of them were classic,33(33/40,82.5%) were large cell subtype and 3(3/6,50%) were nodular subtype.4.4 CRABP-Ⅱand synaptophysin expression was closely correlated (Spear-man rank-correlation, rs=0.317; p=0.001) among the 104 medullobla-stoma tissues, and this correlation was more remarkable in the classic subtype (rs= 0.576; p=0.000). The overlapped CRABP-Ⅰand CRABP-Ⅱsilencing was found in 18/58 (31.0%) of the classic,10/40 (25%) anaplastic/large cell and 3/6 (50%) anaplastic/desmoplastic tumors and, therefore, showed no close correlation of their expression (p>0.05).Conclusion:1. RA chemosensitivity was different in MB cells, Med-3 was sensitive and UW228-2 resistant. 2. Different RA chemosentivity of MB was correlated with RA signaling pathway components especially CRABP-Ⅱexpression or no.3. CpG islands hypermethylation in promoter region of CRABP-Ⅱwas resp-onsible for its silence in MB.4. Expression recovery by demethylation of CRABP-Ⅱcould reverse RA resistant UW228-2 cell to sensitive.5. siRNA of CRABP-Ⅱcould silence its expression and makes the RA sensitive Med-3 cell resistant.6. Regulation experiments indicated that CRABP-Ⅱexpression or no determined RA anti-cancer effects.7. This study reveals for the first time that there were 3 types of CRABP expression, positive, partly positive and negative. The diversiform and heterogeneous CRABP expression was possiblely the key factor to explain markedly difference of RA effect on MB and important reason for primary and secondary resistance.8. These results will provide valuable experimental evidence for reasonable and personalized RA therapy in MB clinic and have apparent significance in translational cancer biology.