Final Report Summary - CANCERTIME (The biological clock and cancer)
The endogenous circadian clock generates and regulates many daily physiological and behavioral rhythms with a period of 24 hours at the cell, tissue and organism level. Circadian clocks ensure that the processes they regulate recur in a coordinated manner at the right time of the day optimizing the functioning of the organism. Disruption of the circadian rhythm (e.g. shift work and frequent travel across time zones) has been reported to increase the risk of cancer in several epidemiologic studies. In addition, it is known that every drug has an optimal time of day or night when it is less toxic and most effective, a concept known as chronotherapy. This suggests a role of circadian rhythms in tumorigenesis but a mechanism that links the circadian clock and cancer is still lacking. However, there is controversy in the literature on whether tumors are capable of retaining functional clocks. In order to successfully apply chronotherapy in cancer treatment, there are two fundamental questions to tackle: a) Do tumors maintain their circadian clock functional? b) If so, is it synchronized to normal tissues? The first goal of this fellowship is to perform a comprehensive analysis of molecular clocks in tumors in comparison to normal tissues. This information is critical for future translational applications directed for the improvement of therapeutic indexes of existing regimens. Furthermore, disruption of circadian rhythms leads to an increase in tumor development and it has been suggested that circadian proteins play a role in cell proliferation and apoptosis. Similarly to p53 protein, clock gene expression is also influenced by DNA-damage and clock proteins regulate cell-cycle arrest, DNA repair and apoptosis. There is evidence that DNA-damage induces both p53 and circadian gene-dependent defence mechanisms against cancer. The second goal is to study potential links between the circadian clock and the p53 function and how both processes may interact to block tumor development.
The first objective of Cancertime is to characterize the circadian clock oscillatory function in spontaneous tumors and compare them to normal tissues. Per2-Luciferase knock-in mice (PER2::LUC) crossed to tumor-prone p53-/- mice were used in this study. These mice are characterized by having a high rate of spontaneous tumorigenesis. Importantly, the presence of circadian-driven luciferase reporter allows for simultaneous accessing of daily variations in clock function both in normal tissues and in spontaneously developed tumors. One group of non tumor bearing PER2::LUC/p53-/- mice and one group of PER2::LUC/ p53-/- mice with spontaneously developed solid sarcoma tumors were used to monitor in vivo the circadian profile of mPer2-driven luciferase in normal (liver and salivary glands) and sarcoma tissues. Xenogen IVIS Imaging System was used to measure luciferase expression at six different timepoints every four hours throughout one day. The circadian profiles of mPer2-driven luciferase signal measured in vivo in sarcoma bearing mice in normal and sarcoma tissues show rhythmic circadian profiles (Figure 1). Furthermore, Dr M. Comas Soberats monitored the circadian profile of mPer2-driven luciferase ex-vivo in explant cultures of normal tissues (liver, spleen and lung) as well as in sarcoma. This experiment was performed using a bioluminescence system (Lumicycle, Actimetrics, Wilmette, IL, USA) allowing for several days of real-time luciferase recording. In tumor bearing mice, the period measured over the first three cycles of the oscillation is close to 24 hours, also for sarcoma and thymic lymphoma tissue. Sarcoma and lymphoma peak phase occurs, similarly to liver, lung and spleen, in the dark phase (Figure 2). The results obtained show that tumors are not only oscillating in the context of the organism where they receive hormonal and metabolic signals but explants tissues continue oscillating demonstrating that tumors have functional clocks capable of timing all their functions.
The next objective asks the question of whether transplanted tumors are capable of maintaining circadian synchronization. First, Dr M. Comas Soberats characterized in vitro the circadian profiles of clock genes mRNA expression profiles (Bmal1, Per1, Per2, Rev-erba) in human and mouse breast cancer cell lines (MCF-10A, MCF-7, MDA-MB-231 and 4T1). After synchronization with serum shock, MCF-7, MDA-MB-231 and 4T1 tumor cell lines have a significant circadian profile of expression for all genes but with lower amplitude of oscillation compared to MCF-10A cells, which indicates a less robust clock. Interestingly, these cell lines do not display significant circadian oscillation profiles of clock controlled genes (Ccnd1 and wee1) (Figure 3). Next, MCF-7 and 4T1 cell lines were transplanted into mice as xenografts and allografts respectively using the intraductal transplantation method to study whether the cells in the the in vivo context synchronize to the host organism’s clock. The tumors were then collected every four hours throughout one day. MCF-7 xenografts show a significant circadian profile of clock genes and clock controlled genes mRNA expression. The oscillation of the circadian profile has higher amplitude than those obtained for MCF-7 cells synchronized in vitro. On the contrary, 4T1 intraductal allografts resulted in lower amplitude of the oscillation profiles of mRNA expression of clock genes. In addition, Dr M. Comas Soberats performed the in vivo characterization of circadian profiles of mRNA expression in the MMTV-PyMT model. The circadian profiles obtained for Bmal1, Per1, Per2 and Rev-erba show significant circadian profiles that peak at the same time as the ones obtained for mammary glands (Figure 4).
Once established that tumors have functional circadian clocks, Dr. Comas Soberats tested the prediction of whether these tumors with functional clocks could be synchronized by food availability uncoupling the synchronization from the master clock in the SCN as previously shown for liver by other authors. Dr Comas Soberats inoculated into SCID mice human fibrosarcoma HT1080 cells stably expressing luciferase under control of Bmal1 promoter in the left flank and Per2 promoter in the right flank. One control group had food availability ad libitum whereas the second group was maintained on restricted feeding schedule (food available at lights-on time only) starting from 2 weeks prior to tumor inoculation. The results showed that animals that were maintained on restricted feeding schedule demonstrated a shift in the phase of both Bmal1 and Per2 genes expression to synchronize to food availability and desynchronize from the SCN. This result was further confirmed by RT-PCR on tumor tissue extracts collected. These data clearly demonstrate the ability of the tumor clock to be manipulated by external signals (in this case feeding schedule) (Figure 5).
In conclusion, tumors studied in objective 2 (spontaneous and transplanted) do retain a functional clock with circadian profiles of mRNA expression of clock genes with lower amplitudes than the normal tissue. We also provide evidence that similar to liver, tumors can be synchronized by food availability independent of the central pacemaker in the SCN. These results are in line with the ones obtained in objective 1 reinforcing the idea of taking advantage of chronotherapy to treat cancer. These results suggest that tumors’ sensitivity to cancer treatment can be modulated through the components of the circadian clock.
This objective is completed. In the return phase, Dr M. Comas Soberats has focused on investigating the putative cross-talk between p53 pathway and the circadian clock in tumors. There is good evidence that circadian genes affect apoptosis and cell cycle. Both processes are largely known to be regulated by p53 protein. However, no clear relationship between p53 and circadian system in relation with cancer has been found so far. Finding the link between the circadian clock and the tumor suppressor protein p53 is fundamental to study the regulation of cell cycle, apoptosis and the circadian system with tumor development. This research has the potential to support the development of translational approaches taking chronotherapy into account.
Prior to Dr. Comas Soberats arrival, we carried out a chromatin immune-precipitation (ChIP)-seq and microarray experiment in MCF-7 cells upon pharmacological activation of p53. Among the high confidence p53-binding sites is the Per1 promoter (Figure 6A). These results have been confirmed by Dr. Comas Soberats by ChIP-qPCR in MCF-7 cells and HCT116 cells (Figure 6B). Furthermore, in agreement with this result, we have found by microarray analysis that Per1 mRNA expression significantly increases upon p53 pharmacological activation on MCF-7 cells (Table 1). Dr. Comas Soberats confirmed this microarray results by RT-PCR and luciferase assay in MCF-7, isogenic p53 positive on and p53 null HCT116 cells (on the text HCT116 and HCT116-/- respectively) (Figure 7A). Per1 mRNA expression is activated after pharmacological activation of p53 with 1M RITA, 10M Nutlin-3a and 100M 5-fluorouracil (5-FU) in HCT116 cells (Figure 7B). The p53 null cells HCT116-/- do not show an increase in Per1 mRNA levels after treatment with the same drugs indicating that Per1 up-regulation is p53 dependent. Over-expression of p53 protein also leads to an increase of Per1 mRNA levels (Figure 7C). Other core clock proteins such as Per2 and Cry1 also show an expression increase whereas Bmal1 levels decrease (Figure 7D). Importantly, synchronization of the circadian clock of cells by 2h serum shock shows that up-regulation or down-regulation of these clock genes strongly depends on the clock phase at which treatment starts (Figure 8). Thus, when RITA (small molecule that activates p53) is added 16 and 20h after synchronization Per1, Per2 and Cry1 show significant increase in RNA levels compared to non-treated synchronized cells. However if RITA is added 4, 8 or 12 after synchronization no significant effect is observed in mRNA levels of these genes. Conversely Bmal1 shows an increase when RITA is added 4h after synchronization but a significant decrease at all other times tested (8, 12, 16, 20 and 24 h after synchronization). This proves a strong dependence of circadian phase of up-regulation of clock genes expression by RITA on these cells. In addition, Dr. Comas Soberats looked at p53 target and regulator genes and found that many of them oscillate in a circadian fashion and that RITA up-regulates or down-regulates their expression depending on the status of the circadian clock. Thus, key apoptotic p53 target genes or p53 regulators such as Puma, Bax, NoxA, Stat3, Pias3, Mdmx, Wip1 are highly up-regulated after RITA treatment 16h after synchronization and either no effect or down-regulation is found if RITA treatment starts 4h after synchronization. These two phases respectively correspond with peak and trough of Per1 oscillation. On the contrary, other genes such as NfyA, Mdm2 or Ccnd1 peak at 4h and have a trough at 16h after synchronization. When Dr. Comas Soberats knocks down Per1 no up-regulation after RITA treatment is observed for Puma and Bax, Stat3, Pias3, Ccnd1 and no down-regulation is observed for Mdm4 (Figure 9). Altogether these results indicate that p53 activated by RITA requires Per1 to carry out its function and that Per1 might be also implicated in p53 regulation itself. Dr. Comas Soberats has shown by flow cytometry that knocking down Per1 gene leads to more cell death. Dr. Comas Soberats finds that if RITA treatment starts 4h after synchronization of the circadian cycle there is less cell death compared to 16h after synchronization coinciding with trough and peak of Per1 (Figure 10). Overall, these results indicate that circadian rhythm status plays a role in apoptosis and that there is a crosstalk between p53 and circadian cycle pathways. The relevance of this result resides in the perspective of taking into account the circadian system to design the treatment of cancer patients based on drugs that activate p53 protein to induce apoptosis.