Circadian gene cry controls glioblastoma tumorigenesis through modulation of myc expression.

Glioblastoma (GB) is the most frequent malignant brain tumor among adults and currently there is no effective treatment. It is a very aggressive tumor that grows fast and spreads through the brain causing the death of patients in 15 months. GB cells mutate frequently and generate a heterogeneous population of tumoral cells genetically distinct. Thus, the contribution of genes and signaling pathways relevant for GB progression is of great relevance. We use a Drosophila model of GB that reproduces the features of human GB, and describe the upregulation of the circadian gene cry in GB patients and in a Drosophila GB model. We study the contribution of cry to the expansion of GB cells, to the neurodegeneration caused by GB, and to premature death and determine that cry is required for GB progression. Moreover, we analyze the mechanisms that regulate cry expression by the PI3K pathway. Finally, we conclude that cry is necessary and sufficient to regulate myc expression in GB. These results contribute to the understanding of the signals that impulse GB malignancy and lethality and open novel opportunities for the treatment of GB patients. presynaptic protein Cell Myc,.


Introduction
Glioblastoma (GB) is the most common and aggressive type of glioma of all brain tumors, accounting for 57.3% of all gliomas (Ostrom et al. 2020). It is classified as a WHO grade IV diffuse oligodendroglial and astrocytic brain tumor. It is more frequent in the adult population between 75-84 years of age.
Treatment is based on radiotherapy accompanied by chemotherapy with temozolamide (TMZ) after surgical tumor resection. Despite current treatments, the mean survival of the patients is around 15 months (Wirsching and Weller 2016). And it is estimated that only 6.8% of patients survive five years after diagnosis (Ostrom et al. 2020). Understanding the genetic, molecular and cellular bases of gliomagenesis processes is fundamental for the development of effective therapies against these types of tumors. GB is a very heterogeneous type of tumor in terms of histopathology and genetic expression, even within the same tumor (Rich and Bigner 2004). However, there are common mutations in GB affecting different pathways that show mutual exclusivity: the p53 pathway, the Rb pathway, as well as components of the PI3K pathway (Brennan et al. 2013).
Previous studies from our lab used a GB model developed by Read and collaborators in 2009 in Drosophila that recapitulates key aspects of the disease (Read et al. 2009;Portela et al. 2019a;b, 2020;Formica et al. 2021;Jarabo et al. 2021;Vigneswaran et al. 2021) both genetically and phenotypically, such as the proliferation of glial cells, invasion, inappropriate differentiation and the interaction between genetic routes and the components of the signaling pathways. The model is based on the expression of the constitutively active forms of EGFR and dp110 (orthologues of EGFR and PI3K catalytic subunit in Drosophila, respectively) specifically in glial cells using the control of repo driver (Read et al. 2009). The co-activation of these signaling pathways in Drosophila glial cells increases Myc levels, essential for tumor transformation characterized by an increase in the number of glial cells, in the volume of the membrane, a reduction in the number of synapses and a reduction in the survival. Furthermore, this co-activation regulates processes such as progression and entry into the cell cycle and protein synthesis (Read et al. 2009;.
c-myc is one of the oncogenes most amplified in human cancer, including GB.
About 60-80% of human GB cases have elevated Myc levels (Annibali et al. 2014). Myc regulates cell proliferation, transcription, differentiation, apoptosis and cell migration and plays an essential role in the progression of GB as it is the point where EGFR and PI3K pathways converge, as well as being essential for tumor transformation (Read et al. 2009;Annibali et al. 2014). Furthermore, in vitro and in vivo studies have shown that Myc inhibition prevents glioma formation, inhibits cell proliferation and survival, and even induces disease regression (Annibali et al. 2014). These features are conserved in Drosophila (Read et al. 2009).
In recent years, the study of alterations in circadian rhythm genes has emerged in different types of cancer, including GB. Previous reports suggested that circadian rhythm genes have an important role in different aspects of tumor progression. The central clock organizes the oscillations and rhythmicity of the physiological processes by controlling the expression of a high number of ubiquitously expressed genes. Among them, there are genes related to cell proliferation or differentiation, such as cell cycle components (Lahti et al. 2012), proto-oncogenes and tumor suppressors (Kettner et al. 2014). Besides, CRY1 expression is androgen-responsive, CRY1 regulates DNA repair and the G2/M transition and it is associated with poor outcome in prostate cancer and colorectal cancer.
In Drosophila, the molecular mechanisms that govern circadian behaviors are based on transcriptional feedback loops evolutionarily conserved from insects to vertebrates. Simplified, clock (clk) and cycle (cyc) encode for proteins that form heterodimers and bind to the E-box sequences, which are found in promoter regions of circadian genes. Among many other effectors, Clk/Cyc induce the expression of their own repressors, period (per) and timeless (tim), which also dimerize in the cytoplasm. These proteins accumulate during night and interact with Clk/Cyc inhibiting their binding to DNA (Allada and Chung 2009;Peschel and Helfrich-Förster 2011), thus impeding their own expression. In turn, decreases Per/Tim levels and increases Clk/Cyc. In multicellular organisms, although all cells have their own circadian rhythm, there is a so-called "central clock", which is the structure responsible for coordinating circadian behavior throughout the body. In mammals, it is the suprachiasmatic nucleus (SCN), located in the anterior region of the hypothalamus and made up of about 50,000 neurons in humans (Jarabo and Martin 2017). There are other tissues involved in the maintenance of circadian rhythms apart from the central clock, which are called ''peripheral clocks'', which help to synchronize the central clock (Kettner et al. 2014). For example, there are studies in mammals that also give certain glial cells a fundamental role in maintaining circadian rhythms. These cells are called "glial clocks" and they in turn depend on the SCN for their resynchronization (Chi-Castañeda and Ortega 2016). All the neurons that compose the central clock express these genes to develop the oscillations that organize the cycles of the whole organism in absence of environmental cues.
Furthermore, synchronization of the internal clock with light/dark cycles relies on cryptochrome protein (Cry), a blue light photopigment expressed in certain subsets of clock neurons. Cry binds Tim, triggering its degradation when activated by light. Cry is a receptor of near-UV/blue light and a regulator of gene expression that belongs to the group of DNA photolyases. It was suggested that the last universal common ancestor (LUCA) had one or several photolyases, supporting the evolutionary conservation of cryptochrome genes (Vechtomova et al., 2020). In mammals, the molecular circadian clock is composed of Clk/Bmal1 (instead of Clk/Cyc) and Per (Per1, 2, 3)/Cry (Cry1, 2) (instead of Per/Tim) (Jarabo and Martin 2017). However, the mammalian gene that plays the role of Drosophila cry remains unknown. Interestingly, Drosophila Cry also acts as a transcriptional repressor and binds to Per when expressed in peripheral clocks (Collins et al. 2006).
Regarding GB, studies in patients with primary gliomas found an association between a specific per1 variant with overall glioma risk. Several circadian genes including cry1, exhibited differential expression in GB samples compared to control brains as described in the literature (Madden et al. 2014;Wang et al. 2021) and human databases (https://www.proteinatlas.org; https://cancer.sanger.ac.uk/). Besides, clk expression was found significantly enhanced in high grade gliomas and correlated to tumor progression (Chen et al. 2013). High per1 and per2 expression increases the efficacy of radiotherapy also in GB cells (Zhanfeng et al. 2015). Different studies show a relationship between Cry and Myc. C-Myc levels have been found to decrease in mice in cry1/cry2 null mutants (Liu et al. 2020).
Besides, Cry expression is induced by Myc in GB cells in culture (Altman et al. 2015).
Taking into account the deregulation in the expression of circadian rhythm genes in tumor tissues in GB, and the pre-established relationship between cry and myc, which is a key player in GB, here we validate the role of Cry in the tumorigenesis and progression of GB.

Expression of cry in glioblastoma
To determine if cry expression was affected in glioma samples, we extracted RNA from heads of 7 days old adult control flies, expressing LacZ in glial cells (repo-Gal4), or PI3K+EGFR constitutively active forms to generate a glioma.
Quantitative RT-PCR results indicate that cry mRNA level is 50 times higher in glioma samples as compared to controls ( Figure 1A). However, to determine if cry upregulation occurs in glial cells we used a specific fluorescent GFP-cry reporter line that generates a GFP tagged form of Cry, and visualized larvae brains in confocal microscopy. The images show that GFP signal in glial cells is higher in glioma samples than in controls ( Figure 1B, C), and this signal is restored to control levels upon cry RNAi in glial cells ( Figure 1D). We quantified the GFP signal that overlaps with glial membrane (red, mRFP, red in Figure   1B´, C´, D´ and E´) and the quantifications indicate that GFP-cry signals is higher in glioma, and this increase is prevented upon cry knockdown in glioma cells ( Figure 1F).
We analyzed human mRNA expression databases for Glioblastoma multiforme (http://gliovis.bioinfo.cnio.es/). The results indicate that cry in GB patients (CRY) is transcriptionally upregulated ( Figure 1G). Besides, CRY upregulation correlates with worse prognosis ( Figure 1H). All together, these results indicate that cry is transcriptionally upregulated in GB cells in Drosophila and patients, and suggest a role in GB malignancy and aggressiveness.

Cry mediates GB progression and neurodegeneration
To determine the contribution of cry to GB progression and the consequences, we used a previously validated protocol (Portela et al. 2019bJarabo et al. 2021) to quantify tumor growth and the associated neurodegeneration. We stained control adult brains, and compared with GB, GB and cry RNAi, and wt brains expressing cry RNAi in glial cells. We used a specific antibody against repo to visualize the nuclei of all glial cells, and quantified the fluorescent confocal images to count the number of glial cells ( expression is sufficient to increase GFP-cry signal, but nor EGFR or Myc overexpression. These results suggest that PI3K controls cry transcription in a Myc independent signaling pathway, and EGFR does not participate in cry regulation in glial cells.

Cry regulates Myc expression in glial cells
Next, to determine the epistatic relation between Cry and Myc, we analyzed  Figure 5L). Nevertheless, none of these genetic modifications is sufficient to expand glial membrane volume ( Figure 5M). These results suggest that cry or Myc is sufficient to trigger glial cell number increase in adult brains, but not to expand the volume of glial membrane network.

Discussion
Different studies have established a relation between alterations in circadian rhythm genes and cancer. Specifically, one of the genes associated with different types of cancer is cry (Shafi et al. 2021;Yang et al. 2021;Mampay et al. 2021). So this study aims to investigate the role of cry in a Drosophila GB model.
We have described the upregulation of cry1 in human GB samples, and in a well studied Drosophila model of GB. This model is based on the two most frequent mutated pathways in GB, PI3K and EGFR, which converge in Myc as a convergence point. This pathway is of great relevance to promote GB cells expansion, GB progression and in consequence, the deterioration of neighboring neurons and a premature death.
The results indicate that cry upregulation in GB cells depends on PI3K expression, and is required for GB cells increase and synapse loss. Cry is sufficient to increase the number of glial cells. However, cry expression is expendable for normal wt glial growth during development, which makes it a potential target for GB treatment.
In addition, we show that Cry is necessary to induce myc expression in GB cells, and cry expression is sufficient to induce myc upregulation. This agrees with in vitro studies that revealed an increase in Myc levels as a result of cry upregulation (Altman et al. 2015). Therefore, we propose that cry is part of the PI3K-Myc signaling pathway in GB, where cry upregulation would be associated with glial proliferation. However, PI3K is a highly promiscuous enzyme that participates in numerous signaling pathways, and the results suggest that Cry contribution is restricted to the malignant features of GB dependent on myc such as GB cell number increase and neurodegeneration. However, cry expression is independent of expansion of glial membrane characteristic of GB progression.
In addition, cry expression in glial cells partially reduces life span, but is less aggressive than GB. This result suggests that Cry plays a central role in GB and it is required for GB formation, and cry mutations might be responsible for many features of GB. The human gene expression databases indicate that cry1 expression levels correlate negatively with life span, and it is associated with a poor prognosis. In consequence, these results suggest that further studies on the contribution of Cry1 to human GB progression could lead to novel strategies to treat GB patients.
The studies of other groups describe the effect of haloperidol on cry1 expression in GB cells. But these results obtained in cell culture suggest that the doses required to treat patients might be toxic, in consequence specific delivery strategies combined with haloperidol are worth of study. We have observed significant effects of cry knockdown in normal glial cells, in line with Bolukbaso and cols that recently described the extension of lifespan by foxo upregulation in glial cells (Bolukbasi et al. 2021). We have observed an effect of cry RNAi expression in the number of glial cells ( Figure 5L). Given that cry and foxo respond to PI3K pathway, it is tempting to speculate that cry expression is relevant for life span extension by PI3K pathway, and associated behaviours such as diet restriction.
In The glioma-inducing line contains the UAS-dEGFR λ , UAS-dp110 CAAX transgenes that encodes for the constitutively active forms of the human orthologs PI3K and EGFR, respectively (Read et al. 2009). Repo-Gal4 line drives the Gal4 expression to glial cells and precursors (Lee and Jones 2005;Casas-Tintó et al. 2017) combined with the UAS-dEGFR λ , UAS-dp110 CAAX line allow us to generate a glioma thanks to the Gal4 system (Brand and Perrimon 1993). Elav-LexA line drives the expression to neurons, allowing us to manipulate neurons in a glioma combining LexA and Gal4 expression systems (Lai and Lee 2006 Quantification Fluorescent reporter-relative cry signals within brains were determined from images taken at the same confocal settings avoiding saturation. For the analysis of co-localization rates, "co-localization" tool from LAS AF Lite software (Leica) was used taking the co-localization rate data for the statistics analyzing the co-localization between green signal (both cases) and signal coming from glial tissue from three slices per brain in similar positions of the z axis.
Glial network was marked by a UAS-myristoylated-RFP reporter specifically expressed under the control of repo-Gal4. The total volume was quantified using Imaris surface tool (Imaris 6.3.1 software). Glial nuclei were marked by staining with the anti-Repo (DSHB). The number of Repo+ cells and number of synapses (anti-nc82; DSHB) were quantified by using the spots tool in Imaris 6.3.1 software. We selected a minimum size and threshold for the spot in the control samples of each experiment: 0.5 μm for active zones and 2 μm for glial cell nuclei. Myc glial signal was quantified using Imaris surface tool (Imaris 6.3.1 software) creating a mask for the glial nuclei signal and selecting exclusively the myc signal corresponding to glial nuclei. Then we applied the same conditions to the analysis of the corresponding experimental sample.

Statistics
The results were analyzed using the GraphPad Prism 5 software (www.graphpad.com). Quantitative parameters were divided into parametric and nonparametric using the D'Agostino and Pearson omnibus normality test, and the variances were analyzed with F test. t test and ANOVA test with Bonferroni's post hoc were used in parametric parameters, using Welch's correction when necessary. The survival assays were analyzed with Mantel-Cox test. The P limit value for rejecting the null hypothesis and considering the differences between cases as statistically significant was P < 0.05 (*). Other Pvalues are indicated as ** when P < 0.01 and *** when P < 0.001.