Novel genetic model of pediatric Diffuse Intrinsic Pontine Glioma in Drosophila melanogaster

Diffuse Intrinsic Pontine Glioma (DIPG) is a lethal pediatric type of brain tumor that grows in the bm and originated from glial cells. Its location and infiltrative nature impede surgical resection and make the treatment difficult and low effective. In consequence, affected children have a short life expectancy of 12 months. The most frequent mutation is a substitution of lysine to methionine at residue 27 of histone H3 (H3K27M). Secondary mutations in additional genes, including Myc, are required for the malignancy of glial cells. The lack of studies and tumor aggressiveness make it necessary to generate new experimental models that reproduce the fundamental aspects of the disease and allow to expand the knowledge about DIPG. Drosophila melanogaster presents advantages as an experimental model and stands out for its genetic tools, easy handling, and great genetic and cellular homology with humans. Drosophila has contributed to the investigation of different diseases, including glioblastoma (GB) and neurodegenerative diseases as Alzheimeŕs or Parkinsońs. Here we present a new genetic model of DIPG generated in Drosophila melanogaster. It is based on the overexpression of H3K27 and Myc in glial cells that produce an increase in the number of glial cells in the ventral nerve cord and the expansion of glial membranes in early developmental stages. However, this novel DIPG model does not produce tumoral features in adult brains, in line with the pediatric nature of this disease. We have evaluated the activation of different signaling pathways active in other glial tumors, in this model of DIPG. The results show that, unlike GB, JNK pathway is not upregulated in DIPG, and it is not determinant for the progression of DIPG. Besides, glial cells in the DIPG model accumulate MMP1 and MMP2 and increase the accumulation of Liprin-γ, previously associated to the formation of synaptic structures in GB cells. The results show that DIPG is a unique entity that differs from other high-grade gliomas such as GB and will require of a different therapeutic approach.


INTRODUCTION
Tumors or neoplasms are abnormal masses of tissue that grow uncontrollably, excessively, autonomously, and irreversibly. According to its clinical behavior we find benign tumors, which do not generate secondary growth, and malignant tumors, of an infiltrative nature and with the capacity to generate secondary implants or metastasis.
Tumors of the Central Nervous System (CNS) represent the seventh neoplasia in frequency within adult population (Louis et al. 2021). They are those that develop in the brain, spinal cord or meninges. Based on the histological characteristics and molecular parameters, the WHO differentiates four grades of CNS tumors according to their aggressiveness, grade I is the less aggressive and IV the most aggressive.
Gliomas are the most common tumors of the CNS, originating from neoplastic glial cells. According to the WHO classification we find 4 types of gliomas: diffuse adulttype gliomas, pediatric-type diffuse low-grade gliomas, pediatric-type diffuse highgrade gliomas, and circumscribed astrocytic gliomas (Louis et al. 2021). In this classification the authors indicate the clinical and molecular differences between diffuse gliomas present mainly in adults ("adult type") and those present mainly in children ("pediatric type"), diffuse low-grade gliomas of pediatric type and diffuse high-grade gliomas of the pediatric type.

High Grade Supratentorial Gliomas
High-grade gliomas (HGG, Supratentorial High-grade Gliomas) represent a spectrum of diseases, with histopathological features shared between childhood and adult tumors (Louis et al. 2021), as well as mutations in the same canonical cancer pathways such as the receptor tyrosine kinase (RTK)/RAS/PI3K, the TP53 pathway, and the RB pathway, although the effectors most affected by the mutation vary, between childhood and adult tumors. Even though the effectors commonly affected by mutations differ between pediatric and adult tumors (Sturm et al. 2014).
Recent molecular profiling data demonstrate that childhood tumors are biologically distinct from adults and suggest that there are unique molecular processes that drive tumorigenesis in the developing brain, which vary by age and anatomical location.
These distinctions reflect the need for very different therapeutic approaches to effectively counter underlying genetic mutations (Jones & Baker 2014).

Diffuse midline glioma
Diffuse intrinsic pontine glioma (DIPG) or diffuse midline glioma has been classified as a grade IV entity within the Pediatric-type HGG (Louis et al. 2021). It is a pediatric brain stem glioma that originates from the ventral pons, accounts for 75-80% of brainstem tumors in children, it has a peak incidence in middle childhood and a life expectancy of less than 1 year (Lapin et al. 2017;Wu et al. 2014).
Diagnosis is based on the diffuse infiltration of the midline crossing, typical histological features, and the presence of alterations in residue 27 of Histone H3 (H3K27) (Louis et al. 2016(Louis et al. , 2021. DIPG cannot be removed surgically due to its location and the infiltrative nature of the disease. Palliative radiotherapy is the only therapy, although it only provides temporary improvements on neurological and radiological function (Lapin et al. 2017;Wu et al. 2014). And despite the trials, DIPG are still considered fatal pediatric tumors without cure (Lapin et al. 2017).

Genetic basis
The acquisition of samples of primary tumors has facilitated the elaboration of genomic profiles and advances in the knowledge of key oncogenic factors. This has favored the identification of the genes responsible for the appearance and progression of this type of tumor.
Proteomic and transcriptomic analysis of DIPG suggests that it is a unique type of glioma that shares biological similarities with HGGs, such as glioblastoma multiforme (GB) (Jones & Baker 2014).
Likewise, there is a great intra-and intertumoral heterogeneity within the markers that define the disease, making it difficult to develop effective therapeutic strategies (Hoffman et al. 2016). Analysis by sequencing of DIPG samples showed mutations, including single nucleotide variations, insertions and deletions, and structural variations.

Histone H3
The most relevant genetic characteristic of this tumor type is the presence of mutations in histone H3, present in 80% of DIPG cases (Lapin et al. 2017 mutations are related to a median overall survival of 9 months and less response to radiotherapy. In both cases, DIPGs with mutations in H3 are associated with poorer outcomes compared with tumors with wild type histone H3 (Lapin et al. 2017).
The specific mutation in histone H3 is caused by the substitution of a lysine to methionine at the residue 27 (H3K27M). H3K27 is a key center for transcriptional regulation: acetylation (H3K27ac) is associated with active enhancers, whereas trimethylation (H3K27me3) represses transcription. This methylation is carried out by the Polycomb repressive complex 2 (PRC2) (Nagaraja et al. 2019). H3K27M has a greater affinity for PRC2, inhibiting it and therefore reducing H3K27me3. However, H3K27M also causes an increase in H3K27me3 in specific tissues, suggesting that both expression and repression of abnormal genes is important (Pajovic et al. 2020).
The tumorigenesis-initiating role of oncohistone in DIPG relies on remodeling of the transcriptome: H3K27M induces epigenetic changes that activate multiple members of the RAS/MAPK and MYC cascade, as well as their downstream transcriptional targets.
However, the activation during early development of the epigenetic pathway of H3K27M is not sufficient to produce tumorigenesis in DIPG, and it is later reinforced by pathwayactivating gene mutations (Pajovic et al. 2020). This evidence explains the coexistence of these mutations in human DIPG.

Myc
C-myc is one of the most amplified oncogenes in human cancer and plays a fundamental role in tumor transformation. This gene encodes the protein MYC, a transcription factor that regulates cell proliferation, differentiation, apoptosis, and cell migration. Rising levels of MYC drive tumor initiation, progression, and recurrence, and are necessary for the tumor maintenance (Hutter et al. 2017).
One of the transcriptional features of H3K27M tumors is the epigenetic activation of MYC. Loss of H3k27me3 induces overexpression of MYC target genes in DIPG (Pajovic et al. 2020). MYC amplifications and overexpression occur in 20% of the cases. MYC alterations are present in DIPG and act as genome-wide transcriptional regulators (Lapin et al. 2017).

Glioblastoma
The most common type of glioma is glioblastoma multiforme (GB), classified by the WHO as a astrocytic and diffuse oligodendroglial grade IV brain tumor. It is the most aggressive and lethal brain tumor, with an incidence of three per 100,000/year. The median survival of patients with GB is 12-15 months, with less than 5% chance of survival after 5 years (Gallego 2015;Louis et al. 2016;McGuire 2016;Rogers et al. 2017). Temozolomide (TMZ) is the only treatment for GB, however, recent studies restrict its use in patients with GB based on the methylation status of methylguanine DNA methyltransferase (MGMT) (Wick et al. 2018). Besides, the Genetic and molecular heterogeneity complicates the diagnosis and treatment of these tumors.
Recent studies have revealed that GB cells extend membrane ultra-long tubes that interconnect tumor cells, known as tumor microtubes (TM) that mediate cell-cell communication and probably contribute to resistance to treatment with radiotherapy, chemotherapy, and surgery . TMs are actin-based filopodia that infiltrate the brain and reach long distances within the brain (Osswald et al. 2015). These TM are associated with a worse prognosis in human gliomas and contribute to the invasion and proliferation, causing effective colonization of the brain by GB cells.

Genetic basis
The most common genetic lesions in GB patients include mutations in the epidermal growth factor receptor (EGFR), loss of PTEN (PI3K antagonist) and mutation of catalytic subunit of PI3K (Furnari et al. 2007;Louis et al. 2016;Wirsching et al. 2016). Furthermore, it is common to find constitutively active Akt, an important effector of PI3K.
In Drosophila, the combination of EGFR and PI3K constitutively active mutations causes a glioma-like condition that reproduces the features of human gliomas, including glial expansion, invasion of the brain, neuronal dysfunction, loss of survival. This coactivation regulates processes such as progression and entry into the cell cycle and protein synthesis (Brand & Perrimon, 1993;Read et al., 2009).
So, in this model, MYC plays a central role as it is the point where both routes converge and is essential for tumor transformation.

Tumor schedule
Among the signaling pathways involved in the progression of GB, the canonical WNT pathway stands out, which is activated after the ligand "Wingless-related integration site" (WNT) binds to the family of Low-density lipoprotein (LRP) or Frizzled (FZD) receptors on the plasma membrane, promoting the expression of cell proliferation genes such as cyclin D1 and Myc (He et al. 1998;Shtutman et al. 1999 Another signaling pathway that has been associated with GB proliferation is the cJun-Nterminal Kinase pathway (JNK) and is currently a drug target for GB (Matsuda et al. 2012). The JNK pathway includes a mitogen-activated protein kinase (MAPK), which belongs to the group of protein kinases stress-activated kinases (SAPKs), a group of kinases that can be activated by any stimulus internal or external that causes cellular stress. The MAPK cascade triggers dual phosphorylation of cytosolic JNK and initiates the phosphorylation of cytoplasmic and nuclear proteins (Chang & Karin 2001), including cytoskeletal and mitochondrial proteins, nuclear transcription factors, protein membrane or nuclear hormone receptors (Bogoyevitch & Kobe 2006). It presents a high homology from Drosophila to mammals (Mark & Richardson 2020). In mammals the pathway involves four kinases and mitogens or cytokines that induce MAP3K activation. In Drosophila, JNK signaling is initiated by the interaction of the Eiger ligand (Egr), the only member of the TNF ligand superfamily (Igaki et al. 2002;Moreno et al. 2002), with the receptors TNF (TNFR) Grindewal (Grnd), or Wengen (Wng) (Igaki et al. 2009). The ligand-receptor interaction initiates a cascade of phosphorylations (Mark & Richardson 2020). A dual role for the JNK pathway in cell death and survival has been proposed, depending of cell type and context (Mark & Richardson 2020). This double role is especially relevant for CNS pathologies where signals associated with cellular stress increase, such as neurodegeneration and tumorigenesis (Musi et al. 2020;Portela et al. 2019).
In addition, the JNK pathway is the main regulator of matrix metalloproteases (MMPs) expression and cell motility in several organisms and tissues, including tumors such as GB . MMPs are a family of endopeptidases capable of degrading the extracellular matrix (ECM). The members of the MMP family include the "classical" MMPs, the membrane bound MMPs (MT MMP), ADAMs (a disintegrin and metalloproteinase; adamlysins), and ADAMTS (a disintegrin and metalloproteinase with thrombospondin motif). In humans, there are more than 20 members in the MMP family and ADAMTS, including collagenases, gelatinases, stromelysins, some elastases, and aggrecanases (Malemud 2006 Cancer cells produce MMPs that facilitate tumor progression and invasiveness, and upregulation of MMPs in GB is associated with the diffuse growth and has been proposed to play a role in cell migration and infiltration in GB (Nakada et al., 2003).
Specifically, among the 23 MMPs present in humans, MMP9, MMP2, and MMP14 are directly involved in the growth and invasion of GB cells (Munaut et al. 2003). Previous studies in Drosophila have shown that Egr is expressed in non-tumor brain tissue but accumulates in tumoral cells and activates the JNK pathway.

Consequently, GB cells produce MMPs that facilitate infiltration of TM and GB
progression (Jarabo et al. 2021;Portela et al. 2019Portela et al. , 2020. Erg or Grnd deletion experiments in Drosophila show that JNK inhibition rescues tumor proliferation and invasiveness (Portela et al. 2020). In vitro and in vivo experiments showed that JNK inhibitors SP600125 and AS602801 affected GB self-renewal and potential for tumor initiation (Matsuda et al. 2012;Okada et al. 2016), although further studies are required to understand the contribution of JNK pathway and use it as an anticancer target.
In recent years the central role of TM in GB biology has emerged as a fundamental mechanism for the progression of GB, becoming an attractive field of study for possible GB treatments. Furthermore, the JNK pathway is a drug target for GB (Matsuda et al., 2012).

Drosophila melanogaster as a disease model
DIPG study has mainly used in vitro cell models and tumor xenografts in the brain of mice (Lapin et al. 2017;Wu et al. 2014 In this report we describe the generation of a new DIPG genetic model in Drosophila that recapitulates features of human disease. We have validated this model and used it to determine the molecular mechanisms involved in DIPG progression. In particular, we analyze the expression and contribution of JNK pathway, MMPs and synaptic genes to DIPG progression.

UAS/GAL4 expression binary system
We used the binary expression system UAS/GAL4 in Drosophila melanogaster to induce gene expression or gene knockdown. It is based on the GAL4 transcriptional activator from Saccharomyces cerevisiae that allows directing its expression to a specific tissue using specific promoters, and the activating sequence UAS (Upstream Activating Sequence) that can be fused with a gene of interest. By crossing two parental lines, one carrying the gene of interest fused to the UAS sequence, with another parental line containing the gene encoding GAL4 under the control of a tissue-specific promoter, the offspring expresses the GAL4 activator that upon binding to the UAS sequences, activates the expression of the gene o genes of interest in a specific tissue (Brand & Perrimon, 1993).
In addition, we incorporated a thermosensitive form of the repressor protein Gal80 (Gal80 TS ). It provides temporal control of the expression upon temperature shift. Gal80 TS is active at 17ºC preventing the binding of GAL4 to the UAS sequence.
While, at a temperature of 29ºC the GAL80 protein is inactive, so GAL4 binds to the UAS sequence and promotes the expression of the gene of interest (SE McGuire et al., 2003).
The repo (reverse polarity protein) promoter sequence was used to direct the expression of Gal4 to Drosophila glial cells (Xiong et al. 1994).

Larval brain dissection
To dissect the larvae, we added phosphate buffered saline (PBS) to a dissection dish.
Then we hold the larva with dissecting forceps and cut the underside of the larva. After that, the head of the larva is held with tweezers and, in turn, the other forceps are inserted through the opening of the section. Next, with the tweezers that hold the head, this is pushed until the internal wall and viscera are exposed to the exterior. Finally, the viscera of the larva are removed, the brain is located and separated from the rest of the body.

Adult brain dissection
First, the flies were anesthetized with CO2 and immobilized on a dissection plate with tungsten pins. A drop of 1X PBS is added and proceed to the dissection: first we remove the proboscis and the cuticle from the head. Once the brain is exposed, the tracheal system is cleaned, and the remains of optical pigment are removed. Finally, the brain is separated from the rest of the body.

Immunostaining
Fixation with 4% formaldehyde (FA) at room temperature for 20 minutes.

Image acquisition and analysis
We used a Leica SP5 confocal microscope (Leica Microsystems) to acquire the images. The images were taken in a 20X and 63X immersion objectives, each 1.5μm in the Z axis. The resolution used was 1024 x 1024 pixels. The lasers used were UV 405 nm, Argon 488 nm, DPSS 561 nm and HeNe 633 nm.
The IMARIS program allows to quantify the amount of fluorescent signal from a 3D image taken using confocal microscopy, and to quantify the number of cells that contain a marker, or the volume occupied by the fluorescent signal of said image.
The Image J software allows you to quantify the pixel intensity of the fluorescent signal in an image 3D taken by confocal microscopy and to quantify the amount of MMP present in each region. We analyzed 6 different brains of each genotype.
We calculated the average of the pixel intensity values and the dispersion curves.

Survival and viability assay
We did survival assay with adult males selected from the four genotypes of interest

Statistical analysis
We used GraphPad Prism program for the statistical analysis. We used t-test statistical analysis for the samples in pairs with a normal distribution and the Mann-Whitney test for samples with a non-normal distribution. In the cases in which more than two genotypes were analyzed, we used ANOVA statistical analysis followed by multiple In all cases, the minimum value chosen to consider a difference as statistically significant was p<0.05.

Co-expression of H3K27 and MYC in glial cells increases the number of glial cells and the glial membrane volume in larval stage.
To The statistical analysis of the results obtained by 3D reconstruction of the confocal microscopy images showed significant differences between control and DIPG samples in the number of glial cells and in the total volume of glial membrane ( Figure 1C, D).
These results indicate that the expression in glial cells of H3K27+MYC combination increases the number of glial cells and the volume of the glial membrane in the larval stages of Drosophila.

Co-expression of H3K27and MYC in glial cells during adult stages does not affect the number of glial cells or glial membrane volume.
DIPG is classified as a pediatric-type glioma, with the highest incidence in the middle childhood, the validation of the experimental model requires that it discriminates the development of the tumor based on the age of the individual, and therefore, that reproduces the characteristics of this pediatric glioma in Drosophila.
To generate DIPG cells in adult stages, we took advantage of the Gal80 TS repression system (see Materials and Methods). The crosses and selected offspring were kept at 17ºC (system UAS/GAL4 inactive) until adult stages. Then, once the flies reach the adult mature stages, we maintained the flies of the progeny 7 days at 29ºC (active UAS/GAL4 system).
We The quantification and statistical analysis of the results did not show significant differences regarding the number of glial cells nor glial membrane volume between both genotypes in adult brains ( Figure 1H, I). Therefore, we did not detect significant differences upon the co-expression of H3K27+MYC in adult glial cells in the number of glial cells nor to the volume of the glial membrane.

Co-expression of H3K27+MYC in glial cells does not cause premature death.
To determine the life span of this genetic  Figure 1J).
Therefore, the co-expression of H3K27+MYC does not synergize and does not cause premature death in adult flies.

The JNK pathway is upregulated in various tumors, including GB. Recent studies in
Drosophila show that activation of JNK through Grnd receptor is necessary for GB progression . In consequence, we analyzed if JNK activation is associated with tumor progression in this model of DIPG.
To determine if the JNK pathway is activated in DIPG cells, we used the previously validated Tre-RFP reporter, whose transcription is specifically activated in response to JNK signaling . The analysis and quantification of Tre-RFP positive cells in the brain did not show significant differences between the Control and H3K27+MYC samples (Figure 2A-C). These results suggest that the co-expression of H3K27 and MYC in glial cells does not increase JNK signaling, unlike what occurs in GB cells.
Erg or Grnd deletion experiments in Drosophila show that JNK inhibition rescues tumor proliferation and invasiveness in GB models (Portela et al. 2020). To confirm that JNK signaling is not involved in the progression of DIPG, we reduced JNK signaling using flies that express the dominant negative form of Grnd (UAS-grndMINOS) (Portela et al. 2020). We analyzed fundamental aspects of the DIPG including the number of glial cells and the volume of the glial membrane ( Figure 2D-G). The results did not show significant differences in the number of glial cells or the volume of glial membrane between H3K27+MYC and H3K27+MYC+grndMINOS. In addition, we observed significant differences in both variables between the control and H3K27+MYC genotypes; and control and H3K27+MYC+grndMINOS ( Figure 2H-I). Therefore, the blockage of JNK signaling does not modify these fundamental aspects of DIPG cells biology, supporting that, unlike in GB, JNK signaling in glial cells is not required in DIPG tumor progression.
Likewise, we did a larval viability assay to further determine the contribution of JNK pathway. We compared the viability of control genotypes, H3K27+MYC, grndMINOS and H3K27+MYC+grndMINOS. The quantification of the percentage of individuals that reached the adult stage showed significant differences between the control and H3K27+MYC genotypes, and between control and H3K27+MYC+grndMINOS.
However, we did not find significant differences between H3K27+MYC and H3K27+MYC+grndMINOS ( Figure 2J). Thus, the inhibition of JNK signaling in glial cells does not rescue the reduction of life span caused by DIPG progression, concluding that JNK signaling is not determinant for the progression of the DIPG.

DIPG cells increase MMP protein accumulation.
A fundamental aspect of DIPG is its ability to infiltrate the brain, a feature in common with GB. Previous studies revealed that GB cells produce MMPs that mediate the degradation of the ECM, thus favoring the infiltration and progression of the tumor (Malemud 2006;Nakada et al. 2003;Portela & Casas-Tintó 2020;Portela et al. 2020).
To determine if MMPs proteins increase in DIPG model cells, we stained larval brain samples for MMP1 and MMP2 of the following genotypes: control (LacZ) and ( Figure 3C) was higher than in the case of MMP2 ( Figure 3G).
Finally, we determined the relative concentration of MMPs with respect to DIPG cells membrane. To visualize DIPG cells membrane we used the fluorescence of the myrRFP ( Figure 3A'-F'). The graphical representation of MMP and RFP intensity in the ventral ganglion of H3K27+MYC larval brains showed MMP1 and MMP2 peaks corresponding to a displaced region with respect to the DIPG cell membrane ( Figure   3D-H). These results suggest that MMPs are accumulated in the immediate extracellular region of DIPG cells as occurs in GB cells (Conte et al. 2021).

DIPG cells accumulate synaptic proteins as GB cells.
Drosophila models of GB have proved the accumulation of synaptic proteins, including

ImpL2 is produced by GB cells and mediates the communication from GB cells to
neurons. ImpL2 is an antagonist of the insulin pathway produced by GB cells. It attenuates insulin signaling in neighboring neurons which promotes mitochondrial dysfunction and synapse loss in neurons. In consequence, neurons undergo degeneration and GB progress more effectively (Jarabo et al. 2021).
To assess if DIPG cells in Drosophila reproduces this behavior, we used a validated MIMIC GFP reporter that replicates ImpL2 expression (Nagarkar-Jaiswal et al. 2015), in control and H3K27+Myc expressing brains ( Figure 4D-E´´). The quantification of GFP signal in DIPG brains show a reduction that suggest a downregulation of ImpL2 in DIPG cells. (Figure 4F), opposed to the results described for GB cells.

DISCUSSION
DIPG is a type of deadly pediatric glioma that currently lacks effective treatments. It has a peak of incidence in middle childhood and a life expectancy of less than 1 year (Lapin et al. 2017;Wu et al. 2014). Therefore, a better understanding of the molecular mechanisms involved in the biology of DIPG is necessary for the development of new therapeutic strategies.
Therefore, the generation of a robust experimental model capable of reproducing the fundamental features of the disease. The most relevant characteristic of this tumor is the presence of mutations in histone H3, which is occurs in 80% of cases and is associated with worse outcomes. However, early activation of the H3K27M pathway is not sufficient to produce tumorigenesis, and one of the characteristic features of H3K27M tumors is epigenetic activation of the MYC pathway (Lapin et al. 2017).
Therefore, the experimental model used in this work is based on the expression in glial cells of the H3K27M+MYC combination.
One of the fundamental aspects of the disease is glial expansion, for this reason we On the other hand, the dispersion curves obtained with the pixel intensity averages of MMP1, MMP2 and RFP in the ventral ganglion of the brain of H3K27MYC larvae, show a displacement of the peak of the MMPs, both MMP1 and MMP2, with respect to that of RFP ( Figure 7D, 7H  Zuccarini et al. 2018). In addition, this route seems to have an essential role in neurological tumors such as medulloblastoma, the most common primary tumor of the CNS in children, and subependymal giant cell astrocytoma (Jozwiak et al. 2007 DIPG is a very complex, lethal, and incurable disease, which has a very limited number of animal models. Therefore, it is necessary to generate a robust model to dig into the knowledge of the DIPG. The results of this work suggest that the co-expression of H3K27+MYC in Drosophila glial cells is a valid model for the study of DIPG in vivo, since it has been proven to reproduce characteristics of the tumor, including glial invasion, tissue specificity and pediatric nature. The study of the underlying molecular mechanisms shows that, as indicated in the literature, DIPG is a unique entity that shares some similarities with GB, but which requires specific therapeutic approaches, for which it is it is necessary to expand the knowledge of the biology of the DIPG.